深网 deepoo.net

从此走进深度人生 Deep net, deep life.

标签： 历史

  • 谭其骧：首都变迁的原因

    一、中原期与东移近海期

    总述上述七大首都（长安、洛阳、邺、开封、杭州、南京、北京）的兴替过程，可以看到，中国的建都史大致可分为前后两期。从殷周直到北宋这二千四百年是为前期，其时一统政权和统治北半个中国的大地区性政权的首都殷（邺）、长安、洛阳、开封，都在中原地区（北纬35°左右1度许，东经108°—114°）；江南的南京只做过统治南半个中国的地区性政权的都城，而位于华北平原北端的北京，则根本还够不上做较大政权的都城。所以这前期又可以叫做中原期。自十二世纪初叶赵宋南渡以后至今八百多年是为后期，一统政权和大地区性政权的首都都离开了中原：或向南移到了江南，杭州做了一百五十年的南宋都城，南京做了五十年的明朝初期首都，又做了此后二百二十年的陪都，直到近代还做过太平天国和民国的首都；或向北移到了北京，先还只是北半个中国金朝的首都，随后又发展成为元、明、清三代的大一统王朝的首都，直到近代还做过民国的首都，今天仍然是我们中华人民共和国的首都。杭州、南京、北京都在前期四大首都之东，距海不远，所以这后期又可以叫做东移近海期。

    为什么前期的大政权要选择中原内地的长安、洛阳、邺、开封为首都，后期的大政权要选择东部近海的杭州、南京、北京为首都？又为什么前期和后期在各个时代要选择不同的城市为首都？这需要我们对历史上择都的条件和首都在历史上所发生的作用作一番分析。

    二、七大古都的历史地位

    历代统治者主要是根据经济、军事、地理位置这三方面的条件来考虑，决定建立他们的统治中心——首都的。经济条件要求都城附近是一片富饶的地区，足以在较大程度上解决统治集团的物质需要，无需或只需少量仰给于远处。军事条件要求都城所在地区既便于制内，即镇压国境以内的叛乱，又利于御外，即抗拒境外敌人的入侵。地理位置要求都城大致位于王朝全境的中心地区，距离全国各地都不太远，道里略均，便于都城与各地区之间的联系，包括政令的传达、物资的运输和人员的来往。设若地理位置并不居中，但具有便利而通畅的交通路线通向四方，特别是重要的经济中心和军事要地，则不居中也就等于居中。所以地理位置这个条件也可以说成是交通运输条件。当然历史上任何时候都并不存在完全符合理想、三方面条件都十分优越的首都，所以每一个王朝的宅都，只能是根据当时的主要矛盾，选择比较而言最有利的地点。首都的选定一般都反映了该时期总的形势，反过来，首都的位置也对此后历史的发展产生一定的影响。

    明白了这个道理，那就不难理解历代首都的迁移，是历史发展的必然结果。

    先谈一谈从中原内地移向东部近海这个历史上前后期的大变动问题。这很简单。自殷周至隋唐，黄河中下游两岸是全国经济最发达的地区，又接近于王朝版图的地理中心，一个政权若能牢固掌握这一片地区，就尤足以控制全国，这就是这一段长达2400年之久的时期的首都离不开中原地区的原因。由于首都在中原，所以当时开凿的运河也都指向中原。五代北宋200年间，经济重心虽已南移江淮，但中原还是可以通过水运通向四方，所以首都仍然能够留在这个水运系统的枢纽地——开封。北宋覆亡以后，出现了南北分裂的局面，于是中原水运又因停止使用而归于淤废，从此以后，无论从经济、军事、交通哪一方面说，中原都处于不利的地位，这就是800年来首都再也不可能迁回到中原之故。

    再让我们逐一阐述一下七大首都何以先后被选为首都。

    中原四大首都中长安的条件最优，所以它作为首都的时间最长，以此为首都的周、秦、西汉、隋、唐也是历史上最兴旺的王朝。长安的条件优在哪里呢？汉高祖即位时都雒阳，听了娄敬、张良的话才西都关中，这两人的话很说明问题。

    娄敬说：“秦地被山带河，四塞以为固，卒然有急，百万之众可具也。因秦之故，资甚美膏腴之地，此所谓天府者也。陛下入关而都之，山东虽乱，秦之故地可全而有也。夫与人斗，不搤其亢，拊其背，未能全其胜也。今陛下入关而都，案秦之故地，此亦搤天下之亢而拊其背也。”

    张良说：“关中左崤函，右陇蜀，沃野千里，南有巴蜀之饶，北有胡苑之利，阻三面而守，独以一面东制诸侯。诸侯安定，河渭漕挽天下，西给京师；诸侯有变，顺流而下，足以委输，此所谓金城千里，天府之国也。”

    秦地，指崤山、函谷关以西战国秦国故地。关中，有广狭二义，广义等于秦地，狭义专指关中盆地，即八百里秦川。秦地对山东六国故地而言地居上游，关中盆地四面有山河（东崤、函、黄河，西陇山，南秦岭，北渭北山地）之固，所以建都关中，凭山河之固则退可以守，据上游之胜则进可以攻，对叛乱势力能“搤其亢”而“拊其背”，在军事上地位十分优越，是之谓“金城”。关中盆地“沃野千里”，是一片“甚美膏腴之地”，又可以取给于南方的巴蜀和北方的胡苑（胡人的牧区）以补不足。若山东诸侯有变，关中的物资足以供应顺流而下的王师，在经济上也有所恃而无恐，是之谓“天府”。关中在当时是这样一个金城天府之国，所以汉高祖便作出了在它的中心地带丰镐、秦咸阳的附近建立作为王朝首都的长安城的决定。

    历史证明这一决定是完全正确的。娄敬、张良抓住了当时初建的汉王朝内部最突出的问题，即中央与山东诸侯之间、统一与分裂势力之间的矛盾问题，他们之所以主张建都关中，主要着眼于都关中足以东制诸侯。此后自高祖至文、景，果然先后顺利地镇压住了多次异姓、同姓诸侯的叛乱，巩固了统一。他们还没有能够预计到日后形势的发展。武帝以后，汉与匈奴之间的矛盾代替了王朝中央与诸侯之间的矛盾，成为当时的主要矛盾，汉朝经过武、昭、宣三代的经营，终于取得了匈奴降服、置西域数十国于都护统辖之下的伟大胜利，这和建都长安便于经营西北这一因素也是分不开的。所以建都长安，确是既有利于制内，又有利于御外。

    隋唐时形势略与西汉相似，关中仍然以沃野著称，对内需要能制服山东和东南潜在的割据势力，对外需要能抵御西北方的强大边疆民族政权突厥与吐蕃的入侵，因而也和西汉一样定都于长安。

    但是，长安作为首都也有不利的一面。它的地理位置比较偏西，距离当时人口最稠密、经济最发达的黄河下游两岸远了一些，距离中唐以后财赋所出的江淮地区那就更远。关中尽管富饶，毕竟“土地狭”，不足以满足京师和西北边防所需大量饷给。西汉时问题虽已很显著，还不很严重，因为关中的不足主要仰给于山东，山东距关中还不算太远。到了隋唐，特别是中唐以后，两河藩镇割据，京师所需百物绝大部分都取之于数千里外的江淮地区，节级转运，劳费惊人，民间至传言“斗钱运斗米”，这一矛盾就越来越尖锐。勉强维持到唐末，终于通过朱全忠强迫昭宗迁都，结束了长安作为首都的历史。五代以后，黄河流域益形衰落，江南的经济地位和河朔的军事地位逐步上升，中原王朝内部便不再是东西对峙的问题，变成了南北争胜之局；主要的外患也不再来自西北，改为来自东北的契丹、女真和蒙古，从而长安又丧失了它在军事上的制内御外作用，所以首都一经撤离，就再也不可能搬回来了。

    洛阳在军事、经济两方面条件都比长安差。伊洛之间虽然也有一片平原，可是远不及关中平原的肥沃广袤；四周也有关河之固——东据成皋，西阻崤、渑，背倚大河，面向伊、洛，但诚如张良所说：“虽有此固，其中小，不过数百里，田地薄，四面受敌，此非用武之国也。”东汉都雒阳，所幸光武完成统一后王朝内部并不存在割据势力，故都洛百数十年得平安无事。但至末年董卓擅行废立，关东州郡起兵讨卓，以当时董卓之强，也就不得不离开这个“四面受敌”之地，西迁长安。

    东汉一代无论对内对外，武功都远不及西汉。特别是对西北边境，大有鞭长莫及之势。西域三绝三通，合计设有都护、长史的时间不过二十余年。安帝后历次羌乱，兵连师老，费用至数百亿，并、凉为之虚耗，三辅亦遭残破。当然，东汉国力之不竞是由多种原因造成的，但首都建在远离边境的雒阳，以致对经营边境有所忽略，不能不是原因之一。

    洛阳的优点主要在于它位居古代的“天下之中”。远在西周初年，周公所以要在这里营建成周雒邑，作为镇抚“东土”的大本营，就是因为它“在于土中”，“诸侯四方纳贡职，道里均矣”。西周为犬戎所破，平王东迁，即于此宅都。后来项羽烧了咸阳，汉高祖初即帝位时也曾都此数月，等到赤眉烧了长安，光武即定都于此。洛阳虽然比不上长安那样是“金城天府之国”中的首都，但它有了这一条为长安所不及，它的不大的四塞之固又为邺与开封所无，所以它在前期中原四大首都中的地位仅次于长安。曹丕舍弃了乃父曹操经营了十多年的邺都而迁都董卓劫迁献帝以来荒芜了30年的洛阳，北魏孝文帝自平城南迁，一度想都邺，而终于定都永嘉乱后荒废达180年之久的洛阳，足见曹丕和拓跋宏都认为都洛胜于都邺，他们考虑问题的着眼点显然是地理位置。邺地处河北，在中原范围内稍东稍北，曹魏为了对付西南的蜀汉和东南的孙吴，拓跋魏企图并吞南朝，混一诸夏，都洛当然比都邺合适。

    隋唐建都长安，隋炀帝、唐高宗都要另建洛阳为东都，经常来往于两都间。炀帝以居洛为常，洛阳是实际上的首都。高宗晚年亦多居洛，其后武周代唐，改东都为神都，正式定为首都。可见隋唐时代洛阳还有比长安更优越的一面，否则杨广、李治、武曌不会作出那样的决定。这不仅是因为它的地理位置在全国范围内比长安来得适中，更重要的在于它是当时的水运枢纽，东南取道通济渠、邗沟、江南运河，可通向富饶的江淮地区，东北取道永济渠可通向河北大平原，直抵王朝东北部的军事重镇涿郡即幽州（今北京），特别是江淮漕运自通济渠东来可以径抵洛阳城中输入含嘉仓，比之于都长安时需从洛阳或洛口再或水或陆，多走上千里路程才能到达目的地，省事省费实不可胜计。隋唐时代皇帝之所以屡次要东幸或移都洛阳，实际就是为了要解决皇室、百官和卫士等的给养问题。武则天死后中宗虽西还长安，不久玄宗开元初年起又屡次因关中岁歉而东幸洛阳。玄宗是颇厌惮往来的劳累的，但又不得不如此。直到开元二十二年裴耀卿改进了漕运办法，每岁可运二百数十万石至长安；二十五年牛仙客献计在关中用岁稔增价和籴之法，史称“自是关中蓄积羡溢，车驾不复幸东都矣”。长安的首都地位才得稳定下来，不至于为洛阳所夺。

    邺处于古代“山东”（一般指黄河流域东部大河南北、太行山东西）地区的中心，背靠山西高原，东南北三面是古代经济最发达的黄淮海大平原，所以它在军事上是无险可守的（曹操在邺城西北隅因城为基，筑铜雀等三台，这是人造的防御工事，当然比不上天然的山河之固），不及长安，也不及洛阳；在地理位置上不如洛阳那么适中。但以经济条件而言，则在长安、洛阳之上，凡是控制山东地区而不能奄有整个黄河流域的政权，一般都要宅都于此。商人七次迁都，自都殷（邺的前身）后凡273年竟不复迁。曹操情愿离开他经营多年的兖州和许，定都于邺；后来虽然统一了黄河流域，仍都此不迁，直到儿子曹丕手里才迁都洛阳。十六国时后赵、前燕，北魏分裂后的东魏、北齐都据有山东之地，也都定都于此。北魏明元帝神瑞二年因比岁霜旱，平城附近民多饥死，朝议欲迁都邺，以崔浩谏不宜动摇根本，乃分简尤贫者，使就食山东，而罢迁都之议。其后孝文帝南迁经邺，崔光清即建议定都于此，理由是：“邺城平原千里，漕运四通，有西门、史起旧迹，可以饶富。”孝文则认为“石虎倾于前，慕容灭于后，国富主奢，暴成速败”，不从。其实孝文这几句道貌岸然的话未必是他的真意，他之所以执意要都洛而不都邺，目的端在都洛便于南伐。但这几句话却充分反映了那个时期邺都经济条件的优越。

    自中唐以后国家财赋愈益依赖江淮漕运，所以五代北宋时，居水运枢纽的开封遂代替安阳（邺）、长安、洛阳，成为择都的首选。

    后期金、元、明、清之所以要选中北京定都，那是由于这几个政权都需要兼顾塞外与中原，而大运河漕运又足以解决都燕的供给。明初之所以都南京，那是由于元末明太祖以此为根据地经营四方完成一统的已成之势，并且正好就近控制东南财赋之地之故。至于南宋有半壁江山，不都南京而都杭州，上文已提到，除了由于自五代以来杭州在东南城市中最为繁盛这一因素外，主要是宋高宗绝意恢复中原的心理在起作用。

    《谭其骧历史地理十讲》（葛剑雄 孟刚选编）

  • 谭同学：民族走廊中的隙地开发与人群互动——以平川瑶为中心的讨论

    一、引言

    无论从地理形态还是社会文化上看，中国都是融多样性为一体的大国。依地理形态而言，施坚雅认为可分出长江上游、长江中游、长江下游、东南沿海、岭南、云贵、华北与西北等巨型区域。①冀朝鼎则综合地理、水利、政治、经济等因素，从“基本经济区”②理解中国历史。二者虽然不乏区别，但在方法上都有“从地方动力去理解国家历史”③的特点。区域“本身也是一个社会历史过程”，其“界临地区往往自成一个区域”。④而且，区域界限并不绝对，往往因为政治、经济和社会互动，具有变动的可能性。⑤

    区域之间有“界”，以绵延的山脉最为常见。“作为整体的山地，一般处于一些较大区域的边缘，构成区域的自然边界……高大广袤的山地对于区域边界的划分有着特别重要的意义，它对文化传播的阻隔作用远远大于长江大河”。⑥这些地域不仅地理上处于区域边缘，且因交通不便，常是国家统治薄弱的边缘。其中的人群还常有刻意“自我边缘化”，强化“蛮”的倾向，⑦以求不承担或少承担赋役。⑧在此意义上，从国家治理角度看区域间的边界地带，也有“空隙”的性质。对此，许倬云有较系统的论述：王朝国家体系“其最终的网络，将是细密而坚实的结构。然而在发展过程中，纲目之间，必有体系所不及的空隙。这些空隙事实上是内在的边陲。在道路体系中，这些不及的空间有斜径小道，超越大路支线，连紧各处的空隙。在经济体系中，这是正规交换行为之外的交易。在社会体系中，这是摈于社会结构之外的游离社群。在政治体系中，这是政治权力所不及的‘化外’，在思想体系中，这是正统之外的‘异端’”。⑨

    在借鉴许倬云论述的基础上，鲁西奇主张称此类区域间的空隙地带为“隙地”，并视其为“内地的边缘”。⑩进而，他将“隙地”的特征总结为：国家权力相对缺失；国家政治控制方式多元化；可耕地资源相对匮乏，经济形态多样化；人口来源复杂多样，很多属于“边缘人群”；社会关系网络多凭借武力，或以利相聚，或以义相结，或以血缘、地缘相类，具有强烈的“边缘性”；文化多元，异于正统意识形态的原始巫术、异端信仰与民间秘密宗教流行。11赵世瑜则认为，这种非均质化“地理缝隙”的一个重要标志是，在“编户齐民”之外，需要“代理人”治理。12此外，吴重庆还指出，隙地作为一种分析视角，也有助于理解近代革命根据地建设，以及当代农村人口“空心化”反向流动等现象。13

    从隙地看中国，无论在历史上还是在现实性上，都不失其价值。不过，作为区域间界限的隙地虽有其边缘性，却不绝对封闭。相反，在某些条件下，它们可以成为人们跨区域流动的“走廊”。历史上许多民族都有跨区域，甚至跨越多个区域迁徙的经历。为此，费孝通曾用“民族走廊”的概念，来指不同民族长期沿一定的自然环境（如河谷或山脉）迁徙，交往、交流、交融而又保持社会文化多样化的格局。14他还提议深入研究南岭、藏彝、西北三大民族走廊，以更好地理解中华民族“在历史上是怎样运动的”。15从宏观上看，民族走廊在宏观上或多或少有隙地的特征。若再往细处看，其内部往往在地理形态、生态条件、生计方式和社会文化等方面也具有多样性。因此，在民族走廊多样化的区块之间，会有一系列小尺度的隙地。

    其实，中国很多区域都有过多种民族迁徙、互动的历史。缘何民族走廊中的少数民族社会文化多样性会格外突出，或者说民族走廊究竟是如何形成的？指出其多样性本身，虽然对经验提炼有重要洞见，但更重要的是理清形成这种结果的过程和机制。从这个角度看，其人群自我边缘化以（部分）回避赋役的因素固然不可忽视，却难以解释为何他们在赋役无实质差别，甚至深受儒家“礼”仪浸淫的情况下，依然坚守少数民族认同。因此，宏观上具有大尺度隙地特征，内部又包含大量小尺度隙地的民族走廊，在形成、运转的机制层面，仍有值得进一步细究的地方。对这一问题的探索，在理解民族认同、民族关系的历史，以及民族走廊发展的现实思考上，均有价值。以下笔者将以对南岭民族走廊西端南侧桂林市恭城瑶族自治县平川河峡谷“平川瑶”的调查为基础，16结合相关文献，尝试探讨该问题。

    恭城县北部栗木镇、观音乡与桂林市灌阳县（水陆交替可达湘江），东北部与湖南省永州市江永县（古称永明），南部与桂林市平乐县、贺州市富川县接壤。平川河发端于观音乡与江永县交界的高山，向东南沿海拔800米—1300米左右高山所夹峡谷平川源（河谷海拔250米—350米），流经水滨、狮塘、蕉山、洋石、杨梅，在观音村的岩口寨出峡谷，再约2公里进入栗木镇地界，在该镇上宅村北侧汇入栗木河。栗木河往南约15公里，即东西向连接恭城、江永两县的恭城河，恭城河往南在平乐县汇入桂江。平川河无法通航甚至放排，从上游水滨村牛眼塘寨经山路到最近的集市栗木圩约35公里（1970年始有机耕路，1988年方通车）。河谷少量耕地可种单季水稻，接近河谷的坡地可种玉米、红薯、土豆，山地除了原生杂木，可种杉树、桐树、油茶树。

    二、隙地开发正当性终源于国家正统

    20世纪70年代，平川源曾发掘出一个陶罐，内有五十余枚古钱币，“开元通宝”居多，另有部分“宋元通宝”“大定通宝”。所有古钱都是发行量较大、流通实用型的，且都不晚于宋、金。蕉山村存有一个五足双耳石香炉，刻着龙凤、舞狮、麒麟、宝相花、龙犬等纹样（被考古人员断为唐代风格石雕）。17由此可知，明代之前平川源应已有一定数量的居民。

    明初，恭城县东部与湖南永明县交界地带发生叛乱，波及桂东北、湘西南，朝廷从桂西河池调兵镇剿。光绪《恭城县志》记道：

    明洪武初，势江源贼目梁朝天，湖南贼首雷虎子、马公三等纠党，由八角岩谋叛，攻破县城，杀戮官吏，时全州、永明二官俱被害。有莫祥才者，山东人也，统带庆远府之河池州宜山县、南丹州等处黄、韦、陈、周、石、唐、欧、赖、莫、贲、谭、覃、徐、祝、陆、廖、雷、马、梁、蒙、容、李、罗等二十三姓之药弩手三百、民壮五百，将贼剿平，克复城池，即以功授莫祥才白面寨巡检司，其弩手、民壮均给照，赐地方、租税，俾子孙永享焉。18

    县志未提及瑶兵。但是，1984年恭城县西岭乡新合村出土了一块题为《猺目万历二年石碑古记》的碑刻（以下简称《猺目碑记》），详细提到了瑶兵。19其碑文道：

    申告恳赏给照，七姓良猺赵中金、邓金通、赵进珠、邓启音、郑元安、盘金童。七姓猺目乃系广（东）德庆州肇庆府铁莲山风（封）川县，入广西恭城县到平源。雷伍（虎）子反，所有招主黄□□、黄明、李富山闻之广东有好良猺，即行招德（得）大朝兵马，之因洪武下山，景太（泰）元年闰三月初三日进平源，剿杀强首雷通天、李通地，贼首退散。给赏良猺，把手（守）山隘，开垦山场，安居乐土。恳给立至守把隘口，又到嘉靖□十七年七月十一日，被东乡贼脚阴家洞，抢得万名（民）不安。本县提调猺名邓贵明、郑海成、赵进旺，□（统）带猺丁拿得生工七名李，□□同解。本县赏给白银五十两，给猺目回源，守真山源隘口地方。后至万历十五年三月十八日，贼首越过苏被口並沙江，立剿（扰）万名（民）不安。本县提调猺名郑进旺、郑德元、赵殊禄，捅（统）带猺丁拿得生工名十，解报本县，即时打死。赏给白艮（银）七十两，给猺目回家，用心固守地方，至万历二十年。守把隘口地方，奉公守法，照越过地方，屡蒙恩赏。但良猺把守隘口地方，山场四至界内土名：赵中金把手（守）到平源，郑元安把守瓮塘源……五猺隘口山场与猺目，永远耕种、管业，开垦先立升科报税，不於（予）另招别猺影（侵）占猺源地界。    当夫上巡马脚不遗被猺，远任前公擅冷（令）后代子孙永远当差科派，那时有无凭只（证）德（得）报恩开垦，攻（功）劳实与朝。报□（万）历祠前，赴本县父台前，伏乞申详上司道府各处衙门计政存案，恳给印照付，猺目各收为据：子孙永远世代沾恩。详给施土司恩泽，历靖申告本县照验，准给申告准凭。    景泰元年闰三月初一进倒不（平）源

    洪武下山、万历二年八月十八日恳给印照20

    此碑错讹甚多。其中，“广东”缺“东”字，“银”错为“艮”，“侵”错为“影”，“平”错为“不”，因字形相近，疑为笔误；“风”（封）、“太”（泰）、“手”（守）、“名”（民）、“剿”（扰）、“於”（予）、“德”（得），字形差别较大，疑为汉语方言恭城话谐音别字；“只”（证）、“伍”（虎），疑为过山瑶勉语口音别字。碑文口吻、立场皆为“良猺”，新合村至今为过山瑶聚居村庄。综合看，撰写碑文者可能是文化程度不高的过山瑶。过山瑶中当至少有部分源于封川县（今封开县）铁莲山或附近山区，否则难以说出细致地名。口述者未必识字，只会发音“封川”，后来撰碑文、刻字者之文化程度恐不够知晓数百公里外的准确县名，而以为是“风川”。

    碑文无确切立碑时间信息，但内容表述为明万历二十年（1592年）之后一段时间，地方官不再强调甚至不再承认以前官方曾准许“良猺”世代享有土地及赋役优惠，以至后来“良猺”再次伸张自己的“权利”。其中疑点颇多。

    其一，若从广东封川县招瑶兵，水路距离约为河池两倍，陆路翻山越岭亦不比河池近，动静不可谓不大。且不说恭城“招主”难以获知封川“良猺”信息，至少志书不至于单记河池兵（详至弩兵23姓），而不记瑶兵（连《猺目碑记》所记赵、邓、郑、盘等常见“良猺”姓氏，都无一被提及）。明万历二十五年（1597年）恭城即首修县志，光绪版县志已是第四版21（前三版已散佚），记有其他几次剿“反”“贼”。前三版如有瑶兵记录，光绪版不应独删此记。

    其二，若“良猺”是明洪武年间，哪怕是洪武最后一年（1398年）下山，却到景泰元年（1450年）才“进平源，剿杀强首雷通天、李通地”（雷、李之名也像是俚语外号），中间隔了五十多年，耗时未免太长。

    其三，在恭城话中，“进平源”意为进入平川源，但碑文“入广西恭城县到平源”，“把手（守）到平源”，“进倒不（平）源”中所提“到／倒平源”（源自西南官话方言恭城话口语，无从判断“到”或“倒”哪种写法准确），却只表示临近平川源峡谷口的平地。

    不管真是官方通过查阅档案确认很久之前曾授予“良猺”“恩泽”，还是讨价还价之后妥协，结果是认可其占有5个“猺隘口山场”（含平川源隘口），“永远耕种、管业”，不允许另外再招其他“猺”来占用。而“良猺”也接受了“开垦先立升科报税”，只是不用“当差”。

    《恭城县志》记载，“雷虎子”事发明初，针对的是官府，故用词为“反”“叛”。《猺目碑记》所述时间却是明嘉靖、万历年间，“贼脚”“贼首”亦未针对官府，而是“抢”“民”，甚至只是“越过”被“良猺”认定属于自己“永远耕种、管业”的地界。“良猺”乃至官府视其为“贼”，但实属新流入当地的人群。当其土地开发范围跨过“猺源”隘口，进入河谷乃至峡谷口外平地时，与“良猺”发生了冲突。“良猺”作为胜利者，将这些冲突附会于五十年甚至更长时间之前镇剿“雷虎子”的历史，运用为国立功的叙事，证明其占有土地和免征差役的正当性。

    无独有偶，平川源的瑶民述及迁徙史，也说是明初“来恭城打雷虎子”（源流地则五花八门）。曾任水滨大队副大队长、水滨村村委会副主任的蒋礼发存有一本破损、散乱的手抄本《上五排历史》22（“排”是明嘉靖九年［1530年］至清宣统元年［1909年］官府在部分瑶山设置的村级管理单位，小村则数村为一排）。其中一篇《平川上五排嘉靖九年照碑记》（以下简称《嘉靖碑记》，碑已毁，但村中有几位老人表示民国时期见过）记道：

    计嘉靖九年（1530年）正月十五日给蒋政聪、周贵清、周福珠、俸仁聪等，各告称：祖公在于平川源上下二涧居住，洪武廿五年（1392年）被永明县雷午（虎）子越来作恶，洪武廿六年告军征剿，蒙上司行榜，仰本县责令本里故民欧（阳）用诚、周福谦招抚周庆陆、俸富三下山向化圣朝。23

    这里所说“上下二涧”，涉及明嘉靖九年实施的排瑶制。它以平川源及峡谷口10个大寨为中心，设10个排。下涧指的是下五排，包括老洼（今观音）、洋石、杨梅、井头、白藤底（今大坑底）诸寨。上涧指的是上五排，包括蕉山、狮塘、水滨、古骨圩（含矮寨）、大畔源诸寨（清乾隆二十七年［1762年］，第一排大畔源寨划归湖南永明县后，将较晚成村的狮尾、黄茅岭［今莲花］、石坪寨设为第一排）。其中，“雷虎子”写为“雷午子”，亦为过山瑶勉语口音所留痕迹（今水滨村只有牛眼塘寨1位老人还会说过山瑶勉语），所记“雷虎子”被征剿时间（明洪武二十六年［1393年］），与光绪《恭城县志》所记“洪武初”相比，有显著出入。此说附会色彩十分鲜明。

    不过，《嘉靖碑记》所载另一事多有印照。碑文记道：

    具记永乐三年（1405年）造册附籍，纳粮四石九斗三升，住种杀功解报，守护地方，至今一百七十余年，并无为非生祸。因被嘉靖六年（1527年）成江附籍良猺周良通等，（将）田地与獞人常金朝、常金龙、龙汝鉴占种。嘉靖七年三月十七日又被周镛、欧阳爵、卢姗等放傲，将本源盗卖王铭等，聚兵杀占、攻破山寨，杀死男妇一千余命，赶散良猺（往）湖广永明地方避住。（周贵）清等将情具告，蒙道行提周镛等，责令协同委官并县哨入源晓谕。军门杀伐利害，抚退王铭。回巢（源）照旧招佃，周贵清等复业本源住种。24

    平川源峡谷口外栗木镇上宅村的《周氏大宗族谱》对此事记道：

    嘉靖七年戊子（1528年），平川源被（恭城北乡栗木）大合（村）招主欧阳爵、本族地主周镛，受银三百两，（将）平（川）源田地尽数卖（恭城东乡）东寨贼（王）铭类，占夺平（川）源，杀死大小男妇一千余命。田地主（周）福谦、周祚、周郁、周郡通族等用呈具告回民瑶兵，备调发监三十四俍兵，四方普洗本乡三寨；胡北洗平三寨，胡伯抽巢，乡境得宁。25

    两则记载略有差异：其一，《嘉靖碑记》提到明嘉靖六年（1527年）就已有过“附籍良猺”将田地租给“獞人”耕种，次年才发生“良猺”土地“尽数”被“盗卖”和被驱赶、杀戮；其二，周氏族谱所记，大合村“招主欧阳爵”和“本族地主周镛”卖土地，属公卖而非“盗卖”。

    类似的事接二连三发生，说明当时有土地的一方，不管是汉人“招主”还是“附籍良猺”地主，将原本租给“良猺”的土地，收回佃权，改租或卖给新来的“猺人”或“獞人”，已非鲜例。新来的“獞人”未经过“良猺”村寨集体同意，从地主个人手中租、买土地之后，即自行耕种（被认作“占种”）。新来的“猺”“贼”则除了自行耕种，还要向原租种的“良猺”再收一道租，以至引发流血冲突。官方提审卖主，军队介入，但最后只是“抚退”而非剿灭“贼”。这更说明，问题实质是争夺土地经营权。周氏族谱既称王铭为“贼”，并记其占平川源、杀人之事，却不提“盗卖”，或为祖先讳。

    在当时的土地开发过程中，“良猺”可能确实贡献不小，且是以组织化的群体形式存在，以至于与土地所有者达成了默契，有集体性的优先耕种权。《嘉靖碑记》提及明永乐三年（1405年）纳粮的标准，或为暗示“良猺”耕种这些土地，原本赋税、租金比较低，因此夺佃、加租都不可接受。该碑记在后文中还提到，事件平息后上、下五排只需各“纳粮税”“六担”，由周、欧阳两姓代收，26此亦证明“良猺”为“附籍”。

    三、土地承载弹性空间及其自我维系

    经明嘉靖年间变故后，平川源“良猺”获得了官方认可的平川源土地经营权，以及相当一部分土地所有权（这可算是官方对欧阳、周氏等山主的惩罚，以此补偿受损的平川源“良猺”）。但是，平川源人口损失不少，而已开垦出来的土地得有适当数量的劳动力耕种，才有经济收益。于是，已有一定山主地位的平川源“良猺”，向官府申请并获得准许，可以村寨集体为单位，主动招徕其他缺少土地，甚至还处于流动状态的“猺”，从深山下到河谷或临近河谷的坡地进行耕种。对此，《嘉靖碑记》载道：

    （明嘉靖）九年（1530年）正月二十五日立赏蒋庆才、庆广招板瑶赵广富。正月二十七招二十五家。李朝聪招板猺赵老担，何涧清招板猺赵广聪，李庆惠招板猺盘大三……嘉靖九年，蒋政威（招）廿五家，田户开在赵广聪名下，蒋世姗招廿五户，开在赵保仔名下。27

    板瑶属于过山瑶的一个支系（但与此前流入平川源“附籍”的过山瑶，显然不属于同一群体），据说因“以头盖夹板而名”，源自广东北部。28但是，仅上五排一年之内就能招到板瑶上百家，甚至在正月3天就招徕到三十余家。由此推测，原本就在平川源及其周边深山游耕、游猎的板瑶，数量必定不少。否则，恐难短时间内有这么多人能够召之即来。依费孝通于1935年所做调查，桂东北大瑶山区的瑶民有控制人口的习惯，一般一对夫妇抚育2个孩子29（部分家庭或有老人，估算平均每家5口左右）。以此为参照粗略推算，该年上五排招徕板瑶即可能达到五百人以上。若下五排情形亦相似，则整个平川源招徕板瑶约一千人。这个数字大致接近此前平川源在冲突中损失的“一千余命”。若这种招徕行动，并不能将周边深山中带有一定流动性的人口悉数全引下山，则说明原本在深山中靠游耕、游猎生存的人口可能远超过千人。平川源及其周边山地能承载的人口有相当的弹性空间，由此可见一斑。

    平川瑶招主得在自己名下给招徕的板瑶开“田户”，意味着这些板瑶主要不是在深山中耕种林间旱地，而是在河谷种田，或在接近河谷的坡地进行开垦。虽然板瑶与平川瑶在语言、服饰、生活习惯上不同，但仅从土地耕作的角度来说，并不必然构成矛盾。然而，一种在水滨村口口相传的说法表明，这部分板瑶中的大多数，后来被平川瑶以武力赶出了平川源。

    水滨村不少村民曾为笔者讲述这段口传历史。其概要为：上五排招徕的大部分板瑶不习惯耕地农作，在清朝初期可能已放弃佃耕，而集中在平川河上游支流冷水源山谷中刀耕火种（冷水源乃从海拔300米左右的平川河谷急剧抬升到1200米左右的陡峭高山溪流，水温明显比平川河低得多，故得此名，属大村水滨寨地界）；冷水源有百来户板瑶，很强势，甚至敢葬人到岗子上寨（属水滨寨大家族周姓的土地）；约在清乾隆年间，水滨寨周姓联合其他寨瑶民，与冷水源板瑶打了一架，死伤不少（不同的人口述数字不同，少则十几个，多则一百多个），冷水源板瑶败走，不知其踪。

    板瑶在桂东北大瑶山区颇为有名，原因之一是入山较晚，没有或极少拥有土地。费孝通于1935年调查发现，板瑶因无地或少地而地位极低，故对耕地格外渴望。30由此反观平川瑶关于板瑶离开平川源的说法，似多有可疑之处。毋宁说，情形更可能是，平川源人口慢慢增加之后，平川瑶开始收回佃权，相当一部分板瑶不得已退到山上，而且是周边地带耕作条件相对较差的冷水源。在暴力驱赶之下，这部分板瑶最后失去了在平川源的土地经营权。但是，少量未聚在冷水源的板瑶，则可能既有通过入赘、过继等方式融入平川瑶村寨者，亦有继续耕种于周边深山者。

    平川源山脉连绵不断，耕地只占极少数，绝大部分土地是开发程度很低的山地，甚至未开发的原始森林。大部分板瑶离开，自然还有新的人群流入。

    清康雍两朝全面推行人丁不单收税的政策，康雍乾之际社会总体稳定，以及红薯、玉米、土豆等旱作物扩散，31致使人口快速膨胀。康熙早期全国人口“可能已经大大超过1亿5千万”，主要“平原和低山区已经人满为患”，32至乾隆晚期又“不止翻了一番”，达到3亿多，33大量人口不得不转向深山区。

    清乾隆年间，不仅有新的以刀耕火种为主的过山瑶，还有来自宝庆府（大致为今湖南邵阳）擅长犁耕锄掘农业的农民，不断涌入平川源及其周边山地。除全国人口，尤其平原人口膨胀的大背景之外，还与宝庆府在乾隆年间特别频繁地发生灾害，灾民难有就地喘息、恢复生产的机会有关。以下略摘几处道光版《宝庆府志》记录为证。

    乾隆“十一年（1746年），武冈、新化大水”；“十二年四月，城步大水……是岁城步大火”；“十三年，城步大疫、新宁水灾……六月新化水灾”；“十四年三月，新宁、武冈水灾……庐舍湮溺甚重”。34以及，乾隆三十年（1765年）“新宁大荒，城步大水大饿……斗米银六钱”；“三十二年秋，新化大水”；“三十三年秋，新化水灾……邵阳大旱，斗米银四钱”；“三十五年，新化旱，城步麦无收”；“三十八年，新化虫伤稼”；“四十年，新化大水”；“四十三年，宝庆大旱大饥，邵阳斗米银八钱、饿殍相望，城步大旱，饥民多聚集肆掠”；四十四年，“城步大饥，斗米银六钱，新化旱”；“四十五年，新宁、武冈、邵阳、新化大水”；“四十六年春，城步大水”；“四十七年春，雷震城步……夏四月，新宁地震”。35

    宝庆人流入平川源，主要靠开荒山耕种桐籽树、油茶树为生。这从水滨村周姓族谱中保留的《立批山场契约》（以下简称《乾隆契约》）可见一斑。该契约写道：

    立批山场人广西恭城坪川源水边村、大田头、旱地四脚（房）人等……鸣锣公议，今将承祖山场座落土名大冷水、小冷水一所……四抵分明。情愿凭中说合，将来批与新化宝庆客人谢代宗、桥柏、坤宗、李咸有叔侄兄弟，耕种开挖，六成生理。当日三面言定，批山价银六十四千。二家言定开山，就日交足，并无短少分厘。每年议定，地钱照户收租，每户租钱二百八十文，风（丰）年不加，次（歉）年不少，其（期）限钱十月十五送至上门。自批之后，青山地山载种桐树、茶树，一概任从客人耕管，主家不得异言幡（翻）悔，任从客人招流（留）耕种人等，主家族内再无异言，如有个民差俞（干预）不与客人相干。若有众姓叔侄人等，不许另生枝节。新化客人谢代宗、李咸有二人不许招流（留）吃酒、打架、赌博，长人不许首流（收留），并无耕种，不许宝山乱横。又有主家茶（查）出，送官禀报，自耳（理）其罪。今恐无凭，立写批字，付与客人收执为据是实。

    请中人：俸奇通、何昌万、蒋子民。请代笔人：蒋子亮

    乾隆五十九年（1794年）十月十五日立批，永远耕种。

    值得注意，《乾隆契约》表明：其一，来自湖南新化县的宝庆人租佃山地，仍得经过水滨村周姓4个“脚”（房支）集体同意；其二，宝庆人是每年按户集资的，但对水滨周姓人而言则属于宗族公款；其三，宝庆人还可另行招留新来的人耕种。

    宝庆人原本即熟悉犁耕、锄掘，其山地耕种技术远远高于此前的过山瑶，甚至也高于平川源本地瑶民。其经营山地的模式是“用‘打锣唱歌’的形式，大面积开垦山地，第一年以种粮为主，次年则植入杉树、桐树、油茶和毛竹，并套种粮食作物，第三年则长树长竹、培植成林”。36据水滨村不少老人估算，宝庆人的套种技术比起当地瑶民种桐籽树、油茶树之后就等着收桐籽、油茶籽的方式，在开荒头十来年经济效益起码高四五倍。1952年土改时，水滨村215户，划出地主、富农共12户，其中8户是宝庆人。37此时，宝庆人居于高山，却相对富裕，证明其土地开发技术的确比较先进。宝庆人也不像此前两拨名称不详的过山瑶，以及板瑶那样，主要生计方式是游耕，而是一旦有山场可开荒，便能就地长期生存下来。

    按《乾隆契约》，宝庆人可再招徕新人进山开垦。加之其开垦效率和收益比较高，进入平川源的宝庆人也日益增多。而本地瑶民当中，也有人抵制不住利益诱惑，不经过村寨集体公议，即将山场私自租给宝庆人开垦。久而久之，又引发了新的冲突。

    现存于平川源狮塘村的一块无题碑刻，记录了一份于清嘉庆二十二年（1817年）订立的契约（以下简称《嘉庆契约》）。其文如下：

    立写天理仁义合同人周姓，李、孟、蒋、卢姓等。今因却被无齿（耻）之徒盗批双水六底业山，并行批飘以（与）湖广楚南新化宝庆之歹（徒），再于加（嘉）庆十一年（1806年）盗批。不料周姓四围（房支）众等查实不服，捉挐批主。成（呈）赴县主不印（应），具（状）往府台宪主详徐，宋（宪）主不重粮田。众等往省投告，详县、宋（宪）主不周。众等归家鸣锣集议，合口同心，情愿将冷水源大罡头一概付众，言（延）请下排四姓村老、二甲商议：水源将来下应粮田，大罡头将来二村牧牛，其出众之物，不能私已受用；水源、六底、大罡方以为上下官务之费，钱文艮（银）两每村占一半。二村合议：虎羊同群，鸡鹊同巢，情愿甘心，甘心情愿，将冷水源抄群出众（全部充公），勒石题名，平半耕管，以清藤面分水为界，二村同心抚做；其后二村不得幡（翻）悔，下村狮公塘不得退速（缩）、为悮（违误），上村周姓不得异言。如有此情，任从证立之主合同执照。上有天神共照，中有二村排甲在场，一干人等立合同，二纸一样、各执一张，存照子孙永远，证立之后，恐有无名之辈，不许入境□（采）伐，不得假湧赫□。

    《嘉庆契约》所述，即本地瑶民私租水滨寨周姓所属冷水源山场给宝庆人，周姓宗族知晓后报官，但从县、府再到省，官司打了11年未果，最后水滨寨以出让冷水源一半山场为代价，请狮塘村四姓瑶民相助，合力赶走通过私人“盗批”租得土地的宝庆人。

    《嘉庆契约》未提及如何对待经过瑶民村寨集体商议租得土地的宝庆人。依笔者对水滨村的调查推测，当时宝庆人并未全部离开，他们中的少数通过入赘、过继等方式融入了平川瑶村寨，其他的则继续耕种于周边深山。不过，此后可能少有新的宝庆人流入，新流入者主要是灌阳人（邻县灌阳的瑶人和汉人，但其瑶人所持语言与平川瑶语不同）。据曾长期担任水滨大队支书的周明统回忆，1958年观音人民公社成立时，平川源动员了1100多人下山，到河谷地带兴建村寨，或加入人口较少的瑶寨居住。其中，宝庆人480多人，其他主要是灌阳人和少量过山瑶。（访谈时间：2020年7月）

    这个1100多人的数字，加上《嘉靖碑记》所提及招徕板瑶约一千人的信息，说明平川源周边山地应至少有养活一千余人的弹性空间。当河谷人口过少时，容易从深山中招徕流动人群，到河谷耕作。当河谷人口接近饱和，尤其是深山中流动人群数量超过土地承载的弹性空间时，则容易出现土地经营权纷争。

    当然，平川瑶内部同样也存在土地竞争。一旦形成纠纷，能内部协调的则内部解决，不能的则诉诸官司。但是，由于国家难以日常化地深入平川源展开治理，讼争往往十分漫长。例如，杨梅村与邻村洋石曾为一块有水源的山场（名为牛角湾），自清嘉庆年间开始即多有纠纷、讼争，直到民国29年（1940年）方由广西高等法院第七分院判决。38平川瑶为掌控土地所有权和经营权，日常更多依赖的还是自身社会团结的力量。

    四、多元社会结合与礼之践诸于野

    从现有可考信息看，明初至永乐三年（1405年），平川源外的大家族（自称“本地人”）与平川源内的“良猺”多为山主、佃户关系。“良猺”经“造册”登记，“附籍”于“本地人”，由其代向官府转缴赋税（这说明，“本地人”更早就已登记为“民”）。后者属于官府治理“良猺”的代理人。依习惯，山地为“良猺”村寨集体租赁经营（未提及水田），地主不能未经“良猺”村寨集体商议，就售卖或转租给新来的人群。其赋税也是以村寨为单位额定缴纳，寨内人口、土地数量变动，对官府和“本地人”而言并不重要。

    平川源“良猺”社会结合首靠姓氏、家族，人口较多的成单姓村寨，甚至一姓分成两三个村寨，人口较少的则多姓结为一寨。不过，姓氏、家族未必完全一致，如古骨圩寨蒋姓与白荆铺寨蒋姓并非同一家族，据传前者先到平川源，被称为“大蒋”，后者被称为“小蒋”。

    百余年后，明嘉靖六年（1527年）“良猺”与新来人群发生流血冲突，官府保护了前者的土地使用权，让其获得了一部分土地所有权。此后，对于租赁的山地，虽然“良猺”依然得给“本地人”山主缴纳租金，但获得了招徕其他人耕种，即转租土地的权利。官府虽然还无力对其“编户齐民”，但已不满于依靠平川源外“本地人”代为治理，于是自嘉靖九年（1530年）开始实施“排瑶制”。平川源被分为10个排，每排设“猺目”，“猺目”作为“户长”直接向官府纳粮缴税，用“猺人法”39治理村寨。排，是由外置入的行政框架，但其管辖范围和头目设置，照顾到了民间以姓氏、家族为社会单位的习惯，久而久之也成了平川源重要的社会单元。迄今为止，在平川源居民的口语中，还经常会用排、上五排和下五排，来指代不同范围的地界和人群。

    在地理分隔明显的条件下，单姓村寨变大后，亲缘网络也随之扩大，内部通婚成为一种需要。例如，据清道光年间狮塘村李姓所修族谱记载：原居高山寨，康熙四十六年（1707年）首次修族谱（已散佚）；本有8个房支，人口增多后曾经族老商议，将第一、二、三房改为姓孟，以便“异姓婚配”；后传至第15代，第二、五、六房绝后，第三、七房人口也少，但第七房在第7代有一户“接”（过继）了永明县一个名叫“卢万洪”的人为子，其后代承李、卢两姓，狮塘始有卢姓（后又搬到老寨，与盘姓结为一寨）；清中期，李姓第四房一户“接”了长房一人为子，继而人丁兴旺，与部分孟姓一道开辟了名为“老虎塘”的新寨子。40

    除了分宗、过继之外，入赘也是平川瑶调整社会结合的重要方式。据传，观音村老洼、洼里两寨村民即外来陈姓人入赘老洼寨盘姓瑶家，留下的后代。其族谱记道：“嘉靖年间”，陈仁意、仁忠兄弟“流落”到老洼打铁，仁忠的独子被该寨某瑶民“招”为女婿。老洼寨李姓、王姓，也自认是外来人员入赘瑶家而留下的后代。41石坪寨是清末从平川河对面的狮尾寨何姓分出来的，但至笔者入村做调查时，俸姓人口已近该寨一半。究其缘由，也是从蕉山村招了一位俸姓女婿上门，繁衍而成。古骨圩寨“大蒋”，据族谱记载，在明万历年间招了狮塘村某杨姓村民为上门女婿，其后代承蒋、杨二姓（1949年，蒋、杨两姓还合建了宗祠）。莲花寨俸姓村民自述原姓周，明初自湖南道州来到该地，改姓俸，清嘉庆年间宗族人口增至2个房支，为“通婚之便”，第二房恢复周姓（二姓族谱同修，字辈排行亦共用）。

    此类案例说明，自清康熙、乾隆年间开始，平川源已有某种程度的“同姓不婚”和宗族的“礼”仪，至嘉庆、道光年间，这种“礼”仪已成为日常现象。不过，通过部分人改姓、分宗的变通方法，实际上同姓内部仍可通婚。入赘者所生子嗣，虽世代住在女方村中，却可以承继两姓宗祧，甚至完全随父姓。儒家所尚“礼”仪，在特殊地理和经济社会条件下，明显发生了质的改变。

    尽管如此，以“礼”为内核的宗族礼仪、祠堂，以及用谱系明晰亲缘关系的做法，毕竟成了平川源瑶民社会结合的常规方式。甚至于，他们还尝试运用此类“礼”仪，与平川源外“本地人”建立起更宏大的联盟。清道光年间，水滨寨周姓编纂族谱，可谓典型案例。

    宋代，恭城出了一位名人周渭。他曾任监察侍御史，给恭城的“民”减税役，并倡举办学。周渭去世后，宋真宗“敕封为惠烈御史周王”42，恭城有不少村建祠崇祀（今县城附近仍有两座周王庙）。清乾隆年间，恭城县内不少周姓编纂族谱，认为周渭的太祖曾居湖北襄阳，并在唐太宗治下（627年—649年）任金紫光禄大夫，生有18个儿子，字辈为“弘”，后代分布于湘西南、粤北和桂东北（同时期，与恭城县较近的湖南宝庆新宁县、道州宁远县也有类似家谱，记为“十八弘”）。其中，栗木镇上宅村周氏族谱修于乾隆二十年（1755年），西岭乡西岭村周氏族谱修于乾隆二十八年（1763年）。周渭祖籍，宋史并无记载，宋、元乃至明代民间亦无家谱记载。在其去世千余年后，却有了清晰的亲属谱系图和跨越数省的迁徙路线图。毋宁说，在清康乾嘉之际，湘桂边区人群修纂族谱，常有某种形式的附会、联盟。

    清道光壬午年（1822年），平川源水滨寨周姓也修纂了族谱。其谱记道，他们与周渭乃同一宗支，皆为周弘颂的后代，而且金紫光禄大夫实际上有24个儿子，谓之“二十四弘”。水滨寨有村民提出，可能更早就修过族谱，道光版族谱只是照抄之前的记录。考虑到彼时村中识字者并不多，且一代代将《嘉靖碑记》之类的文字保存完好，却未见对此前的家谱有只字记录，此说并不可靠。其宗祠则建得更晚，祠堂门口的石碑上刻有“大清光绪六年（1880年）庚辰岁孟冬穀立  奉旨恩受国子监太学生周显煕立”。族谱追述千年亲属脉络难免失真，却能表明早则在清康乾之际，晚则在嘉道之际，儒家之“礼”已被平川源内一些大姓用来编制群体社会关系网络。子弟被恭城送到国子监就读（另有观音村陈姓族谱提及，在晚清出过“名登仕版”的“千总”“巡检”“例贡”），侧面反映了当地文教水平不低。

    清光绪十五年（1889年），《恭城县志》修纂记录道：原来恭城瑶民“间有纳税，亦百中之一，不当差……今则东、北两乡诸猺咸编户受约束、委（威）顺服从，尽皆纳税，多有读书明理、援例报捐者”43。考虑到嘉庆年间恭城曾修纂过县志（已散佚），这段光绪年间的县志记载说明，平川源瑶民在嘉庆至光绪年间（偏近光绪年间的可能性更大），已完成“编户齐民”（深山中少量过山瑶和宝庆人、灌阳人除外）。宣统元年（1909年），他们与栗木河上游的“本地人”一并被纳入恭城县第四区，在赋役上已无明确区别。

    不过，与儒家“礼”仪一样，梅山教、佛教、巫觋信仰在当地社会文化生活中，也扮演着重要角色。

    笔者在平川源实地调查过程中，常听说上、下五排曾经共有“三十六庵、七十二庙”（一说“三十六庵、四十八庙”）。除了单家独户祭拜外，不少庙为上、下五排共同祭祀（如白马将军庙），有的是几个村庄联合祭祀，有的是一村寨或一家族祭祀。直到民国时期，稍大点的寺、庙、庵都有数量不等的水田（通常1—3石），作为庙产，并有相应的组织——“会”，以及“会首”负责管理。

    许愿、还愿（二者中间还可以“暖愿”），是平川源瑶民常见的信仰行为。其中，较大的如“盘王愿”庙会五年一届，于农历十月十五、十六日举行；“婆王愿”庙会三年一届，农历十月十五、十六日举行（上五排可作为“客人”参观），抬婆王像出游各村；“李王愿”为轮祭，狮塘麒麟庙会为农历八月十五日，蕉山近水庙会为农历七月十四日，水滨天祠庙会为农历十月十五日。“暖愿”时间根据还愿时间定，一般在农历六月农闲时日。虽然平川源瑶民对外都认可“平川瑶”，祭盘王，但在内部，上五排瑶民自称“平顶瑶”或“狗头瑶”，不祭婆王，而下五排瑶民则自称“盘瑶”，不祭李王。

    梅山教信仰则更是贯穿于平川瑶的家祭、祠堂公共祭祀、人生礼仪、岁时节日庆典等各个环节。梅山教源于湖南中西部新化县、安化县一带的梅山，宋代开梅山道后，“梅山蛮”往北（武陵山区）、往西（湘西、黔东）、往南（湘西南、桂东北）迁徙，将其宗教带往各地并各具区域特色。44就平川源而言，上五排称“梅山教”，下五排对内称“梅山教”，对外称“淮南教”。水滨村有师公（民间宗教人士）认为，二者核心仪轨和供奉神灵都相同，称呼有别可能是因为下五排与外界汉人打交道稍多些，有攀附道教的色彩。但也有师公认为二者有实质区别，在还愿仪式中，上五排只吹笙挞鼓，而下五排还会打锣敲钹，并且戴着“鬼头”面具跳“鬼舞”（有巫的色彩）。

    平川源梅山教供奉1200多位神灵。传统上村民常将其与自家祖先像一起绘于布帛卷轴上，在重要祭祀场合当神箓悬挂。1984年，水滨村莲花寨某村民清理旧宅，发现俸姓、盘姓神箓各一卷（前者主绘于清乾隆九年［1744年］，增绘于乾隆四十五年［1780年］，后者绘于乾隆六十年［1795年］），合计长108.98米，成为重要文物（现常被称为“梅山图”）。

    此外，在民间信仰中，不少土地被认为具有神圣性，禁止开发。例如，清同治年间水滨村莲花寨、矮寨所在的两个排，公议立碑禁止村民在开天庙、白马庙之间凿山烧石灰，认为会破坏“神山龙脉”。其碑文如下：

    立碑禁神山后龙。两排六□（姓）众等始祖，历来原立开天、□（白）马二庙，左右后龙神山无敢犯。不料客岁崣山何兴秀不守王章，竟敢在左边擅动神山，打石烧灰……是以众等不服，即伸猺目、地老、大彰公论。而（何兴）秀等之情畏圣，以后不敢再行。两排众等勒碑封禁……如有不法之徒胆敢左右违乱后龙神山、打石烧灰，协同禀官究治，不徇私情私放。毋违封禁，切切矣。45

    平川源自清代中晚期开始编家谱、建祠堂甚至尚科考，认可“礼”的正统性，却未如诸多平原区域一样，46将其他民间信仰变成精神生活的“配角”。相反，当地不仅民间信仰种类繁多，而且瑶民还认为信盘王、梅山教和白马将军，有身份象征意义。究其缘由，水滨村一些老人的看法值得参考。蒋礼发表示，“如果盘王、梅山教都不信，怎么还能说是上、下五排的瑶人？”曾长期任大队、村支书的周明统则说：“现在是新中国、新社会，哪个边边角角都有党的光辉，样样都变好了，不讲这些（标准）了。原来要是不讲（信）盘王、不讲（信）梅山教，你怎么有资格在上、下五排做主人，怎么（占）有山、（占）有田？”言下之意，传统时期国家难以日常化管理平川源具体事务，按当地习惯，只有平川瑶人才能占有土地，而盘王、梅山教信仰则是其身份标志。

    五、民族认同更迭及其在隙地的层累

    明初，莫祥才带庆远府河池宜山、南丹之兵到恭城剿“雷虎子”。因其时宜山多聚“獞”“獠”和“狑”，南丹多聚“性颇轻悍”的“狼”和“㺜”（“㺜”的“语言与獞同而声音稍柔”，“服饰略同獞”）47，莫祥才之兵常被称为“狼兵”。这些“狼兵”被安置在恭城东南山隘口白面寨，以防“猺”（当地现有几个村，村民自称其后裔，属壮族）。此类做法，应与明前期、中叶桂东北招“獞”防“猺”、以“狼”制“獞”的政策有关。48在官方和文人记录中，此类冲突被简便地称作“猺乱”。49但若不细究土地、赋役、里甲制度以及“军”“民”“猺”“獞”“狼兵”等人群互动，就难以全面理解这些动乱。50

    言及莫祥才本人，光绪《恭城县志》称其为“山东人”。后世白面寨周边莫姓编纂族谱，更详记其出生地为山东青州府淄博临淄九德峰村，由此推断祖上应为汉人。但是，考虑到最早的《恭城县志》编于明万历二十五年（1597年），距离明初已有二百多年，莫姓族谱编纂更晚。因此，此类记录亦非没有可疑之处。

    据科大卫考证，在明代早期、中叶的广西，尤其是河池所在的桂西，土著被招募和编成军队称为“狼兵”，配备的指挥官一般也是土著首领。51莫祥才在河池统带300名弩兵，其职位应不会太高，甚至在恭城立功后，所授的“白面寨巡检司”也是一个基层武职。作为基层官员带兵，难以绕开日常语言沟通。从社会文化层面看，如莫祥才乃数千公里外的山东淄博人，到遍地是“獞”“獠”“狑”“狼”和“㺜”的广西河池担任基层军官，如何有效“统带”？若真如此，志书既然记他在恭城立功后的武职，按常理也应记他在河池的军职，实际却只字未提。此外，志书还记道，其所带弩兵有23个姓氏。其中，除莫、贲、覃、祝、陆、蒙等后世壮族常见姓氏外，其余皆为常见汉姓。在这样的区域，一支小规模弩兵姓氏如此之多，且汉姓占大部分，亦令人存疑。

    种种迹象表明，莫祥才可能属于河池的基层土官，在当时的族类观念中，属于“獞”“獠”“狑”“狼”或“㺜”中的某类。在二百多年后恭城县修纂志书时，因其后代已登记为“民”，并接受了儒家“礼”仪，自称为汉人（甚至他称也可能已是汉人），而附会祖先源自颇有“礼”仪象征意义的齐鲁大地，隐去了其在河池的官职。此外，志书还将当时弩兵后代自认，甚至他认的各种汉姓，附加到了关于明初的历史追述中。

    由此看，历史上的民族身份表述，不太可能是本质主义的。《猺目碑记》所载叙事，亦如此。它应属过山瑶附会征剿“雷虎子”的历史，以证明自己为“良猺”，且有占“猺山”隘口及其周边土地，以及减税、免役的正当性。立碑者及其所代表的人群，显然已十分清晰地认识到，哪怕这些隘口及周边山地极为偏僻，国家仍毫无疑义是至上的“正统”。其“到／倒平源”的表述表明，至少混杂了部分源自广东封川县的过山瑶，在紧靠平川源峡谷口的平地上建村寨。

    光绪《恭城县志》另有记载：“永乐二年（1404年），拨军屯田、设寨堡，守东、西、北（乡）”，是谓“耕兵”。52平川源峡谷口为北乡的主要“猺源”隘口，应有耕兵设寨。耕兵作为“军”户，不是本地“民”壮，在招“獞”防“猺”的政策背景下，亦不可能是“猺”，只可能是“獞”。

    《猺目碑记》中所涉过山瑶也居此地，时间若是“洪武下山”打“雷虎子”，较之于“獞人”耕兵稍早，若是“景泰元年”则稍晚。相近时间到平川源峡谷口外平地的过山瑶与“獞人”耕兵是否合寨混居，已不得而知，但起码应居住在临近村寨。在紧靠平川源峡谷口平地上，现有周家塘、老氹、岩口等3个自然村寨（老氹为岩口所分出），语言既不同于栗木平地“本地人”所说的“本地话”，也不同于平川源瑶语。这或可说明历史上过山瑶、“獞人”耕兵、平川瑶人与“本地人”，在此有过复杂交融。虽然此三寨人口，在清嘉庆至光绪年间“编户”时已被记为“平川猺”，但日常实践中的民族认同势必呈更复杂的“图层”叠加之状。直至当代，他们也只自称／他称为瑶族，至于是瑶族什么支系已说不清（但肯定不是平川瑶），更不是由“獞”改名而来的壮族。

    明初“雷虎子”起事在恭城河上游山区“势江源”，其后进犯恭城县城，水路、陆路均只需经过恭城中南部，而平川源在恭城最北端的群山中。再参考光绪《恭城县志》记载莫祥才带兵剿“雷虎子”的经过，平川源居民大概率既未参与“谋叛”，亦未参与“平叛”。即使是在该事件之后，官府授权部分“良猺”进入平川源居住，亦不至于驱赶或杀戮原居民。但此后原居民未再有单独的记录和表述，应是融入了“良猺”。其文化和民族身份已无从考据，但无疑成了被“良猺”文化和民族身份覆盖的“图层”。

    水滨村村民告知笔者，平川瑶语与临近的湖南江永县西北部瑶语能大致相通（但需要认真听，加上揣摩意思），而且都信奉梅山教，而与江永县西南部通过恭城河和恭城东部相连地带的瑶语完全不同（且后者不信梅山教）。由此看，其祖上自永明县西北部移入平川源的可能性比较大。他们与平川源峡谷口外、部分源自广东封川县的过山瑶，不属同一支系。但不管是明代之前平川源遗民的原因，还是明早期湖南永明县瑶民移入之后又有少量其他过山瑶融入，直到嘉靖年间，平川源瑶语中有少量特殊词汇为过山瑶勉语口音。以至于与西岭乡新合村《猺目碑记》将“雷虎子”记为“雷伍子”发音一样，平川源上五排《嘉靖碑记》将之记为“雷午子”（在其他语境下，平川瑶语将“虎”字发音为“hao35”，将“午”字发音为“pu41”，皆迥异于“伍”［nge13］）。此外，狮塘村杨姓于清道光年间所修族谱明确承认，祖上本为汉人，元末于长沙被陈友谅乱军所杀，家人不断迁逃，明洪武二年入平川源，入源后第三代一男丁过继给盘姓瑶家为子，后代承盘、杨二姓，才成瑶民。这说明，从明初到明中期，平川源“良猺”内部有其他人群（包括部分过山瑶、汉人）混融的痕迹，但时间长了，自称与他称都变为“平川猺”。

    当时“良猺”所说的“贼”也不同于“雷虎子”那样“反”“叛”国家的人群，而是土地开发越过“良猺”认定界限的“猺”。后者势必流入该区域较晚，在深山中游耕（通常加上游猎、采集），尚未侵犯“良猺”的土地界限时，双方并无矛盾。待其人口规模或游耕范围扩大，进入“良猺”认定拥有权属的地界时，才发生矛盾。广义上说，此类人群也可被称作“过山瑶”（但与此后招徕的板瑶，应属过山瑶不同支系）。进山较晚的过山瑶被较早定居下来的自称“良猺”的过山瑶，以“贼”的名义赶走。过了若干年，县官要求“良猺”当差，“良猺”依官方渠道“申”“报”“乞”“告”，最终达成纳税但不当差的协议。其申告理由，乃附会参与征剿“雷虎子”。如此一来，两类瑶民之间争夺土地，胜利方即表述成了为国立功，实则是“通过追溯祖先的历史来决定谁有没有入住权、是不是村落的成员”53。但是，虽然“良猺”获得官方确认占有土地的权利，且表面上不用服差役，却不得再如以往那样，开垦新土地后不“升科报税”。较之于以往的优免权，新“升科”这部分其实可算一种变相的“役”。54

    如同定居于“猺源”隘口的过山瑶一样，平川源的“良猺”也能认识到，占有土地若要变成合法“权利”，就得国家认可，国家才是产权的终极定义者。明永乐二年（1404年），平川源峡谷口外由“军”户设寨堡，有耕兵守值后，次年平川源内“良猺”就“造册附籍，纳粮”，恐非巧合。只不过，“附籍”意味着官府并不日常化地深入“猺山”治理“良猺”，而是靠峡谷口外平地“本地人”大家族间接治理。由此，平川源“良猺”虽仿照峡谷口外扼守隘口的过山瑶，声称因剿“雷午子”才获得平川源的居住权，但仍不忘强调，此乃“本里故民”周、欧阳等大姓“招抚”的结果，而后者之所以“招抚”，又源于“本县（官府）责令”。其“礼法话语建构”与资源、人员流动统合，实为边地与国家整合的方式。55

    由于不断有新的人群流入“猺山”寻求生存机会，加之“招主”依仗开发山地谋利，新流入人群与原已稳定居住下来的瑶民，易发生矛盾。明嘉靖九年（1530年），平川源“良猺”与峡谷口外“本地人”大家族新招徕的“獞”“猺”发生冲突，之后招徕“板猺”耕种。在约两百年后的清乾隆年间，“良猺”又与“板猺”冲突，再招徕宝庆人耕种。约在百年后的嘉庆年间，“良猺”与宝庆人也发生了冲突。但是，事实上第一、二拨具体支系名称不详的过山瑶，以及后来的“板猺”、宝庆人，只是因未经过“良猺”村寨公议而靠私人“盗批”租得土地的那部分（尽管是大部分）离开平川源河谷地带和靠近河谷的坡地而已。那些经过“良猺”村寨公议而租得土地的人，尽管是少数，却并未全部离开，而是有少量通过入赘、过继的方式融入“良猺”村寨，其他的则长期游移于周边深山，且多有混融。

    虽然不断有其他民族人群更迭认同，融入平川源，但其认同一层层叠加、“层累”56的方向却是有“山主”地位的“良猺”，而不是其他。观音村盘姓族祖上为科考（依规定，未编户的“猺”不得参加），于清咸丰初年改姓陈，对外自称汉人，但传了7代后，在民国年间又恢复姓盘。57杨梅村一家族祖上据传为湖北武昌汉人，明初入平川源，因“此时平源多属盘姓，不得已乃改盘姓”，民国十二年（1923年）立碑改姓杨，但承认是瑶人。58

    六、结论

    中国地大而形态复杂，生态和人类生计方式、社会文化也因此多样。这些因素构成了大小不等的区域，大区域间常有山川、河流等地理“缝隙”。它们既是区域间的界限，在某些条件下也是人们跨区域流动的通道。多民族流经此类地理“缝隙”，构成了民族走廊。民族走廊在宏观上有隙地特征，微观层面则内含各种小尺度的隙地。

    隙地中有大量未开发的土地，典型的如山地及山间小盆地、峡谷，承载人口有一定的弹性空间，这是构成民族走廊的关键。在常规年景，隙地相对封闭，较少外人涉足。周边区域人口膨胀或出现饥荒、战争时，流入隙地的人群规模和速度便会激增。这些人群不管是何种民族，上山首先是为活命，逃避的是具体的战争、饥荒，而非抽象的“逃避国家”59无政府主义。从宏观上看，他们“其实是国家生活在一个更大的经济体系之中，在结构上仍然是国家体系之内，是王朝国家整体性的经济与社会体系的组成部分”60。尽管他们在隙地开发中的确有少纳税甚至免赋役的诉求，但国家才是其财产权的基石。没有国家维系底线秩序，土地开发成果则随时可能为他人侵占。为此，民族走廊中的隙地开发有冲突时，人们哪怕附会，也倾向于援引国家正统权威或“象征体系”61，为自己占有土地、控制土地经营权和享受赋役优免，寻找正当性。

    然而，国家权力发挥作用总会受制于具体的时空条件，因之可以分为两种，一是专制权力，二是基础权力。62前者是后者的基础，却难以用作日常治理；后者细致入微，可用作日常治理，但成本也因此高得多。在民族走廊的隙地开发中，不同人群围绕土地占有、经营，既有合作，又有竞争。土地开发取得效益，需要一定规模的劳动力。在特定的生产技术条件下，土地承载的弹性空间变得狭小时，一拨又一拨新流入隙地的人群，难免加剧土地占有、经营权的竞争。在基础权力有限的情况下，国家深入民族走廊中的隙地开展日常化治理，并非易事。因此，援引国家权威，虽然可声明占有土地及其经营权的正当性，却不能依靠国家深入隙地日常化地厘定土地权利边界。土地权利的日常化维系，还得靠不同人群自身社会团结的力量。

    在这种状态下，民族走廊中隙地人群的动态社会结合，就变得相当关键。一些人群依靠宗教、语言、生活习俗亲近而整合有力，防止新流入隙地的人群占有自己的土地或土地经营权。除了运用过继、入赘、联宗等亲属和“拟制”亲属“联合”63关系网络，村寨地缘共同体亦有举足轻重的地位。以至于，针对外来流动人群哪怕只是获得土地经营权，村寨公议也往往是一个先决条件。国家设定的“附籍”治理关系，尤其是通过民族精英间接治理的组织——排，亦逐步演变成地方实践中的社会结合方式。随着国家在民族走廊隙地中的角色具体化，以及隙地中的主体人群尝试进一步组织化，扩展社会关系网，接近国家权威，儒家“礼”仪也就开始逐步融入其动态社会结合过程。编族谱、建宗祠以明晰亲缘，崇祭祖先，加固亲属或拟制亲属组织，乃至建立跨越村寨、超出隙地范围的区域性联盟。

    然而，儒家“礼”仪在隙地动态社会结合的实践中，也有不得不因地制宜变形的地方。例如，人们可以通过分宗改姓，用形式上的“同姓不婚”，来应对附近村寨无法满足姻亲关系网络需要时，不得不本宗之内开亲。至于过继、入赘等行为，也可形式上满足宗族“礼”仪，但实质上有重要差别。甚至，即使隙地人群深受儒家“礼”仪浸淫，乃至接受国家“编户”，其所承赋役与外界平地上一般的“民”没有实质差别之后，仍倾向于坚守自身原有认同。在社会文化象征上，意识模型相对于无意识模型，更易“操纵”象征效力，64在人群区分和互动中，则是一种“为派系和社会变迁而辩护”65的动态机制。具体到中国社会文化认同，正统之“礼”的社会文化构想或可称“意识形态模型”，“边缘人群”自用或自我期待的构想是与之颇有差别的“自制模型”，而对周边其他人群的构想则可称“观察者模型”。66而依民族走廊隙地中不同人群互动及其认同层累的经验看，三种意识模型可能并非谁“同化”谁的关系。隙地人群既模仿乃至附会正统之“礼”，接触、混融周边人群文化，且认为它们本就是自身文化的有机组成部分。“礼”的文教渗透和实践因地制宜，与其他文化配合得当。这种“我中有你、你中有我”67的格局，将他者一定程度上化为自我，同时又在他者镜像中呈现与他者深度混融的自我，构成意识模型的动态相互镜像化。

    隙地人群在混融多层其他群体文化的基础上，日用正统之“礼”，却仍坚守局部地域主导人群的民间信仰。究其缘由，固然可能与民间信仰转型有一定的滞后性有关，但更重要的在于控制土地。在国家基础权力无法日常化深入民族走廊中隙地的情况下，只有维系隙地中微观层面主体人群的民族身份，才有资格控制土地所有权或经营权，并在土地承载弹性空间变得狭小时，排斥其他新流入隙地的人群。由于某些风俗习惯、民间信仰具有标识民族身份的作用，隙地中的主体人群以及那些尝试通过各种方式融入该群体的人，即使深受儒家“礼”仪影响，也仍倾向于延用而不是中断这些风俗习惯、民间信仰。以至民族走廊中的隙地人群一方面“渐慕华风”68，另一方面又倾向于长期坚守少数民族认同。不了解这一点，界定“华夏边缘”69就难免平面化。

    在历史的长河中，不同人群在民族走廊的隙地中交往、交流、交融。其民族认同也因此一层又一层累积，最终积淀成一种社会记忆。民族认同层累离不开族源叙事，叙事中会有覆盖、改写、附会，甚至无中生有，但积淀成相对稳定的社会记忆之后，便再也无法简单还原。若不细致考究，则难以看清其层累的痕迹。族源叙事虽然未必真实，但层累起来的认同本身却是真实的，在相当长历史时期内有相当强的稳定性。至于其认同层累的方向，究竟导向哪一种民族，则与民族走廊隙地中特定的生态、生计和人群互动过程有关。在这个意义上，尽管民族走廊中不同人群会叙述各种迁徙史（什么民族到了什么地方），但这只是民族认同层累的一个方面，另一方面同样重要的是，到了什么地方慢慢就成了什么民族。对于后一种机制，目前的研究似乎还算不上充分。

    从这个角度看，民族认同研究不宜套用本质主义叙事，只讲述实体般的多民族迁徙史，并且常想方设法溯及远古。如此叙事，讲得再好，即便不是错误的，至少是只讲了历史的一方面。而关于民族认同在地生成机制的叙事，似还有必要花大力气深入研究。从民族走廊及其隙地中长时段、多民族的互动过程看，很显然，多样的人群层累成何种民族认同，与其所经历的地理空间、生态环境、社会互动和文化交流，以及各种制度限定下的政治经济过程，有着密切的联系。这正是民族走廊的形成，及其所孕育的中华民族“多元一体”70和而不同的机制。由此看，从隙地认识民族走廊，从民族走廊认识中国的构成机制，还大有潜力可挖。

    本文转自《开放时代》2025年第1期

  • 马卫东：大一统源于西周封建说

    大一统思想是中国传统文化的重要内容之一，在两三千年的历史岁月中，对于促进中国国家统一、中华民族形成及中华文化繁荣，曾起到过巨大的作用。然而，大一统思想最早形成于什么时代，源于什么样的历史实际，学界却长期存在着不同的看法。多数学者认为，《公羊传》所提出的“大一统”，是战国时代才开始出现的学说，战国以前既无一统的政治格局，也无一统的社会观念。近年来，有的学者提出，中国早在西周时期已是统一王朝，“现在我们不能再以为，只有到了战国时期才开始有统一的意志”，但似乎并没有在史学界引起普遍反响。因此，有必要继续对大一统思想的渊源作进一步深入的探讨。

    本文认为，《公羊传》大一统思想的基本内涵是“重一统”。其具体内容，包括以“尊王”为核心的政治一统；以“内华夏”为宗旨的民族一统；以“崇礼”为中心的文化一统。历史表明，《公羊传》的大一统理论是对西周、春秋以来大一统思想的理论总结。周代的大一统思想，是西周封建和分封制度的产物，它源于西周分封诸侯的历史实际及西周封建所造成的三大认同观念：天子至上的政治认同、华夷之辨的民族认同、尊尚礼乐的文化认同。中国大一统的政治局面和思想观念由西周封建所开创，是西周王朝对中国历史的重大贡献之一。

    一、《公羊》大一统说的内涵及其思想渊源

    “大一统”的概念，最早是由战国时代的《公羊传》提出来的，系对《春秋》“王正月”的解释之辞。《春秋·隐公元年》：“元年，春，王正月。”《公羊传》释曰：“元年者何？君之始年也。春者何？岁之始也。王者孰谓？谓文王也。曷为先言王而后言正月？王正月也。何言乎王正月？大一统也。”

    “大一统”的“大”字，以往多解释为大小的大。其实，这不符合《公羊传》的本义。这里的“大”字应作“重”字讲。按《公羊传》文例，凡言“大”者，多是以什么为重大的意思。如《公羊传·隐公三年》：“君子大居正。”《庄公十八年》：“大其为中国追也。”《襄公十九年》：“大其不伐丧也。”以“大”为“重”，这在先秦两汉文献中不乏其例。《荀子·非十二子》：“大俭约。”王念孙曰：“大亦尚也，谓尊尚俭约也。”《史记·太史公自序》：“大祥而众忌讳。”即重祥瑞而多忌讳。

    “大一统”的“统”字，《公羊传·隐公元年》何休曰：“统者，始也，总系之辞。”许慎《说文解字》释“统”曰：“统，纪也。”又曰：“纪，别丝也。”段玉裁：“别丝者，一丝必有其首，别之是为纪；众丝皆得其首，是为统。”

    刘家和先生在汉人解诂的基础上，深入分析了《公羊传》“一统”的涵义，认为《公羊传》的“一统”，“不是化多（多不复存在）为一，而是合多（多仍旧在）为一。……但此‘一’又非简单地合多为一，而是要从‘头’、从始或从根就合多为一。”

    “大一统”的“一统”，学界往往解释为“统一”，实属误解。关于“一统”与“统一”的区别，台湾学者李新霖先生曾有精辟的论述：“所谓一统者，以天下为家，世界大同为目标；以仁行仁之王道思想，即一统之表现。……所谓统一，乃约束力之象征，齐天下人人于一，以力假仁之霸道世界，即为统一之结果。”

    综合古今诠释，对《公羊传》“大一统”的内涵，我们可以作如下的理解：“大一统”就是“重一统”，具体而言是“重一始”或“重一首”，即通过重视制度建设、张扬礼仪道德，以主体的、原始的、根本的“一”，来统合“多”而为一体（合多为一）；“大统一”则是通过征伐兼并和强力政权消除政治上的“多”，实现国家统治的“一”（化多为一）。可见，从严格的意义上讲，“大一统”和“大统一”并不是两个等同的概念。

    《公羊传》根据《春秋》“王正月”，开宗明义地提出了大一统概念。在阐释历史事件时，又论述了大一统理论的具体内容。从《公羊传》的论述看，《公羊传》大一统理论主要包含三方面内容：以“尊王”为核心的政治一统；以“内华夏”为宗旨的民族一统；以“崇礼”为中心的文化一统。

    强调尊王，维护天子的独尊地位，是《公羊传》大一统理论的核心。《公羊传》首先通过对诸侯独断专行的批评，表达了尊王之义。如《春秋·桓公元年》：“郑伯以璧假许田。”《公羊传》释曰：“其言以璧假之何？易之也。易之，则其言假之何？为恭也。曷为为恭？有天子存，则诸侯不得专地也。”《春秋·僖公元年》：“齐师、宋师、曹师次于聂北，救邢。”《公羊传》释曰：“曷为先言次而后言救？君也。君则其称师何？不与诸侯专封也。”《春秋·宣公十一年》：“冬十月，楚人杀陈夏徵舒。”《公羊传》释曰：“此楚子也，其称人何？贬。曷为贬？不与外讨也。……诸侯之义，不得专讨。”在《公羊传》看来，诸侯的“专地”、“专封”、“专讨”都是违背“一统”的行为，所以《春秋》特加贬损，以维护周天子的权威。在《公羊传》中，关于尊王的论述很多，如“王者无外”（《公羊传·隐公元年》、《公羊传·成公十二年》），“不敢胜天子”（《公羊传·庄公六年》），“王者无敌”（《公羊传·成公元年》）等等，无不是主张“尊王”的慷慨之辞。在周代，天子是最高权力的代表，也是政治一统的标志。《公羊传》的尊王思想，实际上就是主张建立以天子为最高政治首脑，上下相维、尊卑有序的政治秩序，通过维护周天子的独尊地位来实现国家的政治一统。

    以华夏族为主体民族、尊崇华夏文明的“内华夏”思想，是《公羊传》大一统理论的另一重要内容。《公羊传·成公十五年》：“《春秋》，内其国而外诸夏，内诸夏而外夷狄。王者欲一乎天下，曷为以外内之辞言之？言自近者始也。”何休：“明当先正京师，乃正诸夏。诸夏正，乃正夷狄，以渐治之。叶公问政于孔子，孔子曰‘近者说，远者来’。”可见，如何处理华夷关系是大一统理论的应有之义。在华夷关系上，《公羊传》一方面确认华夷之辨，屡言“不与夷狄之执中国”（《公羊传·隐公七年》、《公羊传·僖公二十一年》），“不与夷狄之获中国”（《公羊传·庄公十年》），“不与夷狄之主中国”（《公羊传·昭公二十三年》、《公羊传·哀公十三年》），等等，反对落后的夷狄民族侵犯华夏国家。另一方面，又认为华夷之间的界限并非不可逾越，无论是华夏还是夷狄，只要接受了先进的周礼文化，就可成为华夏的成员，即唐代韩愈在《原道》一文中所概括的“诸侯用夷礼则夷之，进于中国则中国之”。因此，《公羊传》的“内华夏，外夷狄”思想，实际上就是主张建立以华夏族为主体民族，华夷共存、内外有别的民族统一体，并逐渐用先进的华夏文明融合夷狄民族，从而实现国家的民族一统。

    尊尚周礼文化的崇礼思想，也是《公羊传》大一统理论的重要内容之一。《公羊传》认为，天子与诸侯有严格的等级秩序和礼制规范。如《公羊传·隐公五年》：“天子八佾，诸公六，诸侯四。……天子三公称公，王者之后称公，其余大国称侯，小国称伯子男。”《公羊传》强调诸侯要严格遵守周礼，不得逾越，以维护天子的独尊地位。《公羊传》还通过天子、天王、王后、世子、王人、天子之大夫等名例表明尊王之义。如《公羊传·成公八年》：“其称天子何？元年春王正月，正也。”《公羊传·桓公八年》：“女在其国称女，此其称王后何？王者无外，其辞成矣。”《公羊传·僖公五年》：“曷为殊会王世子？世子贵也。”《公羊传·僖公八年》：“王人者何？微者也。曷为序乎诸侯之上？先王命也。”《公羊传》张扬周礼的目的，旨在“欲天下之一乎周也”（《公羊传·文公十三年》），即通过诸侯国和周边民族对周礼的认同，实现国家的文化一统，进而促成并维护国家的政治一统和民族一统。

    由上可知，《公羊传》大一统理论的最大特色就是“合多为一”。具体言之，在政权组织上，首先确认周王室为最高的政权机关，同时承认诸侯国地方政权的合法地位，由王室统合各诸侯国而实现国家的政治一统；在民族结构上，首先确认华夏族的主体民族地位，同时承认夷狄非主体民族，由华夏统合夷狄而实现国家的民族一统；在文化认同上，首先尊尚周礼文化为先进文化，同时涵容各具特色的地域文化，由周礼文化统合各地域文化而实现国家的文化一统。

    《公羊传》由阐释《春秋》而提出大一统学说，其理论直接源于《春秋》。《春秋》是孔子据《鲁春秋》编作的一部史书。在《春秋》一书中，孔子通过对春秋历史的笔削裁剪，表达了自己的政治观点，即所谓的《春秋》大义。其中，“大一统”便是《春秋》的首要之义。《孟子·滕文公下》：“《春秋》，天子之事也。”《史记·太史公自序》：“夫《春秋》，上明三王之道，下辨人事之纪，别嫌疑，明是非，定犹豫，善善恶恶，贤贤贱不肖，存亡国，继绝世，补敝起废，王道之大者也。”又《太史公自序》：“周道衰废……孔子知言之不用，道之不行也，是非二百四十二年之中，以为天下仪表，贬天子，退诸侯，讨大夫，以达王事而已矣。”《孟子》和《史记》所说的“天子之事”、“王道之大”、“以达王事”，即指《春秋》集中表达了孔子的大一统思想。

    除《春秋》一书外，孔子的大一统思想，在《论语》、《礼记》等文献中亦多有反映。如：《论语·季氏》：“天下有道，则礼乐征伐自天子出；天下无道，则礼乐征伐自诸侯出。”《礼记·坊记》：“子曰：‘天无二日，土无二王，家无二主，尊无二上。’”《礼记·曾子问》：“孔子曰：‘天无二日，土无二王，尝禘郊社，尊无二上。’”《论语·颜渊》：“四海之内，皆兄弟也。”《论语·子路》：“叶公问政，子曰：‘近者悦，远者来。’”《论语·子罕》：“子欲居住九夷，或曰：‘陋，如之何？’子曰：‘君子居之，何陋之有？’”以上的诸多论述，都是孔子大一统思想的体现。孔子的大一统思想，是《公羊传》大一统理论的直接来源。

    孔子所生活的春秋时代，天子日益衰微，诸侯势力坐大，“礼乐征伐自天子出”的政治格局趋于瓦解，社会陷入了诸侯争霸、战乱频仍的混乱局面。有鉴于此，孔子大声疾呼，推崇“一统”，渴望国家重新实现安定和统一。孔子的大一统思想也有其思想渊源。《论语·为政》：“殷因于夏礼，所损益可知也；周因于殷礼，所损益可知也；其或继周者，虽百世可知也。”《论语·八佾》：“周监于二代，郁郁乎文哉!吾从周。”《论语·阳货》：“如有用我者，吾其为东周乎!”可见，孔子的“大一统”思想，实质上是主张恢复上有天子、下有诸侯的西周式的、一统的社会秩序。《史记·太史公自序》载孔子曰：“我欲载之空言，不如见于行事之深切著明也。”这说明，孔子的大一统思想，应当有其更早的历史渊源。

    二、《公羊》“尊王”思想源于西周天子至上的政治认同

    从文献记载看，《春秋》和《公羊传》所阐述的大一统思想，早在西周、春秋时代已是一种重要的社会观念。“每一个时代的理论思维，从而我们时代的理论思维，都是一种历史的产物”，大一统思想亦不例外。历史表明，周代的大一统思想是西周封建和分封制度的产物，反映了周代社会的政治关系和意识形态。

    首先，西周封建和分封制度，加强了周天子的权力，使周天子确立了“诸侯之君”的地位。而周天子“诸侯之君”地位的确立，导致了西周一统政治格局与天子至上政治认同观念的形成。《公羊传》以“尊王”为核心的政治一统思想，源于西周一统政治形成的历史实际及周代对王权至上的认同观念。

    夏商时期，王权已经存在。在商代甲骨文和有关文献中，商王屡称“余一人”、“予一人”，表明商代的王权已经形成。然而，商代与西周的王权不可同日而语。在商王统治期间，邦畿之外方国林立。商王对外用兵，征服了一些方国，将其纳入王朝的“外服”。《尚书·酒诰》：“越在外服，侯、甸、男、卫、邦伯。”被征服的方国同商王朝有一定程度的隶属关系。然而，商代的“服国”不是出于商王朝的分封，其服国所辖的土地和人民并非商王赐予，而是其固有的土著居民；服国的首领原是方国的首长，同商王没有血缘关系；服国内仍保持着本族人的聚居状态；服国与商王朝的隶属关系在制度上也缺少明确的规定和保证。因此，商王在“外服”行使的政治权力是有限的。商王和服国首领之间，“犹后世诸侯之于盟主，未有君臣之分也”。在商王和服国首领君臣关系尚未确立的条件下，商王朝无法形成“礼乐征伐自天子出”的政治格局。

    西周的封建和分封制度的实行，“造成了比夏、商二代更为统一的国家，更为集中的王权”。分封制度下西周王权的加强，主要体现在天子与诸侯间君臣关系的确立以及相关的制度规定上。

    西周分封的基本内容，是“受民”、“受疆土”。“受民”、“受疆土”活动本身，便是对君主制的一种确认，即下一级贵族承认其所受的土地和民人，是出于上一级君主的封赐。分封的直接后果之一，是导致了天子与诸侯、诸侯与卿大夫之间君臣关系的确立。《左传·昭公七年》：“王臣公，公臣大夫，大夫臣士。”《仪礼·丧服传》郑玄：“天子、诸侯及卿大夫，有地者皆曰君。”《礼记·曲礼下》：“诸侯见天子曰臣某侯某。”周初经过分封，周天子由夏、商时的“诸侯之长”变成了名副其实的“诸侯之君”。

    天子与诸侯间的君臣关系，集中表现在西周天子的权利和诸侯所承担的义务上。对天子的权利和诸侯的义务，周王室有许多制度规定：

    策命与受命。周天子在分封诸侯时，要举行策命仪式，诸侯接受了策命，就等于接受了天子的统治。如周初封鲁，要求鲁公“帅其宗氏，辑其分族，将其类丑，以法则周公”；封卫，要求康叔“启以商政，疆以周索”；封晋，要求唐叔“启以夏政，疆以戎索”（《左传·定公四年》）。足证受命的诸侯要奉行天子的政令。诸侯国新君嗣位，也要经过天子的策命。《诗·大雅·韩奕》载韩侯嗣位，“王亲命之，缵戎祖考，无废朕命，夙夜匪解，虔共尔位”。周代的策命礼仪，实际是对分封制下天子和诸侯君臣关系的一种确认。

    制爵与受爵。在分封制下，周天子为诸侯规定了不同等级的爵命。《左传·襄公十五年》：“王及公、侯、伯、子、男、甸、采、卫、大夫各居其列。”《国语·周语中》：“昔我先王之有天下也，规方千里以为甸服。……其余以均分公、侯、伯、子、男，使各有宁宇。”《国语·楚语上》：“天子之贵也，唯其以公侯为官正也，而以伯子男为师旅。”爵命是诸侯的法定身份。诸侯阶层依据爵命分配权力、财富并对天子承担规定的义务。

    巡守与述职。在分封制下，天子有巡守的权利，诸侯有“述职”的义务。《孟子·告子下》：“天子适诸侯曰巡狩。”其具体内容便是“春省耕而补不足，秋省敛而助不给。入其疆，土地辟，田野治，养老尊贤，俊杰在位，则有庆，庆以地。入其疆，土地荒芜，遗老失贤，掊克在位，则有让。一不朝，则贬其爵；再不朝，则削其地；三不朝，则六师移之”（《孟子·告子下》）。可见，天子是通过巡守这一政治活动，来行使在政治上对诸侯的统治权力的。《孟子·告子下》：“诸侯朝于天子曰述职。”其具体内容，便是定期朝见天子，接受天子的政令。《国语·周语上》：“诸侯春秋受职于王。”《左传·僖公十二年》：“若节春秋来承王命。”《国语·鲁语上》：“先王制诸侯，使五年四王、一相朝。终则讲于会，以正班爵之义，帅长幼之序，训上下之则，制财用之节，其间无由荒怠。”述职是诸侯对天子履行义务的主要形式。

    征赋与纳贡。在经济上，天子有向诸侯征赋的权利，诸侯有向天子纳贡的义务。《国语·吴语》：“春秋贡献，不解于王府。”贡赋的多少，原则上根据诸侯的爵位高低来确定。《左传·昭公十三年》：“昔天子班贡，轻重以列。列尊贡重，周之制也。”不纳贡赋，要受到天子的惩罚。如春秋时齐桓公伐楚，理由之一是楚国“包茅不入，王祭不共”（《左传·僖公四年》）。

    调兵与从征。在军事上，天子有权从诸侯国征调军队，诸侯有从征助讨的义务。如在周初征讨东夷的战争中，鲁侯伯禽曾奉命“遣三族伐东国”。成王东征时，“王令吴伯曰：以乃师左比毛父。王令吕伯曰：以乃师右比毛父”。诸侯从征助讨，是义不容辞的义务。此外，诸侯征讨“四夷”或有罪之国有功，则应“献捷”、“献功”于周天子。《左传·庄公三十一年》：“凡诸侯有四夷之功，则献于王。”《左传·文公四年》：“诸侯敌王所忾而献其功。”诸侯向天子“献捷”、“献功”，实质上是对天子最高军事权力的一种确认。

    除了从制度上对最高王权进行确认外，西周统治者还从理论上对王权的至上性进行了阐述。西周统治者认为，周王的权力来源于上天。《诗·大雅·大明》：“有命自天，命此文王。”《诗·大雅·下武》：“三后在天，王配于京。”《诗·大雅·假乐》：“假乐君子，……受禄于天。”周王被视为上帝的儿子，代表上帝统治人间。《尚书·召诰》：“皇天上帝，改厥元子。”因此，周初统治者创造了“天子”一词，作为王的尊称。

    据统计，周法高《金文诂林》一书收集的青铜器，有65件有“天子”的称号。在《尚书》、《诗经》等先秦文献中，“天子”的称呼也屡见不鲜。如《诗·大雅·江汉》：“虎拜稽首，天子万年。……作召公考，天子万寿。明明天子，令闻不已。”刘家和先生深入分析了“天子”称号的历史意义：

    天只有一个，天下只有一个，天命也只有一个。……所以天之元子或天子在同一时间内应该也只能有一个，他就是代表唯一的天而统治唯一的天下的唯一的人。

    周代统治者通过王权神授理论，论证了王权的至上性。此外，还把“天命”和“德”联系起来，论证了王权至上的正当性。《尚书·召诰》：“王其德之用，祈天永命。”《尚书·大诰》：“天棐忱辞，其考我民。”《尚书·泰誓》：“天视自我民视，天听自我民听。”《尚书·康诰》：“天畏棐忱，民情大可见。”也就是说，上帝的旨意是通过“民情”表现出来的，周天子因为深得民心才获得了天命。周代统治者通过这种道德化的天命观，使王权获得了“天意”与“民心”的双重依据，有效地强化了周天子的绝对权威。

    西周天子与诸侯之间君臣关系的确立和王权的加强，使周天子在分封诸侯时，能够将周王室统一的社会制度推行到各个诸侯国。统一的社会制度在各个诸侯国的施行，表现在政治制度方面，主要是诸侯国都要实行分封制度、宗法制度、世卿世禄制等；在经济制度方面，诸侯国都要实行井田制度等；在军事制度方面，各诸侯国要实行国人当兵、野人不当兵及“三时务农一时讲武”的制度等。周天子与诸侯之间君臣关系的确立、统一的社会制度在各个诸侯国的施行，标志着西周政治一统格局已经形成。

    在分封制度下，各诸侯国一方面实行王室规定的统一的社会制度，另一方面又享有相当大的地方自治权。政治上，诸侯国有设置采邑地方政权和任命官吏的权力；经济上，诸侯国除向周王室交纳一定的贡赋外，其他经济收入一律归诸侯国所有；军事上，诸侯国有组建军队、任命将帅、调遣与指挥军队的权力。因此，西周分封制政体，不同于后世郡县制基础上的中央集权制政体。在中央集权制政体下，郡守、县令的任命权掌握在皇帝之手，郡县的财政归国家所有，郡县更无组建、调遣军队的权力。可见，西周分封制政体和后世的中央集权制政体，虽然本质上都是“一元”政治，但中央集权制政体的“一”之下，不存在着“多”，即不存在实行地方自治的郡县地方政权（周边少数民族地区的藩属政权除外）。而西周分封制政体的“一”之下，则存在着“多”，即存在着实行地方自治的诸侯国和采邑地方政权。

    为了实现分封制下的“一元”统治，西周王朝规定了本大末小的原则，使王室在各级政权机关中居于绝对的支配地位。据文献记载，天子的王畿有千里之广，诸侯国中的大国只有百里之地，而次国和小国尚不足百里。天子握有十四师的兵力，而诸侯大国不过三师、二师，小国仅一师。强大的经济和军事力量，保证了周王室在西周的政治格局中，成为了主体的、原始的、根本的“一”，能够统合其他的“多”（诸侯国）而为一体，建立起本大末小、强干弱枝的一统政治，即“礼乐征伐自天子出”的政治局面。

    随着分封制度的实行，王权至上观念也在畿内地区和各诸侯国境内得到极力宣扬，并且首先在上层社会形成了对王权至上的普遍认同。在周代文献中，对王权至上的认同和颂扬，记载颇多。如《尚书·洪范》：“惟辟作福，惟辟作威，惟辟玉食。臣无有作福、作威、玉食。”《诗·小雅·北山》：“溥天之下，莫非王土；率土之滨，莫非王臣。”《诗·大雅·下武》：“媚兹一人，应侯顺德。”《诗·大雅·文王有声》：“自西自东，自南自北，无思不服。”《诗·大雅·假乐》：“百辟卿士，媚于天子。”《大克鼎》：“天子其万年无疆，保乂周邦，畯尹四方”等等，都是周人尊王、王权至上观念的反映。

    孔子和《公羊传》以“尊王”为核心的政治一统思想，与西周以来天子至上的王权认同观念是一脉相承的，而这种天子至上的政治认同观念，又源于西周一统政治形成和确立的历史实际。周代的一统政治和一统观念，归根结底，都是西周封建诸侯与分封制度的产物。近代国学大师王国维在论述周初的分封诸侯时，曾有如下的论断：“新建之国皆其功臣昆弟甥舅，本周之臣子，而鲁卫晋齐四国又以王室至亲，为东方大藩，夏殷以来古国方之蔑矣。由是天子之尊非复诸侯之长，而为诸侯之君。……此周初大一统之规模，实与其大居正之制度，相待而成者也。”王国维先生以“大一统”源于周初封建，可谓是不易之论。

    三、《公羊》“内华夏”思想源于西周华夷之辨的民族认同

    西周封建诸侯和分封制度的实行，促成了华夏族的形成与华夏族主体民族地位的确立，而所谓的“华夷之辨”，则是反映了这一历史实际的民族认同。《公羊传》以“内华夏”为宗旨的民族一统思想，源于西周封建所造成的华夏族形成的历史实际以及周代社会“华夷之辨”的民族认同观念。

    关于华夏族，以往有些论著认为，它是随着夏代国家的形成而形成的。实际上并非如此。夏朝虽已产生了凌驾于社会之上的权力机构，但国家仍建立在氏族联合的基础之上。《史记·夏本纪》所载的夏后氏、有扈氏、有男氏、斟寻氏、彤城氏、褒氏、费氏、杞氏、缯氏、辛氏、冥氏、斟戈氏等，都是组成国家的不同氏族。即便商王朝的外服方国，也还是一些“自然形成的共同体”，其居民都是固有的土著居民。处于早期国家阶段的夏、商，组成国家的各氏族、方国都保持着相对单一的族属和血缘，它们与居于统治地位的夏族、商族之间存在着严格的血缘壁垒，彼此的生活方式、语言习惯、礼仪风俗有很大的差别。在这种国家形态下，难以形成一个具有民族自觉意识、共同文化和共同地域的更高形态的民族。

    华夏族作为中华民族统一体的主体民族，形成于西周大规模的封建之后，是周代封建和分封制度的产物。

    周人在克商以前，以周为首的反商联盟有了较大的发展。《逸周书·程典解》：“文王合六州之侯，奉勤于商。”周人把这个联盟称作“区夏”或“有夏”。《尚书·康诰》：“惟乃丕显考文王，……用肇造我区夏。”《尚书·君奭》：“惟文王尚克修和我有夏。”《尚书·立政》：“帝钦罚之，乃伻我有夏式商受命，奄甸万姓。”据沈长云先生研究，“‘夏者，大也’，《尔雅·释诂》及经、传疏并如此训。《方言》说得更清楚：‘自关而西，秦晋之间，凡物之壮大者而爱伟之，谓之夏。’……（周人）使用‘夏’这个人皆爱伟之的称谓来张大自己的部落联盟，来壮大反商势力的声威”。可见，周人是用“夏”来称呼以周邦为首的反商联盟。在周王朝大规模分封之前，这个在“夏”的名义下组成的军事联盟，尚未具有民族的含义。

    华夏族是在周初封建之后的历史进程中逐渐形成的。周初封邦建国时，所面临的最基本形势便是地广人稀。据朱凤瀚先生估算，周人当时的人口约十五万人。除了相当一部分留在王畿，剩下分到数十个国中，各国受封人口之少可想而知。周初分封的这种特殊的政治环境，造就了受封诸国“强烈的‘自群’意识”。周王室适应这一需要，于分封和分封之后的历史进程中，在周王室和各诸侯国的名称上冠以“夏”这个“人皆爱伟之的称谓”，即“诸夏”或“诸华”。所谓“诸夏”或“诸华”，是各诸侯国以整体的名义，一体向境内及周边其他各族所宣示的自称。后来，各诸侯国原有的各族居民，逐渐地接受了周人的礼乐文化，周王室和各诸侯国及其境内的居民，初步具有了“共同的语言、共同的经济基础、共同的地域、共同的文化意识”的民族要素。

    “诸夏”或“诸华”的共同标准语言——“雅言”。《论语·述而》：“子所雅言，《诗》、《书》、执礼，皆雅言也。”雅言即夏言，本是宗周地区的方言语音。随着分封的推行，雅言逐渐成为各诸侯国在举行礼仪活动等场合使用的标准语言。

    “诸夏”或“诸华”各国实行周王室规定的统一的政治、经济和军事制度。井田制度的普遍推行，表明各诸侯国已经具有了“共同的经济基础”。

    “诸夏”或“诸华”逐渐形成了原有各族居民的共同地域。周初封建打破了受封地区的血缘聚居局面，使不同族属的居民在同一地区实现了混居。《大盂鼎》云：“赐汝邦司四伯，人鬲自驭至于庶人六百又五十又九夫；赐夷司王臣十又三伯，人鬲千又五十夫。”鲁、卫、晋受封时，带去了“殷民六族”、“殷民七族”和“怀姓九宗”。这些不同族属的居民经过长时间的杂居、融合，到了西周后期，“在周封各诸侯国中已经基本看不到原有居民的身影，鲁国没有了‘商奄之民’，卫国没有了殷人……他们已共同融合为鲁人、卫人，标志着周封各诸侯国民族融合的完成”。这种情形，使得中原地区连成一片，逐渐演变成原有各族居民共同的地域。

    “诸夏”或“诸华”形成了共同的文化意识。随着分封，“诸夏”或“诸华”的居民逐渐接受了宗周的礼乐文化。《左传·定公十年》孔颖达疏：“中国有礼仪之大，故称夏。”《战国策·赵策二》：“中国者，聪明睿知之所居也，万物财用之所聚也，贤圣之所教也，仁义之所施也，诗书礼乐之所用也，异敏技艺之所试也，远方之所观赴也，蛮夷之所义行也。”“诸夏”或“诸华”居民对周礼文化的普遍认同，标志着“诸夏”或“诸华”共同文化意识的形成。

    总之，西周封建之后，受封诸侯国的各族居民经过融合，逐渐形成了一个有着“共同的语言、共同的经济基础、共同的地域、共同的文化意识”的民族——华夏族。

    华夏民族的形成，西周王朝的强大及其对境内和周边民族统治的加强，使华夏族的主体民族地位得以确立。而西周王朝的非主体民族，则是居于王朝境内和周边地区的“蛮夷戎狄”。华夏族的主体民族地位的确立，使华夏族在西周的民族格局中，成为了主体的、原始的、根本的“一”，能够统合其他的“多”（戎狄蛮夷）而为一体，共同组成了西周统一王朝的民族统一体。

    华夏族作为西周王朝主体民族的地位，在周王朝周边民族与周王朝的朝贡关系上有集中的反映。《逸周书·王会》记载了周成王召集的成周之会，参加这次盛会的有众多的东西南北的周边民族，各族都向周王献纳了方物。《王会》篇编撰于春秋末，周初是否有如此之多的民族参加了成周之会，史料上缺乏更多确切的说明。但西周时期许多周边民族与周王朝保持着朝贡关系，应当属实。《国语·鲁语下》：“昔武王克商，通道于九夷、百蛮，使各以其方贿来贡，使无忘职业。于是肃慎氏贡楛矢石砮，其长尺有咫。”《国语·周语上》：“今自大毕、伯士之终也，犬戎氏以其职来王。”《兮甲盘》：“王命甲政司成周四方积，至于南淮夷。淮夷旧我帛晦人，毋敢不出其帛、其积、其进人、其贾。”以上文献记载表明，臣服于周的民族与周王朝建立了朝贡关系。周朝还设官掌管戎狄蛮夷朝贡之事。《周礼·怀方氏》：“掌来远方之民，致方贡，致远物，而送逆之，达之以节。”《周礼·象胥》：“掌蛮、夷、闽、貉、戎、狄之国使，掌传王之言而谕说焉，以和亲之。”周边民族与周王朝的朝贡关系的建立，实质上是非主体民族对华夏主体民族统治地位在政治上的一种确认。

    华夏族形成之后，与周王朝境内和周边非主体民族的关系日益密切而广泛，民族融合的进程因此而大大地加速。《国语·郑语》记史伯所述西周末年的形势说：“当成周者，南有荆蛮、申、吕、应、邓、陈、蔡、随、唐；北有卫、燕、狄、鲜虞、潞、洛、泉、徐、蒲；西有虞、虢、晋、隗、霍、杨、魏、芮；东有齐、鲁、曹、宋、滕、薛、邹、莒；是非王之支子母弟甥舅也，则皆蛮、荆、戎、狄之人也。”可见，剩下的戎狄蛮夷已可得而数。春秋时期，大部分戎狄蛮夷在强国开疆拓土的过程中被征服而融合。西方的戎族，多被秦国所灭。北方狄族，多被晋国所灭。东方的夷族，多被齐、鲁所并。南方的群蛮，先后被楚国所灭。到了春秋末年，中原地区的戎狄蛮夷，已基本上融入华夏民族之中。

    随着华夏族的形成、华夏族主体民族地位的确立和华夏族的不断壮大，在西周、春秋时期，形成了“华夷之辨”的民族认同观念。周代文献中的“中国”、“华夏”、“四夷”、“五服”、“九服”等概念，都不同程度地反映了这种观念。

    “中国”一词，最早出现于成王时期的青铜器《何尊》铭文：“余其宅兹中国。”本义指京师洛邑地区。后来随着周人统治地域的扩大，“中国”一词的意义也逐渐改变，成为华夏诸国的代称。如《左传·庄公三十一年》：“凡诸侯有四夷之功，则献于王，王以警于夷，中国则否。”《左传·僖公二十五年》：“德以柔中国，刑以威四夷。”以中国指称华夏，正是华夏中心意识的一种反映。

    “华夏”一词，乃周人本其“尚文（彩）”之风尚，在沿用已久的“夏”字之前冠“华”而成的。《尚书·武成》：“华夏蛮貊。”孔安国传：“冕服采章曰华。”《左传·定公十年》：“裔不谋夏，夷不乱华。”孔颖达疏：“中国有礼仪之大，故称夏；有服章之美，谓之华。华夏一也。”华夏的称谓，体现了华夏族在文化上的优越感。

    五服与九服之说屡见于周代文献。《尚书·禹贡》：“五百里甸服。……五百里侯服。……五百里绥服。……五百里要服。……五百里荒服。”《国语·周语上》：“先王之制，邦内甸服，邦外侯服，侯卫宾服，蛮夷要服，戎狄荒服。甸服者祭，侯服者祀，宾服者享，要服者贡，荒服者王。”《周礼·职方氏》：“乃辨九服之邦国。方千里曰王畿，其外方五百里曰侯服，又其外方五百里曰甸服，又其外方五百里曰男服，又其外方五百里曰采服，又其外方五百里曰卫服，又其外方五百里曰蛮服，又其外方五百里曰夷服，又其外方五百里曰镇服，又其外方五百里曰藩服。”《荀子·正论》：“故诸夏之国同服同仪，蛮、夷、戎、狄之国同服不同制。封内甸服，封外侯服，侯卫宾服，蛮夷要服，戎狄荒服。甸服者祭，侯服者祀，宾服者享，要服者贡，荒服者终王。”五服、九服之说都把周王朝统辖的天下划分为三个层次：畿内、诸夏和夷狄，其意义与《春秋》的“内其国而外诸夏、内诸夏而外夷狄”基本一致，是华夷之辨原则在地域观念上的体现。

    在周人的观念中，华夷之辨主要表现在华夷之间在语言、习俗与经济生活等方面的区别。《论语·宪问》：“微管仲，吾其被发左衽矣。”孔子所说的“被发左衽”，即是夷狄的风俗。《礼记·王制》：“中国、夷、蛮、戎、狄，皆有安居、和味、宜服、利用、备器。五方之民，言语不通，嗜欲不同。”《礼记·檀弓》：“有直情而径行者，戎狄之道也，礼道则不然。”可见，周人主要以礼仪风俗作为区分华夷的标准。

    应当说明的是，华夷之辨的民族认同是双向的。《左传·襄公十四年》：“我诸戎饮食衣服，不与华同，贽币不通，言语不达。”《战国策·赵策二》：“远方之所观赴也，蛮夷之所义行也。”《史记·楚世家》载西周晚年楚国国君熊渠宣称：“我蛮夷也，不与中国之号谥。”至春秋中叶，楚武王仍云“我，蛮夷也”（《史记·楚世家》）。《史记·仲尼弟子列传》载子贡出使越国，越王亲往郊迎，曰：“此蛮夷之国，大夫何以俨然辱而临之？”《史记·秦本纪》载秦穆公曰：“中国以礼乐诗书法度为政，然尚时乱，今戎夷无此，何以为治？”这些例证都说明，西周、春秋时期中原地区之外的其他国家和民族，对华夷之别同样也是认同的。

    在周人的民族观念中，与华夷之辨相辅相成的，是华夷一统思想。《左传·昭公二十三年》：“古者，天子守在四夷。”《会笺》：“守在四夷，亦言其和柔四夷以为诸夏之卫也。”《左传·昭公九年》：“我自夏以后稷，魏、骀、芮、岐、毕，吾西土也；及武王克商，蒲姑、商、奄，吾东土也；巴、濮、楚、邓，吾南土也；肃慎、燕亳，吾北土也。”可见在周人的观念中，王朝的疆域包括周边各族在内。前文所引周代文献中的五服、九服之说，也无不把戎狄蛮夷包括在周王朝统辖的范围之内，诚如陈连开先生所言：“对于《禹贡》、《职方》中‘五服’、‘九服’的名称、内容，古今学者多有诠释，各家说法不尽相同，但都表达了以天子为首，以王畿为中心，包括华夷的统一思想。”

    《春秋》与《公羊传》的“内华夏、外夷狄”思想，与西周以华夏族为主体民族，华夷共存、内外有别的民族一统思想是一脉相承的。这种以“内华夏”为宗旨的民族一统思想，源于周初封建所造成的华夏族形成的历史实际以及周代社会对华夷之辨的认同观念。

    四、《公羊》“崇礼”思想源于西周尊尚礼乐的文化认同

    制礼作乐，是西周王朝统治集团为巩固政权而采取的一项重要措施。西周礼乐制度建设的成就，导致了尊尚礼乐的文化认同观念的形成。《公羊传》以“崇礼”为中心的文化一统思想，源于西周制礼作乐的历史实际以及周代社会尊尚礼乐的文化认同观念。

    关于周公制礼作乐，先秦文献中有明确的记载。《左传·文公十八年》：“先君周公制周礼曰：则以观德，德以处事，事以度功，功以食民。”《左传·哀公十一年》：“且子季孙若欲行而法，则周公之典在。”除《左传》外，《尚书·洛诰》还记载了成王对周公说：“四方迪乱，未定于宗礼，亦未克敉公功。”对制礼作乐的意义表示高度的重视。

    事实上，周公的制礼作乐，还处于周礼的草创阶段。经过后来数代君臣的补充和完善，西周中期以后周礼才渐趋完备。《诗经》中多次出现“以洽百礼”的诗句，反映了当时礼制的繁芜。据刘雨先生研究，西周金文材料所载的礼制，“周礼多数是在穆王前后方始完备”。詹子庆先生也认为，“从金文材料反映出，西周中期以后，各种礼仪制度化，如世官制、宗法分封制、昭穆制、册命制、舆服制等都有了定式”。因此，西周礼乐的系统化、完备化和程式化，是在西周中、后期才得以完成的。

    西周制礼作乐，对夏、殷之礼有继承，也有革新。《论语·八佾》：“周监于二代，郁郁乎文哉，吾从周。”《论语·为政》又说：“殷因于夏礼，所损益可知也；周因于殷礼，所损益可知也。”周礼与殷礼的不同之处，是殷礼亲亲，周礼尊尊。《史记·梁孝王世家》褚少孙补：“殷道亲亲，周道尊尊，其义一也。”“亲亲”与“尊尊”是殷周社会的两条重要政治原则。“亲亲”指血缘关系。“尊尊”指阶级关系。从“殷道亲亲”到“周道尊尊”的变化过程，“也就是阶级关系逐步支配并改造了血缘关系的过程”。因此，周礼最显著的特征体现为日益严密的等级制度，即《礼记·中庸》所说的：“亲亲之杀，尊贤之等，礼所生也。”

    西周制礼作乐，还赋予了周礼“德”的内容。周代的各种典礼都蕴含一定的道德意义，即所谓的“礼义”。《礼记·经解》：“故朝觐之礼，所以明君臣之义也；聘问之礼，所以使诸侯相尊敬也；丧祭之礼，所以明臣子之恩也。乡饮酒之礼，所以明长幼之序也；昏姻之礼，所以明男女之别也。”因此，周礼兼具政治统治和道德教化的功能，对维护和巩固西周政权发挥了重要作用。王国维先生说：“古之所谓国家者，非徒政治之枢机，亦道德之枢机也。……是故天子诸侯卿大夫士者，民之表也。制度典礼者，道德之器也。周人为政之精髓，实存于此。”

    西周封建诸侯和分封制度的实行，使周礼首先得到了受封诸侯国的认同。在分封制度下，各级政权之间的等级隶属关系集中反映在周王室制定的礼乐制度上。《左传·庄公十八年》：“名位不同，礼亦异数。”《左传·襄公二十六年》：“自上以下，隆杀以两，礼也。”周代的等级制度，在各种礼制中都有体现。如《国语·楚语下》：“天子举以大牢，祀以会；诸侯举以特牛，祀以太牢；卿举以少牢，祀以特牛；大夫举以特牲，祀以少牢；士食鱼炙，祀以特牲；庶人食菜，祀以鱼。”是为祭祀的等差；《礼记·礼器》：“天子七庙，诸侯五，大夫三，士一。”是为宗庙的等差；《周礼·小胥》：“正乐县之位，王宫县，诸侯轩县，卿大夫判县，士特县。”是为乐舞的等差；《周礼·大宗伯》：“以玉作六瑞，以等邦国：王执镇圭，公执桓圭，侯执信圭，伯执躬圭，子执谷璧，男执蒲璧。”是为命圭的等差；《周礼·典命》：“掌诸侯之五仪……上公九命为伯，其国家、宫室、车旗、衣服、礼仪皆以九为节。侯伯七命，其国家、宫室、车旗、衣服、礼仪皆以七为节。子男五命，其国家、宫室、车旗、衣服、礼仪皆以五为节。”是为不同等级的诸侯在宫室、车旗、衣服、礼仪等方面的等差。当然，《周礼》、《礼记》所提供的史料，有的要作具体分析，但绝大部分史料的来源是有根据的，可作为了解周礼的等级制度的参考资料。西周时期，受封诸侯国遵行周礼，既是诸侯国对其与周王室之间等级隶属的一种确认，也是受封诸侯国对周礼文化的一种认同。

    西周受封诸侯前往边陲建立邦国，带去了祝宗卜史等官吏、周之典籍以及各种天子赏赐的礼器等，也就把先进的周礼文化传播到了那个地区。西周诸侯受封建国后，又确立了以礼治国的方针，大力地推广周礼文化。周代文化以各诸侯国为中心，向四周辐射，使周礼逐渐得到了各国土著居民和周边民族的认同。如：

    鲁国原为东夷族的聚居区，东夷风俗盛行。鲁公伯禽受封之后，征服了徐戎、淮夷各族，“淮夷蛮貊，及彼南夷，莫不率从”（《诗·鲁颂·宫》）。同时“变其俗，革其礼，丧三年然后除之”（《史记·鲁周公世家》），对东夷风俗进行了改革，推行三年之丧等周礼。后来，被征服的东夷各族逐渐认同周礼文化，加速了东夷地区华夏化的进程。春秋时期，鲁国是“犹秉周礼”的礼仪之邦，后来成了儒家的发源地。

    齐国是在薄姑氏旧地上分封的国家，也处于东夷族的包围之中。太公至国，“修政，因其俗，简其礼”（《史记·齐太公世家》），因地制宜地推行周礼。春秋时期，齐桓公在建立霸业的过程中，“招携以礼，怀远以德”（《左传·僖公七年》），以周礼怀柔周边小国，周礼文化得到进一步传播。春秋后期齐相晏婴，原为“莱之夷维人也”（《史记·管晏列传》），却提出“礼之可以为国也久矣，与天地并”（《左传·昭公二十六年》）的主张，继承了齐人以礼治国的传统。经过几代人的努力，齐国成了“冠带衣履天下”（《汉书·地理志》）的文明大国。

    燕国原为商的势力范围，有山戎、孤竹、秽貊等族散居其地。燕国受封后，“修召公之法”（《史记·燕召公世家》），积极推广周礼文化，使周文化与当地的土著文化相互交融。1975年发现的昌平白浮墓，年代约在西周中期，墓主人为臣属于燕国的异族首领之一。“墓主的着装、佩戴的兵器遵循着本民族的习惯，而使用的青铜礼器和埋葬习俗已纳入西周燕国的轨道。”这反映出周礼文化与燕地土著文化融合的情形。春秋战国时期，周礼文化进一步传播到东北地区。《后汉书·东夷列传》：“东夷率皆土著，喜饮酒歌舞，或冠弁衣锦，器用俎豆。所谓中国失礼，求之四夷者也。”当地的民族文化，已融入了周礼文化的因素。

    晋国所封的唐地，“戎狄之民实环之”（《国语·晋语二》）。唐叔虞受封时，周成王令他“启以夏政，疆以戎索”（《左传·定公四年》）。春秋时期，随着晋国的对外扩张，周礼文化也向外辐射，对周边民族产生了深刻影响。晋卿狐偃原为狄族出身，但从其思想来看，他已经完全华夏化了。他倡导以礼教民，在城濮之战前，向晋文公陈述“民未知义”、“民未知信”、“民未知礼”（《左传·僖公二十七年》），强调周礼的基本精神。《左传·襄公十四年》载，戎子驹支面对范宣子的指责，义正词严地用历史事实驳斥晋国执政，最后赋《诗·小雅·青蝇》而退，大有中原饱学之士的风度。春秋后期，晋国周边的戎狄蛮夷基本融入了华夏族，这种民族融合是在“礼”的认同基础上才得以实现的。

    其他如楚、秦、吴、越等国，虽一度被视为蛮夷之邦，但后来逐渐接受了中原文化，也陆续加入了华夏的行列。这些国家都有独特的地域文化，不过始终都受到了周礼文化的影响。如楚大夫申叔时教太子诗、书、礼、乐及春秋、世、令、语、故志、训典等（《国语·楚语上》），与中原各国贵族教育的内容基本一致。吴国的公子季札受聘至鲁，“请观于周乐”，听乐工每奏一曲，都能逐一评论（《左传·襄公二十九年》），显示了很高的周文化修养。类似深谙周礼的人物，在秦、越亦不乏其例。这表明，周礼文化已传播到了楚、秦、吴、越等国，并逐渐得到了上述诸国的认同。

    西周时期尊尚礼乐的文化认同，使周礼文化在西周的文化格局中，成为了主体的、原始的、根本的“一”，能够统合其他的“多”（地域文化）而为一体，形成西周时期的文化一统格局。而文化一统又是促成政治一统的黏合剂，也是促进民族融合的催化剂。《春秋》与《公羊传》以崇礼为中心的文化一统思想，与周代尊尚礼乐的文化认同是一脉相承的。这种以崇礼为中心的文化一统思想，源于西周制礼作乐的历史实际以及周代社会尊尚礼乐的文化认同观念。

    东周以降，西周“礼乐征伐自天子出”的一统局面已被“礼乐征伐自诸侯出”所取代。但是，在思想上对“一统”的认同，仍在很大程度上支配着东周时期人们对历史走向和国家前途的认识，是人们重建统一王朝的精神动力。春秋大国争霸，仍以“尊王攘夷”为旗帜，藉天子的名义维护自己势力范围内的一统秩序。战国时期，“上无天子，下无方伯，力功争强，胜者为右”，重建统一王朝已成为历史发展的大势所趋。当时统治者梦寐以求和思想家大声疾呼的，无不是实现天下的统一。

    由于历史形势发生了变化，战国时期的大一统观念有了新的内容。《史记·李斯列传》：“今诸侯服秦，譬若郡县。夫以秦之强，大王之贤，由灶上骚除，足以灭诸侯，成帝业，为天下一统，此万世之一时也。”李斯所说的“天下一统”，实际上是“大统一”，即以武力兼并为手段，建立以郡县制为基础的中央集权式的统一国家。秦灭六国，建立了空前统一的大秦帝国。从此，中国古代的大一统思想进入了一个新的阶段。

    在中国历史上，自西周王朝以后，曾经历了春秋战国、魏晋南北朝、宋辽金西夏等几个分裂的时期，但始终没有像欧洲那样，形成多个独立的民族国家，而是在经过分裂、对峙和融合后，又出现了秦汉、隋唐、元明清等崭新的统一王朝，使中国社会一步一步地跨上更高的台阶。“一统”始终是中国历史发展的常态，而造就中国一统常态的重要原因之一，正是根植于中国传统文化中的大一统思想和精神。因此，弄清大一统思想的渊源及其历史发展，对我们深入理解在中国延绵两三千年之久、并对中华民族的历史产生过巨大影响的大一统思想，是十分必要的。

    本文原载《文史哲》2013年第4期

  • 陈伟：书于竹木：简牍文化及其载述的国家信史

    简牍及其周边

    简牍是指用于书写的竹、木片和写在竹、木片上的文献。从许慎《说文解字》开始，历代学者提出多种解释，大致认为简用竹制作，形状细长，也称牒、札；牍用木制作，比较宽厚，也称方、板。岳麓书院藏秦简中的令条规定：上呈皇帝的文书“对”（答问）、“请”（请示）、“奏”（报告），采用牍的时候，一牍不超过五行字（“用牍者，一牍毋过五行”）。三行、四行、五行牍的具体宽度，分别约等于3.45、3.83、4.34厘米。又说，“牍厚毋下十分寸一（约0.23厘米），二行牒厚毋下十五分寸一（约0.15厘米）”。综合起来看，容纳文字是在三行以上还是在两行以下，是牍与牒（也就是简）的主要区别。牍可以书写三至五行，比较宽厚；牒或曰简只能书写一或二行，比较窄而薄。这是对呈报皇帝文书的特别要求，但对了解一般简牍的状况也有参考意义。

    近年的发现显示，两行书写的简多用木制，但也有竹制；单行书写的简多为竹制，但也有木制。牍多用木制，但湖北、湖南也出土了竹制的牍。因而，简单地说“竹简”“木牍”，其实不够准确。

    单行和双行书写的简，往往用绳线连系成册以承载长篇文献。《史记·留侯世家》说黄石公“出一编书”，《汉书·诸葛丰传》说“编书其罪”，就涉及这一情形。这也是后世书籍观念中的编（也写作“篇”）、册（也写作“策”）的源头。牍的书写面比较大，可以单独承载不太长的文献，早先认为不存在编连的问题。不过，近期一再发现内容相关但形态各异的文书、簿籍编连成册。现在看来，只是典籍类文献才由形制相同的简书写编卷，而形态各异的文书簿籍造册归档时，并非如此规整。

    简牍上的文字，绝大多数是用毛笔蘸墨写成，偶尔也有红色字迹，即所谓“丹书”。古书中有所谓“漆书”，指的应是墨书。笔、墨、砚、刀，是简牍时代的文房四宝。写错的字，可用刀刮去再写。《史记·孔子世家》说：“至于为《春秋》，笔则笔，削则削，子夏之徒不能赞一辞。”当时处理文案的官员，因而被称作“刀笔吏”。《汉书·萧何曹参传》就说“萧何、曹参皆起秦刀笔吏”。

    《尚书·多士》记“惟殷先人有册有典”。甲骨文已有“册”字。由于“册”的字形类似简册，有学者推测商代已使用简牍。《诗经·小雅·出车》咏叹远征的军人“岂不怀归，畏此简书”。《左传》襄公二十五年记齐大臣崔杼作乱时，“南史氏闻大史尽死，执简以往”；襄公二十七年宋大夫向戌将赏赐文书拿给子罕看，子罕不以为然，“削而投之”。这些是西周、春秋时使用竹简的可靠记载。

    我国现代意义上的简牍发现，始于20世纪初，其后层出不穷，出土地点从西北地区扩展到大多数省份，迄今已发现200多批，总数超过30万枚、300万字。这些简牍的年代主要是战国中期至秦汉魏晋，最早一例是在公元前433年或稍晚入葬的随州曾侯乙墓竹简。春秋以前的简牍由于年代久远不易保存，加之埋藏条件的原因，目前尚未能得见。

    目前，已多次发现西汉纸张的遗存，居延、敦煌、放马滩等地所见的纸还带有文字或地图，显然是用于书写。不过，在东晋末年之前，简牍仍然是主要书写载体。《初学记》卷21“纸七”录《桓玄伪事》称：“古无纸故用简，非主于敬也。今诸用简者，皆以黄纸代之。”这是纸张取代简牍成为官方书写载体的标志。

    简牍的取材、制作、书写，都比较方便。《论衡·量知》就说：“截竹为筒，破以为牒，加笔墨之迹乃成文字。大者为经，小者为传记。断木为椠，析之为板，力加刮削，乃成奏牍。”《汉书·路温舒传》记载路温舒小时候放羊，自己制作木简，练习书写。可见简牍的便易性降低了识字、教育的门槛。商代、西周，学在官府，知识圈狭小，文献的种类、篇幅也有限，简牍的优势不容易发挥。春秋以降，私学勃兴，著述蜂起。战国时各国相继变法，建立以郡县制、官僚制为基础的新兴国家，文书、律令的行用骤然增长，简牍真正有了用武之地。在这个意义上可以不夸张地说，在我国春秋、战国、秦汉时期的政治发展和文化繁荣中，简牍扮演了重要角色。

    由于竹木带有天然纹路，并便于刻齿、挖槽，还可封泥、钤印，因而简牍还可衍生为具有保密、防伪功能的券、符、传、检、署等物件，在公私事务中发挥特别作用。

    （1）检、署

    署是在往来文书、信函上写明收件方以及传递方式的木片，同时也对文件内容起到屏蔽作用，类似于今天的信封。署与文件捆紧后，在捆扎处可敷设胶泥，再盖上印章，不开封不能看到里面的内容。

    检是封缄文书、物品的物件。《急就篇》卷三：“简札检署椠牍家。”颜师古注：“检之言禁也，削木施于物上，所以禁闭之，使不得辄开露也。”检有多种式样，但都带有封泥、钤印的凹槽。用检的文书，比只用署的文书保密效果更好。岳麓秦简“卒令丙三”说：“书当以邮行，为检令高可以旁见印章，坚约之，书检上应署，令并负以疾走。不从令，赀一甲。”这提示我们，检用于以邮行的文书，而不用于其他方式传递的文书。

    （2）券、符、传

    券是财务往来的凭据。一式两份或三份（“三辨券”），用同一木板或枝条剖分而成。券上通常有刻齿，用不同形态的齿表示不同数值，与所记载的数字对应，加强券的可靠性。

    符是从事一些特定事务的凭证。通常一式两份，通过“合符”来验证。西北汉简中发现较多出入符。居延汉简65.9长14.6厘米，刻齿在书写面的左侧，释文为：“始元七年闰月甲辰居延与金关为出入六寸符券齿百从第一至千左居官右移金关符合以行事……”表明这款符用于出入金关，一次制作1000套，各套的左符留在官署，右符放在金关。通关者领取左符到金关验符通行。居延汉简65.10刻齿在书写面的左侧，右半残缺，存留的一行文字与65.9相同。最近有学者测试，二者紧密契合，可能是一套符中的左符和右符。

    传是旅行证件。对因公出行者来说，传同时还是接受交通、食宿安排的凭据。云梦睡虎地秦简《法律答问》记：“今咸阳发伪传，弗智（知），即复封传它县，它县亦传其县次，到关而得。”显示传跟公函一样，封缄后由使用者携带，需要时拆开查验。

    从文物到文献

    简牍的出土位置，主要有墓葬、水井、工作或生活遗址。出土简牍的墓葬分布广泛，湖北地区发现最多。云梦睡虎地11号秦墓，1000多枚竹简集中放在棺内。而在大多数墓中，简牍是放置在棺外，比如椁室中。古井中堆积简牍，主要见于湖南。古人工作或生活遗址出土简牍，主要是在西北地区。

    简牍的揭取和保护通常由专业人员负责，在细心提取简牍的同时，还详细记录各个个体之间的相互关系，为后期的缀合、编连提供参照。在完成清洗、脱色后，需要及时拍摄图像，尽可能充分地获取各种信息。

    简牍文献的整理，是尽可能完整、系统地获取简牍中的文献信息，实现简牍从文物到文献的转换。主要工作环节可用以下几个例子说明。

    认字，是把简牍上书写的古代文字辨认出来。利用文字学、古文字学研究成果，简牍上的大多数字，学者可以认读。但也有一些难字需要推敲考订。郭店简中有一个字出现三次，整理者释为“蚄”，很难讲通。其实这个字是《说文》“杀”的古文，在简文中读“杀（shài）”，衰减的意思。《唐虞之道》7号简“孝之杀爱天下之民”，《语丛三》40号简“爱亲则其杀爱人”，是说把对亲人的爱推广给其他人，属于儒家仁爱的观念。《语丛一》103号简“礼不同、不奉（丰）、不杀”，与《礼记·礼器》所记孔子语相同，是这一释读的直接证据。

    断读，相当于标点，是通过阅读中的停顿，反映文章中的意群和脉络，从而正确地领会文意。断读分原则性断读和喜好性断读两种。喜好性断读，是指出于个人习惯，断句或长或短，不求划一。原则性断读，是说当断必断、当连必连，否则就会导致文句不通或使文意产生歧义。

    张家山汉简《二年律令》65～66号简整理本释文：“群盗及亡从群盗，……矫相以为吏，自以为吏以盗，皆磔。”注释说：“矫相，疑指矫扮他人。”简文中，“相以为吏”与“自以为吏”相对，是形容“盗”的两种情形。矫，指假托、诈称，同时修饰这两种情形。因而中间的逗号应改为顿号，读作“矫相以为吏、自以为吏以盗”，是说相互诈称官吏或者自我诈称官吏而进行盗窃。岳麓秦简《学为伪书》案卷中那位叫学的少年犯供述说：他父亲服劳役受欺侮，经常训斥他。“归居室，心不乐，即独挢（矫）自以为五大夫冯毋择子”，伪造书信进行诈骗。这就属于类似表述。

    编连与缀合，是简牍类文献整理的特殊作业。简牍出土时，原有的编绳大多朽断无存，简牍个体还往往开裂破碎。编连与缀合就是在这些情形下，重建业已丢失的、书写在不同简牍个体及其残片上的文本之间的联系和顺序。编连是对不同简牍个体之间顺序的安排。缀合则是针对同一支简牍而言，在简牍断裂之后，重新把残片拼合起来，以恢复原先的完整形态。在这里，简牍物质形态上的拼复与编次，与文本形态上的连接与整合相互依存，融为一体。

    郭店简《语丛一》31号简与97号简，分别书写“礼因人之情而为之”和“即（节）文者也”。整理本把二者分别看待。《礼记·坊记》说：“礼者，因人之情而为之节文，以为民坊者也。”《管子·心术上》说：“礼者因人之情，缘义之理，而为之节文者也。”《礼记·檀弓下》：“辟踊，哀之至也，有算，为之节文也。”相形之下，31号简显然应当与97号简连读，表述儒家对礼的起源的观念（礼基于人的情感并用仪节来调适）。在我提出这一看法的时候，“文”字还没有得到正确释读。而当学者随即释出“文”字后，这两枚简前后相次就更加确定了。

    缀合，是克服简牍破碎化，提升残片文献价值的关键步骤。我们在研撰《里耶秦简牍校释》过程中，把缀合的推进作为工作目标之一。下文引述亭“赀三甲”的木牍，由四个残片拼合后，方可知其大概。

    云梦睡虎地77号墓出土的西汉简牍《质日》，有的年份损坏严重。我们课题组同事用“寸简寸心”相激励，孜孜以求，一点一点地推进。经反复推敲，用8个残片缀合成一枚下半支简（“己酉 戊申道丈田来治籍 丁未将作司空”），并排定到《十一年质日》的2号位，就是集体攻关的一个实例。

    简牍文献记载的国家信史

    早前，因为简牍出土数量不足，并且大多支离破碎，其学术价值一般只说是证史、补史，处于辅助、补充的位置。现在由于资料的快速积累，尤其是有像睡虎地秦汉简这样数量多、保存也比较好的大宗材料，通过适当整理和互勘合校，简牍文献已经在行政与政区制度、律令与司法制度、经济制度、文书制度、算术与医药、风俗习惯等领域的创新性研究中成为主要的资料依据。

    简牍资料在秦郡县制方面提供了较多新知识，这里举三点说明。

    首先，新发现郡名“洞庭”“苍梧”。《史记·秦本纪》记载：“秦王政立二十六年，初并天下为三十六郡，号为始皇帝。”从南朝宋的裴骃开始，学者对三十六郡所指便聚讼不已。1947年，谭其骧先生发表《秦郡新考》，成为权威性意见。然而，秦简牍中有一些全新的发现。秦始皇二十七年的一件文书说：“今洞庭兵输内史，及巴、南郡、苍梧输甲兵……”（里耶秦简16-5）洞庭、苍梧与人们熟悉的巴郡、南郡并列，显然也是秦郡名。秦始皇三十四年的一件文书（里耶秦简8-758）说“苍梧为郡九岁”，表明在秦王政二十五年统一前夕，就已设立苍梧郡。在传世文献中，秦洞庭、苍梧二郡，毫无踪影。

    里耶秦简对洞庭郡及其属县有较多记录，因而可以推定秦洞庭郡其实相当于传世文献中的黔中郡。《汉书·武帝纪》记武帝元鼎六年“遂定越地”，设南海、苍梧等九郡。有学者认为，秦苍梧郡是西汉苍梧郡的前身，位于南岭以南。根据张家山汉简《奏谳书》所录秦案卷等简牍的证据,秦苍梧郡其实相当于传世文献中的长沙郡。

    其次，昭示中央直达基层的管理体制。在郡县制下，国家之于地方，“如身之使臂，臂之使指”，出土简牍使我们领略到这种体制实际运行的精致与效率。

    里耶秦简8-228记载丞相书的传递，从朝廷所在的内史开始，在传达至各县的同时，还传给南郡，南郡又传给洞庭，从而使这份文书迅速传播到郡县。里耶秦简9-2283是洞庭太守避免征发徭役的指令，从大概是郡治所在的新武陵分四条路线（“别四道”）传达给各县。迁陵县收到文书后，一面向上一站酉阳县回报，一面安排县内各官署传达：“迁陵丞欧敢告尉：告乡、司空、仓主听书从事。尉别书都乡、司空，司空传仓，都乡别启陵、贰春。皆勿留脱。”“别书”指另行抄录传递，在当时应是文书传播中的有效方式。

    最后，展现不同郡县间的行政、经济联系。秦代不同郡县之间可能有相当密切的联系。前面引述里耶秦简属于苍梧郡的指令，因为与洞庭各县有关，传达到洞庭郡迁陵县各乡。里耶秦简8-657则是由于琅邪尉的治所迁到即墨，琅邪郡通报各地。

    里耶秦简中常常出现的“校券”，是不同郡县间钱物往来的凭据。13-300记载迁陵县十四匹传马经过雉县（属南阳郡）时，借用了食料。雉县出具“稗校券”，要求迁陵接受“移计”，“署计年、名”反馈给雉县。这意味着，迁陵不需要交付钱物，而是借助“计”的形式确认债务，再通过中央财政平账。里耶秦简所记一段相关内容颇有故事性。亭来自僰道（属犍为郡），在迁陵担任“冗佐”（一种低层吏员）期间犯事，“赀三甲”，计4032钱。亭自称家里有能力赔偿。迁陵县出具校券，请僰道县索取。结果亭的妻子胥亡说：“贫，弗能入。”要求让亭在迁陵作劳役抵偿。于是迁陵要求僰道退还校券。

    这类事例显示，秦郡县制之下，除了中央与地方的纵向关系之外，地方郡县之间还存在密切的横向联系。这降低了各地政府的运行成本，增强了国家的凝聚力，也给民众带来一些便利。

    文书在秦汉国家治理中，发挥着重要作用。

    睡虎地秦简《秦律十八种》是秦统一之前的律典。其中在多种场合强调“以书”，显示当时已形成文书行政的规范。如《田律》要求“辄以书言”春雨和庄稼抽穗的情况；《金布律》要求官府输送财物时，“以书告其出计之年”；又要求在废旧公物需要及时处置的场合，“以书”呈报；《内史杂律》规定需要请示时，“必以书，毋口请”。

    里耶秦简是秦统一之后洞庭郡迁陵县的档案。较多文书写明“听书从事”，或者提出“书到时”如何运作的具体要求。

    民间的重要事务，如结婚、遗嘱、牛马奴隶等交易，也需要由官府用文书确认。岳麓秦简《识劫婉》案卷中，女主人翁婉，原本是一位叫沛的富豪的妾。沛的妻子在世时，婉已为沛生下两个孩子。沛的妻子去世后，沛免除婉妾的身份，成为庶人，又生了两个孩子。婉自述说，沛把她免为庶人后，娶她为妻，并让她参加宗族、乡里的活动。然而乡署的官员表示：沛免婉为庶人时，在户籍上登记“免妾”。但后来娶婉为妻，并没有报告，婉在户籍上的身份还是“免妾”。

    律令是秦汉帝国建立、运行的重要制度支撑。以睡虎地秦律发端，近五十年来，秦至西汉早期的律令简册层出不穷，蔚为大观。

    对于秦汉律的整体认识，学界颇有歧异，或比较笼统地称之为“律典”，或以为只有一条一条制定的单行律令，而不存在国家颁布的统一法典。

    较早出土的睡虎地秦律、张家山汉简《二年律令》，均已呈现出篇章分明的结构。云梦睡虎地汉律、荆州胡家草场汉律和益阳兔子山汉律目录大致相同，进一步展示出集篇为卷、两卷并存的格局。兔子山律目分为“狱律”“旁律”两部分，其中“狱律”包含告、盗、贼、囚、亡等十七篇，“旁律”包含田、户、仓、金布、市贩等二十七篇。当时的律分“罪名之制”和“事律”两类，大抵“罪名之制”是对犯罪行为的处罚规定，类似于刑事法律；“事律”是对违反制度行为的处罚规定，类似于行政法规。西汉早期律典中，“旁律”诸篇均属事律；“狱律”虽然以“罪名之制”诸篇为主，但却夹杂几篇“事律”（效、关市、厩律等）。这种安排很不好理解，或许与萧何制定“律九章”的历史有关。

    虽然律篇、律条的增删修订不断发生，但在一定时期内，全国存在一个统一的律典。这可以从几个方面来看。

    第一，在睡虎地秦律、里耶秦简和睡虎地汉简中，一再出现“雠律令”的记载。可见律令一有变动，就立即在全国组织校勘，保持同步。

    第二，秦汉时实行奏谳制度，重要案件向上级报告，疑难之狱请上级裁断。向上呈报时必须“具傅所以当者律令”（《岳麓书院藏秦简〔伍〕》66），把判决依据的律令一一附录在判决之后。可见全国上下遵循同一律令，中央立法机构掌握最终解释权。

    第三，张家山汉简《功令》规定各县道狱史在升任郡治狱卒史前，需要集中到中央司法部门（廷尉）参加“律令有罪名者”等内容的考试。考试作答、评分必定要有标准答案，这也显示统一律典的存在。

    第四，某些律篇、律条的变更，会带来律典的全面修订。例如张家山336号墓出土的《汉律十六章》，较多律篇与《二年律令》相同，但律条多有增删和补充，不再出现《收律》，相关律条皆删去“收”和“收孥相坐”的刑罚。这是文帝元年“除收帑诸相坐律令”的结果。胡家草场汉律是汉文帝十三年刑制改革后的律典，与此前的张家山《汉律十六章》和睡虎地汉律相比，刑罚制度判然有别。这证明律典中各篇各条存在密切关联，构成一个有机整体。

    刘邦军至咸阳，萧何“独先入收秦丞相、御史律令图书藏之”，并“作律九章”，奠定汉承秦制的基础。《史记·曹相国世家》记曹参去世后，民众歌颂说：“萧何为法，顜若画一。曹参代之，守而勿失。”司马贞《索隐》解释“顜”字说：“训直，又训明，言法明直若画一也。”《汉书·曹参传》写作“讲”，颜师古注：“讲，和也。画一，言整齐也。”“画一”之歌反映了当时人对律令整齐划一的真实感受。

    秦汉时期法的主要形态有律、令两种。令的资料目前公布的还不多，姑且不论。律就其具备的基本特征而言，称之为“律典”或者“早期律典”是适宜的。

    本文节编自《光明日报》（ 2025年01月04日 10版）

  • 颜荻：秘索思与逻各斯的动力学：古希腊文明精神溯源

    引言 

    古往今来，任何一部文明史都是不同文明互鉴的历史。深刻认识文明互鉴的实践，是一种特有的文明自觉。文明研究有三个关键议题：其一，文明的起源性构造及根源性影响；其二，文明发展的动力原则及生成逻辑；其三，文明对自身历史的认识及系统化表达。无论在中国还是西方，三个议题都贯穿于文明发展的历程之中。可以说，任何一个角度的文明研究都应怀有这三个部分的问题意识并予以展开。 

    就西方文明而言，几乎所有起源性问题都可追溯至古希腊。古希腊作为开端，其始源性构造奠定了西方文明的最初样态。在始源性构造中，有一个议题十分重要，即“秘索思（mythos）与逻各斯（logos）”。它不仅深刻关涉上述三个关键的文明研究内容，且对古希腊文明乃至整个西方文明形成奠基性影响。 

     Mythos一般指“语词”“神话”“故事”与“虚构的言辞”，logos则指“理性”“秩序”“逻辑”和“规则”。二者首先从古希腊历史的发端处，以语言这一最基本的文明形式塑造了古希腊人对自身、社会、世界乃至宇宙的根本想象，同时作为两种不同的思维模式，其动态互动构成古希腊文明乃至西方文明的基本生成逻辑。传统研究将此互动过程经典地描述为“从mythos到logos的转变”，其发展路向通常被认为最终打开了西方理性主义与逻各斯中心主义（logocentrism）的大门，因而对近代以来的启蒙运动与科学主义兴起，乃至现代性的产生与发展形成深远影响。与之相应，在这一过程中西方文明所逐渐形成的对自身历史的认识与系统表达，可称之为历史书写。在logos成为一种权威表达方式时，西方的历史叙事乃至历史观也随之逻各斯化。历史越来越被看作一个理性发展的过程，以至到近代，这一观念进一步与进化论和目的论关联，发展出一系列西方文明对自身价值的评估与判断。 

    因此，mythos与logos一向是西方古典学与相关学科研究的经典课题。无论是围绕mythos与logos的词源学经典讨论，还是从文学、哲学、史学等出发的文本意义考察，均成果丰硕。基于“从mythos到logos转向”的基本框架，相关研究从不同侧面不断巩固“logos对mythos的胜利”这一主流观点，从而形成对mythos与logos关系及其奠基性意义的网络式理解。 

     然而，“logos的胜利”却无法涵盖所有现象。在人类似乎进入由理性、秩序、逻辑与规则构成的科学、中立、通约化的普遍历史世界时，mythos一直作为动力隐隐存在着。自19世纪开始，从“原始思维”到“理性文明”的表述，同时受到不同学科的严厉批评与审查。其中结构主义人类学强调神话作为“深层心智”绝非“野蛮的初级思维”，仪式/功能主义社会学对神话进行了社会功能阐释，神话哲学则努力在哲学中直接复兴神话的意义。这表明，mythos与logos的内在蕴涵显然比既有的线性阐释模式复杂得多。究其根本，在于mythos与logos间相互勾连、冲突与纠缠的状态，在其出现之初便已开始。二者在起源时所构成的此消彼长的动力学原则对西方文明发挥着根本而持续的作用。因此，要厘清整个西方文明在思想史层面的复杂发展脉络，就需回到始源，重新探讨mythos与logos的发生史。从这一视角出发，不仅能观察到西方文明所深含的内在力量，还能在此力量所具有的开放性与包容性中理解西方文明不断塑造与再造的过程，直至通解当下现代性所面临的复杂问题。 

    一、“颠倒”的秘索思与逻各斯 

    从最早的古希腊文献来看，mythos与logos最初即一组有关“言辞”的对立统一的概念。不过，在古希腊早期历史中，mythos与logos的意涵与现在所熟知的意义恰恰相反。早有学者如布鲁斯·林肯指出，logos在古风时期的语境中，所指涉的绝非后人所理解的“理性”与“真实性”，而是与“欺骗”“错误”和“谎言”相关联；反而，现在看似表达“虚构”与“假象”之意的mythos被认为具有更高的真理性甚至神圣的权威性，从而，在mythos与logos的起源之初，两者所表之意，实际正是后来意义的颠倒。 

    赫西俄德与荷马为此提供了经典的例证。例如，在赫西俄德《劳作与时日》中，几乎所有的logos都与虚构和谎言相关，诗人不仅以logos而非我们通常认为的mythos来指代“五代神话”这个虚构的故事，而且特地选用形容词haimulios（欺骗的）来对不同语境中的logos进行修饰。而在《奥德赛》中，足智多谋的奥德修斯，在与佩涅罗佩相认前夕，也讲述了（legein）（<logos）许多谎言（polla pseudea），那些谎言就像真的一样，令王后信任与哭泣。 

    在布鲁斯·林肯所列举的所有相关例证中，可以发现，“秘索思与逻各斯之争”正是始于这两者所包含的积极与消极意义的对立。而伴随着两极的分化，这两个词汇又被进一步赋予相应的性别化特质，从而，在譬喻性的层面上被完全对立。由于logos总带有欺骗与谎言的负面性质，因此，在古希腊整体的厌女（misogyny）语境下，自然与“女性化”的特性相关联。潘多拉“迷人的logoi，以及诡诈的性格”就是典型。而与logos相反，mythos则具有“男性化”的特质。一位英雄的理想就是成为一位“实践的行动者与mythos的言说者”，由此，mythos被显现为一种与英雄精神相关的特质，并时刻与这一男性化的、公共的、强大的力量正向关联。 

    Mythos与logos性别化的对立所反映的不仅是两性本身的问题，而且是在一个更广泛的社会文化意义上，将两者带向了不同的存在之域。与“男性”相关的mythos，其背后意味着“权力”“权威”以及由此而建立的“神圣性”与“真理性”，而logos则恰恰相反。在《荷马史诗》中，当阿伽门农面对克律塞斯（Chryses）的祈求要在集会中力排众议严词拒绝时，他必须使用mythos。因为，越是男性化的、越强大的人，越拥有言说mythos的资格与能力，反之，则被认为应当在mythos的领域保持沉默。与logos相关联的女人便更没有言说mythos的权利。特勒马库斯就明确告诉母亲：“你还是回到里屋，操持你自己的事……mythos是男人关心的事——所有男人，尤其是我，因为我是家中的掌权者。” 

     正如理查德·马丁所指出的，mythos总是一种力量之语，它是一个拥有权力或权威的人所说出的强权化的甚至粗暴的言辞。这种极端男性化的特质与史诗尤其荷马精神高度契合。战争作为英雄荣誉的来源，成为史诗必然歌颂的对象，而正是此“强有力”的话语，不仅标志着英雄取得胜利的强势力量，而且，连同英雄的行动一起，构成了诗歌中那些值得传颂的语言与故事。英雄之诗，从根本上而言，就是力量之诗。换言之，关于英雄的mythos，就是力量的mythos。它光明、正大、直接、不加掩饰，与欺骗、阴暗、迂回的logos形成强烈反差，由此，前者在英雄世界的价值体系中，在对伟大的英雄精神的渴望与追求下，被崇尚为一种揭示英雄本质的、本真性的语言形式，一种与“真实”所关联的“动人”的话语结构。在这个近乎二元对立的价值判别中，mythos——无论是言辞本身，还是其所构成的叙事——便拥有了绝对的权威性与崇高性，甚至与神圣世界关联起来。 

    在此，我们必然会想起赫西俄德《神谱》中缪斯女神谈论mythos的经典段落： 

    女神们首先向我讲出这些话语（mythos）， 

    那些奥林波斯的缪斯，持大盾的宙斯的女儿们： 

    “荒野的牧人啊，你这可鄙的家伙，只知吃喝， 

    我们知道如何讲述（legein）谎言如真实一般， 

    也知道如何如我们所愿唱诵（gēruein）真实（alēthēs）。” 

    神圣之音，mythos，在缪斯作为神明的神圣权威中展开。她们对诗人说话，诗人聆听她们的语词。她们告诫诗人，女神可以讲述谎言，也能够唱颂真实，她们凭自己的意愿，在谎言与真话之间作出选择。若是谎言，则是将其讲出（legein），而若是真理，她们则会为之唱颂（gēruein）。“说”与“唱”标定了谎言与真实的界限，而女神们赐给赫西俄德的是一首“动人的歌”，所以诗人笃信，他从缪斯处所继承的，必然是神明们所歌颂的真实。在神圣的启示下，诗歌作为一种唱颂/言说形式，便接近了最高的真实性与永恒性，它从神圣世界获得了权威的力量，从而在世俗世界中，自然而然成为一种富有权威的真实性表达。 

    在神圣世界的关照下，诗人通过诗歌所唱颂的史诗、故事和神明谱系便与“真实”和“真理”深度勾连。Mythos成为一种罗伯特·福勒所谓的元诗学（metapoetic），一种先验的、不可辩驳的真理，而其所关联的所有语词、言说与话语都与虚假的、错误的、荒谬的logos世界相区离。而当这些“真实的”叙说在世代吟游诗人的口耳相传中成为古希腊的记忆时，mythos所构成的具有“真实性”的“历史”出现了。而这种深嵌于神圣权威之“真理”的真实性，已经超越了历史实证主义意义上的真实，在一种超历史的意义上成为最本真的存在。荷马与赫西俄德，也由此成为所有古希腊人的先师，其mythos之言说，构成了古希腊共同体“真知”的基底，从而塑造着古希腊人对其自身精神与历史意义的根本认识。 

    Mythos与logos在“真”与“假”的二元对立中展开了最初的对话：mythos表达真实的、男性化的、阳刚的、权威性的、公共的、动人的话语体系，logos则表达虚假的、女性化的、阴柔的、边缘性的、私人的、充满冲突的言说。从荷马与赫西俄德到公元前6世纪晚期，这一两极化的表达占据着古希腊世界的主流，mythos也因其所拥有的真理性与权威地位而一直被奉为圭臬。而当mythos的真实性开始受到质疑时，这一图景便开始转变。从希罗多德与前苏格拉底哲人，到修昔底德与智术师群体，最终到柏拉图，mythos逐渐被质疑为不可知的、不真实的、非权威性的话语，而logos则越来越被尊崇为可知的、可控的乃至权威的言说。如此转变使得mythos与logos两者发生结构性倒转，此倒转将影响西方文明对两者意义与关系的根本判摄。而mythos与logos之变是一个逐步发生的漫长过程。 

    二、被“悬置”的秘索思 

    对传统mythos意义的“反叛”，现存文献最早可以追溯至公元前6世纪晚期爱奥尼亚（Ionian）的阿那克里翁（Anakreon）。尽管阿那克里翁本人是一位抒情诗人，但他对mythos的使用却已颇为大胆与前卫。在其残篇中，最具代表性的例子是他在谈及人们反抗萨摩斯的（Samos）僭主波吕克拉特斯（Polykrates）时，用复数mythiētai（说mythos之人）来指涉那些反叛的领袖们。由此，mythos被阿那克里翁纳入政治行动的语境，在动乱的煽动性言辞下，政治领袖所言之mythos就不再是拥有神圣权威的史诗式话语，而是俗化为被世俗政治所利用的“工具与武器”。 

    无论阿那克里翁是否受同时代爱奥尼亚学派（Ionian School）的影响，他作为抒情诗人对mythos意义的创新性用法都可以被视作一个具有标志性意义的节点：当mythos不再与神圣世界确切关联而可以被人事所利用时，这样的言说本身是否还具有美德与权威就被打上了一个问号。这意味着，mythos从前所具有的天然的真理性受到质疑，进而受到优劣评判。在批评与赞扬的表述下，“好的”mythos就变成了一个被竞相争夺的对象，而“坏的”mythos则受到贬斥。这正是阿那克里翁之后的几十年所蔓延开来的景象。 

    诗人品达就常对mythos进行优劣之分，他会批评“有些人所说的mythoi……隐含着谎言和欺骗”而捍卫自己mythos的优越性，将其诗歌视为一种aretai（美德）的表达。在对自我与他人的扬抑之中，品达不断为自身的诗歌立法，以赢得诗人的“桂冠”。前苏格拉底哲学家也参与进了对mythos话语权的争夺中。克洛丰的色诺芬尼（Xenophanes）就曾批评“荷马与赫西俄德将人类中所有有害的、应当受到责难的东西都归因于了神明的力量”，而自己重提一套“好的”mythos的标准。巴门尼德更是明确强调要“听我的mythos！”这与思培多克勒捍卫自己的mythos的方式如出一辙。 

    诗人与哲人同时对自我mythos地位的捍卫，从某种程度上显现出后世所谓“哲学与诗歌之争”的雏形。但此时，哲学仍借用诗歌mythos的权威为自我正名，尚未求诸logos。然而，一旦人人都有权利声称自己的mythos才是更好的言说，mythos原本凌驾于一切的权威便决定性地让位于评判者自身。缪斯不再在场，“人的时代”悄然降临。而伴随着mythos本身超越性的下降，一个必然的问题便是：mythos一词还能否完全承担起其权威性的功能？或者说，mythos是否还具有不可置疑的真理性与说服性来作为人们认识与理解世界的基础？ 

    从阿那克里翁到品达，再到前苏格拉底的哲学家，这些言说者尽管各有其立场与态度，但在面对上述问题时，他们对mythos一词的表达都越来越收缩与谨慎。若在公元前6世纪晚期至公元前5世纪早期，mythos还被部分作为一个正面、积极的词汇来使用，那么，到了希罗多德之时，他已不再能，或不再愿意用mythos来指代其自我表达。他将mythos束之高阁，转身求诸logos，赋予logos以更高的力量与权威。这可以说是logos之变的一个重要转折。 

    希罗多德的写作代表了神话（或mythos）时代对理性（或logos）时代的退让，从他开始，可以明显看到作家对传统mythos整体性的保留态度。在《历史》开篇，希罗多德点明：他希望去探究希腊人与波斯人纷争的原因，于是，详细记述了两者关于同一神话/故事的富有争议的说法。然而，在包括腓尼基人的说法被一一陈列后，这位历史学家以一句总结摒弃了对前述几种mythos的考察：“这两种说法，哪一种更合乎事实，我不想去讨论。下面，我将指出我本人确切知道的那个最先向希腊人发难的人，继而继续我的叙述（logos）。”由此，希罗多德转向了吕底亚国王克洛伊所斯（Kroisos）的故事，并借此将其历史探索追溯到公元前6世纪中叶这个可知的历史时代——它成为希罗多德历史叙述的真正起点，一个“不去论述神话”的历史性开端。 

    有一种历时化（chronological）的意识，清楚表明了希罗多德的记述愿意开展的范围与界限：在对历史“时间”的反复强调下，“历史”停留在“不可知其时”的神话叙事的边缘。对他而言，“神话”过于久远，无法验真与证伪，于是，选择将其悬置——只有那些可以客观知道并验真的时期与事件才是他本人希望去讲述（legein）的对象。这便意味着，在某种程度上，希罗多德将远古的“神话”与故事搁置在了其历史叙述框架之外，或至少，他本人的logos将不会包含传统意义上的mythos，而力图成为一种新的关于过去的叙事。 

    这并不是说希罗多德就此将神话直接贬损为欺骗性、虚假的叙事，而是在“悬置”的方法论原则中，对“神话”或我们称之为mythos的话语体系作出了一个不同于史诗传统的界定。福勒曾敏锐地指出，希罗多德在谈论公元前6世纪中叶的一起历史事件时，引人注目地使用了上文提到的“人的时代”（tēs anthrōpēiēs geneēs）这个不同寻常的短语：“波律克拉铁斯，据我们所知，是在希腊人中第一个想取得海洋统治权的人……不过，在我们所谓的‘人的时代’, 波律克拉铁斯就是第一人。”人类时代的“第一”要从头开始计算，它与神话人物所存在的“前人类时代”或“神话时代”相分离。这意味着，荷马与赫西俄德笔下的英雄与诸神，包括缪斯，都被修昔底德悬置在人类历史周期之外，将其归之于经验事实“不可知晓”“不可确信”或“不可触及”的领域。 

     这是希罗多德在他所处的“人的时代”对mythos作出的 “评判”，但其“悬置”方法使得这一评判相对温和，因为它将史诗传统与希罗多德自身的历史立场之间的张力模糊化了。不过，对于希罗多德而言，仍有一个他必须面对的问题，即，如何解释那些“不可确信”的神话人物所拥有的确定无疑的、流传至今的名字与故事。对此，希罗多德用一句几乎惊世骇俗的评论作出了解释：“每一个神从什么地方生产出来，或者他们是不是都一直存在，他们的外形是怎样的，这一切都可以说是希腊人在不久之前才知道的。因为我认为，赫西俄德与荷马的时代比我的时代不会早过四百年，是他们，把诸神的家世教给了希腊人，把他们的名字、尊荣和记忆教给了所有人并且说出了他们的外形。”希罗多德并不否认神明的存在，但他在可知与不可知的边界上，重新界定了赫西俄德与荷马的位置。这两位诗人“创造”了神灵的名字，正如荷马也同样“创造”了希罗多德本人未曾见过的传说中的欧凯阿诺斯（Ocean）河流一样。他们作为“人”本身，并不一定受到所谓的缪斯的神启，毋宁说，大多数神话故事与人物，不过是诗人自身的创造，它们即便很难证伪，也很难证实。由此，诗人所赋予希腊人的mythos，在希罗多德看来，就应当被排除在人类历史的考察范畴之外，而换个角度来说，书写人类历史的历史学家，也应当自觉地将mythos之言说与内容束之高阁，以确保其可知历史的可确证的真实性。 

     希罗多德在此将诗人的mythos与神圣世界作出了区分，神圣世界仍具有崇高的权威与神圣性，但诗人作为传统中讲述与唱颂mythos之人，却受到实质性质疑。在此意义上，我们或许可以理解，为何希罗多德特别有意识地将自己的叙述指涉为logos，并刻意避免使用mythos一词：他的logos是排除对传统mythos讲述的言辞，而他本人，则是区别于传统诗人的历史学家，是能够给希腊人带去一种新的（也更真切的）记忆的言说者。由此，希罗多德便能够从“可知性”与“真实性”出发为其自身的“历史的logos”赋予更高的位置。于是，当他拒绝采信关于居鲁士（Kyros）出生的三种说法时，他宣称将要告诉我们一个“真正的故事”（ton eonta logon）；而那些希罗多德称之为logioi andres的人，则被认为是具有学养的权威人士，他们不仅通晓过去的故事，而且知道哪些才是值得聆听的。所有这一系列对logos的使用都表明传统诗人权威在明显下降。 

    从古典时期早期的诗人阿那克里翁与品达，到前苏格拉底哲学家，再到希罗多德，可以看到，mythos整体的权威性与神圣性越来越低，随之而来的，是logos以及与之相匹配的 historia(历史)的兴起。尽管，在这一阶段，mythos仍处于某种“中间状态”，希罗多德也仍在书中收集了大量传统神话故事，但mythos还是在historia的判断性“悬置”中受到了无形的挤压与价值重估。这恰恰是希罗多德在其“历史与诗歌之争”的框架下为mythos与logos之变所带来的一个具有深远意义的方向性影响，该影响到智术师与修昔底德之时，将会开展出全部的力量。 

    三、智术主义与逻各斯势力的兴起 

     随着启蒙运动与社会变迁的发展，普罗塔格拉 “人是万物的尺度”的宣言打破了mythos与logos最后微妙的平衡。当把人作为宇宙的中心来度量世界时，诸神便隐退天际，传统中神圣的mythos随之黯然失色。智术师是一个彻底转向logos之言说的群体。当mythos的真理性与说服性一再受到质疑，以“人”为万物中心的智者们，最终选择了彻底摈弃将mythos作为人们理解世界的基础，转而在logos处建立其认识论的根基。对智术师而言，logos之所以被认为是可靠的，是因为它是纯粹的人事：它更多与人类的语言和修辞相关联，与遥远的神话无涉。如此介乎人类现实行动之间的言说，在智术师看来，最能呈现真实的人类社会。高尔吉亚将logos与真实性（reality）联系起来，并在其《海伦颂》（ Encomium of Helen ）中，用logos的修辞学力量为海伦传统的mythos开脱，便是这一观念的典型体现。 

     对logos作为言辞力量的强调，是智术师处理与理解logos的一个显著特征。虽然荷马时代已有katalegein(准确地说)一类将logos作为言说之意的词汇，但公元前5世纪，logos在智术师运动下成为一种社会现象、方法论乃至世界观。就社会与政治背景而言，古希腊城邦对公共辩论的强调强化了logos的重要性，但更重要的是，在人本主义的思想逐渐兴起、传统mythos愈受质疑的大趋势下，logos所进入古希腊社会视野之中的意义。当人们返诸己身，以期对人类自身的行动作出自我解释时，logos作为影响政治行动乃至广泛人类行动的推动力，便获得了作为真实性基础的权威。换言之，通过理解logos在人类社会中所展现的力量，便能够理解人类社会最根本的真实性，而这种真实性又将成为指导人们行动的基础，它足够聚焦当下，不再需要神圣世界与遥远历史的参与。由此，logos与mythos彻底分离。而这一步，智术师们走得要比希罗多德激进许多，在他们对logos的强势追随下，mythos及其背后的整个传统世界与之渐行渐远，甚至隐没。 

    或许并不令人意外的是，在现存智术师的残篇中，mythos出现的情况少之又少。在讨论logos的诡辩与欺骗力量时，高尔吉亚未将该词与mythos相对比，而在《海伦颂》中，他所对比的却是poiēsis（诗歌）。这似乎显示出智术师试图超越既有“mythos与logos”之传统并重新界定两者关系的“野心”。 

    这一野心在智术主义的语境下是可以理解的。因为对智术师（例如高尔吉亚与普罗塔格拉）而言，logos总被认为拥有双重力量：既是一种说理的话语方式，也是一种欺骗性的话术。无论之前人们认为logos与mythos何者真实、何者虚假，在智术师这里，logos囊括了这两个方面，从而在“真实性”问题上不再与传统意义上的mythos相对。两者的关系因而需要被纳入一个新的框架。新的框架是什么？柏拉图的《普罗塔格拉》提供了一个可能是主流的智术师的回答。在这篇被认为很大程度上忠实于智术师本身作品的文本中，mythos被指涉为“给孩子们讲述的虚构的故事”，而logos则为“逻辑论辩”。这意味着，mythos与logos的对立不再是欺骗与真诚、谎言与真理之间的对立，而是“现实”与“虚构”之间的对立。 

    在智术师的现实主义关怀下，mythos被整体文本化（textualization）地处理几乎是一个必然的结果。由于被理解为虚构的，mythos只可能是一种人为的文学现象，而不再是来自缪斯的神启。在公元前5世纪日渐发达的书写体系下，随着口头传播的mythos被越来越多地记录下来当作文本资料和参考资料扩散流通，真实性本就受到质疑的mythos愈加丧失其传统宗教与社会的意义。从而，无论是在智术师群体中，还是在其他领域，mythos都越来越被排除在历史与现实的追问之外。 

    修昔底德无疑深受这一思潮的影响。与希罗多德相比，这位更年轻的历史学家除了精通智术师的作品以外，也更加坚决地将mythos排除在其文本写作之外，从而，在许多人（尤其实证史学家）看来，修昔底德是真正的“历史”书写的开端。尽管就对待mythos的立场而言，希罗多德与修昔底德之间是程度而非性质的差异，但在同时代智术师传统的强烈影响下，修昔底德对mythos与logos的争判更加毫无保留地偏向了logos，即其所代表的“非虚构”的、“现实理性”的一面。 

     修昔底德明确宣称，mythos，连同那些久远的传统记忆都不应当被纳入历史，因为记忆是脆弱、模糊的，甚至是具有欺骗性的——它永远是对历史的挑选、解释与重构。因此，“在这样的领域，很难去相信它们所呈现出来的信息”。作为一位史学家，修昔底德呼吁每一个人仔细甄别所有信息，去觉察那些记忆或传统说法中无法证实甚至不真实的成分，并识别出它们在经年累月后最终与mythōdes(神话)相结盟并倾泻出的那些不可信的言说。对修昔底德而言，现实与记忆之间存在着一个明显的“可信”与“不可轻信”的对立关系，而后者在诗人与故事记录者的笔下又更加严重。因为当诗人“夸大其词地为事件赋予流光溢彩”或当故事记录者“为了听者的愉悦而非为了事实”将未经证实的东西拼凑在一起时，那些令人怀疑的说法就彻底令人难以相信了。为此，修昔底德坚决提出，“如果我们希望能够看清过去的事实，借以预知未来”，就不应当像诗人和故事记录者那样为迎合人们的兴趣而写作，而是应当彻底地回到可信且可证实的“现实”之中。 

    那么，如何确保“现实历史”的真实性？修昔底德走得比智术师更远。他从logos（言辞）转向了ergon（行动），将所有历史书写都建立在现实行动事件的基础上。在《伯罗奔尼撒战争史》中，伯里克利有一个著名说法，即“真理寓于行动之中”，这可以说正是修昔底德的立场。如果说logos还有欺骗的可能性，那么，现实中“当下”的ergon则既不虚构也不虚假。《伯罗奔尼撒战争史》几乎不关注过去与传统，它处理古代（ta palaia），最多是为了通过看似逼真的证据来构建权力逐渐发展的模型。修昔底德所要创建的，是基于权力与战争概念的“行动的理论”，他将关注的视野聚焦于当下，以至于所有远离当下行动的诉说，都被谨慎地悬置甚至排除在外。这位理性主义与实证主义的历史学家不同于那些讲述故事的诗人，他就此将“神话”（mythos）与历史隔绝开来。 

    从智术师对logos的推崇，到其对mythos的文本化理解，再到修昔底德对mythos的排除，在公元前5世纪至公元前4世纪一系列启蒙运动思潮的推动下，mythos已被赋予完全不同于古风时期的位置与地位。一定程度上，mythos在修昔底德的笔下受到了最为激烈的挑战，这也是其在整个古希腊思想历史中所遭受的最为严峻的一次重击。在知识论层面，修昔底德对mythos的处理尤其具有颠覆性，几乎完全否认了mythos之于现实世界的意义，否认了mythos存在的正当性。这使得mythos几乎被驱逐出历史舞台，或至少被足够地边缘化。 

    但修昔底德的观念代表较极端化的立场，甚至，他是与大多数同代人充满分歧的少数派。与修昔底德同时期，存在mythos的另一个面向，且在希腊民间社会更加流行。这一面向在最大程度上保留了对mythos的敬意与推崇，其首要特点正是非历史性以及对神话的演绎，即悲剧。从悲剧中可以看到，尽管mythos无可辩驳地受到了冲击，但它对古希腊社会的影响力仍然强大。由此，历史学家、智术师与悲剧作家之间构成了一种对抗与竞争关系，这显示了秘索思与逻各斯之争在当时更加复杂且充满互动的动力学图景，而这种竞争最终对柏拉图关于mythos/logos问题的判摄形成了重要影响。 

    四、悲剧意识与秘索思逻各斯的此消彼长 

    与修昔底德的历史书写相比，悲剧是一种更加大众化与平民化的文体。虽然，悲剧作家是一群具有高度自觉性的知识精英，但由于悲剧演绎在古希腊尤其雅典城邦是一项面向公民、竞赛性的公共活动，因此，悲剧的受众决定了其与大众阶层更广泛的连接，也由此在一定意义上，可以被视为与陶瓶、壁画、建筑等艺术形式相似的大众文化的代表。尽管以精英与大众、贵族与平民、少数人与多数人等二元架构来与 “秘索思与逻各斯之争”相对应过于粗糙与简略，但悲剧对mythos的敬意与推崇在很大程度上反映了当时社会大众对mythos及其所代表的传统神话的态度与立场。 

    一个有趣的现象是，事实上悲剧经历了一个从历史剧到神话剧的转变。这一转变发生在普利尼克斯（Phrynichus）因其历史剧《米利都的陷落》（The Capture of Miletus）被罚之后。该剧以历史事件为题材，由于其生动呈演了前一年米利都被波斯人攻陷的悲惨遭遇而引得在场希腊观众动容痛哭，所以城邦重金惩罚了普利尼克斯。自此之后，几乎所有悲剧都改为神话题材，不再触碰现实历史，以此避免“悲剧”过于令人悲伤。就这样，现实历史题材在悲剧这个文体刚出现时就被禁止，所有故事又回到神话之中。 

    这是“虚构”的mythos在悲剧领域得到高度肯定的一刻，它在此后成为界定悲剧之所以为悲剧的一个核心要素。在悲剧舞台上，“虚构”是一个被刻意强化的特质。不仅演员会戴上面具、穿上戏服，运用大量台词、“假扮”成剧中人物，而且整个悲剧剧场也与外界隔离开来，被有意制造为一个独立于历史社会的虚构空间。而正是在此空间中，神话的故事被改编、演绎与观看，由此，观众对此“虚构性”形成高度的自觉。“有距离地观看”恰恰构成了虚构之于悲剧的价值，而正是在这多重的距离之下，悲剧及其mythos成为一个被凝视、审查与思考的对象。 

    当然，这里的mythos已不是古风时期意义上的高贵而神圣的话语。虽然同样属于“诗歌”与“神话”范畴，但悲剧特别强调作者对传统神话的独创性改编，这意味着悲剧的mythos是一个极具作者性与创造性的话语表达，而非来自缪斯的神启。就此而言，悲剧的mythos接续的仍是古典时期“去神圣化”的批评传统，它在本质上完全属于智术师意义下“虚构的、非真实的故事”序列。不过，与智术师和历史学家不同，悲剧作家不仅承认并且大大突出了虚构的价值与力量，还试图在“虚构”中，恢复mythos的“真理性”。 

     对悲剧作家而言，真知寓于虚构的故事情节之中。正如亚里士多德所言，悲剧是“对一系列行为的模仿”。戏剧如同镜像一般，通过对故事人物的悲剧性命运的“模仿”，展开了对真实世界中的人性与生命本质的深刻探讨。在一系列无解的悲剧冲突中，世界和人都被展现为充满问题、矛盾与含混性的存在，而恰恰借由“虚构”所带来的距离，那些本被现实世界所掩盖或回避的问题、黑暗与矛盾被充分而安全地暴露出来供观众审视。索福克勒斯的《僭主俄狄浦斯》是以虚构的mythos传达真理的典型，通过对俄狄浦斯悲剧命运的揭示，索福克勒斯表明了理性知识之于真理的局限性。埃斯库罗斯的“奥瑞斯提亚”(Oresteia)对“正义”的根基发出了诘问，在阿伽门农家庭悲剧的演绎中，揭露了绝对正义达成的困难与悖论。欧里庇得斯《美狄亚》《埃勒克特拉》和《希波吕托斯》同样如此，这些剧目都从不同侧面探讨了人与人之间最根本的关系纽带如何可能以及如何不可能。 

    对在场观众而言，这些深植于人性与社会的根本问题指向了他们所身处的真实世界，而恰恰是在这虚构的时空中，真理得以以一种超历史乃至于超人的方式显现出来。它向人们表明，舞台上的mythos，以一种historia和logos所不能达到的方式揭露了真相，此真相不仅比现实历史世界所显现出来的更加深刻，而且也比理性思辨所触及的更加复杂。我们在悲剧中不断看到诸如此类忠告：“你有视力，但你却没有看到你所陷入的困境”，“你根本不知道你过的是什么生活，不知道你在做什么，不知道你是什么人”。对于人们日常所熟悉的知识样态、伦理道德、社会结构乃至于人们自身，悲剧都重新发问，并以一种毁灭性的方式呈现出人类世界中被小心翼翼回避、保护与掩盖起来的难以承受的真相。由此，悲剧作为一种虚构的文学形式，重新给予了mythos最高的真理性。 

     那么logos呢？Logos作为悲剧中的对话与言辞被纳入了mythos的表意系统之中，成为一种工具性的——尽管十分强势的——存在。Logos对悲剧而言不可或缺，它贯穿于整个戏剧演绎，是人物思想表达与交锋最直接的通道。悲剧情节的推进，都在语言的诉说、往来、游戏与较量中达成。而语言的误解、诱惑、欺骗与劝说又构成了悲剧情节中最重要的反转与高潮。可以说，在悲剧中，是logos成就了mythos，这恰恰是悲剧作为一种对话式诗歌文体与史诗或抒情诗最大的区别。在此意义上悲剧充分吸收并利用了公元前5世纪理性主义与修辞学传统，为mythos注入了当代最前沿的活力。然而，悲剧对logos作为言辞乃至逻辑思辨的力量又始终保持谨慎。无论是“奥瑞斯提亚”中对克吕泰莫涅斯特拉修辞术的尖锐批评，还是《僭主俄狄浦斯》中俄狄浦斯诘问判案的反讽性演绎，三大悲剧作家的作品都一再表明，logos是危险的。在如此种种对人物语言的危险性的揭示下，悲剧的mythos毋宁将公元前5世纪的logos整体纳入了其对人性之真理的探讨之中，由此，mythos拥有了对logos进行审查与盘问的权力，进而，前者对后者建立起一种“真理”意义的权威。 

    这是悲剧对公元前5世纪智术师传统、理性主义和实证主义历史观向mythos发起的多重挑战的回应。从悲剧在雅典乃至泛希腊世界的受欢迎程度来看，这一回应无疑十分强劲有力，并且得到了民间社会的大力支持。在每年举行的酒神节中，悲剧在循环往复的宗教与仪式的时空中不断强化着其对古希腊社会的整体性影响。而这一影响首先发生在公民教育上。通过集体的排演与观看，城邦公民不仅形成了个体层面的对悲剧问题的反思，而且通过共同的投票，形成了对悲剧意义的共同意见，从而建立起一种公共的、政治的、社会性的思想基础。这恰恰是自荷马以来mythos对古希腊社会而言最重要的意义，正是悲剧将其延续下来。 

    从mythos所面临的败退之势来说，悲剧在启蒙运动的大背景下，对mythos精神的重新强化是相当不容易的事，但这也表明，mythos在希腊世界中拥有强劲且充满韧性的生命力，使得古希腊的根本特质深深扎根于mythos传统之中，即便深受启蒙运动的冲击，mythos也没有被新兴的思想浪潮所湮灭。比三大悲剧作家再晚一辈的柏拉图目睹了这一切，恰因如此，这位哲学家也显现出了最深的忧虑，他不仅明确发起了“诗歌与哲学之争”，并且还要从根源处对mythos与logos的关系进行彻底的哲学改造。 

    五、柏拉图对秘索思与逻各斯关系的哲学改造 

     柏拉图对传统mythos的批评几乎人所共知。在《理想国》第二、三卷中，他指出，传统诗人所编造的mythos都是虚假的故事，因为他们把伟大的神描写得丑陋不堪、把英雄塑造为无恶不作的恶棍，这样的mythos既不虔诚、也不真实，需要被排除在理想的城邦之外。从上文讨论中可以看出，柏拉图此处所针对的正是史诗与抒情诗传统之下的诗歌，尤其那些将英雄特质极端化的悲剧。对柏拉图而言，诗歌尤其悲剧以虚构的形式所展露出的引以为傲的“悲剧性真理”恰是最糟糕的，因为这些故事对不幸与罪恶“不加拣选地”模仿，并且夸大了欲望、痛苦、快乐这些灵魂中最低劣的部分，因此，这样的诗作极容易将mythos置于伦理的险境。倘若城邦中普通的公民无法分辨模仿的真伪与高下却跟随这些故事行事，那么人们的灵魂不仅不会变得更优秀，还将处于道德败坏的危险之中。因此，最好的办法，就是将那些“讲不道德的故事的”诗人驱逐出去，“至于我们，为了对自己有益，要任用较为严肃和正派的诗人或讲故事的人，模仿好人的语言，按照我们开始立法时所定的规范来说唱故事以教育战士们”。 

     柏拉图之所以对诗歌如此警惕，不完全是因为“虚构”本身对真理形成了威胁，尽管，它的确因其作为对真相的模仿而多少远离真实。他最深的忧虑在于——正如他所目睹的——传统mythos不仅道德含混，而且对公民的影响巨大。这正是柏拉图在“古已有之”的“诗歌与哲学之争”中看到的最大问题。柏拉图深知，在一座城邦中，要彻底驱逐诗歌与故事（mythos）是一件多么困难的事：“故事的制造者”（muthopoioi）在城邦中无处不在。她们首先是母亲和保姆，然后是老男人和老女人，还有忙着照顾新生儿的那些不知疲倦地喋喋不休的人，她们“向他们的耳朵里灌输迷人的话语”，为他们讲述口传的以假乱真的故事。由神话和美丽的故事所承载的整个模仿的情感结构吸引了年轻人的眼睛和耳朵，他们会被那些自发的“神话家”迷住，最终“变成身体、声音和思想的性格和第二本性”。从孩子的睡前故事，到所有公民都热衷于观看的戏剧演出，以情动人的文教无处不在，mythos强劲的生命力令其教育如此深入人心，若其真的道德败坏，那么它将对公民及社会形成毁灭性影响。因此，既然深知无法驱逐mythos本身，那么，至少应当将那些对城邦有害的mythos及其制造者排除在城邦之外，方能对城邦形成最大的保护。这正是柏拉图所谓“驱逐诗人”的真正原因。 

    需要指出的是，柏拉图并未驱逐所有诗人与mythos。在其哲学建构中，更重要的是用新的mythos去替代那些传统的、被驱逐的mythos。“任用较为严肃和正派的诗人或讲故事的人，模仿好人的语言”正是柏拉图在驱逐传统诗人之后，立即给出的一个替代性方案。那么，为何柏拉图要使用这样一个“不彻底的”方案？ 

     从知识论的角度来看，这是因为，mythos仍是柏拉图哲学思辨与教育不可或缺的存在。正如柏拉图笔下的苏格拉底在《理想国》中所承认的，尽管知识最终通过logos获得，但在获取知识的哲学式的辩证法中，人们却必须“不使用任何感觉的对象，而只是通过纯粹的观念来推动达致观念的结果”，这种用非物理术语来对抽象概念和形式进行理解的方法无疑是困难甚至难以自证的。因此，logos的局限性本身就要求mythos作为一种语词性的、哲学的形象，作为“认知的桥梁”，承担起对真理的“可见和可感知的表达”。由此，mythos不仅要成为哲学上的“发言人”，甚至还要成为哲学论证尤其辩证法开始之前真理交流的第一原则（即起点或公理），去完成那些logos或辩证法难以达成的事情。《理想国》中的洞穴神话与厄尔神话等都是典型的例子，由此可以看出，神话对于哲学认知过程的开始和结束都是必要的。从某种程度上而言，它也可以解释为何mythos本身在公众世界中具有（比logos更加）普遍性的吸引力与知识传播的能力，无论结果好坏。 

     在此意义上，便可以理解柏拉图既要“驱逐诗人”又要“留下诗人”的看似矛盾的态度，而我们看到，这一态度远比被动的妥协要积极得多。那么，他所谓“正派的故事”和“好人的语言”是什么？在柏拉图的论证框架下，这两者自然就是由哲学/logos所引领的语言，而这正是柏拉图认为mythos本身所无法达成的东西。哲学之所以比mythos更加权威，是因为其思辨的logos包含了经由理性而得来的“理相”（eidos）。这些“理相”构成了真正的现实，且在那个“真理”的世界中永恒不变。因此，这些具有绝对稳定性的存在可以指明什么是真正的善，并引导人们走向德性。当然，柏拉图哲学“真理性”的自我辩护是一个相当复杂的体系性问题，无法在此展开，但倘若柏拉图假设了他的辩护是成功的，那么，在其理想的城邦建设中，哲学，logos，就成为包括诗歌在内的一切教育与立法的先导与模型，从而使得mythos必然处于一个从属地位。 

     基于此，城邦便可以容纳mythos，并且对其不可或缺的辅助力以及不可抗拒的影响力加以利用。于是，柏拉图提出：“logoi分为两种：一种真实，另一种虚假。必须让人在这两方面都得到教育，而且，首先得在虚假的方面……要首先对孩子们讲神话故事，因为总的来说，这些故事说的是假话，但其中也有真实的东西。”真实的logos，柏拉图指的是哲学的理性辩证；而虚假的logos，即智术师/历史学家意义上的mythos。在哲学向那“虚假的logos”注入“真实的东西”（即哲学真理）后，mythos便得以作为构成城邦logoi（复数）整体的一部分，继续对公民施加“第二本性”般的影响，并作为哲学教育的起点对公民实施真正的知识教育。当mythos成为logos，神话/诗歌成为哲学的一部分时，logos不仅实现了对mythos最好的规训，而且，哲学反过来也成为诗歌，成为“最伟大的一种缪斯的艺术”。 

    某种意义上，柏拉图对logos至高地位的赋予显现出其从智术师处接续而来的批评传统，在mythos与logos问题的整体框架下，柏拉图无疑是作为一位革新的思想家站在了启蒙运动的风口浪尖。然而，这位苏格拉底的学生对智术师传统是有所保留的。他不仅通过对“德性”的强调，用一个完全道德化的“善”的logos取代了智术师笔下“可善可恶”的logos，而且也在其对mythos的处理中，修正了智术师（以及修昔底德）彻底背离mythos传统的进路，将mythos在其理想的城邦中保留下来。这意味着，mythos在柏拉图的哲学中不仅获得了一席之地，而且，还在一个显性的“秘索思与逻各斯之争”中被一位哲学家重新赋予存在的根本价值。柏拉图本人以戏剧对话（mythos）的方式来呈现其哲学，便是最好的例证。 

     柏拉图之所以将mythos纳入其哲学体系，不单是因为其知识论上的前驱性意义及其对公民教育的影响力。《蒂迈欧》中梭伦的故事暗示，这一切或许还与mythos在古希腊的本质相关。这个故事讲述了梭伦前往埃及的见闻。梭伦在与当地最有经验的祭司谈话时发现，“不论他自己还是其他希腊人，可以说都对古老的事物一无所知”。对此，一位年迈的祭司道出了一句箴言：希腊人之所以不知过往，不是因为无知，而是因为“希腊人永远都是孩子”。.祭司的意思是，由于古希腊人总是用口头的方式传播故事，因此，并没有像古埃及那样的书写传统将一切记录下来。在古埃及的对比下，柏拉图指明，古希腊的历史实际总是留存于口头的记忆之上的，神话记忆而非历史书写构成了古希腊之所以为古希腊的本质。就此而言，mythos直抵古希腊精神的核心。它不仅不可能被驱逐，而且还在存在论意义上，牢固地锚定在了古希腊的内核之中。在如此社会里，神话就是历史，它为历史的起源不断输送能量，并塑造着古希腊人的历史与文明意识。我们看到，在此，虚构的故事就不仅是在知识与教育的意义上被需要，而且是在整个古希腊文明的意义上被需要。 

    恰是在这一点上，柏拉图有意识地将mythos融汇进了自己的理想城邦的建构之中，并且，以一种相当积极的方式对其本质进行了最大程度的利用。《理想国》中著名的“高贵的谎言”就是一个典型例证：这个被哲学规训的具有真理性的起源神话成为整个理想的文明城邦建立与教育的起点。这个看似“荒唐的”的传说，“虽然那些听故事的人未必会相信，但后代的后代，子子孙孙迟早会相信的”。在世代的流传下，高贵的谎言成为历史的起源，成为城邦立法最根本的、先验的无可辩驳的基础，从而，mythos也在这个对logos而言最理想的城邦之中成为一个最不可或缺的存在。 

    从柏拉图对mythos的批评来看，他一方面明显继承了智术师与理性主义传统对logos的尊崇，另一方面，也对mythos强韧的力量有着充分的自觉。因此，尽管在柏拉图的理论体系中，logos是绝对高于mythos的存在，后者必须受前者所指导，但无论是在教育意义上，还是在存在论与知识论意义上，柏拉图都承认，mythos对古希腊而言绝不可或缺。 

    由此，虽然在柏拉图这里，我们看到“logos对mythos的胜利”，但我们也看到，这一胜利建立在对mythos的承认、接纳甚至为己所用的基础上。就此而言，柏拉图可以说是从智术师和修昔底德的极端立场的某种后退。在人类的城邦与社会中，这位哲学家试图找到一种mythos与logos间平衡与共存乃至互补的关系，令其各司其职。这一后退，不仅是战略性的，而且深植于其对哲学思辨的理性认识以及古希腊文明本质的深刻理解之中。古希腊人，或人类，对mythos和logos两种精神的需求表明其任何一方都不能，也不可能，被完全否定与排除。恰因如此，可以看到，无论是在柏拉图之前，还是自柏拉图之后跌宕起伏的历史中，mythos与logos总是相互勾连牵延、此消彼长，时而彼此竞争，时而互为补充，直至今日。 

    结语 

    Mythos与logos自古希腊文明伊始，就在口述传统催生下，深深扎根于其文明精神的核心。二者在普遍的二元思维架构中，构成古希腊内在精神的两个面向，一同推动该文明向前发展。 

    Mythos与logos不单是两个词汇与概念，背后隐含的是认知世界及自身处境的表达方式与路径。二者的关系不仅关涉话语体系的构建方式，还包含对“真实与虚假”“神圣与世俗”“诗与思”等一系列问题的思考。因此，该议题既指向神话与历史、神话与哲学、历史与哲学，也在形而上层面与认识论、存在论乃至宇宙论问题关联在一起。正是在多层面的勾连与张力中，mythos与logos开启了一个极为丰富的希腊世界。 

    二者“二元辩证统一”的动力学关系构成古希腊极为关键的文明特质。两者之所以不断此消彼长，是由古希腊开放的宇宙论、世界观与不同时期的社会和思想共同造就的。对变迁动力追根溯源，除却神话与思想、诗歌与哲学、感性与理性这些对立概念本身内在的冲突与竞争，社会文化自身的发展、古希腊民间风俗的变化、传统宗教与世俗生活的抗衡，乃至外来文化与新兴思想的渗透等，也均是推动两者变化的重要因素。 

    进一步看，恰恰是这一相互制约又互相定义的动态特质，对此后西方文明的展开产生了本源性影响。古希腊之后，不仅秘索思与逻各斯之争一直根植于西方思想发展脉络之中，而且两者地位在不同时期的变化也持续影响着西方感性与理性演变的周期以及西方哲学在认识论上的几次重大转型。现代社会兴起后，这一动力学原则更进一步与理性主义、科学主义结合，强化了西方科学与宗教并立的辩证传统，至今仍是西方文明体系的核心结构，对现代性及其内在复杂性的形成产生了深远影响。 

    在mythos与logos经历地位反转与意义更迭后，西方对自身整体文明传统的自我认识与系统性表达也随之形成。Mythos被驱逐出历史叙事的范畴，logos与实证精神合流成为现代西方历史观的真正开端。尽管在古希腊时期，这一历史认识论仍属激进、并非主流，然而经由漫长的中世纪、文艺复兴而进入现代世界之后，它在现代社会发挥出巨大能量。实证主义的理性化书写、古史研究对虚构叙事的全盘否定都显现出自古希腊时期便已在神话、传说与历史之间划出的巨大鸿沟。尽管对神话的复兴一直若隐若现，但整体而言，理性的历史观仍占据上风。这两种历史观的反复纠缠是系统性的，而这正是在现代世界中不断显现的问题。 

    本文节编自《中国社会科学》2024年第10期

  • 胡宝国：魏西晋时期的九品中正制

    魏晋南北朝时期的九品中正制度由于存在时间很久，各个时期多有变化。因此，有必要对这一制度进行分阶段的考察。在这篇文章中，只讨论魏西晋时期的九品中正制。

    一、释“上品无寒门，下品无势族”

    创立于曹魏时期的九品中正制在西晋一朝遭到了大规模的抨击。当时许多人批评中正制度，其中尤以刘毅“上品无寒门，下品无势族”(1)一语最具代表性。涉及到九品中正制度的论著，大都据此得出结论：当时世家大族垄断了上品。本文认为，这一结论仍有值得商榷之处。(2)

    晋武帝时，段灼上表称：“今台阁选举，涂塞耳目，九品访人，唯问中正。故据上品者，非公侯之子孙，则当涂之昆弟也。”(3)段灼与刘毅都指出一部分人垄断了上品。刘毅称为“势族”，段灼称为“公侯之子孙”、“当涂之昆弟”，二者应该是相等的。只不过段灼说得更具体些。所谓“公侯”，即指封爵，“当涂”是指高官要位。当时也有一些人并未直接批评中正制度，而是指斥高官子弟垄断了某些官位。刘颂对晋武帝说：“泰始之初，陛下践阼，其所服乘皆先代功臣之胤，非其子孙，则其曾玄。”(4)愍怀太子被废，阎缵上疏为之申冤，更具体指出，东宫官属如太子洗马、舍人以及“诸王师友文学”等职任人不当，“皆豪族力能得者”(5)。刘毅与段灼，刘颂与阎缵所选择的批评角度虽然不同，但却有相通之处。九品之品与具体官职存在着一定的关系。

    《晋书》卷九○《邓攸传》：邓攸“尝诣镇军贾混。……混奇之，以女妻焉。举灼然二品，为吴王文学”。《晋书》卷五二《华谭传》：“太康中，刺史稽喜举谭秀才。……寻除郎中，迁太子舍人，本国中正。”《晋书》卷四六《李重传》：“李重……弱冠为本国中正，逊让不行，后为始平王文学。”《晋书》卷六一《周浚传》：“（周馥）起家为诸王文学，累迁司徒左西属。司徒王浑表‘馥理识清正，兼有才干，主定九品，检括精详’。”

    担任中正者，本人必须是二品。司徒左西属是司徒府的官吏，“主定九品”，有时还可兼中正，自然也应是二品。(6)我们看到，被中正定为二品的人往往可以任太子舍人、诸王文学，这些职务正是阎缵所提到的。因此，阎缵批评“豪族”垄断这些职务与刘毅、段灼批评他们垄断上品当是一回事。换言之，正是因为他们垄断了上品，所以才能位居上述职务。

    但是，“势族”、“公侯之子孙”、“当涂之昆弟”究竟是些什么人呢？按通常的解释，这不过是世家大族的代名词而已，世族垄断上品的结论就是由此得出的。但考察一下上述批评中正制度的人的家世，事情就会复杂起来。《晋书》卷四五《刘毅传》：“刘毅字仲雄，东莱掖人，汉阳城景王章之后，父喈，承相属。”《晋书》卷四六《刘颂传》：“刘颂字子雅，广陵人，汉广陵厉王胥之后也。世为名族。同郡有雷、蒋、谷、鲁四姓，皆出其下。时人为之语曰：‘雷、蒋、谷、鲁，刘最为祖。’”《晋书》卷四八《段灼传》：“段灼字休然。敦煌人也，世为西土著姓。”同卷《阎缵传》：“阎缵字续伯，巴西安汉人也。”《华阳国志》卷一《巴志》：“安汉县号出人士，大姓陈、范、阎、赵。”以上四人，刘毅为“汉阳城景王章之后”，其父曾任丞相属，究竟属于哪一阶层，难以确定。其他三人或曰“名族”，或称“著姓”，或为“大姓”，当是世族。

    所谓世族，通常是指累世做官的家族。由于在一个地区长久不衰地任官，即被当地人目之为“著姓”、“大姓”、“名族”，或者也可称作地方郡姓。汉代以来，有一些著姓、名族的政治势力及影响并未局限在本地区，如汝南袁氏、弘农杨氏，这些家族世代在中央居高位，在全国范围内都有影响，这样的世族，可以称之为高等世族，以别于地方世族、地方郡姓。

    身为世族的刘颂、段灼、阎缵为什么要攻击世族垄断上品呢？其实，“世族”并不等于“势族”。我们可以通过元康年间举寒素一事加以推断。

    《晋书》卷九四《范粲传》：“元康中，诏求廉让冲退履道寒素者，不计资。”何谓寒素？何谓不计资？据《晋书》卷四六《李重传》载，诏令下达后，“燕国中正刘沈举霍原为寒素”，但司徒府未通过。司徒左长史荀组认为，“寒素者，当谓门寒、身素、无世祚之资。原为列侯，显佩金紫，先为人间流通之事，晚乃务学……草野之誉未洽，德礼无闻。不应寒素之目。”与荀组不同，李重则积极为霍原辩护：“如诏书之旨，以二品系资，或失廉退之士，故开寒素，以明尚德之举……沈为中正，亲执铨衡，陈原隐居求志，笃古好学……如诏书所求之旨，应为二品。”据此，可以得出如下认识：一、此诏是为了解决九品中正制实施中所出现的问题而发的。具体说，就是要冲破某些人仅凭“资”独占二品这种局面，其措施就是举寒素。按此传先云举霍原为寒素，后又云“应为二品”，可知举寒素意即举寒素者为二品。(7)二、前引刘毅说，势族垄断了二品，此传又称“二品系资”，可知势族获得二品即是凭借“资”。因此，有资者即为势族，反之则是寒素，势族是与寒素相对而言的。三、按荀组的说法，寒素应包括两项内容：门寒、身素，又可概括地称之为“无世祚之资”。门寒一词较空洞，留待下面讨论。所谓身素当是指本人无官无爵。荀组正是从此出发反对举霍原为寒素的。其理由主要有二：第一，“原为列侯”，第二，德行不够。德行较抽象，很难说清，所以第一条理由才是重要的。霍原为列侯，不符合“身素”一项，此外，霍原家世虽不可考，但本人未出仕却有封爵，应该说是从祖先那里袭来的，因此，霍原属于“公侯之子孙”，也即是势族，自然也就不能算“门寒”了。可见，荀组虽然仅指出“原为列侯”，但实际意味着霍原二项条件均不符合，所以才反对举他为寒素。

    《晋书》中明言被举为寒素者还有二人。《晋书》卷六八《纪瞻传》：“祖亮，吴尚书令。父陟，光禄大夫……永（元？）康初，州又举（瞻）寒素，大司马辟东阁祭酒。”《晋书》卷九四《范粲传》：“元康中，诏求廉让冲退履道寒素者，不计资，以参选叙，尚书郎王琨乃荐（范）乔曰：‘乔禀德真粹，立操高洁……诚当今之寒素。著历俗之清彦。’时张华领司徒，天下所举凡十七人，于乔特发优论。”(8)据此，当时被举为寒素者共十七人，由于史料缺乏，已无法全部了解他们的情况。但《李重传》却为我们透露了一点消息。元康年间，李重任尚书吏部郎，“务抑华竞，不通私谒，特留心隐逸。由是群才毕举，拔用北海西郭汤、琅邪刘珩、燕国霍原、冯翊吉谋等为秘书郎及诸王文学”(9)。霍原被举为寒素后并未出仕，此处误记。但我们怀疑其他三人均系被举为寒素者，因为他们被“拔用”的时间也是在元康年间，且既称“拔用”，显然地位不高，又与霍原相提并论，最后又被任命为“诸王文学”之类。如前所述，这些职务往往是由二品人士担任的。

    至此，我们知道被举为寒素者除霍原外还有五人。其中西郭汤、刘珩事迹不详，范乔情况较为特殊。其父范粲在魏末官至侍中，但始终不与司马氏合作，“阳狂不言”三十六载。(10)范乔被举为寒素前未出仕。纪瞻父祖均为吴国高官，纪瞻本人为“江南之望”。(11)吉谋家世也略有可考。《三国志》卷二二《魏书·裴潜传》注引《魏略》云：“冯翊甲族桓、田、吉、郭。”同书卷二三《常林传》注引《魏略》云：“吉茂字叔畅，冯翊池阳人也，世为著姓。”

    由此可见，被举为寒素者中起码有两名世族，即纪瞻与吉谋，他们被推举没有引起争论，看来是符合“门寒、身素、无世祚之资”这些条件的。换言之，他们并非势族。所以，世族并不等于势族。势族垄断上品不意味着世族垄断上品。所谓势族，乃是指现实有势力的家族，即那些魏晋政权中的公侯与当涂者。这些人中固然也有两汉以来的著姓、大族，如琅琊王氏、太原王氏、河内司马氏、河东裴氏等等，但也有像石苞、邓艾、石鉴这样一些起自寒微者。(12)他们显然不能以世族目之。固然势族只要稳定地、一代一代地延续下去，终有一天会演变为世族，但那毕竟是以后的事。在魏晋时期，势族不等于世族。势族的地位也并不十分稳固。在瞬息万变的政治斗争中，一些势族衰落了，一些人又上升为势族，虽然势族垄断了上品，但他们当中具体的家族由于现实政治地位不稳定，品也不稳定。《晋书》卷三三《何曾传附子何劭传》：

    劭初亡，袁粲吊岐（何劭子），岐辞以疾。粲独哭而出曰：“今年决下婢子品！”王诠谓之曰：“知死吊死，何必见生！岐前多罪，尔时不下，何公新亡，便下岐品，人谓中正畏强易弱。”粲乃止。

    何岐虽最终未被降品，但可看出其品并不稳定。《晋书》卷四三《王戎传》：“（戎）自经典选，未尝进寒素，退虚名，但与时浮沉，户调门选而已。”按“户调门选”，须“与时浮沉”，说明门户地位常有浮沉。刘毅云：“今之中正……高下逐强弱，是非由爱憎，随世兴衰，不顾才实，衰则削下，兴则扶上，一人之身，旬日异状。”(13)这是对现实政治的真实描述。另一方面，原有的著姓大族只要未跻身于公侯、当涂者之列，就不能算作势族。所以纪瞻、吉谋可以被举为寒素，而安汉大姓阎缵在势族面前只能自称“臣素寒门”。(14)

    稍后的例子也可以证明此点。东晋初年，王敦叛乱中刁协被杀，事后左光禄大夫蔡谟为刁协争追赠官位，在致庾冰的信中说：“又闻谈者亦多谓宜赠。凡事不允当而得众助者，若以善柔得众，而刁令粗刚多怨；若以贵也，刁氏今贱；若以富也，刁氏今贫。人士何故反助寒门而为此言之，足下宜深察此意。”(15)渤海刁氏是很显赫的家族，刁协父刁攸“武帝时御史中丞”，但一旦官场失意却被称为寒门，因此，这一时期寒门一词的含义与宋齐以后不同。地方郡姓在本地虽然绝对不属于寒门，但与“势族”相比，却只能处于寒门的地位。

    西晋时期，人们批评九品中正制度的另一个方面是，九品评定全由中正，不遵乡里舆论。刘毅在论九品疏中一开始就指斥说：“今立中正，定九品，高下任意，荣辱在手”，在以后所论中正制度的“八损”中，他不厌其烦地屡次指出这一点，批评中正不听乡里舆论，“采誉于台府，纳毁于流言”，以私情定品。前引段灼上疏也指斥：“今九品访人，唯问中正。”所以，许多反对九品中正制度的人都主张废除中正制，在土断的基础上行乡举里选。

    综上所述，西晋一朝，人们对中正制度的批评主要集中在两点。第一，势族凭资垄断上品。第二，中正不遵乡论。晋武帝时，卫瓘与汝南王亮的上疏可以说是对中正制度弊端的总结：

    魏氏承颠覆之运，起丧乱之后，人士流移，考详无地，故立九品之制，粗且为一时选用之本耳。其始造也，乡邑清议，不拘爵位，褒贬所加，足为劝励，犹有乡论馀风。中间渐染，遂计资定品，使天下观望，唯以居位为贵。(16)

    按卫瓘的说法，中正制度两方面的弊端是有联系的。正是由于中正不遵乡论，才导致“计资定品”。值得注意的是，中正制度初建时并非如此，只是“中间渐染”。这说明九品中正制度在魏晋时期曾经有过重大变化。

    二、魏、西晋中正制度的演变

    《通典》卷一四选举二历代制中载：“晋依魏氏九品之制，内官吏部尚书，司徒左长史。外官州有大中正，郡国有小中正，皆掌选举。”按此则魏晋时期的九品中正制没有任何变化。这是不准确的。赵翼《廿二史劄记》卷八中正条：“魏文帝初定九品中正之法，郡邑设小中正，州设大中正，由小中正品第人才，以上大中正，大中正核实以上司徒，司徒再核，然后付尚书选用，此陈群所建白也。”这个说法虽然系统化，但比《通典》更不准确。魏晋时期的九品中正制是有变化的。郡中正与州中正之设置并非同时。对此，唐长孺已有精确的考证。按他的意见，中正制度刚建立时，只有郡中正，州中正的设立“至迟不出嘉平二年（250），至早不出正始元年（240），也即是说在曹芳时”(17)。唐先生的这一论断是完全正确的。但是《晋书》卷四四《郑袤传附郑默传》还有须要解释的史料：

    初，帝以贵公子当品，乡里莫敢与为辈。求之州内，于是十二郡中正佥共举默……及武帝出祀南郊，诏使默骖乘。因谓默曰：“卿知何以得骖乘乎？昔州里举卿相辈，常愧有累清谈。”

    晋武帝当品事发生于魏末，但究竟在哪一年，史无明文。《晋书》卷三《武帝纪》：“武皇帝……魏嘉平中（249—254），封北平亭侯，历给事中，奉车都尉。”既云“嘉平中”，则武帝出仕年代肯定在公元250年以后。一般来说，获得中正品第之后即可出仕，尤其是晋武帝这样的贵公子，不大可能已经得到中正品第无官做，也不大可能出仕后尚无中正品第。因此，他出仕与获得中正品第应该大致同时，即都是在“嘉平中”。按《郑默传》载，晋武帝与郑默是由“州内”推举的。但“求之州内”却没有州中正推举，反而由一州之内的全体郡中正“佥共举默”，(18)当时似乎并没有州中正。《晋书》的记载疑有错误。汤球所辑王隐《晋书》卷六亦载此事：“默为散骑常侍。世祖出祀南郊。侍中已陪乘，诏曰：‘使郑常侍参乘。’谓默曰：‘卿知何以得参乘？昔州内举卿，十二郡中正举以相辈，常愧有累清谈。’”汤球注明此段文字辑自《艺文类聚》卷四八、《初学纪》卷一二所引王隐《晋书》。查此二书，《艺文类聚》引作：“郑默为散骑常侍，世祖祠南郊，侍中已陪乘。诏曰：‘使郑常侍默。’曰：‘卿知何以得参乘？昔州内举卿相辈，常愧有累清谈。’”《初学纪》引作：“郑默，字思元，为散骑常侍，武帝出南郊，侍中以陪乘。诏曰：‘使郑常侍参乘。’”二书均无“十二郡中正举以”七字。汤球可能是从其他地方辑出而在注出处时疏忽了。如此推测无大错，则王隐《晋书》与唐修《晋书》记载此事有所不同。即王隐《晋书》在“十二郡中正”诸字之后无“佥共”二字。虽只差二字，但却是非常重要的。因为有时史籍中说若干郡中正只不过是某州中正的代名词。《世说新语·贤媛》篇注引王隐《晋书》云：“后（羊）晫为十郡中正，举陶侃为鄱阳小中正，始得上品也。”羊晫举陶侃在西晋后期。《晋书》卷一五《地理志》下：“惠帝元康元年……割扬州之豫章、鄱阳、庐陵、临川、南康、建安、晋安、荆州之武昌、桂阳、安成十郡，因江水之名而置江州。”羊晫所任“十郡中正”即指任此十郡的中正。其中包括鄱阳郡，所以羊晫可以推举鄱阳人陶侃为郡中正。“十郡中正”，实际就是江州大中正。《太平御览》卷二六五中正条引《晋书》云：“杨晫、陶侃共载诣顾荣。州大中正温雅责晫与小人共载。晫曰：‘江州名少风俗，卿己不能养进寒儁，且可不毁之。’杨晫代雅为州大中正，举侃为鄱阳小中正。”杨晫当为羊晫，此处明言为江州大中正。据此推论前述“十二郡中正”实际当是司州中正的异称。唐修《晋书》记载此事大概是参考了王隐《晋书》，又觉得“十二郡中正举以相辈”费解，故增“佥共”二字，但意思就大不相同了。由以上的分析可知，唐先生关于州中正建立时间的考证还是不可动摇的。

    下面讨论另一个问题。据前引杜佑语，似乎不仅州中正与郡中正是在制度初创时就已同时存在，而且司徒府参预九品评定工作也是从那时开始的。赵翼更明言“此陈群所建白也。”这一说法也是不正确的。首先，史料中从未发现曹魏时司徒府参预品评工作。魏明帝时，傅嘏在难刘劭考课法时说：“方今九州之民，爰及京城，未有六乡之举，其选才之职，专任吏部”。(19)可见，当时选举工作在中央是由吏部一手包办的。其次，杜佑自己在《通典》卷二○职官二中也说：西晋“太始三年……司徒加置左长史。掌差次九品，铨衡人伦”。既然说“加置”，时间又如此具体，在这之前当无左长史。杜氏自相矛盾。《晋书》卷二四《职官志》也有明确记载：“司徒加置左右长史各一人。”《艺文类聚》卷三一引潘尼《答傅咸诗序》：“司徒左长史傅长虞，会定九品，左长史宜得其才。屈为此职，此职执天下清议，宰割百国，而长虞性直而行，或有不堪。余与之亲，作诗从规焉。”诗中有句云：“悠悠群吏，非子不整，嗷嗷众议，非子不靖。”这是西晋司徒左长史参预评定九品的例子。

    综合上文，魏晋之际州中正的建立与司徒府参预九品评定工作是九品中正制的一大变化。这一变化的出现是有原因的。《太平御览》卷二六五中正条引晋宣帝除九品州置大中正议曰：“案九品之状，诸中正既未能料究人才，以为可除九制（品？），州置大中正。”同卷又引《曹羲集》九品议：“伏见明论欲除九品而置州中正，欲检虚实。一州阔远，略不相识，访不得知，会复转访本郡先达者耳，此为问中正而实决于郡人。”据此，置州中正的建议是由司马懿提出的，而曹羲则持不同意见。据同卷引应璩《新论》，应璩也反对建立州中正。他说：“百郡立中正，九州置都士，州闾与郡县希疏，如马齿不相识面，何缘别义理？”应璩的观点与曹羲的观点在某些方面是一致的。他们都认为不必设州中正，因为一州之地过于辽阔，州中正对郡县的情况不了解。所谓“略不相识”、“如马齿不相识面”都是这个意思。但应璩仅仅担心义理难辩，而曹羲所担心的是，由于州中正不清楚下属郡县的情况，结果还得回去访问“本郡先达”，名曰州中正负责，但“实决于郡人”，这样就失去了建立州中正的意义。曹羲的担心是有道理的。中正制初创时就规定“各使诸郡选置中正”(20)。既然中正的推举权在“诸郡”，推举出来的中正当然是最能体现“诸郡”意志的人。九品评定最终“决于郡人”，“决于本郡先达”就不可避免。所以，如果州中正建立后也落得同样下场就等于毫无意义了。由此可以看到，司马懿的本意原是想不理会“本郡先达”的意见，改变中正品评“决于郡人”的现状。曹羲所提出的问题在魏末究竟是如何解决的，由于史料缺乏，还不清楚。但西晋“诸郡”推举中正的权力终于被剥夺而转交给司徒府。中正品评人物必须由司徒府最终核实，“决于郡人”的局面一去不复返了。(21)

    在此，须着重指出，所谓“郡人”、“本郡先达”绝不包括一郡内的所有人，只能是那些地方上的郡姓、著姓、大族。司马懿所要打击的正是他们。明乎此，我们终于可以理解西晋时期一批地方郡姓为何要攻击中正制度了。但是，魏末作为皇权的实际执行者司马懿、曹爽兄弟等人为何要打击地方郡姓呢？由此为何又导致了势族垄断上品？

    如前所述，势族中有不少人就是两汉以来的著姓、大族，就此而论，他们与地方郡姓似乎并无区别。过去的研究往往将他们视为一体。这是不无道理的，但又不完全对。固然，自汉代以来，郡姓、大族一般都是在本地发展起来的，但是其中一部分郡姓并没有就此止步，而是跨出州郡，走向中央，累世公卿，如汝南袁氏“四世五公”，弘农杨氏“四世三公”。这些人的利益已经不仅仅是与地方州郡相联系了，更多的则是与中央政权联系在一起。没有统一的东汉帝国，“四世三公”就只能是一场空幻的梦。因此，董卓之乱以后，他们都企图重建统一国家。建安元年（196），曹操“挟天子”后，许多人纷纷归附到他的旗帜下，就是由于他们认为曹操“乃心王室”。(22)地方郡姓与中央政权联系并不密切，他们的力量在于州郡、在于宗族乡里。因此，董卓之乱爆发后，大量的地方郡姓并没有离开本土。这一方面使他们以后难以上升，另一方面又使他们能够有效地控制宗族乡里，并进而建立自己的武装。在各个地区，他们往往是不安定的因素。西晋时期，地方郡姓依然垄断着州郡僚佐的职务，操纵着乡里舆论。(23)虽然与势族相比，他们处于寒门的地位，但在本地仍不失为著姓、大族。他们的一切特权也就是来源于此。愈是依靠门第过活，便愈要排挤那些没有门第的人。因此，轻视寒人的风气在地方州郡中自汉末历魏晋而不衰。(24)

    总之，地方郡姓由于远离政治斗争中心，所以在汉末以来的历次动乱中都没有受到重大损失，这个阶层基本上没有什么变化。

    与此不同，汉末的高等世族既然寄生在东汉中央政权的躯体上，当统一帝国崩溃后，他们便四散逃亡了。虽然他们都希望重建统一国家，但究竟借助于哪一种力量、哪一派军阀来实现其目的，每个人的选择并不一样，有人投靠曹操，有人追随刘表，有人与孙氏父子共患难，也有人跟着刘备辗转他乡。尽管他们的主观动机一致，但客观行动却使本阶层陷入了分裂中，今天的史家虽然可以根据血统把他们集合在自己的笔下，但在现实斗争中，血统并没有使他们团结在一起。高等世族能否存在下去，也不在于他们的血统。袁绍凭借“四世三公”的地位当了讨伐董卓的盟主，但当大族一旦发现他并非救世主，便又纷纷离开了他。随着官渡之战的结束，这个家族终于迎来了自己的末日。在动乱的年代里，他们能否存在下去，关键在于自身的能力。荀彧帮助曹操艰苦创业，几度难关；司马懿战诸葛、平辽东，战功赫赫，因此他们的家族才能延续下去，成为魏晋政治舞台上的重要角色。也正是由于他们并非依靠门第过活，所以对于那些卑微之士也并不特别压抑。颍川戏志才、郭嘉先世无闻，有“负俗之讥”，但荀彧“取士不以一揆”(25)，大胆拔用了他们。司马懿“知人拔善，显扬侧陋，王基、邓艾、周泰、贾越之徒皆起自寒门而著绩于朝”(26)。司马师为了任用石苞公开提倡曹操当年唯才是举的方针：“苞虽细行不足而有经国才略。夫贞廉之士，未必能经济世务，是以齐桓忘管仲之奢僭而录其匡合之大谋；汉高舍陈平之污行而取其六奇之妙算。苞虽未可以上俦二子，亦今日之选也。”(27)魏晋政权的势族基本就是由战火中锻炼出来的高等世族与这些有“经国才略”的卑微之士组成的。此时，他们的利益又与魏晋中央政权紧密相连了。

    由以上分析可以看到，汉末以来，地方郡姓与中央高等世族经历了不同的道路，不能把二者混为一谈。正始之初，司马懿与曹爽等人同受托孤之任，双方斗争尚未展开。此时，他们事实上行使的是皇权，加强中央对地方的控制是当务之急，而地方郡姓操纵选举显然是与之背道而驰的。因此必须予以打击。

    打击地方郡姓的措施是成功的，但由此导致势族垄断上品却是司马懿始料不及的。正如西晋刘毅所说：“置州都者，取州里清议，咸所归服。将以镇异同，一言议，不谓一人之身了一州之才，一人不审便坐之。”(28)州中正一人说了算是不符合司马懿本意的。司马懿反对地方郡姓操纵选举，但并不反对乡里清议，他所要做的正是要使乡里清议摆脱地方郡姓的控制。然而，这一时期势族正处于向上发展的阶段，加强中央集权的措施在很大程度上被他们改造成一项特权制度。西晋皇权无力根本扭转这一局面，只能在一定意义上加以限制，试图使中正制度不至于完全背离当初创建它的目的。

    与东晋相此，西晋中正主持清议的事例还是不少的。《廿二史劄记》卷八“九品中正”条所载中正清议事例，基本属于西晋时期。这反映出当时皇权还是比较强大的，仅仅根据势族地位而不顾才德定品，在理论上是不能成立的。正是在这样的背景下，才会有前述元康年间举寒素事发生。也是在元康年间，西晋王朝曾发动了一场清议活动。此事《晋书》失载，有幸《通典》保存了这段材料。《通典》卷六○礼二○嘉五周丧不可嫁女娶妇议：

    惠帝元康二年，司徒王浑奏云：“前以冒丧婚娶，伤化悖礼，下十六州推举，今本州中正各有言上。太子家令虞濬有弟丧，嫁女拜时；镇东司马陈湛有弟丧，嫁女拜时；上庸太守王崇有兄丧，嫁女拜时；夏侯俊有弟子丧，为息恒纳妇，恒无服；国子祭酒邹湛有弟妇丧，为息蒙娶妇拜时，蒙有周服；给事中王琛有兄丧，为息稜娶妇拜时；并州刺史羊暨有兄丧，为息明娶妇拜时；征西长史牵昌有弟丧，为息彦娶妇拜时。湛职儒官，身虽无服，据为婚主。按《礼》‘大功之末可以嫁子，小功之末可以娶妇’。无齐缞嫁娶之文，亏违典宪，宜加贬黜，以肃王法。请台免官，以正清议。”……诏曰：“下殇小功，可以嫁娶，俊等简忽丧纪，轻违《礼经》，皆宜如所正。”

    按清议工作，本应由中正主动进行，而此次大规模的清议活动却是在司徒“下十六州推举”的情况下才发生的。这说明中正对清议事不够负责，但也还不能违抗朝廷的命令。清议当否最终由皇帝审批，说明皇权还是有一定力量的。

    综上所述，西晋皇权对势族垄断上品的特权虽不得不认可，但另一方面，皇权还是企图对势族加以限制，这个目的在一定程度上实现了。中正制度在执行中所起的互相矛盾的作用反映出时代的矛盾性。西晋是以后高门世族形成的时期。势族的力量在发展，中正“计资定品”是发展趋势，但势族还不能彻底超越皇权的限制。皇权也还可以有限度地利用中正制度来维护统治秩序。

    三、九品中正制度的作用

    以往的研究者认为，此制度在客观上保证了世家大族的世袭特权，东晋南朝以后，流于形式。根据本文第一节所论述的观点，西晋时，它仅仅是保证了当时的高官显贵的世袭特权，从而在势族的形成以及势族向世族（或称士族）的演变过程中起了重要作用。但只是这样泛泛而论是不够的。因为，单从保障某些高级官吏的世袭特权这一点看，九品中正制并非创举，大家所熟知的汉代的任子制也具有同样的作用。过去，人们在研究九品中正制时，大都将其与汉代的察举制联系起来考虑，这对于探讨中正制度建立的原因无疑是有益的。但是，中正制度在实际运行中既然已经在相当大程度上转化成一种特权制度，它就不再是仅仅与察举制相联系了，而更多的则是与汉代的任子制存在某种继承关系。只有对这两个制度进行比较，才可以更清楚九品中正制的作用。

    任子制与九品中正制虽有相同之处，但也还存在某些差异。首先，在人数上，任子制有严格限制。西汉初年，二千石以上的官吏可以送弟或子到京师为郎官，这叫作任子为郎。《汉书》卷一一《哀帝纪》颜师古注引应劭曰：“任子令者，《汉仪注》：吏二千石以上视事满三年，得任同产若子一人为郎。”东汉安帝在建光元年（121）又下诏发展了西汉的任子制，申明“以公卿、校尉、尚书子弟一人为郎、舍人”(29)。不仅可以任子为郎，而且也可以任子为舍人，这是一个变化。但任子弟一人为官的规定还是一循西汉。在这种制度下，有任子特权的官吏不可能使其后代全部由任子一途入仕。东汉高门世族袁安位至司徒，其子袁敞“以父任为太子舍人”(30)，但另一子袁赏直到袁安死尚未入仕。袁安本传称：袁安死后“数月，窦氏败，帝始亲万机，迫思前议者邪正之节，乃除安子赏为郎”。袁安孙袁汤“桓帝初为司空”，据袁安本传注引《风俗通》云：汤“有子十二人”，但见于记载的只有四人：“长子平，平弟成，左中郎将，并早卒。成弟逢，逢弟隗，皆为公。”(31)袁汤数子入仕，但并不能据此认为他们都是凭借着任子特权。弘农杨氏家族与袁氏家族情况相似，延光三年（124）杨震“因饮酖而卒，时年七十馀……岁馀，顺帝即位，樊丰、周广等诛死，震门生虞放、陈翼诣阙追讼震事，朝廷咸称其忠，乃下诏除二子为郎。”(32)由以上袁、杨家族任子情况看，任子有限额的规定还是执行得比较认真。袁安子袁赏、杨震二子都是在其父死后，按特殊情况授予郎官的。袁、杨家族尚且如此，一般官吏的任子数量也很难超过制度的规定。虽然高官子弟除去任子制度外，还可以从其他途径入仕，如察举、征辟等等，但这毕竟不属于特权制度，其他人士如一般的地方郡姓也可由此途上升。

    与任子制不同，九品中正制建立时并不是一项特权制度，因此也不可能规定高官子弟可以获得上品的人数。没有人数限制而在实际执行中又确实成为特权制度，这就构成了九品中正制度的一大特点。在此情况下，高官子弟大都可以获得上品，步入清途。说得明确些，高官子弟是以族的规模进入政治舞台的，官之为族终于实现了。这在汉代是缺乏保障的。汉代某些高官家族后来演化为累世相承地做官的世家大族，与其说是靠任子制，倒不如说是靠累世通经，察举入仕更为接近事实。魏晋时期，察举制依然存在，但正如严耕望所说：“晋世公卿另有捷径，故即在西晋，汉代经制之秀孝两途已渐不见重视，东晋以下更无论矣。”严氏更引日本学者宫崎市定所述王谢大族不应秀才之举以为佐证。(33)晋代高官子弟对秀、孝两途的不重视正是由于保障其世袭特权的九品中正制没有人数限制。他们不必再以察举制作为入仕的补充手段了。

    制度是对现实的反映，任子制与九品中正制的上述差异表明，汉代高门世族与魏晋以降的高门世族在保障整个宗族的世袭特权方面所具有的能力是不同的。汉代高门世族在皇权、外戚、宦官的限制下还不可能把任子制发展为九品中正制。宗族政治力量有限，在复杂激烈的斗争中要想壮大力量，就必须到本宗族以外寻求支持。史称袁绍能“折节下士”，其目的不过是为了争取“士多附之”而已。不仅袁绍如此，袁氏家族“自安以下，皆博爱宾客，无所拣择，宾客入其门，无贤愚皆得所欲，为天下所归”(34)。汉末袁绍被认为是最有力量的，但这并不是由于自身“四世五公”的空名，而是在于“树恩四世，门生故吏遍于天下”(35)。建安年间，在袁绍家乡汝南“拥兵拒守”，反抗曹操的并不是袁绍的宗族成员，而是“布在诸县”的“门生宾客”。(36)众所周知，汉代的门生故吏与其宗师举主存在着一种类似父子的关系，宗师举主有势，门生故吏可因此飞黄腾达；宗师举主被贬，他们亦同时被贬，宗师举主死后，他们要为之服丧。非血缘关系被罩上了一层宗法面纱。这表明，社会中宗法观念在发展，世族可以借此壮大自己的势力。但另一方面，宗法观念、宗族力量还不够十分发展，盘踞中央的高门世族还不可能使自己的整个家族都不受限制地进入政治舞台。

    魏晋南朝，门生、故吏、宾客依然存在，但他们参加政治活动的记载则不多见了，地位明显下降。(37)高门世族也并不以广召门生、宾客为重要任务，也从来没有人认为高门世族的政治力量是体现在他所控制的门生、故吏、宾客方面。这些变化说明世族自身的宗族力量大大加强了，因此，在政治斗争中，高门世族靠的是本宗族成员占据高官要职，靠的是世族与世族的政治联盟，而联盟的手段则是婚姻。

    以上讨论了任子制与九品中正制不同的一个方面，以及这种不同产生的历史原因。除此之外，任子制与九品中正制还存在另一个不同的方面。汉代的任子制不具有垄断性，除去任子为郎外，拥有赀产十万钱而又非商人者，也可凭赀产为郎，叫作赀选。在察举制下，被举为秀才、孝廉者也多除郎中。此外，还有献策为郎等多种途径。所以，汉代高官子弟不可能垄断郎官。而在九品中正制度下，“上品无寒门，下品无势族”，低等世族很难进入上品之列。高门世族在很大程度上切断了低等世族上升之路。垄断的特征，一方面造成了高低两等世族长期较为稳定的并存局面，另一方面，随着时间的推移，随着门阀政治理论的确立，又必然地出现了族之为官的转变，即某些家族的子弟理所当然地居高位。从依据现实的政治地位以培植本宗族的力量，到依靠族姓地位以巩固自己的力量——官之为族，族之为官，这就是魏晋南朝高门世族所走过的历程。

    综上所述，没有人数限制、封闭性是九品中正制度区别于任子制的关键所在。在此制度下，高门世族的宗族政治力量必然呈现出日益扩张的趋势。毫无疑问，在不断扩大基础上的世袭特权具有更稳固的特征，因为某一分支的衰落不会影响整个宗族政治权力的继续传袭。南朝一些高门世族的家世，往往可以追溯到晋代，其原因必定是复杂多样的，但九品中正制的实行显然是原因之一。

    本文转自《北京大学学报》1987年第1期

  • 刘屹：道荒宏雪岭——重识横跨葱岭的三条古道

    一、问题的提出

    尽管“丝绸之路”的概念，目前看来并非像人们一直以为的是由李希霍芬（Ferdinand von Richthofen, 1833—1905）首创，但李希霍芬仍是最早将“丝绸之路”所经的线路标识在地图上，从而给人以“丝绸之路”确实以某种交通路线状态存在的直观印象之人。李希霍芬主要根据《汉书》的记载，标画出公元前128年至公元150年间的中亚交通路线。在其中，西域南道和北道，分别对应了西越葱岭的南北两条道路：西域北道从疏勒向西，可沿阿赖山脉，进入费尔干纳盆地，再向西抵达撒马尔罕；西域南道则从莎车出发，向西南方向登葱岭，再横穿葱岭上的瓦罕走廊，西去昆都士（Kunduz）和巴尔赫（Balkh）。这很可能是第一张标绘了葱岭东西两侧交通路线的地图。但是，由于李希霍芬本人没有来中国的甘肃和新疆进行过实地考察，他在画这幅中亚彩图时，明显缺乏对葱岭地区实际道路交通状况的充分了解，以至于有的路段画得有些想当然。而李希霍芬这一最早的“丝绸之路”路线图，对后来的“丝绸之路”地图产生了不小的影响。很多由此衍生的“丝绸之路”地图，在涉及葱岭地区的交通路线时，基本上都沿用李希霍芬这一并不准确的描绘。换言之，迄今我们所能看到的“丝绸之路”路线图，在葱岭路段的线路都有很大改进的必要。

    李希霍芬的这幅《中亚地图》还用红线勾勒出一条从地中海东岸一路到中国内地的路线，这是依据托勒密（Claudius Ptolemaeus，约100—168）《地理志》（Geography）所转载的叙利亚商主马厄斯·提提阿努斯（Maes Titianus）所属商队一路东行所留下的记录。这个商队活动于公元前1世纪末或是公元2世纪初，堪称从西方角度关于“丝绸之路”实际道路情况的最早和最重要的记录。1941年，日本学者白鸟库吉（1865—1942）也专门分析了这条商队通行葱岭的道路。白鸟氏受限于当时的条件，对葱岭地区道路的考订也有需要订正的地方。马厄斯商队的记录，对研究“丝绸之路”具有重要的价值，至今仍然受到西方学者的关注。

    此后，关注东西交通、丝绸之路的学者日益增多，但对于葱岭地区道路的考察，仍然是整个“丝绸之路”地理交通研究方面最为欠缺的一环。以笔者有限的知见，只有日本学者桑山正进在研究迦毕试和犍陀罗的历史时，对中国史书记载的求法僧西行求法经行葱岭时的路线，做过一些有益的探索。但葱岭地区的道路并非其研究的重点，因而在整体上相较前人的研究突破性不大。

    虽然中国学者对“丝绸之路”的研究热情经久不衰，但受限于国境线，出国实地考察又极为不便，所以大多数关注葱岭古代交通的中国学者，主要依据的是传世文献记载，只有少数人能够实地考察葱岭地区，但通常也仅限于中国国境线以内的部分。由于人为地截断了葱岭古道的贯通性，对域外的道路交通情况缺乏必要的了解，所以借助这些成果，很难窥见整个葱岭交通道路的全豹。

    近年来，也有一些国内学者努力将视野扩展到国境线以外的地理交通状况，他们的成果极大地弥补了国内学者对葱岭地区境外地理状况和相关研究信息的缺憾。但由于这一领域对国内学者来说长期缺乏必要的前期积累，所以仍留下一些不太准确的描述，或是未能解决的关键性问题。近年来还有勇敢践行域外葱岭古道的中国学者，也为葱岭古道的研究提供了重要的实地考察经验。还有西方学者如傅鹤里（Harry Falk），虽然未曾亲履其境，但善于利用谷歌地图（Google Earth）等现代科技手段，也在探索葱岭古道方面做出了重要推进。本文在利用卫星地图，认定葱岭古道除了传统的南道、北道之外，还有更重要的“中道”等方面，都可说是直接得益于傅鹤里研究的启发。

    总之，对于葱岭这一“丝绸之路”上重要路段的研究，国内外学者一百多年间断断续续地一直在努力推进。但国内学者通常受阻于国境线，对葱岭古道的认识难窥全豹。国外学者往往对汉文史料的理解和掌握存在明显的不足。两方面的研究亟须互为补充，才有可能真正取得对葱岭古道研究的突破性进展。这种突破性一是要建立在对葱岭古道上个别重要地点的重新比定。如对“悬度”位置的重新确认如果可以成立，就会极大强化关于瓦罕走廊在古代葱岭东西两侧交通上重要地位的认识。二是要有对整个葱岭古道的全新认识。以往的研究受李希霍芬的影响很深，以至于似乎横越葱岭的道路只有南北两条，实际上还有一条“中道”更值得重视。而这条“中道”在李希霍芬以降直到今天的各种“丝绸之路”路线图中，却很少得到体现。

    当然，即便能够取得突破性进展，也只是阶段性的推进。毕竟关于葱岭地区道路的研究，将牵涉历史、地理、地质、民族、语言、宗教、国际关系等方方面面。真正综合性的研究仍然有待未来条件具备时才能展开。

    二、“葱岭”与“葱岭古道”

    关于“葱岭”的得名，郦道元《水经注》引佚名的《西河旧事》云：“葱岭在敦煌西八千里，其山高大，上生葱，故曰葱岭。”葱岭上生野葱之说，还见于《水经注》引郭义恭《广志》的记载。葱岭生葱的景象，已得到现代亲履其境者的证实。但葱岭上能够生长野葱的景象，与人们想象中葱岭是终年积雪和寒风凛冽之地，形成鲜明的反差。这不由得令人想到葱岭的地理范围究竟应该如何界定？文献中的记载来自玄奘《大唐西域记》云：

    葱岭者，据赡部洲中。南接大雪山，北至热海、千泉，西至活国，东至乌铩国。东西南北，各数千里。崖岭数百重，幽谷险峻。恒积冰雪，寒风劲烈。多出葱，故谓葱岭。又以山崖葱翠，遂以名焉。

    玄奘所说的葱岭“四至”相当于：北起今吉尔吉斯斯坦的伊塞克湖、塔拉斯一线，南至瓦罕走廊南端的兴都库什山，东起今新疆莎车，西至今阿富汗的昆都士一带。历来谈及古代葱岭的地理范围，都要引述玄奘这一说，并将古之“葱岭”与今之“帕米尔高原”相对应。然玄奘所说的“葱岭”范围，与现代地理概念上的“帕米尔高原”并不完全重合。玄奘之所以用乌铩、活国、热海、千泉、大雪山来界定葱岭的四至，是因为这些地方都是他经行过的。玄奘是历史记录中为数不多的，几乎绕着葱岭走过一圈的旅行者。但对于没有这样旅行经验的人来说，未必也能想象得到，或是都认同玄奘关于葱岭四至的说法。所以，虽然玄奘给我们留下了关于葱岭四至的宝贵记录，但这一记录具有他强烈的个人色彩，需要我们谨慎看待。

    例如，乌铩和活国这两个地点，一东一西，都在帕米尔高原以下的地势平缓、海拔较低地区，理论上就不应属于帕米尔高原。现代地理概念上的帕米尔高原，北部应以外阿赖山脉（Trans-Alay Range）为界，以北就进入费尔干纳盆地（Fergana Valley）了，属于另一个地理区域。东部一般以公格尔峰（Kongur Tagh）一带的西昆仑山脉为界。西部一般以喷赤河（Panj River）自南向北流的河段为界。这三个地理方位上的界线，都与玄奘所言不符。只有玄奘所谓“南接大雪山”，即葱岭的南界应在兴都库什山与喀喇昆仑山之间的连接山脉，与现代地理学概念上的帕米尔高原的南界是符合的。“大雪山”以南就是印度河流域的上印度河谷地带（即印巴争议的克什米尔地区，巴控的吉尔吉特—巴尔蒂斯坦地区，Gilgit-Baltistan），属于另一个地理区域。但现在无论是学界还是社会公众认知，往往把上印度河谷地带也算作葱岭或帕米尔高原的范围。这是需要澄清的。况且，“帕米尔高原”的得名，是由于高原上有所谓“八帕”。这“八帕”的地理范围也不包括上印度河谷地区。因此，如果把“葱岭”界定为今天的帕米尔高原，则其东缘为西昆仑山，西缘为南北流向的喷赤河，北缘是外阿赖山，南缘是兴都库什山。本文讨论的“葱岭”，也主要是指这个地理范围之内。

    所谓“葱岭古道”，本指所有跨越葱岭地区的道路。这些道路既有东西向，也有南北向，而且彼此间犬牙交错，并非呈规则性的直线分布。就“丝绸之路”研究的关注点而言，本文主要讨论东西方向上横跨葱岭的道路。最早李希霍芬标示出了南、北两条路线，现在则应该按照方位，进一步将葱岭古道分为“北、中、南”三条道路。

    葱岭北道，即从今新疆伊尔克什坦口岸西行，进入今塔吉克斯坦境内的阿赖山谷。这条道路早在《汉书·西域传》就有体现：

    休循国，王治鸟飞谷，在葱岭西，去长安万二百一十里。户三百五十八人，口千三十，胜兵四百八十人。东至都护治所三千一百二十里，至捐毒衍敦谷二百六十里，西北至大宛国九百二十里，西至大月氏千六百一十里。民俗衣服类乌孙，因畜随水草，本故塞种也。

    “休循”是从伊犁河流域迁来的塞人所建立的国家。“鸟飞谷”或是指阿赖山谷。如果要说个更具体的地点，应在阿赖山谷的萨雷塔什（Sary Tash），这里也是北上进入费尔干纳盆地的重要岔路口。“捐毒”也是见于《汉书·西域传》的塞人小国，位于与休循接壤的东边，应该在今新疆境内。可见，捐毒和休循就扼守了这条东西方向上横穿阿赖山谷的葱岭古道“北道”。阿赖山谷非常宽阔，水草也多。走这条路既可北上进入费尔干纳盆地的塔吉克斯坦奥什（Osh），也可西行至杜尚别（Dushanbe）。《汉书》既然说从休循“西至大月氏千六百一十里”，说明当时通过这条路是可以通往已经迁徙至阿姆河以北地区的大月氏，自然也包括中亚传统的粟特地区。因而这条“葱岭北道”，在古代主要是从西域北道向西的天然延伸，从西域经此路可去往费尔干纳盆地和阿姆河北岸的粟特地区、阿姆河南岸的巴克特里亚地区（吐火罗）。

    李希霍芬标示出的这条路线，此后也成为丝路交通路线图上关于葱岭地区道路最有代表性的一条。近代以来的外国探险家，如斯文赫定、斯坦因、伯希和等人，也都曾经由这条道路进出中国。但是傅鹤里对这条道路的实际利用率提出质疑，认为在古代很难见到有通行这条道路的记载，这是因为这条道路的降水（雪）量大，又盗匪横行，所以不应作为横穿葱岭的主要道路来看待。他这个意见是有偏颇之处的。

    葱岭中道，即从今塔什库尔干出发，向西南行，而非向南行，越过纳兹塔什山口（Nezatash Pass），进入今塔吉克斯坦的穆尔加布（Murghab）地区，西行至霍罗格（Khorog），再前往阿富汗的法扎巴德（Fayzabad）、昆都士一带。这条路古代主要是从西域通往古代的吐火罗地区，即今天阿富汗北部地区。因这条路的西段有衮特河（Gunt River），故又被称为“衮特路”。衮特河发源于雅什库里湖（Yashilkul，汉文史籍中葱岭上的“三池”之一），自东向西流，与喷赤河交汇处，即霍罗格。这一地带在历史上被称作“识匿”或“赤匿”，即今天的舒格楠（Shighan）地区。这条路以往几乎不被学界所重视，讨论到与这条路相关的历史记录，也大都没有意识到这是一条完全可以单独列出的重要通路。直到近年，傅鹤里才强调了这条路的重要性。在沙俄和苏联时期，从杜尚别经霍罗格到穆尔加布，最终到奥什，修筑了今天帕米尔高原上唯一一条连续贯通的高原公路（原M41）。这条路部分修建于19世纪末沙俄与英国对中亚展开争夺的“大博弈”时期，部分修建于1930年代，居然沿用至今，成为一条几乎横贯帕米尔高原的公路（不含中国境内的塔克敦巴什帕米尔）。通常情况下，现代公路往往就是沿着古代交通路线而修建的。由于这条公路并未连接到中国的边境线，所以国内学者一般对这条路没有给予充分重视。

    葱岭南道，也是李希霍芬根据《汉书》的记载大致勾勒而出的。或从塔什库尔干出发，或从于阗的皮山出发，皆可行至瓦罕走廊的东端入口，再自东向西，横穿大部分属于今天阿富汗境内的瓦罕走廊。到瓦罕走廊的西端，既可沿兴都库什山继续西行抵达阿富汗的喀布尔、巴米扬、贾拉拉巴德地区；也可从兴都库什山的几个山口南下，经巴基斯坦的奇特拉尔(Chitral)，前往斯瓦特、白沙瓦一带。这条路古代主要是从西域通往巴米扬、迦毕试、犍陀罗等佛教圣地，因而在历史记录中出现的频率较高。也是以往学者们关注度最高的一条道路。

    古代的行旅不是今日的旅游，尤其是翻越葱岭这样的高寒高原地区，一定要充分准备，精选线路。除非有必须要绕远才能到达的特定目的地，否则一般不会选择绕远的道路。葱岭古道上这三条道路的选择，主要是根据旅行者的出发地和目的地来定。例如，从中亚粟特地区出发的商人和商队，大概率会选择“葱岭北道”进入西域。这对于他们是最便捷的道路。但中国的求法僧西行求法，却基本上不会选择这条“北道”。因为求法僧要去兴都库什山以南的犍陀罗和印度，选择“葱岭南道”或上印度河谷的道路，才是近便的道路。求法僧如果走“葱岭北道”去犍陀罗和印度，就要先到粟特和吐火罗地区，再南下兴都库什山，这样的选择与从“葱岭南道”西出瓦罕走廊后就从兴都库什山口南下相比，无疑是费时和绕远的。历史上只有个别的求法僧为了去被誉为“小王舍城”的巴尔赫去参礼，才会选择这条路。

    此外，民间商人和商队的活动一般是很难进入古代历史记录的。“葱岭北道”与“葱岭南道”相比，的确很少见到有经行此路的历史记录。求法僧主要选择“葱岭南道”西去东归，因而对于“南道”留下较多的记录。商人和商队则有强烈和明确的逐利意识，在安全有保证的前提下，他们不会选择需要绕远、增加运输和时间成本的道路。粟特商人当然不会只走“葱岭北道”，他们也曾在上印度河谷地区道路的岩刻中留下过踪迹。既然如此，就不能排除粟特商人也会经行过“葱岭中道”和“南道”的可能性。这完全要看他们商业活动的目的地是哪里。像葱岭这样特殊地理环境下的道路，自古至今一直都在那里存在，很多路段甚至千百年来也几乎没什么变化。不能因为没有，或很少见到某条道路的历史记录，就认为这条道路的利用率不如那些频繁见诸记载的道路要低。也不能因为某条道路的记载在某个特定时期明显多于另一条道路，就认为两条道路之间存在此消彼长的兴衰轮替。

    三、中国古人对“葱岭古道”的经行

    笔者已尝试按照朝代先后的顺序，梳理了中国古人经行葱岭古道所留下的历史记录。在此则按照葱岭上三条古道的地理方位，重新爬梳一下这些记录，以期加深对这三条葱岭古道在历史上分别被使用情况的认识。

    有记录的、最早通行“葱岭北道”的中国人，应是张骞。他在第一次出使时，被匈奴扣押十多年后逃脱，继续西行，就是经鸟飞谷至大宛。也就是从疏勒向西，进入阿赖山谷，从萨雷塔什转而向北，进入费尔干纳盆地。然后张骞应从盆地的西侧进入康居所在的索格底亚那地区，再南下到阿姆河北岸的大月氏王庭，进而渡河到阿姆河南岸的蓝市城（巴克特拉，Bactra）。当时大月氏已经征服“希腊—巴克特里亚王国”，但王庭尚未迁至蓝市城。在张骞返国后，大月氏王庭才南迁到巴克特拉。当张骞返国时，特意要避开匈奴在西域的势力范围，所以他不会再走经行鸟飞谷的来时路，既可能走“葱岭中道”也可能走“葱岭南道”。总之东归途中下了葱岭，就选择走经过于阗的西域南道，一直到羌中地区才又被匈奴捕获。

    此外，李广利伐大宛，史书虽未记载其具体的出征路线，但从西域进入费尔干纳盆地，十有八九是要从葱岭北道的衍敦谷、鸟飞谷进兵。这是中国历史上第一次派遣远征军经行葱岭北道，尽管只是走了半途，就转而北上费尔干纳盆地。陈汤攻伐郅支单于时，有“三校从南道逾葱岭径大宛”，也应走的是阿赖山谷这条道路。不过，这些军事行动都不能算是横穿“葱岭北道”。其他可能有大量的粟特商胡是通过“葱岭北道”进入西域乃至中原内地的，只是我们现在看不到直接的文献记录而已。

    2.葱岭中道

    至于“葱岭中道”，前述叙利亚商主马厄斯所属下属的商队，从巴克特拉出发，经葱岭的Komedoi地区，即汉文的“识匿”地区，抵达“石堡”，即塔什库尔干。如果是走“葱岭北道”就无需在“石堡”停留。所以应该是走的“巴克特拉—霍罗格—雅什库里”一线，再通过纳兹塔什山口，抵达塔什库尔干。这虽然不是古代中国人经行的记录，但可以证明这条“葱岭中道”在当时的确是商队经常会选择的一条道路。此外，《汉书·西域传》记载西汉末年时，从皮山出发，经“悬度”到罽宾的途中，将会经过葱岭上的“三池”。这“三池”就是帕米尔高原上六大湖泊中比较靠南的三个，即最南的切克马廷库里（Chaqmaqtin-kul）、中间的佐库里（Zorkul，又名萨雷库里，Sirikul，汉文史籍称“大龙池”）和靠北的雅什库里。雅什库里也是前述衮特河的发源地。可见早在西汉时，汉使已有经行雅什库里的经验。汉使走雅什库里这条路，不仅仅是为了就近水源，因为另两个淡水湖就在“葱岭南道”的途中，完全没必要为了取水而绕远走到雅什库里。而选择经过雅什库里的道路，就意味着前行是要去往霍罗格一带。从霍罗格可以选择向北去阿姆河北岸的粟特地区，还可以南下至伊什卡申，继而向西去往吐火罗的法扎巴德、昆都士；或向南越过兴都库什山，南下犍陀罗。因此，所谓“三池”的记录，实际上就暗示了葱岭的“中道”和“南道”都已有汉使经过。

    在公元1世纪末，贵霜新继位的君主因向东汉求娶公主，被拒，遂由“副王谢”率七万大军进攻西域，围攻疏勒未成而退兵。要动用7万大军穿行葱岭，需要尽可能在葱岭西部（霍罗格和伊什卡申一带都是从西向东横穿葱岭前必要休整、准备的据点）获得足够的给养，再选择相对比较适合大军通行的大路。考虑到当时贵霜都城不是在巴克特拉，就是在犍陀罗的弗楼沙（白沙瓦前身），贵霜军不太可能先北上到阿姆河，再通行阿赖山谷进入西域。他们应是先进入葱岭，到霍罗格和伊什卡申一带，再沿“葱岭中道”至塔什库尔干，这是最有可能的线路。至于“葱岭南道”虽然也可以通行，但要让7万大军鱼贯穿行瓦罕走廊的狭长地带，在军事上恐非明智之选。

    到6世纪初，宋云、惠生出使的去程中，葱岭一段的路程，走的是“汉盘陀—钵和—嚈哒王庭（昆都士）”。以往的研究，包括经常被引用的桑山正进所画的宋云使团的行程路线图，也没有体现出宋云等人的去程走的应该是“葱岭中道”。本文想强调的是：宋云使团很可能是经过“葱岭中道”，而非走“葱岭南道”的瓦罕走廊后，抵达嚈哒王庭所在的昆都士。首先，在宋云使团的记录中，也明确提到了“三池”。如果只走横穿葱岭的单程，这“三池”是没必要都要走到的。宋云很可能去程经过雅什库里，回程则经过佐库里。其次，“汉盘陀”即“渴槃陀”，亦即塔什库尔干。宋云等人从塔什库尔干出发，也是经过纳兹塔什山口，从塔克敦巴什帕米尔进入到小帕米尔。这时既可以向北走“中道”，也可以向南走“南道”。由于“宋云行记”中记载了“波知国，境土甚狭，七日行过。”这应是指瓦罕走廊的狭长地带，七天就可走完。而且波知国只有“二池”，应是指佐库里和切克马廷库里。如果“波知国”指的是瓦罕走廊，则“钵和国”就不可能还在瓦罕走廊上。所以，“钵和国”合理的位置应该在“葱岭中道”上。宋云使团出使的首要目的地是位于昆都士的嚈哒王庭，走“葱岭中道”不仅路途最短，而且路况也比较好走。

    此后，明确走“葱岭中道”的，还有8世纪中期至末期的车朝奉（730—812）。他于751—790年间也游历葱岭东西，并在“罽宾”出家，“悟空”是其法号。回国后，将其经历口述，由圆照于795年记录，作为贞元新译《十力经》《十地经》等经的序，收入大藏，亦名《悟空入竺记》。《悟空入竺记》记载其去程经过葱岭时的路线是：疏勒—葱山—杨兴岭—播蜜川—五赤匿国—护密。这其中，“葱山”应即唐朝在葱岭东部的重要据点——葱岭守捉，或曰“葱岭镇”。亦即说车朝奉一行是从疏勒西登葱岭，到达葱岭镇（塔什库尔干）。“杨兴岭”很可能是纳兹塔什一带的山口，因为“播蜜川”是佐库里湖所在的峡谷，从塔什库尔干到佐库里之间，相对有标识度的山岭，就是纳兹塔什山口。不只是车朝奉，玄奘和慧超也都是经过播蜜川后抵达塔什库尔干的。这说明经行佐库里的道路相对于经行切克马廷库里的道路要更经常被使用。其实这与“石山悬度”的位置有关。因为走切克马廷库里向西通行瓦罕走廊，就一定要经过“石山悬度”；反之，若选择走石山“悬度”东去塔什库尔干，也一定会通过切克马廷库里。最初汉使通罽宾时，之所以走石山“悬度”这条险路，是因为距离最近。后来随着对葱岭地区道路认识的加深，可以替代“悬度”的道路也会出现。但如果像玄奘那样由大象驮着经书，是肯定不会选择“石山悬度”，也就不可能走切克马廷库里之路。车朝奉一行在经过播蜜川后，经过“五赤匿国”。“五赤匿”就是“五识匿”，即是今塔吉克斯坦的舒格楠一带，属于“葱岭中道”的西段。然后从“五识匿”南下到“护密”，亦即“胡蜜”，这是瓦罕走廊西端，今伊什卡申一带。关于“五识匿”和“护密”的位置关系，还可通过慧超《往五天竺国传》得到清晰的理解（详见下文）。这也说明“葱岭中道”与“葱岭南道”之间并非截然分隔，车朝奉一行就是先走了“南道”的东段，然后又走“中道”的西段，再从“中道”回到“南道”的西段。因为他们的目的地是去罽宾犍陀罗地区，所以最终要从瓦罕走廊西端南下。

    至于车朝奉在返程经过葱岭时，他走的是“拘密支—若瑟知国—式匿国—疏勒”一线。其中“拘密支”，Komidai，玄奘记作“拘谜陀”，又作“居密”“俱蜜”，位于葱岭的西部，五识匿地区之北。可见车朝奉还是由葱岭西部向东，经过式匿国，抵达疏勒。其中省略了从式匿到葱岭镇的路段，应该是与去程相差不大，所以没什么特别可记的。车朝奉之所以来去都选择了“葱岭中道”，很可能是因为吐蕃势力已经浸染到上印度河谷的大、小勃律，乃至瓦罕走廊有时也被吐蕃所控制。这种情况下，走“中道”比走“南道”会安全一些。

    此后，清乾隆年间平定大小和卓之乱时，清军追击叛军在葱岭北部的喀拉湖、穆尔加布和雅什库里，与叛军激战，三战三捷。平定叛乱后，乾隆命人在雅什库里湖边树立《平定回部伊西洱库尔淖尔勒铭碑》。这也是最远的一座“乾隆纪功碑”。雅什库里一带可以作为战场，双方投入万人以上规模的部队作战，也说明这一地带相对瓦罕走廊更为开阔，更适合展开大规模的军事行动。

    3.葱岭南道

    以往的研究，有一种从整体上忽视葱岭古道在丝绸之路东西交通上的重要性的倾向。如桑山正进认为：原本上印度河谷道路是中印之间交通的主要通道；由于种种原因，上印度河谷道路被通行瓦罕走廊、走兴都库什山北麓的道路所取代，导致巴米扬地区开始建造大佛像。其实如果梳理历史记录就会发现，葱岭古道很可能较之上印度河谷道路更重要，持续发挥作用的时间也更长。见于历史记载的选择走“葱岭南道”的行者，似乎远多于上述“北道”和“中道”。因而“葱岭南道”一直是西域通往中亚和印度的主干道，甚至上印度河谷道路最兴盛之时，也无法与“葱岭南道”分庭抗礼。

    早在公元前130年左右，从伊犁河流域被大月氏赶出故地的塞人，在“塞王”的带领下，“南越悬度”“南君罽宾”。因为“悬度”已经可以比定为在瓦罕走廊上的一段石山险路，所以塞人就是从塔里木盆地西缘出发，通过瓦罕走廊，实现横穿葱岭，再南下去攻占犍陀罗地区。这也是从葱岭东侧的西域出发，去往葱岭西侧的犍陀罗地区最短的一条路径。因为塞人骑兵要对犍陀罗的希腊人政权发动突袭，所以不可能选择上印度河谷地区那种“悬絙而度”的绳索桥，也不可能在河谷山坳中绕来绕去浪费时间。石山“悬度”虽然凶险，但不是不可逾越。所以塞人进占罽宾，就是通过快速穿越“葱岭南道”的瓦罕走廊而实现的。

    塞人占领犍陀罗地区，建立起塞人的罽宾王国。到西汉末，大批的西汉国使和护送所谓“罽宾使者”回国的汉军将士，都是经历“悬度”险路完成使命的。这其中，只有文忠和赵德等极少数人在历史上留下了姓名。汉使和汉军的马匹不适合葱岭上的高原险路，通行“悬度”时损失较大。所以杜钦建议汉朝放任罽宾，不再参与其国政事；罽宾再有来使，汉朝只负责将其护送到皮山即止，不要再冒着危险将所谓的“罽宾使者”护送回罽宾。这样就避免了在通行“悬度”时的无谓牺牲。

    公元97年，甘英从龟兹出发，“逾悬度，乌弋山离”，去往大秦。既然“逾悬度”，显然也是走了“南道”的瓦罕走廊。因为这样走，出了瓦罕走廊，再沿着兴都库什山西行，就可到达乌弋山离。可以说是最近的道路。甘英无需去往兴都库什山以南的地区，所以南北朝的求法僧才说甘英不曾走过上印度河谷的绳索桥和傍梯险路。

    根据僧传的记载，公元4—5世纪，法显、智猛、昙无竭等大部分求法僧，都是从塔什库尔干南下，不去横穿瓦罕走廊，而是在瓦罕走廊东段的山口，就南下到上印度河谷地区。选择这样的路途，主要是为了去陀历国（Darel,达丽尔山谷）参拜陀历大像。而且通过口耳相传，使得这条路成为南北朝时期大多数求法僧都会选择的道路。但这条路并不能一直保持畅通，如果发生地震，形成堰塞湖，就会破坏道路交通，乃至有的路段会断路两三百年之久。这也就是为何还会有个别求法僧，如北魏的道荣，仍然会在去程和回程都选择走“葱岭南道”。

    大魏使者谷巍龙的题字出现在乌秅，而其出使的目的地是粟特地区的“迷密”（米国）。这并不意味着谷巍龙接下去会沿印度河谷道路一路到犍陀罗，之后北上兴都库什山，经过吐火罗地区，再到粟特地区。前述《汉书·西域传》就有这样的道路，即从乌秅西行会经过“石山悬度”。而要从乌秅西行到“悬度”，就要通过乌秅西北的山口进入到瓦罕走廊东端，再向西经过悬度，横穿瓦罕走廊。到走廊的西端，或者继续西行，就是当年甘英去往乌弋山离的路线，只不过谷巍龙还要继续从乌弋山离北上粟特地区。或者从瓦罕走廊西端沿喷赤河北上，再转西，都可以抵达粟特地区。谷巍龙之所以没选择“葱岭北道”，或是直接从西域北道或南道西上葱岭，大概是因为与北魏敌对的柔然势力控制着西域北道，所以谷巍龙走了西域南道，且从于阗南下到拉达克地区，再转向乌秅。

    520年左右，宋云完成了觐见嚈哒王的使命，带着北魏使团，携带170部佛经，回国复命。因为他是从乾陀罗，即罽宾犍陀罗之地返国，自然会走从犍陀罗去西域的传统道路，那就是“葱岭南道”。宋云带那么多佛经，驮畜行走“悬度”不易，故其返程很可能也是从帕米尔河与瓦罕河交汇处的Gaz Khun村就转而向北，绕开“悬度”，经行佐库里所在的波谜罗川，再抵达汉盘陀（塔什库尔干）。

    大约100年后，玄奘的回程，也是从瓦罕走廊西端开始横穿走廊，经过达摩悉铁帝国（瓦罕走廊西部，汉杜德）、波罗蜜川（播蜜川、大帕米尔）。即从帕米尔河与瓦罕河交汇处的Gaz Khun村以东，就选择相对好走一些的经行“大龙池”（佐库里）道路，大体上走的是“南道”。

    距玄奘经行葱岭差不多一百年，723—727年间，新罗僧慧超，也在从天竺返回唐朝的路途中，走了葱岭古道。他具体的路线是：胡蜜—识匿—葱岭镇。《往五天竺国传》云：

    又从吐火罗国东行七日，至胡蜜王住城。当来于吐火罗国。逢汉使入蕃。略题四韵取辞。五言：君恨西蕃远，余嗟东路长。道荒宏雪岭，险涧贼途倡。鸟飞惊峭嶷，人去难偏梁。平生不扪泪，今日洒千行。

    “胡蜜”又称“护密”或“休密”，本是贵霜时期的五翕侯之一，应该镇守的就是瓦罕走廊西端的伊什卡申一带。慧超到胡蜜时，恰逢“汉使”即唐朝的官使经行胡蜜去往“西蕃”。具体是谁，要出使哪国，都已不可知。就在这域外雪岭之地，两个从东土大唐来的旅人，一个西去，一个东归，意外相遇，而又都喜好汉语诗文，遂以诗相酬，共同抒发在域外偶遇知音、怀念故乡的悲情愁绪。此后，慧超记载了他没有亲履其地，而是听闻传说的“识匿国”：

    又胡蜜国北山里，有九个识匿国。九个王各领兵马而住。有一个王，属胡蜜王。自外各并自住，不属余国。近有两个王，来投于汉国，使命安西，往来〔不〕绝。……彼王常遣三二百人，于大播蜜川，劫彼兴胡，及于使命。纵劫得绢，积在库中，听从坏烂，亦不解作衣著也。此识匿等国，无有佛法也。

    通常认为这里的“九个识匿国”“九个王”应该是“五个识匿国”“五个王”之误。五识匿地区就是今天的舒格楠地区。五识匿中，有的归属胡蜜，有的归顺唐朝，与安西都护府来往频密。但当时唐朝最西境，就是下文提及的“葱岭镇”，亦即“葱岭守捉”，今天的塔什库尔干。从“葱岭守捉”向西，就是识匿地区。应该是比较靠东的两个识匿王更乐于与唐朝往来。“大播蜜川”即玄奘东归时经过葱岭的“波谜罗川”，亦即佐库里湖。这是说五识匿国经常派人劫掠来往的“兴胡”，即通过经商兴利的胡商，主要是指粟特商人。说明粟特商人显然也是经常通行佐库里所在的“葱岭南道”。不仅劫掠胡商，包括来往的国使，也不放过。故前文有诗云：“险涧贼途倡。”这种劫掠行为属于识匿国的“国家行为”，他们劫得大量的“绢”，也不会用来制作衣服，还是习惯穿他们传统的皮裘之衣。实际上在葱岭这样苦寒之地，丝绸、绫绢之类的原料不可能被用于制作当地人的衣服。由此可见，丝绸的确是胡商冒险经行此路运营的主要货品。而识匿国不信佛法，故慧超也不会选择“中道”。慧超选择的道路是：

    又从胡蜜国东行十五日，过播蜜川，即至葱岭镇。此即属汉。

    慧超从瓦罕走廊西端的胡蜜，一路东行，就是走瓦罕走廊，经播蜜川（佐库里），抵达葱岭守捉所在的塔什库尔干。这也是开元时期唐朝西境的极限了。

    公元747年，高仙芝征讨小勃律之役，其大军从龟兹出发，上葱岭后，《旧唐书》记云：

    又二十余日至葱岭守捉，又行二十余日至播密川，又二十余日至特勒满川，即五识匿国也。仙芝乃分为三军：使疏勒守捉使赵崇玭统三千骑趣吐蕃连云堡，自北谷入；使拨换守捉使贾崇瓘自赤佛堂路入；仙芝与中使边令诚自护密国入，约七月十三日辰时会于吐蕃连云堡。堡中有兵千人，又城南十五里，因山为栅，有兵八九千人。城下有婆勒川，水涨不可渡。

    小勃律即吉尔吉特。此前唐军曾三度征讨，都未获胜。天宝六载，高仙芝率一万大军从安西都护府（龟兹）一路西行百日，登上葱岭。在葱岭守捉休整后出发，并未直接从瓦罕走廊东端南下巴罗吉尔山口和达尔科特山口去进攻吉尔吉特，而是直接挥师西进到“葱岭中道”西段的五识匿国地区。“特勒满川”一般认为是帕米尔河。此前唐朝已使位于瓦罕走廊西段的护密国归降，故高仙芝此行并非去攻占五识匿和护密，当地应有亲唐势力接应唐军。他也无需带领一万大军全数西进五识匿地区，应该早在葱岭守捉休整时，就定好分进合击的战术：赵崇玭从“北谷”进军吐蕃占领的连云堡（即萨尔哈德）。所谓“北谷”应即从佐库里一带穿行山谷能够抵达萨尔哈德的道路。今天从萨尔哈德出发，如果不想走石山“悬度”之路，就要向北绕远穿行山谷，也可去往佐库里或切克马廷库里。贾崇瓘则走“赤佛堂路”，有说是在瓦罕走廊东段从帕米尔去往贾帕尔桑河谷的道路。“赤佛堂”的地名或许和《汉书·西域传》所记的“赤土身热之阪”有关。亦即说贾崇瓘这路唐军负责从切克马廷库里这一路夹击连云堡。无论“赤佛堂路”具体地点何在，都不影响学者们认为贾崇瓘这一路实际上是唐军攻击连云堡的“东路军”。赵崇玭和贾崇瓘这两路，不可能是唐军到了五识匿后再回过头去走“北谷”和“赤佛堂路”，应是高仙芝率军绕行到五识匿和护密去实施战略迂回，留下另外两军分别从北面和东面，约定日期，合击连云堡。连云堡南十五里还有吐蕃的一座城寨，下有“婆勒川”。姚大力认为“婆勒”就是Baroghil的音译。在萨尔哈德向南翻越巴罗吉尔山口时，当时吐蕃也派重兵把守。唐军攻下连云堡后继续南下吉尔吉特，就不在本文讨论的葱岭道路范围。总之，高仙芝打小勃律之前，先要拔掉从“葱岭南道”南下小勃律的必经之地连云堡。但如果直接走瓦罕走廊，从东向西进军连云堡，一旦被吐蕃扼守住石山“悬度”，大军就无法前进。高仙芝采取的是通过“葱岭中道”迂回到连云堡的北方和西方，再实现三面合击的战术安排。由此也可见“中道”与“南道”之间存在紧密的关联性。

    此后，随着“安史之乱”的爆发，吐蕃不仅反攻夺占了葱岭古道上的“中道”和“南道”，而且唐朝连西域、河西诸地也逐渐丧失。中原人出于各种政治、军事或是信仰的目的，艰难跋涉于雪岭葱外的时代，遂暂告一段落。

    四、结语

    以上通过将“葱岭古道”细分为“北道”“中道”和“南道”，并将历史上与葱岭有关的每个历史事件和每个具体的旅行者事迹，还原到“葱岭古道”具体的每一条道路上去。这样做希望可以加深我们对这些人物和事件的理解。

    例如，玄奘返程中经过的“大龙池”到底是佐库里，还是切克马廷库里？只要考虑到“石山悬度”的位置，就不难确认“大龙池”一定是指佐库里，因为这条路相对于经过石山“悬度”才能抵达的切克马廷库里之路，要好走得多。再如高仙芝征伐小勃律之战，按照以往的看法，唐军似乎是从瓦罕走廊东端直接进军连云堡，再南下坦驹岭的。但这样一来，唐军必须要经过石山“悬度”才能抵达连云堡。这对上万人的远征部队而言，肯定是危险的选择。高仙芝之所以能够成功，与此前夫蒙灵詧替他打通了护密道路有很大的关系。这使得高仙芝的军队可以得到瓦罕走廊西端护密国的支持，甚至五识匿地区也不会给唐军制造麻烦。所以高仙芝能够采取迂回到连云堡以西，从东、北、西三面合击连云堡的战术。这一点似乎是以往研究高仙芝征伐小勃律的学者都没有意识到的。

    此外，还可得出以下几点关于“葱岭古道”的全新认识：

    其一，从地理上说，葱岭的四至应以今帕米尔高原为界，玄奘的记录并不符合葱岭的实际情况。上印度河谷地区在地质板块上属于兴都库什山以南的印度板块，不属于帕米尔高原的范围，应排除在“葱岭古道”之外，单独作为一个研究对象。

    其二，“葱岭古道”进一步应划分出“北、中、南”三条道路。这其中，“北道”与“中道”和“南道”相比，具有一定的独立性。或者说从“中道”和“南道”较难在东西横向通道上与“北道”产生关联性。但“中道”和“南道”之间，则往往可以根据需要进行穿插经行。实际上，葱岭上的道路组合是多样化的，不是简单的三条线能够涵盖的。古人根据出发地和目的地的不同，可以灵活选择自己要走的道路。但基本的原则是会选择保证安全和距离短、耗时少的路程。玄奘之所以在回程中选择走葱岭而非去程时走天山以北再到中亚粟特地区的道路，就是因为正常情况下从西域到印度去的道路就应该走葱岭古道。

    其三，所谓“瓦罕走廊”，只是葱岭上的“南道”而已，不应被视作通行葱岭南部地区的唯一选择。与之相比，霍罗格与塔什库尔干之间的葱岭“中道”在历史上所起的作用，可能更值得我们关注。今后的“丝绸之路”路线图在经过葱岭地段时，至少应该画出三条东西横贯的路线，而不是只有两条。

    本文转自《中华民族共同体研究》2024年第4期

  • 杨联陞：传统中国对城市商人的统制

    本文主旨，在就传统中国政府对城市商人之统制（包括控制与利用），提出若干看法，以供讨论。所谓商人，系用广义，一切行商坐贾、铺户店号，乃至当铺钱业牙行，均在讨论之列。所谓城市，亦取广义，兼指城镇，不论大小。所谓传统中国，时限可长可短。在本文多指帝国时代末期，自清初至鸦片战争一段，但亦有时兼及前后。

    中国传统，远自二千余年以前，早已以农为本，视工商为末业，政府对四民之待遇，因有重轻。然就全帝国时代而言，亦不可一概而论。如《史记》、《汉书》所载，政府对商人之统制，包括贾人有市籍，不得为吏，不得名田，重其租税，乃至其车马服饰，亦受限制。此种政策，虽起于汉初（或更早），至武帝时，因财政关系，已有孔仅、桑弘羊等，由市井跃登朝列。其他限制，似亦渐成具文。此后在理论上，虽仍轻商，实则对于商人之控制与利用，力图兼顾。唐、宋以来，此种情形，更为显著，议论亦略有改变。读史者当就各时代分别观之，始能得其真象。如就清初至中叶一段论之，则对商人之控制，已不甚严，租税负担，亦非特重，政府且颇以恤商自许。利用则积前代之经验，特重“保”（如保商、保结、连环保）“包”（如包办、包额）诸术，颇有成效。

    在清代商人入仕，远较前代为易。在隋、唐与辽代，工商及其子弟，均不得应科举。但此限制至北宋已见宽弛。据《宋会要·选举》，庆历四年（1044年）定“诸科举人，每三人为一保，所保之事有七”，其七为“身是工商杂类及曾为僧道者”并不得取应。细玩“身是”与“曾为”字样，则不但工商子孙可以应举，即曾为工商而今已改儒业者，似亦可以应举。更早者为淳化三年（992年）所定，“如工商杂类人内有奇才异行卓然不群者，亦并解送”。虽属特例，已开商贾应举之门矣。

    金元时代，对商人应科举，似乎已无限制。明清更有所谓“商籍”，专为盐商子弟在本籍之外盐商营业之地报考生员，而且特为保留名额。据何炳棣教授之计数，盐商子弟，成进士者，明代近一百九十人，举人三百四十人。清代进士至乾隆之末，已达四百二十余人，举人八百二十余人，其中在18世纪，人数尤众。按明清商籍，盖仿元代河东之运学运籍。当异族入主之世，商人往往特受优待，亦可注意也。

    科举之外，尚有捐纳一途，为富商入仕之捷径。清代捐纳制度，近人已有专书详论。在清代主要自为财政关系，然如雍正上谕所言，捐纳进身，可救偏重科举之弊，则其中亦不无政治意味也。

    宋、元以降，商人入仕之途渐广，此与一般社会经济之发展，关联自极密切，在思想上，亦有反映。如宋元儒者，已不讳言治生，明末黄梨洲，已有工商皆本之论，清代沈垚（《落帆楼文集》）更谓“古者四民分，后世四民不分。古者士之子恒为士，后世商之子方能为士。此宋明以来变迁之大较也”。其言虽近于偏激，亦有相当根据。

    秦汉所谓市籍，至少延至唐代。中唐以后，政府对于市场之管制，大见松弛，对商人之特别注籍，似亦不及以前之注意。明代户籍，分军民匠灶四大类。商人似亦属于民户。清代《嘉庆会典》有“军民商灶”之别，然此所谓商，即上文商籍之商，专指盐商而言，不得误解为一般商人。惟以商人当行及纳税（如门摊、铺税等）之故，政府对于孰为商人，及各商资力之大小，亦当有相当了解。保甲调查，亦分住户铺户，此在19世纪之纪录特为显著，京师所在，固不待言，如《津门保甲图说》（1846年）所记天津各区人户，分类详细，数目似亦相当可信也。

    政府就商人收取关卡通过税及落地税等，几于无代无之。关卡之弊，记述议论，亦复多有。工商当行，在政府视为应尽之义务。然行户采买，名为给值，实多白取。所谓和买、坐办等，皆是此类，深为商民之患。就一般税役而论，明清虽有以货币代实物之趋势，实际负担，仍属不小。惟清代在未创设厘金之前，税额较之前代，似为稍轻。

    牙行中之官牙，领有牙帖（纳费），实只相当于唐代之市司，除介绍买卖外，并可评定物价，有时且可为商人之居停主人。在水路则有埠头，亦称船埠头，其作用与牙行同。牙行之作用，与同业商人自组之行，有时相辅，有时相竞，其关系殊为微妙。在政府用为统制之工具，则无甚异同。政府对物价与币值之控制，普通最重视米粮价格与银钱比价，对米粮与货币之流通，有时亦加管制。惟自宋元以后，亦不时有人论及过分统制之恶果，提倡自由流通，此亦经济发展之反映也。

    政府利用商人之一常法，为发商生息。此在若干情形之下，对商人可能有利。但商人须负责偿还本息，往往为难。至于盐商洋商等之捐输报效，名曰情愿，号为踊跃，实际则多出强迫，不过政府与商人分利之美名而已。

    一般言之，清政府对商人，尚属宽大。商人之苦于苛虐者，罢市、请愿，乃至短期暴动，虽有其例，大规模之变乱，则未有商人为领袖者。此中因素，虽甚复杂，与政府对都市商人统制之和缓，似不无关系也。

    一、导论

    这篇关于政府对城市商人之统制的文章并不是一篇研究论文，文中所提出的数点建议只是一个社会史学者所做的一般性的观察，希望或可作为进一步讨论的基础。文中“商人”一词是用的广义，包括各种商人与生意人，固定的与流动的，甚至牙人（经纪人），经营当铺、钱庄的人，以及投资于传统手工业的生意人。这样使用的理由是中国传统上把这些人都称做“商”。“铺户”一词，是登记职业用的，差不多包括所有从事各行生意的人。“店”这个字或指商店或指旅店。因此商人一词必须使用广义才能把一些有意味与相关的事实包括在内。“城市”一词也是用的广义，兼指城、镇与郊区，而不限于城墙以内的地区。事实上，通称为“镇”的市场中心，大抵是没有城墙的。商人只要是在城市做生意的都可称为城市商人，虽然他并不一定住在城里。“统制”这一词包括与商人的地位、活动以及税役等有关的规定与限制。

    本文的讨论集中于清初到鸦片战争（1644—1840年）这一段时期，换言之，即是传统中国开始受到西方势力的空前冲击以前的两个世纪。这段时期特别令人感觉兴趣的理由，其中之一是这段时期内，中国的统治者是几位相当开明而且非常能干的异族皇帝；这段时期中国正经验到社会与经济方面重要的变迁，即是中国大陆学者称为“资本主义萌芽”或初期成分者。[1]此外，中国在这段时期仍保留有许多传统的面貌。

    一般对传统中国只有初步了解的研究者，可能认为旧社会商人的地位是这样的：农人所从事的职业是“本业”，相对的，商人与工匠的职业被视为次等的、非基本性的“末业”。此外，商人多被视为奸狡、惟利是图，因而受到轻视。他们的投机、操纵物价、屯积货财，都被认为不但害及消费者（特别是无助的农民），也对整个经济有害。商人的这些活动有违于公正与安定的原则，因而各种规限与税役必须加在商人身上，对于他们的地位必须加以降抑。但是，像这种一般性的说法至多不过是粗略的说明罢了。

    这种一般性的说法所以流行的一个原因，是受到古代中国某些时期的史籍的影响。差不多三十年前，如果中国学生曾读过一点点中国的正史，很可能不是《史记》，便是《汉书》；前者的范围是从中国古代至西元前100年左右，后者则从西元前206年到西元23年。上述的说法大部分便取材自这两部史书中谈到食货与商人的篇章。[2]那时候大学里中国通史的课程仍然只着重于古代史方面。比如就制度史来说，教授们认为只要说明与讨论汉代的制度史就可以，因为后代差不多都是因袭汉代的模式，只有很少的修改与出入。

    当然，中国古代史与中国第一个官僚帝国确有许多值得研究之处。简单地说，在战国时代（西元前403—西元前221年），政治、社会与经济上巨大的动乱与变迁中，游士、游侠与行商坐贾这些人变得非常流动而活跃。他们成为各独立邦国以及后来帝国的政治资本。因此，他们可能是艾森斯塔教授（S.N.Eisenstadt）所称的“自由浮动资源”的最好的例子，对于他所谓“历史性官僚帝国”之成立，有过重要作用。[3]

    到西元前221年秦统一各国，这个中国史上第一个帝国要面对的问题是如何处理这些自由浮动分子。明显的办法是统制，包括操纵与利用——为了政府的利益，绝对不能让他们自由集附到另一个政治中心，或是自己形成一个有影响力的集团。秦朝只是短短的十几年（西元前221—西元前207年），未能完成这项工作。它的失败也许由于过分注重法家思想，过分独裁。汉朝从这里学到教训，成绩较好。当温驯的儒家学者（借用顾里雅教授H.G.Creel的定义：儒乃懦弱者也）成群地协助或加入汉朝的统治集团，中国官僚帝国的模式便开始形成了。

    汉代是否真正采用压制商人的政策是值得讨论的。支持这方面看法的人会说，商人得缴纳额外的重税，他们不准拥有土地，不准穿着丝绸，他们的子孙不得做官，他们的活动在政府有专卖权的一些基本货物上受到限制。事实上，上面这些说法，除了有关纳税那一项之外，大多数是不难修改的。一个富裕的商人可以很容易放弃他登录的商人身份，变成一个地主，而仍然做谷物、丝帛或其他生意。汉高祖命商人不得衣帛，这道命令恐怕当时并未认真执行过，以后更是完全被忽略了。雄心勃勃的汉武帝即位后，即打破政府不任用商人为官的规定，两个在盐铁买卖上非常成功的商人成为他的主要参谋。把盐铁收归国有的建议，就是他们提出来的。他们主管专卖事业之后，就引进更多的生意人担任政府官职以协助他们办事。桑弘羊，贾人之子，精明而有谋略，深得武帝信任，由侍中官升御史大夫（副相）。由此看来，中国第一个持久的帝制朝代——汉朝，对商人的态度就已经是模棱两可的，至少在一段相当时期内，政府是有意兼用一种对商人限制、征税而又加以利用的政策。

    在后来的朝代里，商人的命运也走着一条曲折的路途。为了解某一段时期商人的地位，一般历史背景的知识是需要的，因为只有与其他时期商人的地位相比较，才可能对某一时期的情形得到一个有意义的评价。

    二、政府对城市商人的统制

    如果回顾一下清代最初两百年间政府对城市商人的统制，很明显的是，这段时期我们见不到什么特别的障碍妨害商人改善他们的地位；政府对商业活动的控制是有限的，加于他们身上的课税与勒索，相对来说较轻（或至少不特别重），另外是，在统制的执行上，往往都离不开“保”与“包”这两个古老而特别重要的观念。我们可以先从最后一点谈起，以作为了解的背景。

    “保”与“包”这两个观念与“报”不可相混。关于“报”我已有另文谈及。这三个观念都是传统中国盛行的观念，而且还继续到现代。在“保”与“包”这两个观念中，“包”流行较晚，大致是自宋代以降，这点也许可以反映出中国从宋代以来就对有限而可确保的利益或结果越来越感到兴趣。

    保的观念几乎在政治、社会与经济生活的每一方面都可以发现。参加科举考试、进入官场、担保贷款、申请护照等等，都需要某种地位的人或某级以上的店铺担保。几个人或店铺联合起来担保的称为“连环保”，执行地方警卫与地方统制的保甲制度，是中国史上最为人熟知的制度之一。包的观念最常见的是包税（常与另一个字“额”连用）此外还用于包车、包船、包工乃至包饭等等。

    我们可以就商业活动范围之内举出更多的例子：政府核准的牙行的一个作用是保证某种程度内的公平交易。政府要求商人行会的领袖负责保证会员的行会，而且要供应清廷官方所需要的应用物品（这些往往牵制到所谓规费以及类似的勒索）。有引票经营盐运的商人首领称做“总商”，责任重大。经营出入口贸易的“公行”，有时称为“保商”，必须负责一个港口的对外贸易。大规模的商业组织，政府往往要他们成为多头制，以便维持制衡。这种预防办法，类似政治圈内所使用的，例如数名省级的高级官员并列。这是中国统治者从历史上得到的经验，知道倚重惟主管首领是问与联合负责的原则。

    （一）地位与登记

    在清朝统治下，阻止商人爬上政治阶梯的障碍，显然很少。中国帝制早期的几百年内，统治阶级经常妒忌地守住他们的政治权力，商人即使想占一席地位都极端困难。隋代（581—618年）所建立的进士制度，一直成为学者经由考试进入官场的最佳途径。但是这项考试，在隋唐（618—907年），以及辽代（907—1125年），对商人、工匠及其子孙是不开放的。[4]这种歧视政策到宋代（960—1279年）似乎减轻了不少。1444年颁布的规定要进士级的考生之间组成相互担保的团体，每一组三人（首都区开封府内五人）。担保的条例有一项是“身是工商杂类，及曾为僧道者”不得取应。条文中所用的“身是”与“曾为”两词似乎指出，出于商人家庭而自己不是商人，或甚至曾为商人而目下已非商人，都准许参加考试。如果我的解释正确，这点值得研究中国历史的人记在心里。同时要注意的是，在金（1115—1234年）、元（1206—1368年）两个异族入主的朝代，似乎没有禁止商人或工匠参加考试的规定。因此我们可以说，最近数百年中商人已经得到了政治解放。

    事实上，在明清两代，盐商还有一项特权，可以令其子弟注册入“商籍”，参加生员考试，以进入商人居住地与经商地的学府，而不必如一般人须返回本籍才能参加考试[5]。此外，学府中特别为商籍学生保留名额，这些生员以后多半在省城参加考试。这种特权无疑地为清代盐商的后代造就了几百位进士，与更多的举人。何炳棣教授在他的研究中，曾举有数字。[6]把这些资料大略地再检查一遍，可以发现这些举人进士大多数是在18世纪通过考试的。

    令人感兴趣的是，为盐商家庭子弟设置学校的制度，可以追溯到元代。1299年，一位蒙古籍的盐政在河东为盐商家子弟设立了一个学校，称为“运学”。注册的学生称为“运籍”，这名词是“商籍”的前身。这件事以后在16世纪末，被人提出来当作在别处成立类似设施的前例。[7]也许，就元朝来说，给予商人特权是很自然的事，因为蒙古的统治阶级十分依赖维吾尔商人与中国商人给他们带来的巨大利润。

    除了考试以外，商人获得荣耀乃至官位的另一途径是“捐纳”，这是一种花钱买头衔、职位的制度。卖官鬻爵自然不是新事，它甚至可以追溯到汉代，但清代的制度无疑地是最完备，而且是最被倚重的一项主要收入。在18世纪早期更是重要。这个制度显然也包含有政治动机。正如雍正皇帝曾公开承认，有才能的人不由正途，而借着捐纳等非正途出身，可以平衡由科举出身者造成的过分影响力。在理论上，正规的捐纳，虽然本身不是正途，却是让生员得官或小官取得晋升的主要台阶，当然也有例外的情形。实际上，所有的富人都能为他们的父母买一个荣衔，并有不少替自己捐买监生、荣衔甚至官职者。富有的商人任意利用这种机会不难想像得出，18世纪的盐商就可以举出很多例来。商人捐官这件事，在19世纪下半叶曾经遭到章奏强烈的反对，但是清政府不能也不肯放弃这笔每年给国库带来几百万两银子的财源。有人曾说，这一大笔收入使得清代早期统治者不必重视商税，结果是商人得利。此外让人感兴趣的一点是，大约从1851年开始，旧式银行称做“银号”者，为人办理捐纳而大赚其钱。[8]

    在结束我们对商人地位的讨论之前，我们需要注意到明清两代社会系统的流动性，这点何炳棣教授已有畅论。[9]其中很有趣的是，我们可以看到家庭分工的例子，父亲或兄弟经营家中的田产或生意，而让儿子或另一个兄弟去读书、参加考试。清代学者沈垚（1798—1840年）曾上溯到宋代，认为这种经济基础是帮助考生成功的重要因素。沈垚认为，从那时候起，所谓四民的士、农、工、商已有了结合与混合的现象。[10]另一位清代学者钱大昕（1728—1804年）也注意到，宋元时代的儒家学者已经鼓励学生首先应获得适当的生活方式（谋生方法），这样才可以使他们在进入官场前专心读书，日后在任位上才能维持正直与清廉。[11]农夫的职业当然是基本的，一个诚实的商人或制造有用而非奢侈品的工匠，他们的职业也可视为基本的，黄宗羲（1610—1695年）曾强调过这一点。[12]这种态度上的改变，无疑地反映出当时的社会环境。在一个较为流动的社会里，不只富商成为有威势、有影响力的人物，就是普通商人也发现他们的地位改善了。另一方面，我们也不能说，古老的轻商观念，此时已经归于消灭。举例来说，乾隆皇帝在1742年下诏免除米与豆在国内所有的通过税，诏令中他依然提出“重本抑末”的老调作为理由。[13]

    与商人地位密切关连的问题是他们在人民中如何登记。中国历史上，登记（著籍）一直是政府统制人民的一项重要手段。从帝制中国开始，正规商人就得登记在“市籍”项下。秦汉时代由于用兵频繁，有时那些名字登记在市籍下的人是第一批被征召入伍的，然后是那些以前曾入市籍的人，再其次是那些父亲或祖父入市籍的人。[14]

    市籍的登记至少继续到唐代，那时候由政府密切统制与监督的城内集中市场颇为繁荣。关于唐代的市场制度，杜希德教授（Denis Twitchett）曾有精辟的论述。[15]但是到了唐代后期，这种市场制度开始衰落，大多数城市市场的规定都被忽略或遗忘，很可能不久以后市籍登记便终止了。

    在明代，户口的登记主要分为四大项：军、民、匠、灶（制盐者）。[16]工匠有专籍，因为他们必须轮班应差。明中叶以降，班匠可以纳银代差，渐渐得到解放。

    军民工匠四种户籍在名义上延至清初。《嘉庆会典》列举“军、民、商、灶”[17]，这一条很容易引致误解，因为此处之商即上述之“商籍”，单指盐商而言，而非指一般的商人。

    户口的登记从1772年正式成为保甲制度的一部分。然而，保甲制度起初并未认真执行，直到1813年冬天，国内发生一连串暴动事件，特别是这年秋天“天理教”的一次暴动，震动了北京皇城，以后保甲制度才比较认真。清代的保甲制度并不是划一的，大致来说是“门牌”的登录以及登记入籍。登记的事项包括户长的“生理”或“行业”。这分为“住户”或“民户”，与“铺户”两个主要项目。有趣的是，铺户的登记只包括那些不与家人同住的店家（我们可以称为离家商人）。店主与家人同住的则归入民户。我们需记住，在中国帝制时代，远离家乡的老百姓很可能引起别人的猜疑，他们得随身携带执照或护照之类的文件以证明他们的身份。

    根据1851年秋天的官方报告，北京的内城（西洋文献称之为“鞑靼城”，因为大多数居民均为旗人）住户七六四四三户，铺户一五三三三户。[18]在北京的外城或所谓“中国城”，铺户的数目可能更多些。另外从天津在1846年施行保甲制度下登记的民众，我们可以发现某些有趣的项目与细数。[19]生意人分成三个项目：“盐商”、“铺户”与“负贩”。在天津城围内登记的九九一四户中，盐商一五九户，铺户三一三二户，负贩一九三五户。在东郊，即东城门外，登记有七○七七户，其中一一○户为盐商，二九七五户为铺户，一三三○户为负贩。在北部的六六三五户中，盐商五二户，铺户三一九六户，负贩七九九户。其他西郊、南郊、东北郊与西北郊四个郊区，登记的户数较少。但在这些区中，生意人三项登记的总数仍超过总户数的三分之一或接近半数。这些显然相当可靠的数字，很可指出在我们所讨论的这段时期的末期，天津市的商业化程度。

    （二）限制、征税与利用

    唐代的各种民法与刑法包括许多关于市场的详细规定，但清代的《大清会典事例》与明代的会典相似，对于贸易与商业方面较少提及。会典中的“市廛”即市场统制一节，仅包括短短的五项：经纪业务、公平价格、市场的独占（把持行市）、度量衡，以及市场上出售的衣料与用具的品质标准。除了第一款内规定私营经纪业务为非法（私充牙行埠头），这点是从《明会典》中抄袭而来，其他各款都依照唐代标准而制定。[20]关于上述最后两项事务的规定，其起源最为古老，也可能最不受人重视。晚清的法律专家薛允升氏（1820—1901年）曾特别感慨这方面执行的松懈，他强调维持货物品质与统一度量衡的重要性，但并未引起作用。[21]

    根据禁止私充经纪的一款，在城镇乡村的各行业的经纪人（诸色牙行），以及类似泊船地方（船埠头）的经理人，应从殷实人中选出来担任。政府发给他们盖有官方印记的登记簿，让他们记录来往商人或船主的姓名、固定住址、通行证号码，以及货物的数量。登记簿每月要送交政府当局检查。那些未经官方核准而营经纪业务的人应受杖刑六十大板，他们所收取的佣金（牙钱）应予没收，如果官方认可的经纪人或埠头（官牙埠头）有掩饰藏匿，应受杖刑五十大板，然后免职。关于物价一款，将制定公平价格的责任给予经纪人（行人，即牙行），而非唐律上所规定的市场官员（市司）。[22]

    经纪人的作用是在买者与卖者中间协调商定一个合理的价格，除此之外，许多经纪人也充当店家，招待来往商人的食住与寄放货物，当然也照章收费。这些费用是在交易时所收的佣金（牙钱、用钱、或称行用）之外的。经纪人也可能充任商人买卖的代理人，为他们接洽贷款，安排他们的交通与货物运输问题。因此经纪人在贸易商业上能担任不少职务。[23]政府要借着经纪人以钳制商人是很自然的事。

    在理论上，只有有执照的经纪人才准许担任这些职务。根据规定，这种执照（称做“牙帖”）只有省级当局才能发给，并有固定的名额，这个执照每隔五年检查一遍，并重新发给（北京从1725年开始），同时，名额亦可能变更。[24]实际上，省区与地方官员常常不顾名额而自行发给执照，因为这项业务是州县政府收入相当可观的一个来源。对省府与清朝政府而言，从经纪人的执照所收取的费用只是非常小的数目，但是，自太平天国叛乱以来，情况有了重大改变，从那时起，特别捐也由经纪人收取，并与厘金合在一起。在湖北与湖南，从经纪人处收取的年度捐税估计有他们的牙帖费的一百倍之多。[25]

    这些经纪人，特别是那些私营的，带给商人的麻烦实多于帮助。当某一行业的商人组成一个行会后，通常都会被与他们这一行打交道的经纪人控制住。通常借着使官准牙人或为本行会员而达到目的。有关这类做法的例子我们在北京18世纪时组成的行会记录上可以看到。[26]

    在这里要强调的一点是，“行”这个字在中文里经常是表示“行业”而非“行会”，除非我们将行会的意思扩大到包括那些没有会馆或公所，甚至没有行规的原始行会。政府热衷于让商人按行业组织起来的主要理由是配合它对各种物资的需要，这种要求可能来自清廷当局或任何大小衙门。商人有义务应付这种要求，称之为“当行”，意思是“本行的当值”。理论上，政府需要的物品应该用“时价”或“实价”买进。事实上，真正照办的很少，即使政府付给相当的价钱，经办人在中间索取的陋规也成为当值商行的一个沉重的负担。1738年，清廷诏令全国各大小衙门纠正这种陋习。[27]在雍正皇帝名义下发布的《州县须知》，警告地方政府官员，不得向商人与百姓强索物品。[28]然而这些命令与警告实际上完全没效。举例来说，为了供应清廷光禄寺所需用的猪肉与鸡，北京城内宛平与大兴两县特别从这两行里挑选了殷实的商人来负责供给，结果害他们从1752年到1756年之间，每年都赔上两三千或三四千两银子，直到这两行在1756年被废除为止。[29]

    在明代末叶以后，这种“当行”制度照规定本可纳银替代。16世纪时，北京城的铺户分为九等，每户每年要付一钱至九钱的银子称做“行银”，以免当行。到1582年冬天，政府批准一项奏折，免除最下三等的铺户缴纳这笔行银。中间三等的铺户，其资金从三百两银到五百两以上的，以及上三等的铺户，其资金多至数千两银者，则需继续缴纳。同一年早期，政府也批准北京城内两县中一三二家官方认可的行业中，三二家小号得以免除缴纳这笔银钱。[30]

    到清代，北京城内的两县获准从内城以外的铺户收取这笔银钱。上等的铺户每年缴付五两银，中等每年二两五钱，下等的铺户则免缴。北京内城九门内的铺户得以免缴的理由是他们得负责整理街道，特别是填土、洒水的工作。

    大多数城市中对商店开设的地点都没有严格规定，只要不太靠近衙门损其尊严就行。但是暂时性的货摊与浮摊不准见于大街上。在皇都里的规定就比较严格，举例来说，北京的内城不准开设戏院与旅店。1756年所做的调查，显示城里有十五家旅店，其中有好几家“关东店”，显然是为在满洲做生意的商人开设的。[31]还有四四家店铺，夜间也经营旅馆业。所有这些店铺都得迁到外城去。另外七二家经营猪肉、酒、鸡、水果与烟草的店铺则准许留在内城。[32]叫卖的负贩有时不准喊出某些被认为是忌讳的字眼。在1648年与1649年，北京城内的负贩曾被禁止叫卖，因为多尔衮嫌他们的声音太吵。[33]

    有关这方面我们可以再加上一点是北京城内一般都实行宵禁，特别是在内城。为了便利警卫，许多较小的街道，特别是通往大道的傍道都树立起栅栏，夜晚关闭，禁止通行。根据《金吾事例》，1729年北京外城有四四○个官准的栅栏；1763年内城有一○九九个栅栏，皇城内有一九六个。这些栅栏似乎一直维持到19世纪初年。[34]栅栏与宵禁令人想起唐代首都长安城内坊门夜闭的严格规定。

    北京城内的九个城门的征收货物税都是在恶名昭著的崇文门税关管制之下。这从明代起就如此，一直继续到民国时。更早的朝代当然也有类似的税。记得南唐时代曾有官吏幽默地对皇帝说，首都不下雨的原因是雨恐怕在城门要缴税。结果，皇帝下令减轻这些税捐。[35]

    像清朝其他的税关，崇文门的税关也有年度的定额。在本文讨论的这段时期内一般定额是十万两银多一点，这笔数目不算大，留给税吏足够的余地去充实他们自己的腰包。[36]税关的主管者照规定都是旗人，他们在这位置做了几年后，大概都得到类似的下场：借某一个罪名免职，其大部分财产充公，但也罕见完全破产之例。清朝皇帝与这些权贵税吏之间的关系正像渔夫与他豢养的鱼鹰之间的关系。

    州县政府的一个重要收入来源称做“落地税”的，是对所有进入其管辖的地方市场的货品所征收的税。这些税通常都是包给衙门的衙役或牙子，自然有滥用职权与腐败的情事。1735年清廷曾下令废止所有乡、镇、近郊的落地税，仅保留县城与州城的。[37]这道命令是否曾广泛执行以及行之多久却是值得怀疑的事。

    总结来说，清代最初两百年内对地区间以及地方贸易的税收并不特别重，尤其当我们比较一下明代万历朝（1573—1619年）的下半期，朝廷的宦官在征收全国商业税那种无情的勒索时，或是比较一下从1850年代加之于各省的厘金，给朝廷从1869年到1908年每年都带来一千四百万至二千一百万两银子的收入时，就可明白。[38]

    在物价管制方面，政府关心的主要是谷价的稳定，以及铜钱与银两的兑换率。为了防止大量囤积铜钱与米谷，政府曾试用各种方式，下令禁止这种事情发生。当谷价太高的时候，最有效的办法显然就是抛售政府所存积的米谷。在北京，官方的米局特别用来供给旗人。由于北京城人口众多，因而有严格的规定管制米谷运出京城。原则上只有少量的米，村民买来供自己食用的才准许运出京城。此外，不论米或谷都不准运出城或甚至京畿地区。[39]清廷对未去壳的谷子管制更为严格，原因是谷子能保存得更长久。

    银与钱的兑换是钱铺的主要生意。通常，北京城的钱铺得五家一组连合互保。18世纪有一段时期清廷依靠官方认可的钱币经纪人（称做钱行）来稳定兑换率。[40]大体来说，雍正与乾隆两朝在北京的成效相当好。兑换率的波动幅度是从八○○到一一○○文铜钱对一两银，但大多数时间都维持在八五○或九五○上下。[41]谈到钱铺间的连保，可注意的是类似的要求初期并未应用到旧式的银行（称做银号）上面，直到1860年数家半官方的银行宣告破产以后，银号才需要连保。由此看出，尽管银本位经济已经继续了几个世纪，政府对银的控制总是落后一步。

    利用城市商人的一个主要方式，是托付给他们一笔公家资金作为投资之用。这种制度称做“发商生息”，在前几个朝代就有了。受到这种资金的商人绝大多数都是当铺与盐商。政府收取的利息是月息一分至二分。一般来说，这笔利息是指定作为特殊用途的。[42]雍正皇帝特别爱好这个制度，用所得的利息来资助八旗与绿营军。清廷的内务府也非常依赖发商的利息为其财源。乾隆皇帝时仍继续这个制度，后来他改变主意，1759年时宣告发商生息于政体有损，下令加以限制。1769年，他下令将已经发给长芦盐商的资金改称做“赏借项款”。[43]使用这个新名词的理由是政府所订的利率较法定准许的月息三分利率低得多。但是，旧的名词与制度仍被清廷、省府、州县政府以及半官方或非官方的组织继续使用下去。可注意的是信托资金对商人并不一定有好处。1783年长沙府的当铺为某种原因婉拒从省府接受更多的资金，托辞说他们手头已有足够的信托资金了。[44]

    另一种利用城市商人的方式是“自动捐献”，称做“捐输”或“报效”，这是城市商人资助政府的军备、公共建设、水患、饥荒的救济，皇帝出巡与皇帝生日等的开销。根据两淮地区盐政管理官方记录的数字显示，在1738年至1804年之间，这个地区的盐商在四十多个场合总共捐献了三千七百五十万两银子。[45]根据盐政的报告，盐商们都是“情愿”甚至“踊跃”认捐，恭请皇帝“赏收”。在另一方面，盐商们又不时请求以分期付款的方式来捐献。有几次，皇帝对商人的忠诚报效与急公好义表示嘉奖，而只赏收一半的捐款。商人在这种情形下所得到的直接回报不过是所谓“议叙”与名义好听而已。然而在其他场合，皇帝为显示对商人的仁慈宽大，准许他们免费取得额外的“余盐”，或是允许他们延期偿付滞纳的盐税与信托基金的利息。皇恩的殊荣，甚至免除盐商对政府的负债，1780年减免了一百二十万两银子，1782年与1784年大约是三百八十六万六千两。[46]另一批重要的自动捐款，是由广东的盐商与洋行（行商）所认捐的。从1773年至1832年间的捐款总数大约是四百万两银子，数目虽不是大得惊人，也是一笔巨款。[47]

    如果能比较清代各皇帝所采行的经济政策，特别是有关商业贸易的细节，甚至比较一个皇帝在不同时期的经济政策，将是一件极有趣的事。遗憾的是这样的比较已远超出本文的范围。然而我们可以强调的是康熙、雍正、乾隆三帝都绝不是蒙昧无知不肯用心的专制君主。康熙皇帝有一次在1717年曾夸称他对盐政方面深刻的了解。[48]雍正皇帝无疑地非常通晓一般的财经事务，但1728年有一次也承认他并不特别了解有关茶政上各种渎职情事，以及有关茶与马的贸易，因此不能给负责的官员特定的指示。[49]乾隆皇帝在1748年曾有很合理的意见，认为一般来说还是把市场方面的事交给人民，准许他们自由流通货物较好。政府的干涉，虽然出于好意，常常由于处理不当而产生扰民的障碍。[50]清代皇帝一般都可以称得上对商人宽大而同情的。但在另一方面，他们对商人有时也出诸操纵甚至有喜怒无常的态度。

    三、城市商人的反叛取向

    相对于政府统制，重要的一点是检讨商人是否曾抗议或反叛这种统制，和采用什么方式。有关这方面讨论，我们可以19世纪学者汪士铎（1802—1889年）所做的观察作为起点。他认为，商人与城市的文人一样，似乎是最不倾向反叛的，或者我们可以说，他们表现非常低度的反叛取向。汪士铎在1853—1856年间，因太平天国之乱曾躲藏在长江下游的南京与绩溪之间，这段时期所保存的日记中有如下一段：

    天下最愚，最不听教诲、不讲理者乡人。自守其所谓“理”而不改。教以正，则哗然动怒；导以为非为乱，则挺然称首。其间妇人又愚于男子。山民又愚于通涂之民。惟商贾则巧猾而不为乱，山民之读书者不及也。在外经商之人，又文弱于当地之商贾。知四民之中，最易作乱者农。工次之。武生次之。山中之士次之。商贾之士次之。城士之士，则硜硜然可以决其不为乱[51]。

    这种议论显然是概括而充满偏见的，但我们或可了解这不完全是处于一个大动乱暴力时代所发的愤激之言。无论如何汪士铎是个相当独立敢言的学者，他不受传统儒家思想的束缚，而且是热心于提倡改革、恢复秩序的人。至少他在这段话里提出一个启发性的见解，就是在传统的四种功能团体中，城市商人与城市文人的反叛取向最低。

    进一步说，根据汪士铎的推论做一初步检查，显示其中确有一些历史的真实性。[52]中国历史上曾记载无数次农民叛变，但几乎看不到任何城市商人领导的叛变。从唐宋时代以降，我们看到走私盐商与海盗商人的记载，然而他们行动的范围似限于山林、沼泽、海岛与外海上，有时在他们势力范围内，他们也会打劫城镇，因而可算是城市商人的敌人。在明清时代，关于矿工、伐木者与城市匠人的暴动与罢工事件，也有所闻。

    当然，一个社会中叛乱取向的问题，或广泛地说暴力取向的问题，其研讨不一定只限于功能团体。举例来说，这个问题可以就个人或团体从年龄、性别、地位、财富、角色、功能、教育、风俗、传统或是其他的角度来探讨。甚至汪士铎所作的粗略的推论也提到其中几方面。然而，对这个问题更深一步的方法论，却已远超出本文的范围，而且坦白地说，也不是作者能力所及的。为说明城市商人在清初时代抗议与叛变的性质与程度，我们可以看看下面四个例子，他们所谓的“商人与手工业者反抗清朝封建统治的斗争”。[53]这四个例子记述的事实均是“罢市”，就是商人与生意人拒绝做生意以示抗议。

    （一）1660年在山西潞安的罢市

    这次罢市的背景是源于明代生产御用丝织品称做“皇紬”的制度。在山西潞安做这一行生意的“机户”，必须以固定的官价供应这项货品，而官价显然是经常不足以抵付生产所需的开销。明末清初时代，皇紬年度配额是三千匹（一匹为六丈八尺）。1652年诏令将配额减去一千五百二十匹零四丈八尺，每匹的价格则从十两银子增至十三两。1658年，配额又由一千四百七十九匹零二丈减去一千一百七十九匹零二丈，因此实际上所需要的仅是三百匹。但是到1660年，机户发动一次罢市，据说将其织机焚毁，手里捧着账簿记载着他们的损失，准备向北京城进发，直接向皇帝请愿。

    据潞安一位朝廷官员王鼐的奏折，这些机户在明末时原有三千张以上的织机，但大多数都已破产，因为他们得依照政府命令按行服务，所谓“抱牌当行”，结果是他们生产的丝得不到适当的偿付而大受损失。从1644年到1660年，所留存的织机仅两三百张。据奏折所言，皇帝的削减配额，延长限期，先行付款，以及“合理实价”，使得机户争着愿为皇室服务。但是本省官吏的取用以及外省采购使者的要求勒索，却使他们遭受损失。理论上，机户们可以从他们出售的丝得到官价付款，但是经过层层勒索，特别是付差官差役的催紬费、验紬费及纳紬费，实际所余无几。

    我们在王鼐的奏折中可看到很生动的描写：“臣乡山西，织造潞紬，上供官府之用，下资小民之生。……为工颇细，获利最微。……今年（1660年）四月，臣乡人来言，各机户焚烧紬机，辞行碎牌，痛苦奔逃，携其赔累簿籍，欲赴京陈告，以艰于路费，中道而阻。天有簿籍，必有取用衙门，有衙门必有取用数目。小民含苦未伸，臣闻不胜骇异。”他接着建议严禁本省不得滥行取用，隔省不许擅差私造。从方志记载中，我们不清楚他的建议采行至何种程度，因为只说到山西巡抚下令立碑严禁。推想大概是，差役与差官不许继续强索，而机户也不许再度罢市。[54]

    （二）1660年安徽芜湖的罢市

    这次罢市是抗议芜湖内地税关过度的附加税与其他各种名目的勒索。根据阴历十月十三日御史李见龙弹劾户部郎中兼湖钞关监督郑秉衡的奏折，在郑秉衡的指使下，若干名不法的官吏征收额外的火耗与特别捐款用来充实其官邸的维持费用。郑秉衡还发明了“皇税”一词，对民船上装载的日用必需品甚至如薪柴与米都征以税。结果是，全部地区的商民发动罢市三天，以1660年阴历七月十四日为始。本地生员韦譞佩等向总督与巡抚请愿，结果总督命令知县接受商民所具甘结，同意地方人民发动罢市是因为征收薪柴与米的征税。据奏折上说，御史闻知这事是得自于从芜湖到北京来诉苦的商人，因而有关这事的消息传遍京城。[55]

    这项弹劾似乎并未发生多大效力，因为罢市的事件已经发生了一年，而且显然已不了了之。对我们来说，有关这次罢市最感兴趣的一点，是其行动的有秩序以及商人与士人间的合作。

    （三）1682年浙江杭州的罢市

    这次罢市是抗议土棍（地方流氓）与旗丁（八旗兵丁）的高利贷，他们对那些无力偿债的人捉去儿女以为抵偿，有时甚至牵连到负债人的亲戚与邻居。杭州北门的商民发动罢市抗议，这事传到一位同情人民的道台王梁那里。第二天，当王梁去与其他官吏会合调查这件事的途中，八旗兵王和尚等一共几百个人，拦住他的仪仗，辱骂他并打破他轿子的顶盖。这次不寻常的暴动，迫使总督与满洲将军连合上奏向朝廷陈明情况，结果皇帝下诏严厉处罚王和尚及其同谋者。这时候，总督则下令店铺恢复营业。这个例子中特殊的一点，是它说明了在一个征服王朝下政治与经济生活的复杂性。[56]

    （四）1698年福建浦城的罢市

    下面这段故事主要是根据直隶任邱人、出于书香门第的庞垲的墓志铭而来的。在戊寅年，即1698年（彭泽益误认为1758年），庞垲受命为福建建宁府知府。他到任后不久，传来报告说建宁府所辖的浦城县令，因为政令过于严苛，迫使人民反叛。城中愤怒的百姓趁着黑夜，攻击县府的“册局”，放火烧毁文件与记录，并杀死了一个当值的胥吏。县令害怕逃走，当地人民接着发动一次总罢市。庞垲得知这事，立刻赶到浦城，要求当地的教官与典史召集乡绅、生员与人民在明伦堂集合。在这些人面前，庞垲宣布县令的错误与罪状，并加以谴责，使士绅与人民气平下来。然后，他再提醒他们无法纪行为的不当。他让县府的财务与库房重新核对与收集未被焚的文件。他命令各行生意人恢复营业，城内秩序始告恢复。

    在这时，总督郭世隆不满省中百姓攻击县府（称为围城），发动罢市的事件日益增加，想借此用高压手段压制罢市，以为警戒。由于县令与地方士绅间的强烈不睦，总督欲借不法结党、阴谋叛变的罪名惩罚所有的士绅。庞垲反对这个做法，他强调县令残酷作风的不当。最后，只有一名变乱者被处死刑，另二人流放。浦城百姓为感谢庞垲的大力相助，建立一个书院来纪念他。他死于1735年。[57]

    显然地在福建省其他城市尚有类似的抗议与罢市的事件。当时的总督郭世隆（1643—1716年）出身山西的绿营。[58]上述故事中的县令是鲍鋐，沈阳人，以前曾任笔帖式（满文bitheshi，即书记官，七品、八品或九品），多半是个旗人。[59]从这件事也可以看出旗人与一般汉人的敌对。

    本文选编自《东汉的豪族》

  • 韩琦：安第斯文明的起源：卡拉尔一苏佩

    传统观点认为，南美安第斯文明的母文化是查文·德万塔尔文化（公元前1200—前200年）。但随着考古发掘取得新进展，卡拉尔—苏佩文明取代了前者地位，被认为是安第斯地区的第一个文明，其存在于公元前3000年至前1800年之间，清晰展现着秘鲁中北部地区第一个复杂社会的样貌。

    20世纪中期以来，不断有考古学家对卡拉尔—苏佩遗址进行考察和研究，但直到1994年秘鲁圣马尔克斯大学的考古学家露丝·沙迪团队对苏佩河谷进行调查，并随之进行系统考古发掘，学界才对它形成新的认知。随着考古挖掘的深入和新成果的出版发表，卡拉尔—苏佩文明的古老性和重要性最终得到证实。2009年，卡拉尔—苏佩圣城被联合国教科文组织列入《世界遗产名录》。

    卡拉尔—苏佩位于秘鲁海岸的中北部地区、利马以北约182公里。秘鲁中北部地区的面积为81497平方公里，包括圣塔、内佩纳、塞钦、库莱布拉斯、瓦尔梅、福塔雷萨等十几个沿海河谷。与其他世界文明中心相比，秘鲁海岸似乎不太可能成为文明发祥地，因为东部安第斯山脉和西部太平洋形成的反气旋作用导致这里极度干旱。然而，该地区有50多条从山脉到大海的河流穿过，利用这种水源发展的灌溉对卡拉尔—苏佩文明的出现发挥了决定性作用。

    在众多河谷中，苏佩河谷在文明起源时期脱颖而出，仅在这一小盆地就发现了20多个可以被归属于同一时期的城市定居点，它们几乎都有公共建筑、圆形广场、住宅等，都有用土坯、石头、树干和植物纤维建成的阶梯式金字塔，其中还有雕像、马黛茶杯器、石器、棉纺织品、烧焦的食品及其他用品。从建筑规模看，卡拉尔城最大，城市布局分布有序，纪念性建筑种类繁多。其距离大海23公里，处于苏佩河谷中段的初始部分，被认为是该地区居民点的首都，被称为“圣城”。

    早在公元前3000年之前，就有一些家族群体在苏佩河谷定居，他们建立集中居住区，疏干湿地，开辟农田，修建灌溉渠道。公元前3000年至公元前2600年，首都地区的城市定居点不断壮大，定居者们在空地上修建广场用于公共活动，并有了第一个金字塔。大约在公元前2600年至公元前2300年，人们对卡拉尔圣城进行整体设计，修建了金字塔和下沉式圆形广场。在公元前2300年至公元前2100年间，大型金字塔、广场等公共建筑的规模和体积都有所扩大。到公元前2100年至公元前1800年，由于劳动力的减少，定居者们用较小的石块改建公共建筑，最后掩埋了一些重要的建筑部分，卡拉尔圣城被废弃。

    总体来看，卡拉尔—苏佩文明的主要特征表现在以下几个方面：

    以农业生产、渔业生产和贸易交换为主要经济形式。苏佩河谷的居民发展出技术比较先进的集约化农业。他们使用简单的工具（如木棍和鹿角）来掘土，修建灌溉水渠以将河水引入农田。考古证据表明，他们已经懂得通过对各种植物品种的实验，来改善粮食和经济作物的种类、提高产量。他们种植的作物主要有：土豆、红薯、南瓜、豆类、花生、辣椒、玉米、葫芦、鳄梨、番石榴、马黛茶、烟草等，其中棉花是交易的主要产品。沿海居民则捕鱼并采集各种海洋生物，主要包括凤尾鱼、沙丁鱼、贻贝和蛤蜊等。农业和渔业形成一种长期的经济互补关系。

    居民们通过以物易物的方式交换产品。沿海居民提供海产品，如离太平洋仅有500米的居民点阿斯佩罗被认为是卡拉尔的渔镇，那里的居民开发了包括使用钩子、麻线、船等在内的捕鱼技术，特别是发明了棉纤维渔网。渔民负责将海产品分发到河谷中的定居点，而河谷居民会给渔民提供所需的渔网和衣物、用作钓线的棉纤维、用作漂浮物的葫芦、制造船桨的木材以及水果蔬菜等，高地居民会提供农产品（粮食）和畜产品（羊驼）。这样，该区域形成一个类似专业化生产的贸易网络，而卡拉尔圣城无疑是这一网络的中心。很显然，这个网络还延伸到更远的地方，因为在卡拉尔—苏佩地区发现了来自高原的洛克木棒、秃鹰羽毛，亚马逊丛林的陆生蜗牛、灵长类动物皮、各种鸟类羽毛以及厄瓜多尔赤道海岸的多刺牡蛎。

    灌溉技术的使用、渔网的发明以及活跃的贸易交换提高了生产力，促进了地区经济发展和生产剩余积累，从而使苏佩社会能够以地方政府的形式加强其政治一体化进程，这种政府形式的有效性可以从国家承担的大型纪念性建筑群建设中得到体现。

    先进的城市规划和建筑。卡拉尔圣城拥有复杂的城市布局。该城占地66公顷，包括一个核心区和一个外围区。核心区包括32座公共建筑和一些住宅建筑群，外围区有一些住宅建筑群。核心区又分为两大部分，北部为上城，南部为下城。北部的公共建筑分为A、B、C三组，每组都有两个金字塔、广场、官员住房。其中B组的金字塔最大，长160米，宽150米，高18米，坐北朝南，背靠河谷，面向下沉式圆形广场，是卡拉尔城的主建筑。下城建筑有下沉式露天剧场、露天剧场神庙、长桌神庙、圆形祭坛神庙，以及平民住宅区等。

    金字塔结构的墙壁上抹有泥土，被涂成白色或浅黄色，偶尔涂成红色。每座金字塔都有一个通向顶部的中央阶梯，其上有几个房间。在主房间都有一个圣火祭坛，祭坛中央有一个火炉，火炉下方配有导风的地下管道。圣火祭坛具有仪式功能，被用于火化各种祭品。

    卡拉尔位于地震活跃区，其建造者使用“希克拉斯”技术，即将石块装在芦苇纤维编织的网格袋中，尺寸和重量各不相同，但非常均匀，有一定的松散度，用它们来支撑挡土墙，填充金字塔。这样，当发生强烈地震时，“希克拉斯”会以有限的方式微动，发挥着柔性地基作用，由此实现建筑物的结构稳定。规模宏大的城市和坚固的建筑表明卡拉尔人已经具有先进的组织能力和工程技术。

    社会分层和阶级分化已经出现。卡拉尔—苏佩文明显示出复杂的社会结构，已经出现明显的社会分层。如从事体力劳动的生产者，包括渔民、农民、工匠；精英阶层，包括商人、定居点的领导者和祭司。精英们不再直接为自己的生计进行生产，而是致力于专门的活动，如加强远距离贸易；进行天文观测来测量时间和制定历法；在公共活动的建筑施工中试验和应用算术与几何知识；举行仪式和献祭活动。

    考古发现揭示出精英阶层和普通民众存在较为明显的区分。城市中心各区的公共建筑和住宅建筑在位置、大小和所用材料上都有区别；服装穿着方式和个人佩戴的饰品上，如男性权威人士的项链和大耳环，女性的项链和头巾，也体现了社会区别。一些装饰品、项链是用从遥远的地方（如厄瓜多尔海岸）所获材料制作的，专供少数社会上层人物使用。

    中央集权的国家雏形已经显露。苏佩河谷的人口分布在苏佩河两岸被称作“帕查卡”的城市定居点中，这些定居点的规模和建筑体量各不相同。每个“帕查卡”都由几个“艾柳”组成，这些“艾柳”是通过亲缘关系结合在一起的族群，拥有相同的祖先，通过祖先来确定身份，并由族长领导。族长中有一个主要首领——库拉卡，负责指挥全体居民。这种政府制度在苏佩河谷20多个城市定居点中运行，由于卡拉尔居于核心地位，它发挥领导和组织其他城市定居点的作用，形成一个广泛而有序的互惠、交流网络。

    卡拉尔是一座和平与和谐之城，考古发掘中没有发现战争的痕迹，没有防御城墙，没有武器，没有残缺不全的尸体，这与通过战争产生国家的理论解释有所不同。美国考古学家乔纳森?哈斯认为，卡拉尔人进行了人类建立政府的实验，他们将个人自由交给一个中央集权机构，由中央集权机构决定创建一个作为仪式中心的城市，并要求大家为共同或更大的利益努力工作。人们之所以选择成立“中央政府”，是因为意识到合作将使个人和整个社区受益。考古学家露丝?沙迪认为，对神的崇拜是凝聚力和社会平衡背后的驱动力。人们之所以接受中央集权政府的存在，是因为他们相信统治者可以在人与生者的社会和神与死者的社会之间进行调解，政府的管理对于保证生活是必要的。卡拉尔社会展现出一定的复杂性，种种迹象表明，卡拉尔不仅仅是一个简单的农业社会，而且是一个具有一定组织能力和复杂结构的社会实体，已经具备早期国家的基本要素。

    宗教作为意识形态与政治权力相结合。卡拉尔的金字塔、广场和祭坛等雄伟建筑不仅是宗教仪式的场所，也是社会和政治活动的中心。金字塔象征着与天界的联系，广场则是集体仪式和庆典的场所。卡拉尔人信奉多神教，崇拜多种神灵。这些神灵与自然现象、农业、天气和其他重要生活领域有关。祭祀活动在卡拉尔占据重要地位。人们通过圣火祭坛进行各种形式的献祭，包括毛发、珠子、石英碎片、骨器、木器、纺织品、鱼类、贝类等，这些被认为是向神灵表达敬意和请求庇护的方式。统治者和祭司被视为祖先和神灵的代表或中介，他们通过控制宗教仪式、祭祀活动和宗教建筑来巩固自己的权威。卡拉尔的宗教活动是在音乐伴奏中进行的，在这里出土了一套由秃鹰和鹈鹕翼骨制成并绘有鸟类和猴子图案的横笛（共32支），一套由骆马骨和鹿骨制成的号角（共38支），一套由芦苇和棉线制作的排笛。在没有军事力量的情况下，宗教成为卡拉尔统治者凝聚和控制社会的力量，它使卡拉尔—苏佩河谷的居民团结起来。

    科技知识在文明发展中发挥重要作用。卡拉尔人开发了先进的农业灌溉系统，修建水渠和水库，这对于他们在干旱环境中维持农业生产至关重要。在设计和建造大型纪念性建筑以及修建灌溉水渠时，显然运用了算术和几何知识。有证据表明，卡拉尔人已经具备天文学知识，并将其应用到与经济、宗教活动有关的历法制定中。在卡拉尔上城C组的公共广场中央竖立着一块巨石，是当时用来观测天文的。他们已经发明一种记录信息的工具系统，如在上城C组的画廊金字塔中，考古学家发现一件纺织品遗物，被认为是“基普”，即用作记录工具的一套打结绳线。同时，在上城B组小金字塔的三个石块上还发现了基普的图画。这说明卡拉尔人已经在使用基普，比印加人早数千年之久。考古学家还发现，一些药用植物多次出现在墓葬中，表明卡拉尔人已经了解一些植物的药用价值。在纺织技术方面，他们利用棉花纤维编织连衣裙，采用穿插和缠绕的方法，还制作了渔网、鞋类、包类、绳索等。圣火祭坛下方建造的地下通风系统，能够引导风力保持火焰燃烧，并将烟雾排到室外。需要指出的是，虽然早在公元前4000年厄瓜多尔的瓦尔迪维亚等地就已经开始生产陶器，但卡拉尔人并没有使用或自己生产陶器。他们用葫芦作为器皿，用木头雕刻勺子，用石头雕制盘子。因此，卡拉尔文明属于“前陶瓷”文明，这一点已被考古学家们认定。

    由于强烈地震和灾难性气候变化，卡拉尔—苏佩文明在公元前1800年左右被遗弃。虽然如此，它在农业、城市建筑、社会政治组织、宗教文化等方面对后来安第斯文明的发展产生深远影响。可以说，卡拉尔—苏佩文明是安第斯文明的摇篮。

    本文转自《光明日报》2024年11月25日

  • 冯裕强：集体化时期工分稀释化视域下乡村公共产品供给研究——以广西容县华六大队为例

    改革开放以来，不少学者对人民公社制度进行了批判，认为它是低效率的、平均主义的制度。例如，有学者认为由于集体经济产权不完整，影响社员的生产积极性，最终导致劳动质量降低。①但是，也有不少学者对此进行反思，认为导致平均主义的原因主要是国家进行工业建设，加上当时的国际环境等因素，不得不从农村抽取剩余产品，而且人民公社在20世纪60年代初“去工业化”后，大量劳力只能进行单一的农业生产，产出极为有限。即便如此，农村还进行了大量的农田水利建设，这对后来的发展起到重要的铺垫作用，也应算作当时的劳动效率。②争议难分高下。笔者不揣浅陋，试图从工分稀释化视角对乡村公共产品供给进行考究，以期对人民公社有更全面的了解，同时也为当下乡村振兴提供经验启示。

    “工分稀释化”，虽有学者提到相近的概念或现象③，但至今尚未有学者对其进行明确定义。笔者以为，工分被稀释主要包括两方面：一是工分的直接稀释，即把非农业生产的工分拿回农业之内进行分配，从而导致工分被稀释，分值下降；其次是物资的间接稀释，即国家、集体从生产队抽走大量物资，从而使队内可分配给社员部分总额减少，最终造成工分贬值。

    乡村公共产品是巩固农业基础地位、保障农村社会稳定、促进农村经济可持续发展的重要基础。国内学术界对农村公共产品的界定④大体一致，主要是指乡村中由集体或政府提供，为广大村民的生产、生活服务，具有一定非竞争性和非排他性的社会产品，具体包括农村基础设施、农田水利主干网络、基础教育、公共卫生、社会保障等。

    一、农田水利基础设施

    华六生产大队（以下简称“华六大队”）位于广西壮族自治区容县南部，隶属于石寨公社，距离县城20多公里，是汉族聚居地，面积约为19.33平方公里，共有十个生产队。⑤容县面积为2257平方公里，其中陆地占97.51％，水域占2.49％，⑥境内岭谷相间，丘陵广布，俗称“八山一水一田”。由于地处山区，为了更好发展农业，华六大队在集体化时期修建了大量农田水利设施。据统计，1963年—1966年间，华六大队修建了大陂、三蛤、枪刀山和长冲等水库，⑦大部分生产队都有受益的山塘或水库。

    为了修建这些水库，必然要耗费大量劳动力。华六大队除了平时抽调社员进行水利建设外，还组建了20人—30人的专业队从事农田水利基本建设。曾任记分员的陀某说：“专业队就是专门开田、开荒、种山，每个生产队抽出几个人。比如我们大队有几十个人，天天都在专业队干活，生产队一样要（给他们）记工分。”（TXL，四队记分员，2017年3月16日）⑧专业队的职责很多，包括水利建设、开荒、大队企业、护理林场等，劳动收入归大队所有。曾任林业员的庞某回忆说：“山上的林木就由专业队队员种植，以前（1958年“大炼钢铁”）烧得太光了，没有林木了。每个队要2—3个，都是年轻的男女民兵。”（PWQ，华六大队林业员、党支部书记，2017年3月20日）曾参加过专业队的肖某也说：“县有县的专业（队），乡有乡的专业（队）。像最大的石剑水库、小垌水库，还有乡的红田水库，每个队抽几个人去。那些水库都是那些人去做的，统一调动。”（XYH，六队专业队队员，2017年4月15日）

    生产队一年中要进行大量的农田水利基本建设，那么这些非农业生产用工究竟在总用工中占多大比例？以1975年为例，根据各级单位的统计数据，八队、华六大队、石寨公社（统计7个生产队）和容县（统计233个生产队）的农业生产用工占总劳动日的比重分别是82.63%、83.70%、85.95%和82.70%。⑨一般而言，统计的生产队越多，就越接近整个县的平均水平。总体上看，公社以下各级单位的生产用工占总劳动日的比值都在容县的生产用工占总劳动日的比值——82.7%上下浮动，也就是说整个县大约要用17%的劳动日去从事非农业生产。这并非特例，在山西省东北里生产队，1977年的非农业生产用工占比达7.7%，这还不包括高达18.98%的农田基建工。⑩可见，在集体化时期，大量非农业生产用工存在于全国各地。除了基建用工，还有各级专业队队员、生产队干部、集体抽调的社员都要回生产队记工分，这些人员的劳动对当年生产队收入的增加并未起直接作用，因此，在外面挣的大量工分拿回生产队进行分配，必然会稀释生产队的工分值。

    那么生产队的实际工分值在集体化时期有什么变化呢？本文从容县档案馆保存的历年分配统计表中整理出表111。

    通过表1可以看到各级单位从1963年到1981年社员分配收入和工分值的变化情况。在生产队一级，由于资料的缺失，我们只能比较完整地看到20世纪70年代的数据变化。总体上，八队从1971年至1981年分配给社员的金额、工分值和人均分配收入都呈波浪式上升，在1979年达到最高值； 华六大队与石寨公社在相同的项目上虽然也呈波浪式上升，但是振幅相对小得多，除了个别年份回落，大部分年份是逐年增长的。分配给社员的金额与工分值、人均分配收入总体上呈正相关。分配给社员的金额越大，工分值和人均分配收入越高，就意味着人民公社在增产的同时社员也实现了增收，集体经济运行良好。三个不同区域都在1979年达到最高值，人均分配收入分别达到98.9元、77.95元和84.44元。需要注意的是，这里的分配金额并不是真金白银，而是生产队把一年所有的劳动产品和收入都折算成货币，扣除所有费用和税金之后的纯收入，生产队实际拥有的现金并没有这么多。

    工分值的高低取决于两方面，一是生产队分配给社员的金额，二是工分的总量。分配给社员的金额是用总收入减去各项支出后得到的数据。而生产队的总收入是农业、林业、畜牧业、副业、渔业和其他收入相加的总和。虽然国家规定生产队应该以发展农业生产为主，12但是非农业生产对生产队的收入也有重要影响。那么生产队的农业生产和非农业生产收入占比各有多少呢？我们以1974年为例。

    本文发现，1974年，农业生产收入占总收入的比重，从生产队到公社再到容县是逐渐降低的，但容县比玉林地区的平均值低了近13个百分点，也就是说容县的非农业生产收入占总收入的比重比玉林地区高出约13个百分点（见表2）。通过对各年份数据进行比对，13%是容县非农业生产收入占比超过玉林地区的正常比值。那么是什么原因导致容县的非农业生产收入占比如此之高呢？要解决这一问题，必须了解林业、畜牧业、副业、渔业和其他收入的占比具体是多少，进而明了容县与玉林地区拉开差距的原因。

    经对比，在畜牧业、渔业和其他收入占总收入的比重方面，容县和玉林地区相差不大，差异产生的主要原因在于林业和副业的收入占总收入的比重，容县比玉林地区分别高出6.33个和4.82个百分点（见表3）。林业收入主要来源于山林，容县地处丘陵，全县有480438人，水田面积为29万亩，人均水田面积仅为0.6亩，山地总面积为225万亩，人均山地面积为4.67亩。13“全县179个大队，山区大队98个，占全县大队55%……一九七一年生产木材28234立方米，占全县木材生产31874立方米的90%。”14华六大队就是这98个山区大队其中之一。据1960年普查，华六大队总面积为24038亩（约16平方公里），其中林地面积为17189亩，151974年，华六大队有1778人和1813亩耕地（1683亩水田），16人均有9.67亩山林、0.95亩水田。如此丰富的森林资源，林业产品具体又有什么呢？1974年的统计年报表显示，石寨公社造林719.3亩，其中用材林（松木和杉木）483.4亩，油茶196亩，玉桂14亩；收获的林副产品有：油茶籽515.9担、油桐籽73.45担、松脂9215.95担；收获的水果为：沙田柚41.5担、龙眼65担和荔枝88.6担；另外还有茶叶96.87担、桑蚕茧129.16担等。这些产品收入是属于林业收入还是副业收入？此问题涉及容县林业与副业的收入来源问题。在八队分类账本中，林业收入主要来源于售卖原木，副业收入内容则更多，包括松脂、纸浆、茶叶、砖瓦、石灰等。这与1975年容县林业局统计分类相符。1975年容县产量较大的林副产品有：油桐籽812担、松脂238131担、木柴183541担、木炭2199担、土纸（纸浆）4220担、沙田柚142830担；木材产品35489立方米（原木30911立方米）17。因此，容县的林业收入主要来源于各类木材，副业收入则主要来源于松脂、木柴和沙田柚等，而就收入占比来说，松脂的收入无疑是最大的。早在1963年，容县就申请建立容县松脂基地，通过调配物资和劳动力有计划地造林和割松脂。181972年，十队割松脂收入达4720.05元，除去人工和材料，净收入3794.2元。19正是有了松脂和其他各类林木和林副产品，才使得容县的非农业生产收入占总收入的比例远高于其他县。

    明晰公社的各项收入后，可以发现，表2中分配给社员的部分占总收入的比重，八队与其他各级单位之间差别较大，除了八队的超过60%，其他各级单位都在55%以下。这意味着整个地区人民公社平均分配到社员的部分占比并不高。导致这样的原因与生产队的管理水平有着密切关系。八队与其他单位相比，税率（主要是农业税——公粮）和集体提留基本保持一致，相差不大；其缴纳的公粮基本保持不变，高产年份会稍提高，减产年份会稍减；集体提留主要包括公积金、公益金、储备粮基金、生产费基金和统筹金，这些不管如何都是要拿出来发展生产和上交集体的。关键是在费用支出占总收入比重方面，八队比玉林地区全部生产队的平均水平低5.66个百分点。根据八队的账本和收益分配统计表的金额，本文计算出八队在1977年和1979年分配给社员的部分占总收入的比重分别为66.9%和64.8%，20分配给社员部分占比很高，说明八队在支出控制和经济管理方面做得比较好。

    费用支出主要包括生产费用、管理费用和其他费用，支出越多，能分配给社员的收入就越少，工分值就越低，所以费用支出直接影响工分值的大小。那么，其他生产队的费用支出高的原因到底是什么？在玉林市档案局笔者发现一份1976年的档案——《关于人民公社收益分配的情况问题和意见》，其内容可以较好地说明这一问题。

    该份档案主要是对1975年玉林地区人民公社收益分配中的分配收入出现的一些问题进行总结并提出整改意见。1975年全地区粮食大增产，但是分配给社员的部分占比并不高，主要原因是费用开支大，全自治区费用支出占总收入的27%，但是玉林地区费用支出占总收入的比重高达33%。费用开支大的原因有八点。第一，有的地方发展生产不坚持“自力更生”原则，远途高价购买或调换化肥，费用开支大，生产成本高；第二，有的地方农田基建补助花样多，标准高，集体负担重；第三，有的地方扩建学校，增加民办教师，从而增大了集体费用的开支；第四，有的地方变相增加脱产人员，加重了集体负担；第五，有的地方社员上调家禽、生猪派购任务，要生产队补钱、补粮，增加集体负担；第六，有的地方的乱支乱补、大吃大喝、请客送礼、挥霍学杂费等不正之风还没有彻底刹住；第七，有的地方搞账外分配，或者高价（市场价）买入猪肉，然后按照牌价（较低价格）分配给社员；第八，有的地方存在贪污、挪用、超支欠款等问题。21从这些原因中，可以看到生产队在经营管理中存在的各种问题，虽然说这些现象并不必然存在于每个生产队，但是如果不严格控制支出，必然会严重影响社员的收入。

    在表1中，本文还注意到工分值的变化。八队的工分值从1965年的0.35元逐渐上升到1981年的0.53元，1979年和1981年都突破了0.5元。由于影响工分值的因素非常多，生产队能够保持增长已属不易。1963年，华六大队的工分值为0.19元，此后逐步增长到1981年的0.55元。与华六大队相比，石寨公社的工分值增势更为平缓，在20世纪70年代总体保持在0.4元左右。这三级单位的工分值虽然涨幅不大，总劳动日却大量增加。通过计算可知，八队在1979年的劳动日是1965年2.1倍；华六大队和石寨公社1979年的劳动日都是1965年的1.76倍。工分主要是靠劳动力挣的，劳动力越多意味着工分越多。1979年，八队的劳动力为70人，是1965年45人的1.5倍；华六大队1979年的劳动力为915人，是1965年719人的1.27倍；石寨公社1979年的劳动力为12643人，是1965年9118人的1.39倍。22可见劳动力的增加速度远没有工分的增长速度快，工分的快速增长必然导致工分被稀释。同时应注意到，人口增长特别是劳动力增长自然使劳动工分增加，但是过多的劳动投入，在单位土地上带来的产出，并不会均一地带来同等幅度的增产。以八队和石寨公社为例，经笔者计算，八队1979年亩产1126.93斤，是1965年亩产886.87斤的1.27倍，而同期八队工分总量增长了1.1倍；1979年石寨公社亩产为1146.15斤，是1965年亩产917.31斤的1.25倍，工分总量却增长了0.76倍。23即便扣除部分工分用于非农业生产，工分的增长速度仍高于每亩的增产速度。这便是黄宗智所讲的“过密化”或“内卷化”现象：“在人多地少和土地的自然生产力有限的现实下，单位土地面积上越来越多的人力投入只可能导致其边际报酬的递减。”24

    为避免农业生产上的过度内卷，充分利用劳动力巩固和发展集体经济，1959年年初中共中央农村工作部对当年全国农村人民公社的劳动力进行了分配规划，提出将51.4%的农村劳动力用于农业生产，剩余的48.6%用于国家工业交通、林牧渔副业、社办工业、交通运输、基本建设、生活服务等方面；在农业中，从事粮食生产的约为8000万个劳动力，占农村总劳动力的38.1%，从事其他作物生产的约有2793万个劳动力，占农村总劳动力的13.3%。25也就是说，从事农业生产的劳动力仅占总劳动力的一半，而真正种植粮食的劳动力不到4成。26所以十队的一位妇女说：“强的劳动力又抽出去了呀，就剩下二、三级的婆娘在家了，有的上山搞副业，没有多少劳动力的。” （XJA，十队社员，2017年3月25日）五队老队长补充道：“（种田）天天都是那帮人的。上调的人做不了，他不做这个就去做那个，做田就是做田的，我搞副业就是搞副业，分了工的。”（WGM，五队队长，2017年3月24日）

    笔者在各生产队的账册中，发现不少专业队和副业人员的回队账单。例如，十队“1971年5月10日，收许有昌交款回队12—3月48元”（修建广西金红铁路，简称“6927工程”），27八队“1972年1月26日，收其文11—12月回队款23.6元” （专业队修船坝），28“1977年3月 14日收世天泥水工入队8元”。29当时规定专业队队员和从事副业的人员必须按一定比例将收入交回生产队，生产队按同等劳动力记工分，这样才能参与生产队的分配，同时生产队还要按时给外出的专业队队员寄口粮。例如，1969年广西从玉林抽调民工18000人，参加金红铁路修建工程，其中容县被抽调3000人。工程文件规定，“民工的生活待遇，每人每月30元，其中40%交回生产队，参加生产队分配，60%由民工个人支配。民工的口粮供应，除从生产队带足本人的口粮外，按工种定量标准，不足部分由国家供应”。30由于路途遥远，口粮是无法送去的，生产队只能通过转账的方式给民工购买口粮，如八队“1970年9月20日，支成才转6927（工程）粮200斤，每百斤9.3元，金额18.6元”。31可见，除去各级专业队队员、副业人员、民办教师等精干劳动力，真正进行农业生产的劳动力是很有限的。在非常有限的劳动力从事农业生产的情况下，其产出自然不会太高。

    1974年，广西壮族自治区革命委员会水利电力局提出要大力组建专业队。“不但骨干水利工程要坚持常年施工，而且社社队队都要组织农田基建专业队，大搞常年施工。一个队、一个社、一个县如果抽出百分之十的劳动力，一年坚持施工十个月，就等于抽调百分之五十的劳动力每年突击两个月要完成的工作量。”32在容县，仅从1974年至1975年2月25日，全县动工大小水利工程727处，完工243处；完成造田、造地10896亩（其中造田5337亩，造地5559亩），另开茶山地9059亩；完成改土面积11.63万亩，共用去452.8万工日。33那么在集体化时期，容县在农田水利基本建设上大概用了多少工呢？

    图1显示，新中国成立后，容县在集体化时期的农田水利基本建设完成的劳动工日的变化。由于这是官方统计资料，所以其中的数据只统计较大的工程，如华六大队除了大陂水库，其他四个小水库均未统计在内，34即还有很多大队、生产队自主修建的小型水库、山塘、沟渠等都没有统计在内。即便如此，以上数据也在总体上体现了集体化时期劳动投入的规律。新中国成立初期，由于国力较弱，集体经济制度还未建全，人们只能对小型水利设施进行修缮，投入的材料和劳动都很少，只有23.5万工日。1953年至1959年间，容县从农业互助组过渡到人民公社，完成的工作量明显增加，完成劳动日35也随之剧增，特别是1958年前后，也就是在“大跃进”时期，劳动投入达到一个小高峰，共投入565万工日。在1970年到1978年间，无论是在工作量上，还是在完成劳动日上，都呈现梯度式剧增之势，特别是在1976年，达到历史的高峰，耗费了1007万工日36。1980年以后，农田水利基本建设基本处于停滞状态。另外，在所有完成劳动工日中，水利用工占了绝大部分，主要是用于兴修大小型水利工程。1980年，由于集体经济的解体，大量农田水利基本建设失去了生产队的人力和物力支持。

    综上可知，一方面，在集体化时期，容县乃至整个广西都抽调了大量劳力进行农田水利基本建设。采取的方法是，专业队常年施工与群众性突击相结合。专业队不仅有建设专项水利工程的，还有从事造田、造地等农田水利基本建设的，另外在级别上还分为大队级、公社级和县级的专业队。这样无论是在农忙时，还是在农闲时，大量劳力都被抽调出去进行各类农田水利基本建设。另一方面，这些农田水利基本建设属于共同生产条件的改进投入，对山区生产队的农业生产尤为重要。虽然短期内对农业生产的影响并不明显，但在灾荒之年，它在一定程度上可以避免或者降低灾害带来的减产程度，甚至可以保证部分农田旱涝保收。

    二、生活性公共产品

    人民公社除了为当地提供大量农田水利基础设施外，还为广大社员提供了各类生活性公共产品，包括文化教育、医疗保健和社会救济等。这些公共产品并不是完全由国家来提供，绝大部分是由当地人民群众自力更生、自筹解决的。这些公共产品的积累并不会在短期内提高农业生产率，只有经过较长时段后，才能显现它们的作用和影响力，所以卢晖临主张要“打开视野看效率”，特别是延后的效率。37而要实现这些积累，社员不得不从相对干瘪的腰包中再掏出一部分劳动产品，这样就会导致分配给社员的产品总量减少，体现在工分上就是工分贬值，进而影响社员的生产积极性。

    （一）以民办教师为主的基础教育

    1969年之前，华六大队有两所小学，共4名公办教师，当地整体文化水平较低。据1964年第二次全国人口普查，华六大队有1392人，具有初小（小学一年级至四年级）以上文化水平的只有664人，占总人口的47.7%；石寨公社有18130人，具有初小以上文化水平的有9425人，占总人口的52%，其中只读完初小，13岁—40岁的青壮年有2686人；读完高小（小学五年级至六年级）的有3348人；初中文化水平有606人；高中文化水平有90人；拥有大学文凭的只有9人。38为提高广大人民群众的科学文化水平，1969年广西要求各地将农村公办小学下放给大队、生产队办，农村公办中学下放给当地社、镇革命委员会直接领导和管理。经县、社统一调整后，仍缺教师的大队，根据实际需要，选拔民办教师充实教师队伍。选拔的要求是：家庭出身好，并有一定教学能力，如果是复退军人和知识青年，则优先录用。对于这些民办教师的工资待遇，补助多少由贫下中农讨论决定。39

    1970年，华六大队共有4所小学，1所小学附初中，公办教师7人，民办教师9人。40在集体化时期，公办教师的薪酬全部由国家支付，而民办教师的薪酬由生产队承担（统筹）。华六大队的年终统计表显示，1973年，十队上交了981斤统筹粮和161元统筹金，其中统筹金是为4名大队干部、4名民办教师以及1名兽医上交的。41然而，同年，华六大队共有13名民办教师，一般生产队原则上选派1名教师，十队由于和九队合开一所分校选派了2名。据当时的大队干部介绍，并不是所有的民办教师都可以统筹，只有教得比较好的才有资格统筹。至于没有得到统筹的教师则回各自的生产队记工分，大队再发给少量的补贴。42

    随着教育事业的发展，到1978年，容县有6161名教师，其中民办教师3626名，占教师总数的59%；43华六大队共有7名公办教师，16名民办教师，44民办教师占比约为69.6%。由于小学教师大部分是民办教师，业务水平低，课堂教学中出现差错屡见不鲜，再加上“半天学习、半天劳动”，“以劳代学”的教学安排，学生学习规律被打乱，知识基础较差，甚至出现大量留级现象。为提高教学质量，容县积极采取多项措施，通过举办轮训班，办函授学校、进修学校，开展巡回辅导等提高教师的业务水平。45

    教学质量不高，除了教师能力不足以外，民办教师的工资待遇没有得到很好的保障，使其不能安心教学也是重要原因。“我县民办教师（不包括自筹教师）的生活待遇有两种，一是国家补助加大队统筹，二是国家补助加生产队记工分，不足部分由学校学费或勤工俭学收入补足。不管是采用哪种办法的，都有较长期拖欠民办教师工资的问题。”46拖欠情况包括：一是教师工资统筹不上来。例如1978年，华六大队5个民办教师，总共被拖欠工资551元，人均被拖欠110.2元。二是教师工资未发齐。部分大队给民办教师一年只发十个月的工资，而且每月工资未达到初中教师30元、小学教师24元的标准。三是教师粮食收不齐。部分大队规定，民办教师的粮食要本人到各生产队收，然而，实际上有的粮食收不全，有的收到次等谷。47可见，当时民办教师待遇存在长期拖欠和粮食以次充好等问题，极大地影响了教师的正常生活。

    根据收益分配统计表的数据，华六大队在1980年已经从1979年的13个生产队分为18个生产队，不少生产队内部开始酝酿分田分地了。民办教师的工资和粮食主要由生产队提供，生产队的解体必然会引起民办教师群体的动荡。1981年，容县教育局在汇报普及教育工作时指出：“我县最难解决的问题有：我县民办教师比例大，群众负担较重，近年来，由于生产队体制的改变，民办教师的粮、款很难统筹解决，严重地影响着民办教师生活和工作的安定。”48十队的许某正是由于分队，导致报酬没有兑现，退出了教师队伍。“（19）80年分田到户，这里（十队）分成三个小队，我们没有统筹得上，我就不做了。”（XJA，十队民办教师，2017年3月25日）由于民办教师和其他上调人员的物资、工分很难从生产队进行统筹，教师队伍面临严峻挑战。从表1中1979年至1981年的数据变化便能推断出各级单位大量缩减支出。虽然华六大队和石寨公社的劳动日与分配给社员部分的金额都减少了，但是分配给社员部分的金额占总收入的比重增加了（华六大队增加了7.54个百分点、石寨公社增加了4.93个百分点）。

    为解决民办教师的教学质量和后勤保障问题，1981年，容县教育局和财政局开始整顿民办教师队伍，辞退了思想品质、业务水平和健康状况不能胜任教学工作的教师，业务能力强、业绩突出的民办骨干教师则被吸收为公办教师。49华六大队的民办教师除了部分因为工资低没有坚持下来，大部分后来都转为编制内教师，成为真正的骨干力量。

    容县在1950年至1981年间，小学生人数从13236人增加到78134人，教职工从641人扩大到3749人（含民办教师1910人）；中学生人数从898人增加到30435人，教职工从98人增加到2317人（含民办教师1039人）。50小学生数量增长了近4.9倍，教职工人数增长4.8倍，适龄儿童的入学率高达93.4%。1981年，容县中小学民办教师人数仍占教师总数的48.6%。据统计，当年全国有民办教师近396.7万人51，占教师总数的47%。支撑这支庞大的民办教师队伍的是全国600多万个生产队52。保守估算，一位民办教师的月工资约为24元，国家和生产队各承担12元，生产队每年还要另外提供600斤口粮（100斤口粮折价为9.5元，600斤口粮折价为57元），因此每位民办教师需要生产队每年为其支付201元，全国396.7万名民办教师每年需要生产队支付79736.7万元，10年便接近80亿元，平均每个生产队10年共支付1333元支持基础教育。事实上还有相当一部分民办教师没有得到统筹，需要回生产队记工分参加分配。53当然，这些支出是值得的。据1982年第三次全国人口普查统计，华六大队有1720人，小学以上文化水平的人数为1398人，占总人口的81.28%，比1964年高出33个百分点；整个公社有24492人，小学以上文化水平的有20275人，占比为72.95%，比1964年高出20个百分点。54可以说，民办教师在广大农村地区，极大提高了社员的科学文化水平，这些学生在改革开放后，逐渐成长为社会主义建设的主力军。

    （二）以赤脚医生为主的公共卫生

    除了教育事业，农村的医疗卫生事业也主要由生产队负担。在毛泽东要求“把医疗卫生工作的重点放到农村去”的“六二六”指示的推动下，全国各地都把这项工作当作一项重要的政治任务，迅速组织医疗队，开展农村合作医疗。

    为积极响应中央号召，使广大群众看得上病，看得起病，吃得起药，1966年5月1日，容县人民委员会卫生科根据中央文件，制定《关于实行合作医疗的卫生所的有关意见》，这份文件成为容县后来开展合作医疗的重要纲领。“凡实行合作医疗的区，则在全区范围内看病不收诊费（门诊、出诊）、注射费、处置费；凡有条件的卫生所要开设中、西药柜，以利方便病者，减轻社员合作医疗负担，解决医生部分工资和卫生所办公费，还可以解决部分贫下中农的医药困难的减免；医生到生产队巡回要背下乡中西成药下去，以利方便病者，但要实行保本保值，收入归卫生所；诊费、药价要坚决贯彻执行国家规定的标准收费；为了减少病者负担，每个医生（接生员）都要学会针灸、使用针灸和使用中草药医疗疾病。”55从这些意见中可以知道，容县主张医生要通过各种手段，尽可能地减轻人民群众的负担，收入归集体，强调中西医结合，特别要充分利用中药和针灸为社员治病。

    要健全合作医疗制度，除了上述规定外，还要解决好医生的生活问题。1967年1月，容县发布《关于人民公社成立卫生所，医生、接生员实行合作医疗制度的通知》，规定每个公社成立卫生所，每个卫生所安排医生1—3名（逐步配备中、西医生1—2人），接生员1—2人。医生和接生员领取的粮食和工资全部由公社统筹解决，医生的月工资为15元至30元，接生员的月工资为15元至20元，另外，他们每月领取大米30斤；统筹粮由生产队统一送当地粮所，粮所则每月按量供应大米。56此文件对医生与接生员的待遇进行了相关规定，但是“粮食、工资全部由公社统筹解决”只是把问题抛给了公社，工资到底怎么解决并没有明确规定。统筹粮经粮所再转到医生手中虽然更有保障和方便管理，在现实中却很难实行，尤其像华六大队这样的山区大队，离县城路途遥远且崎岖不平，医生每月领粮既费时又费力，所以后来大部分大队都是让医生到生产队挑粮而不是到粮所领取。为进一步减轻人民群众负担，容县对药品、医疗器械的采购和零售价格作出规定：“今后凡已实行合作医疗大队的卫生室到所在供销社（县医药公司各门市部）采购中、西药品，医疗器械不论金额多少，一律按批发价作价供应。各卫生所一律按当地供销社零售价销售。”57这些规定从成本、服务等方面要求尽量以最低价格为广大人民群众提供服务。

    由于合作医疗制度是个新事物，具体怎么做只能不断探索，寻找适合本地的制度。容县采取树立典型、相互学习的办法让合作医疗尽快办起来。1970年石头公社的合作医疗办得较好，成为各公社学习的对象。该公社的卫生队伍包括：大队医生、采制药人员、接生员和生产队卫生员。大队级人员的报酬向生产队统筹，实行工分加补贴的办法，医生一般每月补助5—10元，其他工作人员每月补助3—4元，卫生员则回生产队记工分。对于合作医疗资金的筹集，由生产队统一计算按参加人数支付。每人交1元，其中个人交0.5元，生产队交0.4元，大队和公社各交0.05元。在收费制度方面，大队卫生所一般收挂号费0.05元，出诊费0.1元，注射费0.1元，接生费每个小儿0.5元，这些费用由病人负担，药费全部由合作医疗开支。如果是重病号到公社以上医院治病或住院，合作医疗支付60%，剩余的40%由病人负担。合作医疗主张自力更生、全民办医，贯彻“三土（土医、土法、土药）”“四自（自种、自采、自制、自用）”方针。石头公社各级单位均设有草药室，以草药为主（用量要求达到70%—80%），中药为辅，适量西药备急，其中草药的来源为：抽专人采集和群众献药相结合，三级有专人采药、制药，采药、制药人员报酬由大队负责。58石头公社根据本社的经济状况，各级单位分摊社员的部分医疗资金，同时大力采用中草药治疗疾病。因为容县山多，药材丰富，可就地取材，加上生产队种植草药，大大节省了药费开支。

    “合作医疗是收每个人的钱，那时没有收钱（看病），试过两年吃药不要钱，之后就不行了，反正像大队的企业那样。”（TFQ，三队赤脚医生，2017年3月17日）华六大队的佟某当时是一名赤脚医生，1947年出生，1965年在容县学医，1968年9月在大队开始行医。对于赤脚医生的报酬，他说：“最初几年就是吃工资，（19）68—（19）72年，24元每月，8毛钱一天。工资是从各生产队统筹，整个大队有副业人员、大队干部、医生，全部按照整个村有多少收入，再分配每个月多少钱，各个生产队抽多少上来，统一分配的。1972年以后就是吃工分，做医生就相当于搞副业一样，每天记12分。算起来就是三四毛1天，有些生产队只有两三毛，那时很穷的。以前我们容县村医大部分都是吃工分。一个月是50斤稻谷，一年600斤。”（TFQ，三队赤脚医生，2017年1月6日）但是，对于1972年以后一直是吃工分的说法，在曾任大队干部的陈某那里得到不一样的答案。“医生就是从利润那里支付工资，粮食就从村里统筹，老师的工资和粮食也是从村里统筹。”（CPY，华六大队会计辅导员，2017年7月6日）陈某1974年12月至1978年冬在大队任会计辅导员。59由于合作医疗制度在不断完善，华六大队根据上级相关政策，既实行过工分加补贴，也实行过工资制。

    1972年，广西制定了《农村合作医疗制度试行草案》，规定“凡是参加合作医疗者，按规定交纳合作医疗基金或以药代金。基金由个人和集体（公益金）负担，负担比例由社队根据情况自行确定。由生产队统一计算按参加人数支付”；要合理解决赤脚医生的报酬和口粮，其报酬要略高于同等劳动力的水平。60“合理解决”意味着赤脚医生的报酬既可以是工资的形式也可以是工分的形式，只要合情合理，并能够调动医护人员的积极性就行，所以1973年，容县各公社的赤脚医生报酬存在各式各样的形式，例如“实行工资制，开支从收入中解决……合作医疗变大队企业，收入归大队，赤脚医生实行工分加补贴，全部向生产队统筹”。61到1974年，石寨公社有23个医生，报酬都是以工资的形式发放，每月工资最低20元，最高30元。62到1977年，《广西壮族自治区农村合作医疗章程》规定赤脚医生的报酬为：“实行‘工分加补贴’的办法，每年由大队根据赤脚医生的政治思想、工作表现、技术水平、劳动态度等情况评定，一般应略高于同等劳动力的收入水平”。63那么“工分加补贴”具体是如何实行呢？这在1978年《关于加强合作医疗基金筹集和稳定赤脚医生报酬的请示报告》中有介绍：“每个赤脚医生每月在队记260分或300分，每天补助贰角生活费，每月补助六元，有的补九元，按该医生所在队分值计算工分所得部分，平均每月加生活补贴不达24—30元的，再从合作医疗收入中补足。”641979年，容县179个大队中，实行合作医疗的有155个，共有627名赤脚医生。在本大队报销的比例，大部分在30%—50%之间，上送报销比例在20%—40%之间，其中华六大队的合作医疗报销额度是30%。65

    在广大赤脚医生的努力下，1982年，容县60岁—90岁的人口从1964的26517人增长到42699人，占当年总人口的比重由7.07%提高到7.69%。661982年全国人口已超10亿人，60岁以上人口比重达到7.62%，比1964年的6.13%高出近1.5个百分点，67这在一定程度上说明我国医疗水平和卫生保健系统更加完善，而这离不开无数赤脚医生和基层医护工作者的默默奉献。1980年全国农村赤脚医生总人数达146.3万人，其中女赤脚医生48.9万人，农村生产队卫生员235.7万人，农村接生员63.5万人。68而这些不脱产医护人员的工资、口粮主要靠生产队解决。仅就工资方面，医生的月工资为24元，一年为288元，146.3万人一年工资共为42134.4万元，10年便是42亿元。事实上生产队所付出的要远远高于这一数字。国家只支付了少量的管理费和药费，以非常低的成本构建了完善的农村医疗卫生系统，保障了社员的身心健康，提高了出勤率，促进了集体经济的健康发展。

    （三）保障困难群众的基本生活

    在大部分人的回忆中，似乎并没有什么困难户，因为大家都很穷。然而，贫富只是相对的。在各生产队的账本中，笔者发现不少困难户领取国家救济金的凭证。例如，笔者在八队的账本中看到1972年3月7日，“大队拨来仕华救济金10元，交丽梅领”，69后面还有刘丽梅的印章。大队保存的阶级档案显示，陆仕华生于1932年9月，1972年已40岁，一家6口人，育有两儿两女，均不满10岁。70从这些情况来看，陆仕华一家的生活非常艰难。

    生产队用来救助军烈属、五保户和困难户的资金、粮食，一般是用公益金。公益金“要根据每一年度的需要和可能，由社员大会认真讨论决定，不能超过可分配的总收入的百分之二至三……生产大队对于生活没有依靠的老、弱、孤、寡、残疾的社员，遭到不幸事故、生活发生困难的社员，经过社员大会讨论和同意，实行供给或者给以补助”。71

    除了上述困难户外，还有一类困难户往往被人们所忽视，那就是“超支户”，顾名思义，即一年的收入不足以抵扣一年开支的农户。社员一年的劳动收入是通过工分来兑现的，生产队通过工分把各种生产、生活资料分配给社员。如果他们一年的工分收入不足以抵扣其一年的开支，那么这一年不仅没有盈余，反而欠生产队的钱粮。本文以八队的陆仕忠一户为例展开说明。

    1976年，陆仕忠一户共有7人，夫妻二人加五个子女，大女儿1962年生，14岁，属于半劳力；第二个是儿子，1964年生，12岁，其他均为10岁以下儿童。72从表4的支出中，可以看到，陆仕忠一家支出金额最高的是口粮，全年消耗口粮3548.1斤，平均每人消耗506.87斤，需支付335.17元，占总支出的91%；当年挣得工分8599.3分，每个工分值为0.38元，全年总收入为326.77元，不足以抵扣总支出（367.93元），超支了41.07元。

    八队在1976年共有6户超支，11户有盈余，4户平收，总户数为21户，超支户约占29%。这个比例在华六大队应该说是非常低的。1976年，华六大队超支户高达186户，占比55.3%，欠款共计11287元，73不管是占比还是欠款数额都在集体化时期达到最高值。由于欠款数额不断累积，到人民公社后期，生产队处于入不敷出的艰难境地。

    为何会产生如此多的超支户？这是一个不得不探讨的问题。

    超支户的存在，表面上看是农户挣的工分不够多，不足以抵扣从生产队获得的生活物资，本质上是因为生产队的物资不足导致工分含金量不高，以至于农户的工分不够支付其生活开销。如果物资充足，每一个工分所含的物资就更多，大部分农户的工分是足以支付其生活开销的。而物资短缺又与农业的产出密切相关。那么农业产出为何不高呢？当笔者把这一问题抛给村里的老人时，往往得到的答案是：没有肥料和农药。

    农谚说“有收无收在于水，收多收少在于肥”。“那个时候由于生产条件落后，种子也很落后，肥料在市面上也很少有卖。一般都没有肥料来卖，到后期才有这个碳铵和这个氨水。（19）80、（19）79年以前都是没有肥料供应的，基本上是山上的草皮泥，也就是这些人上山铲这个草皮泥来烧，烧了以后再撒到水田里面去，过去都是这样耕种的，也没有什么杂交种子，都是落后的种子，一般是（收获）200—300斤每亩，现在（每亩）都有1000—1200斤。”（CPY，华六大队会计辅导员，2017年1月6日）“过去主要是没有肥料，没有这个良种，现在则有良种、有农药，所以生产好，过去喝粥也难有喝。”（HZN，六队队长，2017年1月6日）

    在八队1977年的分类账中，“农业支出”记录了一整年的所有支出项目。经笔者统计，八队当年共购买了复合肥2斤，尿素415.1斤；碳铵10950斤，包括一级碳铵和次级碳铵（肥力较低，价格较便宜）；农药品种有“乐果”“毒杀芬”“六六粉”“敌百虫”等；早稻浸谷种2270斤，晚稻谷种2884斤，共5154斤。74八队当年有130.5亩耕地，其中水田117亩，旱地13.5亩（4.2亩自留地），75除去自留地，集体实际拥有耕地126.3亩，两季共252.6亩，平均每亩施1.64斤尿素和43.35斤碳铵，每亩水田要22斤谷种，农药以“六六粉”为主。投入这些生产要素后，当年八队共收获109523斤稻谷，亩产468斤。76

    此外，农业产出低还受到生态环境的制约。正如黄宗智对新自由主义经济学理论批判的那样，农业不同于工业，不是投入的生产要素越多，单位产出就越多，甚至总量和产出几乎可以无限制扩大。把农业等同于工业，本身就是对农业的误解。农业说到底是人在土地上种植植物的有机问题，而不是一个机器生产的无机问题。因为农业生产受地力和生态环境的限制，土地不可能无限产出。77很可能一场洪涝或者干旱就能把农民辛苦劳作一年的成果化为乌有。

    从表5可知，容县在1969年—1982年的14年间，影响早稻的各类自然灾害频繁发生，发生率从高到低依次排列是：病虫害、龙舟水、倒春寒和夏涝。需要注意的是，表5并未统计对晚稻影响较大的寒露风。当这些灾害组合性地发生时，会给农业造成致命打击。例如1976年，由于倒春寒的发生，当地烂秧严重，既损失了大量稻种，又推迟了播种季节。不巧的是，当年不仅出现龙舟水，病虫害也大发生，由于预防及时和经营管理得较好，早稻损失不大。但是，由于早稻种植推迟，导致晚稻插播也推迟，这样就使晚稻在扬花灌浆期遭遇寒露风。“抽穗扬花期遇到寒露风天气，直接影响抽穗开花的速度，使空秕粒增多，降低千粒重，造成减产。”78当年水稻产量八队比1975年减收5874斤，人均分配口粮减少20斤；华六大队减产110075斤，人均分配口粮减少54.9斤；石寨公社减产1178295斤，人均分配口粮减少56斤。79这最终导致华六大队的超支户数量由1975年144户增加到186户，占比为55.3%。同年，容县减产3423万斤，人均分配口粮减少70.6斤，超支户由35005户增加到36779户，增加了1774户。80这些数据说明，农业生产深受生态环境的制约，尤其是自然灾害对农作物的影响。然而，经济学家们往往有意或无意地忽视了这一重要因素。生产队有超支户、平收户和盈余户，其中最容易由不欠生产队转变成欠生产队的农户是平收户。自然灾害对平收户的影响，就像“一个处身于水深没颈的人，即使是一阵轻波细浪，也可能把它淹没”。81可以说，生产经营中的任何一个环节出现异常，都有可能使平收户变为超支户。这也是为什么在集体化时期，人民公社要进行大量的农田水利基本建设。有了完善的农田水利设施，可较好地降低自然灾害对农作物的损害程度，使得农民在面对寒露风时，并不是无能为力。由于容县历年出现寒露风概率最多的时间是从每年10月11日至11月10日82，所以较好的办法是种植早熟和中熟的稻种，这样就可以让水稻在抽穗扬花期避开寒露风，但这需要优良的稻种。此外，根据广大人民群众长期的耕作经验：“有水不怕寒露风”，在寒露风到来之前往田里灌水，就可以保存地温和增加稻田小环境的温度，从而减轻寒露风对水稻的危害。83而要有大量水源，就需要水库贮存水，以及通过相应的沟渠和设施把水引入田中。

    当然，生态环境并不是造成农户超支的主要原因，它只能在一定程度上限制生产队农业产出的总量。造成农户超支的主要原因是国家与集体从生产队中提取了过多物资。国家之所以提取大量物资，是为了满足工业化的需要。陈云在1950年6月说：“中国是个农业国，工业化的投资不能不从农业上打主意。搞工业要投资，必须拿出一批资金来，不从农业打主意，这批资金转不过来。”84刘少奇也认为：“发展中国经济，使中国工业化，是需要巨大的资金的……但是从哪里并且怎样来筹集这些资金呢？……只有由中国人民自己节约……而要人民节省出大量的资金，就不能不影响人民生活水平提高的速度，就是说，在最近一二十年内人民生活水平提高的速度不能不受到一些限制。这并不是为了别的，只是为了创造劳动人民将来更好的生活”。851955年7月31日，毛泽东强调：“为了完成国家工业化和农业技术改造所需要的大量资金，其中有一个相当大的部分是要从农业方面积累起来的。这除了直接的农业税以外，就是发展为农民所需要的大量生活资料的轻工业的生产，拿这些东西去同农民的商品粮食和轻工业原粮相交换，既满足了农民和国家两方面的物资需要，又为国家积累了资金。”86可见，在集体化时期，人民生活水平的提高和加快工业化进程是矛盾的。国家从长远考虑，只能牺牲人民生活水平的快速提高。

    1960年，《中共中央关于农村人民公社分配工作的指示》指出：中央原来规定的总扣留占40%左右，分配给社员的部分占60%左右。如果当地收入水平较高，如每人分配在100元以上的，扣留可以多于40%；如果收入水平较低，如每人分配在50元以下的，扣留可以少于40%。87也就是说，正常情况下，人民公社要向国家和集体贡献大约四成左右的劳动成果。虽然人民公社制度在不断调整，但这一核心规定一直贯穿于集体化时期。1974年玉林地区分配给社员的部分占总收入的比重只有53.94%，该地区当年分配给社员的部分占比最高的是平南县，为55.77%，最低的是陆川县，为48.54%。88当“上下左右向生产队伸手，四面八方挖生产队墙角”89时，社员辛苦劳作一年，分配总量甚至不足一半，超支户怎能不多？

    除了生产水平低、生态环境制约和国家、集体抽取过多物资，还有一个重要原因直接影响超支户的数量，即人民公社的分配制度。分配制度是生产关系的一部分，采用什么样的分配制度取决于生产发展的水平。1962年通过的《农村人民公社工作条例修正草案》指出，粮食分配应根据本队的情况和大多数社员的意见，分别采取各种不同的办法，可以采取基本口粮和按劳动工分分配粮食相结合的办法，也可以采取按劳动工分分配加照顾的办法等。不管采取何种办法，都应该做到既要调动大多数社员的劳动积极性，又要确保困难户能够吃到一般标准的口粮。90虽然国家要求生产队要遵循按劳分配、多劳多得的原则，避免分配上的平均主义，但是在实际分配中，基本口粮占比往往较大，很难进行真正意义上的按劳分配。

    “在目前口粮不高的情况下，必须首先保证各等人口留粮放在安全线上，过分强调多劳多吃，是不符合粮食分配原则，是不正视当前粮食状况，是没有全面了解社员的要求，其后果，必然引起今后粮食安排的被动，亦不能达到发挥全体社员的劳动积极性。”91所以，华六大队在集体化时期粮食分配的70%按人口定量分配，30%按劳动工分分配。在农业生产水平较低的情况下，生产队首先要保证每一位社员都有口饭吃，也就是学者们所说的生存伦理92，当社员的基本生活都难以保障时，生产队就会面临解体的风险。如果国家政策允许生产队切实贯彻按劳分配，多劳多得，不劳动者不得食的社会主义分配原则，社员的生产积极性可能会大大增加，超支户的数量也可能会减少，但是也可能会导致部分农户的生活非常困难，甚至饿死人。这样的结果不仅国家政策不允许，熟人社会中的道德规范也是不允许的。虽然按三七开的比例分配物资具有一定的平均主义倾向，但它在保证大部分人的基本生活和激励劳动力积极出工参加生产活动上较好地进行了平衡。

    三、小结

    第一，人民公社为广大乡村提供了丰富的公共产品，内容涉及社员生产、生活的方方面面。与当下的政策不同，集体化时期的公共产品均由生产队或生产大队自我供给，生产、运输、管理、消费等各个环节都在本地进行，并没有获得足够的财政和物资支持。然而，这恰恰表明集体经济具有社会经济的属性，即经济活动和参与经济活动中的人及其所在的社会网络是紧密地结合在一起的，它们是相互嵌入的关系，集体经济的效益最终是让所有社员都能够受益，而不是像资本主义经济那样，脱离地方社会和文化，以攫取地方社会资源为目的进行经济活动，虽然经济效益非常可观，但是将所有的问题和矛盾都遗留给当地，以竭泽而渔的方式破坏当地的可持续发展。潘毅认为，社会经济的要旨，就是以人为本，立足社区而不是让资本剥削社区，互助合作，民主参与，人类与土地和谐共生。生产不是为了消费，而是为了解决民生，追求共同富裕，是一种多元化的社会所有制。在本质上，社会经济不是服务于资本累积，而是将社会重新嵌入社会关系中的一种新形态的经济模式。93

    正如毛泽东在《中国农村的社会主义高潮》编者按中所言：“人民群众有无限的创造力。他们可以组织起来，向一切可以发挥自己力量的地方和部门进军，向生产的深度和广度进军，替自己创造日益增多的福利事业。”94作为社会经济重要组成部分的集体经济，在生产力发展水平较低的条件下，在农村修建了大量水利设施，尽可能地提高了土地生产效率，同时增强了生产队抗灾、救灾能力。此外，人民公社还广泛组织群众发展基础教育事业和医疗卫生事业。这些福利事业不仅价格低廉，而且在广度和深度上都动员了社员进行自我教育、自我成长和自我保健，满足了社员自身发展的需要。事实上，农民在集体化过程当中所受到的洗礼要远远高于笔者所看到的，包括管理水平、纪律教育和科技创新等，所有的这些都在塑造着“新型农民”，为改革开放后国家的飞速发展，提供了优质劳动力。所以，笔者以为，要实现乡村的再次振兴，必须把广大人民群众重新组织起来，使经济回归社会，尤其是作为社会经济的集体经济，这是一条可供选择的路径。

    第二，通过研究发现，人民公社时期的农业生产效率客观上的确存在效率低下的问题，例如人们的收入水平较低，生活条件改善缓慢等。但是，在这些事实背后蕴含着错综复杂的原因及逻辑，当本文剥离这些原因后再度审视集体经济制度时，发现导致社员收入不高的原因是工分被稀释了。农户总收入计算公式能很好地对此进行说明。

    由于农户的劳动力在一年或者数年内，基本保持不变或者变化不大，所以，农户总工分事实上是在相对平稳的区间内浮动。因此，影响农户总收入的因素主要是生产队的工分值。而导致生产队工分值变化的因素主要有两个，即生产队的纯收入（总收入－生产成本）与生产队总工分数。当纯收入保持不变时，生产队的总工分越多，即分母越大，工分值越小；当总工分数保持不变时，纯收入越少，工分值也会随之变小。所以，工分稀释化主要包括两个方面，一方面是工分的直接稀释，即把非农业生产的工分拿回农业之内进行分配，从而导致工分被稀释，分值下降；另一方面是实物和现金等物质上的间接稀释，即从生产队中抽走、消耗大量物资，减少生产队的纯收入，进而降低了生产队的工分值。如果把各级单位强加在生产队身上的各种“包袱”给抛弃掉，工分值和社员所得将会大大提高。

    第三，在学界，对人民公社批判最多的就是平均主义和“大锅饭”，其中“大锅饭”几乎成了人民公社的代名词，污名化非常严重。笔者以为把造成平均主义的原因归结为人民公社制度本身是值得商榷的。因为“人民公社低效率的原因是综合的，既有公社自身的原因，也有公社自身之外的原因，但公社自身之外的原因是主要原因，而不是相反”。95当国家和集体从生产队拿走过多的剩余产品时，可供分配的产品自然不足，人均占有量也就无法提高，如此才导致所谓的平均主义。经研究，本文发现，在生产力、人力和物资都非常有限的条件下，人民公社的农业生产仍能保持较平稳的增长，实属不易。同时，人民公社为支援国家工业化建设，提供了大量公购粮和农副产品；为满足人们的生产生活需要，在农村地区提供了丰富的基建、教育、医疗和社会保障等社会公共品。可以说，它的效果是多元的。因此，对人民公社的评价，不能仅局限于某一方面或某一时段，而应放大到整个国家层面和历史的脉络中进行考究，才能得出较为客观的结论。

    本文转自《开放时代》2024年第6期

  • 黄波粼 钟子善：上海农村集体托幼实践的考察（1958—1962）

    从思想史的脉络来看，关于托儿所与幼儿园的构想无疑具有相当长之历史。从柏拉图的《理想国》①，到启蒙运动时期众多的空想社会主义者②，至恩格斯③，后至康有为的《大同书》④、青年毛泽东⑤等都提出过“幼儿公育”的想法。从实践层面而言，清末以降，许多政治、社会力量都认识到了公共育儿的必要性，这使得公育思想在本土化建构中与实体建设并举⑥，托幼事业逐渐成为各种现代性设计不可或缺的一项重要内容。新中国成立后，公共托幼的必要性和重要性更加突出，因为恩格斯曾提出随着生产资料转归社会所有，“孩子的抚养和教育成为公共的事业”⑦，列宁也曾强调托儿所及幼儿园是共产主义“幼芽的标本”⑧。新中国托幼实践的大规模开展，大体发生在1958年至1962年，学界在这方面成果较丰，大多从妇女史的视角来回应马克思主义理论下的“妇女解放”问题。⑨有学者认为，大力推广集体托幼有着培育“共产主义新人”的意义。⑩有学者通过分析20世纪50年代末的农村集体托幼进一步指出，兴办托幼并非只为培育“共产主义新人”，它更是一个塑造“共产主义新农民”的过程。11尽管现有成果已经对新中国集体托幼有了比较深刻的认识，但仍有一些问题需要继续追问。例如，回到具体的历史进程和情境之中，国家诠释的“共产主义”具有哪些意涵？它们又是如何被农民接受的？本文拟以地方档案为主要史料，考察1958年—1962年的上海农村12集体托幼实践，着力展现其围绕各个阶段中心工作而曲折发展的轨迹，通过回溯若干具体措施深入农村的过程，呈现一段塑造共产主义新农民的历史。

    一、 塑造农民对于共产主义精神的政治认同（1958年9月—12月）

    延安时期，毛泽东给延安第一保育院题词：“儿童万岁”13，又强调一定要为教育后代而努力。新中国成立初期，“社会主义老大哥”苏联的集体托幼模式引发国人极大兴趣，14集体托幼在当时看来是一种与共产主义生活相适应的教养模式，被视为“共产主义萌芽”，直接关乎共产主义新一代的培养，“为将来的共产主义社会准备了‘人’的条件”，因而是“一万年都要做的工作”。15

    1958年北戴河会议后，毛泽东就指出人民公社办托儿所的重要性：“是搞钢铁，搞棉花、小麦重要？还是孩子重要？这是涉及下一代的问题。托儿所一定要比家里好些，才能看到人民公社的优越性”。16当年9月，国务院下发的《关于教育工作的指示》明确提出：全国应在3年—5年内，完成“使学龄前儿童大多数都能入托儿所和幼儿园”的任务。17上海农村托幼组织的大幅增长也是从1958年9月开始的。与此同时，上海农村确立将“家务劳动社会化”作为向共产主义过渡的重要内容。其中，“儿童教养集体化”是“家务劳动社会化”的首要目标18，即开办托儿所和幼儿园，集中教养7岁以下的社员子女19。除房屋外，摇篮、凳子、床铺、被子等都是开办园所必须具备的，全托还需食具、毛巾、水瓶、浴室、脚盆等。仅凭公社或生产大队积累下来的少量物资与资金很难满足办所办园所必要的物质条件。如何在“少花钱”乃至“不花钱”的情况下快速搭好托幼机构的架子，将儿童“迎”进来，成为摆在基层干部面前的首要难题。考虑到添置新的设备、物资需要花一大笔钱，且在短期内还不易购买到，因此不仅要坚持“因陋就简，勤俭节约，自给自足”的原则，还要发动群众通过借、调、征用等方式凑起必需物品。宝山县红旗人民公社第五生产大队全托幼儿园的建立就是一个典型例子。首先，大队向农民宣传“儿童集体化”的伟大意义，解释什么是“我为人人、人人为我”20的共产主义精神，并深入浅出地说明全托可以解放劳动力投入生产，能够更好地教养儿童，以及为何须自力更生，即如何遵循“勤俭办园”的原则。接着，动员有空房的或房子较大的社员，紧缩一部分住房，或移居到其他社员家中。最后，腾出一幢房子共八间，四间做宿舍，三间做教室，一间做炊事房，天井做小活动场，场地上扎竹篱做运动场地和花园。此外，要求入园幼儿的家长自带床及其他日常品，公共用具则是发动妈妈们有什么带什么，发扬集体互助精神。对实在困难、拿不出钱的家庭，就与其他孩子的家长协商合用，不够的部分，由干部发动其他社员适当添一些。最终，办园所需物品都是妈妈们自己送来的，共计24张床（包括床板）、50条被头，每个小孩1只矮凳、2只碗、1只匙。连没有小孩的金大妹也借出了长凳、马桶等。21典型事例的示范，鼓舞各地以共济互助精神大力兴办集体托幼。安乐生产队幼儿园的物资筹备过程亦是如此。该园于1958年11月7日建立，是由大队直接领导，社员在“不花钱”的原则下办起来的。22上海农村之所以在1958年秋季出现集体托幼的高潮，很大程度上是因为广泛推行此法。

    无论是房屋还是日常用具，筹备托幼，“物”的基本设施自然是题中之义，而将“人”和“物”两者结合起来考虑也是对建立集体托幼机构的基本要求。“人”的要素首先是要解决部分干部和多数家长“思想不通”的问题。对干部群体而言，并非所有的基层干部都支持开办园所，其原因在于他们认为托幼事业对农业生产不但没什么帮助反而会拖生产的后腿，因而“不划算”。另外，在“热心”办园所的干部眼中，集体托幼对于家长无疑是一件天大的好事，但事实上，多数家长，尤其妇女并不这样认为。尽管她们对于基层传达的集体托幼与自身解放的关系已是耳熟能详，却因“人在田里，心在家里”的切身体验，生出不少顾虑：孩子交给别人看不放心，怕别人照管不好得了病，怕保育员偏心眼等。23

    为了“打通思想”，大队干部通常会先在干部会议上通过算细账——“孩子在田里农作物损失，劳动力不能发挥等赔账”——让干部群体在兴办园所的必要性上达成共识。后要求干部召开妈妈或社员会议，设身处地地以妇女的实际家庭经济利益算家庭帐。如一家六口人，夫妻二人，一个老人带三个孩子，若三个孩子都送托儿所或幼儿园的话，老人就能去挣工分了。按一个老人一天最少挣5分计算，一个月可得150分。此外，针对妈妈们心理上的不安，以及小孩放在家里容易发生危险和事故等，进一步打消妈妈的顾虑。24不过，做通家长的思想工作并非易事，在农村工作“全面开花”的形势下需要耗费一定的时间与精力。

    “打通思想”之余，还须确定由谁来照管园所的幼儿。由于以农业生产为中心任务，当时的保教人员往往由大队干部指定女性半劳动力或辅助劳动力来担任，由此形成“青壮年上前方，老弱做后勤”的人员配备模式。上海县马桥大队、奉贤县南桥大队及松江县张朴生产队53个托儿所和13个幼儿园的保育员绝大多数是老妈妈，年龄最大的72岁，最小的46岁，平均年龄在50岁左右。不少身体残疾，无法参加劳动生产的妇女也当起了保育员。有个托儿所的盲人阿姨王修金，一个人同时带三个孩子。25还有部分园所则是“小囡带小囡”。宝山县红旗人民公社第五生产大队的全托幼儿园由5个教养员负责，年龄最大的21岁，最小的才13岁。26一旦孩子在户外活动，有些保教人员在体力上难免力有不逮。在“物”与“人”的准备环节，两者通常是同步进行的。由于“边组织、边教育、边行动”27，很多园所在短短几天就办了起来。值得注意的是，早在1956年教育部、卫生部、内务部三个部门就曾联合发文，“收3周岁以下的儿童者为托儿所，收3至6周岁的儿童者为幼儿园”28。因此，尽管因陋就简是一贯方针，但这一时期上海农村在开办园所的过程中还是严格遵循了将幼儿园与托儿所分开等相关要求。托儿所以生产队（即一个自然村）为单位开办，便于妈妈接送，幼儿园则以两三个生产队为单位联合开办，或以生产大队为单位开办。

    由此可见，人民公社化初期，兴办集体托幼除了具有基本的公共育儿功能，还有塑造农民对共产主义精神的政治认同的任务。农民群体的政治认同之所以重要，在于他们是党和国家确定的阶级柱石之一。需要明确的是，这里的农民实际指的是村民、农村基层干部、园所保育员及幼儿。尽管这些群体有各自的角色，但实质上仍是农民。1958年12月，中共中央高度肯定湖北省委《关于做好当前人民生活的几项工作的规定》，这份文件指出，办好园所“适用于农村，原则上也适用于城市”29，明确了托幼工作先面向农村的取向，更加凸显了这一时期塑造农民对共产主义精神政治认同的重要性。

    和多数群众运动一样，上海农村在推行集体托幼时因为急于求成造成了不少问题。特别是为了“赶、学、比、超”与应付上级检查，很多公社不管孩子是否有老人带，家长有无需要、有无意愿，大讲“托儿化”“包下来”，以“强迫命令”的形式组织儿童进园进所过集体生活。30从1958年9月30日实现公社化，到10月下旬，不到一个月的时间里上海农村成立了1400多个幼儿园，4300多个托儿所，480个托儿组，收托儿童10万多人。31当年年底，江苏省苏州专区所辖六县与南通专区崇明县先后划入上海，这时，上海农村地区辖有11个县。32由于所辖区域的扩大以及农村福利工作的持续推进，托幼机构数量及入园入所的幼儿人数就更多了。据上海市妇联统计，这一时期共办幼儿园、托儿所29603个，收托孩子 582762名，收托孩子占上海农村学龄前儿童总数的80%。33

    人民公社化初期，上海农村在推行集体托幼时虽在“物”“人”等问题上遇到诸多困难，但在动员及组织农民的过程中完成了塑造具有共产主义精神的“新农民”的第一步。尽管，因推广时急于求成而问题渐显，但上海农村的集体托幼并没有停止，反而在“全民托幼”的大潮中真正步入“实践期”，进入一个崭新阶段。这是因为它始终配合农业生产发展34和各个阶段的政策调整而进行，中心工作起起伏伏，农村集体托幼实践便随之波浪式地向前发展。35

    二、 共产主义议题凸显与农民集体托幼需求（1959年1月—8月）

    国家要求在农村大力推行集体托幼，并不意味着所有农民都会自觉参与。一些家长担心孩子在托儿所可能会“吃不饱”，或被大孩子“欺辱”，36上海市委妇女工作委员会发现“儿童实到数往往少于报名数，有时仅及一半”37。多数社员对生产大队负担全部托费不满意，有人发牢骚说：“领这两个囡，工分弄光”38，认为有孩子入托特别是有多个孩子入托的家庭占尽便宜39。之所以出现此类声音，是因为虽然新中国成立前后的土地改革重组了农村权力关系，使农民感受到了国家权威，但尚未改变其劳作和生活秩序，仍然如传统乡土社会时一样生活，没有产生公共育儿的需求。40随着20世纪50年代后期党和国家在农村的中心工作——人民公社化运动的到来，这种秩序才被彻底改变。自此，参加集体劳动与做好分配成为农民的基本任务和利益所在。为了保障个人在人民公社中的利益并维持其运作发展，他们较为普遍地“自动”产生了集体托幼的需求。

    站在历史的比较视野来看这一问题，会更加清晰。不独中国共产党，20世纪二三十年代进行乡村幼稚园试验的陶行知和国民政府也曾在推行托幼的同时着力推进集体合作，却都没能将二者结构性地联系起来。

    1926年，在关于创设乡村幼稚园问题的文章中，陶行知将平民化视作建立乡村幼稚园的关键之一。如其所言，农民贫且忙，幼稚园应济“农村需要”。41因此，他试图把建设乡村幼稚园与改善农民生计结合起来：通过就地取材、物尽其用解决房屋、用具等设备，以农民生产和生活时间为准安排幼儿活动，并将幼儿的康健放在第一位，以此实现幼稚园“下乡”。42陶行知建设乡村幼稚园的初衷在于通过解决农村幼儿集中教养问题纾解农民的“穷愚”困境，带有造福农民的公益性和福利性。但由于时局动荡与经费短缺，以及农村生产方式43等诸多因素，陶行知的乡村幼稚园试验没有也不可能促生农民内在的托幼需求。在此情况下，他力图通过建设乡村幼稚园为国育儿的愿望也只能止于零星试验。

    20世纪30年代国民政府主导的托幼事业，其政策体系包括宗旨、课程、经费、管理及师资等，无疑为托幼的发展提供了政策保障，这套政策体系的建构是“自上而下”与“自下而上”相结合，既有中央政府为主导的取向，也有专家学术研究成果的推进，如陶行知等的幼稚园试验对中央政策的制定助益良多。44因而，在相当程度上，国民政府认可了陶行知等人乡村幼稚园建设的理论及实践，只不过注重为推行托幼提供政策保障的做法仅在表面上加深了乡村幼稚园建设的“国家化”程度，并没有改变民间“自治”的格局。45诚如时人所批评的，在乡村，大部分的合作都被豪绅所把持，外界无法跳过他们去直接组织农民。46国民政府虽曾试图将政权下沉至乡村，但因轻视乡村又缺乏动员能力，以失败告终，47即国家权力难以“下乡”。概言之，虽然国民政府加大了政策层面的介入力度，但仍属民间“自治”的乡村幼稚园注定很难与农民的集体托幼需求结构性地联系起来。九一八事变后，国民政府的托幼事业虽有发展，但极不平衡。1932年，即便在托幼机构较为发达的上海，其幼稚园的幼儿总数也只有1045人，48而上海县、青浦县、南汇县、松江县、金山县、川沙县保姆所幼儿数合计仅190人，49乡村托幼严重落后。此外，当局更将扩大托幼规模与“培养教化国民”联系起来，50在未能创造出农民内在的集体托幼需求的情况下，此等流于空泛、无所着落的说教51预示着乡村托幼只能停留在专家的实践层面。

    反观20世纪50年代后期，中国共产党在农村推行集体托幼时，在国家强力主导下，首先将分散的小农全部组织到共产主义的人民公社之内。实践中，开始时大办集体托幼的热潮和随后的整顿总体上还算稳步、健康地发展。521959年3月以后迅猛发展，全国形成集体托幼高潮，这使“共产主义”成为农民不得不面对的严肃议题。对于一些无子女农民的不满，干部教育他们要基于集体利益，且要有长远眼光，认识到“办托儿所、幼儿园是人民公社的优越性，现在没有囡，将来有囡，现在没有囡，下代有囡”53。一些保育员认为自己工分低，还被家长看不起，情愿去田里生产，干部就开展“我为人人、人人为我”的辩论，讲明带孩子比生产更重要，使她们明了保育事业的意义，树立托幼工作的光荣感。54此种情形非上海独有，全国各地屡见不鲜。如河北省徐水县就“小孩子要不要由公社来抚养”展开辩论，没有小孩的社员认为自己“吃亏了”，不同意开办，有小孩的社员当即反驳：“你现在养活俺小孩，将来你老了还不是由俺小孩养活你吗？”结果，那些怕吃亏的人最终被“辩倒了”。55所谓“辩论”，事实上就是毛泽东批评的“动不动‘辩你一家伙’”56，实际上已容不得落后分子。上海市嘉定县要求办幼托的大字报就占了50%，奉贤县的家长一定要把未满3岁的幼儿送入全托，为解决孩子穿衣服的后顾之忧，还把布票也交给了幼儿园。57无论抱怨还是从众，对于农民来说，传统的劳作和生活秩序发生了彻底改变。既然进了共产主义的人民公社，就要参加集体劳动并做好分配。这意味着在家带孩子就少了参加集体生产的劳动力，影响全家拿工分及年底收入，于是，农民在传统乡土社会所没有的集体托幼需求被激发出来。

    概言之，与陶行知和国民政府推行的乡村幼稚园相比，中国共产党主导的人民公社化运动在凸显共产主义议题的同时，结构性地激发出农民集体托幼的需求。此种需求虽然源于国家这一外部性动力，却因为农民难以选择而成为其内在需要。毛泽东认为，小孩进托儿所，教育、食宿等都由社会负担，不是破灭家庭，而是废除家长制。58这也回应了费孝通对传统乡土社会孩子由家庭抚育的思考。费孝通曾明言，若“以家庭和保育院来比较的话，大体上家庭里所生长出来的孩子比较健全些”59。他还认为，乡土社会的农民只有在偶然或临时的非常态中才需要伙伴和团体60，除非中国社会乡土性的基层发生了变化61。如果说20世纪50年代中期的农业合作化运动所带来的正是这样一种根本性变化的话，那么50年代末的人民公社化运动所造成的农村集体托幼则是这种根本性变化的更高表现形式。

    关于人民公社化之后农民迫切需要集体托幼的原因，上海市幼儿保健教育委员会的调查报告极具代表性：

    对于有孩子入托特别是有多个孩子入托的家庭而言，孩子越多，用在孩子身上的补贴就越多。日托尚且如此，贴粮贴钱的全托对孩子的补贴则更多，生产大队除补贴25斤左右粮食外，每年用于一个全托儿童的费用更是多达30—40元。62

    类似的情况还发生在七一公社联明生产大队。由于是棉、粮、菜夹种地区，该大队较为富裕，一个劳动力全年平均收入为272.28元，63如果以此为基数，补贴一个全托儿童的30元—40元相当于一个劳动力一年收入的11%—14.9%。也就是说，一个家庭如果有一个孩子进全托，其所耗补贴相当于一个劳动力一个多月的收入。事实上，有些家庭不止一个孩子进全托。加之，托儿所、幼儿园的日常开支由大队公益金担负，在家长负担一定工分后，大队全年托幼经费的支出占公益金的65.6%，远高于困难户补助、工伤等其他福利性开支。64由于集体托幼实行包下来的政策，特别是“吃饭不要钱”65，在上海农村，甚至出现这样的情形：

    不少夫妻一方或两人从事非农业生产的职工家庭以及有亲戚在农村的家庭也将孩子送入托儿所幼儿园中，甚至有少数人代人领养，自己却“拿了领养费”。对于这些未参加集体生产劳动的家庭而言，无疑享受了与社员一样的福利。66

    农民内在的集体托幼需求正是在这样的情境下被激发了出来。换言之，人民公社化初期的集体托幼成为农村干部群众接轨共产主义的重要工具。

    国家不仅激发了农民集体托幼的需求，还通过兴办更完善的托幼组织充分保障这种需求的实现。1959年3月，全国妇联提出“在具有基本条件的地方”必须“积极办好全托”之后，上海市妇联要求配合人民公社化运动和农业生产发展需要，切实将办好全托纳入全面规划。67不仅将原有的临时性、季节性的托儿所全部转为日托，还兴办了大量新的日托与全托。大量园所的兴办确实解决了不少农村妇女生产牵累与孩子照管的难题，使女性安心投入生产。至此，上海农村的集体托幼真正成为一项“运动”。4月，仅嘉定、奉贤、松江三个县，共办起5539个托儿所，11个县共办15635个托儿所，入托幼儿达到233999名。68

    然而，1958年年底至1959年年初麻疹的流行，打乱了工作步调。由于卫生知识、传染病预防及隔离条件缺乏，致麻疹传染面扩大，造成不少幼儿死亡。69家长们非常惊慌，纷纷把孩子抱回家，致使托儿所、幼儿园缩减。不久，上海召开五级干部会议贯彻第二次郑州会议精神，“精简生活服务人员到7%左右”以补充、加强农业生产战线。在此形势下，奉贤县一生产队队长认为劳动力如此紧张，却让十几个人带孩子“不合算”，急欲把幼儿园“砍掉”，将省下来的保教人员充实到大田上去，70部分托幼机构形成了“无人照管孩子”的局面。这些言行不免偏颇，但由此也可以看出，生活还是要让位于生产。这种影响在1959年夏季全面体现出来，上海农村托儿所和幼儿园数量出现下降。如金山县新农公社，原有109个幼儿园，收托 1000 多名孩子，在五级干部会议后仅留下13个幼儿园，收50多名孩子，下降90%。71

    三、 在革命口号中促使农民奔向共产主义（1959年9月—1962年）

    1959年庐山会议后，中央层面开始“反右倾”斗争，这为处于低迷的农村托幼事业带来了转机。在“反右倾”的总形势下，农村福利工作的“倒退”往往被视为“右倾”的表现。1959年9月，上海开始着手恢复、提高农村福利方面的工作，再次确立“一手抓生产，一手抓生活”的工作方针。为了强调政治挂帅，各级党委书记都亲自抓托幼工作。72在干部群众中间，针对错误观点或行为进行了思想教育。73最为常见的教育方式是“调查”。例如，川沙县对一些公社进行调查后，发现不少家长“有需要而未送托”，在深入了解原因的基础上，不仅教育他们应关心儿童的安全，还设法帮助解决实际问题。74据说这个办法很灵验，通过调查及时发现并解决各种问题，使很多家长、保教员及基层干部都期盼将日托转为全托。75

    “反右倾、鼓干劲”开始后，上海市委在基层公社紧锣密鼓地宣传八届八中全会精神，要求各项工作“鼓足干劲”76，提出了不少脱离实际的园所建设任务和办园目标，农村再次掀起“全面跃进”的高潮，加强对托幼的领导成为基层干部的重要工作。如南汇县惠明公社明六生产队周水连听了会议精神传达后，回去立即办了4个托儿所和4个幼儿园。77

    在“大跃进”的氛围下，为了不断开办新幼儿园以完成高指标，1959年12月，上海市教育局制定《农村幼儿园民办公助办法（草案）》（以下简称《办法》），要求继续在做好师资培养，提供教材，充实公社幼教干部，发动“公带民、老带新”等工作的基础上，对一些经济困难的幼儿园给予一定的补助，并对办得好的幼儿园给予一定的奖励，以促进农村幼儿园的繁荣和发展。如补助新建幼儿园3元—5元，补助困难幼儿园每班每季度不超过6元，每季度被评为先进的幼儿园奖励不超过10元。此外，还规定补助及奖励应作充实教育设备之用，如教养员用的参考资料、图书、、教具，或简单的卫生医药箱、毛巾、脸盆、肥皂等。78该《办法》的出台是自1958年以来，上海市财政第一次大规模地对农村人民公社开办的园所给予资金补助。在当时财政极为有限的情况下，这项补助极为可贵，充分表露了政府对托幼工作的期待，在一定程度上也有助于解决实际的资金困难。随着《办法》的逐步落实，上海的农村托幼事业迎来了1960年的大发展。

    1960年4月，为了掀起集体托幼的新高潮，上海市委周密部署，各部门高度重视并做了大量工作。9日，时任上海市委第一书记的柯庆施提出上海必须“逐步分批实现公社化”79；中旬，上海市卫生局与教育局向上海农村派驻“幼托事业工作队”80，宣传托幼的意义，培训保育人员，组织示范教学，制定规章制度等81。该工作队抓住薄弱环节，帮助兴办托幼，同时也起到了“督促”公社及大队干部的作用。如奉贤县肖塘公社原来只有7所幼儿园，在工作队的帮助下办起了82所幼儿园，托儿所也从97所增加到172所。82当月，上海成立市幼儿保健教育委员会，旨在加强对托幼工作的统一领导。83不久，该委员会向上海市委提交了《关于幼儿园、托儿所发展情况和今后打算的报告》，在肯定前期依靠“穷办法、土办法”兴办托幼事业做法的基础上提出，今后必须坚持“边发展、边整顿、边巩固、边提高”的原则，才能促进托幼事业“多、快、好、省”地发展。此外，它还对农村托幼作出系统性规划：“县、公社、生产队各级的幼托工作组织，层层要有专人负责，以加强对这项工作的领导”，要求农村入园入托儿童国庆前达到85%以上。为了完成指标以“出色的成绩迎接国庆”，须掀起三个高潮：结合“六一”评比表扬先进儿童工作者和先进儿童工作集体，造声势、树标兵；7月，对幼儿园、托儿所，组织一次夏令卫生工作大检查，以提高卫生保健水平；9月，再组织一次托幼工作的全面性对口检查。84这些部署不可谓不细致周到。

    为了完成上述指标，上海农村各县首先指定专人负责托幼工作的领导。金山、青浦、松江等县设立生活福利委员会85，嘉定、上海、川沙、浦东、奉贤等县成立了生活福利办公室。通过“六一”评选工作，南汇、崇明两县又在生活福利办公室下设托幼小组或托幼办公室领导托幼工作，宝山县则专门建立了托幼委员会。这些职能部门或小组的建立，使得托幼工作经常被提到议事日程并做统一布署。基层公社也专门配备了负责托幼工作的干部。如泥城公社的党委书记就对托幼工作做到“五抓”（一抓干部群众的思想教育；二抓规划，做到心中有数；三抓统一安排，安排生产的同时安排托幼工；四抓具体问题，如粮食、房屋、设备、工分等；五抓专线领导，层层有人领导，副书记挂帅）。86其次，掀起检查评比、树立标兵的浪潮。公社频繁组织各类检查评比活动，在检查评比后，将检查的情况与各生产队发展托幼的进度表，分发至各生产队，激起“落后”生产队的赶超想法。如金山县紧紧抓住兴塔公社红旗幼儿园这一标兵，开现场会交流经验，全县掀起了“学兴塔、赶兴塔、超兴塔”的高潮。87此外，在浦东、上海、川沙、宝山、松江、崇明、嘉定等县的25个公社和7个镇的妇联联合倡议“积极发展幼儿园，做到凡是无人照管的儿童全部入园入所”之下，大办托幼的友谊竞赛在这些区域不断涌现。88

    在上下贯通的组织领导下，至1960年6月，上海农村掀起了一股大办托儿所、幼儿园的热潮。据《文汇报》称，仅嘉定一个县，一个多月内入园入托幼儿就增加了18000多名，当月，全县入托入园幼儿占学前儿童人数的70%以上。其中，势头较好的如城西、封浜等公社这一比例更是达到90%。89又据上海市妇联农村工作部1960年4月统计，农村入托入园人数上升到 40 万。90另据统计，1960年6月中旬，上海农村共有托儿所20201所，收托幼儿251338人，共有幼儿园8354所，收托幼儿26216人，托儿所的收托比例提高到80%，幼儿园的收托比例提高到72.6%。91在以指标为先的“跃进”氛围中，这些数据可能有浮夸成分，但也从侧面呈现出农民在革命口号下奔向共产主义的历史情境。

    “大跃进”期间，高指标、浮夸风并非农村集体托幼所独有，但这项工作似乎又有其独特价值，当时的革命口号精炼地体现了这种价值—— “一夜托儿化”，“实行寄宿制，消灭三大差别”92。从某种意义上讲，彼时的农村集体托幼在这种夸张的“革命”氛围中，已然浮现毛泽东所期待的“六亿神州尽舜尧”93的美好图景。

    实际上，“大跃进”期间，上海农村的园所大多是匆匆上马，存在诸多问题，比如保育员的卫生知识、业务能力、托儿所设备、环境卫生、管理水平等跟不上，加之其他条件限制，前面描绘的集体托幼成效恐怕与真实情况存在不小落差。当时就有人质疑1958年的集体托幼，比如，“囡多占便宜，我们负担领囡费，做来做去担几个共囡，啥叫按劳取酬”94，引起了有子女入托家庭和无子女入托家庭之间的矛盾95；有的因托儿所“路远不便”，认为没必要再办96；有的在孩子入托后不久就要接回家，态度还十分强硬97。出现这些现象的原因或许是托幼工作不够扎实而使农民心生怨言，但也从侧面说明上海农村此时仍未实现以集体托幼形式塑造共产主义新农民的目标。

    尽管“大跃进”时期的集体托幼成果存在一定的虚报浮夸，98但上海农村的集体托幼实践还是取得了一定成效。以上海农村某生产大队为例，自1958年人民公社建立的三年里，该大队由原来1个农忙托儿所发展为3个常年托儿所。幼儿在托儿所里生活得很健康，也未发生过重大事故，家长说“有了托儿所孩子高兴，妈妈也可安心做生活”99。通过兴办集体托幼，上海农村妇女不仅摆脱了孩子的拖累，全心搞生产，还有时间学习文化并脱盲。其中，有的人当上了保育员、教养员和妇女干部；100有的人以往毫无卫生知识，现在能当保育员；有的人过去一字不识，如今能当教养员；有的人以往从不关心政治，现在当起了妇女干部。这不仅大大解放了妇女，还提高了人民公社托幼事业的工作质量和业务水平。101仅1958年下半年，上海市妇联与卫生、教育等部门联合训练公社托幼工作干部就达 250 余人，102这些人能够在人民公社中胜任各自的工作，客观上展现了农村集体托幼实践的效果。

    1961年，中央提出“调整、巩固、充实、提高”的方针之后，人民公社迎来了大调整，生产工作成为农村的中心任务，托幼工程自然退居“次要”。此后，农村集体托幼进入常态化阶段。随着工作重心的转移，1962年8月，上海市幼儿保健教育委员会撤销103，农村托幼工作恢复到由上海市妇联农村工作部主管，该部将办园办所的决定权下放给社员群众，托幼机构“办不办，怎么办，办什么样子的”成为群众自己决定的一项事务。104不难发现，上海农村托幼实践进行到此时，已经因为中心工作的变化而承载起新的政治话语。“大跃进”时期培育“共产主义接班人”的宏大历史进程也进入尾声。

    四、融入托幼日常的具体措施：培育“共产主义接班人”

    1959年六一儿童节，《人民日报》发表社论，号召将托儿所和幼儿园办成“培养共产主义事业接班人的基地”105。上海在推动农村集体托幼实践落地生根的过程中，既要有切合受教育对象特点的动员技术，又要根据“因陋就简”的现实条件做出周密安排，在此基础上，再为具体的国家任务服务。上海实施若干措施，将共产主义从方方面面点滴渗入农民的日常生活，对“接班人”的共产主义塑造愈发深入。

    （一）注重卫生保健

    在对农民进行集体托幼的有效动员之后，还需培训具有一定业务能力的保育员，这是农村集体托幼能够顺利开展的必备条件。1958年年底以前，由于多种原因，园所的保育员难堪重任，尤其缺乏卫生保健知识。对于实质上仍是农民，业余充当保育员的群体，需要经常性地开展培训工作。1959年春，上海农村开始建构由文教、妇联及卫生部门三者相互协作、配合的培训体系，在编写保健知识丛书的基础上，由各级医务系统如县医院、护士学校、妇幼保健所及公社医院主导，分层分批对保育员进行脱产或不脱产、定期或不定期、长期或短期的卫生保健知识培训，106如培训保育员必须学会和做好除“七害”107、讲卫生，晨间检查，预防传染病，孩子有病会隔离和报告，培养孩子的卫生习惯，孩子的饮食和卫生，安排孩子的生活，各种消毒工作，保护孩子的安全，教养孩子等十件事，使其逐步达到初级保育护士水平。108

    事实上，要做好幼儿疾病预防及应急处理，仅凭培训保育员这一项措施是远远不够的。因此，在“预防为主、防治结合”的要求下，公社还依托地段与区域的专业医务力量，搭建幼儿疾病预防与治疗平台，通过预防接种，建立上下贯通、层层负责的卫生保健网，并以此作为示范向全县推广。如川沙县蔡路公社幼儿园与卫生院取得联系，通过卫生院妇幼科医生或保健员经常来园做卫生保健的业务指导，建立定期检查制度，并按时进行预防接种。109自1959年卫生保健网建立后，对保教人员及专业医务人员执行卫生保健措施起到了监督、引导与帮助的作用。同时，该卫生保健网在预防幼儿常见传染病上也产生了良好效果110，一旦出现病孩也能及时治疗111，幼儿因病致死的情况很少发生。1960年，园所的麻疹发病率较1958年同期成倍下降，很多园所甚至一年来都没有发生过麻疹。112

    此外，上海市妇联农村工作部于1959年3月制订了《关于农村托儿所、幼儿园工作暂行条例（草稿）》，其中，不少详尽的规定为幼儿的卫生保健提供了制度保障。比如，在园所选址方面，必须选择兼顾幼儿安全和方便家长接送的地方，应注意平坦宽敞、清洁卫生、空气流通、阳光充足，水塘、河沟、畜圈、马路及医院等旁边不宜设园所；在食具方面，幼儿的开水壶、碗筷须自备一套，并设自来水冲洗脸、手，避免传染疾病；在预防保健方面，定期为幼儿预防接种和健康检查，有病须立即隔离；在生活制度方面，吃饭、睡眠、游戏、洗脸要有秩序。每日晨间检查，食具、玩具每周须消毒1至2次，饭前便后要洗手。113据上海市妇联农村工作部检查，有1/4的园所落实较好114，这些规定促使幼儿养成了清洁卫生的习惯。115一些模范幼儿园如宝山县吴淞公社卫星幼儿园的日常保健常态化，每日写生活记录，随时掌握幼儿的吃睡、大小便情况，还设有隔离室。116以上举措可谓具体而微。

    应该说，通过卫生保健的宣传、人员培训、体系建构及制度落实等各个环节，农民对于以“新农民”之姿投入托幼卫生保健事业的认知还是有所提高，学习保健知识及养成卫生习惯的积极性也随之被激发出来。农民与集体、国家之间的联结被强化的同时，对共产主义的认同也会显著加深。

    （二） 优先供应饮食

    在构建卫生保健体系的同时，还需优先供应幼儿的饮食。“身体是革命的本钱”，幼儿茁壮成长、体魄强健是推行农村集体托幼的必然要求。正是出于这种考虑，基层干部对幼儿的膳食给予了极大关照，在饮食供应方面，托幼组织在同期的农村福利组织中处于被优先供给的地位。

    上海市妇联农村工作部要求幼儿饮食必须由专人负责，单独做适合幼儿年龄的饭菜，并按50∶1的比例配备炊事员。117幼儿的一日三餐要与成人饮食有所区别，且专门烧各式小菜，以换口味。就连幼儿平日喝的开水，洗脸洗脚用的热水，也都由食堂供应，以减少疾病的发生。118并且，根据幼儿年龄定粮，实行“专人管理、计划用粮”。如奉贤县三官公社胡村生产队的全托幼儿园每人每月定粮，大的孩子（虚岁7岁—8岁）每天12两左右，小的孩子（4岁—6岁）9两左右。119“吃得饱”这一要求基本能得到保障，甚至还略有积余。

    在“吃得饱”的基础上，还讲究营养均衡、荤素搭配，即“吃得好”。一些生产队特意为全托幼儿园配置了一定面积的自留地，供保教人员耕种以提高幼儿的伙食水平。如南汇县大团公社沙庙生产队的全托幼儿园自种高粱、黄豆、玉米、芋艿、卷心菜、红萝卜等，做到了生产队不再贴粮食，蔬菜也可基本自给。120这些粮蔬专供幼儿园使用，不用上交集体。为了保证蛋白质的供应，人民公社还通过各种手段尽量予以满足。不少时候，由生产大队专为幼儿购买含有蛋白质成分较高、易于消化吸收的鱼、蛋、肉类等食物。121园所自养的家禽家畜也可稍作补充，改善幼儿伙食。如宝山县新生生产队幼儿园养了鸡、鸭、猪、羊和兔子，每逢节日都可以吃到荤菜122，有些幼儿园甚至隔一天就能吃到荤菜123。此外，不少公社的供销部专为幼儿提供一定量的糕点、点心、糖果、饼干及线粉124，在夏季，还为幼儿供给特定的食物或饮料以防暑降温125。至于经费，一般由大队公益金担负。126公益金不足时主要靠农民自己解决，保教人员、食堂等工作人员也主要来源于农民，这使得农村集体托幼看似颇有群众“自主办园”的味道。但很显然，这场实践的主导力量、决定性因素还是党和政府。

    由于公社对托幼饮食及营养方面的重视，园所的幼儿一般比散养在家的幼儿待遇更好，有生产队队长称“小囡粮食够吃，荤菜、红枣、饼干样样优先供应，比家里养得壮多了”127。在缺粮少食的困难时期，这些共产主义“幼苗”无疑享受着优先待遇。一个显著的事实就是，送进全托的孩子除本人口粮外，生产队给每个孩子每月平均补贴2斤左右的粮食，散养幼儿则没有。也难怪一些没有子女入托的社员自认为“吃亏”。128应该指出的是，对托幼“优先供应饮食”是符合实际的合理安排。广大农民在回报很少的情况下兢兢业业投入集体托幼实践，其历史作用不应被低估。即便有不足，主要也源于难以逾越的历史条件限制，不应过于苛责。

    （三）强调全面教育

    集体托幼实践的基本诉求在于“教养结合”以培育共产主义接班人。所谓教养结合，即仅关注幼儿的卫生保健、饮食营养是远远不够的，还必须加入“教”的因素。上海市教育局明确要求从幼儿园的性质和“两大任务”129出发，强调幼儿园既是福利机构又是教育机构，特别批评了对“教育机构”性质认识不足的问题。130为此，在幼教人员的配备上，上海市妇联农村工作部要求个人成分、身体状况、工作态度等基本条件满足外，还特别强调“最好是具有一定的文化水平的青壮年”担任。131总的来看，这一时期农村托幼在培养德育和智育方面的“接班人”这一问题上做出了许多尝试性探索。

    为了培育“德才兼备”的共产主义接班人，园所十分注重幼儿的德育。在德育方面，主要包括集体主义教育、人民公社的认同教育及热爱国家与领袖方面的教育。中国福利会幼儿园不仅编写出版书籍，还利用游戏及表演让幼儿明白集体主义的好处与意义，如“集体劳动生产出来的农作物，比个人劳动生产出来的农作物既多又好”，同时将集体主义意识落实在幼儿日常行为中，要求各班互助浇水施肥，“共享”劳动成果，提醒孩子“长出来的菜是大家的，不分你的我的”。132此外，教养员时常带领幼儿观察公社出现的各种新兴事业和群众福利事业，如工厂、食堂、敬老院、俱乐部等，使孩子们懂得“有了共产党和毛主席，成立了人民公社，人们的生活将越过越好”133，教幼儿学唱《打夯歌》《人民公社真正好 》《国旗歌 》《爱毛主席 》等儿歌134，要求中班孩子学会写“毛主席”三个字135。

    在智育方面，要求逐步培养幼儿的感官、语言、思维、动手等诸多能力。1960年前后，上海市教育局出版了大量幼儿教材与教辅资料，要求幼儿学习拼音、认识汉字及学会计算，并针对幼儿年龄，提出了不同层次的要求。如小班学会1—5的数的概念，中班学会10以内的加减法，大班学会20以内不进位、不退位的加减法，学会口编应用题等。136其中，南汇县惠南幼儿园在计算教学方面被列为先进，该幼儿园在教学思想上实现了智育与德育两方面的紧密结合，联系政治形势，生活实际以及生产实际进行计算教学。例如，给幼儿讲“人民公社好”时，把平时使用的教具和增添的新教具结合起来反映农村全面发展，表现农民生活水平的提高，并采取多样化、针对性的教学形式，用蜡光纸剪成钢铁元帅、棉花姑娘、麦公公之类等进行计算教学，用实物和玩具、卡片进行计算，用计算木架数数、计算等。137凡此种种或许可以说明，上海农村集体托幼实践在贯彻落实各项国家任务方面发挥了积极作用。

    无论是幼儿，还是从事托幼事业的工作人员在承接以上三项具体措施的过程中，感受到的不再是宏大高远的共产主义意象，而是融入日常的共产主义。由此，农村集体托幼实践对农民的共产主义塑造润物无声地走向具体深入。

    五、结语

    从全国范围看，人民公社化时期的农村集体托幼实践成绩斐然，帮助大量农民，尤其妇女，有更多精力投入农业生产。但这场集体托幼实践并非只为实现家务劳动社会化以解放妇女，更有着力塑造共产主义新农民之意图。通过考察1958年至1962年的上海集体托幼实践可以发现，由于这场实践始终跟随党和国家各个时期的中心工作，而且实行了注重卫生保健等具体措施，其对农民的共产主义塑造体现为一系列的革命性实践。

    人民公社化初期，无论是推行“儿童集体化”还是倡导“我为人人、人人为我”，都将农村集体托幼与塑造农民对共产主义精神的政治认同结合起来。人民公社化运动在凸显共产主义这一严肃议题的同时，结构性地激发出农民集体托幼的内在需求，集体托幼实践由此真正成为一项“运动”，集体托幼成为农民接轨共产主义的重要工具。“大跃进”及其后的常态化集体托幼促使农民在革命口号中奔向共产主义。而培训保育员，构建卫生保健网，优先供应幼儿饮食，以及强调全面教育，则是推行农村集体托幼的具体措施，共产主义由此融入农民的日常。这些革命性实践为农民诠释出集体主义、人民公社、爱国、爱领袖等诸多意涵，促使他们形成对共产主义的政治认同。农村集体托幼实践塑造共产主义新农民的丰富历史图景也由此展开。

    最后，似有必要对农村集体托幼实践塑造出的共产主义新农民的内涵作进一步讨论。前已述及，新农民不仅包括有孩子的农民、无孩子的农民，也包括农村基层干部、园所的保教人员，还包括园所的幼儿。对这些看似不同群体的农民而言，集体托幼都有一个从最初的诸多顾虑发展成为其内在需求的过程，且这种需求在“大跃进”的情境中达到最高潮。站在国家的角度来看，无论有孩子的农民还是无孩子的农民，都是为了更好地担负起他们作为”国家的农民”的责任，农村的基层干部与园所的保教人员为的是更好地完成“国家的托幼工作”，园所的幼儿则是为了更好地成为“国家的接班人”。因此，虽然陶行知和国民政府的乡村幼稚园实践与人民公社化初期中国共产党主导的农村集体托幼实践都担负着“为国”的重任，但只有后者才能有效完成这一任务。

    本文转自《开放时代》2024年第6期

  • 韩昇：武则天时代的官僚阶层与科举

    本文载于《学术月刊》2024年第11期

    武则天登基以来，内部大狱频兴，朝政空转；外部烽火四起，挫折连连。国势日蹙，完全无法同唐太宗“贞观之治”同日而语，和唐高宗在位时期相比也颇为不如。从人事的角度观察，没有治国统兵的人才是一个重要的原因。要真实反映武则天的用人状况，必须进行全面考察，不可以偏概全。这里从两条线、三个层面切入，观其全貌。

    所谓的两条线，第一条线是理应掌管国政的朝官，第二条线是武则天真正委以重任的近幸宠臣。第二条线还可以细分为武家子弟、宠幸嬖臣；以及酷吏等两个层面。结合第一条线，构成了任用官吏的三个层面。

    一、朝官

    第一条线。朝廷最高执政者为一般所称的宰相，亦即唐朝的“同中书门下三品”，或称“同中书门下平章事”，武则天改“中书”和“门下”为“凤阁”“鸾台”，故中书门下称作“凤阁鸾台”。从嗣圣元年（684）武则天废中宗、垂帘听政以来，直到她被政变推翻的神龙元年（705）的二十一年间，任用了五十三位“凤阁鸾台三品（平章事）”头衔的宰相：

    刘祎之，武承嗣，魏玄同，苏良嗣，韦思谦，韦待价，张光辅，王本立，范履冰，邢文伟，周允元，岑长倩，裴居道，傅游艺，格辅元，乐思晦，崔神基，狄仁杰，杨执柔，崔元琮，李昭德，姚璹，李元素，韦巨源，陆元方，苏味道，王孝杰，杨再思，杜景俭，王方庆，李道广，娄师德，武三思，武攸宁，姚元崇，李峤，魏元忠，吉顼，王及善，豆卢钦望，张锡，韦安石，李怀远，顾琮，李廻秀，唐休璟，韦承庆，朱敬则，韦嗣立，宗楚客，崔玄𬀩，张柬之，苏瓌。

    鸾台（门下省）和凤阁（中书省）的首长亦是宰相。这两个机构掌管皇帝诏敕和军国政令，在皇城内办公，最能接近大内里面的武则天，宛如皇帝左右的鸾凤。

    鸾台纳言：王德真，苏良嗣，韦思谦，裴居道，魏玄同，武承嗣，武攸宁，史务滋，欧阳通，姚璹，娄师德，狄仁杰，李峤，韦安石。

    凤阁内史：裴居道，岑长倩，张光辅，邢文伟，宗秦客，豆卢钦望，王及善，武三思，狄仁杰，李峤，杨再思。

    鸾台凤阁的首长人数亦多，经常变更，受酷吏政治迫害者约三分之一上下。

    宰相是朝廷最高首长，中流砥柱，安危所系，本应最为稳定。唐朝建立以后，高祖任用的裴寂，太宗任用房玄龄、杜如晦等宰相，任职时间都很长，对于安定社稷、保持政策的连续性起到重要的作用。然而，到了武则天时代，这种局面骤然巨变，宰相更迭极为频繁，没有任何一个朝廷部门堪与相比。而且，这个席不暇暖的群体，即使把政争中遭到贬黜的情况排除在外，也至少有四分之一以上受到酷吏的迫害乃至屠戮。宰相被呼来唤去，弃之如同敝屣，则所有官吏的处境可想而知，武则天时代政情的高度不稳和内斗的极端残酷，实态毕露。

    为什么宰相群体更替最频繁呢？因为武则天对他们把控最严也最直接。武则天身处大内，既无政绩也无功绩，无以服人；同时自然没有共创事业的部属，堪以寄任。而且，她作为高宗内眷的妇女身份，不方便经常和朝廷大臣聚乐宴饮，了解外请，增进感情。所以，她只能紧紧控制权力中枢的宰相群体，通过他们掌控全局。宰相作为政令下达、沟通内外最重要的渠道，必须盯紧看牢。由于朝廷乃至社会民情、官吏所思所想，只能通过文件和密报等间接途径获取，又要应对篡唐立周的改朝换代剧变，随时有被颠覆的危险，这些都会极大加剧生性多疑的武则天的猜忌。所以，她采取频繁更替宰相乃至施以毒手的苛酷手段，势所必然。这里是她控制全局的关键，亦是命门所在。

    安危系于此地，宰相的治国能力并不重要，竭尽忠诚才是关键。所以，武则天时代的宰相群体有两个特点：第一，实务政绩型官员很少，大多出自政务官员。第二，武氏子弟实际掌权。武氏子弟是武则天政权最稳定的人事，无论他们是否身处宰相位置，宰相都要听命于他们。推而广之，武则天宠幸的汉子，宰相也要接受其统辖。例如征伐契丹时，武三思为主帅，宰相姚璹为副；征伐突厥时，薛怀义为主帅，宰相李昭德为副等等，乃至造天枢、进颂词之类事务，也是武氏子弟统领宰相实施。

    与此形成鲜明对照的是中央最高行政部门的尚书省，其首长在武则天时代最为稳定，二十一年间仅有六位，分别是左仆射：苏良嗣，武承嗣，王及善；右仆射：韦待价，岑长倩，豆卢钦望。除了岑长倩一人被迫害致死外，其余五人基本平安。左右仆射为尚书省主官，武则天时代曾经改称“左相”“右相”，实际上有名无实，和宰相沾不上边。武承嗣作为武氏子弟，不管担任什么职务都大权在握。尚书省长官之所以相对稳定且平安，根本原因就是没有实权。武则天通过宰相直接指挥六部、九卿，作为六部上级主管部门的尚书省形同摆设，主官在位唯唯诺诺，乏善可陈，故各人本传事迹记载寥寥，滥竽亦可充数。韦待价以军功起家，武则天用他担任天官（吏部）尚书、文昌右相，“素无藻鉴之才，自武职而起，居选部，既铨综无叙，甚为当时所嗤”。韦待价自知非治国之才，“既累登非据，颇不自安，频上表辞职，则天每降优制不许之”。武则天为什么坚持把不懂行政的人放在行政主管的位置上呢？其实就是为了将其虚化为承旨画押的华丽道具，便于她直接掌控朝廷。

    尚书省上层的人事，如表1所示（分为武则天垂帘听政与称帝两个时段）：

    表1

    尚书省长官被弱化乃至虚化，但尚书省的职能不可完全废弃，因而出现上权下移的情况，亦即尚书省的左丞（正四品上）和右丞（正四品下）实际处理都省事务。尚书省主官左、右相（从二品）位高权虚，人少稳定，但左、右丞官员颇多，频有更迭，表明他们才是真正主事者。以下级官员主持事务，是独裁者常用的集权手段。由于官位卑下，受到超常重用时感恩戴德，听话卖力。而且，还因为官位卑下，制度上无权参预重大国务，所以让他们参预何种事务，以及参预到什么程度等等，君主皆可随心所欲，权力收放自如。武则天以此手段控制尚书省。

    外朝机构主要是六部，其人事任用情况如表2：

    表2

    人事变更的频度，依次为吏部、兵部、刑部、礼部、户部、工部。用这个指标观察武则天时代各个官署的情况，可以发现它们在朝廷权力结构上的重要性同人事变更频度成正比，越是重要，掌控越严，人事更迭越发频繁。由此归纳出武则天朝的权力秩序及其结构如图1。

    图1

    这明显是一个以军政为中心的朝廷：一切以皇帝集权独裁为最高目标，由吏部担纲彻底更换官吏队伍，兵部作为权力支柱，刑部作为整肃工具，礼部制造改朝换代的理论与合法性。皇权笼罩于全社会，生产、技术、民生等皆处于从属地位。武则天彻底改变了唐太宗建立的社会发展国策与朝廷架构。

    朝廷中最受重视的吏部和兵部，副职的变动异常的频繁，还多次出现其他部难以见到的官员再任的情况。这明显是武则天直接插手，安插委任亲信；同时表明在至关重要的权力部门，武则天倚重副职，越级操控部务，使之完全听命于皇帝。

    朝官这条线高级官员的选任情况，根据《旧唐书》和《新唐书》的不完全记载，列示如表3：

    表3

    两《唐书》官员传记固然不能覆盖官员全体，然而，达到一定的量亦足以反映用人原则和基本面貌。根据上表所示，至少可以确认以下两点：

    第一，官员大都出自官宦之家。

    唐朝建立后，功臣和高官后裔，特别是军功子弟在仕宦上获得优待。唐高宗仪凤年间，魏元忠上封事指出：

    当今朝廷用人，类取将门子弟，亦有死事之家而蒙抽擢者。

    北魏孝文帝迁都洛阳推行新官制，按照官职高低分为甲乙丙丁“四姓”等级，确立了优先录用官宦子弟的制度规定，北齐、北周、隋朝和唐朝都沿袭这一原则，武则天亦是如此，故官宦出身者出仕比例甚高，功臣子弟更受重用。太宗、高宗朝名将薛仁贵，儿子薛讷，“则天以讷将门，使摄左武威卫将军、安东道经略”。

    从唐朝建立到武则天全面掌权，经过了大约半个世纪，许多功臣业已凋零，功勋门第逐渐变味为官宦之家。官宦出身既是政治可靠的凭证，在入仕升迁上受到重视，也是官场的护身符，在仕途挫折罹难时，能够起到从轻处罚或者事过境迁后东山再起的佑庇作用。开国初期的重视功勋，逐渐蜕变为建政后常规铨选时讲究家世亲缘，武则天对此颇为坚持，有所发挥。岑文本是唐太宗任用的宰辅重臣，其侄子岑长倩因此得到重用，高宗时出任宰相，支持武则天夺权，故长年身居权力中枢，直到武则天欲立武承嗣为皇位继承人之际，因为主张维持亲子继承而得罪武则天，下狱处死。此后在朝廷举荐人才的时候，凤阁侍郎韦嗣立推荐岑长倩族子岑羲入朝任职，并说明其为朝廷罪犯亲属，武则天不但批准了岑羲的任用，而且还为受牵连的高管亲属的任用开了绿灯，“由是缘坐近亲，相次入省”。对落难或者受牵连的官宦子弟网开一面予以任用，显然不是个例，而成为规则，维护着优待官宦子弟入仕的一贯方针。

    第二，注重名门家世，尤其是亲缘关系。武氏子弟不循正常途径入仕，应置于第二条人事线论述。武则天的母亲自称出自天下名门之弘农杨氏，实为隋朝皇族之杨氏。隋室杨氏因为是武则天外家的缘故，一直受到重用。武周“时武承嗣、攸宁相次知政事”，武则天对地官尚书杨执柔说：“‘我令当宗及外家，常一人为宰相。’由是执柔同中书门下三品。”武氏和杨氏联合坐庄朝政，成为一条规则。

    李唐与隋杨乃姻亲，政治上虽为敌手，亲情却深。李渊的母亲和隋文帝独孤皇后为亲姊妹，隋朝灭亡后，李渊对隋杨皇族给予照顾，亲自做媒将隋朝纳言杨达女儿嫁给武士彟，让这位河东木材商人粘上皇亲国戚的边，成为武则天日后飞黄腾达不可或缺的门槛。武则天显然领悟到王朝政治的奥秘，深知金字塔权力结构的顶端是少数门阀士族垄断权力，运用朝廷强力部门作为工具，实现对整个官僚体系的控制。在她的理解中，管理社会的核心不是遵守规则，而是追求权力的无限扩大，笼罩一切。权力需要人来掌握，掌权的人越少则权力集中，越有利于皇权。因此，等级森严的寡头政治成为她的营造蓝图。武氏家族（包括赐予“武”姓的皇子）居于金字塔尖，被选中的士族与近宠佞幸组成朝廷上层。有所不同的是被选中的士族相对稳定，而近宠佞幸与酷吏则频频变动，道理在于这些人作为工具固然必不可少，但落到具体的走狗却需要经常更换。近宠佞幸与酷吏属于第二条人事线研究的对象，留待后述。被选中的少数门阀士族颇受重用，飞黄腾达，纷纷跻身于权力上层，与武氏家族共同构成核心统治集团。例如杨氏家族在“则天时，又以外戚崇宠。一家之内，驸马三人，王妃五人，赠皇后一人，三品以上官二十余人，遂为盛族”；韦氏家族之“巨源与安石及则天时文昌右相待价，并是五服之亲，自余近属至大官者数十人”。

    唐朝是贵族建立的王朝，高祖李渊以此为荣，建政初期曾经对宰臣裴寂说道：

    我李氏昔在陇西，富有龟玉，降及祖祢，姻娅帝王，及举义兵，四海云集，才涉数月，升为天子。至如前代皇王，多起微贱，劬劳行阵，下不聊生。公复世胄名家，历职清要，岂若萧何、曹参起自刀笔吏也。惟我与公，千载之后，无愧前修矣。

    得意之情溢于言表。倚重士族和功勋家族成为唐朝人事的重要原则，唐太宗修《氏族志》和武则天重用士族皆为此原则的一脉相传，除了武氏因武则天而破格崛起之外，老牌士族左右高层政治的局面一仍其旧，未有改观。武则天重用士族，寄任之深甚至扭曲制度。唐朝制度规定，近亲不得同时担任高官要职，以防止某一家族权力过大。对此项规定，武则天采取变通的办法规避，李峤担任宰相，两年后其舅张锡也升任宰相，武则天让李峤转任成均祭酒，“舅甥相继在相位，时人荣之”。士族对于权位的诉求也直言不讳。垂拱年间的宰臣韦思谦把两个儿子韦嗣立和韦承庆径直托付给武则天，说：“臣有两男忠孝，堪事陛下。”武则天欣然接受，对韦嗣立明言：“今授卿凤阁舍人，令卿兄弟自相替代。”果如其言，先是韦嗣立接替韦承庆担任凤阁舍人，然后由韦承庆轮替韦嗣立出任天官侍郎，不久又接下韦嗣立的宰辅要职，等到韦承庆去世，又让韦嗣立接任黄门侍郎，前后四度轮替，宛如左右手传接一般。

    优待功臣后人，讲究官宦家世，倚重名门士族这三条铨选的基本原则，武则天无不坚持贯彻，比起唐太宗时代逐步开放用人的家世条件，有所倒退。陈寅恪未对武则天用人的实际情况进行整体考察，断言武则天破格用人，培养出新兴阶级攘夺替代西魏、北周、杨隋及唐初将相旧家之政权尊位，“故武周之代李唐，不仅为政治之变迁，实亦社会之革命。”此番议论完全得不到事实的支持。武则天称帝充其量只是僭主篡政，酷吏政治绝非社会革命，新兴阶级亦非权力所能制造，只能是社会生产形态所决定的客观存在。

    一朝有一朝的组织原则。武则天朝对于太宗朝组织原则的最大改变，是把对唐朝的忠诚演变为对她个人的效忠。她遴选并重用的官宦士族都遵循这条最高原则。

    从小生活在权贵圈子里成长的功臣高官子弟，对于政治人事嗅觉最为敏感，察言观色得风向之先，其中想飞黄腾达的人跟风最紧。丘和、丘行恭父子建唐时立有大功，皆获陪葬皇陵的殊荣。丘行恭之子丘神勣jì属于最早投靠武则天的功臣子弟，充当鹰犬，出手害死章怀太子，与酷吏周兴、来俊臣齐名；岑文本侄子岑长倩等一批功臣子弟因为支持武则天取代李唐而得到重用，俱见前述。李大亮的族孙李廻秀，在武则天晚年当上宰相，“颇托附权幸，倾心以事张易之、昌宗兄弟，由是深为谠正之士所讥”。

    在逢迎武则天近幸方面，王朝体制内的士族亦不遑多让。崔义玄精通儒经，以学干禄，为唐高宗立武则天为皇后出谋卖力，主持审判长孙无忌。因为这份功劳，两个儿子崔神基和崔神庆都得到武则天的重用。武则天晚年，朝中大臣拼死控告武则天的男宠张昌宗犯罪，崔神庆受命审理此案，竟然为其开脱。张昌宗、张易之兄弟为武则天晚年之最爱，士族官员趋势附炎，卑躬攀附，父子三人皆为宰相的韦氏，韦承庆讨好张氏兄弟；几度进谏武则天的宰相李峤，其实和张氏兄弟交情甚深，以至于武则天倒台后，他们都为此遭到贬黜。宰相杨再思历仕三朝，主持政务十余年，地道的官油子。他善于体察上意，皇上喜欢的，他吹捧得天花乱坠，皇上讨厌的，他诋毁得丑陋无比。有人私下问他身居高位何苦如此呢？他道出为官数十载的心得：正直的官员招灾惹祸，唯有望风顺旨才能保全性命。原来赞美颂圣的合唱队充斥着虚情矫饰的歌手，声嘶力竭的领唱者往往最洞悉内里幽暗。杨再思年轻就通过科举，腹有经纶，黠于应对。张昌宗遭诉，群情汹汹。武则天询问杨再思意见，杨再思说张昌宗炼仙丹给皇上服用，皇上身强体健便是国家万幸，所以张昌宗功劳莫大。避开犯罪事实，只谈皇上重于社稷，利君则利国，情郎瞬间成为英雄，迎合了武则天万难割舍的感情。张易之兄弟大宴朝官，饮酒互捧，张昌宗容貌粉嫩而得武则天欢心，一众官员赞美张昌宗貌似莲花，杨再思挺身纠正道：“人言六郎面似莲花；再思以为莲花似六郎，非六郎似莲花也。”这等话术浸染弥漫成为武周王朝的官风。

    各路出身的王朝官员汇聚在一起，国家正事做不了，真话说不得，有失品格的种种表演，未必都是他们猥琐卑劣，而是那个时代的政治生态所致。当然，他们的所作所为反过来也强化了那种环境，互为因果，最终无人幸免。于是，官场晋升的秘径变成通途，“时朝廷谀佞者多获进用，故幸恩者，事无大小，但近谄谀，皆获进见”。

    拍马溜须而不做事，即使身居高位也不敢有所作为。在朝不为恶，偶尔说些合乎道理的建言，这在正常的社会属于常识底线，但在武周却足以振聋发聩，勇气有加，难能可贵。武周时代朝官的水平，后人颇有评论：

    豆卢钦望、张光辅、史务滋、崔元综、周允元等，或有片言，非无小善，登于大用，可谓具臣。

    苏味道、李峤等，俱为辅相，各处穹崇。观其章疏之能，非无奥赡；验以弼谐之道，罔有贞纯。

    崔融、卢藏用、徐彦伯等，文学之功，不让苏、李，止有守常之道，而无应变之机。

    崔（融）与卢（藏用）、徐（彦伯），皆攻翰墨。文虽堪尚，义无可则。备位守常，斯言罔忒。

    这些评价并非贬低之辞。武则天晚年请狄仁杰举荐宰辅高官，狄仁杰当面询问武则天是否觉得当朝主官乃“文吏”之流，不堪大任？武则天深以为然。一朝皆凡庸，是谁之过？然而，到此地步，不想崩溃只能举贤任能，转机因此萌生，历史总要做出选择。

    二、近幸宠臣

    武周政权的朝官，在大清洗的肃杀氛围中，实际上已经沦为摆设，把朝廷门面装潢得煞有介事，敷衍日常事务，跑腿当差。真正掌握权力的是第二条线，亦即武则天委以重任的近幸宠臣。同第一条朝官线的重要区别，在于他们基本不经过吏部铨选正途入仕。这条线上的人物可以分为两个层面，首先是中心层面，有处于权力中枢、出将入相的武氏子弟，以及武则天信赖有加的男宠团队。其次是前台层面，有刮起血雨腥风、致令人人自危的酷吏集团。这两拨人的权力都来源于武则天。

    首先来看中心层面的武氏子弟。武则天时代构成政治权力的亲属基础者，有下面这些人。在亲属关系上，他们分别是武则天的侄子和侄孙两代人；在政治秩序上，他们分别被封为亲王和郡王。

    亲王

    梁王  武三思  武则天长兄武元庆之子。

    魏王  武承嗣  武则天次兄武元爽之子。

    陈王  武承业  武承嗣弟，追封。

    定王  武攸暨  武则天伯父武士让之孙，始封千乘王，尚太平公主后进封。

    亲王四人：武三思和武承嗣为武则天异母侄子，武承业为追封，三人皆为侄子辈；武攸暨因为尚太平公主而进封亲王，为侄孙辈，乃特例。

    郡王

    武崇训  武三思子，尚安乐公主，封高阳王。

    武崇烈  武崇训弟，封新安王。

    武延基  武承嗣子，始封南阳王，后袭父封，坐私议张昌宗，被杀。

    武延义  武延基弟，袭父封，继魏王。

    武延秀  武延义弟，封淮阳王。

    武延晖  武承业子，袭父封，嗣陈王。

    武延祚  武延晖弟，封咸安王。

    武攸宜  武则天堂兄武惟良之子，封建安王。

    武攸绪  武攸宜弟，封安平王。

    武攸宁  武则天堂兄武怀运之子，武攸暨之兄，封建昌王。

    武攸归  武攸宁弟，封九江王。

    武攸止  武攸归弟，封恒安王。

    武攸望  武攸止弟，封会稽王。

    武懿宗  武则天堂兄武志元之子，封河内王。

    武嗣宗  武懿宗弟，封临川王。

    武尚宾  武则天堂兄武仁范之子，封河间王。

    武重规  武尚宾弟，封高平王。

    武载德  武重规弟，封颍川王。

    作为武氏子弟集团的附庸，可以加上宗秦客、宗楚客、宗晋卿和纪处讷四人。前三人为武则天外甥，纪处讷则是武三思的连襟。

    武氏子弟集团最醒目的特色，是完全未见科举出身者。且不论唐朝高度重视文化，自开国以来就建立起文化程度甚高的官吏队伍，从社会发展而言，以武力开国的王朝到了和平年代，其军功集团的后代也在时代潮流的推动下转而向学，子弟通过科举途径入仕晋升，继续仰仗家世门荫者日渐稀少，受人轻视。武士彟作为唐朝开国功臣，其家族子弟这等学历，透露出武氏家族对于文化的态度，落伍于时代。这批武氏权贵中，最有文化，以至于史家给予记载的是武三思，“略涉文史”，仅此而已。他留下诗歌创作的记录是赞颂张昌宗才高貌美，乃神仙王子晋转世。

    武氏姻亲子弟，宗秦客、宗楚客、宗晋卿和纪处讷四人同样未见学业与科举记载。如果把视野扩大到整个第二条人事线，亦即将武则天男宠团队也一并考察，情况如下：

    薛怀义原名韦小宝，街头摆摊出身，以魁梧雄壮获得宠幸。武则天为了掩盖这段少年劣迹，令其出家为僧，编入女婿薛氏的士族谱中，主持朝廷宗教事业，找人编撰《大云经》，陈说符命，发现武则天是弥勒下凡。

    张易之、张昌宗兄弟，出自贞观名臣张行成家族。张行成少时追随大师刘炫，勤学不倦，应科举及第，历仕太宗、高宗两朝，为一代名臣。张易之兄弟是张行成的族孙，不可思议的是文化家族的子弟竟然不循科举正道，张易之是依靠门荫入仕的，因为白皙美貌，擅长音声。其弟张昌宗首先被太平公主发掘出来，用得称心，转而推荐给武则天，同样表现不俗，大得欢心。张昌宗推荐兄长张易之说：“臣兄易之器用过臣，兼工合炼。”原来这对兄弟兼具炼丹才能。由此可知，他们自小研修道家房中阴阳之术，耽于学业，故难应科举，只好走门荫之路。张易之、张昌宗兄弟几乎专宠，武氏权贵争相为他们牵马前导，招摇过市，加上以权贪赃，惹来妒忌非议，沸沸扬扬。武则天为了遮掩丑闻，让他们主持朝廷文化事业，集中天下美少年和宰辅大臣们，济济一堂，组建文化机构“控鹤监”，更美其名称“奉宸府”，编撰《三教珠英》等大型文集，煌煌千余卷。

    薛怀义的宗教事业，张易之兄弟的文化事业，再往前追溯到北门学士的巨著编撰，有人称之为武则天大力推动的文化盛世。

    做出如此不凡成就的张氏兄弟，虽然没有科举出身，亦非胸无点墨，史书记载张氏兄弟勉强能写成文章，至于和武则天酬唱应对的诗文，自有宋之问、阎朝隐等文学工匠代笔。

    武氏子弟与男宠团队，以及他们同武则天的关系如何呢？在政治风头上，男宠团队风光无限。早先得宠的薛怀义，乃至后来的新欢张易之兄弟，进出内外，武承嗣、武三思一帮武氏子弟争先恐后为其牵马执辔，献诗赞颂，卑辞厚礼，媚态可掬。作为武周皇族却要竭力逢迎男宠，武氏子弟心有不甘，武承嗣的儿子武延基与妻子永泰郡主，以及懿德太子等人私下聚集议论，谈到张易之兄弟任意出入宫中，无不愤恨难耐，摩拳擦掌。这些议论竟然不翼而飞传入武则天耳朵，武则天大怒，逼令武延基自尽。私下非议竟然要付出生命的代价，武则天心中的情感天平清晰可见。然而，这只是表面现象，在政治利益的天平上，武氏子弟才是根本，是武周政权的根基和血脉，武周政权总归要传给姓武的，以至于武则天的亲生子女都要改姓武，试图将他们塞进武氏血脉。武氏子弟充当男宠的马前卒，武则天当然知道，且乐见所为，如果企图反抗则铁腕镇压，绝不留情。这是什么道理呢？并不是男宠金贵，而是武则天将他们当作自己的化身和试金石，测试属下是否绝对驯服而已。服从男宠就是服从武则天，男宠不为人齿，却能够做到诚心悦服，证明对于武则天的驯服臻于精纯，绝对到肝脑涂地，万死不辞。

    武延基是未经风浪的权贵子弟，自命不凡，这恰是心生二志的萌芽，咎由自取。其父辈武承嗣和武三思则迥然不同。武承嗣写不了诗文，却将马牵得十分安稳，让薛怀义和张易之兄弟享尽荣耀。武三思粗通文墨，双眼如炬看出张昌宗乃神仙转世，亲自写诗，还组织编排大型音乐舞蹈表现仙人下凡的绚丽场面，让崔融动情绝唱：“昔遇浮丘伯，今同丁令威。中郎才貌是，藏史姓名非”，把自己感动得涕泗俱下。为什么父子两代差距如此巨大呢？道理就在于武则天同兄弟的关系。武承嗣的父亲武元爽、武三思的父亲武元庆，以及其他诸武的父辈如武惟良、武怀运等等，武则天幼年饱受他们的欺负，尤其是武则天的母亲对他们恨之入骨绝不宽恕，让武则天掌权后给予摧残泄恨，武元庆、武元爽遭黜，配流岭外而死；武惟良、武怀运被诬陷下毒害死外甥女韩国夫人，被处死。武则天的兄弟，自己不死，就只能等待处死。武承嗣和武三思早年都曾随父亲配流边荒，武则天决意篡唐建周以后，出于政治需要才把他们召回京城。骤落暴起，亲尝政治炎凉与绝情，武承嗣和武三思对于姑妈早已胆战心惊，变得十分乖巧，虽然身居高官，却十分清楚权力来自何方，对此顶礼膜拜。这种出格的表现有违自然，看似尽忠，实为恐惧。捆绑到篡唐立周的战船上，构成吴越同舟的共同命运，捍卫武则天就是保卫自身的政治特权，背后的驱动力不是绝对忠诚的感情，而是荣辱与共的利益。

    利益为本，必定得陇望蜀。武承嗣欲望和野心膨胀起来，想独占权力，便策动武则天尽诛李唐子孙，同时组织宵小请愿，试图成为太子，吞下武周的果实。武则天未遂其愿，致令武承嗣怏怏而死。作为政治精算师的武则天，是信任有父仇的侄儿，还是相信亲生的儿子呢？武承嗣越界了，利令智昏，自取灭亡。从他儿子武延基非议张易之兄弟一事，武则天难道看不出来武承嗣不为人知的家庭内部只讲利益不尽忠诚的真情吗？武承嗣和武延基父子之死，显现出武则天的底线：皇位传给姓武的亲生儿子，武氏子弟掌控朝廷，成为武周政权的核心。所以，武则天花费更多的心血培育武氏第三代，几乎都封为郡王，出将入相，以保武周江山长远稳固。武则天的政治算盘在内心早已权衡清楚，决不是晚年在大臣的谏言下幡然醒悟，立子继承。大臣们的谏言因为契合武则天的心意而被采纳，同时也给了跃跃欲试的武氏子弟一个无法扭转的交代。通过和朝中大臣讨论继承人问题，武则天也摸清了大臣们的政治态度。她这个决定是明智的，武氏第三代在内政外交上的庸劣表现，根本不可能作为皇帝撑起大局，与其被推翻，不如回归政治合法性。之所以成为糊不上墙的烂泥，武氏几代人皆无学业与科举，已经有了答案。

    其次来看前台层面的酷吏集团。在唐朝，武则天时代首次出现酷吏，完全改变了政治规则和社会风气，影响深远。唐朝的出现不仅是一次成功的改朝换代，而且是一场重要的政治革新。五胡十六国南北朝分裂时代，恃力使诈成为政治常态，社会上层失德，下层失信，导致国家数百年难以真正统一。唐太宗总结历史教训，致力于重建法律与制度，取信于民。唐朝建立到武周时代将近七十年，垂拱而治，依靠的就是官民互信，制度公平。到唐太宗晚年，“天下刑几措，是时州县有良吏，无酷吏”。武则天僭主当政，威望不足，忧惧群臣不服，便重用一批酷吏大规模整肃异己，构陷告密，开启酷吏政治时代。酷吏政治与武则天执政相始终，甚至长于武周政权的存在时间。武则天倒台之后，酷吏政治随之而去。但是，它没有消亡，而是潜伏在帝制体内，不时兴风作浪。

    酷吏作为僭主独裁的主要工具，威慑并实际控制整个官僚阶层，因此，他们无疑处于政治权利结构的顶层。另一方面，酷吏的所作所为，乃秉承上意，因此，他们常常被轻视为君权行使的道具，而非具有独立意志和利益的集团。事实并不尽然，当工具坐大的时候，便逐渐膨胀起欲望，从狐假虎威，假公济私，直至奴大欺主。武则天对薛怀义隐忍再三，唐朝多少皇帝死于宦官之手，说明任何政治集团一旦成形便有了主张和利益诉求。所以，酷吏集团不可仅仅当作皇权的影子简单处理。当告密和清洗全面铺开之后，海量的案件并非君主所能掌控，检举何人，镇压什么，都与发起的酷吏的感情、学识、见地和利益息息相关。他们的指向性变成强有力的鞭子和精神指挥棒，逼迫并规定着官僚队伍的思想观念、施政行为和价值取向，进而深深地影响文化程度不高的芸芸众生，形成弥漫世间的社会风气。最终出现的结果往往和君主最初的政治蓝图不尽吻合，甚至相去甚远，原因就在于君主和酷吏文化水平和利益见识的落差。君主用工具剪裁世界，酷吏则以其品行见识塑造世界。大千世界从来不是单方面所能制造的，而是各方面合力的产物。

    酷吏的身世塑造其品行和情感，文化见识规定其眼光和行为。这两者又极大地左右着官僚队伍乃至整个社会的文明水平。吏治庸劣从来都是社会堕落的驱动力。

    武则天时代，告密成风，酷吏成群。然而，能够得到武则天重视，挑选出来兴风作浪，成为酷吏代表的主要有以下这些人：

    来俊臣，乡间地痞；左台御史中丞。

    周兴，少习法律；秋官侍郎，尚书左丞。

    傅游艺，吏员；同凤阁鸾台平章事。

    丘神勣，官宦子弟；左金吾卫大将军。

    索元礼，胡人；游击将军。

    侯思止，家奴无赖，文盲；朝散大夫，左台侍御史。

    万国俊，乡间地痞；朝散大夫，肃正台侍御史。

    来子珣，无学，告密入仕；左台监察御史。

    王弘义，告密入仕；左台侍御史。

    郭霸，吏员，革命举；左台监察御史。

    吉顼，进士；天官侍郎，同凤阁鸾台平章事。

    让一个时代陷入血腥恐怖的酷吏，只有吉顼一人是进士出身。少时读过书的仅见周兴，曾经学习法律，为日后翻弄法条打下基础，属于刀笔吏。上述11人中，9人出自乡间地痞无赖，甚至侯思止还是个文盲，却官至左台侍御史，主持监察炼狱。这批人的行迹与文化程度，显然无法通过朝廷正规的仕进考察，所以都由武则天直接提拔重用。武则天用的人，再荒唐也不容议论。侯思止言行举行粗野愚蛮，成为官场笑柄。武则天知道后，怒斥嘲笑者：“我已用之，卿何笑也？”当听说了侯思止那些惊动四座的话语，自己也忍不住喷笑。侯思止丑态百出，在武则天看来却是愚忠可靠，故其官位坐得十分牢靠。11人中，有文化学业者2人，占18%。另一方面，升任宰相的也是2人，同样占18%。文化低同官职高形成鲜明的对照。

    武则天时代用人的两条线、三个层面，第一条朝官线基本遵循入仕正常规则考察录用。在武则天酷吏大清洗的恐怖气氛下，动辄犯咎下狱，故上上下下明哲保身，敷衍了事。他们整体文化水平最高，权力却最小，得过且过，形同摆设。第二条线的中心层面，有武氏子弟和武则天男宠团队，文化程度颇低，职位最高，握有大权，构成武周政权的政治人事基础；前台层面的酷吏集团，基本由地痞无赖出身者组成，通过诬告或者兼进谀词而获重用，飞速蹿升，权势熏天。和中心层面相比，前台层面的酷吏集团是必须的存在，至于具体的个人则需要经常更换，败亡亦在瞬间。他们得意之时极尽残忍，破灭之际人剐其肉，遗臭万年。他们刮起互害之风，自己无一幸免，“既为祸始，必以凶终”。

    从人事结构来看，武则天时代是武氏子弟、男宠团队和酷吏集团联合管控朝官，进而掌控全社会；同时也是无知对文化的压制，权力对于法律制度的践踏。

    三、士族政治与科举

    武则天基本遵循唐朝官员入仕与晋升的铨选原则，另一方面则在权力的上层重用武氏子弟、男宠团队和酷吏，掌控百官的黜骘乃至生杀大权，主宰政局。她重用之人学历低，非贵族名门出身，格外引人注目，以至于有研究者把武则天作为唐朝政局的分水岭，认为武则天大量提拔庶族寒门，改变了门阀士族对于政治的垄断。陈寅恪进一步把视野扩大到北周，认为当年宇文泰组建关陇地区胡汉各族实力人物组成的“关陇集团”，垄断政治直到武则天方才打破。武则天大批提拔科举出身的人入仕，形成“新兴阶级”，如此则武则天不仅在唐朝，乃至在中国古代历史上都是改变历史进程的领袖。以一人之力改变三代王朝的历史方向，这样的功业恐怕空前绝后。

    陈寅恪对于南北朝隋唐史研究的贡献，在于提出了宏大的问题，启发历史学家去思考和论证。学说的成立，首先要通过证伪的检验，其次才是不同视角的分析论辩，在思想碰撞中发展。

    陈寅恪对武则天历史定位的基点是魏晋南北朝以来的士族政治。首先需要厘清的概念是这个历史阶段的士族与士族政治。士族指的是社会的统治阶层。士族与皇帝为主导的政治、军事势力结合，相互依靠，掌控并长期把持中央王朝到地方的政治权力。曹魏建立“九品官人之法”，表面高举“唯才是举”大旗，很快转为重视家世，到了西晋则日益强调家世礼法，从铨选制度上极大强化了官僚士族的特权地位，世袭垄断政治权力，形成固化的士族政治形态。魏晋南北朝的士族，大多起自东汉崩溃以后一再出现的大动乱，在兵荒马乱中聚集亲族乡党据险自保，组成自立武装，割据乡村，概称为“坞壁”。几百年的战乱和外族入侵，使得坞壁得以长期维持，遂演变为世家大族，将地方社会碎片化，以至于重新建立的各个王朝都必须得到他们的支持才能控制地方。世家大族大小不等，大者跨郡连州，千家万户；小者数百家一族，武断乡曲。他们通过联姻构成亲族网络，跻身于王朝官僚之中，凭借在乡势力支持政权，利用国家权力垄断地方。婚和宦是支撑士族长久不衰的两大法宝。士族内部有高下等级之分，这种区分不仅凭借在乡实力和官位高低，还根据文化和声誉，虽然不像确定官品那样清晰严格，但也有必备的条件：连续几代人中出现公卿宰辅一级的高官，属于政治条件；颇有文化学养，遵循礼法家教，属于文化条件。政治和文化两方面条件都具备的世家大族，受到社会普遍的承认与重视，例如北朝隋唐的崔、卢、李、郑、王等山东士族，韦、裴、柳、薛、杨、杜等关中士族，被视为最高的门第。其下还有各个州郡级别的士族等级，构成从朝廷到地方的世家大族等级结构。王朝在此基础上，结合在当朝官职的高低，编撰氏族谱，作为铨选的家庭条件和分配政治权利的依据。北魏孝文帝开其端，划分甲乙丙丁等第；后续王朝全都跟进。唐太宗修《氏族志》，唐高宗和武则天重修《姓氏录》，表明对于士族等级秩序的高度重视。据此可知，说武则天力图打破士族政治，不知从何说起。兼具实力、官品、文化三者优势的士族，得到各大政治势力的积极拉拢，成为其政权的支柱。他们在朝身居高位，在地雄踞一方，并且根据各自的身份地位形成比较固定的通婚圈，备受瞩目，演变成社会上重视的“门阀”。这种政治生态称为“士族门阀政治”。在确定士族身份等第的时候，文化条件颇为重要，品行与学术决定家族的声誉和社会影响。官职高却没有文化被视为权势豪门，地方上有实力缺少文化的家族被称作豪强，总之同具有文化色彩的“士”难以沾边。所以，士族研究从这个角度区分兼具文化学养者为士族，仅凭官职或者强宗势力者为世家大族。当然，这一区分并不是那么严格。作为统治阶层，常见笼统使用士族一词。

    在世家大族或者士族等级秩序的框架之内，其下层被称作“庶族”“寒门”等。以往的研究对于士庶之分并不清晰，如果以五品以上官职划线，那么庶族就是下层官吏直至小地主之家，缺乏权势的中小地主自然被归为“寒门”。他们也被称作“庶族地主”等。然而，无论士族、庶族，他们都属于统治阶层。即使武则天时代出现大量提拔庶族寒门的现象，既不构成“新兴阶级”，也完全称不上“社会革命”，充其量只是统治阶层内部的成分调整。何况武则天任用的官员，如前面列示的三个层面，酷吏多为无业的地痞游民，连“寒门”都构不上；武氏与男宠固然文化水平低，但其家族在唐朝已经上升为功臣权贵，甚至是皇族，无法再用“庶族”指称他们；而朝官的选任与唐朝开国以来的状况没有大的变化。综合三个层面所展示的真实状况，无法支持陈寅恪所谓武则天缔造庶族寒门“新兴阶级”的假说。陈寅恪并未提供实证分析的根据，不知所本，故只能对其结论提出商榷。

    其次，陈寅恪提出的“新兴阶级”，最重要的特点是科举进士出身，工于为文。亦即武则天之前，唐朝铨选重视门第家世，用的是“西魏、北周、杨隋及唐初将相旧家”，而武则天破格录用科举出身者，形成与所谓“关陇集团”对立的“新兴阶级”。

    这里涉及两个问题：1.支撑起北周、隋、唐政权的旧家，亦即所谓的“关陇集团”的存续状况。宇文泰以后来所封的八柱国、十二大将军等二十余家创业家族为核心建立西魏、北周政权，所言甚是。但是，这一创业功勋集团从北周宇文护专政时起就遭受猜忌和镇压；北周武帝辉煌的功业昙花一现，人亡政毁；杨坚政变建隋，抑制并清洗宇文泰组建的关陇集团主要家族，隋炀帝则重用江南士族。李渊建唐，依靠的是河东士族与大姓，唐太宗则强调用人上的五湖四海。这一历史进程呈现了走出关陇的清晰脚印。政权长治久安的人事基础在于用人区域和社会阶层的广泛性，统治者只要不失心智，自然深谙个中道理。

    2.政治史所讲的地域政治集团，是指集中任用某地人的政策与原则。西魏、北周的统治地域仅仅局限于关陇地区，只能任用关陇人事，别无他选，并不是拥有广阔的统治区域而有意识地专门任用关陇人事。所以，所谓“关陇本位政策”或者“关陇集团”，说了等于没说。更何况严酷的政治现实，生死攸关，且政治目标与利益各不相同，从来都是一朝天子一朝臣，甚至是一朝天子数朝臣，哪有一朝大臣数朝皇帝，更加不可思议的竟然是一朝大臣三代王朝，从理论到现实都不成立。从宇文护到隋文帝，执政者仅有关陇地区的从政经历，故人事基础局限在这里。即便如此，他们也在扩大用人的面，压制宇文泰的创业“旧家”。明白无误的变化出现在隋炀帝时代。隋朝成为全国性政权之后，用人的区域日渐扩大。隋炀帝曾经指挥统一江南的战争，皇后又出自南朝皇族萧氏，故他重视南方，拔擢江南士人，委以重任，甚至主导朝政，极大改变了关陇官僚居多的成色。唐朝自创业时起，就以太原组建的班底构成核心人事圈，笼络山东士族。武士彟就在此时进入政治核心圈，崛起于政坛；武则天也是因为功臣之女才选入宫中，日后执掌权柄。武氏是李唐政权下的既得利益者，和其他创业家族命运与共，构成唐朝人事的基本盘。武则天出于一己之私，打压李唐政权的忠实支持者，但她主要用没有社会根基的酷吏集团作为打手整肃官僚，并没有从根本上改变官僚队伍的成分和用人路线。其道理显而易见，武则天要的是至尊皇权，而不是摧毁自己赖以生存的政权根基；她要在最高权贵阶层中长期占有武氏一席之地，并不为酷吏之流痞子政客谋求利益，改变权贵阶层的结构；她处心积虑推进李、武联姻，就是为了补武氏合法性短板从而获得长远安定；她深知根深蒂固的士族阶层的重要性，所以对韦氏、杨氏、崔氏等老牌士族笼络重用，甚至让他们父子兄弟同时身居要职，宽容他们对于男宠团队乃至武氏子弟的轻蔑；她扮演官僚、“旧家”敌对者的角色以煽动下层，却没有改变李唐依靠官宦士族的组织路线。所以，武则天表面上看似泼辣凌厉，其实内心极其精明，她走在极端政策的边缘，却在最关键之处未越雷池一步。从本质上看，她是士族政治的坚定维护者，而非掘墓人。

    北周隋唐的相关性在于三代王朝的建立者同出一源，此偶然现象的关键在于北周、杨隋皆短祚，事起仓促，只要不是被其他政治势力所征服，剩下的便是同一平台脱颖而出的新秀。新的创业者都是受到当政者压迫而心生异志的雄才，而非同一事业的前仆后继者。理念、目标和利益各不相同，如何构成同一体质的政治集团呢？所以，所谓的“关陇集团”把持北周、隋、唐三朝政治的议论，属于想象的建构。

    不存在垄断三朝政治的所谓“关陇集团”，却出现一个确定不移的现象，那就是官员铨选与晋升中，科举出身者日益增多，反映出对于文化的要求越来越高，成为大势所趋。为什么会出现这样的变化呢？这一趋势究竟是个人意志的产物，还是国家社会发展的必然？

    自从东汉末年董卓被杀时起，朝廷就失去了对全国的统制，内战爆发，一步步沦为彻底的分裂割据，直至唐朝建立为止，中国在战乱和分裂中度过了将近四个半世纪的漫长岁月，其间虽然有过西晋和隋朝短暂的统一，却都以失败告终。分裂战乱的时代，真理由思辨的洞彻发现沦落为暴力的胜负角逐所决定，乱世的最高道理就是胜利。所以，这个时期过眼云烟般的繁多政权无不把实力和功绩作为用人的根本标准。曹操一再发布《求贤令》，公开倡导重用反道德能取胜的人，开启其端。魏文帝时代创立“九品官人之法”，任命中央到地方各级中正官来评选人才。这个制度存在根本性的内在冲突，亦即选用士族来贯彻“唯才是举”，不啻缘木求鱼，随着时间推移越来越走向反面。士族出身的中正用“家世”条件评定人才品级，结果“九品中正制”成为加强并固化士族门阀的强有力机器。另一方面，频仍的战争涌现出许多勇武的将领，在军事化的国家机器中占据主流。“九品官人之法”制造门阀士族，军国体制制造军功阶层，源源不断，在王朝政权内混杂合流，形成门阀和军功两大特色，在权斗中共存。权斗源于军功阶层对于文化和士人的蔑视，痛下杀手，必欲将其奴仆化。北魏崔浩事件等等，层出不穷的文化大狱莫不因此而生。共存则是对现实的屈服。攀援上权力宝座的各色人物，无不企图固化既得利益，军功和权势皆不可长久，丛林互噬注定没有胜者，欲求长在，最有效的途径就是转变为门阀，故不可一世的军功阶层不得不向现实低头，与士族合流，借其金字招牌悬挂在列戟的大门之上。北魏孝文帝确定胡汉姓族等第，令其通婚联姻，这样的事例屡见不鲜，道理如出一辙。皇帝亲自做媒，不是开婚姻介绍所的业余爱好，而是营造铁打江山的专业操作。

    军功士族门阀政治，是东汉灭亡以来中国长期不能统一、政权无法稳固和不断发生社会动乱的根源之一。因此，想要建设稳固富强国家的统治者都必须改变这种局面，任何根本性的社会变革，一定依靠法律、制度乃至文化价值观的改造重塑。人治最不可靠，凌驾于法律制度之上的人治，既可行善，亦可为恶，方向上飘忽不定，则易于颠覆。隋文帝建国之后，断然拆除士族门阀政治的制度台柱，“九品及中正至开皇中方罢”。自曹魏创立以来沿用数百年的“九品官人之法”终于被废除。此举有利于扩大政权的社会基础，故为后面的王朝所遵循。

    军功士族门阀政治是战乱时代军国体制的产物，打破门阀政治不仅是制度的变革，更是治国理念的转变。古人说马上得天下，王朝是靠武力打出来的。可是，政权建立之后，国家能否继续采用军国体制管理呢？唐太宗对此有深刻的认识，他还是秦王、天策上将的时候就积极延揽四方文学之士，皆为一时之选，其佼佼者号称“十八学士”。戎马倥偬之际，唐太宗仍然和文士一起深入研讨如何治理国家，从根本上认识到治国不能采用军事命令式的行政强制，更不能听任权力恶性膨胀，凌驾于一切之上，必须讲道理，重规则，建立完善的法律与制度，提升社会文化道德水平，才能实现国家繁荣强大、长治久安的目标。政治路线须要人去落实，什么样的人做什么样的事，所以官吏铨选至关重要。隋文帝废除“九品官人之法”，代之以科举考试，隋炀帝进一步确立进士科的主流地位。唐朝建立之后，强化了科举在铨选中的比重，特别是唐太宗以军事统帅的威望率领创业的军功部属转变崇尚武力的思想观念，积极倡导文治，大力办学，拓展科举入仕的途径，坚持不懈，推动社会形成尊重文化、推重科举的风气。唐朝科举设秀才、明经、进士、明法、明书、明算六科，秀才科后来停止，明法、明书、明算三科为专门科目，故常设科目为明经和进士。盛唐以后进士科越来越显赫，压过明经科，求仕进者趋之若鹜。唐五代时人王定保撰述唐代科举状况，说道：“进士科始于隋大业中，盛于贞观、永徽之际。搢绅虽位极人臣，不由进士者，终不为美，以至岁贡常不减八九百人。”显而易见，唐朝甫建即大力推进科举制，发展迅速，到唐太宗贞观年代已经被世人视为仕进正道，很受尊崇，以至于权贵子弟通过门荫或者军功起家者都以不能由科举入仕而深感不足，一生抱憾。唐太宗以广收天下英才为己任，当年见到新进士缀行而出的场面，欣然说道：“天下英雄入吾彀中矣！”偃武修文打下唐朝将近三百年的根基。文化的盛大，不是用读书人的多少所能表现，最重要的是社会各个阶层都崇尚文化，以此为荣，蔚然成风，这就是王定保盛赞贞观时代的立论所在。读书人虽多却甘当鹰犬，重用近乎文盲的酷吏镇压士人，哪怕编撰出颂扬的诗文，王定保并不以为是文化盛世。把科举作为入仕正道是唐朝既定国策，坚持推动，不待武则天方才肇始。这个重要转变是军功立国后走向长治久安的必由之路，具有客观的必要性与必然性，从这个意义上看，唐太宗并无首创之功，却是及早的觉悟者，避免了有害无益的弯路和折腾。

    科举制同九品官人之法相比，颇具公平性。后者注重家世出身，造成士族高门几乎垄断官场的僵化局面。科举则允许士子报名投考，凭借个人成绩录用，打开了社会下层之人上升的通道。侯君集、孙伏伽等都是自寒素举进士入仕、初唐位居朝廷大臣的著名例子。唐高祖武德五年（622），居住在邺城的陇西人李义琛、李义琰兄弟，及其堂弟李上德三人同年考上进士，成为佳话，载入史册。投名报考的科举制一旦取代九品官人之法，下层士子入仕的比例必然越来越高，毋庸置疑。然而这是一个发展的进程，积累数十年的科考，到武则天时代寒门出身增多，只是结果的呈现。社会的进步必定是人身及身份性限制的日益解除。科举制代表的正是这个方向。

    需要特别指出，论述科举制时往往把录用人数的增多作为制度推进的直接有力证据，实则南辕北辙。科举制不是一般的入学考试，而是入仕当官的铨选，所以不可能大量录取，其名额由每年需要补充的官吏员数来决定。唐太宗励行小朝廷，精兵简政，其组建的朝廷人数有各种记载，这里选取较多的一种，《新唐书·百官一》记载：“初，太宗省内外官，定制为七百三十员。”乱世的社会平均年龄颇低，故初唐朝廷需要世代更新的人数很少，决定了科举录取人数必定较少。依照唐朝“壮室而仕，耳顺而退”的三十年仕宦期，到唐高宗年代，科举录用人数呈现梯度增长，并随时代推移构成阶梯式上升，都是自然而然之势，绝非个人一己托天之功。

    依据官吏世代更新的需要决定科举登科人数的原则，唐高宗显庆二年（657），主持人事的黄门侍郎、知吏部选事刘祥道上疏，核算当时内外文武官，九品以上者13465人，取大数放宽至14000人，每年补充500人都要多出不少，而当年补充了1400人，超过需要2倍多。亦即唐高宗时代，科举登科加上依靠门荫等入仕者早已供大于求，不能再扩大科举登科人数了。由此可知，唐太宗贞观时代科举录取人数较少，反映当时严格执行职官的编制。高宗时代已经在10∶1的求仕压力下增加了很多登科名额，人浮于事。到了武则天时代，为了夺权乃至篡唐建周，武则天“务收人心”，做了重要改变，一是取消考试糊名，致使贿赂公行。例如来俊臣握生杀大权之时，大肆收受请托，每次铨选都要违法安排数百人入仕，至于王公权贵的插手科考，难以尽述。二是不按成绩，任意扩大录取人数，天授二年（691），十道举人，大批拔擢官吏；长寿元年（692）一月，武则天接见各地举人，“无问贤愚，悉加擢用，高者试凤阁舍人、给事中，次试员外郎、侍御史、补阙、拾遗、校书郎”，数以百计。而且还开启了试官制度，大批未能通过考试和品行考察的人，先行任职，把政务和民生当作儿戏。新录用的官吏如此之多，致使各个部门冗官泛滥，社会流传歌谣称曰：“补阙连车载，拾遗平斗量；欋推侍御史，碗脱校书郎。”

    官多至滥，并不会给下层寒士带来普遍的机会。武则天任用士族主持铨选，李峤录用权势之家的亲戚二千余人，以员外郎身份到各部门掌管事务，同在编的主官发生激烈争执，甚至互相殴击。此类官场乱象，是士族权门内部利益之争，和平民有什么关系呢？在古代史上，冗员滥官从来是权力泛滥的表现，带来的是更大的不公平。从武则天时代到中宗、睿宗朝急剧膨胀的“墨敕官”“斜封官”，请托贿赂充斥官场，便是其结果，而庶族寒门更无升迁的希望，如何形成与士族“旧家”对抗的“新兴阶级”呢？冗员滥官是对法制的破坏，绝不是科举制的进步，更不是所谓的“社会革命”。只有一律平等的公平制度，才是下层士子的上升通道。

    武则天时代的官僚阶层，呈现出非制度性越级提拔的武氏子弟、男宠团队、酷吏集团真正掌握权力，控制朝官的状态。朝官则基本遵循铨选途径入仕，没有改变唐初以来士族与官宦子弟为主的基本面貌。魏晋南北朝隋唐时代最重要的社会变革是科举制取代九品官人之法，拆除士族门阀政治的制度性支柱。这场变革肇始于隋朝，成为主流而蔚然大观于唐太宗时代，武则天时代未见制度上的进步，冗员滥官却对制度造成重大伤害，反而阻碍了寒门士子的正常上升。

  •  刘永华《程允亨的十九世纪：危机》

    文章节选自《程允亨的十九世纪：一个徽州乡民的生活世界及其变迁》（ 刘永华 著 三联书店2024-11）

    危机（节选）

    光绪九年程氏兄弟分家之后的最初几年里，无论是清王朝还是程家自身，似乎都没有发生太大的变化。不过分家十几年后，这个世界发生了几次令人震惊的事件。甲午一役，大清海陆军败北，朝廷不仅面临巨额赔款问题，国内改革的呼声也越来越高。维新运动接踵而至，但不久即告失败。随后是庚子年的义和团运动爆发、八国联军侵华和辛丑年的巨额赔款。程家自身的变化发生得要早些。光绪十六年、光绪十八年，允亨的双亲先后去世。光绪十九年，同仓成亲。一年后，新一代出生。程家完成了新一轮的代际继替周期。此后，八国联军攻占北京那年，程家发生了家计危机。这些发生在一个王朝和一个农户层面的国事与家事，并非没有丝毫关联。这两个层面发生的事件，以及其他一些因素，都在程家家计危机的发生中扮演了或大或小的角色。

    国事与家事

    为理解这一时期程家的生计环境，我们先来梳理一下此期对程家生计影响最大的两种商品——大米与茶叶——的价格。太平天国后期，江南、安徽、江西一带米价大幅上涨。

    太平天国结束后，各地米价普遍下跌。从 19世纪70年代中叶至80年代中叶，米价基本保持稳定。此后价格逐渐上涨，1895年的米价比1875年上涨了50%。至清朝覆亡时，米价比 1875年上涨了1.5倍。可以想见， 1895年前，米价上涨相对缓慢，而此后15年的时间里，价格涨幅较大。回到婺北米市， 19世纪后期价格运动的总体方向与其他地区相似，不过 19世纪末以前的涨幅不甚突出。如表 7.3所示，太平天国后，婺北地区的米价大致回落至19世纪40年代的水平（但略低于道光十九年、二十年），这个价位基本维持至90年代中期。至 19世纪末20世纪初，在全国米价攀升的影响下，婺北地区的米价也迅速上涨。光绪二十二年米价是每石2.5元，光绪二十六年攀升至3.06元，比光绪二十二年上涨了22%。那么茶价呢？根据第六章的讨论，太平天国运动结束后，茶价为0.19元/斤左右，较太平天国运动开始后的价格（0.13—0.15元/斤）稍有回升，但远低于运动爆发前的水平（ 0.29元/斤）。分家后，茶价一度有所下跌（ 0.155元/斤），光绪十八年后稍有回升（ 0.166元/斤），但仍较分家前低了将近13%。与此同时，跟分家前相比，程家生产的茶叶总量也有所下降。分家前，每年产茶 2担左右，分家前几年甚至达到年产 3担的峰值。分家后因茶园分割，产量回落至年产1.5担至2担余的规模。因此，分家后的十余年时间里，米价和茶价/茶叶年产量之间的剪刀差有所缩小，但收缩幅度不算大。至 20世纪之交，随着米价的攀升，这个剪刀差才进一步收缩，开始对程家的生计构成威胁。

    除了茶叶收入稍有回落外，这一时期程家的其他现金收入也有一定缩水，其中最重要的是山货。山货在程家现金收入中的地位，在相当长的时间里仅次于茶叶，最高时全年收入可达近19元（分家前）。

    程家历年收入一览表 （单位：元）

    但分家后，仅光绪二十一年、光绪二十二年超过10元（分别是13.20元和11元），其余年份都在9元以下。这一时期投入收购、加工黄精和挖掘葛根、制作葛粉的时间，都出现了大幅下降的情形。截至太平天国前期，在这两种山货的生产与贸易方面，程家共投入253日，占所有生计行事投入天数的8.42%；太平天国结束后至分家前，劳动投入增加至767.5日，在生计投入时间中的占比上升至 13.05%；分家后，劳动投入下降至99.5日，占比降至仅 2.65%，两者的时间投入，无论是绝对数量还是相对比例都大幅下降。细读排日账，山货收入的下降，跟葛粉产量的下降有直接关系。分家后，程家投入葛根挖掘的时间越来越少。这一方面跟分家后程家劳力的减少有一定关系，但更重要的原因，或许是经过数十年的密集挖掘后，葛根资源逐渐减少（同期制作葛巾的时间也减少了，两者应该是有内在关联的）。其结果是，分家后，程家从山货获取的现金收入逐渐下降，这对 19世纪 90年代以后的程家生计来说，无疑是一个不好的消息。

    程家历年粮食产量估算表 （单位：市斤）

    不过，对程家生计带来影响的，并不限于米价、茶价的波动和山货收入下降的问题，借贷在其中也扮演了重要角色。根据托尼（ RichardH. Tawney）的说法，借贷是历史上乡民社会的基本问题之一，他曾经指出：“在所有小农经营耕作的国家里，乡村社会的根本问题并不是工资收入问题，而是借贷问题。”在困扰20世纪二三十年代中国农户生计的各种因素中，他将债务视为其中很重要的一项。他的看法得到了其他研究的证实。据陈翰笙30年代的调查，广东番禺调查的67个村子中，有50个村子的负债农户占70%以上。他估计，整个广东有三分之二的农户负有某种债务。他指出，广东农户的借债，十分之三是因为疾病、婚丧或其他临时的费用，而十分之七只是为了购买粮食养家糊口。由于本书开头谈到的那场光绪二十六年十月发生的危机是由债务引起的，我们有必要梳理一下此前十年（1891—1900）程家的债务状况。

    从排日账看，程家在 19世纪七八十年代并非完全不举债，但这些债务数量不大，在程家的偿还能力范围内，这种状况一直延续至分家后最初几年。光绪十七年，程家仍无数额较大的举债记录。不过此年发开过世。次年，发开的妻子也亡故。当年出现了两笔举债记录。第一笔发生于三月初十日，通过抵押田皮一秤，向廷远祠借入5.5银元。第二笔发生于同月廿五日，从余味山祠借来英洋22元。这两次举债原因不详，不过主要原因估计有二：其一，支付前一年与本年为发开夫妇办理小规模丧葬仪式的开销；其二，支付同仓娶亲的聘金。光绪十八年四月二日，也就是允亨母亲过世不到三个月后，程家举行了订亲仪式，聘金46元，这是一笔不小的开销，已经超出了当时程家一年茶叶销售的毛收入。加上公堂礼、谢媒人钱及举办婚礼酒席等各种费用，这场婚礼的开销当不在60元以下。如计入排日账记录的其他相关开销，此年的仪式与礼物开销高达73元余，为全年总开支的68%。

    程家历年开支结构表（光绪十八年—二十一年） （单位：元）

    允亨的儿媳是在次年正月二十五日进门的。在此前后，发生了一系列借贷行为。第一笔发生于进门十天前，程家以田皮字一张为抵押，向一位村民借入英洋5元。第二笔发生于此次借贷一个月后，也以田皮字一张为押，从一个会社借入英洋 10元。这两次举债很可能是为了支付同仓成亲酒席的开销。这一年的第三笔借贷发生于八月十四日，当天允亨从兄长和一位村民处借来11元，当天归还给余氏云青祠，取回契字（总共花销 24元，另 13元由允亨自筹）。七天后，程家再次以田皮字为押，从余味山祠借入英洋 20元。这几次借贷应该也是为了处理娶亲的费用，而第三次借贷显示，程家期望通过资金的周转，保住自己的一块耕地。

    光绪二十年三月初二日，程家支付了2.3元，赎回一处茶坦的契字，这是当年发生的唯一与借贷有关的行为，而且这次还是取赎而非举债。光绪二十一年发生两次借贷。第一笔发生于正月十五日，程家以庄下田契为押，从一位村民手上借入英洋13元。第二笔发生于六月初七日，这次以牛栏田契为押，从一位村民那里借入英洋 17元。光绪二十二年五月，程家再次以顿底田皮字为押，从一位邻居处借入英洋 10元。这三次举债的用途不详，很可能是为了分拆前两年所借债款的利息及支付光绪二十一年十二月十六日允亨孙子“做三朝”的开销。自光绪二十三年至二十五年，程家的排日账已佚。不过从光绪二十六年程家的债务清单看，这几年程家共借入 6笔款子，其中光绪二十四年（1898）借入4笔，光绪二十五年借入 2笔，总计 65元（详下），占清单所列债务总额的一半余。由于这几年的排日账没有保存下来，这几笔债务的用途已无从知晓。

    那么，这些债务对程家带来多大的经济压力呢？我们来看看沱川的借贷利息问题。综合排日账记录的的借贷案例，沱川借贷利息大概有三种情况。其一，无利。这种情况很少见。前面提到，道光二十年二月十九日，发开从母亲手上借到8两余银子，没有还款记录，应该是无利的。光绪二十七年四月二十日，允亨归还有兴2元，排日账记录“无利”，查借入时间是三月十六日，可能因时间较短，有兴没有收利息。其二，10%左右。这种情况也比较少见。咸丰五年七月七日，发开向彦兄借钱，“言定加一”，也即年息率10%。光绪十八年十二月二日，允亨向春元借入5元，次年五月二日归还，刚好满半年，利息为240文，可求得年息率为9.6%。光绪十九年八月十四日，允亨从兄长允兴处借入6元，光绪二十一年七月初六日支付利息1元，外加铜钱100文，可推得年息率为 9.2%左右。不过光绪二十二年七月十七日支付利息1元，年息率升至 16.7%。这个案例说明，就算关系很近的亲属，也会收取不低的利息（下面余熊能借贷例也是如此）。其三，20%左右。这种情况最为常见。道光二十五年二月十七日，发开向社会借入400文，次年二月二十二日归还本息共480文，可求得年息率为20%。光绪十九年八月十四日，允亨从钦五祠借入5元，光绪二十年八月十四日还支付利息1元，年息率为20%。光绪十八年四月初十日，允亨从外甥余熊能处借入1元，六月十九日还，付利息30文，可求得利息率为 18%。光绪十九年八月九日，允亨向灶子母借入5元，光绪二十年八月八日支付利息1元，年息率为20%。前两例是向会社、祠堂借贷的事例，后两例是向个体借贷的事例，除余熊能事例可能因有亲属关系利息稍低外，其他均为 20%的年息率。此外，排日账中还记录了以田地、房屋为抵押，利息以租谷形式交付的事例，也不多见，兹不赘述。

    参照光绪二十六年程家债务清单，从历年借贷数额看，光绪二十三年以前，程家的借贷总数累计54元；光绪二十四年、光绪二十五年两年累计70元。可见光绪二十三年之前，程家借贷问题还不算严重，光绪二十二年、二十三年甚至没有借入大笔款项（同时，光绪二十一年、二十二年程家的收入不错），如以20%的年息率计算，每年需偿付利息10.8元，其数额尚在基本可控范围内。相比之下，光绪二十四年后，借贷数量明显增加，光绪二十四年、二十五年，共借入70元，特别是光绪二十四年借入了50元，程家的财务状况急转直下。如以20%的年息率计算，每年需偿付利息24.8元，如以光绪二十六年程家的年收入计算，程家每年需支付的借贷利息，就高达年收入的64%，这还没计入米价上涨造成的经济压力及其他小笔借贷的利息。因此可以断定，随着债务的大幅增加，程家仅仅靠生计收入已无力偿清债务。使情况变得更糟的是，田地的抵押，意味着程家每年必须缴纳更多的地租，程家自身的口粮供给能力也受到影响。

    程家历年收入一览表 （单位：元）

    光绪二十六年发生的一笔不成功的交易，直接影响到程家资金的周转，也有必要稍做讨论。程家生产的茶叶，一般是由茶商前来沱川收购。这一年华北爆发义和团运动，茶叶市场似乎不太顺畅。根据当年的海关报告，截至 1900年上半年，中国多数地区贸易正常进行，华北地区只是到了 6月局势才开始变得严峻，但其他地区贸易照常进行，长江流域的局势风平浪静。在上海茶市方面，跟1899年相比， 1900年红茶出口英、德、美、俄四国的数量有相当程度的提高。报告还提到，此年徽州茶（Hyson）的交易数量跟上一年相似。不过报告也显示，1900年中国的绿茶出口量，比上一年少了13300多担（但较之 1898年增加 15100多担）。报告还提到，“绿茶市场于 6月8日开启，开始出售的是少量平水茶，其价格比前一季度开市低了大约10%。茶叶质量与 1899年不相上下；但由于对主要消费市场 —美国—的预期很糟，一开始成交量很小。但是，后来需求增长，7月中旬前，价格已回升了 5%—10%”。国际贸易的波动，尽管对总出口量的影响不大，但可能造成地方茶市的震荡。

    据排日账记载，光绪二十六年五月四日（1900年5月31日），“己早晨挑茶乙头上小沱，遇汪顺意兄家卖，未卖，转回家”。五月十一日（ 6月7日），将茶叶售予休宁大连的一位茶商，总计茶叶177斤余，售价英洋29元余。不幸的是，由于某种原因，这位茶商一直没有支付购茶款。于是从此年六月至次年十二月底，允亨频频前往大连催账，但每次至多讨得一、二元，有时甚至空手而回。上海绿茶市场开市的日期，晚于程家出售春茶的时间，因此不能说开市初期茶市的行情，会直接影响到徽州地方茶市。不过上海茶商对市场的基本判断，会辗转影响到徽州茶市，则不无可能。毕竟，茶叶在一段时间内找不到买主的情形，是程家此前从未遇见过的问题。而且茶叶出口量的下降，也可能给茶市带来震荡。最后买入程家茶叶的吴发祥，是大连人，程家此前对其为人应有一定了解。如果他是一个经常赖账的人，程家应会有所耳闻。因此，此人可能受到茶市波动的影响，本身也折了本，因而无力偿付购茶款。这笔茶款的金额看似不大，不过对当时负债累累的程家来说，却事关自身的资金周转和借贷信用。无论如何，最终悲剧还是发生了。

    此外，允亨自身的消费习惯，也给家计带来一定的压力。对比允亨与发开的排日账，允亨似乎不如发开节俭。他不时请朋友打平伙。他还有饮酒的嗜好，平日经常到食杂店买酒买菜。笔者观察到，分家后允亨买酒的次数似有变化，特别是到了光绪后期，经常买酒喝（参见第八章）。在生计逐渐恶化的时期，这种嗜好无疑会增加开支压力。

    总体而言，程家家计危机的出现，主要原因不在于茶款没有着落导致的资金紧缺问题，而是经过数年的累积，程家举债的数额已经达到难以偿还的危险境地，即使在正常的年份，他们也丧失了偿清债务的能力。而这些债务的产生，并非由于国际的、全国性或区域性的政经变动，而是由于两三场人生礼仪，尤其是娶亲的昂贵开支。假如程家将娶亲时间推迟几年，他们还会借入这么大笔的债款吗？未必。但是我们能说，这场悲剧纯粹是因为允亨个人决策的错误？也许不能这么说，毕竟影响程家生计的米价、茶价波动，是受到区域性乃至全国性的市场影响的。因此，在这场灾难中，包括米价上涨、茶价稍有下降、山货逐渐枯竭在内的经济局势，加上义和团运动带来的短时段的政经局势，以及允亨的个人嗜好及作为家长做出的决策，都在这种灾难的发生过程中扮演了一定角色。

    危机的应对

    光绪二十六年十月的危机，似乎来得有些突然。事发七天前，允亨还在家中筹办一场酒席，并请人前来“做伙头办碗”。次日，接女婿，请来几位亲友吃酒。这似乎是允亨长女的出嫁酒。十九日至二十三日，允亨如常砍柴、休息。然后到了二十四日，便发生了债主带人抬走他家中猪的事。但继续往回看，我们发现，九月十五日，允亨就以 10元的价格，当出了一处田产（参见附录六）。那位债主很可能是了解到程家债台高筑、屡次讨债未果后，才带人抬走他的猪的。

    危机发生后，允亨似乎有些震惊，接下来的两天内，他没有采取任何行动，似乎不知如何应对。二十四日，排日账只交代“己在家嬉”，又记录“欠少云先生娘来取账，旺成经手，带鸟人（鲸）〔掠〕玉猪去”。后来他在一张纸条上交代，带人前来讨债的债主是巧娇嫂，而抬走猪的是一位“烟鬼人”。次日写道，“己在家里事，欠账难身”。终于，十月廿六日，也即危机发生后的第三天，程氏父子委托本族的程敬敷和好友余添丁前来清理债务。当日，他们俩“到余架家、余竹孙家二家账项，了通无阻”。后面这两位是程家的债主，允亨大概请敬敷、添丁去商讨债务事宜。他们还拟了一份程家债务清单，这份清单夹在光绪二十六年排日账内，保存至今：

    借来账项人员述后：
    启架兄家：
    癸巳八月廿乙日借来英洋贰拾元。有顿底田皮约乙纸。
    乙未正月十五日借来英洋拾元。有庄下田皮约乙纸，又加拾贰员。
    六月初七借来亦洋拾柒元。有牛栏田契乙纸。
    祝孙兄家：
    戊戌五月廿九日借来亦洋拾伍元。有顿底田皮约乙纸。
    己亥五月廿八日借来英洋叁元，又利洋贰元。三共贰十元正。
    兴良兄家：
    戊戌七月十七日借来亦洋拾伍元。有顿底田皮约乙纸，中见胞兄。
    素从祠：
    己亥六月六日借来亦洋拾元。有庄下田皮约乙张。
    培掘祠：
    戊戌五月初八日借来亦洋拾伍元。有顿底田皮约乙纸。
    万青兄：
    戊戌二月初乙日借来英洋伍元。有牛栏田契乙纸。
    成林祠：
    甲午三月十六日借来英洋拾贰元。有牛栏田契乙纸，中见胞兄。

    根据这份清单，程家借贷的重要账款共10笔，最早的是光绪十九年（1893）的一笔债款，最晚的是光绪二十五年（1899）的债款，其中光绪十九年借入20元，光绪二十年借入 12元，光绪二十一年借入 22元，光绪二十四年借入50元，光绪二十五年借入 20元，所涉债务共 124元，约当这一年程家茶叶销售毛收入的4倍多。

    为偿清债务，程家采取了一系列措施。首先，十月二十七日，“出当青（布）三丈零七寸，又白布三丈八尺零八寸，又青布三丈五尺零贰寸，托兴娥嫂出当英洋贰元正”。同时，“又去英洋二员上素从祠利，掉字乙纸，（伏）〔复〕写一纸，写屋契字一张，付素从祠”。素从祠是清单所列债权人之一，程家借入 10元，此次除支付利息外，还重新立契，以房屋抵押，估计通过这个办法，取回了此前抵押的庄下田皮契。其次，十月三十日，“己同儿托余添灯兄、敬敷弟卖池鱼卅六斤，每洋四斤，计英洋八员，（低）〔抵？〕账”。将鱼塘养的鱼出售，得价8元抵债。再次，十一月初一日，出售顿底、庄下田皮二处，筹得英洋80元。初五日，又支银 5元还培拙祠（应即账单中的培掘祠），将顿底田皮契赎回，同时将菜园一处抵押给该祠，计价10元。初十日，大概账目基本处理完毕，请余添丁吃酒。

    排日账中夹了一张纸条，上面交代了程家出售田皮等物业、财产的详情，很可能是允亨在料理账目的过程中写下的：

    光绪二十六年十一月初乙日，巧娇嫂倩烟鬼人抢去猪乙口，因身该欠账项甚多，只得向家兄及瑞弟商情，将顿底併庄下贰处田皮共八秤，卖与余慰农兄家，计英洋捌拾员，支洋陆拾员还慰农兄，账项清讫。支洋拾贰元还兴良兄，帐目清讫。支洋伍元还培拙祠，下欠拾员，将门口前菜园押在祠内生殖。支洋柒元还巧娇婶，将猪乙口抵英洋陆元五角。又将塘鱼叁拾乙斤抵英洋柒元五角，三共还贰拾乙元，清讫。支洋贰元还素从祠利钱，下欠英洋拾元正，将身住屋当与祠内，长年加贰行息。

    这份文件交代的信息，远不止于出售田皮，还包括前面提到的卖鱼等信息。出售田皮得到的 80元中， 60元是用于向买主还债，实际仅收到20元现金。然后偿还兴良 12元（上面的清单欠 15元）。程家共欠巧娇21元，猪估价 6.5元，鱼售价得 7.5元，另付 7元，偿清了债务。此外就是需要偿还几个祠堂的欠款，培拙祠欠款是15元，付还 5元，另欠 10元以一块菜园做抵押；素从祠欠款是10元，以房屋做抵押，这一点前面已谈到。对照前面的清单，程家还需偿还启架47元，万青 5元，成林祠 12元，共计64元，仍是一笔不小的欠款。

    经过这场危机，程家无疑已经元气大伤，经济状况濒临破产。允亨自身似乎深受打击。十月二十九日，他在账中写到：“己在家事体多端。”十一月阴雨天气多，他常在家中休息。十二月，他接连生了八天病。手头拮据，他没有找医生诊治。十二月二十六日，他再进大连找吴发祥讨债，仍是一无所获。尽管如此，他也试图恢复正常生活。他继续参与劳动，上山砍柴、帮人扛木材。十一月二十三日，他托敬敷出燕山买来小猪一头，要价 2.8元，他手头没钱，买猪的钱只能暂时先欠着。

  • 拱玉书：楔形文字文明的特点

    就字面意义而言，两河流域文明或美索不达米亚文明就是发生在两河流域的文明。这个定义只指出了文明发生的地点，只回答了“在哪”的问题，没有涉及这个文明的突出特点。这个文明的突出特点是什么？我认为是文字，即楔形文字。如果根据一个文明的特点来给这个文明下个定义，那么，我现在谈及的这个文明应该叫楔形文字文明，即用楔形文字记录语言以储存和传递信息的文明。这个定义可以摆脱地域束缚，把地理上不属于两河流域、却使用楔形文字记录自己的民族语言、因而属于楔形文字文化圈的古代西亚地区的文明都囊括在内。“书同文”是这个文明的最显著的特点，也是最大“公约数”。因此，我首先从文字谈起。

    一、“书同文”。“书同文”就是用同一种文字书写，上古时代的整个西亚地区几乎都用或曾用楔形文字书写，因此可以说，他们“书同文”。但他们的“书同文”只是一种表象，与中华文明中的“书同文”貌合神离。貌合是说，从表面上看，无论是对以古代两河流域为中的西亚地区而言，还是对中华文明而言，“书同文”都意味着在一个跨行政区、甚至跨国界的广大地区使用同一种文字，西亚上古时代的大部分族群都曾使用楔形文字，中华文明使用汉字，此所谓二者貌合。神离是说，楔形文字书写的语言非止一种，而汉字书写的语言只是汉语一种(指在中国境内)。

    两河流域(底格里斯河和幼发拉底河)南部是楔形文字的发祥地。早在公元前3200年前后，苏美尔人就发明了楔形文字，并用它来记录自己的民族语言苏美尔语(苏美尔人把自己的语言叫作eme─gi7“土著语”)。早在公元前2700年前后的早王朝时期，苏美尔人在用楔形文字书写苏美尔语文献的同时，时而也用楔形文字书写阿卡德语文献。阿卡德王朝时期(公元前2334—前2154年)，阿卡德语成为官方语言。在此后的一个多世纪里，除一些文学作品外，几乎所有文献都用阿卡德语书写。由于楔形文字是为苏美尔语发明的，所有独体字(从发生的角度观察)都在形式上是象形字，功能上是表意字，有时兼用来表音(节)，所以，用这种文字体系表达(或记载)苏美尔语不成问题，但表达阿卡德语时却显得蹩脚。于是，这时的书吏对楔形文字的使用方式进行了改革：一、多数表意字不再用来表意，而是用来表音，即表音节；二、弃用大部分表意字，只保留一部分表意字的表意用法。这种改革改变了楔形文字的性质，使楔形文字从表意文字(logographic writing)变成了音节文字(syllabic writing)。不论是作为表意文字的楔形文字，还是作为音节文字的楔形文字，其中的任何单字，不论是独体字，还是复合字，都不能只表辅音，不表元音，而必须是表达音节，或元音—辅音式音节，如in、ap等，或辅音—元音式音节，如ba、ti等，抑或辅音—元音—辅音式音节，如tam、?ul等。公元前14世纪，地中海沿岸的乌迦里特出现了楔形字母文字，30个符号分别代表30个辅音，如b、d、?、t等，其中的27个字母是基本字母，3个字母属于附加字母，只用于一些特殊场合，例如用来表达外来借词。到了公元前6世纪的古波斯时期，在国家权力的干预和组织下，在传统埃兰楔文的基础上，波斯人治下的埃兰书吏创造了一个由36个音节符号、5个表意符号组成的文字体系，这个文字体系是在很短的时间内，专门为古波斯语量身打造的。在形式上和功能上，这套楔形符号体系与“字母文字”几乎没有区别，绝大多数学者认为，这36个符号中的任何符号，都不代表语音的最小单位语素(phoneme)，而代表音节(syllable)。我的看法不同，我认为这套符号体系是字母+表意的混合文字体系(下面将说明理由)。这套符号体系与此前的阿卡德(包括巴比伦和亚述)音节文字和埃兰音节文字都有很大区别。最大的区别在于用字量，阿卡德—巴比伦—亚述音节文字体系用字数量约600个符号，书写中埃兰语和新埃兰语的音节文字体系用字量约120个符号，而用来书写古波斯语的符号体系只有36个“音节”(实为字母)符号，加上5个表意符号，加起来不过41个符号。不论古波斯时期创造的这套文字体系属于字母文字，还是属于音节文字，这套文字体系在人类文明史上都是划时代的创新。

    楔形文字的使用范围不限于两河流域，埃兰和古波斯帝国的统治中心都不在两河流域，曾借用楔形文字的赫梯人所处的位置更是与楔形文字发祥地的苏美尔相去甚远。公元前2500—前2400年间，楔形文字西传到了叙利亚地区，那里的埃布拉(Ebla)古国接受了楔形文字，开始用楔形文字记录自己的民族语言——埃布拉语(Eblaite)。至于埃布拉语属于西塞姆语还是东塞姆语，在学术界仍有争议；但确定无疑的是，它更接近古阿卡德语。在埃布拉语中，双音节或三音节词汇居多，不适合用表意文字表达，于是，埃布拉人把以表意为主的苏美尔楔形文字改造成为以表音(节)为主的音节文字，这与稍后的阿卡德帝国的做法是一样的。不过，目前还不能确定，究竟是阿卡德人效法埃布拉人，把苏美尔人的表意文字体系变成了音节文字体系，还是恰恰相反。两个族群所操的语言十分接近，在政治舞台上活跃的时间也大致相同，二者在文字方面的创新应该不是平行而独立的，更不应该是巧合，而是二者之中一个是创新者，一个是借鉴者。在阿卡德人统治时期，两河流域东边的埃兰人也接受了楔形文字，用来书写与达罗毗荼语(Dravidian)有关联的埃兰语。公元前1500年前后，小亚细亚的赫梯人也开始借用楔形文字来书写自己的民族语言——属于印欧语系的赫梯语。地中海沿岸的乌迦里特人于公元前14世纪甚至发明了楔形字母来书写属于西塞姆语的乌迦里特语。这套字母包括30个辅音字母和一个隔字符。

    可见，古代西亚地区的“书同文”是真实的，但这种“书同文”只流于表面，背后的实际情况是：在“书同文”过程中，楔形文字经历了三次脱胎换骨的根本变化，第一次变化发生于公元前2400年前后，从表意文字体系发展出音节文字体系(或音节—表意体系)；第二次变化发生于公元前14世纪，在音节文字的基础上，地中海沿岸产生楔形字母，即乌迦里特字母(30个辅音符号)；第三次变化发生于公元前6世纪的古波斯帝国，在埃兰音节文字的基础上产生古波斯楔形字母+表意字的混合文字体系，36个字母+5个表意字。第一个在波斯波利斯(Persepolis)完整而准确地临摹古波斯语铭文的尼布尔(Karsten Niebuhr，1733—1815)在完全读不懂铭文的情况下，仅凭直觉判断，认为书写古波斯语的楔形文字是字母(Buchstaben)文字。德国的格罗特芬(G. F. Grotefend，1775─1853)是第一个成功解读古波斯语铭文的人，而他是把这种文字当作字母文字来解读的，因而获得成功，例如，他把书写“大流士”的7个符号解读为d─a─r─h─e─u─sh，显然，在格罗特芬看来，这七个符号就是七个字母，代表语音中的最小单位。从20世纪50年代起，学术著作中的古波斯字母表都成了音节表，a、i、u、ka、ku、ga、gu等等。专门研究古波斯语语法的美国宾大教授肯特(R. G. Kent)认为，每个辅音都自带一个“固有”(inherent)的元音。他一边这样认为，一边又将(仅举一例)“我是大流士”音译为adam：Drayavau?，而不是adama：Drayavau?a，这令人费解。依我浅见，古波斯的这套文字体系属于字母+表意字的混合文字体系，36个字母+5个表意字。在36个字母中，除三个元音(a、i、u)字母外，其余都是辅音字母，不自带“固有”的元音，元音需由阅读者根据语言中的正确形式自行添加。很多(如果不是全部的)文字体系，包括这套古波斯文字体系，都是为某种特定语言发明的，更是为以那种特定语言为母语的人发明的。就古波斯的这套字母而言，只要波斯人掌握了这套辅音字母的发音，就能正确地书写和阅读，也就是说，这套字母文字体系具有与生俱来的助记性质，不完全表达语言。

    楔文的上述变化代表了人类历史上出现的三种主要的文字类型：表意文字、音节文字和字母文字。这三种类型产生的先后顺序是先有表意文字(公元前3200年前后)，若干世纪后产生音节文字(公元前2400前后)，再过千年后产生字母文字(公元前14世纪)，但这不代表文字由低级向高级的发展，更不是文字发展的三阶段。这三种文字类型没有高低之分和优劣之别，它们都是为适应各自所表达的语言的需要而产生的，都是原配语言的完美的可视符号。它们有各自的产生途径和发展规律，它们之间的关系不是取代关系，也不是晋级关系，而是互不干扰、平行发展、各走各路的关系。音节楔形文字产生后，作为表意的楔形文字并未退出历史舞台，而是继续使用。乌迦里特楔形字母产生后，很快就消失了，这也不是字母文字本身的错。古波斯时期的楔形字母+表意字的混合文字体系也很快走完了自己的生命历程。这也不是说这种文字体系本身多么不好而一定短命。某种文字体系的终结往往不是文字本身的原因，而是另有原因。

    楔形文字的种种变化都发生在公元前。从楔文产生的公元前3200年前后，到公元前1世纪，公元前的这最后三千年见证了楔形文字本身的种种变化，包括楔形文字被多个古代民族借用来书写自己的民族语言。上古时代的整个西亚地区族群复杂，政治风云变幻莫测，文明周期相对较短，究其原因，其中有地理原因，这里是欧亚非的交汇点，也是各文明的汇聚点，民族交融和交锋从古到今一直在上演。除这个原因外，可能还存在一个重要原因，那就是，在这个地区，始终没有出现一个在人口数量上具有绝对优势、在文化上足够优秀、文化认同感足够强烈，以至于可以由此产生巨大的文化凝聚力、长期立于不败之地的主体民族(或族群)。

    “书同文”本来可以带来文化上的凝聚力，但由于古代西亚的情况是同文不同语，同文不同种，所以，这种“同文”没有给这里的文化带来凝聚力，也没有给这里的人带来文化认同感。中华文明中的“书同文”是国家推行的政策，具有明确目的，那就是维护大一统，本身自带凝聚力和向心力。古代西亚地区楔形文字文化圈的“书同文”，是后进文化为保持自身文化的延续和发展而采取的拿来而后进行改造的措施，目的是为了在一种强势文化中保留自己的语言和文化，本身自带离心性，即脱离先进文化或至少与先进文化保持平行而不被完全融合或同化的离心性。

    二、这个文明的另一个特点是尊同神。苏美尔人创造的或尊崇的各种神灵也被后来的不同族群所崇拜。苏美尔人尊崇的天神安(An)、“风”神恩利尔(Enlil)、智慧者恩基(Enki)、月神楠纳(Nanna)、战神和爱神伊楠娜(Inanna)、太阳神乌图(Utu)等等，也都是后来的阿卡德人、埃布拉人、巴比伦人和亚述人尊崇的神。多神崇拜始终是楔形文字文明的唯一宗教形式，这个文明的意识形态深深植根于多神崇拜。中巴比伦后期，即公元前1200年前后，开始出现独尊一神的倾向，但一神教始终没有能够打破多神崇拜的传统。很显然，楔形文字文明在宗教方面缺乏创新，或可谓守成有余、创新不足。

    楔形文字文明中各族群崇拜的神绝不限于上面提到的几个或多个自然神，戴梅尔在1914年发表的《巴比伦万神殿》里罗列了3300个神的具体名称，在1950年的第2版中，神的数量增加到5580个，去掉重复的，仍有5367个，这还是仅限于巴比伦尼亚地区，不包括其他地区。舒鲁帕克遗址出土了很多早王朝时期(约公元前2500年)的神表，其中最大的一块神表泥版记载了560个神的名字，这些神都是苏美尔人崇拜的神，至少神的名字是苏美尔语，不包括名字属于非苏美尔语的神。一般说来，每个城市都有一到两个保护神，国王有自己的个人保护神，大概普通百姓也有自己的保护神，至少官员或社会名流如此。拉迦什出土的早王朝时期的文献常提到与邻邦发生冲突，也常提到冲突一方的主神对冲突另一方国王的某种行为不满，于是发动战争，为神而战，胜利也属于神。虽然国王们常常打着神的旗号发动战争，但针对的都不是对方的神，而是人。

    神有等级，有大神，有小神，大神中还有等级，上面提到的神都是大神中的大神。不论是大神还是小神，神之间不存在仇恨，也不存在神之间的相互杀戮，《创世神话》中的神间大战发生在造人之前，与人间没有关系。人间的城市(国家)都有保护神，保护神的地位有高有低，但每个城市(国家)的政治、经济以及宗教地位并非取决于保护神的地位。尼普尔是例外，这里是众神之父(ab─ba─dingir─dingir─ré─ne─ke4)恩利尔的崇拜地，是苏美尔人的宗教中心，取得霸权的国王通常要到这里为恩利尔建立神庙或修缮神庙，为自己的统治或霸权营造合法性。这个所谓的宗教中心是个政权更迭的见证地，是君王政治表演的舞台，与普通百姓的信仰没有关系。在历史文献中也不乏某国之神奉恩利尔之命向另一国开战的例子，如拉迦什向温玛宣战被视为“宁吉苏神，恩利尔的战士，遵(恩利尔)正义之命，与吉萨(温玛)开战”。可见，一个神对某一城市(国家)而言是保护神，而对其他城市(国家)而言可能是威胁和灾难。多神崇拜的宗教信仰和一城一神(有的城市不止一神)的实际操作把历史上、文化上以及宗教等方面都高度认同的同一族群从精神上和物理上分割开来，在精神上和物理上都给这样的族群赋予了潜在的离心力，带来了分裂隐患。多神崇拜应该是楔形文字文明逐渐衰败而最终走向消亡的原因之一。

    三、求一统也是这个文明的特点之一。大一统始终是有抱负的统治者的追求目标。乌鲁克早期文明(即公元前3200年前后)时期的政治大势目前尚无从知晓，早王朝时期(约公元前2800—2350年)的天下大势趋于明朗，这个时期城邦林立，战争频繁，城邦间常常相互攻伐，争夺地区霸权。公元前2330年前后，萨尔贡(Sargon)征服各邦，以阿卡德为都建立统一帝国，统治范围包括西至地中海、南到波斯湾的广大地区。这种统一局面仅仅维持了一个多世纪，之后很多传统的独立城邦就纷纷独立，这时又遭到古提(Gutium)人入侵，以两河为中心的广大西亚地区进入古提人统治时期。由于古提人留下的历史铭文极少，现代学者对这个时期的了解十分有限。根据《苏美尔王表》的记载，古提人的统治历经21王，享国91年零40天，而后遭到乌鲁克人图黑伽尔(Utuhegal)领导的苏美尔联军的驱逐，乌鲁克恢复独立，其他地区的传统城市(国家)也都恢复独立。乌尔娜玛(Urnamma，公元前2111—前2094年)很快把这些城市(国家)又统一在他的治下，建立了中央集权制国家，现代学者名之曰乌尔第三王朝，盛极一时。但仅仅历经五王便亡国，末王被俘往埃兰，两河流域再度陷入分裂，这种局面持续大约两个世纪。此后，汉穆拉比(Hammurapi，公元前1792—前1750年)建立统一帝国，享国约一个半世纪，于公元前1600年前后，灭于赫梯王穆尔什里一世(Mur?iliI)之手。赫梯人没有统治巴比伦尼亚的意图，班师回国。凯喜特人(Kassites)趁虚而入，取得巴比伦尼亚的统治权。凯喜特人既不是塞姆人，也不是苏美尔人，其语言归属问题至今悬而未解。凯喜特人不但接管了前朝天下，还继承和发扬了巴比伦人的文化传统，建立了稳固的政权，历经36王，享国近400年，从公元前1530年到前1155年。凯喜特王朝灭亡后，经海国第一王朝和伊辛第二王朝，西亚地区再次统一，这次是统一在亚述人的统治下，现代学者称这个时期为新亚述时期(约公元前1000—前625年)。公元前7世纪，权力中心又南移到巴比伦尼亚的迦勒底王朝(公元前625—前539年)。公元前539年，波斯人占领巴比伦，两河流域的历史进入波斯人统治时期，即古波斯时期(公元前539—前331年)。之后是亚历山大大帝(公元前336—前323年)的短暂统治。亚历山大去世后，西亚地区再次陷入分裂，在塞琉古统治时期，苏美尔书写传统一度在两河流域南部的文明发祥地乌鲁克复兴。目前发现的最后一块楔文泥版属于公元74年。至此，楔形文字文明彻底成为历史。

    纵观楔形文字文明的整个发展、衰亡的历程可以发现，统一可以实现，但不可持续，原因很多，其中一个重要原因是参与这个文明的族群众多，但没有一个主体族群，即没有一个人数足够多，文化足够强，任何人也打不倒，即使一时倒下，也能再度复兴的主体族群。这是这个地区不断出现统一、分裂、再统一、再分裂，朝代不断更替、权力频频易主、传统逐渐丧失、文化一再受到冲击而最终彻底消亡的重要原因。如果说在楔形文字文化圈中哪个族群在一定程度上可称得上主体族群，那一定是苏美尔人，他们最接近“主体民族”的标准，他们发明了文字，创造了一套宗教体系，在文学艺术和科学技术方面也取得了卓越成就。他们的文明延续千余年，可谓千年不倒(从公元前3200—前1800年)，在倒下后的近两千年里影响仍在。到了纪元前后，这个文明才彻底消失。不可思议的是，这个曾经引领世界千余年的文明消失得如此彻底，以至于“苏美尔”和“苏美尔人”在希伯来《旧约圣经》和西方古典时期的著作中没有留下一点痕迹。没有近现代的考古发掘和文献学家的努力，就没有苏美尔文明的再现和复活。

    四、最后谈谈宽容性。时代的变迁和朝代的更替往往都是在血雨腥风中实现的，即使是邻邦之间争夺土地或水源也会杀得尸横遍野。在历史文献中，很多君王极力鼓吹他们杀敌、洗城的功绩，到了新亚述时期，这种鼓吹更是达到登峰造极的程度。文献中的鼓吹也许就是现实中的真实。毋庸置疑，残酷性和血腥性是战争的常态。但也有少数例外，从这些例外中可以看到一些人性的光芒，值得了解，也值得借鉴。

    早在公元前2800年前后，巴比伦尼亚北部的基什(Ki?)国王阿伽曾率军南下，包围了两河流域南部的乌鲁克。乌鲁克国王吉尔伽美什率众应敌，不但战胜强敌，还俘获敌军的亲征国王。然而，吉尔伽美什没有加害于这位来犯国王，而是让他安全地重返家园。不论出于什么理由和目的，这都是人性善良一面的体现，都是一种包容和宽容。自身强大，战胜敌人，然后原谅敌人，宽容敌人，化敌为友，这是强者的自信，也是强者的智慧和善良。吉尔伽美什被视为古代君王的典范，一定与他的强大、智慧、善良和宽容有关。《吉尔伽美什与阿伽》歌颂的正是他的这样品质。

    古波斯时期的居鲁士(公元前559—前530年)更是把强者和宽容演绎到了极致。公元前539年(一说前538年)，居鲁士的军队占领巴比伦。对巴比伦人而言，波斯人是外族，历史上的外族入侵都是血腥的，阿卡德帝国、乌尔第三王朝建立的帝国以及古巴比伦帝国都是在外族入侵中灭亡的，他们遭到的打击是毁灭性的。然而，居鲁士对巴比伦人却采取了怀柔政策，尤其在宗教方面，居鲁士展现了包容和宽容，这让巴比伦人感激不已。于是，巴比伦书吏撰文赞美居鲁士的功德，他们把铭文写在一个腰鼓形的泥质载体上，这就是“居鲁士圆柱铭文”。铭文不但讲到居鲁士允许尼布甲尼撒统治时期的“巴比伦之囚”返回自己的家园，还讲到居鲁士采取的其他宗教包容政策：把以前被运到苏萨的属于“苏美尔和阿卡德”的神像都毫发无损地送回原神庙。按照苏美尔、巴比伦以及亚述的传统，毁掉一座城市，一定要毁掉神庙，毁灭神像，或把神像作为战利品掠走。居鲁士不但没有这样做，还使那些以前被运到苏萨的神像物归原主，这对巴比伦人而言是莫大的恩惠和宽容，所以，巴比伦人感恩戴德，作文盛赞恩主。居鲁士是一代枭雄，是大征服者，占领巴比伦后不久就去征服马萨盖特人，并战死沙场。可以说，居鲁士对巴比伦人采取的怀柔和宽容超乎寻常。居鲁士为什么唯独对巴比伦人采取了怀柔和宽容政策？也许是出于对先进文化的尊重或敬畏！巴比伦人的悠久历史以及文化、科技(尤其是天文学)、文学等方面的优势世人有目共睹。从《居鲁士圆柱铭文》可知，居鲁士自称马尔都克(Markuk)神是“我的主人”(EN─ia)。马尔都克是巴比伦人的主神，征服者信奉被征服者的主神，这是信仰认同，也是文化认同。征服者认同被征服者的文化和宗教，说明征服者有接受先进文化的意愿和情怀，更说明先进文化自带一种威力，一种同化后进文化的威力。

    本文转自《世界历史》2023年第5期，有节略

  • 欧阳晓莉：两河流域文化元素在古埃及前王朝时期的发现

    古埃及文明在发展之初就受到了西亚的影响。公元前6500—前6000年间，地中海东岸黎凡特地区连年干旱，致使部分居民迁徙到埃及。他们把早已驯化的大麦、小麦、绵羊和山羊带到埃及，从此揭开了当地农业革命的序幕。前王朝时期(对应涅迦达文化，约公元前4000—前3000年)，在上埃及河谷地区从希拉康波利斯经涅迦达到阿拜多斯这段不足250千米的尼罗河两岸，古埃及文明的火种得以点燃并最终以燎原之势扩散到上下埃及全境。正是在涅迦达文化的中晚期，来自古代两河流域的文化元素在古埃及崭露头角。

    一、图案、滚印与青金石

    古代两河流域对古埃及文明影响的讨论始于1923年的一篇新闻报道：著名考古学家皮特里(Flinders Petrie，1853-1942)认为卢浮宫博物馆收购的一件文物——戈贝尔?艾尔—阿拉克刀柄(Gebel el-Arak Knife Handle)——证实了埃及王朝的创立者，即所谓的王朝人种，来自两河流域南部的苏美尔地区和伊朗西南部的苏萨。该刀柄象牙材质，长约25.5厘米，宽约4.5厘米，可能属涅迦达文化III期(约公元前3200—前3100年)。皮特里基于这一刀柄对古埃及历史起源的阐释早已过时，但刀柄正面上方图案中一名成年男子双臂分别搏击两头狮子的主题毋庸置疑来自两河流域，男子连面蓄须、头系宽边发带、上身赤裸、下身着过膝长袍的形象，同样具有显著的两河流域苏美尔文明的特色。此图像的两河流域风格是如此之明显，以至于有学者提出刀柄乃两河流域的工匠在埃及所制。

    同样的驯兽者主题还出现于希拉孔波利斯第100号墓的彩绘壁画。虽然墓室的建筑特点指向涅迦达文化II期(约公元前3500—前3200年)，但壁画风格属于III期。其中一幅画面有一位通体红褐色、腰系白带的男性用双臂分别与两头站立的狮子做搏击状。这被认为是古埃及艺术借用外来元素的最早案例。埃及同时期的本土图像艺术更具自然主义的风格，这类外来主题与之相比具有一定独特性。

    除搏击猛兽的驯兽者外，其他来自两河流域的图像主题还包括：长颈猫科动物、带翅膀的狮身鹰首兽、盘绕花朵的蛇以及行进中的动物行列。在那尔迈调色板这一埃及前王朝时期最著名的文物上，就出现了一对长颈猫科动物的形象：它们的长脖互相交缠，其间的凹槽形成一个调色碟，颈部靠头的位置各系着一根绳，都由一位男性拉住。一对高度相似的长颈缠绕的动物形象同样出现在两河流域乌鲁克时期的一枚碧玉滚印(cylinder seal)之上。那尔迈调色板也出土于希拉孔波利斯，在一处神庙遗址掩埋宝藏的地方被发现。传统解释认为它表现了上埃及国王那尔迈征服下埃及并最终统一埃及全境的进程，但更新的学说强调浮雕反映的并非上下埃及统一的特定历史事件，而是象征着国王率军打败敌人的一般化战斗过程。

    上述传播至埃及的图案在两河流域的重要载体是滚印，它是两河流域独创的印章形制，目的在于使印章图案及铭文在泥球或泥板表面的面积最大化。在埃及迄今为止已发现了20枚左右两河流域风格的滚印，其中部分可能是出自埃及工匠之手的仿制品。涅迦达遗址有两处墓葬，一处同时出土了滚印和青金石，滚印图案类似于两河流域南部拉格什和伊朗西南部苏萨所发现的滚印的图案；另一处墓葬中的滚印，其图案类似于两河流域南部的乌尔和北部的高拉(Tepe Gawra，摩苏尔东北20余公里)以及叙利亚境内布拉克遗址(Tell Brak)所出土滚印的图案。上述墓葬中的滚印年代都约为涅迦达文化II期的中晚阶段。

    青金石质地的印章是两河流域滚印中最昂贵的种类之一。在公元前4千纪的西亚和北非，青金石最可能的来源是阿富汗东部的巴达赫尚省(Badkhshan)。在古埃及前王朝15000处左右墓葬中，约167处出土了青金石(比例略高于1%)。考虑盗墓以及早期发掘报告欠完备的因素，以青金石为陪葬品的墓葬比例应该更高。但到了古埃及第一王朝时期(约公元前3000—前2890年)，青金石的分布范围缩减，仅出现在最顶级的社会精英成员的墓葬中。在随后几个世纪，青金石甚至从古埃及墓葬中完全消失。它再度出现是在第四王朝法老斯奈夫鲁(约公元前2613—前2589年在位)的王后墓中。在本文涉及的时段内，阿富汗东部是唯一已知的青金石产地。因此，追踪考古发现中的青金石有助于重构中亚、伊朗、两河流域和埃及之间的贸易路线。

    在整个埃及前王朝时期，在阿拉伯半岛都未发现青金石的考古遗迹。因此现有模型推断，青金石首先从阿富汗经伊朗高原运抵两河流域，再转运至两河南部居民在叙利亚建立的殖民地(详见下文)，最后经海路从黎凡特北部(如叙利亚遗址杰贝勒?阿鲁达(Jebel Aruda))到达埃及。两河流域最早的青金石实物则回溯至公元前6千纪晚期，是一些发现于北部亚明遗址(Tepe Yarim)的念珠。在公元前5千纪至前4千纪中期之间，两河流域北部垄断了上述陆上青金石之路，仅在高拉一处遗址就发现了500余件青金石的念珠、印章和镶嵌物。虽然青金石原料并非两河流域的特产，青金石文化却首先绽放于该地区并传播到古代近东其他区域。在两河流域，青金石被赋予了非同一般的象征意义，并与神祇和王权密切联系在一起。

    二、“乌鲁克扩张”

    古埃及前王朝对应于两河流域的乌鲁克时期(约公元前3900/3800—前3200/3100年)。此时的两河流域以南部城市乌鲁克为中心，经历了一个社会突变和飞速发展的阶段，学术界称之为“乌鲁克扩张”(Uruk Expansion)。在此之前，两河流域南部与北部、伊朗西南部以及地中海东岸的黎凡特地区，在定居点的绝对规模和定居点相互间的差异上并没有显著不同。但到了公元前4千纪下半期，两河流域南部冲积平原上的政治组织(polity)在整体规模、内部分化程度以及定居点的等级结构上都远超古代近东乃至世界其他地区。进入公元前4千纪晚期，两河流域南部已发展出相互竞争的若干城邦，它们政治上分裂但文化上彼此相近，同时还向外扩张。两河流域南部最早的苏美尔文明从此腾飞，在城市化、社会政治复杂化和经济差异化等方面均领先于世界。

    两河流域南部的经济增长始于公元前5—前4千纪早期，主要贸易品是地方性特产，如羊毛及其纺织品、皮革制品、奶制品、谷物、蔬菜瓜果、亚麻纺织品、各种熏鱼或咸鱼、禽类以及芦苇制品，这些物产分别来自游牧部落、农业定居人群以及生活在底格里斯河和幼发拉底河入海口沼泽地带的居民。第二阶段始于公元前4千纪中期，此时精英阶层的社会意识增强，贸易品生产的专业化程度降低，各群体都利用前一阶段积累的剩余产品和人力资源进行生产，以取代从周边乃至两河流域以外地区的进口。这一使用当地产品替代进口物品的机制促进了经济发展。第三阶段则是公元前4千纪下半期，特点是对外贸易的大幅度增长。此时两河流域的羊毛织物在周边地区大受欢迎，需求强劲。加之驴被驯化为驮兽，其负重至少是人的两倍，大幅提升了长途运输能力。这两个刺激因素在后来公元前2千纪早期两河流域与小亚细亚的古亚述贸易中都有明确证据。

    随着地方贸易和对外贸易的发展，南部乌鲁克的居民开始迁入两河流域周边地区。他们最初组成小规模的移民社群，生活在当地的资源开发或调运中心。他们进而控制了当地的这类中心，将其建设为跨地区的贸易枢纽。还有一种情况就是乌鲁克殖民者白手起家，在一片处女地上建立定居点并沿袭两河流域的社会和城市习俗。判断这些定居点与两河流域南部有关的主要依据是物质文化材料，包括建筑、陶器和雕刻作品的风格，以及陶筹和滚印的使用等。

    最后一类乌鲁克居民白手起家建立的定居点，在考古遗迹中最易辨认，主要坐落于土耳其东南部和叙利亚北部的幼发拉底河河畔，周围环绕着小规模的乌鲁克村落群，为其提供农业和畜牧业产品。最具代表性的遗址是叙利亚境内的哈布巴?卡比拉(Habuba Kabira-süd)和附近规模更小的杰贝勒?阿鲁达，两地距幼发拉底河的传统渡河点迈斯凯内(Meskene)都不远。杰贝勒?阿鲁达是一处乌鲁克晚期新建的定居点，占地略大于3公顷，所在的小山包俯瞰幼发拉底河河谷。它的主要建筑是规模较大的民居，显然是精英人士的住宅，说明该定居点可能是哈布巴?卡比拉的行政中心。杰贝勒?阿鲁达南面约8公里处就是同期大得多的定居点哈布巴?卡比拉。它同样是一处乌鲁克晚期新建的定居点，初期面积约为6公顷，后来扩张加倍，建有防御工事且规划整齐，居住区、产业区及行政区界限分明，显然是人为规划的成果。防御工事外还延伸出一大片居住区，面积最大时达22公顷。

    上述三类遗址——有乌鲁克居民居住的当地定居点、乌鲁克居民控制的当地中心、完全由乌鲁克居民建立的定居点——的功能引发了诸多讨论。它们可能是出于各种原因离开两河流域南部的居民所建立的住所，也可能位于两河流域南部与周边地区进出口产品的商业要道附近，同时还是各种信息和情报的汇聚点。前两类规模较小的定居点在个人或团体的投资范围内，但像杰贝勒?阿鲁达和哈布巴?卡比拉这类新建定居点应该得到了两河流域母邦的财力和人力资助。

    乌鲁克扩张过后，接踵而来的捷姆迭特?那色时期(Jemdet Nasr，约公元前3100—前2900年)是两河流域发展相对缓慢的年代。自公元前2900左右开始，两河流域进入了城邦争霸的早王朝时期。

    三、黎凡特地区的走廊作用

    近几十年间，研究两河流域与埃及早期文明交流的重点在于探讨“乌鲁克扩张”期间两河流域对埃及的影响，以及居间的叙利亚和土耳其的定居点、村落、市镇乃至城邦在两地间的物资流动和文化传播中所发挥的作用。阿拉伯半岛的角色还有待深入探讨，但也不容忽视。

    考古人员在南黎凡特考察或发掘了约40处关乎古埃及历史初期的遗址，它们基本分布在后世古埃及人称之为“荷鲁斯之路”的东地中海沿岸的狭长地带。埃及风格的建筑遗迹和来自埃及的陶器(那尔迈的名字在陶片上最常见)都有发现，还有类似堡垒的设施，其目的可能就是保障通往埃及的贸易路线畅通。位于尼罗河三角洲东部的遗址泰尔—法卡(Tell el-Farkha)被认为是上埃及权贵建立的物品中转站和管理中心，以保证和促进与西亚的贸易。这里出土了大量来自巴勒斯坦和两河流域的器物，还有作为驮兽的驴的遗骸。在阿拜多斯的代表性墓葬U—J墓中(约公元前3200—前3150年)发现了约2000件陶罐，其中仅少数产自埃及本地，大部分则来自今天的巴勒斯坦地区，用于盛装那里出产的葡萄酒；还有用黎巴嫩松木制作的冥器，以及可能来自埃塞俄比亚的黑曜石等坚硬宝石。

    两河流域对埃及的影响也表现在建筑材料和装饰风格上。在公元前3千纪的早王朝和古王朝时期，埃及盛行一种长方形的马斯塔巴大墓(mastaba)，因其梯形体的地上建筑酷似阿拉伯板凳而得名。第一王朝的创建者阿哈(Aha)就在涅迦达为其母建造了巨大的马斯塔巴墓。这类王室大墓往往建有附属的祭庙，祭庙外墙由泥砖砌成且呈现出壁凹式的装饰(niched facade)。带有这类装饰的建筑通常被称为“宫殿正门”，因为它与王宫的出现相关且与王权的关系密切。作为建筑材料的泥砖和壁凹式的装饰风格都类似于两河流域乌鲁克遗址的神庙外墙。这类外墙先后发现于两河流域南部的乌尔和乌鲁克以及北部的高拉，后来又发现于叙利亚境内的乌鲁克扩张时代遗址哈布巴?卡比拉和杰贝勒?阿鲁达(见上文)，其空间分布佐证了乌鲁克的扩张现象和叙利亚在埃及与两河之间的桥梁作用。

    四、两大文明交相辉映

    总体而言，两河流域对早期埃及的影响在物质文化上的表现主要包括青金石、滚印、雕刻艺术中的特定图案主题和建筑装饰上的壁凹式外墙，这些因素在前王朝时期下半段(公元前4千纪下半期)已经出现在埃及，在希拉孔波利斯、涅迦达和阿比多斯三地的墓葬遗存中均有发现，其背后的推动力可能与两河流域南部的“乌鲁克扩张”密切相关。这一时期乌鲁克文化正处在扩张的高潮阶段，两河流域南部的长途贸易发达，进出口产品丰富，居民们或私人筹资，或得到所在城市资助，得以远赴两河流域北部和黎凡特地区建立规模不等的定居点乃至商业殖民地，两河流域的文化也随之得以传播。通过黎凡特这一地理走廊以及从黎凡特到埃及的海路运输，上述特定文化元素到达了埃及。与此相反，在两河流域并未发现来自埃及的文化元素。

    但笔者以为，这一反差并没有强烈到足以宣称两河文明影响了埃及早期文明的进程。无论青金石、滚印还是泥砖，抑或雕刻图案和壁凹式装饰，其绝对数量和空间分布都相当有限，基本局限在精英阶层的墓葬中。与其说它们来自两河流域这点吸引了埃及受众，毋宁说它们的异域风情被埃及的社会上层所借用以传达王权至高的意识形态。此时埃及的社会精英或许在摸索建立权威、彰显等级秩序的不同路径，因而借用了这些来自两河流域的文化要素。

    随着上下埃及的统一和第一王朝的建立，来自两河流域的文化元素在公元前3千纪初从埃及本土一度消失了。埃及与西亚的贸易明显减少，原来派驻在加沙、以色列南部以及努比亚的人员撤回本土，尼罗河第一瀑布和尼罗河三角洲东北部分别成为埃及与努比亚以及西亚之间的界限。埃及文明从此步入自我形塑的时代，以象形文字、金字塔和木乃伊为基本特征的埃及文化其本土特色日益鲜明。同期的两河流域则在“乌鲁克扩张”后进入了发展相对缓慢和地方化倾向愈加突出的阶段，似乎丧失了文化输出的动力和能力。

    纵观古代埃及和两河流域的史前时代，两大文明既各有千秋，又交相辉映。在埃及，当农业和畜牧业发展后，一定规模的定居区域在公元前4000年后才逐步形成。埃及定居点的出现不仅比两河流域晚得多，而且过程缓慢，其城市化也没有后者那样普遍和彻底。它在史前的政治格局起初也朝着两河流域南部众多城邦共生并存的方向发展，但这一趋势不久后便中断，转而向上下埃及统一的国家迈进。而在两河流域，史前的多城邦共存竞争局面一直延续到长达五六百年的早王朝时期(约公元前2900—前2350年)，之后才迎来首个统一南北两部的阿卡德王朝。正因为统治阶层建构权力的模式不同，尽管两个文明之间存在诸多相似之处，但埃及发展成了统一的、以王权为特征的领土国家，两河流域则保持了众多城邦竞争和并存的传统。

    本文转自《世界历史》2023年第1期

  • 许宏：考古学视角下的中国诞生史[辑编]

    从司马迁的记载开始，三代王朝夏、商、周是华夏族的成丁礼，再之前是悠长的婴儿和少年时期，从这个时候开始成熟起来，然后有了一个比较大的王朝国家。但是究竟是夏还是商，现在还有争议，我们看这个表就比较清楚。我们一直以来就存在着历史文献学和考古学两大话语系统，这两大话语系统最初是边界明显的：一边是历史文献上的伏羲、女娲、三皇五帝、夏商周王朝；一边是考古学上的前仰韶、仰韶、龙山、二里头、二里岗时代。这两大话语系统的合流是在殷墟。为什么是在殷墟？有一个绝对不可逾越的条件就是，当时有可以证明自己族属和王朝归属的文字材料出现，这才可以把这两大话语系统整合，以后的西周、东周、秦汉魏晋都可以证明，但在那之前没有文字材料，没有史证。在前殷墟时代，如果我们把考古学遗存跟它的族属、王朝归属相对应的话，都只能是推论和假说。就是因为它没有直接性的文字材料，所以在大的历史分期上，我们习惯于把它分成历史时期（history）——有明确文字记载的时期；原史时期（proto-history）——文字开始出现，但还不足以解决狭义的历史问题；史前时期（pre-history），基本上就是这样一个脉络。

    史前、原史、历史阶段划分与对应史料

    前中国时代与“中国”的初兴

    许宏：考古学视角下的中国诞生史
    华夏文明腹心地区的五颗明珠——五大都邑遗址，都背靠邙山，面向古洛河

    任何事物都有其从无到有，从小到大，发生发展的过程，国家起源以及中国文明的形成也不例外。考古学揭示出的距今五六千年以来的东亚大陆展现了这样的图景。大约距今六千年以前，广袤的东亚大陆上的史前人群，还都居住在不大的聚落中，以原始农业和渔猎为主，过着大体平等、自给自足的生活。各区域文化独立发展，同时又显现出一定的跨地域的共性。到了距今5500～3800年间，也就是考古学上的仰韶时代后期至龙山时代，被称为东亚“大两河流域”的黄河流域和长江流域的许多地区，进入了一个发生着深刻的社会变革的时期。随着人口的增长，这一时期开始出现了阶层分化和社会复杂化现象，区域之间的文化交流和摩擦冲突都日趋频繁。许多前所未见的文化现象集中出现，聚落形态上发生着根本的变化。如大型中心聚落及以其为核心形成的一个个大群落，城墙与壕沟、大型台基和殿堂建筑、大型祭坛、大型墓葬等耗工费时的工程，随葬品丰厚的大墓和一贫如洗的小墓所反映出的社会严重分化等等，都十分令人瞩目。

    众多相对独立的部族或古国并存且相互竞争。如中原及周边的仰韶文化、石峁文化、陶寺文化、王湾三期文化，西北地区的大地湾文化、齐家文化，辽西和内蒙东部的红山文化，山东地区的大汶口文化、龙山文化，江淮地区的薛家岗文化，长江下游的凌家滩文化、崧泽文化、良渚文化，长江中游的屈家岭文化、石家河文化，长江上游的宝墩文化等，在文化面貌上各具特色，异彩纷呈。

    红点是当时邦国中心所在地

    那是一个“满天星斗”的时代，邦国林立是那个时代最显著的特征。有的学者将其称为“古国时代”或“邦国时代”，有的则借用欧美学界的话语系统，将其称之为“酋邦时代”。无论如何，那是一个小国寡民的时代。整个东亚大陆的面积，与现在的欧洲差不多，而当时的这些星罗棋布的古国或部族，也和现在欧洲的样态差不多。那么，问题来了：它们都属于“中国”吗？

    要说清这件事，得先捋一捋相关的概念。关于“文明”的解说五花八门，这里无法详细展开，但说古代文明是人类文化发展的较高阶段或形态，而其标志是“国家”的出现，应会得到大多数人的认可。[……]

    显然，中国有5000年文明史的提法，是把这些都当成了中华文明史也即“中国”诞生史的一部分。其认知脉络是，这些人类群团在相互交流、碰撞的文化互动中，逐渐形成了一个松散的交互作用圈，这也就奠定了后世中华文明的基础。随着1970年代末期以来一系列重要发现的公布，中国在三代王朝文明之前即已出现了城市和国家，它们是探索中国文明起源的重要线索的观点得到了普遍认同。源远流长，单线进化，从未间断，成为中国学术界在中国文明起源问题上的主流看法。

    这当然是有道理的。[……]说中华文明可以上溯到新石器时代甚至旧石器时代的认识，显然出于这样的考虑。但这样无限制地追溯，意义何在？同时，其认知前提是百川归海的单线进化论，而事实果真如此吗？甚而，在不少人心目中，一个默认的前提是，现中华人民共和国境内的古代遗存，理所当然就是中华文明的源头。这样的认识，可以成立吗？

    首先，考古学家观察到的上述许多古国或部族，大都经历了发生、发展乃至最后消亡的全过程，也即它们各自谱写了完整的生命史的篇章，而只是给后起的中原王朝文明以程度不同的文化给养或影响。到公元前2000年前后，它们先后退出历史舞台，在这些人类共同体和后来崛起的中原文明之间，有一个“连续”中的“断裂”。这种断裂究竟是出于天灾还是人祸，原因想必多种多样，学术界还在探索之中。在某些区域，“大禹治水”传说中的大洪水，或许就是原因之一。考古学的研究对象是支离破碎的古代遗存，所以知其然不知其所以然的事，所在多有。

    如前所述，我们知道在现在的中国境内，上古时期曾有众多相互独立的国家并存。而顾名思义，在“国”前冠以“中”字，“中国”也就有了“中央之城”或“中央之邦”的意蕴。这同时也说明“中国”已并非初始阶段的国家，显然，它一定是一个在当时具有相当的影响力、具有排他性的核心。因而，我们也就不能说最初有多个“中国”，作为发达、复杂的政治实体的“中国”也是不能无限制地上溯的。

    许宏：考古学视角下的中国诞生史
    史前时代东亚城址的三大系统

    说到“中国”，还要捋捋这一概念的源起和演化。在出土文物中，“中国”一词最早见于西周初年的青铜器“何尊”的铭文。而在传世文献中，“中国”一词最早出现于东周时期成书的《尚书》和《诗经》等书中。“中国”一词出现后，仅在古代中国就衍生出多种含义，如王国都城及京畿地区、中原地区、国内或内地、诸夏族居地乃至华夏国家等。“中国”成为具有近代国家概念的正式名称，始于“中华民国”，是它的简称；现在也是“中华人民共和国”的简称。其中，最接近“中国”一词本来意义的是“王国都城及京畿地区”，那里是王权国家的权力中心之所在，已形成具有向心力和辐射性的强势文化“磁场”。其地理位置居中，有地利之便，因此又称为“国中”、“土中”或“中原”。

    那么，究竟是什么时候，后世“中国”的雏形或者说“最早的中国”崛起于世呢？

    按古代文献的说法，夏王朝是中国最早的王朝，是破坏了原始民主制的世袭“家天下”的开端。一般认为，夏王朝始建于公元前二十一世纪，国家级重大科研项目“夏商周断代工程”，把夏王朝建立的年代定为公元前2070年左右。在考古学上，那时仍属于龙山时代，在其后约200多年的时间里，中原地区仍然处于邦国林立，战乱频仍的时代，各人类群团不相统属，筑城以自守，外来文化因素明显。显然，“逐鹿中原”的战争正处于白热化的阶段，看不出跨地域的社会整合的迹象。也就是说，至少在所谓的夏王朝前期，考古学上看不到与文献相对应的“王朝气象”。

    与此同时，兴盛一时的中原周边地区的各支考古学文化先后走向衰落；到了公元前1800年前后，中原龙山文化系统的城址和大型中心聚落也纷纷退出历史舞台。代之而起的是，地处中原腹地嵩（山）洛（阳）地区的二里头文化在极短的时间内吸收了各区域的文明因素，以中原文化为依托最终崛起。二里头文化的分布范围首次突破了地理单元的制约，几乎遍布于整个黄河中游地区。二里头文化的因素向四围辐射的范围更远大于此。

    伴随着区域性文明中心的衰落，此期出现了超大型的都邑——二里头遗址。地处中原腹地洛阳盆地的二里头遗址，其现存面积达300万平方米。经半个多世纪的田野工作，在这里发现了中国最早的城市主干道网，最早的宫城，最早的多进院落大型宫殿建筑，最早的中轴线布局的宫殿建筑群，最早的封闭式官营手工业作坊区，最早的青铜礼乐器群、兵器群以及青铜器铸造作坊、最早的绿松石器作坊、最早的使用双轮车的证据，等等。这样的规模和内涵在当时的东亚大陆都是独一无二的，可以说，这里是中国乃至东亚地区最早的具有明确城市规划的大型都邑。

    二里头文化与二里头都邑的出现，表明当时的社会由若干相互竞争的政治实体并存的局面，进入到广域王权国家阶段。黄河和长江流域这一东亚文明的腹心地区开始由多元化的邦国文明走向一体化的王朝文明。作为广域王权国家概念的“中国”，在前一阶段还没有形成。

    要之，我们倾向于以公元前1700年前后东亚地区最早的核心文化——二里头文化，最早的广域王权国家——二里头国家的出现为界，把东亚大陆的早期文明史划分为两个大的阶段，即以中原为中心的“中原（中国）王朝时代”，和此前政治实体林立的“前中国时代”和“前王朝时代”。

    许宏：考古学视角下的中国诞生史
    郑洛地区龙山时代聚落分布（赵春青 2001）

    值得注意的是，这两大阶段也恰是东亚大陆青铜时代和前青铜时代的分野。

    在二里头时代之前的数百年时间里，东亚大陆的多数区域，早期铜器的使用呈现出红铜、砷铜、青铜并存的状况。铜制品多为器形简单的小件工具和装饰品等生活用具，锻、铸均有，制造工艺处于初级阶段，尚未熟练掌握合金比例。如多位学者已分析指出的那样，东亚大陆用铜遗存的出现，应与接受外来影响关系密切。至于东亚大陆部分区域进入青铜时代的时间，依据最新的年代学研究，要晚到公元前1700年前后了。

    考古学观察到的现象是，出土最早的青铜礼容器的中原地区，也是东亚大陆最早出现广域王权国家的地区。青铜礼器的出现和当时的中原社会，都经历了文化交流中的碰撞与裂变的历程。其同步性引人遐思。二者相互作用刺激，导致中原地区自公元前二千纪上半叶，进入了史上空前的大提速时代。早期中国，由此起步。那么，是青铜礼器及其铸造术，催生了最早的“中国”？

    随着二里头文化在中原的崛起，这支唯一使用复杂的合范技术生产青铜容器（礼器）的先进文化成为跃入中国青铜时代的一匹黑马。值得注意的是，这些青铜礼器只随葬于二里头都邑社会上层的墓葬中，在这个金字塔式的等级社会中，青铜礼器的使用成为处于塔尖的统治阶层身份地位的标志。这些最新问世的祭祀与宫廷礼仪用青铜酒器、乐器，仪仗用青铜武器，以及传统的玉礼器，构成独具中国特色的青铜礼乐文明。“国之大事，在祀与戎”（《左传•成公十三年》）。保有祭祀特权与强大的军力，自古以来就是一个国家立于不败之地的根本。从早期王朝流传下来的祭天崇祖的传统，几千年来一直是中国人宗教信仰和实践的主要内容。二里头都城规划中祭祀区的存在，以及以青铜为主的祭祀用礼仪用器，都与大型礼制建筑一样，是用来昭示早期王朝礼制传统的重要标志物。由于军事力量在立国上的重要性，青铜与玉石兵器也成为祭祀礼器和表现身份地位的仪仗用器的有机组成部分。二里头文化青铜礼器产品的使用范围主要限于二里头都邑的贵族。也就是说，二里头都邑不仅垄断了青铜礼器的生产，也独占了青铜礼器的“消费”即使用权。

    其中，酒器是具有中国特色的酒文化乃至它背后的礼仪制度的重要载体。作为统治阶层身份地位的象征，以酒器为中心的礼器群，成为中国最早的青铜礼器群。从这里，我们可以看出中国古代文明主要是建立在社会关系的巨变（在等级秩序下人际关系的大调整）而非人与自然关系巨变的基础上的。而铸造铜爵等造型复杂的酒器，至少需要精确地组合起内模和3件以上的外范，即当时已采用了先进的复合范工艺。克服其中的种种困难，最终铸造出青铜礼器的内在动力，应当就是这一时期新兴王权对宫廷礼仪的整饬。

    二里头遗址发现的青铜钺，是迄今所知中国最早的青铜钺。钺作为象征军事权威的仪仗用器，也是一种用于“大辟之刑”的刑具。甲骨文金文中“王”字的字形，像横置的钺，在最初应指代秉持斧钺之人即有军事统帅权的首领，随着早期国家的出现，逐渐成为握有最高权力的统治者的称号。早于甲骨文时代数百年的二里头都城中出土的玉石钺，和迄今所知中国最早的青铜钺，就应是已出现的“王权”的又一个重要象征。换言之，钺的礼仪化是中国王朝文明形成与早期发展的一个缩影。

    在早期王朝的礼器群中，爵、钺等器种持续兴盛于三代逾千年，甚至成为后世中国社会政治文化的重要符号，个中原因，颇具深意。

    二里头的聚落变迁

    另一个可资观察的角度是都邑的城郭形态。这一问题上的权威观点是，城墙是构成都城的基本政治要素，不存在没有城墙的都城。通过对以先秦至秦汉时期为中心的都城发展历程的初步考察，笔者认为整个中国古代都城史可以依城郭形态的不同，划分为两个大的阶段，即防御性城郭阶段和礼仪性城郭阶段。在自最早的广域王权国家都邑二里头至曹魏邺城前近两千年的时间里，庞大的都邑不设防，有宫城而无外郭城，是都城空间构造的主流，这一现象可以概括为“大都无城”。在二里头、殷墟、周原、丰镐、洛邑、秦咸阳、西汉长安和东汉洛阳等一系列都邑中有清晰的显现。这与广域王权国家强盛的国势及军事、外交优势，作为“移民城市”的居民成分复杂化，对都城所处自然条件的充分利用等，都有一定的关联。处于都城发展史早期阶段的防御性城郭的实用性，导致城郭的有无取决于政治、军事、地理等诸多因素，“大都无城”的聚落形态应即这一历史背景的产物；而后起的带有贯穿全城的大中轴线、实施里坊制的礼仪性城郭，因同时具有权力层级的象征意义，才开启了汉代以后城、郭兼备的都城发展的新纪元。

    在这一早期中国都邑布局的演变过程中，最令人瞩目的是二里头时代的到来，这是“大都无城”传统的肇始。如上所述，二里头遗址是迄今可以确认的中国最早的具有明确规划的都邑，其布局开中国古代都城规划制度的先河。但在逾半世纪的田野工作中，却一直没有发现圈围起整个二里头都邑聚落的防御设施，仅知在边缘地带分布着不相连属的沟状遗迹，应具有区划的作用。

    二里头遗址地理位置

    如果将二里头时代的聚落形态与更早的龙山时代作比较，可知前者最大的变化，一是中心聚落面积的大幅度提升，由龙山时代的10余至数十余万平方米，扩大至300万平方米；二是基本上摒弃了龙山时代普遍筑城的传统，代之而起的环壕成为这一时代的主流防御设施。

    由对考古材料的分析可知，进入二里头时代，聚落内部社会层级间的区隔得到强化，而与此同时，对外防御设施则相对弱化。从聚落形态的角度看，二里头都邑是“大都无城”的一个最早的典范。究其原因，不能不考虑到都邑内的居民。二里头可能是最早集聚了周边人口的中心城市，其人口由众多小规模的、彼此不相关连的血亲集团所组成，这种特征又与其后的殷墟和西周时代的都邑颇为相近。而广域王权国家则是从二里头时代至西周时代社会结构上的共性。以“大都无城”为主要特征的都邑聚落形态与早期王朝阶段社会结构上的关联性，值得进一步探究。显然，“大都无城”，是前中国时代终结、最早的“中国”初兴的一个重要的标志。

    要之，以二里头时代为界，东亚大陆的国家起源进程呈现出非连续性和多歧性。以良渚、陶寺、石峁文明为代表的龙山时代众多区域性邦国文明，各领风骚数百年，最终退出了历史舞台。它们走完了其生命史的全过程，而与后起的中原青铜文明仅有或多或少的间接关系，这就使东亚大陆的国家起源进程呈现出“连续”中的“断裂”的态势。这是我们把东亚大陆国家起源进程划分为两大阶段的重要依据。

    通观东南良渚的水城、中原陶寺的土城、西北石峁的石城，都是因地制宜、适应环境的产物，它们也的确都是区域性文明；这与“大都无城”的二里头形成了鲜明的对比。它们所拥有的“前铜礼器群”还看不到像以二里头为先导的中原王朝礼器群那样严格的礼仪规制尤其是重酒的礼器组合。而以软实力见长的二里头，显然通过社会与文化的整合具有了“普世”的魅力，在众多族群的膜拜与模仿中扩大了自身的影响，其范围远远超出了中原地区。更为重要的是，它的文明底蕴通过二里岗时代、殷墟时代乃至西周时代王朝间的传承扬弃，成为中国古代文明的主流。

  • 葛剑雄：九州的传说和现实

    虽然把“中国”确定为我们整个国家的名称是到19世纪后期才出现的事情，但中国统一的概念却已经存在了三千多年。甚至在中原的统一国家形成之前，政治家和学者已经纷纷推出了各自的统一蓝图。虽然当时还没有一个君主真正能够统治这片广袤的土地，但“溥（普）天之下，莫非王土”的颂歌却在西周时就已经普遍流传，并且被视为真理而接受。

    不过，这首颂歌的作者（或许不止一个）大概不会想到，这种统一观居然统治了中国二千多年，并且到今天还没有消除它的潜在影响。

    在中国儒家的经典著作《尚书》中有一篇《禹贡》，一开始就写道：“禹铺土，随山刊木，奠高山、大川。”意思是说，在洪水横流以后，大禹一面规划治水，一面根据名山大川的分布重新划定区域。接着列出的九个单位是：冀州、兖州、青州、徐州、扬州、荆州、豫州、梁州、雍州，这就是九州。

    葛剑雄 | 九州的传说和现实
    《禹贡》所描述的九州区域图

    在另一篇《舜典》中，又提到在尧、舜时，“肇十有二州”。“肇”是开始的意思。对这句话，西汉的学者谷永和东汉初的学者班固解释为：在尧的时候遭到洪水，全国被大水分割为十二部分。但东汉末年的马融的说法是：舜在大禹治水之后，从禹所划分的九州中又分出幽州、并州、和营州三个单位，因而总共有了十二个州。这一说法获得后世多数学者的赞同。

    从未实行过的九州制

    由于这些记载都出于儒家经典，又得到后世众多学者的肯定，所以从西汉以来就成为不可动摇的定论，几乎没有人表示怀疑。人们一般都认为，从大禹治水开始就有了九州这样的政区，以后又演变为十二州。直到现在，一些在叙述一个地方行政区域的历史时，往往还要从九州讲起，似乎这是历史事实。

    由于全国就分为九州，所以九州又一直被当作全国、“天下”的代名词。如南宋诗人陆游《示儿》诗中的名句“死去原知万事空，但悲不见九州同”，就是取这样的用意；晚清诗人龚自珍“九州生气恃风雷”一句也是如此。

    五四运动以后，学者们向儒家经典提出了挑战。经过反复的争论和研究，历史学界已经把这传统的九州说推翻了。原来，《禹贡》中的记载并不是历史事实，九州也不是中国最早的行政区划。

    《禹贡》虽然托名为大禹所作，其实却是战国后期人的作品。具体的证据很多，最主要的理由是《禹贡》中所记的不少地理状况都是战国时的现象，有的地名和水名甚至要到战国后期才出现，如果真是大禹所作，他岂能未卜先知？而且在《尚书》各篇中，《禹贡》的语言照理应比出现在它以后的《盘庚》（记录商朝中期的君主盘庚迁都事）等篇难懂，事实恰恰相反；这也只能说明《禹贡》问世的时间较晚。

    《禹贡》所讲的内容不符合历史事实，至多只有传说的价值。到目前为止的考古发掘和研究的成果，还只能证实商朝的历史。近年来在河南等地发现的一些文化址，一些学者认为就是属于夏朝。如果这一观点得到进一步的证明和普遍的承认，那末夏朝的主要统治区应该在今河南一带，与文献记载传说中的夏都不超出今山西南部、山东西部和河南的范围是一致的。而《禹贡》所叙述的九州的范围，北至燕山山脉和渤海湾，南至南岭一带，西至陇东高原；至于具体涉及的内容更广；当然不可能是夏朝的事实。

    现有的研究成果足以证明，不仅传说中的大禹时代还不可能有什么行政区划，就是商朝和更后的西周时代也还没有出现行政区划。既然《禹贡》是战国后期的产物，那么九州制是不是当时的制度呢？也不是。大家知道，到战国后期，周天子的权力早已荡然无存，而秦始皇还没有统一六国，七个主要的诸侯国各自为政，又有谁有这样的权威能制定并且实行包括各国的疆域在内的行政区划呢？

    可见，九州制只是当时学者对未来统一国家的一种规划，反映了他们一种政治理想。

    秦始皇在全国推行了郡县制，却没有在郡以上设立州。到了公元前二世末，也就是在《禹贡》问世的一二百年以后的西汉元封五年（前106年），汉武帝将全国除首都附近的七个郡级单位以外政区分属于十三部，即豫州、兖州、青州、徐州、冀州、幽州、并州、凉州、益州、荆州、扬州、交趾、朔方；每部设刺史一人，负责巡察境内的地方官和豪强地主；称为十三刺史部， 简称十三部或十三州。但那时的州还是一种监察区，而且这十一个以州命名的单位中没有《禹贡》九州中的梁州和雍州，增加了凉州、益州、并州和幽州。在公元１世纪后的东汉，州才成为最高一级的行政区域。朔方并入了并州，加上管辖首都一带的司隶校尉部，总数仍为十三。由于交趾改称交州，以州命名的单位就有了十二个，也不是九个。东汉末年曹操曾想按九州来重划政区，却没有成功；从此再也没有人作过这样的尝试。从这一角度来讲，九州从来没有成为中国的现实。

    胎死腹中的五服制

    在《禹贡》中还记载了一种“五服”制：五百里甸服，五百里侯服，五百里绥服，五百里要服，五百里荒服。

    根据这样一种国家模式，在王居住的京城往外，第一等是甸服（以农业为主的直接统治区），第二等是侯服（诸侯统治区），第三等是绥服（必须加以绥抚的地区），第四等是要服（边远地区），第五等是荒服（蛮荒地区）。

    葛剑雄 | 九州的传说和现实
    五服图

    如果说，九州制因为是以名山大川为主要界限，所以还能使人相信为实际行政区域的话，五服制这样四四方方二千五百里的划分就难以自圆其说了。连宋代的儒家学者蔡沈在给《尚书》作注释时也不得不指出：“尧的都城在冀州，冀州的北界在今河北北部和内蒙古南部，恐怕不会有二千五百里。即使算到这么远，也都是沙漠不毛之地了。而东南最富庶的地区反而被列入要服和荒服（离冀州一千五百至二千五百里），根据地势来考察，简直弄不明白是怎么回事！”

    但是五服制中有一点却反映了这样一个事实：在生产力低下、运输相当困难的情况下，王（天子）对臣民的贡品的征收不得不随距离的远近而改变。例如在天子直属区“五百里甸服”的范围内就规定了五种不同的纳贡标准：一百里内割下来的作物连穗带秆起交，二百里内只交谷穗，三百里内交谷子，四百里内交粗米，五百里内交精米。实际实行的制度虽不可能如此刻板，但运输能力显然是必须考虑的因素。

    九州制是对未来的设想，五服制却是对过去的理想化。因为在西周和以前虽然采用类似的分等级统治体制，却并没把每一等级固定为五百里，实际上也不存在这样的可能。所以五服制虽见于《禹贡》，却从来没有哪一个君主或政治家有意实行过，只能胎死腹中。

    大九州说

    正因为九州制仅仅是一种理想，所以在《禹贡》问世以后，还出现了另外几种九州的方案，如《周礼》（也是托名周朝制度的著作）中的《职方》、《尔雅》中的《释地》和《吕氏春秋》中的《有始览》都提出了自己的九州规划，各州名称与《禹贡》不尽相同，划分的范围也有所差异。

    战国时齐国学者邹衍又提出了他的大九州学说，大意是这样的（今译）：儒家所谓的中国，不过只有天下的八十一分之一。中国的名称叫赤县神州，内部有九个州，就是大禹划定的，但这还不能算是真正的州。在中国之外像赤县神州这样的单位共有九个，这才是九州。在九州的周围有大海包围，人类和动物都无法来往。这样的九州合起来又是一个州，像这样的单位也有九个，在它们的周围有更大的海洋包围着，这就到了天地的边缘。

    这种学说与其说是对外部世界的了解，还不如说是出于臆想和推理。比起那种中国就等于天下，除了中国（实际上只是中原）之外就没有文明社会的观点来，大九州学说高明地承认了还存在着不止一个同样发达的人类社会。但恰恰在这一点上又作了实际上的自我否定：由于各州之间都由无边无际的大海阻隔，人民禽兽是无法来往的。所以这种存在只具有理论和思辨上的意义，而不是对中国有影响的现实。

    中原和华夏

    无论是九州的设想，还是大九州的学说，出现在战国后期都不是偶然的。

    《禹贡》所描述的地理范围已经相当广大，涉及今天中国内地的绝大部分。要具备这样丰富的地理知识，活动范围只限于黄河中下游的夏人、商人和西周人是办不到的。而在战国后期，秦、楚、齐、燕、韩、赵、魏这七个主要诸侯国的疆域已经达到了这样的范围，在互相的交流中，各国的学者就可能掌握这些地理知识。《禹贡》中还记录了各地的农业生产条件，如土壤的类型、土地的等级、水文状况等；应纳贡赋的等级和物产等；都是经济发展达到一定水准的反映。例如梁州的贡物中有铁和镂，镂就是钢。如果没有冶金技术的进步，学者的想像力再丰富，也不可能把这种品种载入著作中。

    在七国的竞争中，尽管鹿死谁手还没有最终明朗，但统一已是大势所趋。秦国变得越来越强大，在错综复杂的形势中明显处于主导地位。一些有远见的知识分子纷纷投向秦国，并为秦国战胜其他六国，完成统一事业出谋划策，也为统一后的未来规划蓝图。多数研究者认为《禹贡》是秦国学者的作品，就考虑到这个因素。

    在经过战争、吞并和融合之后，华夏族已经成为黄河流域乃至东亚大陆人数最多、经济文化最发达、实力最强的民族，占据了当时地理条件最优越的地区。而非华夏民族则被迫迁出了黄河流域，或者逐步融入了华夏族，或者接受了华夏文化并以华夏的一支自居。在蒙古高原、青藏高原、长江流域及其以南和大陆附近的茫茫海洋上，还不存在在总体上能与之匹敌的其他民族和政权，而对此范围之外的情况，虽然人们不至于一无所知（例如穿越河西走廊至中亚的陆上交通线和通向东南亚的海上交通线可能已经存在），但肯定相当有限。

    然而随着境外玉石、珠宝、香料等珍奇异物的流入和亲历者见闻的传播，以中原为中心的观念不能不有所动摇。根据九州的理论，中原是文明的中心，九州是文明的范围，但这些珍异并不产在九州，而是来自“非我族类”的夷狄之邦；莫非那里存在着比中原更高的文明？国君、贵族和上层人士享用着来自境外的珍奇，却从不承认会有文明程度超过自己的社会，于是西方的昆仑山、西王母、瑶池和东方的海上神山一类神话便合适地弥补了这一漏洞——原来在中国之外的确存在着一个可望而不可及的神灵世界。但这丝毫不会动摇中国的中心地位，因为西王母尽管伟大，昆仑山尽管崇高，蓬莱尽管奇妙，却都属于神仙的体系，而除了神仙之外，境外就只是一片早期愚昧落后的混沌世界。

    可以认为：在战国时期形成的统一观，是以华夏族（汉族的前身）为主干、以黄河中下游平原地区为中心的，是一种封闭的观念。

    本文摘自《昔日的天下观》《统一与分裂：中国历史的启示》商务印书馆2013年版

  • 钱穆：中国文化之地理背景

    一

    中国是一个文化发展很早的同家，他与埃及、巴比仑、印度，在世界史上上古部分里，同应占到很重要的篇幅。但中国因其环境关系，他的文化，自始即走上独自发展的路径。在有史以前，更渺茫的时代里，中国是否与西方文化有所接触，及其相互间影响何如，现在尚无从深论。但就大体言，中国文化开始，较之埃及、巴比仑、印度诸国，特别见为是一种孤立的，则已成为一种明显的事实。

    中国文化不仅比较孤立，而且亦比较特殊，这里面有些可从地理背景上来说明。埃及、巴比仑、印度的文化，比较上皆在一个小地面上产生。独有中国文化，产生在特别大的地面上。这是双方最相异的一点。人类文化的最先开始，他们的居地，均赖有河水灌溉，好使农业易于产生。而此灌溉区域，又须不很广大，四围有天然的屏障，好让这区域里的居民，一则易于集中而到达相当的密度，一则易于安居乐业而不受外围敌人之侵扰。在此环境下，人类文化始易萌芽。埃及尼罗河流域，巴比仑美索不达米亚平原，印度印度河流域，莫不如此。印度文化进展到恒河流域，较为扩大，但仍不能与中国相比。中国的地理背景，显然与上述诸国不同。

    普通都说，中国文化发生在黄河流域。其实黄河本身并不适于灌溉与交通。中国文化发生，精密言之，并不赖藉于黄河本身，他所依凭的是黄河的各条支流。每一支流之两岸和其流进黄河时雨水相交的那一个角里，却是古代中国文化之摇篮。那一种两水相交而形成的三角地带，这是一个水桠杈，中国古书里称之曰“汭”，汭是在两水环抱之内的意思。中国古书里常称渭汭、泾汭、洛汭，即指此等三角地带而言。我们若把中国古史上各个朝代的发源地和根据地分配在上述的地理形势上，则大略可作如下之推测。唐、虞文化是发生在现在山西省之西南部，黄河大曲的东岸及北岸，汾水两岸及其流入黄河的桠杈地带。夏文化则发生在现在河南省之西部，黄河大曲之南岸，伊水、洛水两岸，及其流入黄河的桠杈地带。

    周文化则发生在现在陕西省之东部，黄河大曲之西岸，渭水两岸，及其流入黄河的桠杈地带。这一个黄河的大隈曲，两岸流着泾、渭、伊、洛、汾、涑几条支流，每一条支流的两岸，及其流进黄河的三角桠杈地带里面，都合宜于古代农业之发展。而这一些支流之上游，又莫不有高山叠岭为其天然的屏蔽，故每一支流实自成为一小区域，宛如埃及、巴比仑般，合宜于人类文化之生长。而黄河的几个渡口，在今山西省河津、临晋、平陆诸县的，则为他们当时相互交通的孔道。

    据中国古史传说，虞、夏文化极相密接，大概夏部族便从洛水流域向北渡过黄河，而与汾水流域的虞部族相接触。其主要的渡口为平陆的茅津渡，稍东的有孟津。周部族之原始居地，据旧说乃自今陕西渭河上流逐步东移。但据本书作者之意见，颇似有从山西汾河下流西渡黄河转到陕西渭河下流之可能。无论如何，周部族在其定居渭河下游之后，常与黄河东岸汾水流域居民交通接触，则为断无可疑之事。因此上述虞夏周三氏族的文化，很早便能融成一体，很难再分辨的了。这可以说是中国古代较为西部的一个文化系统。

    中国古代的黄河，流到今河南省东部，一到郑县境，即折向北，经今河南浚县大伾山下，直向北流，靠近太行山麓，到今天津附近之渤海湾入海。在今安阳县（旧彰德府）附近，便有漳水、洹水流入黄河，这里是古代殷、商氏族的政府所在地。他们本由黄河南岸迁来，在此建都，达二百八十年之久。最近五十年内，在那里发掘到许多牛胛骨与龟版，上刻贞卜文字，正为此时代殷商王室之遗物，因此我们对于此一时期中在此地域的商文化，增多了不少新智识。

    原来的商族，则在今河南省归德附近，那里并非黄河流经之地，但在古代则此一带地面保存很多的湖泽，最有名的如孟诸泽、蒙泽之类。也有许多水流，如睢水、濊水（即涣水）之类。自此（归德）稍向北，到河南中部，则有荥泽、圃田泽等。自此稍东北，山东西部，则有菏泽、雷夏、大野等泽。大抵商部族的文化，即在此等沼泽地带产生。那一带正是古代淮水、济水包裹下的大平原，商代文化由此渐渐渡河向北伸展而至今河南之安阳，此即所谓殷墟的，这可以说是中国古代较为东部的一个文化系统。这一个文化系统，再溯上去，或可发生在中国之极东，燕、齐滨海一带，现在也无从详说了。

    但在有史以前很早时期，似乎上述的中国东西两大系统的文化，早已有不断的接触与往来，因此也就很难分辨说他们是两个系统。更难说这两大系统的文化，孰先孰后。

    现在再从古代商族的文化地域说起。因为有新出土的甲骨文为证，比较更可信据。那时商王室的政治势力，似乎向西直达渭水流域，早与周部族相接触，而向东则达今山东、河北两省沿海，中间包有济水流域的低洼地带。向东北则直至辽河流域，向南则到淮水流域，向西南则到汉水流域之中游，说不定古代商族的文化势力尚可跨越淮、汉以南，而抵达长江北岸。这些地带，严格言之，早已在黄河流域外，而远在商代早已在中国文化区域里。及到周代兴起，则长江流域、汉水、淮水、济水、辽河诸流域，都成为中国文化区域之一部分，其事更属显明。

    我们只根据上文约略所谈，便可见古代中国文化环境，实与埃及、巴比仑、印度诸邦绝然不同。埃及、巴比仑、印度诸邦，有的只籍一个河流和一个水系，如埃及的尼罗河。有的是两条小水合成一流，如巴比仑之底格里斯与阿付腊底河，但其实仍只好算一个水系，而且又都是很小的。只有印度算有印度河与恒河两流域，但两河均不算甚大，其水系亦甚简单，没有许多支流。只有中国，同时有许多河流与许多水系，而且都是极大和极复杂的。那些水系，可照大小分成许多等级。如黄河、长江为第一级，汉水、淮水、济水、辽河等可为第二级，渭水、泾水、洛水、汾水、漳水等则为第三级，此下还有第四级第五级等诸水系，如汾水相近冇涑水，漳水相近有淇水、濮水，入洛水者有伊水，入渭水者有沣水、滈水等。此等小水，在中国古代史上皆极著名。中国古代的农业文化，似乎先在此诸小水系上开始发展，渐渐扩大蔓延，弥漫及于整个大水系。我们只要把埃及、巴比仑、印度及中国的地图仔细对看，便知其间的不同。

    埃及和巴比仑的地形，是单一性的一个水系与单一性的一个平原。印度地形较复杂，但其最早发展，亦只在印度北部的印度河流域与恒河流域，他的地形仍是比较单纯。只有中国文化，开始便在一个复杂而广大的地面上展开。有复杂的大水系，到处有堪作农耕凭藉的灌溉区域，诸区域相互间都可隔离独立，使在这一个区域里面的居民，一面密集到理想适合的浓度，再一面又得四围的天然屏障而满足其安全要求。如此则极适合于古代社会文化之酝酿与成长。但一到其小区域内的文化发展到相当限度，又可藉着小水系进到大水系，而相互间有亲密频繁的接触。因此中国文化开始便易走进一个大局面，与埃及、巴比仑、印度，始终限制在小面积里的情形大大不同。若把家庭作譬喻，埃及、巴比仑、印度是一个小家庭，他们只备一个摇篮，只能长育一个孩子。中国是一个大家庭，他能具备好几个摇篮，同时抚养好几个孩子。这些孩子成长起来，其性情习惯自与小家庭中的独养子不同。这是中国文化与埃及、巴比仑、印度相异原于地理背景之最大的一点。

    其次再有一点，则关于气候方面。埃及、巴比仑、印度全都近在热带，全在北纬三十度左右，物产比较丰足，衣食易给，他们的文化，大抵从多量的闲暇时间里产生。只有中国已在北温带的较北地带，在北纬三十五度左右。黄河流域的气候，是不能和埃及、印度相比的，论其雨量，也远不如埃及、印度诸地之丰富。古代中国北部应该和现在的情形相差不远，我们只看周初时代《豳风·七月》诗里所描写那时的节令物产以及一般农民生活，便知那时情形实与现在山西、陕西一带黄河、渭水附近甚相类似。因此中国人开始便在一种勤奋耐劳的情况下创造他的文化，较之埃及、巴比仑、印度之闲暇与富足的社会，又是绝不相似了。

    二

    根据上述，古代中国因其天然环境之特殊，影响其文化之形成，因有许多独特之点，自亦不难想像而知。兹再约举其大者言之。

    第一：古代文化发展，皆在小环境里开始，其缺点在于不易形成伟大的国家组织。独有中国文化，自始即在一大环境下展开，因此易于养成并促进其对于政治、社会凡属人事方面的种种团结与处理之方法与才能。遂使中国人能迅速完成为一内部统一的大国家，为世界同时任何民族所不及。

    第二：在小环境里产生的文化社会，每易遭受外围文化较低的异族之侵凌，而打断或阻碍其发展。独有中国文化，因在大环境下展开，又能迅速完成同家内部之团结与统一，因此对于外来异族之抵抗力量特别强大，得以不受摧残，而保持其文化进展之前程，逐渐发展。直至现在成为世界上文化绵历最悠久的国家，又为世界任何民族所不及。

    第三：古代文明多在小地面的肥沃区域里产生，因此易于到达其顶点，很早便失却另一新鲜向前的刺激，使其活力无地使用，易于趋向过度的奢侈生活，而招致社会内部之安逸与退化。独有中国文化，因在较苦瘠而较广大的地面产生，因此不断有新刺激与新发展的前途。而在其文化生长过程下，社会内部亦始终能保持一种勤奋与朴素的美德，使其文化常有新精力，不易腐化。直到现在，只有中国民族在世界史上仍见其有虽若陷于老朽，而仍有其内在尚新之气概，此又为并世诸民族所不逮。

    因于上述三点，所以中国文化经过二三千年的发展，完成了他的上古史之后，一到秦、汉统一时代，正为中国文化开始走上新环境、新气象之另一进程，渐渐由黄河流域扩展至长江流域的时代。而与他同时的几个文明古国，如埃及、巴比仑、印度等，皆已在世界文化史上开始退出他们重要的地位，而让给其他的新兴民族来扮演另一幕的主角了。

    三

    若照全世界人类文化已往成绩而论，便只有西方欧洲文化和东方中国文化两大系统，算得源远流长，直到现在，成为人类文化之两大主干。我们不妨乘便再将此两大文化约略作一简单的比较。

    欧洲文化的远祖是希腊，希腊文化灿烂时期，正和中国西周乃至春秋、战国时代相平行。但双方有一极大的不同。希腊诸邦，虽则有他们共同的文化，却从没有他们共同的政治组织。希腊永远是一种互相独立的市府政治，每一市府，各成一单位。中国西周乃至春秋时代，虽亦同样有许多国家，每一国家虽则几乎亦同样以一个城市，即中国古书中称为“国”的为中心，但这些国家，论其创始，大体都由一个中央政府，即西周王室所分封，或经西周王室之正式承认。因此西周时代的中国，理论上已是一个统一国家，不过只是一种“封建式的统一”，而非后代郡县式的统一而已。

    中国此时之所谓“封建”，亦和欧洲中世纪的封建不同。惟其如此，所以一到春秋时代，虽则西周王室东迁，他为中原诸侯共主的尊严早已失去，但还可以有齐桓公、晋文公一辈在列国诸侯中称霸为盟主的起来，代替王室，继续联合和好与统一的工作。这是西方希腊政治所不能完成的。因此西方希腊诸市府，一到中国秦、汉时代，便不免完全为罗马所吞灭，从此西方文化又要走入一新境界。但中国秦、汉时代，却并非如西方般，由外面来了一个新势力，把旧有的中国吞灭，中国秦、汉时代，只是在旧中国的内部，自身有一种改进，由封建式的统一，转变而成“郡县式的统一”，使其统一之性质与功能，益增完密与强固而已。

    我们继此可以说到西方罗马与汉代之不同。罗马政府的性质，论其原始也和希腊市府一般。后来逐步向外伸张，始造成一个伟大的帝国。这一个帝国之组织，有他的中心即罗马城，与其四围之征服地。这是在帝国内部显然对立的两个部分。至于中国汉代，其开始并没有一个像希腊市府般的基本中心，汉代的中国，大体上依然承袭春秋、战国时代来，只在其内部组织上，起了一种新变化。这一种变化，即如上节所说，由封建式的统一转变成为郡县式的统一。

    因此汉代中国，我们只可说他有了一种新组织，却不能说他遇到一个新的征服者。罗马帝国由征服而完成，汉代中国则不然。那时的中国，早已有他二三千年以上的历史，在商、周时代，国家体制早已逐渐完成了。一到汉代，在他内部，另有一番新的政治组织之酝酿与转化。因此在罗马帝国里面，显然有“征服者”与“被征服者”两部分之对立，而在汉代中国，则浑然整然，只是一体相承，并没有征服者与被征服者之区分。西方习惯称罗马为帝国(Empire)，汉代中国决不然，只可称为一国家(Nation)。照西方历史讲，由希腊到罗马，不仅当时的政治形态变了，由市府到帝国，而且整个的国家和人民的大传统也全都变了，由希腊人及希腊诸市府变到罗马人与罗马帝国。而那时的中国，则人民和国家的大传统，一些也没有变，依然是中国人和中国，只变了他内部的政治形态，由封建到郡县。

    我们再由此说到罗马覆亡后的西方中古时期，和中国汉代覆亡后之魏晋南北朝时期，两者中间仍有显著的不同。罗马覆亡，依然和希腊覆亡一样，是遇到了一个新的征服者，北方蛮族。此后的欧洲史，不仅政治形态上发生变动，由帝国到封建，而且在整个的人民和国家的大传统上也一样的发生变动，由南方罗马人转变到北方日耳曼人，又由罗马帝国转变到中世纪封建诸王国。中国汉代的覆灭，并不是在中国以外，另来了一个新的征服者，而仍然是在中国内部起了一种政治形态之动荡。东汉以后，魏、蜀、吴三国分裂，下及西晋统一，依然可以说是一种政治变动，而非整个民族和国家传统之转移。此后五胡乱华，虽有不少当时称为胡人的乘机起乱，但此等胡人，早已归化中国，多数居在中国内地，已经同样受到中国的教育。他们的动乱，严格言之，仍可看作当时中国内部的一种政治问题和社会问题，而非在中国人民与中国国家之外，另来一个新的征服者。

    若依当时人口比数论，不仅南方中国，全以中国汉人为主体，即在北方中国，除却少数胡族外，百分之八九十以上的主要户口依然是中国的汉人。当时南方政治系统，固然沿着汉代以来的旧传统与旧规模，即在北朝，除却王室由胡族为之，其一部分主要的军队由胡人充任以外，全个政府，还是胡、汉合作。中国许多故家大族，没有南迁而留在北方的，依然形成当时政治上的中坚势力，而社会下层农、工、商、贾各色人等，则全以汉人为主干。因此当时北朝的政治传统，社会生活，文化信仰，可以说一样承袭着汉代而仍然为中国式的旧传统。虽不免有少许变动，但这种变动，乃历史上任何一个时代所不免。若单论到民族和国家的大传统，文化上的大趋向，则根本并无摇移。

    因此西方的中古时代，北方蛮族完全以一种新的民族出现而为此下西方历史之主干，旧的罗马人则在数量上已成被压倒的劣势而逐渐消失。反之，在中国史上，魏晋南北朝时代，依然以旧的中国人为当时政治、社会、文化各部门各方面之主干与中坚。至于新的胡人，只以比较的少数加入活动，如以许多小支流浸灌入一条大河中，当时虽有一些激动，不久即全部混化而失其存在了。这一层是中国魏晋南北朝时代和欧洲中古时期的绝大不同处。

    因此西方的中古时期，可以说是一个转变，亦可说是一个脱节，那时的事物，主要的全是新兴的。北方日耳曼民族成为将来历史和文化之主干，这是新兴的。当时所行的封建制度，亦是新兴的。西方的封建，乃罗马政治崩溃后，自然形成的一种社会现象，根本与中国史上西周时代所谓的封建不同。中国的封建制度，乃古代中国统一政治进展中之一步骤、一动象；西方封建，则为罗马政治解消以后一种暂时脱节的现象。那时在西方主持联合与统一工作的，主要者并非封建制度，而为基督教的教会组织。这种教会组织又是新兴的。

    希腊、罗马和基督教会之三者，成为近代西方文化之三主源。在中国魏晋南北朝时代，虽同样有印度佛教之流入，并亦一时称盛，但在历史影响上，复与西方中古时期的基督教绝然不同。基督教是在罗马文化烂熟腐败以后，完全以新的姿态出现而完成其感化北方蛮族的功能的。但魏晋南北朝时代的中国，则以往传统文化并未全部衰歇。孔子的教训，依然为社会人生之最大信仰与最大归趋，只在那时又新增了一个由印度传来的佛教，而一到唐代以后，佛教也到底与儒教思想相合流相混化。因此我们可以说，在欧洲中古时期，论其民族，是旧的罗马民族衰歇而新的日耳曼民族兴起。在中国则只在旧的中国汉民族里面增加了一些新民族新分子，胡人。

    论政治，在欧洲中古时期，是旧的罗马统治崩溃，而新的封建社会兴起。在中国则依然是秦、汉的政治制度之沿续，根本上并无多少转换。论文化与信仰，在欧洲中古时期，则由旧的罗马文化转变到新的基督教文化。在中国，则依然是一个孔子传统，只另外又加进一些佛教的成分。却不能说那时的中国，由旧的孔教而变成为新的佛教了。

    由此言之，西方的中古时期，全是一个新的转变，而魏晋南北朝时代的中国，则大体还是一个旧的沿袭。那些王朝的起灭和政权之转移，只是上面说的一种政治形态之动荡。若论民族和国家的大传统，中国依然还是一个承续，根本没有摇移。

    根据上述，来看近代西方新兴的民族国家，他们在两洋史上，又都是以全新的姿态而出现的。论其民族和国家的大传统，他们复和古代的希腊、罗马不同。但中国史则以一贯的民族传统与国家传统而绵延着，可说从商、周以来，四千年没有变动。所有中国史上的变动，伤害不到民族和同家的大传统。因此中国历史只有层层团结和步步扩展的一种绵延，很少彻底推翻与重新建立的像近代两方人所谓的革命。这是中西两方历史形态一个大不同处，因此而影响到双方对于历史观念之分歧。

    西方人看历史，根本是一个“变动”，常由这一阶段变动到那一阶段。若再从这个变动观念上加进时间观念，则谓历史是“进步”的，人类历史常由这一时代的这一阶段，进展到另一时代的另一阶段。但中国人看历史，则永远在一个“根本”上，与其说是变动，不如说是“转化”。与其说是进步，不如说是“绵延”。中国人的看法，人类历史的运行，不是一种变动，而是一种转化。不是一种进步，而是一种绵延，并不是从这一阶段变动、进步而达另一阶段，只是依然在这一阶段上逐渐转化、绵延。

    变动、进步是“异体的”，转化、绵延则是“同体的”。变动、进步则由这个变成了那个。转化、绵延则永远还是这一个。因此两方人看历史，常偏向于“空间”的与“权力”的“向外伸展”；中网人看历史，常偏向于“时间”的与“生长”的“自我绵延”。西方人的看法，常是“我”与“非我”两个对立。中国人的看法，只有自我一体浑然存在。双方历史形态之不同，以及双方对于历史观念之不同，其后面便透露出双方文化意识上之不同。这一种不同，若推寻根柢，我们依然可以说中西双方全都受着一些地理背景的影响。中国在很早时期，便已凝成一个统一的大国家。在西方则直到近代，由中国人眼光看来，依然如在我们的春秋、战国时代，列国纷争，还没有走上统一的路。

    中国历史正因为数千年来常在一个大一统的和平局面之下，因此他的对外问题常没有像他对内问题那般的重要。中国人的态度，常常是反身向着内看的。所谓向内看，是指看一切东西都在他向己的里面。这样便成为自我一体浑然存在。西方历史则永远在列国纷争、此起彼仆的斗争状态之下，因此他们的对内问题常没有像他们对外问题那般的重要，西方人的态度，则常常是向外看的。所谓向外看，是指看一切东西都在他自己的外面，所以成为我与非我屹然对立。惟其常向外看，认为有两体对立，所以特別注意在空间的“扩张”，以及“权力”和“征服”上。惟其常向内看，认为只有一体浑然，所以特别注意到时间的“绵延”以及“生长”和“根本”上。

    四

    其次说到双方经济形态，中国文化是自始到今建筑在农业上面的，西方则自希腊、罗马以来，大体上可以说是建筑在商业上面。一个是彻头彻尾的农业文化，一个是彻头彻尾的商业文化，这是双方很显著的不同点。

    依西方人看法，人类文化的进展，必然由农业文化进一步变成商业文化。但中国人看法，则并不如此。中国人认为人类生活，永远仰赖农业为基础，因此人类文化也永远应该不脱离农业文化的境界，只有在农业文化的根本上再加绵延展扩而附上一个工业，更加绵延展扩而又附上一个商业，但文化还是一线相承，他的根本却依然是一个农业。

    照西方人看，文化是变动的，进步的，由农到商截然不同。照中国人看，则文化还是根本的与生长的，一切以农为主。这里自然也有地理背景的影响。因为西方文化开始如埃及、巴比仑等，他们本只有一个狹小的农业区，他们的农业文化不久便要达到饱和点，使他们不得不转换方向改进到商业经济的路上去。希腊、罗马乃至近代西方国家莫不如此。在中国则有无限的农耕区域可资发展，因此全世界人类的农业文化，只有在中国得到一个继长增荣不断发展的机会。

    中国历史，在很早时期里，便已有很繁荣的商业了。但因中国开始便成为一个统一的大国，因此他的商业常是对内之重要性超过了对外。若西方各国，则常是对外通商的重要性超过了对内。因此双方对商业的看法，也便有异。西方常常运用国家力量来保护和推进其国外商业。中国则常常以政府法令来裁制国内商业势力之过分旺盛，使其不能远驾于农、工之上。因此在西方国家很早便带有一种近代所谓“资本帝国主义”的姿态，在中国则自始到今常采用一种近代所谓“民主社会主义”的政策。

    再换辞言之，农业文化是自给自足的，商业文化是内外依存的。他是要吸收外面来营养自己的。因此农业文化常觉得内外一体，只求安足。商业文化则常觉彼我对立，惟求富强。结果富而不足，强而不安，因此常要变动，常望进步。农业文化是不求富强但求安足的，因此能自本自根一线绵延。

    我们继此讲到科学和工业，科学知识和机械工业在现世界的中国是远为落后的。但中国已往历史上，也不断有科学思想与机械创作之发现，只因中国人常采用的是民主社会主义的经济政策，“不患寡而患不均”。对于机械生产，不仅不加奖励，抑且时时加以禁止与阻抑，因此中国在机械工业一方面，得不到一个活泼的发展。在中国的机械和工业，是专走上精美的艺术和灵巧的玩具方面去了。科学思想在中国之不发达，当然不止此一因，但科学没有实际应用的机会，自为中国科学不发达的最要原因之一。

    五

    其次我们再说到中西双方对于人生观念和人生理想的异同。“自由”(Liberty & Freedom)一词是西方人向来最重视的。西方全部历史，他们说，即是一部人类自由的发展史。西方全部文化，他们说，即是一部人类发展自由的文化。“人生”、“历史”和“文化”，本来只是一事，在西方只要说到“自由”，便把这三方面都提纲挈领的总会在一处了。在中国则似乎始终并不注重“自由”这个字。西方用来和自由针对的，还有“组织”和“联合”(Organization & Unity)。希腊代表着自由，罗马和基督教会则代表着组织和联合。这是西方历史和西方文化的两大流，亦是西方人生之两大干。我们只把握这两个概念来看两方史，便可一一看出隐藏在两方历史后面的一切意义和价值。

    但中国人向来既不注重自由，因此也便不注重组织和联合，因为自由和联合的后面，还有一个概念存在的，这便是“两体对立”。因有两体对立，所以要求自由，同时又要求联合。但两体对立，是西方人注重向外看，注重在空间方面看的结果。是由西方商业文化内不足的经济状态下产生的现象。中国人一向在农业文化中生长，自我安定，不须向外寻求，因此中国人一向注重向内看，注重在时间方面看，便不见有严重的两体对立，因此中国人也不很重视自由，又不重视联合了。中国人因为常偏于向内看的缘故，看人生和社会只是浑然整然的一体。

    这个浑然整然的一体之根本，大言之是自然、是天；小言之，则是各自的小我。“小我”与“大自然”混然一体，这便是中国人所谓的“天人合一”。小我并不和此大自然体对立，只成为此体之一种根荄，渐渐生长扩大而圆成，则此小我便与大自然融和而浑化了。此即到达天人合一的境界。中国《大学》一书上所说的修身、齐家、治国、平天下，一层一层的扩大，即是一层一层的生长，又是一层一层的圆成，最后融和而化，此身与家、国、天下并不成为对立。这是中国人的人生观。

    我们若把希腊的自由观念和罗马帝国以及基督教会的一种组织和联合的力量来看中国史，便得不到隐藏在中国史内面深处的意义与价值。我们必先了解中国人的人生观念和其文化精神，再来看中国历史，自可认识和评判其特殊的意义和价值了。但反过来说，我们也正要在中国的文化大流里来认识中同人的人生观念和其文化精神。

    继此我们再讲到中两双方的宗教信仰。西方人常看世界是两体对立的，在宗教上也有一个“天国”和“人世”的对立。在中国人观念里，则世界只有一个。中国人不看重并亦不信有另外的一个天国，因此中国人要求永生，也只想永生在这个世界上。中国人要求不朽，也只想不朽在这个[让界上。中同古代所传诵的立德、立功、立言三不朽，便从这种观念下产生。中国人只想把他的德行、事业、教训永远留存在这个世界这个社会上。中国人不想超世界超社会之外，还有一个天国。因此在西方发展为宗教的，在中国只发展成“伦理”。中国人对世界对人生的“义务”观念，反更重于“自由”观念。在西方常以义务与权利相对立，在中国则常以义务与白由相融和。义务与自由之融和，在中国便是“性”（自由）与“命”（义务）之合一，也便是“天人合一”。

    西方人不仅看世界常是两体对立，即其看自己个人，亦常是两体对立的。西方古代观念，认人有“灵魂”“肉体”两部分，灵魂部分接触的是理性的“精神世界”，肉体部分接触的是感官的“物质世界”。从此推衍，便有西方传统的“二元论”的哲学思想。而同时因为西方人认为物质世界是超然独立的，因此他们才能用纯客观的态度来探究宇宙而走上科学思想的大园地。中国人则较为倾向“身心一致”的观念，并不信有灵肉对立。他看世界，亦不认为对我而超然独立，他依然不是向外看，而是向内看。他认为我与世界还是息息相通，融为一体。

    儒家思想完全以“伦理观”来融化了“宇宙观”，这种态度是最为明显了。即在道家，他们是要摆脱儒家的人本主义，而从宇宙万物的更广大的立场来观察真理的，但他们也依然保留中国人天人合一的观点，他们并不曾从纯客观的心情上来考察宇宙。因此在中国道家思想里，虽有许多接近西方科学精神的端倪，但到底还发展不出严格的西方科学来。

    以上所述，只在指出中西双方的人生观念、文化精神和历史大流，有些处是完全各走了一条不同的路。我们要想了解中国文化和中国历史，我们先应该习得中国人的观点，再循之推寻。否则若从另一观点来观察和批评中国史和中国文化，则终必有搔不着痛痒之苦。

    本文原载《中国文化史导论》，商务印书馆，2023年版

  • 何立波：迦太基的兴衰及其多元文明特征

    迦太基是腓尼基城邦的推罗人在非洲建立的一个殖民地，后成为古代地中海世界一个著名的商业民族。迦太基据说是推罗人在柏萨(Byrsa)向非洲土著借来的一块“牛皮之地”，经过不断扩张拓展，成为和希腊并列的西地中海两殖民帝国，也是与罗马并列的西地中海世界的强国。迦太基在商业领域和希腊人有激烈竞争，也和新兴的罗马在西西里出现了冲突，与希腊人和罗马人均兵戎相见。迦太基在公元前146年被罗马所灭，作为一个国家不复存在。但是迦太基文明的混合型特征，以及带来的文明交流和传播的意义却不容忽视。对于迦太基的兴衰及其文明特征，学界已经取得了一些成果，但还有很大的提升空间。①

    一、迦太基海上商业帝国的建立与早期西地中海世界的文明交流

    从民族上说，迦太基人属于古代地中海东岸腓尼基人的一支。迦太基人代表了古代东方民族殖民、商业和航海的高峰，在某种意义上进行了古代国家“重商主义”文明的最早探索。

    (一)迦太基的建立和对外贸易扩张

    一般认为，腓尼基人中的推罗人于公元前814年在非洲建立了迦太基。“迦太基”在腓尼基语中为“”，在希腊语中为“Karchēdon”，在拉丁语中为“Carthago”，意思是“新城”。②

    关于迦太基的建城史，一直有“一张牛皮”的传说。古希腊史学家提迈欧(Timaeus，约前352-前256)最早提出迦太基由推罗妇女狄多(Dido，亦称Elissa)在第一次奥林匹克运动会前的第38年(前814年)所建之说。③罗马帝国史学家阿庇安(Appian，约95-165)提供了更详细的记载，称狄多带领族人来到迦太基所在地，求土著酋长赐一块牛皮之地。在征得土著酋长的同意后，狄多将一张牛皮裁剪成条，圈出一块城镇大小的地皮，即后来迦太基的卫城柏萨。④“柏萨”的腓尼基语为“Bozra”，意思是围城、堡垒；而在希腊语中为“Byrsa”，意思是“藏牛之处”。⑤柏萨位于迦太基城的中央，是一个险峻之地。阿庇安在叙述公元前146年迦太基毁灭时，提到迦太基已繁荣了700年⑥，显然指的是她建立于公元前9世纪晚期的说法。到公元前l世纪末希腊地理学家斯特拉波(Strabo，前64-公元23)到访非洲时，迦太基城仍有方圆360斯塔德(stades)的规模。⑦

    古典学家迈尔斯(Richard Miles)认为，目前迦太基发现的最早的考古文化层的回溯，仅仅能追溯到公元前760年左右。⑧考古学家辛塔斯(Cintas)试图通过对迦太基遗址出土的陶器进行分析，建立迦太基早期历史的时间表，但是罗马人在毁灭迦太基的时候造成较大损毁。1922年，辛塔斯终于在13个陶罐中确认有9个属于希腊陶罐，认定迦太基建城不太可能早于公元前725年。⑨推罗人建立迦太基可能是希腊罗马人的一种想象，斯特拉波、阿庇安等都持这种观点。而实际上，迦太基的早期居民还包括腓尼基各城邦及众多非洲居民。关于早期迦太基，目前发现的最重要的古典档案是“闪米特铭文集”(Corpus inscriptionum semiticarum)第1卷第5684—5号，一般被认为是出现在公元前7世纪，仍保留着明显的推罗语言的特点。⑩

    目前看来，腓尼基人在西地中海的扩张，不会早于公元前750年。(11)在公元前8世纪之前，迦太基在西地中海世界的主要活动是建立殖民地。韦尔(Benjamin W.Wells)认为，殖民地是迦太基国家的支柱，迦太基殖民体系是古典时代的殖民典范。(12)青铜时代晚期的殖民地有两种模式：一种是腓尼基的贸易殖民，专注于奢侈品贸易和经济利益，在地中海西部沿海建立了稳定的定居点；另一种模式是希腊人的殖民运动，在海外建立城邦，带有政治目的。(13)早期迦太基人的贸易和希腊人有所不同，他们更多是以贸易者的身份在海外殖民和经商。他们在迦太基以外的定居地与其说是城市，不如说是“贸易港口”更为合适。

    迦太基从希腊世界和东方输入油、酒等食物及纺织品、陶器、青铜器等手工业品，满足本国居民和殖民地居民需要。目前考古材料证实的迦太基从国外进口商品，最早可以追溯到公元前650年。(14)西西里岛的阿克拉加斯城(Acragas)、意大利的坎佩尼亚，以及爱琴海的罗德斯岛，都是迦太基酒类来源地。油是从阿克拉加斯城运来的。迦太基从希腊大陆、意大利的坎佩尼亚、西西里得到了青铜器物、宝石制品和花瓶。考古学家在一座献给迦太基女神塔妮特(Tanit)的神庙中发现了东方化时代(前720-前580)的6个科林斯式希腊陶罐，时间在公元前740年到前710年间。(15)在迦太基人墓地还发现了从塞浦路斯运来的陶俑和青铜水瓶。埃及也提供了受迦太基人欢迎的装饰品。法国史学家杜丹(Jules Toutain)认为，迦太基的非洲领土及其殖民帝国能出口的商品有限，只有奴隶、矿砂和金属，尤其是西班牙南部的铅和银。(16)迦太基沉船的船骸提供了重要的物质信息，如铜锭、锡锭、玻璃、金银首饰、彩陶、酒和油等。(17)作为一个以商业贸易闻名的古典民族，迦太基人缺乏具有自己民族特色的手工业品，经常仿制和改造从希腊和埃及进口的商品，如他们仿制的希腊陶器就几乎达到了以假乱真的地步。

    (二)迦太基地中海贸易航线的建立和大西洋航海的探索

    迦太基在地中海参与和拓展的贸易航线，大体可分为东西航线和南北航线。东西航线亦称“黎凡特—西班牙线路”，是从希腊、小亚、西亚到西班牙、直布罗陀海峡的航线。伊比利亚半岛和萨丁尼亚岛，是地中海世界主要的铜、锡、银的产地。推罗人首先开辟了从西班牙南部的加底斯(Gades)到推罗的“金属航线”。西西里的希腊史学家狄奥多鲁斯(Diodorus of Sicily，前80-前21)称，迦太基海疆西达直布罗陀的“赫拉克勒斯(Heracles)之柱”、加底斯和大西洋，而加底斯位于有人居住的世界的最远边界。(18)有利的地理位置、肥沃的内陆土地和更优良的港口，让迦太基人很快控制了这条航线，成为地中海最大的铜、锡、银的贸易商。(19)

    南北航线对迦太基来说尤为重要，它将迦太基与西西里、科西嘉、撒丁岛、意大利联系起来，成为第勒尼安海贸易圈的重要坐标。迦太基还位于非洲通往希腊、爱琴海地区的海上航线的通道，有利于它从海外获得粮食、原材料和手工艺品。迦太基早期的粮食依赖进口，主要来自西班牙、意大利、西西里、希腊、黎凡特等地。在迦太基早期居民的陪葬品中，发现了大量希腊陶器(包括爱奥尼亚式和科林斯式陶器)。公元前6世纪初，银价暴跌，金属航线随之衰落，南北航线变得愈加重要起来。迦太基位于东西航线和南北航线的交会点，很快发展成为地中海世界贸易中心。在迦太基国家收入的来源中，关税占主要地位。(20)迦太基在希腊化时期成为地中海世界的一座国际都市，吸引了大量移民。

    迦太基人的贸易具有居间商的性质，航线广泛分布在欧洲、非洲的大西洋沿岸和西地中海地区，形成一个地中海世界贸易网络，迦太基货币也成为地中海西部地区的硬通货币。为保护贸易航线和商业利益，迦太基建立起一支强大的海军舰队，使用先进的三桨座战舰。在内陆殖民地农业的支持、海上贸易的商业支撑和海军舰队的保护下，迦太基建立了一个商业殖民帝国。杜丹强调，迦太基商业的真正范围是海洋，尤其是西部。他们从撒哈拉、苏丹获得了黑奴、象牙、兽皮、黄金、鸵鸟，从高卢获得锡、铅，其买卖都是在沿海或交通要道进行的。杜丹指出，迦太基人在地中海世界贸易中所起的作用，很像17世纪的荷兰人。(21)

    迦太基能够发展成为一个海上商业殖民帝国，是和他们高超的造船技术和航海水平分不开的。古罗马学者老普林尼(Pliny the Elder，23—79)告诉我们，腓尼基人是一个擅长航海、通晓天文并发明字母文字的民族。(22)早在公元前3000年左右，腓尼基人城市比布鲁斯(Byblos)就已制造出拥有弧形船体、能够经受大海考验的船只。古希腊史学家波里比乌斯(Polybius，前204-前122)在谈到迦太基和罗马军队的对比时指出，航海技术长期以来是迦太基人的一种特殊技能，他们比其他任何民族都熟悉大海。(23)迦太基商队以拥有大船而著名，用帆航行。迦太基水手不仅能够沿岸航行，而且能够进行深海航行，在航行中依靠太阳和星辰的位置、熟悉的海岸地形和地貌来辨别航行的方向。

    古希腊罗马作家提到了迦太基人在公元前5世纪的两次大西洋航行，比希腊航海家皮西亚斯(Pytheas of Massalia，前320-前285)在公元前4世纪晚期的首次大西洋航行早了1个世纪。公元4世纪的罗马学者阿维努斯(Avienus)提到，迦太基将军希米尔科(Himilco，前460-前410)船队向北沿着伊比利亚半岛和高卢的西海岸航行数月，在西欧寻找矿石，可能到达了不列颠。(24)老普林尼记录了迦太基将军汉诺(Hanno)的《航行记》，称汉诺船队沿非洲西海岸进行了探险。(25)汉诺船队大致在公元前520年进行远航(26)，远达摩洛哥、毛里塔尼亚、冈比亚等地，并在摩洛哥建立了殖民地。据说最远达几内亚湾，接近了赤道。大西洋航行探索将迦太基人的贸易圈扩展到了大西洋海域。公元前460年，迦太基人开始将摩洛哥的咸鱼运往希腊的科林斯(Corinth)。位于摩洛哥大西洋沿岸的殖民地的建立，与迦太基殖民活动的扩展路径是一致的。迦太基人拥有当时最好的船长和水手，他们能够安全地航行在地中海西部海域和大西洋沿岸。(27)

    二、迦太基人与希腊人、罗马人在海上商业贸易中的冲突和结果

    在迦太基人所开展的对外殖民和贸易活动中，他们和古希腊人、罗马人发生了碰撞和交流。迦太基人在对希腊人、罗马人的战争中失败，奠定了古代西方叙述中迦太基人的失败者形象。

    (一)迦太基和希腊人在西西里的争夺和较量

    在迦太基的对外殖民活动中，西西里是联系地中海东西航道和南北航道的重要交通要冲，地位举足轻重。希腊史学家修昔底德(Thucydides，前460-前400/396)说腓尼基人在希腊之前就来到西西里殖民，居住在沿海，占据岬角和沿海的岛屿。(28)但实际情况可能并非如此。希腊人在东方化时代开始大殖民，在西西里和意大利南部(希腊人称为“大希腊”地区)建立殖民地。希腊人在南意大利的库玛(Cumae)建立殖民地是公元前750年，到达西西里是在公元前8世纪。(29)腓尼基人在西西里建立殖民地不会早于公元前8世纪。(30)在地中海西部，迦太基充当了腓尼基人保护者的角色。迦太基人在西西里西部、科西嘉和撒丁岛建立了很多殖民地，在西西里建立的殖民地有摩提亚(Motya)、帕诺慕斯(Panormus)、索罗伊斯(Soloeis)等。在科西嘉，为对抗共同的敌人希腊人，迦太基人和意大利北部的伊特鲁里亚人(Etruscans)建立了同盟关系。

    狄奥多鲁斯认为，财富是引起人类竞争的主要因素。(31)迦太基人征服一个地方后，通常会征收一大笔贡金。(32)迦太基人坚持要控制通向西班牙的航线，与公元前5世纪的雅典坚持垄断通往黑海的航线颇为相似，二者争夺的对象分别是金属和谷物。迦太基人和希腊人的西西里战争虽有商业和贸易的动机，但无法与17世纪欧洲的商业战争相提并论。

    西西里是迦太基通向西班牙的交通要冲和中转站。从公元前580年开始，腓尼基人和希腊人在西西里西部发生了冲突。希腊人试图进入西西里最西端的利利俾(Lilybaeum)，遭到迦太基人的驱赶。在公元前535年的阿拉里亚(Alalia)战役中，迦太基人和伊特鲁里亚人联合起来击败了希腊人，将第勒尼安海毗邻的海域变为了迦太基的水域。从公元前5世纪开始，迦太基在西西里的主要竞争者是希腊城邦阿克拉伽斯和叙拉古。公元前580年开始的一个多世纪里，为与希腊人争夺西西里岛，迦太基、希腊之间发生了三次战争(布匿—希腊战争)。迦太基在公元前5世纪才有20多万人口，无法建立像其他国家那种公民兵体制，只能是从被征服者和商人中征募雇佣兵，由迦太基人充任将军。(33)也有人认为，精明的迦太基人不肯亲自参加战争，宁愿花钱来雇所谓的“蛮族人”来当兵。(34)

    随着公元前525年波斯帝国征服推罗，日益强大的迦太基与母邦推罗间的联系仅剩下向腓尼基神灵献祭和坚守推罗宗教传统了。公元前4世纪初，伯罗奔尼撒战争后的希腊人停止在西西里殖民，迦太基与希腊双方殖民争霸告一段落。迦太基控制西西里岛的西部，希腊科林斯人在西西里所建立的城邦叙拉古占据了西西里岛东部。西西里的殖民者带来了外来的文化，从多利亚人的农业文化到希腊的陶器和迦太基人的布匿文字都在西西里长期存在。在西西里的外来文化中，希腊文化的影响最大。在人类历史上，最早的洲际划分的理念是由希腊人提出的。波里比乌斯指出，希腊人把“有人居住的世界”()分为欧罗巴、亚细亚和利比亚三大洲。(35)老普林尼说，非洲被希腊人称为“利比亚”，它不包括埃及。(36)斯特拉波强调利比亚疆域很小，那里只有沙漠和野兽；(37)有人居住的世界只有希腊人和野蛮人两种人。(38)狄奥多鲁斯在叙述西西里的历史时，将西西里岛上叙拉古之外的居民称为“蛮族人”，称“除了叙拉古以外的蛮族人拥有了整个西西里”。在书写希腊人和迦太基人在西西里的冲突的时候，狄奥多鲁斯时而使用“迦太基人”的称呼，时而使用“蛮族人”的说法(39)，表达了希腊世界对于迦太基人的看法。

    西西里的外来文化包括布匿文化、希腊文化等，二者分别分布在西部和东部。在迦太基占主导的西西里西部，土著的伊利米人(Elymians)和外来的迦太基人都没有建立起政治组织，更多地表现为一种文化上的传播和交流。在公元前6世纪的西西里西部，迦太基人更注重加强对其他腓尼基人城市的控制，他们在两个世纪后彻底控制腓尼基城邦的货币发行权。到公元前6世纪末，迦太基加强了对西西里西部的统治，在岛屿南部的势力也有了明显提升。迦太基宗教在西西里和撒丁岛的传播有了明显的加强，如献祭和葬仪及托非特(tophet)祭坛。(40)

    与迦太基人相比，希腊人在西西里政治上的存在感更强，建立了叙拉古等城邦国家。叙拉古实行僭主政治，以文化发达而著称。狄奥多鲁斯认为，在公元前5世纪早期的希波战争中，迦太基曾经和波斯帝国结盟，从西部和东部同时对希腊人开战，其中迦太基人负责对西西里和意大利的希腊人作战，波斯人对希腊本土作战。(41)发生在西西里的希梅拉(Himera)战役和希腊大陆的萨拉美斯(Salamis)海战，是在公元前480年的同一天发生的。希罗多德(Herodotus，约前480-前425)、亚里士多德均认为，这有可能只是时间上的巧合而已，波斯人和迦太基人并未在战前协商一致。(42)迦太基对和波斯人结盟未必有兴趣，因为母邦推罗人被迫成为波斯人的仆从国。雅典曾试图联合迦太基人，对抗与他们有矛盾的叙拉古，遭到迦太基的拒绝。在与阿克拉伽斯—叙拉古联军的希梅拉战役中，迦太基的30万军队损失了15万人。(43)迦太基在战后70年间逐渐将注意力转向非洲，还不断向内陆地区推进，试图将利比亚腓尼基化，并将自身非洲化，以巩固其非洲帝国。(44)

    在希波战争后，希腊人加强了对西西里的控制，希腊文化在西西里的影响超过了迦太基。从公元前4世纪起，迦太基人、坎佩尼亚人(Campanians)、奥斯其人(Oscans)和罗马人都活跃在西西里，使得西西里成为一个地中海世界不同文明交汇的大熔炉。公元前2世纪以来，随着罗马人在西西里统治的开始，西西里的政治日益罗马化。但新的政治秩序并未带来立竿见影的效果，希腊文化仍是西西里的强势文化。在西西里，迦太基人的布匿文化并未在第一次布匿战争结束后消失。在公元前1世纪的西西里，布匿文化的影响仍无处不在，陶罐、花瓶等物品上大量出现了布匿语和希腊语的双语铭文。(45)布匿铭文在西西里至少存在到了公元1世纪，布匿口语存在的时间更长。(46)在西部西西里，希腊人和“蛮族人”之间有密切的来往，存在着语言的融合。希腊化时期利利俾出土的铭文记载的两个人的名字，就是集希腊语、拉丁语、布匿语于一身的混合式名字。

    (二)迦太基人和罗马人商业霸权的争夺以及最终的失败

    迦太基在其统治区域内实行贸易垄断制度，达到控制商贸航线和征收关税的目的。据波里比乌斯记载，迦太基与罗马在公元前509年签订了一个条约，对罗马人及其盟邦船只航行到“菲尔角”(Fair Promontory，亦译“美丽岬角”)以西海域进行了严格限制(47)，保持了他们对地中海西部的海上商业霸权。但学者卡列(M.Cary)认为，波里比乌斯所说的签订条约的时间是错误的，正确的时间应是公元前308年。(48)这个条约表明，迦太基在公元前6世纪或4世纪末已有能力对所有外国商船封锁直布罗陀海峡。没有迦太基人的同意，希腊航海家皮西亚斯从加底斯出发经过直布罗陀海峡的大西洋航行就不可能发生。(49)从马耳他(Malta)到西西里、撒丁岛、巴利阿里(Balearics)岛、西班牙的航线，是由迦太基人严密控制的。从公元前6世纪起，希腊人不能直接从西班牙南部获得锡、铜和银。公元前540年，迦太基人联合伊特鲁里亚人将希腊人赶出了科西嘉。(50)

    到公元前6世纪晚期，迦太基的领土从昔兰尼(Cyrene)延伸到了大西洋。(51)在迦太基人看来，撒丁岛和利比亚是自己的领土，而西西里只涉及他们所统治的一部分区域。波里比乌斯告诉我们，迦太基人和罗马人在公元前306年又签订了一个条约，迦太基人明确将利比亚和撒丁岛视为自己的私有财产，将推罗和乌提卡(Utica)也纳入了自己的领土范围，不许罗马人涉足。(52)迦太基人严厉打击海盗，成为古代文明早期和平贸易协约的倡导者。

    迦太基并不缺乏金银，他们从中非得到黄金，从西班牙获取白银。公元前6世纪，钱币在希腊世界已经普及。到公元前6世纪最后25年，希腊人主要居住区都已使用了金属货币。(53)迦太基人重视商业，他们教育青年的指导思想就是要证明商业在迦太基公私生活中的优越地位。迦太基人早期采取的是物物交换或使用希腊等外币。希罗多德记述了迦太基商人同利比亚土著居民“无声的贸易”的交易细节，迦太基人以商品换取土著(可能是摩洛哥土著)的黄金。(54)公元前5世纪末，为了给西西里的雇佣兵开支薪酬，迦太基人用西班牙的银锭来铸造自己的银币。

    公元前3世纪早期是迦太基商业帝国的黄金时期。希腊人海上的势力日益衰落，亚历山大帝国一分为三，罗马人忙于征服中部和北部的意大利人，迦太基人几乎垄断了整个地中海的商业贸易。在罗马人发起挑战之前，迦太基人保持了他们对海洋的控制权，而且把注意力转向通过侵略建立一个陆上帝国，采取了一种更有侵略性的帝国主义政策。随着罗马在公元前3世纪介入西西里事务，迦太基和罗马人的矛盾逐渐激化。从公元前264年争夺西西里开始，罗马和迦太基先后发生了三次战争(布匿战争)。布匿战争揭开了整个地中海世界国家间关系转型的序幕，标志着一个新时代的开始。罗马人开始从陆地向海洋拓展，对迦太基人在西地中海世界的利益构成了严重的挑战。

    罗马是一个传统的陆军强国，早期并未设置海军。第一次布匿战争爆发后，罗马迅速组建一支由公民兵组成的海军，军队士气高。迦太基海军强大但是公民兵不足，主要依靠西班牙人、柏柏尔人(Berbers)及后来的努米底亚人(Numidians)等外族人所组成的雇佣军，军队缺乏凝聚力和爱国精神。波里比乌斯清楚地看到了这一点，认为迦太基军队使用的是外国人和雇佣军，而罗马军队士兵都是本国的公民和土著人，罗马人是在为自己的国家和儿女而战。(55)公元前242年，在第一次布匿战争中战败的迦太基人退出西西里，罗马人获得了除了叙拉古之外的整个岛屿(56)，西西里由此在公元前227年成为罗马的第一个行省。迦太基在第二次布匿战争中在本土战败，被迫和罗马人签订了几乎交出一切的苛刻条约。在反贵族寡头的选民的支持下，汉尼拔于公元前196年当选为迦太基的首席行政长官苏菲特(sufetes)，推行民主改革，重整军队，迦太基逐渐从战败的阴影中走出，元气得以恢复。

    迦太基的复兴让罗马人产生了一种恐惧之情，贵族加图(Marcus Cato，前234-前149)提出：“迦太基必须被毁灭。”(57)罗马在公元前149年发起第三次布匿战争。阿庇安、狄奥多鲁斯等希腊罗马史学家提供了公元前146年迦太基战败被毁的画面，称罗马军队纵火烧毁迦太基城。(58)这次战争让迦太基的人口丧失了三分之一，一个世纪以后也未能得到完全的恢复。(59)罗马人为了抹去迦太基的历史记忆，几乎将迦太基图书馆的所有藏书都转赠给了他们在非洲的盟友努米底亚，企图只留下罗马人的历史书写。阿庇安告诉我们，罗马元老院派往迦太基的10名元老组成的代表团下令将迦太基夷为平地，任何人不得在迦太基居住，否则会受到诅咒。(60)与迦太基人斗争的这段波澜壮阔的经历，成为罗马神话不可或缺的一页。第二次布匿战争激发了罗马人的民族意识和爱国热情，罗马人开始首次书写罗马和迦太基的历史，建构了失败的迦太基和胜利的罗马的民族形象。

    斯特拉波告诉我们，迦太基于公元前146年的最后时刻在利比亚本土仍然拥有300座城市和70万城市人口的实力。(61)古代知识精英有夸大数字的习惯，斯特拉波这个数字可能有些夸张了。《剑桥古代史》提出，此时迦太基城约有20万—30万人。(62)韦尔认为，迦太基城在公元前146年毁灭之际的人口不会超过13万人，因为该城没有那么多人口的居住空间。(63)毁灭迦太基，让罗马商人和意大利商人填补了远距离贸易的空白。他们取代了迦太基商人，来到努米底亚从事贸易活动，获利颇丰。

    (三)古希腊罗马作家对迦太基人的负面形象建构

    在古希腊罗马作家的眼中，迦太基人是一群好斗的恶毒的东方入侵者，是暴虐的、虚伪的和贪婪的。波里比乌斯指出：在迦太基，任何能产生利益的东西都不会被视为可耻的，竞选官职可进行公开的贿赂；而罗马人会谴责不择手段赚钱的方式，竞选行贿是死罪。(64)狄奥多鲁斯对迦太基人和罗马人的文明冲突有详细的报道，他说迦太基人只要失败就会导致精神和意志的崩溃，这在别的民族是很难想象的。(65)他意味深长地说了一句话：“我认为有的民族对人类社会危害甚烈是理所当然的事情，如严霜和冰雪会摧毁刚刚生长的作物。”(66)狄奥多鲁斯的言外之意，迦太基人是一个有害的民族。他批评迦太基当局在国家危机之际提拔新将领，危机过后却又让他们身败名裂的做法。(67)但狄奥多鲁斯也客观地指出了罗马人的残酷、失信和傲慢，指出了罗马人在毁灭迦太基后的问题，如忧患意识的缺乏、行政官员的贪婪和对法律的无视、民意煽动家所造成的危险、内战延长带来的恐惧等问题(68)，这成为罗马共和晚期诸多矛盾之源。

    地中海西部的希腊人和迦太基人之间，存在着一道不可逾越的鸿沟。希腊人表现出一种与生俱来的优越性和激烈的排外情绪，对希腊人与其他民族在政治、文化、宗教等领域日益增强的融合趋势表现出一种对抗的态度。古希腊作家对迦太基人的态度，直接影响了罗马的知识精英。罗马人从未将自己看作希腊人，但他们已经认识到自己在民族文化的分水岭中与希腊文化同属一个阵营。这个分水岭将文明的希腊世界与野蛮人的世界区分开了，而迦太基人显然属于野蛮人世界。罗马史学家李维在关于布匿战争的记载中，一直在将罗马人的美德与迦太基人的恶习进行对比，认为汉尼拔的背信弃义超过了任何一位迦太基人。西里乌斯·伊塔利库斯(Silius Italicus，28—103)是一位罗马元老，写下以第二次布匿战争为主题的史诗《布匿战记》(Punica)，使用了“迦太基人是残忍的”这样的表述。(69)希腊罗马人对迦太基人的“他者”想象——胡言乱语的、贪婪的、不守信用的、残忍的、傲慢的、不敬神的迦太基人，成为古典史学叙述的主流。“迦太基式的信用”(fides Punica)，亦成为背信弃义的同义语。(70)

    在希腊罗马人看来，迦太基的失败与他们对待盟邦的态度有直接的关系。狄奥多鲁斯指出，迦太基人“过于残酷和苛刻”(71)，对待被征服地区经常采取掠夺、索取和高压的政策，因而盟邦痛恨迦太基的高压统治。(72)狄奥多鲁斯还说，迦太基军队在占领西西里阿克拉伽斯城之后，将神庙财产和城市的雕像、物品都运到迦太基，然后焚烧了寺庙，洗劫了全城。在占领另一座城市杰拉(Gela)之后，官兵都发了财。(73)狄奥多鲁斯强调，长期以来，迦太基的利比亚盟邦憎恨迦太基人的压榨和索取，他们一有机会就发起叛乱。(74)在叙述和希腊人争夺西西里挫败原因的时候，狄奥多鲁斯指出这是迦太基人不尊重推罗神灵的结果，认为迦太基人一开始对在推罗受到顶礼膜拜的赫拉克勒斯是尊重的，向其奉献年收入的十分之一。而在后来，他们疏忽了，所以神的恩泽就减少了。狄奥多鲁斯还强调了神对迦太基贵族用别人的孩子冒充自己的孩子献给神灵的愤怒之情。(75)

    古希腊作家普鲁塔克(Plutarchus，约46-120)和狄奥多鲁斯都提到迦太基人用儿童献祭，后者还说迦太基贵族曾用200位贵族儿童作为牺牲品献给神灵。(76)威尔·杜兰(Will Durant)赞同古典作家的这种说法，认为直到2世纪仍存在献给麦勒卡特(Melkart，推罗主神)的活童祭品。(77)《牛津古典辞书》认为从公元前8世纪晚期到公元前146年，童祭在迦太基一直存在，但在实践中一开始就用动物代替了儿童。(78)泽内达·A.拉戈金(Zenaide A.Ragozin)提出，是在与耶稣同时期的罗马皇帝提比略(Tiberius，37-41年在位)时期才彻底废除了迦太基的这个习俗。(79)《剑桥古代史》指出迦太基托非特祭坛的骨灰瓮中发现了儿童烧焦的骨头，还有两篇提到杀婴的铭文。(80)但奥贝特(M.Aubet)不同意这种杀婴献祭的观点，他认为考古学家对托非特祭坛中的人类骨殖和灰烬的分析表明，他们大多属于死婴和新生儿的，这明显带有自然死亡的味道(81)，并非来自儿童献祭所产生的非正常死亡。撒里贝姆特(Andrea Salimbeti)提出，目前的证据无法证实迦太基存在火焚活童的祭祀。(82)本文亦认为，古希腊罗马作家的记载经常存在夸大的成分和道听途说的习惯。根据目前的史料和考古材料，我们对普鲁塔克和狄奥多鲁斯关于迦太基存在活童祭观点的可靠性还无法确认，还有待于进一步的材料和研究。

    三、迦太基的文明特征和文明贡献

    作为东方闪米特人的一支，迦太基人创建了一个庞大的海上殖民和商业帝国，将文明扩展到了地中海西部广大区域甚至是大西洋沿岸，同时带动了北部非洲本土的开发，推动了东西方文明、非洲文明和外来文明的交流和融合。

    (一)迦太基对非洲的开发和文明传播

    迦太基海上商业帝国的辉煌，很容易掩盖其农业的成就。当推罗人和西顿人在非洲沿岸的巴巴里(Barbara)建立最初的停靠港时，当地居民还处于新石器时代，金属器具是从外国输入的。古代非洲(不包括希腊罗马人所说的不属于利比亚的埃及)居民也栽种过某些作物和饲养过某些牲畜，但只有到了迦太基成立和崛起以后，非洲才在迦太基的引领下有了真正的农业。从某种意义上说，迦太基也是非洲的首都。迦太基人的农业开发是成功的，他们大量使用奴隶劳动，对土著的示范和对非洲的物质繁荣都做出了贡献。它把文明传播到它所兼并的国家，超出自己的领土，传播到其附属国和同盟国中去。(83)迦太基人从西亚带来了葡萄、橄榄、无花果、石榴等，在土质和气候适宜的非洲广泛种植，迦太基的石榴被罗马人称为“迦太基苹果”。

    公元前5世纪在西西里的殖民活动受挫后，迦太基加快了开发非洲本土农业的步伐，将撒哈拉最肥沃的土地和大流沙区纳入其领土，变为农业区。从公元前3世纪末开始，谷物和葡萄园的收入就已经成为迦太基统治阶层的主要收入来源了。(84)古希腊罗马作家共66次提及迦太基著名农学家马戈(Mago)的耕作法。(85)柏柏尔人同迦太基人的接触比其他非洲民族都早，受迦太基人的影响也最大。柏柏尔人向迦太基人学会了农业，迈入了农耕文明。菲利普C.内勒(Phillip C.Naylor)认为，迦太基人懂得欣赏不同民族的文化，其复杂的文化是通过万神殿和语言来表达的，给善于接受文化的柏柏尔人留下深刻印象。(86)迦太基文化对柏柏尔人的思想和习俗的影响，成为跨文化交往的典范。努米底亚人在公元前3世纪向迦太基人学会了建设城市，引入了马戈农书所介绍的耕作方法，还吸取了迦太基的文化及宗教。(87)柏柏尔人在公元前202年建立了努米底亚王国，他们所建的城市科塔(Cirta)也出现了迦太基式的托非特祭坛。(88)努米底亚贵族热衷于同迦太基上层通婚，给子女取布匿人的名字。(89)努米底亚和迦太基的铜币设计都是一样的，二者出现了文化同化的现象。

    20世纪以来的考古挖掘表明，迦太基人对北部非洲内陆进行了深度渗透，建立了很多殖民地和临时性商业居留地，创建了一系列城市，使得迦太基文明得以在非洲广泛传播。1970年代，考古学家在非洲卡本半岛(Cap Bon Peninsula，今突尼斯东北)发现了迦太基人所建的殖民地盖赫库阿勒(Kerkouane)的遗迹。该城可能建于公元前6世纪，提供了迦太基人和利比亚土著人交往的线索。盖赫库阿勒居民敬拜的神是腓尼基人和迦太基人的神灵麦勒卡特和其子锡德(Sid)、塔尼特等。该城大部分房屋都有中央庭院，院中建有迦太基人用作洁净仪式的浴室。盖赫库阿勒居民使用迦太基语，但利比亚的元素随处可见，如利比亚土著居民的殡葬仪式。盖赫库阿勒城还发现了描写希腊英雄奥德修斯形象的雅典式黑彩陶制酒壶和爱奥尼亚式杯子，以及具有希腊风格的建筑，反映了该城多元文化的特点。(90)

    斯特拉波告诉我们，迦太基被毁后在很长一段时期内和希腊的科林斯一样，一直处于荒无人烟的状态。(91)罗马有识之士试图推出重建迦太基的计划。据阿庇安记载，在迦太基荒芜了30年后，罗马保民官盖约·格拉古(Gaius Gracchus，前154-前121)在公元前123年提出移民6000人到迦太基的计划，而不顾西庇阿在毁灭迦太基的时候诅咒迦太基将永作牧场的事实。(92)恺撒(Gaius Julius Caesar，前100-前44)曾在公元前44年重提移民迦太基的计划，以解决贫民的土地问题，但未来得及实施便遇刺身亡。奥古斯都(Augustus，前63-公元14)在公元前29年派遣移民重建迦太基，并将非洲行省(亦称阿非利加行省)的首府从乌提卡迁到迦太基。(93)罗马退役军人、殖民者、商人和工程人员接踵而来，将新迦太基建设成为罗马帝国西部最大的城市，也是罗马帝国著名的文化、教育和学术中心。斯特拉波指出，和利比亚境内的其他人和城市一样，迦太基是一座繁荣兴旺的城市。(94)在迦太基以东的港口大莱普提斯(Leptis Magna)，诞生了罗马帝国首位出身非洲的元首塞普提米乌斯·塞维鲁(Septimius Severus，146-211)，塞维鲁王朝也成为罗马帝国历史上的第一个非洲王朝。

    (二)迦太基文明与希腊文明、埃及文明和罗马文明的交流和交融

    迦太基文化受到希腊文明、埃及文明的明显影响，表现出混合性或者折中主义的特点，体现在艺术、宗教和物质文化等方面。至少从公元前6世纪起，迦太基人就将希腊和埃及的建筑风格与迦太基的建筑风格结合起来，形成了混合式的迦太基建筑艺术。

    迦太基与西西里希腊人的战争带来了文化宗教的交流，导致了二者文化融合的范围远远超出了西西里海岸地区而深入内陆，这是西西里殖民化以来的显著特征之一。从公元前4世纪起，希腊文化对迦太基的影响越来越明显，但是迦太基的文化并未失去自身传统，尤其是在语言和宗教方面。公元前4世纪早期，迦太基当局禁止迦太基人学希腊语，但却未见成效，希腊语在北部非洲成了仅次于布匿语的第二语言。(95)大量证据显示，迦太基人会说希腊语，阅读希腊著作，着希腊服饰，崇拜希腊神祇。在西西里，出现了很多具有希腊文化和腓尼基文化特色的器具和艺术品，如西西里出现的铸有六头金牛肖像的酒杯。迦太基人的墓葬中出现了大量希腊钱币，主要是希腊铜币，这可能是对外贸易的需要。斯特拉波说，在迦太基卫城柏萨的高处，有一座阿斯克勒庇俄斯(Asclepius)神庙。(96)阿斯克勒庇俄斯是希腊的医神，它在迦太基出现说明迦太基人对于希腊医神的崇拜和对祛除疾病、战胜瘟疫的渴求，亦反映希腊宗教对迦太基人的影响。

    另外，迦太基人对希腊神赫拉克勒斯的崇拜，亦反映了希腊文化对迦太基文化的影响。从某种意义上说，赫拉克勒斯崇拜具有文化的多样性与跨文化的相互关联性，比其他任何一位神都更适合古代地中海居民。从公元前6世纪起，在地中海中西部，赫拉克勒斯开始越来越多地被与迦太基的麦勒卡特联系在一起。当汉尼拔试图寻找一位天国的精神领袖以将西方的人们联合起来抗衡罗马之际，他选中的就是赫拉克勒斯-麦勒卡特。(97)泽内达·A.拉戈金强调腓尼基宗教亦对希腊(尤其是希腊大陆的东海岸)的宗教有影响(98)，但未提供有力的证据。

    在希腊人的影响下，迦太基人的丧葬方式也发生了改变，从土葬转为了火葬。在奥古斯都重建迦太基城之后，这里到处都有希腊风格的雕像。20世纪以来蓬勃发展的考古学，让我们重构古代迦太基的文化成为可能。考古学家在突尼斯圣路易山丘的斜坡上，挖掘出了为罗马人纵火毁城的灰烬层所覆盖的房屋以及迦太基的一个街区，都反映出这里具有希腊化时代的特点，如迦太基街区的房屋规格较小、由各个房间所环绕着的中央庭院作为建筑物的光源，等等。

    埃及人的宗教和来世观也对迦太基人产生了影响。迦太基人崇拜埃及的奥里西斯(Osiris)神，将迦太基主神巴尔-哈蒙(Baal Hammon)和埃及的阿蒙神相提并论。受埃及人的影响，迦太基人认为人死后会有两个灵魂。迦太基人还向埃及人学会了制作木乃伊的技术。公元前4世纪的迦太基墓葬出土的剃刀上的宗教主题图案，集中反映了埃及和迦太基的神灵和神圣符号。在迦太基人的护身符上，出现了埃及的神灵、动物形象和神圣符号。从埃及进口的圣甲虫和首饰(99)，在迦太基大受欢迎。在迦太基人的陪葬品中，出现了描绘埃及神祇和法老形象的祭品，这被认为有助于驱邪。

    迦太基公元前146年亡国后的这段历史，被称为晚期布匿或新布匿时期。虽然罗马文化成为官方文化，但迦太基人的文化并没有中断，在语言、建筑、绘画、雕刻、美术、教育等领域都有体现。迦太基艺术有其自身的特点，表现为对细节和对称的过度关注。非洲很多地方官员都使用了迦太基式的“苏菲特”的名称。迦太基神庙保留了下来，但祭司取罗马的名字，着罗马的托袈。(100)迦太基的神名改成了罗马的神名，巴尔-哈蒙变成了罗马农神萨图尔(Saturn)，塔尼特变成为了罗马天后朱诺(Juno)。巴尔-萨图尔被视为罗马的丰收之神，成为罗马非洲行省农业发展在宗教文化领域的反映。(101)罗马人将腓尼基人语言称为布匿语，迦太基战争被罗马人称为布匿战争。迦太基被罗马征服之后，拉丁语的重要性超过希腊语，成为迦太基人从小必须学习的语言，城市中受教育者两种语言都会。但在整个罗马帝国的非洲行省，布匿语仍是官方通行语言。公元前8年，非洲大莱普提斯城的迦太基籍城市长官捐助了一座纪念碑，使用了拉丁语和迦太基语的双语铭文。捐助者的名字“汉尼拔·塔帕皮乌斯·鲁福斯”，就是罗马姓氏和迦太基本名的结合。(102)在罗马帝国基督教神学家圣奥古斯丁(St.Aurelius Augustinus，354-430)写作的时代，布匿语仍出现在非洲的拉丁书信和铭文中，并一直使用到了5世纪早期，显示了语言传统强大的惯性和生命力。439年，汪达尔首领盖萨里克(Geiseric，389-477)占领了迦太基，建立了汪达尔王国，罗马-迦太基的时代结束了。(103)

    结语

    迦太基研究专家沃明顿指出：“作为城邦国家的迦太基试图统治一个帝国，并能够维持3个世纪之久的统治，比雅典人的统治还长了三分之一。”(104)在古代世界，迦太基的成功之处不仅体现在拥有的巨额财富，而且包括它统治的长久的稳定和持久，这甚至赢得了其对手的尊敬。西塞罗强调，迦太基如果没有使用智慧和政治谋略的话，就不可能维持一个帝国达600年之久。(105)亚里士多德将迦太基的政体和斯巴达、克里特的政体归为一类，认为它们是最接近理想的混合政体，高度评价迦太基设施优良、政治稳定和制度修明。(106)迦太基也是亚里士多德所研究的唯一非希腊国家。古希腊作家伊索克拉底(Isocrates，前436-前338)也称赞迦太基和斯巴达是世界上治理最好的两个民族。(107)

    财富和政体具有密切的关系。迦太基是世界历史上第一个有组织的商业国家(108)，进行了地中海商业帝国建设的第一次尝试。迦太基的财富来自它在非洲和西地中海的区域性帝国，它在海上力量保护下的海外贸易，为其维持一支强大的海军提供了财政支持。迦太基商业的成功使之把政治权力给了财富寡头，在公民人口不足的情况下供养了一支雇佣军，维护了迦太基政权的稳定。迦太基在公元前6世纪成为一个强大的国家。到公元前5世纪，随着迦太基人在西西里的布匿—希腊战争中遇挫，迦太基人在西地中海的贸易活动暂时走向了衰落。到公元前4世纪中叶，随着非洲的开发和对非洲属国资源的掠夺，迦太基的商业再次繁荣起来。在丧失西班牙的海外领地之后，迦太基仍被波里比乌斯称为世界上最富有的城市。(109)迦太基对外来文化采取了开放的态度，越来越多地受到了希腊文化的影响。

    迦太基建构了一个地中海世界商贸殖民网，也为各殖民地之间的政治、经济、社会文化的交流和非洲的开发做出了积极的贡献。在杜丹看来，迦太基的历史表现出了一种文明的肤浅和虚弱，其主要推动力是获取财富和扩展商业，其诸多胜利转瞬即逝。迦太基没有依靠经济优势去争取政治和文化的进步，未能在文化领域取得辉煌的成果。这也是迦太基在和罗马的文明较量中失败的重要原因。但我们必须看到，作为西方文明源头的古典世界从来不是希腊—罗马文明的特有成果，而是包括迦太基文明在内的不同文化与民族之间互动、交流和融合的结果。

    注释：

    ①迦太基研究先驱彻什的《非洲帝国之迦太基》(Alfre John Church,Carthage of the Empire of Africa,New York:G.P.Putnam’s Sons,1899)，论述了迦太基发展成为非洲帝国的过程。斯密斯的《迦太基和迦太基人》(R.Bosworth Smith,Carthage and the Carthaginians,London:Longmans,Green and Co.,1913)考察了迦太基、迦太基人的发展历程及其对古代世界的影响。沃明顿的《迦太基》(B.H.Warmington,Carthage,London:Robert Hale Limited,1960)从汉诺的非洲航行来追溯迦太基历史。莫斯卡蒂的《迦太基艺术与文明》(Sabatino Moscati,Carthage:Art et Civilization,Milan:Jaca Book,1983)重点介绍了迦太基的文化。皮卡德的《迦太基：从诞生到终结悲剧的布匿历史与文化研究》(Gilbert Charles Picard,Carthage:A Survey of Punic History and Culture from Its Birth to Final Tragedy,London:Sidgwick & Jackson,1987)是一部涉及政治、经济、文化和外交等内容的综合性著作。霍约斯的《迦太基人》(Dexter Hoyos,The Carthaginians,London and New York:Taylor and Francis Group Press,2010)考察了迦太基崛起的进程及商业帝国的建立。迈尔斯的《迦太基必须毁灭：古文明的兴衰》(Richard Miles,Carthage Must Be Destroyed:The Rise and Fall of An Ancient Mediterranean Civilization,London:Allen Lane,2010)借鉴了近现代研究成果，学术价值较高。特里布拉图编的《古代西西里的语言联系》(Olga Tribulato,ed.,Language and Linguistic Contact in Ancient Sicily,Cambridge,U.K.:Cambridge University Press,2012)考察了布匿语等古代西西里语言和文化。国内研究论文有数篇。陈恒的《迦太基建城日期小考》，《常熟高专学报》2001年第1期，对迦太基建城进行了考察；杜建军、刘自强的《论布匿战争爆发的原因》(2002)从政治经济文化进行了探讨。总的来说，关于迦太基的崛起等问题还有进一步研究的较大空间。

    ②Alfred J.Church,Carthage of the Empire of Africa,p.11.

    ③Dionysius of Halicarnassus,The Roman Antiquities,Loeb Classical Library,trans.Earnest Cary,Vol.1,Cambridge,Mass.:Harvard University Press,1937,p.245.本文所引古希腊罗马文献，出自“罗布古典丛书”(Loeb Classical Library)。

    ④Appian,The Punic Wars,I.1,Loeb Classical Library,Vol.1,trans.Horace White,Cambridge,Mass.:Harvard University Press,1912,p.403.

    ⑤Andrea Salimbeti,Raffaele D’ Amato,The Carthaginians,6th-2nd Century BC,Oxford:Osprey Publishing,1991,p.4.

    ⑥Appian,The Punic Wars,XIX 132,Vol.1,p.637.

    ⑦Strabo,Geography,XVII.3.14,Loeb Classical Library,Vol.8,trans.Horace Leonard Jones,Cambridge,Mass.:Harvard University Press,1928,p.183.注：斯塔德(Stadium，复数为Stadia)，古希腊长度单位。1斯塔德合625英尺约等于184.97米。

    ⑧理查德·迈尔斯：《迦太基必须灭亡：古文明的兴衰》，孟驰译，北京：社科文献出版社，2016年，第85页。

    ⑨B.H.Warmington,Carthage,London:Robert Hale Limited,1960,p.23.陈恒采用了辛塔斯所提出的公元前725年的说法，参见陈恒：《迦太基建城日期考》，《常熟高专学报》2001年第1期。

    ⑩Maria Giulia Amadasi Guzzo,”Phoenician and Punic in Sicily,” in Olga Tribulato,ed.,Language and Linguistic Contact in Ancient Sicily,Cambridge,U.K.:Cambridge University Press,2012,p.130.

    (11)B.H.Warmington,Carthage,p.22.

    (12)Benjamin W.Wells,”Business and Politics at Carthage,” The Sewanee Review,Vol.28,No.4(Oct 1920),p.507.

    (13)Olga Tribulato,”So Many Siciliies,” in Olga Tribulato,ed.,Language and Linguistic Contact in Ancient Sicily,pp.15-16.

    (14)B.H.Warmington,Carthage,p.23.

    (15)J.N.Coldstream,Geometric Greece,London:Routledge,1977,p.240.

    (16)杜丹：《古代世界经济生活》，志扬译，北京：商务印书馆，1963年，第187页。

    (17)理查德·迈尔斯：《迦太基必须灭亡：古文明的兴衰》，第37页。

    (18)Diodorus of Sicily,Library of History,XXV.10,Loeb Classical Library,Vol.11,trans.Francis R.Walton,Cambridge,Mass.:Harvard University Press,1957,p.155.

    (19)Simon Hornblower,Antony Spawforth & Esther Eidinow,eds.,The Ox ford Classical Dictionary,Oxford:Oxford University Press,2012,p.284.

    (20)Benjamin W.Wells,”Business and Politics at Carthage,” p.505.

    (21)杜丹：《古代世界经济生活》，第188页。

    (22)Pliny the Elder,Natural History,V.67,Vol.2,p.271.

    (23)Polybius,The Histories,VI.52,Loeb Classical Library,Vol.3,trans.W.R.Paton,Cambridge,Mass:Harvard University Press,1923,p.385.

    (24)Alfred J.Church,Carthage of the Empire of Africa,p.100.

    (25)Pliny the Elder,Natural History,V.8,Vol.2,p.223.

    (26)Alfred J.Church,Carthage of the Empire of Africa,p.95.

    (27)Benjamin W.Wells,”Business and Politics at Carthage,” p.503.

    (28)Thucydides,History of the Peloponnesian War,VI.2,Loeb Classical Library,Vol.3,trans.C.Forster Smith,London:W.Heinemann; New York:G.Putnam’s Sons,1919,p.183.

    (29)B.H.Warmington,Carthage,p.34.

    (30)Olga Tribulato,”So Many Siciliies,” p.14.

    (31)Diodorus of Sicily,Library of History,XXXVI.3,Loeb Classical Library,Vol.12,trans.Francis R.Walton,Cambridge,Mass.:Harvard University Press,1967,p.237.

    (32)Diodorus of Sicily,Library of History,XIV.65,Loeb Classical Library,Vol.6,trans.C.H.Oldfather,Cambridge,Mass.:Harvard University Press,1954,p.190.

    (33)B.H.Warmington,Carthage,pp.40-55.

    (34)R.Bosworth Smith,Carthage and the Carthaginians,London:Longmans,Green and Co.,1913,p.57.

    (35)Polybius,The Histories,III.37,Loeb Classical Library,Vol.2,trans.W.R.Paton,Cambridge,Mass.:Harvard University Press,1923,p.87.

    (36)Pliny the Elder,Natural History,V.1,Vol.2,p.219.

    (37)Strabo,Geography,XVII.3.1,Vol.8,p.155.

    (38)Strabo,Geography,XVI.2.38,Loeb Classical Library,Vol.7,trans.Horace Leonard Jones,Cambridge,Mass.:Harvard University Press,1930,p.287.

    (39)Diodorus of Sicily,Library of History,XX 3-6,Loeb Classical Library,Vol.10,trans.Russel M.Geer,Cambridge,Mass.:Harvard University Press,1954,pp.150-157.

    (40)Maria Eugenia Aubet,The Phoenicians and the West:Politics,Colonies and Trade,Cambridge U.K.:Cambridge University Press,2001,p.284.

    (41)Diodorus of Sicily,Library of History,XI.1,Loeb Classical Library,Vol.4,trans.C.H.Oldfather,Cambridge,Mass.:Harvard University Press,1946,p.123.

    (42)B.H.Warmington,Carthage,p.46.

    (43)R.Bosworth Smith,Carthage and the Carthaginians,p.21.

    (44)J.博德曼、N.G.L.哈蒙德等编：《剑桥古代史》，第四卷，张强等译，北京：中国社会科学出版社，2020年，第848页。

    (45)Maria Giulia Amadasi Guzzo,”Phoenician and Punic in Sicily,” p.126.

    (46)Olga Tribulato,”So Many Siciliies,” pp.17-29.

    (47)Polybius,The Histories,III.22-23,Vol.2,pp.53-55.按照波里比乌斯的说法，菲尔角在迦太基的前端，方向朝北，可能是今天的伯恩角，但是也可能是法里纳角(C.Farina)。

    (48)M.Cary,”A Forgotten Treaty between Rome and Carthage,” The Journal of Roman Studies,Vol.9,1919,p.76.

    (49)A.E.阿斯汀等编：《剑桥古代史》，第八卷，陈恒等译，北京：中国社会科学出版社，2020年，第21页。

    (50)J.博德曼、N.G.L.哈蒙德等编：《剑桥古代史》，第四卷，第492页。

    (51)A.E.阿斯汀等编：《剑桥古代史》，第八卷，第20—22页。

    (52)Polybius,The Histories,III.24,Vol.2,p.57.

    (53)J.博德曼、N.G.L.哈蒙德等编：《剑桥古代史》，第四卷，第486页。

    (54)Herodotus,Histories,IV.196,Loeb Classical Library,Vol.2,trans.A.D.Godley,Cambridge,Mass.:Harvard University Press,1926,p.399.

    (55)Polybius,The Histories,VI.52,Vol.3,p.387.

    (56)Polybius,The Histories,II.1,Vol.2,p.241.

    (57)Plutarch,Lives,”Marcus Cato,” LXX 1,Loeb Classical Library,Vol.2,trans.Bernadotte Perrin,Cambridge,Mass.:Harvard University Press,1914,p.383.

    (58)Appian,The Punic Wars,XIX.127-135,Vol.1,pp.627—637; Diodorus of Sicily,Library of History,XXXIII.24,Vol.11,p.435.

    (59)Benjamin W.Wells,”Business and Politics at Carthage,” p.518.

    (60)Appian,The Punic Wars,XX 135,Vol.1,p.643.

    (61)Strabo,Geography,XVII.3.15,Vol.8,p.185.

    (62)A.E.阿斯汀等编：《剑桥古代史》，第八卷，第172页。

    (63)Benjamin W.Wells,”Business and Politics at Carthage,” p.506.

    (64)Polybius,The Histories,VI.56,Vol.3,p.393.

    (65)Diodorus of Sicily,Library of History,XXIII.11,Vol.11,p.95.

    (66)Diodorus of Sicily,Library of History,XXVI.1,Vol.11,p.179.

    (67)Diodorus of Sicily,Library of History,XX 10,Loeb Classical Library,Vol.10,trans.Russel M.Geer,Cambridge,Mass.:Harvard University Press,1954,p.169.

    (68)Diodorus of Sicily,Library of History,XXXIV/XXXV.33,Vol.12,pp.131-133.

    (69)Silius Italicus,Punica,I,171-172,Loeb Classical Library,Vol.1,trans.J.D.Duff,Cambridge,Mass.:Harvard University Press,1927,p.17.

    (70)J.Starks,”Fides Aeneia:The Transference of Punic Stereotypes in the Aeneid,” Classical Journal,Vol.94,No.3(Feb.-Mar.1999),pp.250-260.

    (71)Diodorus of Sicily,Library of History,XIV.46,Vol.6,p.141.

    (72)Diodorus of Sicily,Library of History,XIV.76,Vol.6,p.217.

    (73)Diodorus of Sicily,Library of History,XIII.96; XⅢ,108,Loeb Classical Library,Vol.5,trans.C.H.Oldfather,Cambridge,Mass.:Harvard University Press,1956,pp.395—397; 429.

    (74)Diodorus of Sicily,Library of History,XX.3,Loeb Classical Library,Vol.10,trans.Russel M.Geer,Cambridge,Mass.:Harvard University Press,1954,p.150.

    (75)Diodorus of Sicily,Library of History,XX.14,Vol.10,p.179.

    (76)Plutarch,Moralia,171C-D,Loeb Classical Library,Vol.2,trans.F.C.Babbitt,Cambridge,Mass.:Harvard University Press,1928,p.493; Diodorus of Sicily,Library of History,XX.14,Vol.10,p.179.

    (77)威尔·杜兰：《凯撒与基督》下册，周杰译，幼狮文化公司译，北京：东方出版社，2003年，第601页。麦勒卡特是推罗人的主神，但在迦太基宗教体系中的地位却有了明显的下降，并非主神。

    (78)Simon Hornblower,Antony Spawforth & Esther Eidinow,eds.,The Ox ford Classical Dictionary,p.284.

    (79)泽内达·A.拉戈金：《亚述：从帝国的崛起到尼尼微的陷落》，吴晓真译，北京：商务印书馆，2020年，第137页。

    (80)F.W.沃克班克等编：《剑桥古代史》，第七卷第二分册，胡玉娟等译，北京：中国社会科学出版社，2020年，第567页。

    (81)María Eugenia Aubet,The Phoenicians and the West:Politics,Colonies and Trade,pp.251-252.

    (82)Andrea Salimbeti,Raffaele D’ Amato,The Carthaginians,6th-2nd Century BC,p.10.

    (83)杜丹：《古代世界经济生活》，第181—183页。

    (84)Benjamin W.Wells,”Business and Politics at Carthage,” p.514.

    (85)理查德·迈尔斯：《迦太基必须灭亡：古文明的兴衰》，第132页。

    (86)菲利普C.内勒：《北非史》，韩志斌等译，北京：中国大百科全书出版社，2013年，第25页。

    (87)Simon Hornblower,Antony Spawforth & Esther Eidinow.eds.,The Ox ford Classical Dictionary,p.284.

    (88)B.H.Warmington,Carthage,p.209.

    (89)夏尔·安德烈·朱利安：《北非史》，第一卷上册，上海新闻出版系统“五七干校”翻译组译，上海：上海人民出版社，1974年，第167页。

    (90)理查德·迈尔斯：《迦太基必须灭亡：古文明的兴衰》，第105—107页。

    (91)Strabo,Geography,XVII.3.15,Vol.8,p.185.

    (92)Appian,The Civil Wars,I.24,Loeb Classical Library,Vol.3,trans.Horace White,Cambridge,Mass.:Harvard University Press,1913,p.49.

    (93)Appian,The Punic Wars,XX.136,Vol.1,p.645.

    (94)Strabo,Geography,XVII.3.15,Vol.8,p.185.

    (95)徐晓旭：《“罗马和平”下不同文化的相遇》，《光明日报》2020年9月21日，第14版。

    (96)Strabo,Geography,XVII.3.14,Vol.8,p.185.

    (97)理查德·迈尔斯：《迦太基必须灭亡：古文明的兴衰》，第27页。

    (98)泽内达·A.拉戈金：《亚述：从帝国的崛起到尼尼微的陷落》，第144页。

    (99)F.W.沃克班克等编：《剑桥古代史》第七卷第二分册，第564页。

    (100)夏尔·安德烈·朱利安：《北非史》，第129页。

    (101)B.H.Warmington,Carthage,p.42.

    (102)理查德·迈尔斯：《迦太基必须灭亡：古文明的兴衰》，第370页。

    (103)Mattew Bunson.ed.,A Dictionary of the Roman Empire,Oxford:Oxford University Press,1995,p.98.

    (104)B.H.Warmington,Carthage,p.42.

    (105)Cicero,De Re Publica,I,”Fragments,” 3,Loeb Classical Library,trans.C.W.Keyes,Cambridge,Mass.:Harvard University Press,1928,p.109.

    (106)亚里士多德：《政治学》，吴寿彭译，北京：商务印书馆，1965年，第106页。

    (107)F.W.沃克班克等编：《剑桥古代史》，第七卷第二分册，第537页。

    (108)Benjamin W.Wells,”Business and Politics at Carthage,” p.499.

    (109)Polybius,The Histories,XVIII.35,Loeb Classical Library,Vol.5,trans.W.R.Paton,Cambridge,Mass.:Harvard University Press,1926,p.163.

    本文转自《外国问题研究》2023年第2期

  • Yuval Noah Harari 《Nexus》

    Contents
    PROLOGUE
    PART I: Human Networks
    CHAPTER 1: What Is Information?
    CHAPTER 2: Stories: Unlimited Connections
    CHAPTER 3: Documents: The Bite of the Paper Tigers
    CHAPTER 4: Errors: The Fantasy of Infallibility
    CHAPTER 5: Decisions: A Brief History of Democracy and Totalitarianism
    PART II: The Inorganic Network
    CHAPTER 6: The New Members: How Computers Are Different from Printing Presses
    CHAPTER 7: Relentless: The Network Is Always On
    CHAPTER 8: Fallible: The Network Is Often Wrong
    PART III: Computer Politics
    CHAPTER 9: Democracies: Can We Still Hold a Conversation?
    CHAPTER 10: Totalitarianism: All Power to the Algorithms?
    CHAPTER 11: The Silicon Curtain: Global Empire or Global Split?
    EPILOGUE

    Prologue

    We have named our species Homo sapiens—the wise human. But it is debatable how well we have lived up to the name.
    Over the last 100,000 years, we Sapiens have certainly accumulated enormous power. Just listing all our discoveries, inventions, and conquests would fill volumes. But power isn’t wisdom, and after 100,000 years of discoveries, inventions, and conquests humanity has pushed itself into an existential crisis. We are on the verge of ecological collapse, caused by the misuse of our own power. We are also busy creating new technologies like artificial intelligence (AI) that have the potential to escape our control and enslave or annihilate us. Yet instead of our species uniting to deal with these existential challenges, international tensions are rising, global cooperation is becoming more difficult, countries are stockpiling doomsday weapons, and a new world war does not seem impossible.
    If we Sapiens are so wise, why are we so self-destructive?
    At a deeper level, although we have accumulated so much information about everything from DNA molecules to distant galaxies, it doesn’t seem that all this information has given us an answer to the big questions of life: Who are we? What should we aspire to? What is a good life, and how should we live it? Despite the stupendous amounts of information at our disposal, we are as susceptible as our ancient ancestors to fantasy and delusion. Nazism and Stalinism are but two recent examples of the mass insanity that occasionally engulfs even modern societies. Nobody disputes that humans today have a lot more information and power than in the Stone Age, but it is far from certain that we understand ourselves and our role in the universe much better.

    Why are we so good at accumulating more information and power, but far less successful at acquiring wisdom? Throughout history many traditions have believed that some fatal flaw in our nature tempts us to pursue powers we don’t know how to handle. The Greek myth of Phaethon told of a boy who discovers that he is the son of Helios, the sun god. Wishing to prove his divine origin, Phaethon demands the privilege of driving the chariot of the sun. Helios warns Phaethon that no human can control the celestial horses that pull the solar chariot. But Phaethon insists, until the sun god relents. After rising proudly in the sky, Phaethon indeed loses control of the chariot. The sun veers off course, scorching all vegetation, killing numerous beings, and threatening to burn the earth itself. Zeus intervenes and strikes Phaethon with a thunderbolt. The conceited human drops from the sky like a falling star, himself on fire. The gods reassert control of the sky and save the world.

    Two thousand years later, when the Industrial Revolution was making its first steps and machines began replacing humans in numerous tasks, Johann Wolfgang von Goethe published a similar cautionary tale titled “The Sorcerer’s Apprentice.” Goethe’s poem (later popularized as a Walt Disney animation starring Mickey Mouse) tells how an old sorcerer leaves a young apprentice in charge of his workshop and gives him some chores to tend to while he is gone, like fetching water from the river. The apprentice decides to make things easier for himself and, using one of the sorcerer’s spells, enchants a broom to fetch the water for him. But the apprentice doesn’t know how to stop the broom, which relentlessly fetches more and more water, threatening to flood the workshop. In panic, the apprentice cuts the enchanted broom in two with an ax, only to see each half become another broom. Now two enchanted brooms are inundating the workshop with water. When the old sorcerer returns, the apprentice pleads for help: “The spirits that I summoned, I now cannot rid myself of again.” The sorcerer immediately breaks the spell and stops the flood. The lesson to the apprentice—and to humanity—is clear: never summon powers you cannot control.

    What do the cautionary fables of the apprentice and of Phaethon tell us in the twenty-first century? We humans have obviously refused to heed their warnings. We have already driven the earth’s climate out of balance and have summoned billions of enchanted brooms, drones, chatbots, and other algorithmic spirits that may escape our control and unleash a flood of unintended consequences.

    What should we do, then? The fables offer no answers, other than to wait for some god or sorcerer to save us. This, of course, is an extremely dangerous message. It encourages people to abdicate responsibility and put their faith in gods and sorcerers instead. Even worse, it fails to appreciate that gods and sorcerers are themselves a human invention—just like chariots, brooms, and algorithms. The tendency to create powerful things with unintended consequences started not with the invention of the steam engine or AI but with the invention of religion. Prophets and theologians have repeatedly summoned powerful spirits that were supposed to bring love and joy but ended up flooding the world with blood.

    The Phaethon myth and Goethe’s poem fail to provide useful advice because they misconstrue the way humans gain power. In both fables, a single human acquires enormous power, but is then corrupted by hubris and greed. The conclusion is that our flawed individual psychology makes us abuse power. What this crude analysis misses is that human power is never the outcome of individual initiative. Power always stems from cooperation between large numbers of humans.

    Accordingly, it isn’t our individual psychology that causes us to abuse power. After all, alongside greed, hubris, and cruelty, humans are also capable of love, compassion, humility, and joy. True, among the worst members of our species, greed and cruelty reign supreme and lead bad actors to abuse power. But why would human societies choose to entrust power to their worst members? Most Germans in 1933, for example, were not psychopaths. So why did they vote for Hitler?

    Our tendency to summon powers we cannot control stems not from individual psychology but from the unique way our species cooperates in large numbers. The main argument of this book is that humankind gains enormous power by building large networks of cooperation, but the way these networks are built predisposes them to use power unwisely. Our problem, then, is a network problem.

    Even more specifically, it is an information problem. Information is the glue that holds networks together. But for tens of thousands of years, Sapiens built and maintained large networks by inventing and spreading fictions, fantasies, and mass delusions—about gods, about enchanted broomsticks, about AI, and about a great many other things. While each individual human is typically interested in knowing the truth about themselves and the world, large networks bind members and create order by relying on fictions and fantasies. That’s how we got, for example, to Nazism and Stalinism. These were exceptionally powerful networks, held together by exceptionally deluded ideas. As George Orwell famously put it, ignorance is strength.

    The fact that the Nazi and Stalinist regimes were founded on cruel fantasies and shameless lies did not make them historically exceptional, nor did it preordain them to collapse. Nazism and Stalinism were two of the strongest networks humans ever created. In late 1941 and early 1942, the Axis powers came within reach of winning World War II. Stalin eventually emerged as the victor of that war,1 and in the 1950s and 1960s he and his heirs also had a reasonable chance of winning the Cold War. By the 1990s liberal democracies had gained the upper hand, but this now seems like a temporary victory. In the twenty-first century, some new totalitarian regime may well succeed where Hitler and Stalin failed, creating an all-powerful network that could prevent future generations from even attempting to expose its lies and fictions. We should not assume that delusional networks are doomed to failure. If we want to prevent their triumph, we will have to do the hard work ourselves.

    THE NAIVE VIEW OF INFORMATION

    It is difficult to appreciate the strength of delusional networks because of a broader misunderstanding about how big information networks—whether delusional or not—operate. This misunderstanding is encapsulated in something I call “the naive view of information.” While fables like the myth of Phaethon and “The Sorcerer’s Apprentice” present an overly pessimistic view of individual human psychology, the naive view of information disseminates an overly optimistic view of large-scale human networks.

    The naive view argues that by gathering and processing much more information than individuals can, big networks achieve a better understanding of medicine, physics, economics, and numerous other fields, which makes the network not only powerful but also wise. For example, by gathering information on pathogens, pharmaceutical companies and health-care services can determine the true causes of many diseases, which enables them to develop more effective medicines and to make wiser decisions about their usage. This view posits that in sufficient quantities information leads to truth, and truth in turn leads to both power and wisdom. Ignorance, in contrast, seems to lead nowhere. While delusional or deceitful networks might occasionally arise in moments of historical crisis, in the long term they are bound to lose to more clear-sighted and honest rivals. A health-care service that ignores information about pathogens, or a pharmaceutical giant that deliberately spreads disinformation, will ultimately lose out to competitors that make wiser use of information. The naive view thus implies that delusional networks must be aberrations and that big networks can usually be trusted to handle power wisely.

    The naive view of information

    Of course, the naive view acknowledges that many things can go wrong on the path from information to truth. We might make honest mistakes in gathering and processing the information. Malicious actors motivated by greed or hate might hide important facts or try to deceive us. As a result, information sometimes leads to error rather than truth. For example, partial information, faulty analysis, or a disinformation campaign might lead even experts to misidentify the true cause of a particular disease.

    However, the naive view assumes that the antidote to most problems we encounter in gathering and processing information is gathering and processing even more information. While we are never completely safe from error, in most cases more information means greater accuracy. A single doctor wishing to identify the cause of an epidemic by examining a single patient is less likely to succeed than thousands of doctors gathering data on millions of patients. And if the doctors themselves conspire to hide the truth, making medical information more freely available to the public and to investigative journalists will eventually reveal the scam. According to this view, the bigger the information network, the closer it must be to the truth.

    Naturally, even if we analyze information accurately and discover important truths, this does not guarantee we will use the resulting capabilities wisely. Wisdom is commonly understood to mean “making right decisions,” but what “right” means depends on value judgments that differ between diverse people, cultures, or ideologies. Scientists who discover a new pathogen may develop a vaccine to protect people. But if the scientists—or their political overlords—believe in a racist ideology that advocates that some races are inferior and should be exterminated, the new medical knowledge might be used to develop a biological weapon that kills millions.

    In this case too, the naive view of information holds that additional information offers at least a partial remedy. The naive view thinks that disagreements about values turn out on closer inspection to be the fault of either the lack of information or deliberate disinformation. According to this view, racists are ill-informed people who just don’t know the facts of biology and history. They think that “race” is a valid biological category, and they have been brainwashed by bogus conspiracy theories. The remedy to racism is therefore to provide people with more biological and historical facts. It may take time, but in a free market of information sooner or later truth will prevail.

    The naive view is of course more nuanced and thoughtful than can be explained in a few paragraphs, but its core tenet is that information is an essentially good thing, and the more we have of it, the better. Given enough information and enough time, we are bound to discover the truth about things ranging from viral infections to racist biases, thereby developing not only our power but also the wisdom necessary to use that power well.

    This naive view justifies the pursuit of ever more powerful information technologies and has been the semiofficial ideology of the computer age and the internet. In June 1989, a few months before the fall of the Berlin Wall and of the Iron Curtain, Ronald Reagan declared that “the Goliath of totalitarian control will rapidly be brought down by the David of the microchip” and that “the biggest of Big Brothers is increasingly helpless against communications technology.… Information is the oxygen of the modern age.… It seeps through the walls topped with barbed wire. It wafts across the electrified, booby-trapped borders. Breezes of electronic beams blow through the Iron Curtain as if it was lace.”2 In November 2009, Barack Obama spoke in the same spirit on a visit to Shanghai, telling his Chinese hosts, “I am a big believer in technology and I’m a big believer in openness when it comes to the flow of information. I think that the more freely information flows, the stronger the society becomes.”3

    Entrepreneurs and corporations have often expressed similarly rosy views of information technology. Already in 1858 an editorial in The New Englander about the invention of the telegraph stated, “It is impossible that old prejudices and hostilities should longer exist, while such an instrument has been created for an exchange of thought between all the nations of the earth.”4 Nearly two centuries and two world wars later, Mark Zuckerberg said that Facebook’s goal “is to help people to share more in order to make the world more open and to help promote understanding between people.”5

    In his 2024 book, The Singularity Is Nearer, the eminent futurologist and entrepreneur Ray Kurzweil surveys the history of information technology and concludes that “the reality is that nearly every aspect of life is getting progressively better as a result of exponentially improving technology.” Looking back at the grand sweep of human history, he cites examples like the invention of the printing press to argue that by its very nature information technology tends to spawn “a virtuous circle advancing nearly every aspect of human well-being, including literacy, education, wealth, sanitation, health, democratization and reduction in violence.”6

    The naive view of information is perhaps most succinctly captured in Google’s mission statement “to organize the world’s information and make it universally accessible and useful.” Google’s answer to Goethe’s warnings is that while a single apprentice pilfering his master’s secret spell book is likely to cause disaster, when a lot of apprentices are given free access to all the world’s information, they will not only create useful enchanted brooms but also learn to handle them wisely.

    GOOGLE VERSUS GOETHE

    It must be stressed that there are numerous cases when having more information has indeed enabled humans to understand the world better and to make wiser use of their power. Consider, for example, the dramatic reduction in child mortality. Johann Wolfgang von Goethe was the eldest of seven siblings, but only he and his sister Cornelia got to celebrate their seventh birthday. Disease carried off their brother Hermann Jacob at age six, their sister Catharina Elisabeth at age four, their sister Johanna Maria at age two, their brother Georg Adolf at age eight months, and a fifth, unnamed brother was stillborn. Cornelia then died from disease aged twenty-six, leaving Johann Wolfgang as the sole survivor from their family.7

    Johann Wolfgang von Goethe went on to have five children of his own, of whom all but the eldest son—August—died within two weeks of their birth. In all probability the cause was incompatibility between the blood groups of Goethe and his wife, Christiane, which after the first successful pregnancy led the mother to develop antibodies to the fetal blood. This condition, known as rhesus disease, is nowadays treated so effectively that the mortality rate is less than 2 percent, but in the 1790s it had an average mortality rate of 50 percent, and for Goethe’s four younger children it was a death sentence.8

    Altogether in the Goethe family—a well-to-do German family in the late eighteenth century—the child survival rate was an abysmal 25 percent. Only three out of twelve children reached adulthood. This horrendous statistic was not exceptional. Around the time Goethe wrote “The Sorcerer’s Apprentice” in 1797, it is estimated that only about 50 percent of German children reached age fifteen,9 and the same was probably true in most other parts of the world.10 By 2020, 95.6 percent of children worldwide lived beyond their fifteenth birthday,11 and in Germany that figure was 99.5 percent.12 This momentous achievement would not have been possible without collecting, analyzing, and sharing massive amounts of medical data about things like blood groups. In this case, then, the naive view of information proved to be correct.

    However, the naive view of information sees only part of the picture, and the history of the modern age was not just about reducing child mortality. In recent generations humanity has experienced the greatest increase ever in both the amount and the speed of our information production. Every smartphone contains more information than the ancient Library of Alexandria13 and enables its owner to instantaneously connect to billions of other people throughout the world. Yet with all this information circulating at breathtaking speeds, humanity is closer than ever to annihilating itself.

    Despite—or perhaps because of—our hoard of data, we are continuing to spew greenhouse gases into the atmosphere, pollute rivers and oceans, cut down forests, destroy entire habitats, drive countless species to extinction, and jeopardize the ecological foundations of our own species. We are also producing ever more powerful weapons of mass destruction, from thermonuclear bombs to doomsday viruses. Our leaders don’t lack information about these dangers, yet instead of collaborating to find solutions, they are edging closer to a global war.

    Would having even more information make things better—or worse? We will soon find out. Numerous corporations and governments are in a race to develop the most powerful information technology in history—AI. Some leading entrepreneurs, like the American investor Marc Andreessen, believe that AI will finally solve all of humanity’s problems. On June 6, 2023, Andreessen published an essay titled “Why AI Will Save the World,” peppered with bold statements like “I am here to bring the good news: AI will not destroy the world, and in fact may save it” and “AI can make everything we care about better.” He concluded, “The development and proliferation of AI—far from a risk that we should fear—is a moral obligation that we have to ourselves, to our children, and to our future.”14

    Ray Kurzweil concurs, arguing in The Singularity Is Nearer that “AI is the pivotal technology that will allow us to meet the pressing challenges that confront us, including overcoming disease, poverty, environmental degradation, and all of our human frailties. We have a moral imperative to realize this promise of new technologies.” Kurzweil is keenly aware of the technology’s potential perils, and analyzes them at length, but believes they could be mitigated successfully.15

    Others are more skeptical. Not only philosophers and social scientists but also many leading AI experts and entrepreneurs like Yoshua Bengio, Geoffrey Hinton, Sam Altman, Elon Musk, and Mustafa Suleyman have warned the public that AI could destroy our civilization.16 A 2024 article co-authored by Bengio, Hinton, and numerous other experts noted that “unchecked AI advancement could culminate in a large-scale loss of life and the biosphere, and the marginalization or even extinction of humanity.”17 In a 2023 survey of 2,778 AI researchers, more than a third gave at least a 10 percent chance to advanced AI leading to outcomes as bad as human extinction.18 In 2023 close to thirty governments—including those of China, the United States, and the U.K.—signed the Bletchley Declaration on AI, which acknowledged that “there is potential for serious, even catastrophic, harm, either deliberate or unintentional, stemming from the most significant capabilities of these AI models.”19 By using such apocalyptic terms, experts and governments have no wish to conjure a Hollywood image of killer robots running in the streets and shooting people. Such a scenario is unlikely, and it merely distracts people from the real dangers. Rather, experts warn about two other scenarios.

    First, the power of AI could supercharge existing human conflicts, dividing humanity against itself. Just as in the twentieth century the Iron Curtain divided the rival powers in the Cold War, so in the twenty-first century the Silicon Curtain—made of silicon chips and computer codes rather than barbed wire—might come to divide rival powers in a new global conflict. Because the AI arms race will produce ever more destructive weapons, even a small spark might ignite a cataclysmic conflagration.

    Second, the Silicon Curtain might come to divide not one group of humans from another but rather all humans from our new AI overlords. No matter where we live, we might find ourselves cocooned by a web of unfathomable algorithms that manage our lives, reshape our politics and culture, and even reengineer our bodies and minds—while we can no longer comprehend the forces that control us, let alone stop them. If a twenty-first-century totalitarian network succeeds in conquering the world, it may be run by nonhuman intelligence, rather than by a human dictator. People who single out China, Russia, or a post-democratic United States as their main source for totalitarian nightmares misunderstand the danger. In fact, Chinese, Russians, Americans, and all other humans are together threatened by the totalitarian potential of nonhuman intelligence.

    Given the magnitude of the danger, AI should be of interest to all human beings. While not everyone can become an AI expert, we should all keep in mind that AI is the first technology in history that can make decisions and create new ideas by itself. All previous human inventions have empowered humans, because no matter how powerful the new tool was, the decisions about its usage always remained in our hands. Knives and bombs do not themselves decide whom to kill. They are dumb tools, lacking the intelligence necessary to process information and make independent decisions. In contrast, AI has the required intelligence to process information by itself, and therefore replace humans in decision making.

    Its mastery of information also enables AI to independently generate new ideas, in fields ranging from music to medicine. Gramophones played our music, and microscopes revealed the secrets of our cells, but gramophones couldn’t compose new symphonies, and microscopes couldn’t synthesize new drugs. AI is already capable of producing art and making scientific discoveries by itself. In the next few decades, it will likely gain the ability even to create new life-forms, either by writing genetic code or by inventing an inorganic code animating inorganic entities.

    Even at the present moment, in the embryonic stage of the AI revolution, computers already make decisions about us—whether to give us a mortgage, to hire us for a job, to send us to prison. This trend will only increase and accelerate, making it more difficult to understand our own lives. Can we trust computer algorithms to make wise decisions and create a better world? That’s a much bigger gamble than trusting an enchanted broom to fetch water. And it is more than just human lives we are gambling on. AI could alter the course not just of our species’ history but of the evolution of all life-forms.

    WEAPONIZING INFORMATION

    In 2016, I published Homo Deus, a book that highlighted some of the dangers posed to humanity by the new information technologies. That book argued that the real hero of history has always been information, rather than Homo sapiens, and that scientists increasingly understand not just history but also biology, politics, and economics in terms of information flows. Animals, states, and markets are all information networks, absorbing data from the environment, making decisions, and releasing data back. The book warned that while we hope better information technology will give us health, happiness, and power, it may actually take power away from us and destroy both our physical and our mental health. Homo Deus hypothesized that if humans aren’t careful, we might dissolve within the torrent of information like a clump of earth within a gushing river, and that in the grand scheme of things humanity will turn out to have been just a ripple within the cosmic dataflow.

    In the years since Homo Deus was published, the pace of change has only accelerated, and power has indeed been shifting from humans to algorithms. Many of the scenarios that sounded like science fiction in 2016—such as algorithms that can create art, masquerade as human beings, make crucial life decisions about us, and know more about us than we know about ourselves—are everyday realities in 2024.

    Many other things have changed since 2016. The ecological crisis has intensified, international tensions have escalated, and a populist wave has undermined the cohesion of even the most robust democracies. Populism has also mounted a radical challenge to the naive view of information. Populist leaders such as Donald Trump and Jair Bolsonaro, and populist movements and conspiracy theories such as QAnon and the anti-vaxxers, have argued that all traditional institutions that gain authority by claiming to gather information and discover truth are simply lying. Bureaucrats, judges, doctors, mainstream journalists, and academic experts are elite cabals that have no interest in the truth and are deliberately spreading disinformation to gain power and privileges for themselves at the expense of “the people.” The rise of politicians like Trump and movements like QAnon has a specific political context, unique to the conditions of the United States in the late 2010s. But populism as an antiestablishment worldview long predated Trump and is relevant to numerous other historical contexts now and in the future. In a nutshell, populism views information as a weapon.20

    The populist view of information

    In its more extreme versions, populism posits that there is no objective truth at all and that everyone has “their own truth,” which they wield to vanquish rivals. According to this worldview, power is the only reality. All social interactions are power struggles, because humans are interested only in power. The claim to be interested in something else—like truth or justice—is nothing more than a ploy to gain power. Whenever and wherever populism succeeds in disseminating the view of information as a weapon, language itself is undermined. Nouns like “facts” and adjectives like “accurate” and “truthful” become elusive. Such words are not taken as pointing to a common objective reality. Rather, any talk of “facts” or “truth” is bound to prompt at least some people to ask, “Whose facts and whose truth are you referring to?”

    It should be stressed that this power-focused and deeply skeptical view of information isn’t a new phenomenon and it wasn’t invented by anti-vaxxers, flat-earthers, Bolsonaristas, or Trump supporters. Similar views have been propagated long before 2016, including by some of humanity’s brightest minds.21 In the late twentieth century, for example, intellectuals from the radical left like Michel Foucault and Edward Said claimed that scientific institutions like clinics and universities are not pursuing timeless and objective truths but are instead using power to determine what counts as truth, in the service of capitalist and colonialist elites. These radical critiques occasionally went as far as arguing that “scientific facts” are nothing more than a capitalist or colonialist “discourse” and that people in power can never be really interested in truth and can never be trusted to recognize and correct their own mistakes.22

    This particular line of radical leftist thinking goes back to Karl Marx, who argued in the mid-nineteenth century that power is the only reality, that information is a weapon, and that elites who claim to be serving truth and justice are in fact pursuing narrow class privileges. In the words of the 1848 Communist Manifesto, “The history of all hitherto existing societies is the history of class struggles. Freeman and slave, patrician and plebeian, lord and serf, guildmaster and journeyman, in a word, oppressor and oppressed stood in constant opposition to one another, carried on an uninterrupted, now hidden, now open, fight.” This binary interpretation of history implies that every human interaction is a power struggle between oppressors and oppressed. Accordingly, whenever anyone says anything, the question to ask isn’t, “What is being said? Is it true?” but rather, “Who is saying this? Whose privileges does it serve?”

    Of course, right-wing populists such as Trump and Bolsonaro are unlikely to have read Foucault or Marx, and indeed present themselves as fiercely anti-Marxist. They also greatly differ from Marxists in their suggested policies in fields like taxation and welfare. But their basic view of society and of information is surprisingly Marxist, seeing all human interactions as a power struggle between oppressors and oppressed. For example, in his inaugural address in 2017 Trump announced that “a small group in our nation’s capital has reaped the rewards of government while the people have borne the cost.”23 Such rhetoric is a staple of populism, which the political scientist Cas Mudde has described as an “ideology that considers society to be ultimately separated into two homogeneous and antagonistic groups, ‘the pure people’ versus ‘the corrupt elite.’ ”24 Just as Marxists claimed that the media functions as a mouthpiece for the capitalist class, and that scientific institutions like universities spread disinformation in order to perpetuate capitalist control, populists accuse these same institutions of working to advance the interests of the “corrupt elites” at the expense of “the people.”

    Present-day populists also suffer from the same incoherency that plagued radical antiestablishment movements in previous generations. If power is the only reality, and if information is just a weapon, what does it imply about the populists themselves? Are they too interested only in power, and are they too lying to us to gain power?

    Populists have sought to extricate themselves from this conundrum in two different ways. Some populist movements claim adherence to the ideals of modern science and to the traditions of skeptical empiricism. They tell people that indeed you should never trust any institutions or figures of authority—including self-proclaimed populist parties and politicians. Instead, you should “do your own research” and trust only what you can directly observe by yourself.25 This radical empiricist position implies that while large-scale institutions like political parties, courts, newspapers, and universities can never be trusted, individuals who make the effort can still find the truth by themselves.

    This approach may sound scientific and may appeal to free-spirited individuals, but it leaves open the question of how human communities can cooperate to build health-care systems or pass environmental regulations, which demand large-scale institutional organization. Is a single individual capable of doing all the necessary research to decide whether the earth’s climate is heating up and what should be done about it? How would a single person go about collecting climate data from throughout the world, not to mention obtaining reliable records from past centuries? Trusting only “my own research” may sound scientific, but in practice it amounts to believing that there is no objective truth. As we shall see in chapter 4, science is a collaborative institutional effort rather than a personal quest.

    An alternative populist solution is to abandon the modern scientific ideal of finding the truth via “research” and instead go back to relying on divine revelation or mysticism. Traditional religions like Christianity, Islam, and Hinduism have typically characterized humans as untrustworthy power-hungry creatures who can access the truth only thanks to the intervention of a divine intelligence. In the 2010s and early 2020s populist parties from Brazil to Turkey and from the United States to India have aligned themselves with such traditional religions. They have expressed radical doubt about modern institutions while declaring complete faith in ancient scriptures. The populists claim that the articles you read in The New York Times or in Science are just an elitist ploy to gain power, but what you read in the Bible, the Quran, or the Vedas is absolute truth.26

    A variation on this theme calls on people to put their trust in charismatic leaders like Trump and Bolsonaro, who are depicted by their supporters either as the messengers of God27 or as possessing a mystical bond with “the people.” While ordinary politicians lie to the people in order to gain power for themselves, the charismatic leader is the infallible mouthpiece of the people who exposes all the lies.28 One of the recurrent paradoxes of populism is that it starts by warning us that all human elites are driven by a dangerous hunger for power, but often ends by entrusting all power to a single ambitious human.

    We will explore populism at greater depth in chapter 5, but at this point it is important to note that populists are eroding trust in large-scale institutions and international cooperation just when humanity confronts the existential challenges of ecological collapse, global war, and out-of-control technology. Instead of trusting complex human institutions, populists give us the same advice as the Phaethon myth and “The Sorcerer’s Apprentice”: “Trust God or the great sorcerer to intervene and make everything right again.” If we take this advice, we’ll likely find ourselves in the short term under the thumb of the worst kind of power-hungry humans, and in the long term under the thumb of new AI overlords. Or we might find ourselves nowhere at all, as Earth becomes inhospitable for human life.

    If we wish to avoid relinquishing power to a charismatic leader or an inscrutable AI, we must first gain a better understanding of what information is, how it helps to build human networks, and how it relates to truth and power. Populists are right to be suspicious of the naive view of information, but they are wrong to think that power is the only reality and that information is always a weapon. Information isn’t the raw material of truth, but it isn’t a mere weapon, either. There is enough space between these extremes for a more nuanced and hopeful view of human information networks and of our ability to handle power wisely. This book is dedicated to exploring that middle ground.

    THE ROAD AHEAD

    The first part of this book surveys the historical development of human information networks. It doesn’t attempt to present a comprehensive century-by-century account of information technologies like script, printing presses, and radio. Instead, by studying a few examples, it explores key dilemmas that people in all eras faced when trying to construct information networks, and it examines how different answers to these dilemmas shaped contrasting human societies. What we usually think of as ideological and political conflicts often turn out to be clashes between opposing types of information networks.

    Part 1 begins by examining two principles that have been essential for large-scale human information networks: mythology and bureaucracy. Chapters 2 and 3 describe how large-scale information networks—from ancient kingdoms to present-day states—have relied on both mythmakers and bureaucrats. The stories of the Bible, for example, were essential for the Christian Church, but there would have been no Bible if church bureaucrats hadn’t curated, edited, and disseminated these stories. A difficult dilemma for every human network is that mythmakers and bureaucrats tend to pull in different directions. Institutions and societies are often defined by the balance they manage to find between the conflicting needs of their mythmakers and their bureaucrats. The Christian Church itself split into rival churches, like the Catholic and Protestant churches, which struck different balances between mythology and bureaucracy.

    Chapter 4 then focuses on the problem of erroneous information and on the benefits and drawbacks of maintaining self-correcting mechanisms, such as independent courts or peer-reviewed journals. The chapter contrasts institutions that relied on weak self-correcting mechanisms, like the Catholic Church, with institutions that developed strong self-correcting mechanisms, like scientific disciplines. Weak self-correcting mechanisms sometimes result in historical calamities like the early modern European witch hunts, while strong self-correcting mechanisms sometimes destabilize the network from within. Judged in terms of longevity, spread, and power, the Catholic Church has been perhaps the most successful institution in human history, despite—or perhaps because of—the relative weakness of its self-correcting mechanisms.

    After part 1 surveys the roles of mythology and bureaucracy, and the contrast between strong and weak self-correcting mechanisms, chapter 5 concludes the historical discussion by focusing on another contrast—between distributed and centralized information networks. Democratic systems allow information to flow freely along many independent channels, whereas totalitarian systems strive to concentrate information in one hub. Each choice has both advantages and shortcomings. Understanding political systems like the United States and the U.S.S.R. in terms of information flows can explain much about their differing trajectories.

    This historical part of the book is crucial for understanding present-day developments and future scenarios. The rise of AI is arguably the biggest information revolution in history. But we cannot understand it unless we compare it with its predecessors. History isn’t the study of the past; it is the study of change. History teaches us what remains the same, what changes, and how things change. This is as relevant to information revolutions as to every other kind of historical transformation. Thus, understanding the process through which the allegedly infallible Bible was canonized provides valuable insight about present-day claims for AI infallibility. Similarly, studying the early modern witch hunts and Stalin’s collectivization offers stark warnings about what might go wrong as we give AIs greater control over twenty-first-century societies. A deep knowledge of history is also vital to understand what is new about AI, how it is fundamentally different from printing presses and radio sets, and in what specific ways future AI dictatorship could be very unlike anything we have seen before.

    The book doesn’t argue that studying the past enables us to predict the future. As emphasized repeatedly in the following pages, history is not deterministic, and the future will be shaped by the choices we all make in coming years. The whole point of writing this book is that by making informed choices, we can prevent the worst outcomes. If we cannot change the future, why waste time discussing it?

    Building upon the historical survey in part 1, the book’s second part—“The Inorganic Network”—examines the new information network we are creating today, focusing on the political implications of the rise of AI. Chapters 6–8 discuss recent examples from throughout the world—such as the role of social media algorithms in instigating ethnic violence in Myanmar in 2016–17—to explain in what ways AI is different from all previous information technologies. Examples are taken mostly from the 2010s rather than the 2020s, because we have gained a modicum of historical perspective on events of the 2010s.

    Part 2 argues that we are creating an entirely new kind of information network, without pausing to reckon with its implications. It emphasizes the shift from organic to inorganic information networks. The Roman Empire, the Catholic Church, and the U.S.S.R. all relied on carbon-based brains to process information and make decisions. The silicon-based computers that dominate the new information network function in radically different ways. For better or worse, silicon chips are free from many of the limitations that organic biochemistry imposes on carbon neurons. Silicon chips can create spies that never sleep, financiers that never forget, and despots that never die. How will this change society, economics, and politics?

    The third and final part of the book—“Computer Politics”—examines how different kinds of societies might deal with the threats and promises of the inorganic information network. Will carbon-based life-forms like us have a chance of understanding and controlling the new information network? As noted above, history isn’t deterministic, and for at least a few more years we Sapiens still have the power to shape our future.

    Accordingly, chapter 9 explores how democracies might deal with the inorganic network. How, for example, can flesh-and-blood politicians make financial decisions if the financial system is increasingly controlled by AI and the very meaning of money comes to depend on inscrutable algorithms? How can democracies maintain a public conversation about anything—be it finance or gender—if we can no longer know whether we are talking with another human or with a chatbot masquerading as a human?

    Chapter 10 explores the potential impact of the inorganic network on totalitarianism. While dictators would be happy to get rid of all public conversations, they have their own fears of AI. Autocracies are based on terrorizing and censoring their own agents. But how can a human dictator terrorize an AI, censor its unfathomable processes, or prevent it from seizing power to itself?

    Finally, chapter 11 explores how the new information network could influence the balance of power between democratic and totalitarian societies on the global level. Will AI tilt the balance decisively in favor of one camp? Will the world split into hostile blocs whose rivalry makes all of us easy prey for an out-of-control AI? Or can we unite in defense of our common interests?

    But before we explore the past, present, and possible futures of information networks, we need to start with a deceptively simple question. What exactly is information?

    PART I  Human Networks

    CHAPTER 1 What Is Information?

    It is always tricky to define fundamental concepts. Since they are the basis for everything that follows, they themselves seem to lack any basis of their own. Physicists have a hard time defining matter and energy, biologists have a hard time defining life, and philosophers have a hard time defining reality.

    Information is increasingly seen by many philosophers and biologists, and even by some physicists, as the most basic building block of reality, more elementary than matter and energy.1 No wonder that there are many disputes about how to define information, and how it is related to the evolution of life or to basic ideas in physics such as entropy, the laws of thermodynamics, and the quantum uncertainty principle.2 This book will make no attempt to resolve—or even explain—these disputes, nor will it offer a universal definition of information applicable to physics, biology, and all other fields of knowledge. Since it is a work of history, which studies the past and future development of human societies, it will focus on the definition and role of information in history.

    In everyday usage, information is associated with human-made symbols like spoken or written words. Consider, for example, the story of Cher Ami and the Lost Battalion. In October 1918, when the American Expeditionary Forces was fighting to liberate northern France from the Germans, a battalion of more than five hundred American soldiers was trapped behind enemy lines. American artillery, which was trying to provide them with cover fire, misidentified their location and dropped the barrage directly on them. The battalion’s commander, Major Charles Whittlesey, urgently needed to inform headquarters of his true location, but no runner could break through the German line. According to several accounts, as a last resort Whittlesey turned to Cher Ami, an army carrier pigeon. On a tiny piece of paper, Whittlesey wrote, “We are along the road paralell [sic] 276.4. Our artillery is dropping a barrage directly on us. For heaven’s sake stop it.” The paper was inserted into a canister on Cher Ami’s right leg, and the bird was released into the air. One of the battalion’s soldiers, Private John Nell, recalled years later, “We knew without a doubt this was our last chance. If that one lonely, scared pigeon failed to find its loft, our fate was sealed.”

    Witnesses later described how Cher Ami flew into heavy German fire. A shell exploded directly below the bird, killing five men and severely injuring the pigeon. A splinter tore through Cher Ami’s chest, and his right leg was left hanging by a tendon. But he got through. The wounded pigeon flew the forty kilometers to division headquarters in about forty-five minutes, with the canister containing the crucial message attached to the remnant of his right leg. Though there is some controversy about the exact details, it is clear that the American artillery adjusted its barrage, and an American counterattack rescued the Lost Battalion. Cher Ami was tended by army medics, sent to the United States as a hero, and became the subject of numerous articles, short stories, children’s books, poems, and even movies. The pigeon had no idea what information he was conveying, but the symbols inked on the piece of paper he carried helped save hundreds of men from death and captivity.3

    Information, however, does not have to consist of human-made symbols. According to the biblical myth of the Flood, Noah learned that the water had finally receded because the pigeon he sent out from the ark returned with an olive branch in her mouth. Then God set a rainbow in the clouds as a heavenly record of his promise never to flood the earth again. Pigeons, olive branches, and rainbows have since become iconic symbols of peace and tolerance. Objects that are even more remote than rainbows can also be information. For astronomers the shape and movement of galaxies constitute crucial information about the history of the universe. For navigators the North Star indicates which way is north. For astrologers the stars are a cosmic script, conveying information about the future of individual humans and entire societies.

    Of course, defining something as “information” is a matter of perspective. An astronomer or astrologer might view the Libra constellation as “information,” but these distant stars are far more than just a notice board for human observers. There might be an alien civilization up there, totally oblivious to the information we glean from their home and to the stories we tell about it. Similarly, a piece of paper marked with ink splotches can be crucial information for an army unit, or dinner for a family of termites. Any object can be information—or not. This makes it difficult to define what information is.

    The ambivalence of information has played an important role in the annals of military espionage, when spies needed to communicate information surreptitiously. During World War I, northern France was not the only major battleground. From 1915 to 1918 the British and Ottoman Empires fought for control of the Middle East. After repulsing an Ottoman attack on the Sinai Peninsula and the Suez Canal, the British in turn invaded the Ottoman Empire, but were held at bay until October 1917 by a fortified Ottoman line stretching from Beersheba to Gaza. British attempts to break through were repulsed at the First Battle of Gaza (March 26, 1917) and the Second Battle of Gaza (April 17–19, 1917). Meanwhile, pro-British Jews living in Palestine set up a spy network code-named NILI to inform the British about Ottoman troop movements. One method they developed to communicate with their British operators involved window shutters. Sarah Aaronsohn, a NILI commander, had a house overlooking the Mediterranean. She signaled British ships by closing or opening a particular shutter, according to a predetermined code. Numerous people, including Ottoman soldiers, could obviously see the shutter, but nobody other than NILI agents and their British operators understood it was vital military information.4 So, when is a shutter just a shutter, and when is it information?

    The Ottomans eventually caught the NILI spy ring due in part to a strange mishap. In addition to shutters, NILI used carrier pigeons to convey coded messages. On September 3, 1917, one of the pigeons diverted off course and landed in—of all places—the house of an Ottoman officer. The officer found the coded message but couldn’t decipher it. Nevertheless, the pigeon itself was crucial information. Its existence indicated to the Ottomans that a spy ring was operating under their noses. As Marshall McLuhan might have put it, the pigeon was the message. NILI agents learned about the capture of the pigeon and immediately killed and buried all the remaining birds they had, because the mere possession of carrier pigeons was now incriminating information. But the massacre of the pigeons did not save NILI. Within a month the spy network was uncovered, several of its members were executed, and Sarah Aaronsohn committed suicide to avoid divulging NILI’s secrets under torture.5 When is a pigeon just a pigeon, and when is it information?

    Clearly, then, information cannot be defined as specific types of material objects. Any object—a star, a shutter, a pigeon—can be information in the right context. So exactly what context defines such objects as “information”? The naive view of information argues that objects are defined as information in the context of truth seeking. Something is information if people use it to try to discover the truth. This view links the concept of information with the concept of truth and assumes that the main role of information is to represent reality. There is a reality “out there,” and information is something that represents that reality and that we can therefore use to learn about reality. For example, the information NILI provided the British was meant to represent the reality of Ottoman troop movements. If the Ottomans massed ten thousand soldiers in Gaza—the centerpiece of their defenses—a piece of paper with symbols representing “ten thousand” and “Gaza” was important information that could help the British win the battle. If, on the other hand, there were actually twenty thousand Ottoman troops in Gaza, that piece of paper did not represent reality accurately, and could lead the British to make a disastrous military mistake.

    Put another way, the naive view argues that information is an attempt to represent reality, and when this attempt succeeds, we call it truth. While this book takes many issues with the naive view, it agrees that truth is an accurate representation of reality. But this book also holds that most information is not an attempt to represent reality and that what defines information is something entirely different. Most information in human society, and indeed in other biological and physical systems, does not represent anything.

    I want to spend a little longer on this complex and crucial argument, because it constitutes the theoretical basis of the book.

    WHAT IS TRUTH?

    Throughout this book, “truth” is understood as something that accurately represents certain aspects of reality. Underlying the notion of truth is the premise that there exists one universal reality. Anything that has ever existed or will ever exist in the universe—from the North Star, to the NILI pigeon, to web pages on astrology—is part of this single reality. This is why the search for truth is a universal project. While different people, nations, or cultures may have competing beliefs and feelings, they cannot possess contradictory truths, because they all share a universal reality. Anyone who rejects universalism rejects truth.

    Truth and reality are nevertheless different things, because no matter how truthful an account is, it can never represent reality in all its aspects. If a NILI agent wrote that there are ten thousand Ottoman soldiers in Gaza, and there were indeed ten thousand soldiers there, this accurately pointed to a certain aspect of reality, but it neglected many other aspects. The very act of counting entities—whether apples, oranges, or soldiers—necessarily focuses attention on the similarities between these entities while discounting differences.6 For example, saying only that there were ten thousand Ottoman soldiers in Gaza neglected to specify whether some were experienced veterans and others were green recruits. If there were a thousand recruits and nine thousand old hands, the military reality was quite different from if there were nine thousand rookies and a thousand battle-hardened veterans.

    There were many other differences between the soldiers. Some were healthy; others were sick. Some Ottoman troops were ethnically Turkish, while others were Arabs, Kurds, or Jews. Some were brave, others cowardly. Indeed, each soldier was a unique human being, with different parents and friends and individual fears and hopes. World War I poets like Wilfred Owen famously attempted to represent these latter aspects of military reality, which mere statistics never conveyed accurately. Does this imply that writing “ten thousand soldiers” is always a misrepresentation of reality, and that to describe the military situation around Gaza in 1917, we must specify the unique history and personality of every soldier?

    Another problem with any attempt to represent reality is that reality contains many viewpoints. For example, present-day Israelis, Palestinians, Turks, and Britons have different perspectives on the British invasion of the Ottoman Empire, the NILI underground, and the activities of Sarah Aaronsohn. That does not mean, of course, that there are several entirely separate realities, or that there are no historical facts. There is just one reality, but it is complex.

    Reality includes an objective level with objective facts that don’t depend on people’s beliefs; for example, it is an objective fact that Sarah Aaronsohn died on October 9, 1917, from self-inflicted gunshot wounds. Saying that “Sarah Aaronsohn died in an airplane crash on May 15, 1919,” is an error.

    Reality also includes a subjective level with subjective facts like the beliefs and feelings of various people, but in this case too facts can be separated from errors. For example, it is a fact that Israelis tend to regard Aaronsohn as a patriotic hero. Three weeks after her suicide, the information NILI supplied helped the British finally break the Ottoman line at the Battle of Beersheba (October 31, 1917) and the Third Battle of Gaza (November 1–2, 1917). On November 2, 1917, the British foreign secretary, Arthur Balfour, issued the Balfour Declaration, announcing that the British government “view with favor the establishment in Palestine of a national home for the Jewish people.” Israelis credit this in part to NILI and Sarah Aaronsohn, whom they admire for her sacrifice. It is another fact that Palestinians evaluate things very differently. Rather than admiring Aaronsohn, they regard her—if they’ve heard about her at all—as an imperialist agent. Even though we are dealing here with subjective views and feelings, we can still distinguish truth from falsehood. For views and feelings—just like stars and pigeons—are a part of the universal reality. Saying that “Sarah Aaronsohn is admired by everyone for her role in defeating the Ottoman Empire” is an error, not in line with reality.

    Nationality is not the only thing that affects people’s viewpoint. Israeli men and Israeli women may see Aaronsohn differently, and so do left-wingers and right-wingers, or Orthodox and secular Jews. Since suicide is forbidden by Jewish religious law, Orthodox Jews have difficulty seeing Aaronsohn’s suicide as a heroic act (she was actually denied burial in the hallowed ground of a Jewish cemetery). Ultimately, each individual has a different perspective on the world, shaped by the intersection of different personalities and life histories. Does this imply that when we wish to describe reality, we must always list all the different viewpoints it contains and that a truthful biography of Sarah Aaronsohn, for example, must specify how every single Israeli and Palestinian has felt about her?

    Taken to extremes, such a pursuit of accuracy may lead us to try to represent the world on a one-to-one scale, as in the famous Jorge Luis Borges story “On Exactitude in Science” (1946). In this story Borges tells of a fictitious ancient empire that became obsessed with producing ever more accurate maps of its territory, until eventually it produced a map with a one-to-one scale. The entire empire was covered with a map of the empire. So many resources were wasted on this ambitious representational project that the empire collapsed. Then the map too began to disintegrate, and Borges tells us that only “in the western Deserts, tattered fragments of the map are still to be found, sheltering an occasional beast or beggar.”7 A one-to-one map may look like the ultimate representation of reality, but tellingly it is no longer a representation at all; it is the reality.

    The point is that even the most truthful accounts of reality can never represent it in full. There are always some aspects of reality that are neglected or distorted in every representation. Truth, then, isn’t a one-to-one representation of reality. Rather, truth is something that brings our attention to certain aspects of reality while inevitably ignoring other aspects. No account of reality is 100 percent accurate, but some accounts are nevertheless more truthful than others.

    WHAT INFORMATION DOES

    As noted above, the naive view sees information as an attempt to represent reality. It is aware that some information doesn’t represent reality well, but it dismisses this as unfortunate cases of “misinformation” or “disinformation.” Misinformation is an honest mistake, occurring when someone tries to represent reality but gets it wrong. Disinformation is a deliberate lie, occurring when someone consciously intends to distort our view of reality.

    The naive view further believes that the solution to the problems caused by misinformation and disinformation is more information. This idea, sometimes called the counterspeech doctrine, is associated with the U.S. Supreme Court justice Louis D. Brandeis, who wrote in Whitney v. California (1927) that the remedy to false speech is more speech and that in the long term free discussion is bound to expose falsehoods and fallacies. If all information is an attempt to represent reality, then as the amount of information in the world grows, we can expect the flood of information to expose the occasional lies and errors and to ultimately provide us with a more truthful understanding of the world.

    On this crucial point, this book strongly disagrees with the naive view. There certainly are instances of information that attempt to represent reality and succeed in doing so, but this is not the defining characteristic of information. A few pages ago I referred to stars as information and casually mentioned astrologers alongside astronomers. Adherents of the naive view of information probably squirmed in their chairs when they read it. According to the naive view, astronomers derive “real information” from the stars, while the information that astrologers imagine to read in constellations is either “misinformation” or “disinformation.” If only people were given more information about the universe, surely they would abandon astrology altogether. But the fact is that for thousands of years astrology has had a huge impact on history, and today millions of people still check their star signs before making the most important decisions of their lives, like what to study and whom to marry. As of 2021, the global astrology market was valued at $12.8 billion.8

    No matter what we think about the accuracy of astrological information, we should acknowledge its important role in history. It has connected lovers, and even entire empires. Roman emperors routinely consulted astrologers before making decisions. Indeed, astrology was held in such high esteem that casting the horoscope of a reigning emperor was a capital offense. Presumably, anyone casting such a horoscope could foretell when and how the emperor would die.9 Rulers in some countries still take astrology very seriously. In 2005 the junta of Myanmar allegedly moved the country’s capital from Yangon to Naypyidaw based on astrological advice.10 A theory of information that cannot account for the historical significance of astrology is clearly inadequate.

    What the example of astrology illustrates is that errors, lies, fantasies, and fictions are information, too. Contrary to what the naive view of information says, information has no essential link to truth, and its role in history isn’t to represent a preexisting reality. Rather, what information does is to create new realities by tying together disparate things—whether couples or empires. Its defining feature is connection rather than representation, and information is whatever connects different points into a network. Information doesn’t necessarily inform us about things. Rather, it puts things in formation. Horoscopes put lovers in astrological formations, propaganda broadcasts put voters in political formations, and marching songs put soldiers in military formations.

    As a paradigmatic case, consider music. Most symphonies, melodies, and tunes don’t represent anything, which is why it makes no sense to ask whether they are true or false. Over the years people have created a lot of bad music, but not fake music. Without representing anything, music nevertheless does a remarkable job in connecting large numbers of people and synchronizing their emotions and movements. Music can make soldiers march in formation, clubbers sway together, church congregations clap in rhythm, and sports fans chant in unison.11

    The role of information in connecting things is of course not unique to human history. A case can be made that this is the chief role of information in biology too.12 Consider DNA, the molecular information that makes organic life possible. Like music, DNA doesn’t represent reality. Though generations of zebras have been fleeing lions, you cannot find in the zebra DNA a string of nucleobases representing “lion” nor another string representing “flight.” Similarly, zebra DNA contains no representation of the sun, wind, rain, or any other external phenomena that zebras encounter during their lives. Nor does DNA represent internal phenomena like body organs or emotions. There is no combination of nucleobases that represents a heart, or fear.

    Instead of trying to represent preexisting things, DNA helps to produce entirely new things. For instance, various strings of DNA nucleobases initiate cellular chemical processes that result in the production of adrenaline. Adrenaline too doesn’t represent reality in any way. Rather, adrenaline circulates through the body, initiating additional chemical processes that increase the heart rate and direct more blood to the muscles.13 DNA and adrenaline thereby help to connect cells in the heart, cells in the leg muscles, and trillions of other cells throughout the body to form a functioning network that can do remarkable things, like run away from a lion.

    If DNA represented reality, we could have asked questions like “Does zebra DNA represent reality more accurately than lion DNA?” or “Is the DNA of one zebra telling the truth about the world, while another zebra is misled by her fake DNA?” These, of course, are nonsensical questions. We might evaluate DNA by the fitness of the organism it produces, but not by truthfulness. While it is common to talk about DNA “errors,” this refers only to mutations in the process of copying DNA—not to a failure to represent reality accurately. A genetic mutation that inhibits the production of adrenaline reduces the fitness of a particular zebra, ultimately causing the network of cells to disintegrate, as when the zebra is killed by a lion and its trillions of cells lose connection with one another and decompose. But this kind of network failure means disintegration, not disinformation. That’s true of countries, political parties, and news networks as much as of zebras.

    Crucially, errors in the copying of DNA don’t always reduce fitness. Once in a blue moon, they increase fitness. Without such mutations, there would be no process of evolution. All life-forms exist thanks to genetic “errors.” The wonders of evolution are possible because DNA doesn’t represent any preexisting realities; it creates new realities.

    Let us pause to digest the implications of this. Information is something that creates new realities by connecting different points into a network. This still includes the view of information as representation. Sometimes, a truthful representation of reality can connect humans, as when 600 million people sat glued to their television sets in July 1969, watching Neil Armstrong and Buzz Aldrin walking on the moon.14 The images on the screens accurately represented what was happening 384,000 kilometers away, and seeing them gave rise to feelings of awe, pride, and human brotherliness that helped connect people.

    However, such fraternal feelings can be produced in other ways, too. The emphasis on connection leaves ample room for other types of information that do not represent reality well. Sometimes erroneous representations of reality might also serve as a social nexus, as when millions of followers of a conspiracy theory watch a YouTube video claiming that the moon landing never happened. These images convey an erroneous representation of reality, but they might nevertheless give rise to feelings of anger against the establishment or pride in one’s own wisdom that help create a cohesive new group.

    Sometimes networks can be connected without any attempt to represent reality, neither accurate nor erroneous, as when genetic information connects trillions of cells or when a stirring musical piece connects thousands of humans.

    As a final example, consider Mark Zuckerberg’s vision of the Metaverse. The Metaverse is a virtual universe made entirely of information. Unlike the one-to-one map built by Jorge Luis Borges’s imaginary empire, the Metaverse isn’t an attempt to represent our world, but rather an attempt to augment or even replace our world. It doesn’t offer us a digital replica of Buenos Aires or Salt Lake City; it invites people to build new virtual communities with novel landscapes and rules. As of 2024 the Metaverse seems like an overblown pipe dream, but within a couple of decades billions of people might migrate to live much of their lives in an augmented virtual reality, holding there most of their social and professional activities. People might come to build relationships, join movements, hold jobs, and experience emotional ups and downs in environments made of bits rather than atoms. Perhaps only in some remote deserts, tattered fragments of the old reality could still be found, sheltering an occasional beast or beggar.

    INFORMATION IN HUMAN HISTORY

    Viewing information as a social nexus helps us understand many aspects of human history that confound the naive view of information as representation. It explains the historical success not only of astrology but of much more important things, like the Bible. While some may dismiss astrology as a quaint sideshow in human history, nobody can deny the central role the Bible has played. If the main job of information had been to represent reality accurately, it would have been hard to explain why the Bible became one of the most influential texts in history.

    The Bible makes many serious errors in its description of both human affairs and natural processes. The book of Genesis claims that all human groups—including, for example, the San people of the Kalahari Desert and the Aborigines of Australia—descend from a single family that lived in the Middle East about four thousand years ago.15 According to Genesis, after the Flood all Noah’s descendants lived together in Mesopotamia, but following the destruction of the Tower of Babel they spread to the four corners of the earth and became the ancestors of all living humans. In fact, the ancestors of the San people lived in Africa for hundreds of thousands of years without ever leaving the continent, and the ancestors of the Aborigines settled Australia more than fifty thousand years ago.16 Both genetic and archaeological evidence rule out the idea that the entire ancient populations of South Africa and Australia were annihilated about four thousand years ago by a flood and that these areas were subsequently repopulated by Middle Eastern immigrants.

    An even graver distortion involves our understanding of infectious diseases. The Bible routinely depicts epidemics as divine punishment for human sins17 and claims they can be stopped or prevented by prayers and religious rituals.18 However, epidemics are of course caused by pathogens and can be stopped or prevented by following hygiene rules and using medicines and vaccines. This is today widely accepted even by religious leaders like the pope, who during the COVID-19 pandemic advised people to self-isolate, instead of congregating to pray together.19

    Yet while the Bible has done a poor job in representing the reality of human origins, migrations, and epidemics, it has nevertheless been very effective in connecting billions of people and creating the Jewish and Christian religions. Like DNA initiating chemical processes that bind billions of cells into organic networks, the Bible initiated social processes that bonded billions of people into religious networks. And just as a network of cells can do things that single cells cannot, so a religious network can do things that individual humans cannot, like building temples, maintaining legal systems, celebrating holidays, and waging holy wars.

    To conclude, information sometimes represents reality, and sometimes doesn’t. But it always connects. This is its fundamental characteristic. Therefore, when examining the role of information in history, although it sometimes makes sense to ask “How well does it represent reality? Is it true or false?” often the more crucial questions are “How well does it connect people? What new network does it create?”

    It should be emphasized that rejecting the naive view of information as representation does not force us to reject the notion of truth, nor does it force us to embrace the populist view of information as a weapon. While information always connects, some types of information—from scientific books to political speeches—may strive to connect people by accurately representing certain aspects of reality. But this requires a special effort, which most information does not make. This is why the naive view is wrong to believe that creating more powerful information technology will necessarily result in a more truthful understanding of the world. If no additional steps are taken to tilt the balance in favor of truth, an increase in the amount and speed of information is likely to swamp the relatively rare and expensive truthful accounts by much more common and cheap types of information.

    When we look at the history of information from the Stone Age to the Silicon Age, we therefore see a constant rise in connectivity, without a concomitant rise in truthfulness or wisdom. Contrary to what the naive view believes, Homo sapiens didn’t conquer the world because we are talented at turning information into an accurate map of reality. Rather, the secret of our success is that we are talented at using information to connect lots of individuals. Unfortunately, this ability often goes hand in hand with believing in lies, errors, and fantasies. This is why even technologically advanced societies like Nazi Germany and the Soviet Union have been prone to hold delusional ideas, without their delusions necessarily weakening them. Indeed, the mass delusions of Nazi and Stalinist ideologies about things like race and class actually helped them make tens of millions of people march together in lockstep.

    In chapters 2–5 we’ll take a closer look at the history of information networks. We’ll discuss how, over tens of thousands of years, humans invented various information technologies that greatly improved connectivity and cooperation without necessarily resulting in a more truthful representation of the world. These information technologies—invented centuries and millennia ago—still shape our world even in the era of the internet and AI. The first information technology we’ll examine, which is also the first information technology developed by humans, is the story.

    CHAPTER 2 Stories: Unlimited Connections

    We Sapiens rule the world not because we are so wise but because we are the only animals that can cooperate flexibly in large numbers. I have explored this idea in my previous books Sapiens and Homo Deus, but a brief recap is inescapable.

    The Sapiens’ ability to cooperate flexibly in large numbers has precursors among other animals. Some social mammals like chimpanzees display significant flexibility in the way they cooperate, while some social insects like ants cooperate in very large numbers. But neither chimps nor ants establish empires, religions, or trade networks. Sapiens are capable of doing such things because we are far more flexible than chimps and can simultaneously cooperate in even larger numbers than ants. In fact, there is no upper limit to the number of Sapiens who can cooperate with one another. The Catholic Church has about 1.4 billion members. China has a population of about 1.4 billion. The global trade network connects about 8 billion Sapiens.

    This is surprising given that humans cannot form long-term intimate bonds with more than a few hundred individuals.1 It takes many years and common experiences to get to know someone’s unique character and history and to cultivate ties of mutual trust and affection. Consequently, if Sapiens networks were connected only by personal human-to-human bonds, our networks would have remained very small. This is the situation among our chimpanzee cousins, for example. Their typical community numbers 20–60 members, and on rare occasions the number might increase to 150–200.2 This appears to have been the situation also among ancient human species like Neanderthals and archaic Homo sapiens. Each of their bands numbered a few dozen individuals, and different bands rarely cooperated.3

    About seventy thousand years ago, Homo sapiens bands began displaying an unprecedented capacity to cooperate with one another, as evidenced by the emergence of inter-band trade and artistic traditions and by the rapid spread of our species from our African homeland to the entire globe. What enabled different bands to cooperate is that evolutionary changes in brain structure and linguistic abilities apparently gave Sapiens the aptitude to tell and believe fictional stories and to be deeply moved by them. Instead of building a network from human-to-human chains alone—as the Neanderthals, for example, did—stories provided Homo sapiens with a new type of chain: human-to-story chains. In order to cooperate, Sapiens no longer had to know each other personally; they just had to know the same story. And the same story can be familiar to billions of individuals. A story can thereby serve like a central connector, with an unlimited number of outlets into which an unlimited number of people can plug. For example, the 1.4 billion members of the Catholic Church are connected by the Bible and other key Christian stories; the 1.4 billion citizens of China are connected by the stories of communist ideology and Chinese nationalism; and the 8 billion members of the global trade network are connected by stories about currencies, corporations, and brands.

    Even charismatic leaders who have millions of followers are an example of this rule rather than an exception. It may seem that in the case of ancient Chinese emperors, medieval Catholic popes, or modern corporate titans it has been a single flesh-and-blood human—rather than a story—that has served as a nexus linking millions of followers. But, of course, in all these cases almost none of the followers has had a personal bond with the leader. Instead, what they have connected to has been a carefully crafted story about the leader, and it is in this story that they have put their faith.

    Joseph Stalin, who stood at the nexus of one of the biggest personality cults in history, understood this well. When his troublesome son Vasily exploited his famous name to frighten and awe people, Stalin berated him. “But I’m a Stalin too,” protested Vasily. “No, you’re not,” replied Stalin. “You’re not Stalin and I’m not Stalin. Stalin is Soviet power. Stalin is what he is in the newspapers and the portraits, not you, no—not even me!”4

    Present-day influencers and celebrities would concur. Some have hundreds of millions of online followers, with whom they communicate daily through social media. But there is very little authentic personal connection there. The social media accounts are usually run by a team of experts, and every image and word is professionally crafted and curated to manufacture what is nowadays called a brand.5

    A “brand” is a specific type of story. To brand a product means to tell a story about that product, which may have little to do with the product’s actual qualities but which consumers nevertheless learn to associate with the product. For example, over the decades the Coca-Cola corporation has invested tens of billions of dollars in advertisements that tell and retell the story of the Coca-Cola drink.6 People have seen and heard the story so often that many have come to associate a certain concoction of flavored water with fun, happiness, and youth (as opposed to tooth decay, obesity, and plastic waste). That’s branding.7

    As Stalin knew, it is possible to brand not only products but also individuals. A corrupt billionaire can be branded as the champion of the poor; a bungling imbecile can be branded as an infallible genius; and a guru who sexually abuses his followers can be branded as a chaste saint. People think they connect to the person, but in fact they connect to the story told about the person, and there is often a huge gulf between the two.

    Even the story of Cher Ami, the heroic pigeon, was partly the product of a branding campaign aimed at enhancing the public image of the U.S. Army’s Pigeon Service. A 2021 revisionist study by the historian Frank Blazich found that though there is no doubt Cher Ami sustained severe injuries while transporting a message somewhere in Northern France, several key features of the story are doubtful or inaccurate. First, relying on contemporary military records, Blazich demonstrated that headquarters learned about the exact location of the Lost Battalion about twenty minutes prior to the pigeon’s arrival. It was not the pigeon that put a stop to the barrage of friendly fire decimating the Lost Battalion. Even more crucially, there is simply no proof that the pigeon carrying Major Whittlesey’s message was Cher Ami. It might well have been another bird, while Cher Ami might have sustained his wounds a couple of weeks later, during an altogether different battle.

    According to Blazich, the doubts and inconsistencies in Cher Ami’s story were overshadowed by its propaganda value to the army and its appeal to the public. Over the years the story was retold so many times that facts became hopelessly enmeshed with fiction. Journalists, poets, and filmmakers added fanciful details to it, for example that the pigeon lost an eye as well as a leg and that it was awarded the Distinguished Service Cross. In the 1920s and 1930s Cher Ami became the most famous bird in the world. When he died, his carefully preserved corpse was placed on display at the Smithsonian Museum, where it became a pilgrimage site for American patriots and World War I veterans. As the story grew in the telling, it took over even the recollections of survivors of the Lost Battalion, who came to accept the popular narrative at face value. Blazich recounts the case of Sherman Eager, an officer in the Lost Battalion, who decades after the war brought his children to see Cher Ami at the Smithsonian and told them, “You all owe your lives to that pigeon.” Whatever the facts may be, the story of the self-sacrificing winged saviour proved irresistible.8

    As a much more extreme example, consider Jesus. Two millennia of storytelling have encased Jesus within such a thick cocoon of stories that it is impossible to recover the historical person. Indeed, for millions of devout Christians merely raising the possibility that the real person was different from the story is blasphemy. As far as we can tell, the real Jesus was a typical Jewish preacher who built a small following by giving sermons and healing the sick. After his death, however, Jesus became the subject of one of the most remarkable branding campaigns in history. This little-known provincial guru, who during his short career gathered just a handful of disciples and who was executed as a common criminal, was rebranded after death as the incarnation of the cosmic god who created the universe.9 Though no contemporary portrait of Jesus has survived, and though the Bible never describes what he looked like, imaginary renderings of him have become some of the most recognizable icons in the world.

    It should be stressed that the creation of the Jesus story was not a deliberate lie. People like Saint Paul, Tertullian, Saint Augustine, and Martin Luther didn’t set out to deceive anyone. They projected their deeply felt hopes and feelings on the figure of Jesus, in the same way that all of us routinely project our feelings on our parents, lovers, and leaders. While branding campaigns are occasionally a cynical exercise of disinformation, most of the really big stories of history have been the result of emotional projections and wishful thinking. True believers play a key role in the rise of every major religion and ideology, and the Jesus story changed history because it gained an immense number of true believers.

    By gaining all those believers, the story of Jesus managed to have a much bigger impact on history than the person of Jesus. The person of Jesus walked from village to village on his two feet, talking with people, eating and drinking with them, placing his hands on their sick bodies. He made a difference to the lives of perhaps several thousand individuals, all living in one minor Roman province. In contrast, the story of Jesus flew around the whole world, first on the wings of gossip, anecdote, and rumor; then via parchment texts, paintings, and statues; and eventually as blockbuster movies and internet memes. Billions of people not only heard the Jesus story but came to believe in it too, which created one of the biggest and most influential networks in the world.

    Stories like the one about Jesus can be seen as a way of stretching preexisting biological bonds. Family is the strongest bond known to humans. One way that stories build trust between strangers is by making these strangers reimagine each other as family. The Jesus story presented Jesus as the heavenly father of all humans, encouraged hundreds of millions of Christians to see each other as brothers and sisters, and created a shared pool of family memories. While most Christians were not physically present at the Last Supper, they have heard the story so many times, and they have seen so many images of the event, that they “remember” it more vividly than they remember most of the family dinners in which they actually participated.

    Interestingly, Jesus’s last supper was the Jewish Passover meal, which according to the Gospel accounts Jesus shared with his disciples just before his crucifixion. In Jewish tradition, the whole purpose of the Passover meal is to create and reenact artificial memories. Every year Jewish families sit together on the eve of Passover to eat and reminisce about “their” exodus from Egypt. They are supposed not only to tell the story of how the descendants of Jacob escaped slavery in Egypt but to remember how they personally suffered at the hands of the Egyptians, how they personally saw the sea part, and how they personally received the Ten Commandments from Jehovah at Mount Sinai.

    The Jewish tradition doesn’t mince words here. The text of the Passover ritual (the Haggadah) insists that “in every generation a person is obligated to regard himself as if he personally had come out of Egypt.” If anyone objects that this is a fiction, and that they didn’t personally come out of Egypt, Jewish sages have a ready answer. They claim that the souls of all Jews throughout history were created by Jehovah long before they were born and all these souls were present at Mount Sinai.10 As Salvador Litvak, a Jewish social media influencer, explained to his online followers in 2018, “You and I were there together.… When we fulfill the obligation to see ourselves as if we personally left Egypt, it’s not a metaphor. We don’t imagine the Exodus, we remember it.”11

    So every year, in the most important celebration of the Jewish calendar, millions of Jews put on a show that they remember things that they didn’t witness and that in all probability never happened at all. As numerous modern studies show, repeatedly retelling a fake memory eventually causes the person to adopt it as a genuine recollection.12 When two Jews encounter each other for the first time, they can immediately feel that they both belong to the same family, that they were together slaves in Egypt, and that they were together at Mount Sinai. That’s a powerful bond that sustained the Jewish network over many centuries and continents.

    INTERSUBJECTIVE ENTITIES

    The Jewish Passover story builds a large network by taking existing biological kin bonds and stretching them way beyond their biological limits. But there is an even more revolutionary way for stories to build networks. Like DNA, stories can create entirely new entities. Indeed, stories can even create an entirely new level of reality. As far as we know, prior to the emergence of stories the universe contained just two levels of reality. Stories added a third.

    The two levels of reality that preceded storytelling are objective reality and subjective reality. Objective reality consists of things like stones, mountains, and asteroids—things that exist whether we are aware of them or not. An asteroid hurtling toward planet Earth, for example, exists even if nobody knows it’s out there. Then there is subjective reality: things like pain, pleasure, and love that aren’t “out there” but rather “in here.” Subjective things exist in our awareness of them. An unfelt ache is an oxymoron.

    But some stories are able to create a third level of reality: intersubjective reality. Whereas subjective things like pain exist in a single mind, intersubjective things like laws, gods, nations, corporations, and currencies exist in the nexus between large numbers of minds. More specifically, they exist in the stories people tell one another. The information humans exchange about intersubjective things doesn’t represent anything that had already existed prior to the exchange of information; rather, the exchange of information creates these things.

    When I tell you that I am in pain, telling you about it doesn’t create the pain. And if I stop talking about the pain, it doesn’t make the pain go away. Similarly, when I tell you that I saw an asteroid, this doesn’t create the asteroid. The asteroid exists whether people talk about it or not. But when lots of people tell one another stories about laws, gods, or currencies, this is what creates these laws, gods, or currencies. If people stop talking about them, they disappear. Intersubjective things exist in the exchange of information.

    Let’s take a closer look. The calorific value of pizza doesn’t depend on our beliefs. A typical pizza contains between fifteen hundred and twenty-five hundred calories.13 In contrast, the financial value of money—and pizzas—depends entirely on our beliefs. How many pizzas can you purchase for a dollar, or for a bitcoin? In 2010, Laszlo Hanyecz bought two pizzas for 10,000 bitcoins. It was the first known commercial transaction involving bitcoin—and with hindsight, also the most expensive pizza ever. By November 2021, a single bitcoin was valued at more than $69,000, so the bitcoins Hanyecz paid for his two pizzas were worth $690 million, enough to purchase millions of pizzas.14 While the calorific value of pizza is an objective reality that remained the same between 2010 and 2021, the financial value of bitcoin is an intersubjective reality that changed dramatically during the same period, depending on the stories people told and believed about bitcoin.

    Another example. Suppose I ask, “Does the Loch Ness Monster exist?” This is a question about the objective level of reality. Some people believe that dinosaur-like animals really do inhabit Loch Ness. Others dismiss the idea as a fantasy or a hoax. Over the years, many attempts have been made to resolve the disagreement once and for all, using scientific methods such as sonar scans and DNA surveys. If huge animals live in the lake, they should appear on sonar, and they should leave DNA traces. Based on the available evidence, the scientific consensus is that the Loch Ness Monster does not exist. (A DNA survey conducted in 2019 found genetic material from three thousand species, but no monster. At most, Loch Ness may contain some five-kilo eels.15) Many people may nevertheless continue to believe that the Loch Ness Monster exists, but believing it doesn’t change objective reality.

    In contrast to animals, whose existence can be verified or disproved through objective tests, states are intersubjective entities. We normally don’t notice it, because everybody takes the existence of the United States, China, Russia, or Brazil for granted. But there are cases when people disagree about the existence of certain states, and then their intersubjective status emerges. The Israeli-Palestinian conflict, for example, revolves around this matter, because some people and governments refuse to acknowledge the existence of Israel and others refuse to acknowledge the existence of Palestine. As of 2024, the governments of Brazil and China, for example, say that both Israel and Palestine exist; the governments of the United States and Cameroon recognize only Israel’s existence; whereas the governments of Algeria and Iran recognize only Palestine. Other cases range from Kosovo, which as of 2024 is recognized as a state by around half of the 193 UN members,16 to Abkhazia, which almost all governments see as a sovereign territory of Georgia, but which is recognized as a state by Russia, Venezuela, Nicaragua, Nauru, and Syria.17

    Indeed, almost all states pass at least temporarily through a phase during which their existence is contested, when struggling for independence. Did the United States come into existence on July 4, 1776, or only when other states like France and finally the U.K. recognized it? Between the declaration of U.S. independence on July 4, 1776, and the signing of the Treaty of Paris on September 3, 1783, some people like George Washington believed the United States existed, while other people like King George III vehemently rejected this idea.

    Disagreements about the existence of states cannot be resolved by an objective test, such as a DNA survey or a sonar scan. Unlike animals, states are not an objective reality. When we ask whether a particular state exists, we are raising a question about intersubjective reality. If enough people agree that a particular state exists, then it does. It can then do things like sign legally binding treaties with other governments as well as NGOs and private corporations.

    Of all genres of stories, those that create intersubjective realities have been the most crucial for the development of large-scale human networks. Implanting fake family memories is certainly helpful, but no religions or empires managed to survive for long without a strong belief in the existence of a god, a nation, a law code, or a currency. For the formation of the Christian Church, for example, it was important that people recollect what Jesus said at the Last Supper, but the crucial step was making people believe that Jesus was a god rather than just an inspiring rabbi. For the formation of the Jewish religion, it was helpful that Jews “remembered” how they together escaped slavery in Egypt, but the really decisive step was making all Jews adhere to the same religious law code, the Halakha.

    Intersubjective things like laws, gods, and currencies are extremely powerful within a particular information network and utterly meaningless outside it. Suppose a billionaire crashes his private jet on a deserted island and finds himself alone with a suitcase full of banknotes and bonds. When he was in São Paulo or Mumbai, he could use these papers to make people feed him, clothe him, protect him, and build him a private jet. But once he is cut off from other members of our information network, his banknotes and bonds immediately become worthless. He cannot use them to get the island’s monkeys to provide him with food or to build him a raft.

    THE POWER OF STORIES

    Whether through implanting fake memories, forming fictional relationships, or creating intersubjective realities, stories produced large-scale human networks. These networks in turn completely changed the balance of power in the world. Story-based networks made Homo sapiens the most powerful of all animals, giving it a crucial edge not only over lions and mammoths but also over other ancient human species like Neanderthals.

    Neanderthals lived in small isolated bands, and to the best of our knowledge different bands cooperated with one another only rarely and weakly, if at all.18 Stone Age Sapiens too lived in small bands of a few dozen individuals. But following the emergence of storytelling, Sapiens bands no longer lived in isolation. Bands were connected by stories about things like revered ancestors, totem animals, and guardian spirits. Bands that shared stories and intersubjective realities constituted a tribe. Each tribe was a network connecting hundreds or even thousands of individuals.19

    Belonging to a large tribe had an obvious advantage in times of conflict. Five hundred Sapiens could easily defeat fifty Neanderthals.20 But tribal networks had many additional advantages. If we live in an isolated band of fifty people and a severe drought hits our home territory, many of us might starve to death. If we try to migrate elsewhere, we are likely to encounter hostile groups, and we might also find it difficult to forage for food, water, and flint (to make tools) in unfamiliar territory. However, if our band is part of a tribal network, in times of need at least some of us could go live with our distant friends. If our shared tribal identity is strong enough, they would welcome us and teach us about the local dangers and opportunities. A decade or two later, we might reciprocate. The tribal network, then, acted like an insurance policy. It minimized risk by spreading it across a lot more people.21

    Even in quiet times Sapiens could benefit enormously from exchanging information not just with a few dozen members of a small band but with an entire tribal network. If one of the tribe’s bands discovered a better way to make spear points, learned how to heal wounds with some rare medicinal herb, or invented a needle to sew clothes, that knowledge could be quickly passed to the other bands. Even though individually Sapiens might not have been more intelligent than Neanderthals, five hundred Sapiens together were far more intelligent than fifty Neanderthals.22

    All this was made possible by stories. The power of stories is often missed or denied by materialist interpretations of history. In particular, Marxists tend to view stories as merely a smoke screen for underlying power relations and material interests. According to Marxist theories, people are always motivated by objective material interests and use stories only to camouflage these interests and confound their rivals. For example, in this reading the Crusades, World War I, and the Iraq War were all fought for the economic interests of powerful elites rather than for religious, nationalist, or liberal ideals. Understanding these wars means setting aside all the mythological fig leaves—about God, patriotism, or democracy—and observing power relations in their nakedness.

    This Marxist view, however, is not only cynical but wrong. While materialist interests certainly played a role in the Crusades, World War I, the Iraq War, and most other human conflicts, that does not mean that religious, national, and liberal ideals played no role at all. Moreover, materialist interests by themselves cannot explain the identities of the rival camps. Why is it that in the twelfth century landowners and merchants from France, Germany, and Italy united to conquer territories and trade routes in the Levant—instead of landowners and merchants from France and North Africa uniting to conquer Italy? And why is it that in 2003, the United States and Britain sought to conquer the oil fields of Iraq, rather than the gas fields of Norway? Can this really be explained by purely materialist considerations, without any recourse to people’s religious and ideological beliefs?

    In fact, all relations between large-scale human groups are shaped by stories, because the identities of these groups are themselves defined by stories. There are no objective definitions for who is British, American, Norwegian, or Iraqi; all these identities are shaped by national and religious myths that are constantly challenged and revised. Marxists may claim that large-scale groups have objective identities and interests, independent of stories. If that is so, how can we explain that only humans have large-scale groups like tribes, nations, and religions, whereas chimpanzees lack them? After all, chimpanzees share with humans all our objective material interests; they too need to drink, eat, and protect themselves from diseases. They too want sex and social power. But chimpanzees cannot maintain large-scale groups, because they are unable to create the stories that connect such groups and define their identities and interests. Contrary to Marxist thinking, large-scale identities and interests in history are always intersubjective; they are never objective.

    This is good news. If history had been shaped solely by material interests and power struggles, there would be no point talking to people who disagree with us. Any conflict would ultimately be the result of objective power relations, which cannot be changed merely by talking. In particular, if privileged people can see and believe only those things that enshrine their privileges, how can anything except violence persuade them to renounce those privileges and alter their beliefs? Luckily, since history is shaped by intersubjective stories, sometimes we can avert conflict and make peace by talking with people, changing the stories in which they and we believe, or coming up with a new story that everyone can accept.

    Take, for example, the rise of Nazism. There certainly were material interests that drove millions of Germans to support Hitler. The Nazis would probably never have come to power if it wasn’t for the economic crisis of the early 1930s. However, it is wrong to think that the Third Reich was the inevitable outcome of underlying power relations and material interests. Hitler won the 1933 elections because during the economic crisis millions of Germans came to believe the Nazi story rather than one of the alternative stories on offer. This wasn’t the inevitable result of Germans pursuing their material interests and protecting their privileges; it was a tragic mistake. We can confidently say that it was a mistake, and that Germans could have chosen better stories, because we know what happened next. Twelve years of Nazi rule didn’t foster the Germans’ material interests. Nazism led to the destruction of Germany and the deaths of millions. Later, when Germans adopted liberal democracy, this did lead to a lasting improvement in their lives. Couldn’t the Germans have skipped the failed Nazi experiment and put their faith in liberal democracy already in the early 1930s? The position of this book is that they could have. History is often shaped not by deterministic power relations, but rather by tragic mistakes that result from believing in mesmerizing but harmful stories.

    THE NOBLE LIE

    The centrality of stories reveals something fundamental about the power of our species, and it explains why power doesn’t always go hand in hand with wisdom. The naive view of information says that information leads to truth, and knowing the truth helps people to gain both power and wisdom. This sounds reassuring. It implies that people who ignore the truth are unlikely to have much power, whereas people who respect the truth can gain much power, but that power would be tempered by wisdom. For example, people who ignore the truth about human biology might believe racist myths but will not be able to produce powerful medicines and bioweapons, whereas people who understand biology will have that kind of power but will not use it in the service of racist ideologies. If this had indeed been the case, we could sleep calmly, trusting our presidents, high priests, and CEOs to be wise and honest. A politician, a movement, or a country might conceivably get ahead here and there with the help of lies and deceptions, but in the long term that would be a self-defeating strategy.

    Unfortunately, this is not the world in which we live. In history, power stems only partially from knowing the truth. It also stems from the ability to maintain social order among a large number of people. Suppose you want to make an atom bomb. To succeed, you obviously need some accurate knowledge of physics. But you also need lots of people to mine uranium ore, build nuclear reactors, and provide food for the construction workers, miners, and physicists. The Manhattan Project directly employed about 130,000 people, with millions more working to sustain them.23 Robert Oppenheimer could devote himself to his equations because he relied on thousands of miners to extract uranium at the Eldorado mine in northern Canada and the Shinkolobwe mine in the Belgian Congo24—not to mention the farmers who grew potatoes for his lunch. If you want to make an atom bomb, you must find a way to make millions of people cooperate.

    It is the same with all ambitious projects that humans undertake. A Stone Age band going to hunt a mammoth obviously needed to know some true facts about mammoths. If they believed they could kill a mammoth by casting spells, their hunting expedition would have failed. But just knowing facts about mammoths wasn’t enough, either. The hunters also needed to make sure all of them agreed on the same plan and bravely did their bit even in the face of mortal danger. If they believed that by pronouncing a spell they could guarantee a good afterlife for dead hunters, their hunting expeditions had a much higher chance of success. Even if objectively the spell was powerless and did not benefit dead hunters in any way, by fortifying the courage and solidarity of living hunters, it nevertheless made a crucial contribution to the hunt’s success.25

    While power depends on both truth and order, in most cases it is the people who know how to maintain order who call the shots, giving instructions to the people who merely know the truth about things like mammoths or nuclear physics. Robert Oppenheimer obeyed Franklin Delano Roosevelt rather than the other way around. Similarly, Werner Heisenberg obeyed Adolf Hitler, Igor Kurchatov deferred to Joseph Stalin, and in contemporary Iran experts in nuclear physics follow the orders of experts in Shiite theology.

    What the people at the top know, which nuclear physicists don’t always realize, is that telling the truth about the universe is hardly the most efficient way to produce order among large numbers of humans. It is true that E = mc², and it explains a lot of what happens in the universe, but knowing that E = mc² usually doesn’t resolve political disagreements or inspire people to make sacrifices for a common cause. Instead, what holds human networks together tends to be fictional stories, especially stories about intersubjective things like gods, money, and nations. When it comes to uniting people, fiction enjoys two inherent advantages over the truth. First, fiction can be made as simple as we like, whereas the truth tends to be complicated, because the reality it is supposed to represent is complicated. Take, for example, the truth about nations. It is difficult to grasp that the nation to which one belongs is an intersubjective entity that exists only in our collective imagination. You rarely hear politicians say such things in their political speeches. It is far easier to believe that our nation is God’s chosen people, entrusted by the Creator with some special mission. This simple story has been repeatedly told by countless politicians from Israel to Iran and from the United States to Russia.

    Second, the truth is often painful and disturbing, and if we try to make it more comforting and flattering, it will no longer be the truth. In contrast, fiction is highly malleable. The history of every nation contains some dark episodes that citizens don’t like to acknowledge and remember. An Israeli politician who in her election speeches details the miseries inflicted on Palestinian civilians by the Israeli occupation is unlikely to get many votes. In contrast, a politician who builds a national myth by ignoring uncomfortable facts, focusing on glorious moments in the Jewish past, and embellishing reality wherever necessary may well sweep to power. That’s the case not just in Israel but in all countries. How many Italians or Indians want to hear the unblemished truth about their nations? An uncompromising adherence to the truth is essential for scientific progress, and it is also an admirable spiritual practice, but it is not a winning political strategy.

    Already in his Republic, Plato imagined that the constitution of his utopian state would be based on “the noble lie”—a fictional story about the origin of the social order, one that secures the citizens’ loyalty and prevents them from questioning the constitution. Citizens should be told, Plato wrote, that they were all born out of the earth, that the land is their mother, and that they therefore owe filial loyalty to the motherland. They should further be told that when they were conceived, the gods intermingled different metals—gold, silver, bronze, and iron—into them, which justifies a natural hierarchy between golden rulers and bronze servants. While Plato’s utopia was never realized in practice, numerous polities through the ages told their inhabitants variations of this noble lie.

    Plato’s noble lie notwithstanding, we should not conclude that all politicians are liars or that all national histories are deceptions. The choice isn’t simply between telling the truth and lying. There is a third option. Telling a fictional story is lying only when you pretend that the story is a true representation of reality. Telling a fictional story isn’t lying when you avoid such pretense and acknowledge that you are trying to create a new intersubjective reality rather than represent a preexisting objective reality.

    For example, on September 17, 1787, the Constitutional Convention signed the U.S. Constitution, which came into force in 1789. The Constitution didn’t reveal any preexisting truth about the world, but crucially it wasn’t a lie, either. Rejecting Plato’s recommendation, the authors of the text didn’t deceive anyone about the text’s origins. They didn’t pretend that the text came down from heaven or that it had been inspired by some god. Rather, they acknowledged that it was an extremely creative legal fiction generated by fallible human beings.

    “We the People of the United States,” says the Constitution about its own origins, “in Order to form a more perfect Union … do ordain and establish this Constitution.” Despite the acknowledgment that it is a human-made legal fiction, the U.S. Constitution indeed managed to form a powerful union. It maintained for more than two centuries a surprising degree of order among many millions of people who belonged to a wide range of religious, ethnic, and cultural groups. The U.S. Constitution has thus functioned like a tune that without claiming to represent anything has nevertheless made numerous people act together in order.

    It is crucial to note that “order” should not be confused with fairness or justice. The order created and maintained by the U.S. Constitution condoned slavery, the subordination of women, the expropriation of indigenous people, and extreme economic inequality. The genius of the U.S. Constitution is that by acknowledging that it is a legal fiction created by human beings, it was able to provide mechanisms to reach agreement on amending itself and remedying its own injustices (as chapter 5 explores in greater depth). The Constitution’s Article V details how people can propose and ratify such amendments, which “shall be valid to all Intents and Purposes, as Part of this Constitution.” Less than a century after the Constitution was written, the Thirteenth Amendment abolished slavery.

    In this, the U.S. Constitution was fundamentally different from stories that denied their fictive nature and claimed divine origin, such as the Ten Commandments. Like the U.S. Constitution, the Ten Commandments endorsed slavery. The Tenth Commandment says, “You shall not covet your neighbor’s house. You shall not covet your neighbor’s wife, or his male slave or female slave” (Exodus 20:17). This implies that God is perfectly okay with people holding slaves, and objects only to the coveting of slaves belonging to someone else. But unlike the U.S. Constitution, the Ten Commandments failed to provide any amendment mechanism. There is no Eleventh Commandment that says, “You can amend commandments by a two-thirds majority vote.”

    This crucial difference between the two texts is clear from their opening gambits. The U.S. Constitution opens with “We the People.” By acknowledging its human origin, it invests humans with the power to amend it. The Ten Commandments open with “I am the Lord your God.” By claiming divine origin, it precludes humans from changing it. As a result, the biblical text still endorses slavery even today.

    All human political systems are based on fictions, but some admit it, and some do not. Being truthful about the origins of our social order makes it easier to make changes in it. If humans like us invented it, we can amend it. But such truthfulness comes at a price. Acknowledging the human origins of the social order makes it harder to persuade everyone to agree on it. If humans like us invented it, why should we accept it? As we shall see in chapter 5, until the late eighteenth century the lack of mass communication technology made it extremely difficult to conduct open debates between millions of people about the rules of the social order. To maintain order, Russian tsars, Muslim caliphs, and Chinese sons of heaven therefore claimed that the fundamental rules of society came down from heaven and were not open to human amendment. In the early twenty-first century, many political systems still claim superhuman authority and oppose open debates that may result in unwelcome changes.

    THE PERENNIAL DILEMMA

    After we understand the key role of fiction in history, it is finally possible to present a more complete model of information networks, which goes beyond both the naive view of information and the populist critique of that view. Contrary to the naive view, information isn’t the raw material of truth, and human information networks aren’t geared only to discover the truth. But contrary to the populist view, information isn’t just a weapon, either. Rather, to survive and flourish, every human information network needs to do two things simultaneously: discover truth and create order. Accordingly, as history unfolded, human information networks have been developing two distinct sets of skills. On the one hand, as the naive view expects, the networks have learned how to process information to gain a more accurate understanding of things like medicine, mammoths, and nuclear physics. At the same time, the networks have also learned how to use information to maintain stronger social order among larger populations, by using not just truthful accounts but also fictions, fantasies, propaganda, and—occasionally—downright lies.

    The naive view of information
    A more complete historical view of information

    Having a lot of information doesn’t in and of itself guarantee either truth or order. It is a difficult process to use information to discover the truth and simultaneously use it to maintain order. What makes things worse is that these two processes are often contradictory, because it is frequently easier to maintain order through fictions. Sometimes—as in the case of the U.S. Constitution—fictional stories may acknowledge their fictionality, but more often they disavow it. Religions, for example, always claim to be an objective and eternal truth rather than a fictional story invented by humans. In such cases, the search for truth threatens the foundations of the social order. Many societies require their populations not to know their true origins: ignorance is strength. What happens, then, when people get uncomfortably close to the truth? What happens when the same bit of information reveals an important fact about the world, and also undermines the noble lie that holds society together? In such cases society may seek to preserve order by placing limits on the search for truth.

    One obvious example is Darwin’s theory of evolution. Understanding evolution greatly advances our understanding of the origins and biology of species, including Homo sapiens, but it also undermines the central myths that maintain order in numerous societies. No wonder that various governments and churches have banned or limited the teaching of evolution, preferring to sacrifice truth for the sake of order.26

    A related problem is that an information network may allow and even encourage people to search for truth, but only in specific fields that help generate power without threatening the social order. The result can be a very powerful network that is singularly lacking in wisdom. Nazi Germany, for example, cultivated many of the world’s leading experts in chemistry, optics, engineering, and rocket science. It was largely Nazi rocket science that later brought the Americans to the moon.27 This scientific prowess helped the Nazis build an extremely powerful war machine, which was then deployed in the service of a deranged and murderous mythology. Under Nazi rule Germans were encouraged to develop rocket science, but they were not free to question racist theories about biology and history.

    That’s a major reason why the history of human information networks isn’t a triumphant march of progress. While over the generations human networks have grown increasingly powerful, they have not necessarily grown increasingly wise. If a network privileges order over truth, it can become very powerful but use that power unwisely.

    Instead of a march of progress, the history of human information networks is a tightrope walk trying to balance truth with order. In the twenty-first century we aren’t much better at finding the right balance than our ancestors were in the Stone Age. Contrary to what the mission statements of corporations like Google and Facebook imply, simply increasing the speed and efficiency of our information technology doesn’t necessarily make the world a better place. It only makes the need to balance truth and order more urgent. The invention of the story taught us this lesson already tens of thousands of years ago. And the same lesson would be taught again, when humans came up with their second great information technology: the written document.

    CHAPTER 3 Documents: The Bite of the Paper Tigers

    Stories were the first crucial information technology developed by humans. They laid the foundation for all large-scale human cooperation and made humans the most powerful animals on earth. But as an information technology, stories have their limitations.

    To appreciate this, consider the role storytelling plays in the formation of nations. Many nations have first been conceived in the imagination of poets. Sarah Aaronsohn and the NILI underground are remembered by present-day Israelis as some of the first Zionists who risked their lives in the 1910s to establish a Jewish state in Palestine, but from where did NILI members get this idea in the first place? They were inspired by an earlier generation of poets, thinkers, and visionaries such as Theodor Herzl and Hayim Nahman Bialik.

    In the 1890s and first decade of the twentieth century, Bialik, a Ukrainian Jew, published numerous poems and stories bewailing the persecution and weakness of European Jews and calling on them to take their fate in their hands—to defend themselves by force of arms, immigrate to Palestine, and there establish their own state. One of his most stirring poems was written following the Kishinev Pogrom of 1903, in which forty-nine Jews were murdered and dozens more were injured.1 “In the City of Slaughter” condemned the murderous antisemitic mob who perpetrated the atrocities, but it also criticized the Jews themselves for their pacifism and helplessness.

    In one heart-wrenching scene, Bialik described how Jewish women were gang-raped, while their husbands and brothers hid nearby, afraid to intervene. The poem compares the Jewish men to terrified mice and imagines how they quietly prayed to God to perform some miracle, which failed to materialize. The poem then tells how even after the pogrom was over, the survivors had no thought of arming themselves and instead entered Talmudic disputations about whether the raped women were now ritualistically “defiled” or whether they were still “pure.” This poem is mandatory reading in many Israeli schools today. It is also mandatory reading for anyone wishing to understand how after two millennia of being one of the most pacifist groups in history, Jews built one of the most formidable armies in the world. Not for nothing was Bialik named Israel’s national poet.2

    The fact that Bialik lived in Ukraine, and was intimately familiar with the persecution of Ashkenazi Jews in eastern Europe but had little understanding of conditions in Palestine, contributed to the subsequent conflict there between Jews and Arabs. Bialik’s poems inspired Jews to see themselves as victims in dire need of developing their military might and building their own country, but hardly considered the catastrophic consequences for the Arab inhabitants of Palestine, or indeed for the Mizrahi Jewish communities native to the Middle East. When the Arab-Israeli conflict exploded in the late 1940s, hundreds of thousands of Palestinians and hundreds of thousands of Mizrahi Jews were driven out of their ancestral homes in the Middle East, partly as a result of poems composed half a century earlier in Ukraine.3

    While Bialik was writing in Ukraine, the Hungarian Jew Theodor Herzl was busy organizing the Zionist movement in the 1890s and early years of the twentieth century. As a central part of his political activism, Herzl published two books. The Jewish State (1896) was a manifesto outlining Herzl’s idea of establishing a Jewish state in Palestine, and The Old New Land (1902) was a utopian novel set in the year 1923 describing the prosperous Jewish state that Herzl envisioned. The two books—which fatefully also tended to ignore realities on the ground in Palestine—were immensely influential in shaping the Zionist movement. The Old New Land appeared in Hebrew under the title Tel Aviv (a loose Hebrew translation of “Old New Land”). The city of Tel Aviv, established seven years after the book’s publication, took its name from the book. While Bialik is Israel’s national poet, Herzl is known as the visionary of the state.

    The yarns Bialik and Herzl wove ignored many crucial facts about contemporary reality, most notably that around 1900 the Jews of Palestine comprised only 6–9 percent of the region’s total population of about 600,000 people.4 While disregarding such demographic facts, Bialik and Herzl accorded great importance to mythology, most notably the stories of the Bible, without which modern Zionism is unimaginable. Bialik and Herzl were also influenced by the nationalist myths that were created in the nineteenth century by almost every other ethnic group in Europe. The Ukrainian Jew Bialik and the Hungarian Jew Herzl did for Zionism what was earlier done by the poets Taras Shevchenko for Ukrainian nationalism,5 Sándor Petőfi for Hungarian nationalism,6 and Adam Mickiewicz for Polish nationalism.7 Observing the growth of other national movements all around, Herzl wrote that nations arise “out of dreams, songs, fantasies.”8

    But dreams, songs, and fantasies, however inspiring, are not enough to create a functioning nation-state. Bialik inspired generations of Jewish fighters, but to equip and maintain an army, it is also necessary to raise taxes and buy guns. Herzl’s utopian book laid the foundations for the city of Tel Aviv, but to keep the city going, it was also necessary to dig a sewage system. When all is said and done, the essence of patriotism isn’t reciting stirring poems about the beauty of the motherland, and it certainly isn’t making hate-filled speeches against foreigners and minorities. Rather, patriotism means paying your taxes so that people on the other side of the country also enjoy the benefit of a sewage system, as well as security, education, and health care.

    To manage all these services and raise the necessary taxes, enormous amounts of information need to be collected, stored, and processed: information about properties, payments, exemptions, discounts, debts, inventories, shipments, budgets, bills, and salaries. This, however, is not the kind of information that can be turned into a memorable poem or a captivating myth. Instead, tax records come in the shape of various types of lists, ranging from a simple item-by-item record to more elaborate tables and spreadsheets. No matter how intricate these data sets may become, they eschew narrative in favor of dryly listing amounts owed and amounts paid. Poets can afford to ignore such mundane facts, but tax collectors cannot.

    Lists are crucial not only for national taxation systems but also for almost all other complex financial institutions. Corporations, banks, and stock markets cannot exist without them. A church, a university, or a library that wants to balance its budget soon realizes that in addition to priests and poets who can mesmerize people with stories, it needs accountants who know their way around the various types of lists.

    Lists and stories are complementary. National myths legitimize the tax records, while the tax records help transform aspirational stories into concrete schools and hospitals. Something analogous happens in the field of finance. The dollar, the pound sterling, and the bitcoin all come into being by persuading people to believe a story, and tales told by bankers, finance ministers, and investment gurus raise or lower their value. When the chairperson of the Federal Reserve wants to curb inflation, when a finance minister wants to pass a new budget, and when a tech entrepreneur wants to draw investors, they all turn to storytelling. But to actually manage a bank, a budget, or a start-up, lists are essential.

    The big problem with lists, and the crucial difference between lists and stories, is that lists tend to be far more boring than stories, which means that while we easily remember stories, we find it difficult to remember lists. This is an important fact about how the human brain processes information. Evolution has adapted our brains to be good at absorbing, retaining, and processing even very large quantities of information when they are shaped into a story. The Ramayana, one of the foundational tales of Hindu mythology, is twenty-four thousand verses long and runs to about seventeen hundred pages in modern editions, yet despite its enormous length generations of Hindus succeeded in remembering and reciting it by heart.9

    In the twentieth and twenty-first centuries, the Ramayana was repeatedly adapted for film and television. In 1987–88, a seventy-eight-episode version (running to about 2,730 hours) was the most watched television series in the world, with more than 650 million viewers. According to a BBC report, when episodes were aired, “streets would be deserted, shops would be closed, and people would bathe and garland their TV sets.” During the 2020 COVID-19 lockdown the series was re-aired and again became the most watched show in the world.10 While modern TV audiences need not memorize any texts by heart, it is noteworthy how easy they find it to follow the intricate plots of epic dramas, detective thrillers, and soap operas, recalling who each character is and how they are related to numerous others. We are so accustomed to performing such feats of memory that we seldom consider how extraordinary they are.

    What makes us so good at remembering epic poems and long-running TV series is that long-term human memory is particularly adapted to retaining stories. As Kendall Haven writes in his 2007 book Story Proof: The Science Behind the Startling Power of Story, “Human minds … rely on stories and on story architecture as the primary roadmap for understanding, making sense of, remembering, and planning our lives.… Lives are like stories because we think in story terms.” Haven references more than 120 academic studies, concluding that “research overwhelmingly, convincingly, and without opposition provides the evidence” that stories are a highly efficient “vehicle for communicating factual, conceptual, emotional, and tacit information.”11

    In contrast, most people find it hard to remember lists by heart, and few people would be interested in watching a TV recitation of India’s tax records or annual budget. Mnemonic methods used to memorize lists of items often work by weaving the items into a plot, thereby turning the list into a story.12 But even with the help of such mnemonic devices, who could remember their country’s tax records or budget? The information may be vital—determining what quality of health care, education, and welfare services citizens enjoy—but our brains are not adapted to remembering such things. Unlike national poems and myths, which can be stored in our brains, complex national taxation and administration systems have required a unique nonorganic information technology in order to function. This technology is the written document.

    TO KILL A LOAN

    The written document was invented many times in many places. Some of the earliest examples come from ancient Mesopotamia. A cuneiform clay tablet dated to the twenty-eighth day of the tenth month of the forty-first year of the reign of King Shulgi of Ur (ca. 2053/4 BCE) recorded the monthly deliveries of sheep and goats. Fifteen sheep were delivered on the second day of the month, 7 sheep on the third day, 11 sheep on the fourth, 219 on the fifth, 47 on the sixth, and so on until 3 sheep were delivered on the twenty-eighth. In total, says the clay tablet, 896 animals were received that month. Remembering all these deliveries was important for the royal administration, to monitor people’s obedience and to keep track of available resources. While doing so in one’s head was a formidable challenge, it was easy for a learned scribe to write them down on a clay tablet.13

    Like stories and like all other information technologies in history, written documents didn’t necessarily represent reality accurately. The Ur tablet, for example, contained a mistake. The document says that a total of 896 animals were received during that month, but when modern scholars added up all the individual entries they reached a total of 898. The scribe who wrote the document apparently made a mistake when he calculated the overall tally, and the tablet preserved this mistake for posterity.

    But whether true or false, written documents created new realities. By recording lists of properties, taxes, and payments, they made it far easier to create administrative systems, kingdoms, religious organizations, and trade networks. More specifically, documents changed the method used for creating intersubjective realities. In oral cultures, intersubjective realities were created by telling a story that many people repeated with their mouths and remembered in their brains. Brain capacity consequently placed a limit on the kinds of intersubjective realities that humans created. Humans couldn’t forge an intersubjective reality that their brains couldn’t remember.

    This limit could be transcended, however, by writing documents. The documents didn’t represent an objective empirical reality; the reality was the documents themselves. As we shall see in later chapters, written documents thereby provided precedents and models that would eventually be used by computers. The ability of computers to create intersubjective realities is an extension of the power of clay tablets and pieces of paper.

    As a key example, consider ownership. In oral communities that lacked written documents, ownership was an intersubjective reality created through the words and behaviors of the community members. To own a field meant that your neighbors agreed that this field was yours and behaved accordingly. They didn’t build a hut on that field, graze their livestock there, or pick fruits there without first asking your permission. Ownership was created and maintained by people continuously saying or signaling things to one another. This made ownership the affair of a local community and placed a limit on the ability of a distant central authority to control all landownership. No king, minister, or priest could remember who owned each field in hundreds of distant villages. This also placed a limit on the ability of individuals to claim and exercise absolute property rights, and instead favored various forms of communal property rights. For example, your neighbors might acknowledge your right to cultivate a field but not your right to sell it to foreigners.14

    In a literate state, to own a field increasingly came to mean that it is written on some clay tablet, bamboo strip, piece of paper, or silicon chip that you own that field. If your neighbors have been grazing their sheep for years on a piece of land, and none of them ever said that you own it, but you can somehow produce an official document that says it is yours, you have a good chance of enforcing your claim. Conversely, if all the neighbors agree that it is your field but you don’t have any official document that proves it, tough luck. Ownership is still an intersubjective reality created by exchanging information, but the information now takes the form of a written document (or a computer file) rather than of people talking and gesturing to each other. This means that ownership can now be determined by a central authority that produces and holds the relevant documents. It also means that you can sell your field without asking your neighbors’ permission, simply by transferring the crucial document to someone else.

    The power of documents to create intersubjective realities was beautifully manifested in the Old Assyrian dialect, which treated documents as living things that could also be killed. Loan contracts were “killed” (duākum) when the debt was repaid. This was done by destroying the tablet, adding some mark to it, or breaking its seal. The loan contract didn’t represent reality; it was the reality. If somebody repaid the loan but failed to “kill the document,” the debt was still owed. Conversely, if somebody didn’t repay the loan but the document “died” in some other way—perhaps the dog ate it—the debt was no more.15 The same happens with money. If your dog eats a hundred-dollar bill, those hundred dollars cease to exist.

    In Shulgi’s Ur, in ancient Assyria, and in numerous subsequent polities, social, economic, and political relations relied on documents that create reality instead of merely representing it. When writing constitutions, peace treaties, and commercial contracts, lawyers, politicians, and businesspeople wrangle for weeks and even months over each word—because they know that these pieces of paper can wield enormous power.

    BUREAUCRACY

    Every new information technology has its unexpected bottlenecks. It solves some old problems but creates new ones. In the early 1730s BCE, Narâmtani, a priestess in the Mesopotamian city of Sippar, wrote a letter (on a clay tablet) to a relative, asking him to send her a few clay tablets he kept in his house. She explained that her claim to an inheritance was being contested and she couldn’t prove her case in court without those documents. She ended her message with a plea: “Now, do not neglect me!”16

    We don’t know what happened next, but just imagine the situation if the relative searched his house but could not find the missing tablets. As people produced more and more documents, finding them turned out to be far from easy. This was a particular challenge for kings, priests, merchants, and anyone else who accumulated thousands of documents in their archives. How do you find the right tax record, payment receipt, or business contract when you need it? Written documents were much better than human brains in recording certain types of information. But they created a new and very thorny problem: retrieval.17

    The brain is remarkably efficient in retrieving whatever information is stored in its network of tens of billions of neurons and trillions of synapses. Though our brain archives countless complex stories about our personal life, our national history, and our religious mythology, healthy people can retrieve information about any of them in less than a second. What did you eat for breakfast? Who was your first crush? When did your country gain its independence? What’s the first verse in the Bible?

    How did you retrieve all these pieces of information? What mechanism activates the right neurons and synapses to rapidly call up the necessary information? Though neuroscientists have made some progress in the study of memory, nobody yet understands what memories are, or how exactly they are stored and retrieved.18 What we do know is that millions of years of evolution streamlined the brain’s retrieval processes. However, once humans have outsourced memories from organic brains to inorganic documents, retrieval could no longer rely on that streamlined biological system. Nor could it rely on the foraging abilities that humans evolved over millions of years. Evolution has adapted humans for finding fruits and mushrooms in a forest, but not for finding documents in an archive.

    Foragers locate fruits and mushrooms in a forest, because evolution has organized forests according to a discernible organic order. Fruit trees photosynthesize, so they require sunlight. Mushrooms feed on dead organic matter, which can usually be found in the ground. So mushrooms are usually down at soil level, whereas fruits grow further up. Another common rule is that apples grow on apple trees, whereas figs grow on figs trees. So if you are looking for an apple, you first need to locate an apple tree, and then look up. When living in a forest, humans learn this organic order.

    It is very different with archives. Since documents aren’t organisms, they don’t obey any biological laws, and evolution didn’t organize them for us. Tax reports don’t grow on a tax-report shelf. They need to be placed there. For that, somebody first needs to come up with the idea of categorizing information by shelves, and to decide which documents should go on which shelf. Unlike foragers, who need merely to discover the preexisting order of the forest, archivists need to devise a new order for the world. That order is called bureaucracy.

    Bureaucracy is the way people in large organizations solved the retrieval problem and thereby created bigger and more powerful information networks. But like mythology, bureaucracy too tends to sacrifice truth for order. By inventing a new order and imposing it on the world, bureaucracy distorted people’s understanding of the world in unique ways. Many of the problems of our twenty-first-century information networks—like biased algorithms that mislabel people, or rigid protocols that ignore human needs and feelings—are not new problems of the computer age. They are quintessential bureaucratic problems that have existed long before anyone even dreamed of computers.

    BUREAUCRACY AND THE SEARCH FOR TRUTH

    Bureaucracy literally means “rule by writing desk.” The term was invented in eighteenth-century France, when the typical official sat next to a writing desk with drawers—a bureau.19 At the heart of the bureaucratic order, then, is the drawer. Bureaucracy seeks to solve the retrieval problem by dividing the world into drawers, and knowing which document goes into which drawer.

    The principle remains the same regardless of whether the document is placed into a drawer, a shelf, a basket, a jar, a computer folder, or any other receptacle: divide and rule. Divide the world into containers, and keep the containers separate so the documents don’t get mixed up. This principle, however, comes with a price. Instead of focusing on understanding the world as it is, bureaucracy is often busy imposing a new and artificial order on the world. Bureaucrats begin by inventing various drawers, which are intersubjective realities that don’t necessarily correspond to any objective divisions in the world. The bureaucrats then try to force the world to fit into these drawers, and if the fit isn’t very good, the bureaucrats push harder. Anyone who ever filled out an official form knows this only too well. When you fill out the form, and none of the listed options fits your circumstances, you must adapt yourself to the form, rather than the form adapting to you. Reducing the messiness of reality to a limited number of fixed drawers helps bureaucrats keep order, but it comes at the expense of truth. Because they are fixated on their drawers—even when reality is far more complex—bureaucrats often develop a distorted understanding of the world.

    The urge to divide reality into rigid drawers also leads bureaucrats to pursue narrow goals irrespective of the wider impact of their actions. A bureaucrat tasked with increasing industrial production is likely to ignore environmental considerations that fall outside her purview, and perhaps dump toxic waste into a nearby river, leading to an ecological disaster downstream. If the government then establishes a new department to combat pollution, its bureaucrats are likely to push for ever more stringent regulations, even if this results in economic ruin for communities upstream. Ideally, someone should be able to take into account all the different considerations and aspects, but such a holistic approach requires transcending or abolishing the bureaucratic division.

    The distortions created by bureaucracy affect not only government agencies and private corporations but also scientific disciplines. Consider, for example, how universities are divided into different faculties and departments. History is separate from biology and from mathematics. Why? Certainly this division doesn’t reflect objective reality. It is the intersubjective invention of academic bureaucrats. The COVID-19 pandemic, for example, was at one and the same time a historical, biological, and mathematical event. But the academic study of pandemics is divided between the separate departments of history, biology, and mathematics (among others). Students pursuing an academic degree must usually decide to which of these departments they belong. Their decision limits their choice of courses, which in turn shapes their understanding of the world. Mathematics students learn how to predict future morbidity levels from present rates of infection; biology students learn how viruses mutate over time; and history students learn how religious and political beliefs affect people’s willingness to follow government instructions. To fully understand COVID-19 requires taking into account mathematical, biological, and historical phenomena, but academic bureaucracy doesn’t encourage such a holistic approach.

    As you climb the academic ladder, the pressure to specialize only increases. The academic world is ruled by the law of publish or perish. If you want a job, you must publish in peer-reviewed journals. But journals are divided by discipline, and publishing an article on virus mutations in a biology journal demands following different conventions from publishing an article on the politics of pandemics in a history journal. There are different jargons, different citation rules, and different expectations. Historians should have a deep understanding of culture and know how to read and interpret historical documents. Biologists should have a deep understanding of evolution and know how to read and interpret DNA molecules. Things that fall in between categories—like the interplay between human political ideologies and virus evolution—are often left unaddressed.20

    To appreciate how academics force a messy and fluid world into rigid bureaucratic categories, let’s dig a little deeper in the specific discipline of biology. Before Darwin could explain the origin of species, earlier scholars like Carl Linnaeus first had to define what a species is and classify all living organisms into species. To argue that lions and tigers evolved from a common feline ancestor, you first have to define “lions” and “tigers.”21 This turned out to be a difficult and never-ending job, because animals, plants, and other organisms often trespass the boundaries of their allotted drawers.

    Evolution cannot be easily contained in any bureaucratic schema. The whole point of evolution is that species continually change, which means that putting each species in one unchanging drawer distorts biological reality. For example, it is an open question when Homo erectus ended and Homo sapiens began. Were there once two Erectus parents whose child was the first Sapiens?22 Species also keep intermingling, with animals belonging to seemingly separate species not only having sex but even siring fertile offspring. Most Sapiens living today have about 1–3 percent Neanderthal DNA,23 indicating that there once was a child whose father was a Neanderthal and whose mother was a Sapiens (or vice versa). So are Sapiens and Neanderthals the same species or different species? And is “species” an objective reality that biologists discover, or is it an intersubjective reality that biologists impose?24

    There are numerous other examples of animals breaking out of their drawers, so the neat bureaucratic division fails to accurately categorize ring species, fusion species, and hybrids.25 Grizzly bears and polar bears sometimes produce pizzly bears and grolar bears.26 Lions and tigers produce ligers and tigons.27

    When we shift our attention from mammals and other multicellular organisms to the world of single-cell bacteria and archaea, we discover that anarchy reigns. In a process known as horizontal gene transfer, single-cell organisms routinely exchange genetic material not only with organisms from related species but also with organisms from entirely different genera, kingdoms, orders, and even domains. Bacteriologists have a very difficult job keeping tabs on these chimeras.28

    And when we reach the very edge of life and consider viruses like SARS-CoV-2 (responsible for COVID-19), things become even more complicated. Viruses straddle the supposed rigid boundary between living beings and lifeless matter—between biology and chemistry. Unlike bacteria, viruses aren’t single-cell organisms. They aren’t cells at all, and don’t possess any cellular machinery of their own. Viruses don’t eat or metabolize, and cannot reproduce by themselves. They are tiny packets of genetic code, which are able to penetrate cells, hijack their cellular machinery, and instruct them to produce more copies of that alien genetic code. The new copies burst out of the cell to infect and hijack more cells, which is how the alien code turns viral. Scientists argue endlessly about whether viruses should count as life-forms or whether they fall outside the boundary of life.29 But this boundary isn’t an objective reality; it is an intersubjective convention. Even if biologists reach a consensus that viruses are life-forms, it wouldn’t change anything about how viruses behave; it will only change how humans think about them.

    Of course, intersubjective conventions are themselves part of reality. As we humans become more powerful, so our intersubjective beliefs become more consequential for the world outside our information networks. For example, scientists and legislators have categorized species according to the threat of extinction they face, on a scale ranging from “least concern” through “vulnerable” and “endangered” to “extinct.” Defining a particular population of animals as an “endangered species” is an intersubjective human convention, but it can have far-reaching consequences, for instance by imposing legal restrictions on hunting those animals or destroying their habitat. A bureaucratic decision about whether a certain animal belongs in the “endangered species” drawer or in the “vulnerable species” drawer could make the difference between life and death. As we shall see time and again in subsequent chapters, when a bureaucracy puts a label on you, even though the label might be pure convention, it can still determine your fate. That’s true whether the bureaucrat is a flesh-and-blood expert on animals; a flesh-and-blood expert on humans; or an inorganic AI.

    THE DEEP STATE

    In defense of bureaucracy it should be noted that while it sometimes sacrifices truth and distorts our understanding of the world, it often does so for the sake of order, without which it would be hard to maintain any large-scale human network. While bureaucracies are never perfect, is there a better way to manage big networks? For example, if we decided to abolish all conventional divisions in the academic world, all departments and faculties and specialized journals, would every prospective doctor be expected to devote several years to the study of history, and would people who studied the impact of the Black Death on Christian theology be considered expert virologists? Would it lead to better health-care systems?

    Anyone who fantasizes about abolishing all bureaucracies in favor of a more holistic approach to the world should reflect on the fact that hospitals too are bureaucratic institutions. They are divided into different departments, with hierarchies, protocols, and lots of forms to fill out. They suffer from many bureaucratic illnesses, but they still manage to cure us of many of our biological illnesses. The same goes for almost all the other services that make our life better, from our schools to our sewage system.

    When you flush the toilet, where does the waste go? It goes into the deep state. There is an intricate subterranean web of pipes, pumps, and tunnels that runs under our houses and collects our waste, separates it from the supply of drinking water, and either treats or safely disposes of it. Somebody needs to design, construct, and maintain that deep web, plug holes in it, monitor pollution levels, and pay the workers. That too is bureaucratic work, and we would face a lot of discomfort and even death if we abolished that particular department. Sewage water and drinking water are always in danger of mixing, but luckily for us there are bureaucrats who keep them separate.

    Prior to the establishment of modern sewage systems, waterborne infectious diseases like dysentery and cholera killed millions of people around the world.30 In 1854 hundreds of London residents began dying of cholera. It was a relatively small outbreak, but it proved to be a turning point in the history of cholera, of epidemics more generally, and of sewage. The leading medical theory of the day argued that cholera epidemics were caused by “bad air.” But the physician John Snow suspected that the cause was the water supply. He painstakingly tracked and listed all known cholera patients, their place of residence, and their source of water. The resulting data led him to identify the water pump on Broad Street in Soho as the epicenter of the outbreak.

    This was tedious bureaucratic work—collecting data, categorizing it, and mapping it—but it saved lives. Snow explained his findings to local officials, persuading them to disable the Broad Street pump, which effectively ended the outbreak. Subsequent research discovered that the well providing water to the Broad Street pump was dug less than a meter from a cholera-infected cesspit.31

    Snow’s discovery, and the work of many subsequent scientists, engineers, lawyers, and officials, resulted in a sprawling bureaucracy regulating cesspits, water pumps, and sewage lines. In today’s England, digging wells and constructing cesspits require filling out forms and getting licenses, which ensure that drinking water doesn’t come from a well someone dug next to a cesspit.32

    It is easy to forget about this system when it works well, but since 1854 it has saved millions of lives, and it is one of the most important services provided by modern states. In 2014, Prime Minister Narendra Modi of India identified the lack of toilets as one of India’s biggest problems. Open defecation is a major cause for spreading diseases like cholera, dysentery, and diarrhea, as well as exposing women and girls to sexual assaults. As part of his flagship Clean India Mission, Modi promised to provide all Indian citizens with access to toilets, and between 2014 and 2020 the Indian state invested around ten billion dollars in the project, building more than 100 million new latrines.33 Sewage isn’t the stuff of epic poems, but it is a test of a well-functioning state.

    THE BIOLOGICAL DRAMAS

    Mythology and bureaucracy are the twin pillars of every large-scale society. Yet while mythology tends to inspire fascination, bureaucracy tends to inspire suspicion. Despite the services they provide, even beneficial bureaucracies often fail to win the public’s trust. For many people, the very word “bureaucracy” carries negative connotations. This is because it is inherently difficult to know whether a bureaucratic system is beneficial or malicious. For all bureaucracies—good or bad—share one key characteristic: it is hard for humans to understand them.

    Any kid can tell the difference between a friend and a bully. You know if someone shares their lunch with you or instead takes yours. But when the tax collector comes to take a cut from your earnings, how can you tell whether it goes to build a new public sewage system or a new private dacha for the president? It is hard to get all the relevant information, and even harder to interpret it. It is similarly difficult for citizens to understand the bureaucratic procedures determining how pupils are admitted to schools, how patients are treated in hospitals, or how garbage is collected and recycled. It takes a minute to tweet allegations of bias, fraud, or corruption, and many weeks of arduous work to prove or disprove them.

    Documents, archives, forms, licenses, regulations, and other bureaucratic procedures have changed the way information flows in society, and with it the way power works. This made it far more difficult to understand power. What is happening behind the closed doors of offices and archives, where anonymous officials analyze and organize piles of documents and determine our fate with a stroke of a pen or a click of a mouse?

    In tribal societies that lack written documents and bureaucracies, the human network is composed of only human-to-human and human-to-story chains. Authority belongs to the people who control the junctions that link the various chains. These junctions are the tribe’s foundational myths. Charismatic leaders, orators, and mythmakers know how to use these stories in order to shape identities, build alliances, and sway emotions.34

    In human networks connected by written documents and bureaucratic procedures—from ancient Ur to modern India—society relies in part on the interaction between humans and documents. In addition to human-to-human and human-to-story chains, such societies are held together by human-to-document chains. When we observe a bureaucratic society at work, we still see humans telling stories to other humans, as when millions of Indians watch the Ramayana series, but we also see humans passing documents to other humans, as when TV networks are required to apply for broadcasting licenses and fill out tax reports. Looked at from a different perspective, what we see is documents compelling humans to engage with other documents.

    This led to shifts in authority. As documents became a crucial nexus linking many social chains, considerable power came to be invested in these documents, and experts in the arcane logic of documents emerged as new authority figures. Administrators, accountants, and lawyers mastered not just reading and writing but also the skills of composing forms, separating drawers, and managing archives. In bureaucratic systems, power often comes from understanding how to manipulate obscure budgetary loopholes and from knowing your way around the labyrinths of offices, committees, and subcommittees.

    This shift in authority changed the balance of power in the world. For better or worse, literate bureaucracies tended to strengthen the central authority at the expense of ordinary citizens. It’s not just that documents and archives made it easier for the center to tax, judge, and conscript everybody. The difficulty of understanding bureaucratic power simultaneously made it harder for the masses to influence, resist, or evade the central authority. Even when bureaucracy was a benign force, providing people with sewage systems, education, and security, it still tended to increase the gap between rulers and ruled. The system enabled the center to collect and record a lot more information about the people it governed, while the latter found it much more difficult to understand how the system itself worked.

    Art, which helps us understand many other aspects of life, offered only limited assistance in this case. Poets, playwrights, and moviemakers have occasionally focused on the dynamics of bureaucratic power. However, this has proven to be a very difficult story to communicate. Artists usually work with a limited set of story lines that are rooted in our biology, but none of these biological dramas sheds much light on the workings of bureaucracy, because they have all been scripted by evolution millions of years before the emergence of documents and archives. To understand what “biological dramas” are, and why they are a poor guide for understanding bureaucracy, let’s consider in detail the plot of one of humanity’s greatest artistic masterpieces—the Ramayana.

    One important plotline of the Ramayana concerns the relations between the eponymous prince, Rama, his father, King Dasharatha, and his stepmother, Queen Kaikeyi. Though Rama, being the eldest son, is the rightful heir to the kingdom, Kaikeyi persuades the king to banish Rama to the wilderness and bestow the succession instead on her son Bharata. Underlying this plotline are several biological dramas that go back hundreds of millions of years in mammalian and avian evolution.

    All mammal and bird offspring depend on their parents in the first stage of life, seek parental care, and fear parental neglect or hostility. Life and death hang in the balance. A cub or chick pushed out of the nest too soon is in danger of death from starvation or predation. Among humans, the fear of being neglected or abandoned by one’s parents is a template not just for children’s stories like Snow White, Cinderella, and Harry Potter but also for some of our most influential national and religious myths. The Ramayana is far from being the sole example. In Christian theology damnation is conceived as losing all contact with the mother church and the heavenly father. Hell is a lost child crying for his or her missing parents.

    A related biological drama, which is also familiar to human children, mammalian cubs, and avian chicks, is “Father loves me more than he loves you.” Biologists and geneticists have identified sibling rivalry as one of the key processes of evolution.35 Siblings routinely compete for food and parental attention, and in some species the killing of one sibling by another is commonplace. About a quarter of spotted hyena cubs are killed by their siblings, who typically enjoy greater parental care as a result.36 Among sand tiger sharks, females hold numerous embryos in their uterus. The first embryo that reaches about ten centimeters in length then eats all the others.37 The dynamics of sibling rivalry are manifested in numerous myths in addition to the Ramayana, for instance in the stories of Cain and Abel, King Lear, and the TV series Succession. Entire nations—like the Jewish people—may base their identity on the claim that “we are Father’s favorite children.”

    The second major plotline of the Ramayana focuses on the romantic triangle formed by Prince Rama, his lover, Sita, and the demon-king Ravana, who kidnaps Sita. “Boy meets girl” and “boy fights boy over girl” are also biological dramas that have been enacted by countless mammals, birds, reptiles, and fish for hundreds of millions of years. We are mesmerized by these stories because understanding them has been essential for our ancestors’ survival. Human storytellers like Homer, Shakespeare, and Valmiki—the purported author of the Ramayana—have displayed an amazing capacity to elaborate on the biological dramas, but even the greatest poetical narratives usually copy their basic plotline from the handbook of evolution.

    A third theme recurring in the Ramayana is the tension between purity and impurity, with Sita being the paragon of purity in Hindu culture. The cultural obsession with purity originates in the evolutionary struggle to avoid pollution. All animals are torn between the need to try new food and the fear of being poisoned. Evolution therefore equipped animals with both curiosity and the capacity to feel disgust on coming into contact with something toxic or otherwise dangerous.38 Politicians and prophets have learned how to manipulate these disgust mechanisms. In nationalist and religious myths, countries or churches are depicted as a biological body in danger of being polluted by impure intruders. For centuries bigots have often said that ethnic and religious minorities spread diseases,39 that LGBTQ people are a source of pollution,40 or that women are impure.41 During the Rwanda genocide of 1994, Hutu propaganda referred to the Tutsis as cockroaches. The Nazis compared Jews to rats. Experiments have shown that chimpanzees, too, react with disgust to images of unfamiliar chimpanzees from another band.42

    Perhaps in no other culture was the biological drama of “purity versus impurity” carried to greater extremes than in traditional Hinduism. It constructed an intersubjective system of castes ranked by their supposed level of purity, with the pure Brahmins at the top and the allegedly impure Dalit (formerly known as untouchables) at the bottom. Professions, tools, and everyday activities have also been classified by their level of purity, and strict rules have forbidden “impure” persons to marry “pure” people, touch them, prepare food for them, or even come near them.

    The modern state of India still struggles with this legacy, which influences almost all aspects of life. For example, fears of impurity created various complications for the aforementioned Clean India Mission, because allegedly “pure” people were reluctant to get involved in “impure” activities such as building, maintaining, and cleaning toilets, or to share public latrines with allegedly “impure” persons.43 On September 25, 2019, two Dalit children—twelve-year-old Roshni Valmiki and her ten-year-old nephew Avinash—were lynched in the Indian village of Bhakhedi for defecating near the house of a family from the higher Yadav caste. They were forced to defecate in public because their houses lacked functioning toilets. A local official later explained that their household—while being among the poorest in the village—was nevertheless excluded from the list of families eligible for government aid to build toilets. The children routinely suffered from other caste-based discrimination, for example being forced to bring separate mats and utensils to school and to sit apart from the other pupils, so as not to “pollute” them.44

    The list of biological dramas that press our emotional buttons includes several additional classics, such as “Who will be alpha?” “Us versus them,” and “Good versus evil.” These dramas, too, feature prominently in the Ramayana, and all of them are well known to wolf packs and chimpanzee bands as well as to human societies. Together, these biological dramas form the backbone of almost all human art and mythology. But art’s dependence on the biological dramas have made it difficult for artists to explain the mechanisms of bureaucracy. The Ramayana is set within the context of large agrarian kingdoms, but it shows little interest in how such kingdoms register property, collect taxes, catalog archives, or finance wars. Sibling rivalry and romantic triangles aren’t a good guide for the dynamics of documents, which have no siblings and no romantic life.

    Storytellers like Franz Kafka, who focused on the often surreal ways that bureaucracy shapes human lives, pioneered new nonbiological plotlines. In Kafka’s Trial, the bank clerk K. is arrested by unidentified officials of an unfathomable agency for an unnamed crime. Despite his best efforts, he never understands what is happening to him or uncovers the aims of the agency that is crushing him. While sometimes taken as an existential or theological reference to the human condition in the universe and to the unfathomability of God, on a more mundane level the story highlights the potentially nightmarish character of bureaucracies, which as an insurance lawyer Kafka knew all too well.

    In bureaucratic societies, the lives of ordinary people are often upended by unidentified officials of an unfathomable agency for incomprehensible reasons. Whereas stories about heroes who confront monsters—from the Ramayana to Spider-Man—repackage the biological dramas of confronting predators and romantic rivals, the unique horror of Kafkaesque stories comes from the unfathomability of the threat. Evolution has primed our minds to understand death by a tiger. Our mind finds it much more difficult to understand death by a document.

    Some portrayals of bureaucracy are satirical. Joseph Heller’s iconic 1961 novel, Catch-22, illustrated the central role bureaucracy plays in war. The ex–private first class Wintergreen in the mail room—who decides which letters to forward—is a more powerful figure than any general.45 The 1980s British sitcoms Yes Minister and Yes, Prime Minister showed the ways that civil servants use arcane regulations, obscure subcommittees, and piles of documents to manipulate and control their political bosses. The 2015 comedy-drama The Big Short (based on a 2010 book by Michael Lewis) explored the bureaucratic roots of the 2007–8 financial crisis. The movie’s arch-villains are not humans but collateralized debt obligations (CDOs), which are financial devices invented by investment bankers and understood by nobody else in the world. These bureaucratic Godzillas slumbered unnoticed in the depths of bank portfolios, until they suddenly emerged in 2007 to wreak havoc on the lives of billions of people by instigating a major financial crisis.

    Artworks like these have had some success in shaping perceptions of how bureaucratic power works, but this is an uphill battle, because since the Stone Age our minds have been primed to focus on biological dramas rather than bureaucratic ones. Most Hollywood and Bollywood blockbusters are not about CDOs. Rather, even in the twenty-first century, most blockbusters are essentially Stone Age stories about the hero who fights the monster to win the girl. Similarly, when depicting the dynamics of political power, TV series like Game of Thrones, The Crown, and Succession focus on the family intrigues of the dynastic court rather than on the bureaucratic labyrinth that sustains—and sometimes curbs—the dynasty’s power.

    LET’S KILL ALL THE LAWYERS

    The difficulty of depicting and understanding bureaucratic realities has had unfortunate results. On the one hand, it leaves people feeling helpless in the face of harmful powers they do not understand, like the hero of Kafka’s Trial. On the other hand, it also leaves people with the impression that bureaucracy is a malign conspiracy, even in cases when it is in fact a benign force providing us with health care, security, and justice.

    In the sixteenth century, Ludovico Ariosto described the allegorical figure of Discord as a woman who walks around in a cloud of “sheaves of summonses and writs, cross-examinations and powers of attorney, and great piles of glosses, counsel’s opinions and precedents—all of which tended to the greater insecurity of impoverished folk. In front and behind her and on either side she was hemmed in by notaries, attorneys and barristers.”46

    In his description of Jack Cade’s Rebellion (1450) in Henry VI, Part 2, Shakespeare has a commoner rebel called Dick the Butcher take the antipathy to bureaucracy to its logical conclusion. Dick has a plan to establish a better social order. “The first thing we do,” advises Dick, “let’s kill all the lawyers.” The rebel leader, Jack Cade, runs with Dick’s proposal in a forceful attack on bureaucracy and in particular on written documents: “Is not this a lamentable thing, that of the skin of an innocent lamb should be made parchment? That parchment, being scribbled o’er, should undo a man? Some say the bee stings: but I say, ’tis the bee’s wax; for I did but seal once to a thing, and I was never mine own man since.” Just then the rebels capture a clerk and accuse him of being able to write and read. After a short interrogation that establishes his “crime,” Cade orders his men, “Hang him with his pen and inkhorn about his neck.”47

    Seventy years prior to Jack Cade’s Rebellion, during the even bigger 1381 Peasants’ Revolt, the rebels focused their ire not only on flesh-and-blood bureaucrats but also on their documents, destroying numerous archives, burning court rolls, charters, and administrative and legal records. In one incident, they made a bonfire of the archives of the University of Cambridge. An old woman named Margery Starr scattered the ashes to the winds while crying, “Away with the learning of the clerks, away with it!” Thomas Walsingham, a monk in St. Albans Abbey who witnessed the destruction of the abbey’s archive firsthand, described how the rebels “set fire to all court rolls and muniments, so that after they had got rid of these records of their ancient service their lords would not be able to claim any right at all against them at some future time.”48 Killing the documents erased the debts.

    Similar attacks on archives characterized numerous other insurgencies throughout history. For example, during the Great Jewish Revolt in 66 CE, one of the first things the rebels did upon capturing Jerusalem was to set fire to the central archive in order to destroy records of debts, thereby wining the support of the populace.49 During the French Revolution in 1789, numerous local and regional archives were destroyed for comparable reasons.50 Many rebels might have been illiterate, but they knew that without the documents the bureaucratic machine couldn’t function.

    I can sympathize with the suspicion of government bureaucracies and of the power of official documents, because they have played an important role in my own family. My maternal grandfather had his life upended by a government census and by the inability to find a crucial document. My grandfather Bruno Luttinger was born in 1913 in Chernivtsi. Today this town is in Ukraine, but in 1913 it was part of the Habsburg Empire. Bruno’s father disappeared in World War I, and he was raised by his mother, Chaya-Pearl. When the war was over, Chernivtsi was annexed to Romania. In the late 1930s, as Romania became a fascist dictatorship, an important plank of its new antisemitic policy was to conduct a Jewish census.

    In 1936 official statistics said that 758,000 Jews lived in Romania, constituting 4.2 percent of the population. The same official statistics said that the total number of refugees from the U.S.S.R., Jews and non-Jews, was about 11,000. In 1937 a new fascist government came to power, headed by Prime Minister Octavian Goga. Goga was a renowned poet as well as a politician, but he quickly graduated from patriotic poetry to fake statistics and oppressive bureaucracy. He and his colleagues ignored the official statistics and claimed that hundreds of thousands of Jewish refugees were flooding into Romania. In several interviews Goga claimed that half a million Jews had entered Romania illegally and that the total number of Jews in the country was 1.5 million. Government organs, far-right statisticians, and popular newspapers regularly cited even higher figures. The Romanian embassy in Paris, for example, claimed there were a million Jewish refugees in Romania. Christian Romanians were gripped by mass hysteria that they would soon be replaced or become a minority in a Jewish-led country.

    Goga’s government stepped in to offer a solution to the imaginary problem invented by its own propaganda. On January 22, 1938, the government issued a law ordering all Jews in Romania to provide documented proof that they were born in Romanian territory and were entitled to Romanian citizenship. Jews who failed to provide proof would lose their citizenship, along with all rights to residence and employment.

    Suddenly Romania’s Jews found themselves in a bureaucratic hell. Many had to travel to their birthplace to look for the relevant documents, only to discover that the municipal archives were destroyed during World War I. Jews born in territories annexed to Romania only after 1918—like Chernivtsi—faced special difficulties, because they lacked Romanian birth certificates and because many other documents about their families were archived in the former Habsburg capitals of Vienna and Budapest instead of in Bucharest. Jews often didn’t even know which documents they were supposed to be looking for, because the census law didn’t specify which documents were considered sufficient “proof.”

    Clerks and archivists gained a new and lucrative source of income as frantic Jews offered to pay large bribes to get their hands on the right document. Even if no bribes were involved, the process was extremely costly: any request for documentation, as well as filing the citizenship request with the authorities, involved paying fees. Finding and filing the right document did not guarantee success. A difference of a single letter between how a name was spelled on the birth certificate and on the citizenship papers was enough for the authorities to revoke the citizenship.

    Many Jews could not clear these bureaucratic hurdles and didn’t even file a citizenship request. Of those who did, only 63 percent got their citizenship approved. Altogether, out of 758,000 Romanian Jews, 367,000 lost their citizenship.51 My grandfather Bruno was among them. When the new census law was passed in Bucharest, Bruno did not think much about it. He was born in Chernivtsi and lived there all his life. The thought that he needed to prove to some bureaucrat that he was not an alien struck him as ridiculous. Moreover, in early 1938 his mother fell ill and died, and Bruno felt he had much bigger things to worry about than chasing documents.

    In December 1938 an official letter arrived from Bucharest canceling Bruno’s citizenship, and as an alien he was promptly fired from his job in a Chernivtsi radio shop. Bruno was now not only alone and jobless but also stateless and without much prospect for alternative employment. Nine months later World War II erupted, and the danger for paperless Jews was mounting. Of the Romanian Jews who lost their citizenship in 1938, the vast majority would be murdered over the next few years by the Romanian fascists and their Nazi allies (Jews who retained their citizenship had a much higher survival rate).52

    My grandfather repeatedly tried to escape the tightening noose, but it was difficult without the right papers. Several times he smuggled himself onto trains and ships, only to be caught and arrested. In 1940 he finally managed to board one of the last ships bound for Palestine before the gates of hell slammed shut. When he arrived in Palestine, he was immediately imprisoned by the British as an illegal immigrant. After two months in prison, the British offered a deal: stay in jail and risk deportation, or enlist in the British army and get Palestinian citizenship. My grandfather grabbed the offer with both hands and from 1941 to 1945 served in the British army in the North African and Italian campaigns. In exchange, he got his papers.

    In our family it became a sacred duty to preserve documents. Bank statements, electricity bills, expired student cards, letters from the municipality—if it had an official-looking stamp on it, it would be filed in one of the many folders in our cupboard. You never knew which of these documents might one day save your life.

    THE MIRACLE DOCUMENT

    Should we love the bureaucratic information network or hate it? Stories like that of my grandfather indicate the dangers inherent in bureaucratic power. Stories like that of the London cholera epidemic indicate its potential benevolence. All powerful information networks can do both good and ill, depending on how they are designed and used. Merely increasing the quantity of information in a network doesn’t guarantee its benevolence, nor make it any easier to find the right balance between truth and order. That is a key historical lesson for the designers and users of the new information networks of the twenty-first century.

    Future information networks, particularly those based on AI, will be different from previous networks in many ways. While in part 1 we are examining how mythology and bureaucracy have been essential for large-scale information networks, in part 2 we will see how AI is taking up the role of both bureaucrats and mythmakers. AI tools know how to find and process data better than flesh-and-blood bureaucrats, and AI is also acquiring the ability to compose stories better than most humans.

    But before we explore the new AI-based information networks of the twenty-first century, and before we examine the threats and promises of AI mythmakers and AI bureaucrats, there is one more thing we need to understand about the long-term history of information networks. We have now seen that information networks don’t maximize truth, but rather seek to find a balance between truth and order. Bureaucracy and mythology are both essential for maintaining order, and both are happy to sacrifice truth for the sake of order. What mechanisms, then, ensure that bureaucracy and mythology don’t lose touch with truth altogether, and what mechanisms enable information networks to identify and correct their own mistakes, even at the price of some disorder?

    The way human information networks have dealt with the problem of errors will be the main subject of the next two chapters. We’ll start by considering the invention of another information technology: the holy book. Holy books like the Bible and the Quran are an information technology that is meant to both include all the vital information society needs and be free from all possibility of error. What happens when an information network believes itself to be utterly incapable of any error? The history of allegedly infallible holy books highlights some of the limitations of all information networks and holds important lessons for the attempt to create infallible AIs in the twenty-first century.

    CHAPTER 4 Errors: The Fantasy of Infallibility

    As Saint Augustine famously said, “To err is human; to persist in error is diabolical.”1 The fallibility of human beings, and the need to correct human errors, have played key roles in every mythology. According to Christian mythology, the whole of history is an attempt to correct Adam and Eve’s original sin. According to Marxist-Leninist thinking, even the working class is likely to be fooled by its oppressors and misidentify its own interests, which is why it requires the leadership of a wise party vanguard. Bureaucracy, too, is constantly on the lookout for errors, from misplaced documents to inefficient procedures. Complex bureaucratic systems usually contain self-disciplinary bodies, and when a major catastrophe occurs—like a military defeat or a financial meltdown—commissions of inquiry are set up to understand what went wrong and make sure the same mistake is not repeated.

    In order to function, self-correcting mechanisms need legitimacy. If humans are prone to error, how can we trust the self-correcting mechanisms to be free from error? To escape this seemingly endless loop, humans have often fantasized about some superhuman mechanism, free from all error, that they can rely upon to identify and correct their own mistakes. Today one might hope that AI could provide such a mechanism, as when in April 2023 Elon Musk announced, “I’m going to start something, which I call TruthGPT or a maximum truth-seeking AI that tries to understand the nature of the universe.”2 We will see in later chapters why this is a dangerous fantasy. In previous eras, such fantasies took a different form—religion.

    In our personal lives, religion can fulfill many different functions, like providing solace or explaining the mysteries of life. But historically, the most important function of religion has been to provide superhuman legitimacy for the social order. Religions like Judaism, Christianity, Islam, and Hinduism propose that their ideas and rules were established by an infallible superhuman authority, and are therefore free from all possibility of error, and should never be questioned or changed by fallible humans.

    TAKING HUMANS OUT OF THE LOOP

    At the heart of every religion lies the fantasy of connecting to a superhuman and infallible intelligence. This is why, as we shall explore in chapter 8, studying the history of religion is highly relevant to present-day debates about AI. In the history of religion, a recurrent problem is how to convince people that a certain dogma indeed originated from an infallible superhuman source. Even if in principle I am eager to submit to the gods’ will, how do I know what the gods really want?

    Throughout history many humans claimed to convey messages from the gods, but the messages often contradicted each other. One person said a god appeared to her in a dream; another person said she was visited by an angel; a third recounted how he met a spirit in a forest—and each preached a different message. The anthropologist Harvey Whitehouse recounts how when he was doing fieldwork among the Baining people of New Britain in the late 1980s, a young man called Tanotka fell sick, and in his feverish delirium began making cryptic statements like “I am Wutka” and “I am a post.” Most of these statements were heard only by Tanotka’s older brother, Baninge, who began telling about them to other people and interpreting them in a creative way. Baninge said that his brother was possessed by an ancestral spirit called Wutka and that he was divinely chosen to be the main support of the community, just as local houses were supported by a central post.

    After Tanotka recovered, he continued to deliver cryptic messages from Wutka, which were interpreted by Baninge in ever more elaborate ways. Baninge also began having dreams of his own, which allegedly revealed additional divine messages. He claimed that the end of the world was imminent, and convinced many of the locals to grant him dictatorial powers so that he could prepare the community for the coming apocalypse. Baninge proceeded to waste almost all the community’s resources on extravagant feasts and rituals. When the apocalypse didn’t materialize and the community almost starved, Baninge’s power collapsed. Though some locals continued to believe that he and Tanotka were divine messengers, many others concluded that the two were charlatans—or perhaps the servants of the Devil.3

    How could people distinguish the true will of the gods from the inventions or imaginations of fallible humans? Unless you had a personal divine revelation, knowing what the gods said meant trusting what fallible humans like Tanotka and Baninge claimed the gods said. But how can you trust these humans, especially if you don’t know them personally? Religion wants to take fallible humans out of the loop and give people access to infallible superhuman laws, but religion repeatedly boiled down to trusting this or that human.

    One way around this problem was to create religious institutions that vetted the purported divine messengers. Already in tribal societies communication with superhuman entities like tribal spirits was often the domain of religious experts. Among the Baining people, specialized spirit mediums known as agungaraga were traditionally responsible for communicating with the spirits and thereby learning the hidden causes of misfortunes ranging from illness to crop failure. Their membership in an established institution made the agungaraga more trustworthy than Tanotka and Baninge, and made their authority more stable and widely acknowledged.4 Among the Kalapalo tribe of Brazil religious rituals were organized by hereditary ritual officers known as the anetaū. In ancient Celtic and Hindu societies similar duties were the preserve of druids and Brahmins.5 As human societies grew and became more complex, so did their religious institutions. Priests and oracles had to train long and hard for the important task of representing the gods, so people no longer needed to trust just any layperson who claimed to have met an angel or to carry a divine message.6 In ancient Greece, for example, if you wanted to know what the gods said, you went to an accredited expert like the Pythia—the high priestess at the temple of Apollo in Delphi.

    But as long as religious institutions like oracular temples were staffed by fallible humans, they too were open to error and corruption. Herodotus recounts that when Athens was ruled by the tyrant Hippias, the pro-democracy faction bribed the Pythia to help them. Whenever any Spartan came to the Pythia to consult the gods on either official or private matters, the Pythia invariably replied that the Spartans must first free Athens from the tyrant. The Spartans, who were Hippias’s allies, eventually submitted to the alleged will of the gods and sent an army to Athens that deposed Hippias in 510 BCE, leading to the establishment of Athenian democracy.7

    If a human prophet could falsify the words of a god, then the key problem of religion wasn’t solved by creating religious institutions like temples and priestly orders. People still needed to trust fallible humans in order to access the supposedly infallible gods. Was it possible to somehow bypass the humans altogether?

    THE INFALLIBLE TECHNOLOGY

    Holy books like the Bible and the Quran are a technology to bypass human fallibility, and religions of the book—like Judaism, Christianity, and Islam—have been built around that technological artifact. To appreciate how this technology is meant to work, we should begin by explaining what a book is and what makes books different from other kinds of written texts. A book is a fixed block of texts—such as chapters, stories, recipes, or epistles—that always go together and have many identical copies. This makes a book something different from oral tales, from bureaucratic documents, and from archives. When telling a story orally, every time we tell the story it might be a little different, and if many people tell the story over a long time, significant variations are bound to creep in. In contrast, all copies of a book are supposed to be identical. As for bureaucratic documents, they tend to be relatively short, and often exist only as a single copy in one archive. If a long document has many copies placed in numerous archives, we would normally call it a book. Finally, a book that contains many texts is also different from an archive, because each archive contains a different collection of texts, whereas all copies of a book contain the same chapters, the same stories, or the same recipes. The book thereby ensures that many people in many times and places can access the same database.

    The book became an important religious technology in the first millennium BCE. After tens of thousands of years in which gods spoke to humans via shamans, priests, prophets, oracles, and other human messengers, religious movements like Judaism began arguing that the gods speak through this novel technology of the book. There is one specific book whose many chapters allegedly contain all the divine words about everything from the creation of the universe to food regulations. Crucially, no priest, prophet, or human institution can forget or change these divine words, because you can always compare what the fallible humans are telling you with what the infallible book records.

    But religions of the book had their own set of problems. Most obviously, who decides what to include in the holy book? The first copy didn’t come down from heaven. It had to be compiled by humans. Still, the faithful hoped that this thorny problem could be solved by a once-and-for-all supreme effort. If we could get together the wisest and most trustworthy humans, and they could all agree on the contents of the holy book, from that moment onward we could excise humans from the loop, and the divine words would forever be safe from human interference.

    Many objections can be raised against this procedure: Who selects the wisest humans? On the basis of what criteria? What if they cannot reach a consensus? What if they later change their minds? Nevertheless, this was the procedure used to compile holy books like the Hebrew Bible.

    THE MAKING OF THE HEBREW BIBLE

    During the first millennium BCE, Jewish prophets, priests, and scholars produced an extensive collection of stories, documents, prophecies, poems, prayers, and chronicles. The Bible as a single holy book didn’t exist in biblical times. King David or the prophet Isaiah never saw a copy of the Bible.

    It is sometimes claimed, erroneously, that the oldest surviving copy of the Bible comes from the Dead Sea Scrolls. These scrolls are a collection of about nine hundred different documents, written mostly in the last two centuries BCE and found in various caves around Qumran, a village near the Dead Sea.8 Most scholars believe they constituted the archive of a Jewish sect that lived nearby.9

    Significantly, none of the scrolls contains a copy of the Bible, and no scroll indicates that the twenty-four books of the Old Testament were considered a single and complete database. Some of the scrolls certainly record texts that are today part of the canonical Bible. For example, nineteen scrolls and fragmentary manuscripts preserve parts of the book of Genesis.10 But many scrolls record texts that were later excluded from the Bible. For example, more than twenty scrolls and fragments preserve parts of the book of Enoch—a book allegedly written by the patriarch Enoch, the great-grandfather of Noah, and containing the history of the angels and demons as well as a prophecy about the coming of the Messiah.11 The Jews of Qumran apparently gave great importance to both Genesis and Enoch, and did not think that Genesis was canonical while Enoch was apocryphal.12 Indeed, to this day some Ethiopian Jewish and Christian sects consider Enoch part of their canon.13

    Even the scrolls that record future canonical texts sometimes differ from the present-day canonical version. For example, the canonical text of Deuteronomy 32:8 says that God divided the nations of the earth according to “the number of the sons of Israel.” The version recorded in the Dead Sea Scrolls has “the number of the sons of God” instead, implying a rather startling notion that God has multiple sons.14 In Deuteronomy 8:6 the canonical text requires the faithful to fear God, whereas the Dead Sea version asks them to love God.15 Some variations are much more substantial than just a single word here or there. The Psalms scrolls contain several entire psalms that are missing from the canonical Bible (most notably Psalms 151, 154, 155).16

    Similarly, the oldest translation of the Bible—the Greek Septuagint—completed between the third and the first centuries BCE, is different in many ways from the later canonical version.17 It includes, for example, the books of Tobit, Judith, Sirach, Maccabees, the Wisdom of Solomon, the Psalms of Solomon, and Psalm 151.18 It also has longer versions of Daniel and Esther.19 Its book of Jeremiah is 15 percent shorter than the canonical version.20 Finally, in Deuteronomy 32:8 most Septuagint manuscripts have either “sons of God” or “angels of God” rather than “sons of Israel.”21

    It took centuries of hairsplitting debates among learned Jewish sages—known as rabbis—to streamline the canonical database and to decide which of the many texts in circulation would get into the Bible as the official word of Jehovah and which would be excluded. By the time of Jesus agreement was probably reached on most of the texts, but even a century later rabbis were still arguing whether the Song of Songs should be part of the canon or not. Some rabbis condemned that text as secular love poetry, while Rabbi Akiva (d. 135 CE) defended it as the divinely inspired creation of King Solomon. Akiva famously said that “the Song of Songs is the Holy of Holies.”22 By the end of the second century CE widespread consensus was apparently reached among Jewish rabbis about which texts were part of the biblical canon and which were not, but debates about this matter, and about the precise wordings, spelling, and pronunciation of each text, were not finally resolved until the Masoretic era (seventh to tenth centuries CE).23

    This process of canonization decided that Genesis was the word of Jehovah, but the book of Enoch, the Life of Adam and Eve, and the Testament of Abraham were human fabrications.24 The Psalms of King David were canonized (minus psalms 151–55), but the Psalms of King Solomon were not. The book of Malachi got the seal of approval; the book of Baruch did not. Chronicles, yes; Maccabees, no.

    Interestingly, some books mentioned in the Bible itself failed to get into the canon. For example, the books of Joshua and Samuel both refer to a very ancient sacred text known as the book of Jasher (Joshua 10:13, 2 Samuel 1:18). The book of Numbers refers to “the Book of the Wars of the Lord” (Numbers 21:14). And when 2 Chronicles surveys the reign of King Solomon, it concludes by saying that “the rest of the acts of Solomon, first and last, are written in the chronicles of Nathan the prophet, and in the prophecy of Ahijah the Shilonite, and in the visions of Iddo the seer” (2 Chronicles 9:29). The books of Iddo, Ahijah, and Nathan, as well as the books of Jasher and the Wars of the Lord, aren’t in the canonical Bible. Apparently, they were not excluded on purpose; they just got lost.25

    After the canon was sealed, most Jews gradually forgot the role of human institutions in the messy process of compiling the Bible. Jewish Orthodoxy maintained that God personally handed down to Moses at Mount Sinai the entire first part of the Bible, the Torah. Many rabbis further argued that God created the Torah at the very dawn of time so that even biblical characters who lived before Moses—like Noah and Adam—read and studied it.26 The other parts of the Bible also came to be seen as a divinely created or divinely inspired text, totally different from ordinary human compilations. Once the holy book was sealed, it was hoped that Jews now had direct access to Jehovah’s exact words, which no fallible human or corrupt institution could erase or alter.

    Anticipating the blockchain idea by two thousand years, Jews began making numerous copies of the holy code, and every Jewish community was supposed to have at least one in its synagogue or its bet midrash (house of study).27 This was meant to achieve two things. First, disseminating many copies of the holy book promised to democratize religion and place strict limits on the power of would-be human autocrats. Whereas the archives of Egyptian pharaohs and Assyrian kings empowered the unfathomable kingly bureaucracy at the expense of the masses, the Jewish holy book seemed to give power to the masses, who could now hold even the most brazen leader accountable to God’s laws.

    Second, and more important, having numerous copies of the same book prevented any meddling with the text. If there were thousands of identical copies in numerous locations, any attempt to change even a single letter in the holy code could easily be exposed as a fraud. With numerous Bibles available in far-flung locations, Jews replaced human despotism with divine sovereignty. The social order was now guaranteed by the infallible technology of the book. Or so it seemed.

    THE INSTITUTION STRIKES BACK

    Even before the process of canonizing the Bible was completed, the biblical project had run into further difficulties. Agreeing on the precise contents of the holy book was not the only problem with this supposedly infallible technology. Another obvious problem concerned copying the text. For the holy book to work its magic, Jews needed to have many copies wherever they lived. With Jewish centers emerging not only in Palestine but also in Mesopotamia and Egypt, and with new Jewish communities extending from central Asia to the Atlantic, how to make sure that copyists working thousands of kilometers apart would not change the holy book either on purpose or by mistake?

    To forestall such problems, the rabbis who canonized the Bible devised painstaking regulations for copying the holy book. For example, a scribe was not allowed to pause at certain critical moments in the copying process. When writing the name of God, the scribe “may not respond even if the king greets him. If he was about to write two or three divine names successively, he may pause between them and respond.”28 Rabbi Yishmael (second century CE) told one copyist, “You are doing Heaven’s work, and if you delete one letter or add one letter—you destroy the entire world.”29 In truth, copying errors crept in without destroying the entire world, and no two ancient Bibles were identical.30

    A second and much bigger problem concerned interpretation. Even when people agree on the sanctity of a book and on its exact wording, they can still interpret the same words in different ways. The Bible says that you should not work on the Sabbath. But it doesn’t clarify what counts as “work.” Is it okay to water your field on the Sabbath? What about watering your flowerpot or herd of goats? Is it okay to read a book on the Sabbath? How about writing a book? How about tearing a piece of paper? The rabbis ruled that reading a book isn’t work, but tearing paper is work, which is why nowadays Orthodox Jews prepare a stack of already ripped toilet paper to use on the Sabbath.

    The holy book also says that you should not cook a young goat in its mother’s milk (Exodus 23:19). Some people interpreted this quite literally: if you slaughter a young goat, don’t cook it in the milk of its own mother. But it’s fine to cook it in the milk of an unrelated goat, or in the milk of a cow. Other people interpreted this prohibition much more broadly to mean that meat and dairy products should never be mixed, so you are not allowed to have a milkshake after fried chicken. As unlikely as this may sound, most rabbis ruled that the second interpretation is the correct one, even though chickens don’t lactate.

    More problems resulted from the fact that even if the technology of the book succeeded in limiting changes to the holy words, the world beyond the book continued to spin, and it was unclear how to relate old rules to new situations. Most biblical texts focused on the lives of Jewish shepherds and farmers in the hill country of Palestine and in the sacred city of Jerusalem. But by the second century CE, most Jews lived elsewhere. A particularly large Jewish community grew in the port of Alexandria, one of the richest metropolises of the Roman Empire. A Jewish shipping magnate living in Alexandria would have found that many of the biblical laws were irrelevant to his life while many of his pressing questions had no clear answers in the holy text. He couldn’t obey the commandments about worshipping in the Jerusalem temple, because not only did he not live near Jerusalem, but the temple didn’t even exist anymore. In contrast, when he contemplated whether it was kosher for him to sail his Rome-bound grain ships on the Sabbath, it turned out that long sea voyages were not considered by the authors of Leviticus and Deuteronomy.31

    Inevitably, the holy book spawned numerous interpretations, which were far more consequential than the book itself. As Jews increasingly argued over the interpretation of the Bible, rabbis gained more power and prestige. Writing down the word of Jehovah was supposed to limit the authority of the old priestly institution, but it gave rise to the authority of a new rabbinical institution. Rabbis became the Jewish technocratic elite, developing their rational and rhetorical skills through years of philosophical debates and legal disputations. The attempt to bypass fallible human institutions by relying on a new information technology backfired, because of the need for a human institution to interpret the holy book.

    When the rabbis eventually reached some consensus about how to interpret the Bible, Jews saw another chance to get rid of the fallible human institution. They imagined that if they wrote the agreed interpretation in a new holy book, and made numerous copies of it, that would eliminate the need for any further human intercession between them and the divine code. So after much back-and-forth about which rabbinical opinions should be included and which should be ignored, a new holy book was canonized in the third century CE: the Mishnah.32

    As the Mishnah became more authoritative than the plain text of the Bible, Jews began to believe that the Mishnah could not possibly have been created by humans. It too must have been inspired by Jehovah, or perhaps even composed by the infallible deity in person. Today many Orthodox Jews firmly believe that the Mishnah was handed to Moses by Jehovah on Mount Sinai, passed orally from generation to generation, until it was written down in the third century CE.33

    Alas, no sooner had the Mishnah been canonized and copied than Jews began arguing about the correct interpretation of the Mishnah. And when a consensus was reached about the interpretation of the Mishnah and canonized in the fifth to sixth centuries as a third holy book—the Talmud—Jews began disagreeing about the interpretation of the Talmud.34

    The dream of bypassing fallible human institutions through the technology of the holy book never materialized. With each iteration, the power of the rabbinical institution only increased. “Trust the infallible book” turned into “trust the humans who interpret the book.” Judaism was shaped by the Talmud far more than by the Bible, and rabbinical arguments about the interpretation of the Talmud became even more important than the Talmud itself.35

    This is inevitable, because the world keeps changing. The Mishnah and Talmud dealt with questions raised by second-century Jewish shipping magnates that had no clear answer in the Bible. Modernity too raised many new questions that have no straightforward answers in the Mishnah and Talmud. For example, when electrical appliances developed in the twentieth century, Jews struggled with numerous unprecedented questions such as whether it is okay to press the electrical buttons of an elevator on the Sabbath?

    The Orthodox answer is no. As noted earlier, the Bible forbids working on the Sabbath, and rabbis argued that pressing an electrical button is “work,” because electricity is akin to fire, and it has long been established that kindling a fire is “work.” Does this mean that elderly Jews living in a Brooklyn high-rise must climb a hundred steps to their apartment in order to avoid working on the Sabbath? Well, Orthodox Jews invented a “Sabbath elevator,” which continually goes up and down buildings, stopping on every floor, without you having to perform any “work” by pressing an electrical button.36 The invention of AI gives another twist to this old story. By relying on facial recognition, an AI can quickly direct the elevator to your floor, without making you desecrate the Sabbath.37

    This profusion of texts and interpretations has, over time, caused a profound change in Judaism. Originally, it was a religion of priests and temples, focused on rituals and sacrifices. In biblical times, the quintessential Jewish scene was a priest in blood-splattered robes sacrificing a lamb on the altar of Jehovah. Over the centuries, however, Judaism became an “information religion,” obsessed with texts and interpretations. From second-century Alexandria to twenty-first-century Brooklyn, the quintessential Jewish scene became a group of rabbis arguing about the interpretation of a text.

    This change was extremely surprising given that almost nowhere in the Bible itself do you find anyone arguing about the interpretation of any text. Such debates were not part of biblical culture itself. For example, when Korah and his followers challenged the right of Moses to lead the people of Israel, and demanded a more equitable division of power, Moses reacted not by entering a learned discussion or by quoting some scriptural passage. Rather, Moses called upon God to perform a miracle, and the moment he finished speaking, the ground split, “and the earth opened its mouth and swallowed them and their households” (Numbers 16:31–32). When Elijah was challenged by 450 prophets of Baal and 400 prophets of Asherah to a public test in front of the people of Israel, he proved the superiority of Jehovah over Baal and Asherah first by miraculously summoning fire from the sky and then by slaughtering the pagan prophets. Nobody read any text, and nobody engaged in any rational debate (1 Kings 18).

    As Judaism replaced sacrifices with texts, it gravitated toward a view of information as the most fundamental building block of reality, anticipating current ideas in physics and computer science. The flood of texts generated by rabbis was increasingly seen as more important, and even more real, than plowing a field, baking a loaf of bread, or sacrificing a lamb in a temple. After the temple in Jerusalem was destroyed by the Romans and all temple rituals ceased, rabbis nevertheless devoted enormous efforts to writing texts about the proper way to conduct temple rituals and then arguing about the correct interpretation of these texts. Centuries after the temple was no more, the amount of information concerning these virtual rituals only continued to increase. The rabbis weren’t oblivious to this seeming gap between text and reality. Rather, they maintained that writing texts about the rituals and arguing about these texts were far more important than actually performing the rituals.38
    This eventually led the rabbis to believe that the entire universe was an information sphere—a realm composed of words and running on the alphabetical code of the Hebrew letters. They further maintained that this informational universe was created so that Jews could read texts and argue about their interpretation, and that if Jews ever stop reading these texts and arguing about them, the universe will cease to exist.39 In everyday life, this view meant that for the rabbis words in texts were often more important than facts in the world. Or more accurately, which words appeared in sacred texts became some of the most important facts about the world, shaping the lives of individuals and entire communities.

    THE SPLIT BIBLE

    The above description of the canonization of the Bible, and the creation of the Mishnah and Talmud, ignores one very important fact. The process of canonizing the word of Jehovah created not one chain of texts but several competing chains. There were people who believed in Jehovah, but not in the rabbis. Most of these dissenters did accept the first block in the biblical chain—which they called the Old Testament. But already before the rabbis sealed this block, the dissenters rejected the authority of the entire rabbinical institution, which led them to subsequently reject the Mishnah and Talmud, too. These dissenters were the Christians.

    When Christianity emerged in the first century CE, it was not a unified religion, but rather a variety of Jewish movements that didn’t agree on much, except that they all regarded Jesus Christ—rather than the rabbinical institution—as the ultimate authority on Jehovah’s words.40 Christians accepted the divinity of texts like Genesis, Samuel, and Isaiah, but they argued that the rabbis misunderstood these texts, and only Jesus and his disciples knew the true meaning of passages like “the Lord himself will give you a sign: the almah will conceive and give birth to a son, and will call him Immanuel” (Isaiah 7:14). The rabbis said almah meant “young woman,” Immanuel meant “God with us” (in Hebrew immanu means “with us” and el means “God”), and the entire passage was interpreted as a divine promise to help the Jewish people in their struggle against oppressive foreign empires. In contrast, the Christians argued that almah meant “virgin,” that Immanuel meant that God will literally be born among humans, and that this was a prophecy about the divine Jesus being born on earth to the Virgin Mary.41

    However, by rejecting the rabbinical institution while simultaneously accepting the possibility of new divine revelations, the Christians opened the door to chaos. In the first century CE, and even more so in the second and third centuries CE, different Christians came up with radically new interpretations for books like Genesis and Isaiah, as well as with a plethora of new messages from God. Since they rejected the authority of the rabbis, since Jesus was dead and couldn’t adjudicate between them, and since a unified Christian church didn’t yet exist, who could decide which of all these interpretations and messages was divinely inspired?

    Thus, it was not just John who described the end of the world in his Apocalypse (the book of Revelation). We have many additional apocalypses from that era, for example the Apocalypse of Peter, the Apocalypse of James, and even the Apocalypse of Abraham.42 As for the life and teachings of Jesus, in addition to the four Gospels of Matthew, Mark, Luke, and John, early Christians had the Gospel of Peter, the Gospel of Mary, the Gospel of Truth, the Gospel of the Savior, and numerous others.43 Similarly, aside from the Acts of the Apostles, there were at least a dozen other Acts such as the Acts of Peter and the Acts of Andrew.44 Letters were even more prolific. Most present-day Christian Bibles contain fourteen epistles attributed to Paul, three attributed to John, two to Peter, and one each to James and Jude. Ancient Christians were familiar not only with additional Pauline letters (such as the Epistle to the Laodiceans) but with numerous other epistles supposedly written by other disciples and saints.45

    As Christians composed more and more gospels, epistles, prophecies, parables, prayers, and other texts, it became harder to know which ones to pay attention to. Christians needed a curation institution. That’s how the New Testament was created. At roughly the same time that debates among Jewish rabbis were producing the Mishnah and Talmud, debates between Christian priests, bishops, and theologians were producing the New Testament.
    In a letter from 367 CE, Bishop Athanasius of Alexandria recommended twenty-seven texts that faithful Christians should read—a rather eclectic collection of stories, letters, and prophecies written by different people in different times and places. Athanasius recommended the Apocalypse of John, but not that of Peter or Abraham. He approved of Paul’s Epistle to the Galatians, but not of Paul’s Epistle to the Laodiceans. He endorsed the Gospels of Matthew, Mark, Luke, and John, but rejected the Gospel of Thomas and the Gospel of Truth.46

    A generation later, in the Councils of Hippo (393) and Carthage (397), gatherings of bishops and theologians formally canonized this list of recommendations, which became known as the New Testament.47 When Christians talk about “the Bible,” they mean the Old Testament together with the New Testament. In contrast, Judaism never accepted the New Testament, and when Jews talk about “the Bible,” they mean only the Old Testament, which is supplemented by the Mishnah and Talmud. Interestingly, Hebrew to this day lacks a word to describe the Christian holy book, which contains both the Old Testament and the New Testament. Jewish thought sees them as two utterly unrelated books and simply refuses to acknowledge that there might be a single book encompassing both, even though it is probably the most common book in the world.

    It is crucial to note that the people who created the New Testament weren’t the authors of the twenty-seven texts it contains; they were the curators. Due to the paucity of evidence from the period, we do not know if Athanasius’s list of texts reflected his personal judgment, or whether it originated with earlier Christian thinkers. What we do know is that prior to the Councils of Hippo and Carthage there were rival recommendation lists for Christians. The earliest such list was codified by Marcion of Sinope in the middle of the second century. The Marcion canon included only the Gospel of Luke and ten epistles of Paul. Even these eleven texts were somewhat different from the versions later canonized at Hippo and Carthage. Either Marcion was unaware of other texts like the Gospel of John and the book of Revelation, or he did not think highly of them.48

    The church father Saint John Chrysostom, a contemporary of Bishop Athanasius’s, recommended only twenty-two books, leaving 2 Peter, 2 John, 3 John, Jude, and Revelation out of his list.49 Some Christian churches in the Middle East to this day follow Chrysostom’s shorter list.50 The Armenian Church took about a thousand years to make up its mind about the book of Revelation, while it included in its canon the Third Epistle to the Corinthians, which other churches—like the Catholic and Protestant churches—consider a forgery.51 The Ethiopian Church endorsed Athanasius’s list in full, but added four other books: Sinodos, the book of Clement, the book of the Covenant, and the Didascalia.52 Other lists endorsed the two epistles of Clement, the visions of the Shepherd of Hermas, the Epistle of Barnabas, the Apocalypse of Peter, and various other texts that didn’t make it into Athanasius’s selection.53

    We do not know the precise reasons why specific texts were endorsed or rejected by different churches, church councils, and church fathers. But the consequences were far-reaching. While churches made decisions about texts, the texts themselves shaped the churches. As a key example, consider the role of women in the church. Some early Christian leaders saw women as intellectually and ethically inferior to men, and argued that women should be restricted to subordinate roles in society and in the Christian community. These views were reflected in texts like the First Epistle to Timothy.

    In one of its passages, this text, attributed to Saint Paul, says, “A woman should learn in quietness and full submission. I do not permit a woman to teach or to assume authority over a man; she must be quiet. For Adam was formed first, then Eve. And Adam was not the one deceived; it was the woman who was deceived and became a sinner. But women will be saved through childbearing—if they continue in faith, love and holiness with propriety” (2:11–15). But modern scholars as well as some ancient Christian leaders like Marcion have considered this letter a second-century forgery, ascribed to Saint Paul but actually written by someone else.54

    In opposition to 1 Timothy, during the second, third, and fourth centuries CE there were important Christian texts that saw women as equal to men, and even authorized women to occupy leadership roles, like the Gospel of Mary55 or the Acts of Paul and Thecla. The latter text was written at about the same time as 1 Timothy, and for a time was extremely popular.56 It narrates the adventures of Saint Paul and his female disciple Thecla, describing how Thecla not only performed numerous miracles but also baptized herself with her own hands and often preached. For centuries, Thecla was one of the most revered Christian saints and was seen as evidence that women could baptize, preach, and lead Christian communities.57

    Before the Councils of Hippo and Carthage, it wasn’t clear that 1 Timothy was more authoritative than the Acts of Paul and Thecla. By choosing to include 1 Timothy in their recommendation list while rejecting the Acts of Paul and Thecla, the assembled bishops and theologians shaped Christian attitudes toward women down to the present day. We can only hypothesize what Christianity might have looked like if the New Testament had included the Acts of Paul and Thecla instead of 1 Timothy. Perhaps in addition to church fathers like Athanasius, the church would have had mothers, while misogyny would have been labeled a dangerous heresy perverting Jesus’s message of universal love.

    Just as most Jews forgot that rabbis curated the Old Testament, so most Christians forgot that church councils curated the New Testament, and came to view it simply as the infallible word of God. But while the holy book was seen as the ultimate source of authority, the process of curating the book placed real power in the hands of the curating institution. In Judaism the canonization of the Old Testament and Mishnah went hand in hand with creating the institution of the rabbinate. In Christianity the canonization of the New Testament went hand in hand with the creation of a unified Christian church. Christians trusted church officials—like Bishop Athanasius—because of what they read in the New Testament, but they had faith in the New Testament because this is what the bishops told them to read. The attempt to invest all authority in an infallible superhuman technology led to the rise of a new and extremely powerful human institution—the church.

    THE ECHO CHAMBER

    As time passed, problems of interpretation increasingly tilted the balance of power between the holy book and the church in favor of the institution. Just as the need to interpret Jewish holy books empowered the rabbinate, so the need to interpret Christian holy books empowered the church. The same saying of Jesus or the same Pauline epistle could be understood in various ways, and it was the institution that decided which reading was correct. The institution in turn was repeatedly shaken by struggles over the authority to interpret the holy book, which resulted in institutional schisms such as that between the Western Catholic Church and the Eastern Orthodox Church.

    All Christians read the Sermon on the Mount in the Gospel of Matthew and learned that we should love our enemies, that we should turn the other cheek, and that the meek shall inherit the earth. But what did that actually mean? Christians could read this as a call to reject all use of military force,58 or to reject all social hierarchies.59 The Catholic Church, however, viewed such pacifists and egalitarian readings as heresies. It interpreted Jesus’s words in a way that allowed the church to become the richest landowner in Europe, to launch violent crusades, and to establish murderous inquisitions. Catholic theology accepted that Jesus told us to love our enemies, but explained that burning heretics was an act of love, because it deterred additional people from adopting heretical views, thereby saving them from the flames of hell. The French inquisitor Jacques Fournier wrote in the early fourteenth century an entire treatise on the Sermon on the Mount that explained how the text provided justification for hunting heretics.60 Fournier’s view was not a fringe notion. He went on to become Pope Benedict XII (1334–42).

    Fournier’s task as inquisitor, and later as pope, was to ensure that the Catholic Church’s interpretation of the holy book would prevail. In this, Fournier and his fellow churchmen used not only violent coercion but also their control of book production. Prior to the advent of letterpress printing in Europe in the fifteenth century, making many copies of a book was a prohibitive enterprise for all but the most wealthy individuals and institutions. The Catholic Church used its power and wealth to disseminate copies of its favored texts while prohibiting the production and spread of what it considered erroneous ones.

    Of course, the church couldn’t prevent the occasional freethinker from formulating heretical ideas. But because it controlled key nodes in the medieval information network—such as copying workshops, archives, and libraries—it could prevent such a heretic from making and distributing a hundred copies of her book. To get an idea of the difficulties faced by a heretical author seeking to disseminate her views, consider that when Leofric was made bishop of Exeter in 1050, he found just five books in the cathedral’s library. He immediately established a copying workshop in the cathedral, but in the twenty-two years before he died in 1072, his copyists produced only sixty-six additional volumes.61 In the thirteenth century the library of Oxford University consisted of a few books kept in a chest under St. Mary’s Church. In 1424 the library of Cambridge University boasted a grand total of only 122 books.62 An Oxford University decree from 1409 stipulated that “all recent texts” studied at the university must be unanimously approved “by a panel of twelve theologians appointed by the archbishop.”63

    The church sought to lock society inside an echo chamber, allowing the spread only of those books that supported it, and people trusted the church because almost all the books supported it. Even illiterate laypersons who didn’t read books were still awed by recitations of these precious texts or expositions on their content. That’s how the belief in a supposedly infallible superhuman technology like the New Testament led to the rise of an extremely powerful but fallible human institution like the Catholic Church that crushed all opposing views as “erroneous” while allowing no one to question its own views.

    Catholic information experts such as Jacques Fournier spent their days reading Thomas Aquinas’s interpretation of Augustine’s interpretation of Saint Paul’s epistles and composing additional interpretations of their own. All those interrelated texts didn’t represent reality; they created a new information sphere even bigger and more powerful than that created by the Jewish rabbis. Medieval Europeans were cocooned inside that information sphere, their daily activities, thoughts, and emotions shaped by texts about texts about texts.

    PRINT, SCIENCE, AND WITCHES

    The attempt to bypass human fallibility by investing authority in an infallible text never succeeded. If anyone thought this was due to some unique flaw of the Jewish rabbis or the Catholic priests, the Protestant Reformation repeated the experiment again and again—always getting the same results. Luther, Calvin, and their successors argued that there was no need for any fallible human institution to interpose itself between ordinary people and the holy book. Christians should abandon all the parasitical bureaucracies that grew around the Bible and reconnect to the original word of God. But the word of God never interpreted itself, which is why not only Lutherans and Calvinists but numerous other Protestant sects eventually established their own church institutions and invested them with the authority to interpret the text and persecute heretics.64

    If infallible texts merely lead to the rise of fallible and oppressive churches, how then to deal with the problem of human error? The naive view of information posits that the problem can be solved by creating the opposite of a church—namely, a free market of information. The naive view expects that if all restrictions on the free flow of information are removed, error will inevitably be exposed and displaced by truth. As noted in the prologue, this is wishful thinking. Let’s delve a little deeper to understand why. As a test case, consider what happened during one of the most celebrated epochs in the history of information networks: the European print revolution. The introduction of the printing press to Europe in the mid-fifteenth century made it possible to mass-produce texts relatively quickly, cheaply, and secretly, even if the Catholic Church disapproved of them. It is estimated that in the forty-six years from 1454 to 1500 more than twelve million volumes were printed in Europe. By contrast, in the previous thousand years only about eleven million volumes were hand copied.65 By 1600, all kinds of fringe people—heretics, revolutionaries, proto-scientists—could disseminate their writings much more rapidly, widely, and easily than ever before.

    In the history of information networks, the print revolution of early modern Europe is usually hailed as a moment of triumph, breaking the stranglehold that the Catholic Church had maintained over the European information network. Allegedly, by allowing people to exchange information much more freely than before, it led to the scientific revolution. There is a grain of truth in this. Without print, it would certainly have been much harder for Copernicus, Galileo, and their colleagues to develop and spread their ideas.

    But print wasn’t the root cause of the scientific revolution. The only thing the printing press did was to faithfully reproduce texts. The machine had no ability to come up with any new ideas of its own. Those who connect print to science assume that the mere act of producing and spreading more information inevitably leads people to the truth. In fact, print allowed the rapid spread not only of scientific facts but also of religious fantasies, fake news, and conspiracy theories. Perhaps the most notorious example of the latter was the belief in a worldwide conspiracy of satanic witches, which led to the witch-hunt craze that engulfed early modern Europe.66

    Belief in magic and in witches has characterized human societies in all continents and eras, but different societies imagined witches and reacted to them in very different ways. Some societies believed that witches controlled spirits, talked with the dead, and predicted the future; others imagined that witches stole cattle and located hidden treasure. In one community witches were thought to cause disease, blight cornfields, and concoct love potions, while in another community they supposedly entered houses at night, performed household chores, and stole milk. In some locales witches were thought to be mostly female, while in others they were generally imagined to be male. Some cultures were terrified of witches and persecuted them violently, but others tolerated or even honored them. Finally, there were societies in every continent and era that gave witches little importance.67

    For most of the Middle Ages, most European societies belonged to the latter category and were not overly concerned about witches. The medieval Catholic Church didn’t see them as a major threat to humanity, and some churchmen actively discouraged witch-hunting. According to the influential tenth-century text Canon Episcopi—which defined medieval church doctrine on the matter—witchcraft was mostly illusion, and belief in the reality of witchcraft was an unchristian superstition.68 The European witch-hunt craze was a modern rather than a medieval phenomenon.

    In the 1420s and 1430s churchmen and scholars operating mainly in the Alps region took elements from Christian religion, local folklore, and Greco-Roman heritage and amalgamated them into a new theory of witchcraft.69 Previously, even when witches were dreaded, they were considered a strictly local problem—isolated criminals who, inspired by personal malevolence, used magical means to commit theft and murder. In contrast, the new scholarly model argued that witches were a far more formidable threat to society. There was allegedly a global conspiracy of witches, led by Satan, which constituted an institutionalized anti-Christian religion. Its purpose was nothing less than the complete destruction of the social order and of humankind. Witches were said to gather at night in huge demonic assemblies, where they worshipped Satan, killed children, ate human flesh, engaged in orgies, and cast spells that caused storms, epidemics, and other catastrophes.

    Inspired by such ideas, the first mass witch hunts and witch trials were led by local churchmen and noblemen in the Valais region of the western Alps between 1428 and 1436, leading to the execution of more than two hundred supposed male and female witches. From this Alpine heartland, rumors about the global witch conspiracy trickled to other parts of Europe, but the belief was still far from mainstream, the Catholic establishment did not embrace it, and other regions didn’t launch large-scale witch hunts like those in the Valais.

    In 1485, a Dominican friar and inquisitor called Heinrich Kramer embarked on a witch-hunting expedition in another Alpine region—the Austrian Tyrol. Kramer was a fervent convert to the new belief in a global satanic conspiracy.70 He also seems to have been mentally unhinged, and his accusations of satanic witchcraft were colored by rabid misogyny and odd sexual fixations. Local church authorities, led by the bishop of Brixen, were skeptical of Kramer’s accusations and alarmed by his activities. They stopped his inquisition, released the suspects he arrested, and expelled him from the area.71

    Kramer hit back through the printing press. Within two years of his banishment, he compiled and published the Malleus Maleficarum—The Hammer of the Witches. This was a do-it-yourself guidebook to exposing and killing witches in which Kramer described in detail the worldwide conspiracy and the means by which honest Christians could uncover and foil the witches. In particular, he recommended the use of horrific methods of torture in order to extract confessions from people suspected of witchcraft, and was adamant that the only punishment for the guilty was execution.

    Kramer organized and codified previous ideas and stories and added many details from his own fertile and hate-filled imagination. Relying on ancient Christian misogynist teachings like those of 1 Timothy, Kramer sexualized witchcraft. He argued that witches were typically female, because witchcraft originated in lust, which was supposedly stronger in women. He warned readers that sex could cause a pious woman to become a witch and her husband to become bewitched.72

    An entire chapter of the Hammer is dedicated to the ability of witches to steal men’s penises. Kramer discusses at length whether the witches are really able to take away the male member from its owner, or whether they are only able to create an illusion of castration in men’s minds. Kramer asks, “What is to be thought of those witches who in this way sometimes collect male organs in great numbers, as many as twenty or thirty members together, and put them in a bird’s nest, or shut them up in a box, where they move themselves like living members, and eat oats and corn, as has been seen by many?” He then relates a story he heard from one man: “When he had lost his member, he approached a known witch to ask her to restore it to him. She told the afflicted man to climb a certain tree, and that he might take which he liked out of the nest in which there were several members. And when he tried to take a big one, the witch said: You must not take that one; adding, because it belongs to a parish priest.”73 Numerous notions about witches that are still popular today—for instance, that witches are predominantly women, that witches engage in wild sexual activities, and that witches kill and mutilate children—were given their canonical form by Kramer’s book.

    Like the bishop of Brixen, other churchmen were initially skeptical of Kramer’s wild ideas, and there was some resistance to the book among church experts.74 But The Hammer of the Witches became one of the biggest best sellers of early modern Europe. It catered to people’s deepest fears, as well as to their lurid interest in hearing about orgies, cannibalism, child murders, and satanic conspiracies. The book had gone through eight editions by 1500, another five by 1520, and sixteen more by 1670, with many vernacular translations.75 It became the definitive work on witchcraft and witch-hunting and inspired a host of imitations and elaborations. As Kramer’s fame grew, his work was embraced by the church experts. Kramer was appointed papal representative and made inquisitor of Bohemia and Moravia in 1500. Even today his ideas continue to shape the world, and many current theories about a global satanic conspiracy—like QAnon—draw upon and perpetuate his fantasies.

    While it would be an exaggeration to argue that the invention of print caused the European witch-hunt craze, the printing press played a pivotal role in the rapid dissemination of the belief in a global satanic conspiracy. As Kramer’s ideas gained popularity, printing presses produced not only many additional copies of The Hammer of the Witches and copycat books but also a torrent of cheap one-page pamphlets, whose sensational texts were often accompanied by illustrations depicting people attacked by demons or witches burned at the stake.76 These publications also gave fantastic statistics about the size of the witches’ conspiracy. For example, the Burgundian judge and witch-hunter Henri Boguet (1550–1619) speculated that there were 300,000 witches in France alone and 1.8 million in all of Europe.77 Such claims fueled mass hysteria, which in the sixteenth and seventeenth centuries led to the torture and execution of between 40,000 and 50,000 innocent people who were accused of witchcraft.78 The victims included individuals from all walks of life and ages, including children as young as five.79

    People began denouncing one another for witchcraft on the flimsiest evidence, often to avenge personal slights or to gain economic and political advantage. Once an official investigation began, the accused were often doomed. The inquisitorial methods recommended by The Hammer of the Witches were truly diabolical. If the accused confessed to being a witch, they were executed and their property divided between the accuser, the executioner, and the inquisitors. If the accused refused to confess, this was taken as evidence of their demonic obstinacy, and they were then tortured in horrendous ways, their fingers broken, their flesh cut with hot pincers, their bodies stretched to the breaking point or submerged in boiling water. Sooner or later they could stand it no longer and confessed—and were duly executed.80

    To take one example, in 1600 authorities in Munich arrested on suspicion of witchcraft the Pappenheimer family—father Paulus, mother Anna, two grown sons, and a ten-year-old boy, Hansel. The inquisitors began by torturing little Hansel. The protocol of the interrogation, which can still be read in the Munich archives, has a note from one of the interrogators regarding the ten-year-old boy: “May be tortured to the limit so that he incriminates his mother.”81 After being tortured in unspeakable ways, the Pappenheimers confessed to numerous crimes including killing 265 people by sorcery and causing fourteen destructive storms. They were all condemned to death.

    The bodies of each of the four adult family members were torn with red-hot pincers, the men’s limbs were broken on the wheel, the father was impaled on a stake, the mother’s breasts were cut off, and all were then burned alive. The ten-year-old Hansel was forced to watch all this. Four months later, he too was executed.82 The witch-hunters were extremely thorough in their search for the devil and his accomplices. But if the witch-hunters really wanted to find diabolical evil, they just had to look in the mirror.

    THE SPANISH INQUISITION TO THE RESCUE

    Witch hunts seldom ended by killing just one person or one family. Since the underlying model postulated a global conspiracy, people accused of witchcraft were tortured to name accomplices. This was then used as evidence to imprison, torture, and execute others. If any officials, scholars, or churchmen voiced objections to these absurd methods, this could be seen as proof that they too must be witches—which led to their own arrest and torture.

    For example, in 1453—when belief in the satanic conspiracy was just beginning to take hold—a French doctor of theology called Guillaume Edelin bravely sought to quash it before it spread. He repeated the claims of the medieval Canon Episcopi that witchcraft was an illusion and that witches couldn’t really fly at night to meet Satan and make a pact with him. Edelin was then himself accused of being a witch and arrested. Under torture he confessed that he personally had flown on a broomstick and signed a pact with the devil and that it was Satan who commissioned him to preach that witchcraft was an illusion. His judges were lenient with him; he was spared execution and got life imprisonment instead.83

    The witch hunts illustrate the dark side of creating an information sphere. As with rabbinical discussions of the Talmud and scholastic discussions of Christian scriptures, the witch hunts were fueled by an expanding ocean of information that instead of representing reality created a new reality. Witches were not an objective reality. Nobody in early modern Europe had sex with Satan or was capable of flying on broomsticks and creating hailstorms. But witches became an intersubjective reality. Like money, witches were made real by exchanging information about witches.

    An entire witch-hunting bureaucracy dedicated itself to such exchanges. Theologians, lawyers, inquisitors, and the owners of printing presses made a living by collecting and producing information about witches, cataloging different species of witches, investigating how witches behaved, and recommending how they could be exposed and defeated. Professional witch-hunters offered their services to governments and municipalities, charging large sums of money. Archives were filled by detailed reports of witch-hunting expeditions, protocols of witch trials, and lengthy confessions extracted from the alleged witches.

    Expert witch-hunters used all that data to refine their theories further. Like scholars arguing about the correct interpretation of scripture, the witch-hunters debated the correct interpretation of The Hammer of the Witches and other influential books. The witch-hunting bureaucracy did what bureaucracy often does: it invented the intersubjective category of “witches” and imposed it on reality. It even printed forms, with standard accusations and confessions of witchcraft and blank spaces left for dates, names, and the signature of the accused. All that information produced a lot of order and power; it was a means for certain people to gain authority and for society as a whole to discipline its members. But it produced zero truth and zero wisdom.

    As the witch-hunting bureaucracy generated more and more information, it became harder to dismiss all that information as pure fantasy. Could it be that the entire silo of witch-hunting data did not contain a single grain of truth in it? What about all the books written by learned churchmen? What about all the protocols of trials conducted by esteemed judges? What about the tens of thousands of documented confessions?

    The new intersubjective reality was so convincing that even some people accused of witchcraft came to believe that they were indeed part of a worldwide satanic conspiracy. If everybody said so, it must be true. As discussed in chapter 2, humans are susceptible to adopting fake memories. At least some early modern Europeans dreamed or fantasized about summoning devils, having sex with Satan, and practicing witchcraft, and when accused of being witches, they confused their dreams and fantasies with reality.84

    Consequently, even as the witch hunts reached their ghastly crescendo in the early seventeenth century, and many people suspected that something was clearly wrong, it was difficult to reject the whole thing as pure fantasy. One of the worst witch-hunting episodes in early modern Europe occurred in the towns of Bamberg and Würzburg in southern Germany in the late 1620s. In Bamberg, a city of fewer than 12,000 at the time,85 up to 900 innocent people were executed from 1625 to 1631.86 In Würzburg another 1,200 people were tortured and killed, out of a population of around 11,500.87 In August 1629, the chancellor of the prince-bishop of Würzburg wrote a letter to a friend about the ongoing witch hunt, in which he confessed his doubts about the matter. The letter is worth quoting at length:

    As to the affair of the witches … it has started up afresh, and no words can do justice to it. Ah, the woe and the misery of it—there are still four hundred in the city, high and low, of every rank and sex, nay, even clerics, so strongly accused that they may be arrested at any hour.… The Prince-Bishop has over forty students who are soon to be pastors; among them thirteen or fourteen are said to be witches. A few days ago a Dean was arrested; two others who were summoned have fled. The notary of our Church consistory, a very learned man, was yesterday arrested and put to the torture. In a word, a third part of the city is surely involved. The richest, most attractive, most prominent, of the clergy are already executed. A week ago a maiden of nineteen was executed, of whom it is everywhere said that she was the fairest in the whole city, and was held by everybody a girl of singular modesty and purity. She will be followed by seven or eight others of the best and most attractive persons.… And thus many are put to death for renouncing God and being at the witch-dances, against whom nobody has ever else spoken a word.

    To conclude this wretched matter, there are children of three and four years, to the number of three hundred, who are said to have had intercourse with the Devil. I have seen put to death children of seven, promising students of ten, twelve, fourteen, and fifteen.… [B]ut I cannot and must not write more of this misery.

    The chancellor then added this interesting postscript to the letter:

    Though there are many wonderful and terrible things happening, it is beyond doubt that, at a place called the Fraw-Rengberg, the Devil in person, with eight thousand of his followers, held an assembly and celebrated mass before them all, administering to his audience (that is, the witches) turnip-rinds and parings in place of the Holy Eucharist. There took place not only foul but most horrible and hideous blasphemies, whereof I shudder to write.88

    Even after expressing his horror at the insanity of the witch hunt in Würzburg, the chancellor nevertheless expressed his firm belief in the satanic conspiracy of witches. He didn’t witness any witchcraft firsthand, but so much information about witches was circulating that it was difficult for him to doubt all of it. Witch hunts were a catastrophe caused by the spread of toxic information. They are a prime example of a problem that was created by information, and was made worse by more information.

    This was a conclusion reached not just by modern scholars but also by some perceptive observers at the time. Alonso de Salazar Frías, a Spanish inquisitor, made a thorough investigation of witch hunts and witch trials in the early seventeenth century. He concluded that “I have not found one single proof nor even the slightest indication from which to infer that one act of witchcraft has actually taken place,” and that “there were neither witches nor bewitched until they were talked and written about.”89 Salazar Frías well understood the meaning of intersubjective realities and correctly identified the entire witch-hunting industry as an intersubjective information sphere.

    The history of the early modern European witch craze demonstrates that releasing barriers to the flow of information doesn’t necessarily lead to the discovery and spread of truth. It can just as easily lead to the spread of lies and fantasies and to the creation of toxic information spheres. More specifically, a completely free market of ideas may incentivize the dissemination of outrage and sensationalism at the expense of truth. It is not difficult to understand why. Printers and booksellers made a lot more money from the lurid tales of The Hammer of the Witches than they did from the dull mathematics of Copernicus’s On the Revolutions of the Heavenly Spheres. The latter was one of the founding texts of the modern scientific tradition. It is credited with earth-shattering discoveries that displaced our planet from the center of the universe and thereby initiated the Copernican revolution. But when it was first published in 1543, its initial print run of four hundred failed to sell out, and it took until 1566 for a second edition to be published in a similar-sized print run. The third edition did not appear until 1617. As Arthur Koestler quipped, it was an all-time worst seller.90 What really got the scientific revolution going was neither the printing press nor a completely free market of information, but rather a novel approach to the problem of human fallibility.

    THE DISCOVERY OF IGNORANCE

    The history of print and witch-hunting indicates that an unregulated information market doesn’t necessarily lead people to identify and correct their errors, because it may well prioritize outrage over truth. For truth to win, it is necessary to establish curation institutions that have the power to tilt the balance in favor of the facts. However, as the history of the Catholic Church indicates, such institutions might use their curation power to quash any criticism of themselves, labeling all alternative views erroneous and preventing the institution’s own errors from being exposed and corrected. Is it possible to establish better curation institutions that use their power to further the pursuit of truth rather than to accumulate more power for themselves?

    Early modern Europe saw the foundation of exactly such curation institutions, and it was these institutions—rather than the printing press or specific books like On the Revolutions of the Heavenly Spheres—that constituted the bedrock of the scientific revolution. These key curation institutions were not the universities. Many of the most important leaders of the scientific revolution were not university professors. Nicolaus Copernicus, Robert Boyle, Tycho Brahe, and René Descartes, for example, held no academic positions. Nor did Spinoza, Leibniz, Locke, Berkeley, Voltaire, Diderot, or Rousseau.

    The curation institutions that played a central role in the scientific revolution connected scholars and researchers both in and out of universities, forging an information network that spanned the whole of Europe and eventually the world. For the scientific revolution to gather pace, scientists had to trust information published by colleagues in distant lands. This kind of trust in the work of people whom one had never met was evident in scientific associations like the Royal Society of London for Improving Natural Knowledge, founded in 1660, and the French Académie des Sciences (1666); scientific journals like the Philosophical Transactions of the Royal Society (1665) and the Histoire de l’Académie Royale des Sciences (1699); and scientific publishers like the architects of the Encyclopédie (1751–72). These institutions curated information on the basis of empirical evidence, bringing attention to the discoveries of Copernicus rather than to the fantasies of Kramer. When a paper was submitted to the Philosophical Transactions of the Royal Society, the lead question the editors asked was not, “How many people would pay to read this?” but, “What proof is there that it is true?”

    At first, these new institutions seemed as flimsy as cobwebs, lacking the power necessary to reshape human society. Unlike the witch-hunting experts, the editors of the Philosophical Transactions of the Royal Society could not torture and execute anyone. And unlike the Catholic Church, the Académie des Sciences did not command huge territories and budgets. But scientific institutions did accrue influence thanks to a very original claim to trust. A church typically told people to trust it because it possessed the absolute truth, in the form of an infallible holy book. A scientific institution, in contrast, gained authority because it had strong self-correcting mechanisms that exposed and rectified the errors of the institution itself. It was these self-correcting mechanisms, not the technology of printing, that were the engine of the scientific revolution.

    In other words, the scientific revolution was launched by the discovery of ignorance.91 Religions of the book assumed that they had access to an infallible source of knowledge. The Christians had the Bible, the Muslims had the Quran, the Hindus had the Vedas, and the Buddhists had the Tipitaka. Scientific culture has no comparable holy book, nor does it claim that any of its heroes are infallible prophets, saints, or geniuses. The scientific project starts by rejecting the fantasy of infallibility and proceeding to construct an information network that takes error to be inescapable. Sure, there is much talk about the genius of Copernicus, Darwin, and Einstein, but none of them is considered faultless. They all made mistakes, and even the most celebrated scientific tracts are sure to contain errors and lacunae.

    Since even geniuses suffer from confirmation bias, you cannot trust them to correct their own errors. Science is a team effort, relying on institutional collaboration rather than on individual scientists or, say, a single infallible book. Of course, institutions too are prone to error. Scientific institutions are nevertheless different from religious institutions, inasmuch as they reward skepticism and innovation rather than conformity. Scientific institutions are also different from conspiracy theories, inasmuch as they reward self-skepticism. Conspiracy theorists tend to be extremely skeptical regarding the existing consensus, but when it comes to their own beliefs, they lose all their skepticism and fall prey to confirmation bias.92 The trademark of science is not merely skepticism but self-skepticism, and at the heart of every scientific institution we find a strong self-correcting mechanism. Scientific institutions do reach a broad consensus about the accuracy of certain theories—such as quantum mechanics or the theory of evolution—but only because these theories have managed to survive intense efforts to disprove them, launched not only by outsiders but by members of the institution itself.

    SELF-CORRECTING MECHANISMS

    As an information technology, the self-correcting mechanism is the polar opposite of the holy book. The holy book is supposed to be infallible. The self-correcting mechanism embraces fallibility. By self-correcting, I refer to mechanisms that an entity uses to correct itself. A teacher correcting a student’s essay is not a self-correcting mechanism; the student isn’t correcting their own essay. A judge sending a criminal to prison is not a self-correcting mechanism; the criminal isn’t exposing their own crime. When the Allies defeated and dismantled the Nazi regime, this was not a self-correcting mechanism; left to its own devices, Germany would not have denazified itself. But when a scientific journal publishes a paper correcting a mistake that appeared in a previous paper, that’s an example of an institution self-correcting its own errors.

    Self-correcting mechanisms are ubiquitous in nature. Children learn how to walk thanks to them. You make a wrong move, you fall, you learn from your mistake, you try doing it a little differently. Sure, sometimes parents and teachers give the child a hand or offer advice, but a child who relies entirely on such external corrections or keeps excusing mistakes instead of learning from them will find it very difficult to walk. Indeed, even as adults, every time we walk, our body engages in an intricate process of self-correction. As our body navigates through space, internal feedback loops between brain, limbs, and sensory organs keep our legs and hands in their proper place and our balance just right.93

    Many other bodily processes require constant self-correction. Our blood pressure, temperature, sugar levels, and numerous other parameters must be given some leeway to change in accordance with varying circumstances, but they should never go above or below certain critical thresholds. Our blood pressure needs to increase when we run, to decrease when we sleep, but must always keep within certain bounds.94 Our body manages this delicate biochemical dance through a host of homeostatic self-correcting mechanisms. If our blood pressure goes too high, the self-correcting mechanisms lower it. If our blood pressure is dangerously low, the self-correcting mechanisms raise it. If the self-correcting mechanisms go out of order, we could die.95

    Institutions, too, die without self-correcting mechanisms. These mechanisms start with the realization that humans are fallible and corruptible. But instead of despairing of humans and looking for a way to bypass them, the institution actively seeks its own errors and corrects them. All institutions that manage to endure beyond a handful of years possess such mechanisms, but institutions differ greatly in the strength and visibility of their self-correcting mechanisms.

    For example, the Catholic Church is an institution with relatively weak self-correcting mechanisms. Since it claims infallibility, it cannot admit institutional mistakes. It is occasionally willing to acknowledge that some of its members have erred or sinned, but the institution itself allegedly remains perfect. For example, in the Second Vatican Council in 1964, the Catholic Church acknowledged that “Christ summons the Church to continual reformation as she sojourns here on earth. The Church is always in need of this, insofar as she is an institution of men here on earth. Thus if, in various times and circumstances, there have been deficiencies in moral conduct or in church discipline, or even in the way that church teaching has been formulated—to be carefully distinguished from the deposit of faith itself—these can and should be set right at the opportune moment.”96

    This admission sounds promising, but the devil is in the details, specifically in the refusal to countenance the possibility of any deficiency in “the deposit of faith.” In Catholic dogma “the deposit of faith” refers to the body of revealed truth that the church has received from scriptures and from its sacred tradition of interpreting scripture. The Catholic Church acknowledges that priests are fallible humans who can sin and can also make mistakes in the way they formulate church teachings. However, the holy book itself can never err. What does this imply about the entire church as an institution that combines fallible humans with an infallible text?

    According to Catholic dogma, biblical infallibility and divine guidance trump human corruption, so even though individual members of the church may err and sin, the Catholic Church as an institution is never wrong. Allegedly, never in history did God allow the majority of church leaders to make a serious mistake in their interpretation of the holy book. This principle is common to many religions. Jewish Orthodoxy accepted the possibility that the rabbis who composed the Mishnah and Talmud might have erred in personal matters, but when they came to decree religious doctrine, God ensured that they would make no mistake.97 In Islam there is an analogous principle known as Ijma. According to one important Hadith, Muhammad said that “Allah will ensure my community will never agree on error.”98

    In Catholicism, alleged institutional perfection is enshrined most clearly in the doctrine of papal infallibility, which says that while in personal matters popes may err, in their institutional role they are infallible.99 For example, Pope Alexander VI erred in breaking his vow of celibacy, having a mistress and siring several children, yet when defining official church teachings on matters of ethics or theology, he was incapable of mistake.

    In line with these views, the Catholic Church has always employed a self-correcting mechanism to supervise its human members in their personal affairs, but it never developed a mechanism for amending the Bible or for amending its “deposit of faith.” This attitude is manifest in the few formal apologies the Catholic Church issued for its past conduct. In recent decades, several popes apologized for the mistreatment of Jews, women, non-Catholic Christians, and indigenous cultures, as well as for more specific events such as the sacking of Constantinople in 1204 and the abuse of children in Catholic schools. It is commendable that the Catholic Church made such apologies at all; religious institutions rarely do so. Nevertheless, in all these cases, the popes were careful to shift responsibility away from scriptures and from the church as an institution. Instead, the blame was laid on the shoulders of individual churchmen who misinterpreted scriptures and deviated from the true teachings of the church.

    For example, in March 2000, Pope John Paul II conducted a special ceremony in which he asked forgiveness for a long list of historical crimes against Jews, heretics, women, and indigenous people. He apologized “for the use of violence that some have committed in the service of truth.” This terminology implied that the violence was the fault of “some” misguided individuals who didn’t understand the truth taught by the church. The pope didn’t accept the possibility that perhaps these individuals understood exactly what the church was teaching and that these teachings just were not the truth.100

    Similarly, when Pope Francis apologized in 2022 for the abuses against indigenous people in Canada’s church-run residential schools, he said, “I ask for forgiveness, in particular, for the ways in which many members of the church … cooperated … in projects of cultural destruction and forced assimilation.”101 Note his careful shifting of responsibility. The fault lay with “many members of the church,” not with the church and its teachings. As if it were never official church doctrine to destroy indigenous cultures and forcefully convert people.

    In fact, it wasn’t a few wayward priests who launched the Crusades, imposed laws that discriminated against Jews and women, or orchestrated the systematic annihilation of indigenous religions throughout the world.102 The writings of many revered church fathers, and the official decrees of many popes and church councils, are full of passages disparaging “pagan” and “heretical” religions, calling for their destruction, discriminating against their members, and legitimizing the use of violence to convert people to Christianity.103 For example, in 1452 Pope Nicholas V issued the Dum Diversas bull, addressed to King Afonso V of Portugal and other Catholic monarchs. The bull said, “We grant you by these present documents, with our Apostolic Authority, full and free permission to invade, search out, capture, and subjugate the Saracens and pagans and any other unbelievers and enemies of Christ wherever they may be, as well as their kingdoms, duchies, counties, principalities, and other property … and to reduce their persons into perpetual servitude.”104 This official proclamation, repeated numerous times by subsequent popes, laid the theological basis for European imperialism and the destruction of native cultures across the world. Of course, though the church doesn’t acknowledge it officially, over time it has changed its institutional structures, its core teachings, and its interpretation of scripture. The Catholic Church of today is far less antisemitic and misogynist than it was in medieval and early modern times. Pope Francis is far more tolerant of indigenous cultures than Pope Nicholas V. There is an institutional self-correcting mechanism at work here, which reacts both to external pressures and to internal soul-searching. But what characterizes self-correcting in institutions like the Catholic Church is that even when it happens, it is denied rather than celebrated. The first rule of changing church teachings is that you never admit changing church teachings.

    You would never hear a pope announcing to the world, “Our experts have just discovered a really big error in the Bible. We’ll soon issue an updated edition.” Instead, when asked about the church’s more generous attitude to Jews or women, popes imply that this was always what the church really taught, even if some individual churchmen previously failed to understand the message correctly. Denying the existence of self-correction doesn’t entirely stop it from happening, but it does weaken and slow it. Because the correction of past mistakes is not acknowledged, let alone celebrated, when the faithful encounter another serious problem in the institution and its teachings, they are paralyzed by fear of changing something that is supposedly eternal and infallible. They cannot benefit from the example of previous changes.

    For instance, when Catholics like Pope Francis himself are now reconsidering the church’s teachings on homosexuality,105 they find it difficult to simply acknowledge past mistakes and change the teachings. If eventually a future pope would issue an apology for the mistreatment of LGBTQ people, the way to do it would be to again shift the blame to the shoulders of some overzealous individuals who misunderstood the gospel. To maintain its religious authority the Catholic Church has had no choice but to deny the existence of institutional self-correction. For the church fell into the infallibility trap. Once it based its religious authority on a claim to infallibility, any public admission of institutional error—even on relatively minor issues—could completely destroy its authority.

    THE DSM AND THE BIBLE

    In contrast to the Catholic Church, the scientific institutions that emerged in early modern Europe have been built around strong self-correcting mechanisms. Scientific institutions maintain that even if most scientists in a particular period believe something to be true, it may yet turn out to be inaccurate or incomplete. In the nineteenth century most physicists accepted Newtonian physics as a comprehensive account of the universe, but in the twentieth century the theory of relativity and quantum mechanics exposed the inaccuracies and limitations of Newton’s model.106 The most celebrated moments in the history of science are precisely those moments when accepted wisdom is overturned and new theories are born.

    Crucially, scientific institutions are willing to admit their institutional responsibility for major mistakes and crimes. For example, present-day universities routinely give courses, and professional journals routinely publish articles, that expose the institutional racism and sexism that characterized the scientific study of subjects like biology, anthropology, and history in the nineteenth and much of the twentieth centuries. Research on individual test cases such as the Tuskegee Syphilis Study, and on governmental policies ranging from the White Australia policy to the Holocaust, have repeatedly and extensively studied how flawed biological, anthropological, and historical theories developed in leading scientific institutions were used to justify and facilitate discrimination, imperialism, and even genocide. These crimes and errors are not blamed on a few misguided scholars. They are seen as an institutional failure of entire academic disciplines.107

    The willingness to admit major institutional errors contributes to the relatively fast pace at which science is developing. When the available evidence justifies it, dominant theories are often discarded within a few generations, to be replaced by new theories. What students of biology, anthropology, and history learn at university in the early twenty-first century is very different from what they learned there a century previously.

    Psychiatry offers numerous similar examples for strong self-correcting mechanisms. On the shelf of most psychiatrists you can find the DSM—the Diagnostic and Statistical Manual of Mental Disorders. It is occasionally nicknamed the psychiatrists’ bible. But there is a crucial difference between the DSM and the Bible. First published in 1952, the DSM is revised every decade or two, with the fifth edition appearing in 2013. Over the years, the definition of many disorders has changed, new ones have been added, while others have been deleted. Homosexuality, for example, was listed in 1952 as a sociopathic personality disturbance, but removed from the DSM in 1974. It took just twenty-two years to correct this error in the DSM. That’s not a holy book. That’s a scientific text.

    Today the discipline of psychiatry doesn’t try to reinterpret the 1952 definition of homosexuality in a more benign spirit. Rather, it views the 1952 definition as a downright error. More important, the error is not attributed to the shortcomings of a few homophobic professors. Rather, it is acknowledged to be the result of deep institutional biases in the discipline of psychiatry.108 Confessing the past institutional errors of their discipline makes psychiatrists today more careful not to commit new such errors, as evidenced in the heated debate regarding transgender people and people on the autistic spectrum. Of course, no matter how careful they are, psychiatrists are still likely to make institutional mistakes. But they are also likely to acknowledge and correct them.109

    PUBLISH OR PERISH

    What makes scientific self-correcting mechanisms particularly strong is that scientific institutions are not just willing to admit institutional error and ignorance; they are actively seeking to expose them. This is evident in the institutions’ incentive structure. In religious institutions, members are incentivized to conform to existing doctrine and be suspicious of novelty. You become a rabbi, imam, or priest by professing doctrinal loyalty, and you can advance up the ranks to become pope, chief rabbi, or grand ayatollah without criticizing your predecessors or advancing any radical new notions. Indeed, many of the most powerful and admired religious leaders of recent times—such as Pope Benedict XVI, Chief Rabbi of Israel David Lau, and Ayatollah Khamenei of Iran—have won fame and supporters by strict resistance to new ideas and trends like feminism.110

    In science it works the other way around. Hiring and promotions in scientific institutions are based on the principle of “publish or perish,” and to publish in prestigious journals, you must expose some mistake in existing theories or discover something your predecessors and teachers didn’t know. Nobody wins a Nobel Prize for faithfully repeating what previous scholars said and opposing every new scientific theory.

    Of course, just as religion has room for self-correcting, so science has ample room for conformism, too. Science is an institutional enterprise, and scientists rely on the institution for almost everything they know. For example, how do I know what medieval and early modern Europeans thought about witchcraft? I have not visited all the relevant archives myself, nor have I read all the relevant primary sources. In fact, I am incapable of reading many of these sources directly, because I do not know all the necessary languages, nor am I skilled in deciphering medieval and early modern handwriting. Instead, I have relied on books and articles published by other scholars, such as Ronald Hutton’s book The Witch: A History of Fear, which was published by Yale University Press in 2017.

    I haven’t met Ronald Hutton, who is a professor of history at the University of Bristol, nor do I personally know the Bristol officials who hired him or the Yale editorial team who published his book. I nevertheless trust what I read in Hutton’s book, because I understand how institutions like the University of Bristol and Yale University Press operate. Their self-correcting mechanisms have two crucial features: First, the self-correcting mechanisms are built into the core of the institutions rather than being a peripheral add-on. Second, these institutions publicly celebrate self-correcting instead of denying it. It is of course possible that some of the information I gained from Hutton’s book may be incorrect, or I myself may misinterpret it. But experts on the history of witchcraft who have read Hutton’s book and who might be reading the present book will hopefully spot any such errors and expose them.

    Populist critics of scientific institutions may counter that, in fact, these institutions use their power to stifle unorthodox views and launch their own witch hunts against dissenters. It is certainly true that if a scholar opposes the current orthodox view of their discipline, it might sometimes have negative consequences: articles rejected, research grants denied, nasty ad hominem attacks, and in rare cases even getting fired from their job.111 I do not wish to belittle the suffering such things cause, but it is still a far cry from being physically tortured and burned at the stake.

    Consider, for example, the story of the chemist Dan Shechtman. In April 1982, while observing through an electron microscope, Shechtman saw something that all contemporary theories in chemistry claimed simply could not exist: the atoms in a mixed sample of aluminum and manganese were crystallized in a pattern with a five-fold rotational symmetry. At the time, scientists knew of various possible symmetrical structures in solid crystals, but five-fold symmetry was considered against the very laws of nature. Shechtman’s discovery of what came to be called quasicrystals sounded so outlandish that it was difficult to find a peer-reviewed journal willing to publish it. It didn’t help that Shechtman was at the time a junior scientist. He didn’t even have his own laboratory; he was working in someone’s else facility. But the editors of the journal Physical Review Letters, after reviewing the evidence, eventually published Shechtman’s article in 1984.112 And then, as he describes it, “all hell broke loose.”

    Shechtman’s claims were dismissed by most of his colleagues, and he was blamed for mismanaging his experiments. The head of his laboratory also turned on Shechtman. In a dramatic gesture, he placed a chemistry textbook on Shechtman’s desk and told him, “Danny, please read this book and you will understand that what you are saying cannot be.” Shechtman boldly replied that he saw the quasicrystals in the microscope—not in the book. As a result, he was kicked out of the lab. Worse was to come. Linus Pauling, a two-time Nobel laureate and one of the most eminent scientists of the twentieth century, led a brutal personal attack on Shechtman. In a conference attended by hundreds of scientists, Pauling proclaimed, “Danny Shechtman is talking nonsense, there are no quasicrystals, just quasi-scientists.”

    But Shechtman was not imprisoned or killed. He got a place in another lab. The evidence he presented turned out to be more convincing than the existing chemistry textbooks and the views of Linus Pauling. Several colleagues repeated Shechtman’s experiments and replicated his findings. A mere ten years after Shechtman saw the quasicrystals through his microscope, the International Union of Crystallography—the leading scientific association in the field—altered its definition of what a crystal is. Chemistry textbooks were changed accordingly, and an entire new scientific field emerged—the study of quasicrystals. In 2011, Shechtman was awarded the Nobel Prize in Chemistry for his discovery.113 The Nobel Committee said that “his discovery was extremely controversial [but] eventually forced scientists to reconsider their conception of the very nature of matter.”114

    Shechtman’s story is hardly exceptional. The annals of science are full of similar cases. Before the theory of relativity and quantum mechanics became the cornerstones of twentieth-century physics, they initially provoked bitter controversies, including personal assaults by the old guard on the proponents of the new theories. Similarly, when Georg Cantor developed in the late nineteenth century his theory of infinite numbers, which became the basis for much of twentieth-century mathematics, he was personally attacked by some of the leading mathematicians of his day like Henri Poincaré and Leopold Kronecker. Populists are right to think that scientists suffer from the same human biases as everyone else. However, thanks to institutional self-correcting mechanisms these biases can be overcome. If enough empirical evidence is provided, it often takes just a few decades for an unorthodox theory to upend established wisdom and become the new consensus.

    As we shall see in the next chapter, there were times and places where scientific self-correcting mechanisms ceased functioning and academic dissent could lead to physical torture, imprisonment, and death. In the Soviet Union, for example, questioning official dogma on any matter—economics, genetics, or history—could lead not only to dismissal but even to a couple of years in the gulag or an executioner’s bullet.115 A famous case involved the bogus theories of the agronomist Trofim Lysenko. He rejected mainstream genetics and the theory of evolution by natural selection and advanced his own pet theory, which said that “re-education” could change the traits of plants and animals, and even transform one species into another. Lysenkoism greatly appealed to Stalin, who had ideological and political reasons for believing in the almost limitless potential of “re-education.” Thousands of scientists who opposed Lysenko and continued to uphold the theory of evolution by natural selection were dismissed from their jobs, and some were imprisoned or executed. Nikolai Vavilov, a botanist and geneticist who was Lysenko’s former mentor turned critic, was tried in July 1941 along with the botanist Leonid Govorov, the geneticist Georgii Karpechenko, and the agronomist Aleksandr Bondarenko. The latter three were shot, while Vavilov died in a camp in Saratov in 1943.116 Under pressure from the dictator, the Lenin All-Union Academy of Agricultural Sciences eventually announced in August 1948 that henceforth Soviet institutions would teach Lysenkoism as the only correct theory.117

    But for precisely this reason, the Lenin All-Union Academy of Agricultural Sciences ceased being a scientific institution, and Soviet dogma on genetics was an ideology rather than a science. An institution can call itself by whatever name it wants, but if it lacks a strong self-correcting mechanism, it is not a scientific institution.

    THE LIMITS OF SELF-CORRECTION

    Does all this mean that in self-correcting mechanisms we have found the magic bullet that protects human information networks from error and bias? Unfortunately, things are far more complicated. There is a reason why institutions like the Catholic Church and the Soviet Communist Party eschewed strong self-correcting mechanisms. While such mechanisms are vital for the pursuit of truth, they are costly in terms of maintaining order. Strong self-correcting mechanisms tend to create doubts, disagreements, conflicts, and rifts and to undermine the myths that hold the social order together.

    Of course, order by itself isn’t necessarily good. For example, the social order of early modern Europe endorsed, among other things, not only witch hunts but also the exploitation of millions of peasants by a handful of aristocrats, the systematic mistreatment of women, and widespread discrimination against Jews, Muslims, and other minorities. But even when the social order is highly oppressive, undermining it doesn’t necessarily lead to a better place. It could just lead to chaos and worse oppression. The history of information networks has always involved maintaining a balance between truth and order. Just as sacrificing truth for the sake of order comes with a cost, so does sacrificing order for truth.

    Scientific institutions have been able to afford their strong self-correcting mechanisms because they leave the difficult job of preserving the social order to other institutions. If a thief breaks into a chemistry lab, or a psychiatrist receives death threats, they don’t complain to a peer-reviewed journal; they call the police. Is it possible, then, to maintain strong self-correcting mechanisms in institutions other than academic disciplines? In particular, can such mechanisms exist in institutions like police forces, armies, political parties, and governments that are charged with maintaining the social order?

    We’ll explore this question in the next chapter, which focuses on the political aspects of information flows and examines the long-term history of democracies and dictatorships. As we shall see, democracies believe that it is possible to maintain strong self-correcting mechanisms even in politics. Dictatorships disavow such mechanisms. Thus, at the height of the Cold War, newspapers and universities in the democratic United States openly exposed and criticized American war crimes in Vietnam. Newspapers and universities in the totalitarian Soviet Union were also happy to criticize American crimes, but they remained silent about Soviet crimes in Afghanistan and elsewhere. Soviet silence was scientifically unjustifiable, but it made political sense. American self-flagellation about the Vietnam War continues even today to divide the American public and to undermine America’s reputation throughout the world, whereas Soviet and Russian silence about the Afghanistan War has helped dim its memory and limit its reputational costs.

    Only after understanding the politics of information in historical systems like ancient Athens, the Roman Empire, the United States, and the Soviet Union will we be ready to explore the revolutionary implications of the rise of AI. For one of the biggest questions about AI is whether it will favor or undermine democratic self-correcting mechanisms.

    CHAPTER 5 Decisions: A Brief History of Democracy and Totalitarianism

    Democracy and dictatorship are typically discussed as contrasting political and ethical systems. This chapter seeks to shift the terms of the discussion, by surveying the history of democracy and dictatorship as contrasting types of information networks. It examines how information in democracies flows differently than in dictatorial systems and how inventing new information technologies helps different kinds of regimes flourish.

    Dictatorial information networks are highly centralized.1 This means two things. First, the center enjoys unlimited authority, hence information tends to flow to the central hub, where the most important decisions are made. In the Roman Empire all roads led to Rome, in Nazi Germany information flowed to Berlin, and in the Soviet Union it streamed to Moscow. Sometimes the central government attempts to concentrate all information in its hands and to dictate all decisions by itself, controlling the totality of people’s lives. This totalizing form of dictatorship, practiced by the likes of Hitler and Stalin, is known as totalitarianism. As we shall see, technical difficulties often prevent dictators from becoming totalitarian. The Roman emperor Nero, for example, didn’t have the technology necessary to micromanage the lives of millions of peasants in remote provincial villages. In many dictatorial regimes considerable autonomy is therefore left to individuals, corporations, and communities. However, the dictators always retain the authority to intervene in people’s lives. In Nero’s Rome freedom was not an ideal but a by-product of the government’s inability to exert totalitarian control.

    The second characteristic of dictatorial networks is that they assume the center is infallible. They therefore dislike any challenge to the center’s decisions. Soviet propaganda depicted Stalin as an infallible genius, and Roman propaganda treated emperors as divine beings. Even when Stalin or Nero made a patently disastrous decision, there were no robust self-correcting mechanisms in the Soviet Union or the Roman Empire that could expose the mistake and push for a better course of action.

    In theory, a highly centralized information network could try to maintain strong self-correcting mechanisms, like independent courts and elected legislative bodies. But if they functioned well, these would challenge the central authority and thereby decentralize the information network. Dictators always see such independent power hubs as threats and seek to neutralize them. This is what happened to the Roman Senate, whose power was whittled away by successive Caesars until it became little more than a rubber stamp for imperial whims.2 The same fate befell the Soviet judicial system, which never dared resist the will of the Communist Party. Stalinist show trials, as their name indicates, were theater with preordained results.3

    To summarize, a dictatorship is a centralized information network, lacking strong self-correcting mechanisms. A democracy, in contrast, is a distributed information network, possessing strong self-correcting mechanisms. When we look at a democratic information network, we do see a central hub. The government is the most important executive power in a democracy, and government agencies therefore gather and store vast quantities of information. But there are many additional information channels that connect lots of independent nodes. Legislative bodies, political parties, courts, the press, corporations, local communities, NGOs, and individual citizens communicate freely and directly with one another so that most information never passes through any government agency and many important decisions are made elsewhere. Individuals choose for themselves where to live, where to work, and whom to marry. Corporations make their own choices about where to open a branch, how much to invest in certain projects, and how much to charge for goods and services. Communities decide for themselves about organizing charities, sporting events, and religious festivals. Autonomy is not a consequence of the government’s ineffectiveness; it is the democratic ideal.

    Even if it possesses the technology necessary to micromanage people’s lives, a democratic government leaves as much room as possible for people to make their own choices. A common misconception is that in a democracy everything is decided by majority vote. In fact, in a democracy as little as possible is decided centrally, and only the relatively few decisions that must be made centrally should reflect the will of the majority. In a democracy, if 99 percent of people want to dress in a particular way and worship a particular god, the remaining 1 percent should still be free to dress and worship differently.

    Of course, if the central government doesn’t intervene at all in people’s lives, and doesn’t provide them with basic services like security, it isn’t a democracy; it is anarchy. In all democracies the center raises taxes and maintains an army, and in most modern democracies it also provides at least some level of health care, education, and welfare. But any intervention in people’s lives demands an explanation. In the absence of a compelling reason, a democratic government should leave people to their own devices.

    Another crucial characteristic of democracies is that they assume everyone is fallible. Therefore, while democracies give the center the authority to make some vital decisions, they also maintain strong mechanisms that can challenge the central authority. To paraphrase President James Madison, since humans are fallible, a government is necessary, but since government too is fallible, it needs mechanisms to expose and correct its errors, such as holding regular elections, protecting the freedom of the press, and separating the executive, legislative, and judicial branches of government.

    Consequently, while a dictatorship is about one central information hub dictating everything, a democracy is an ongoing conversation between diverse information nodes. The nodes often influence each other, but in most matters they are not obliged to reach a consensus. Individuals, corporations, and communities can continue to think and behave in different ways. There are, of course, cases when everyone must behave the same, and diversity cannot be tolerated. For example, when in 2002–3 Americans disagreed about whether to invade Iraq, everyone ultimately had to abide by a single decision. It was unacceptable that some Americans would maintain a private peace with Saddam Hussein while others declared war. Whether good or bad, the decision to invade Iraq committed every American citizen. So also when initiating national infrastructure projects or defining criminal offenses. No country can function well if every person is allowed to lay a separate rail network or to have their own definition of murder.

    In order to make decisions on such collective matters, a countrywide public conversation must first be held, following which the people’s representatives—elected in free and fair elections—make a choice. But even after that choice has been made, it should remain open to reexamination and correction. While the network cannot change its previous choices, it can elect a different government next time.

    MAJORITY DICTATORSHIP

    The definition of democracy as a distributed information network with strong self-correcting mechanisms stands in sharp contrast to a common misconception that equates democracy only with elections. Elections are a central part of the democratic tool kit, but they are not democracy. In the absence of additional self-correcting mechanisms, elections can easily be rigged. Even if the elections are completely free and fair, by itself this too doesn’t guarantee democracy. For democracy is not the same thing as majority dictatorship.

    Suppose that in a free and fair election 51 percent of voters choose a government that subsequently sends 1 percent of voters to be exterminated in death camps, because they belong to some hated religious minority. Is this democratic? Clearly it is not. The problem isn’t that genocide demands a special majority of more than 51 percent. It’s not that if the government gets the backing of 60 percent, 75 percent, or even 99 percent of voters, then its death camps finally become democratic. A democracy is not a system in which a majority of any size can decide to exterminate unpopular minorities; it is a system in which there are clear limits on the power of the center.

    Suppose 51 percent of voters choose a government that then takes away the voting rights of the other 49 percent of voters, or perhaps of just 1 percent of them. Is that democratic? Again the answer is no, and it has nothing to do with the numbers. Disenfranchising political rivals dismantles one of the vital self-correcting mechanisms of democratic networks. Elections are a mechanism for the network to say, “We made a mistake; let’s try something else.” But if the center can disenfranchise people at will, that self-correcting mechanism is neutered.

    These two examples may sound outlandish, but they are unfortunately within the realm of the possible. Hitler began sending Jews and communists to concentration camps within months of rising to power through democratic elections, and in the United States numerous democratically elected governments have disenfranchised African Americans, Native Americans, and other oppressed populations. Of course, most assaults on democracy are more subtle. The careers of strongmen like Vladimir Putin, Viktor Orbán, Recep Tayyip Erdoğan, Rodrigo Duterte, Jair Bolsonaro, and Benjamin Netanyahu demonstrate how a leader who uses democracy to rise to power can then use his power to undermine democracy. As Erdoğan once put it, “Democracy is like a tram. You ride it until you arrive at your destination, then you step off.”4

    The most common method strongmen use to undermine democracy is to attack its self-correcting mechanisms one by one, often beginning with the courts and the media. The typical strongman either deprives courts of their powers or packs them with his loyalists and seeks to close all independent media outlets while building his own omnipresent propaganda machine.5

    Once the courts are no longer able to check the government’s power by legal means, and once the media obediently parrots the government line, all other institutions or persons who dare oppose the government can be smeared and persecuted as traitors, criminals, or foreign agents. Academic institutions, municipalities, NGOs, and private businesses are either dismantled or brought under government control. At that stage, the government can also rig the elections at will, for example by jailing popular opposition leaders, preventing opposition parties from participating in the elections, gerrymandering election districts, or disenfranchising voters. Appeals against these antidemocratic measures are dismissed by the government’s handpicked judges. Journalists and academics who criticize these measures are fired. The remaining media outlets, academic institutions, and judicial authorities all praise these measures as necessary steps to protect the nation and its allegedly democratic system from traitors and foreign agents. The strongmen don’t usually take the final step of abolishing the elections outright. Instead, they keep them as a ritual that serves to provide legitimacy and maintain a democratic facade, as happens, for example, in Putin’s Russia.

    Supporters of strongmen often don’t see this process as antidemocratic. They are genuinely baffled when told that electoral victory doesn’t grant them unlimited power. Instead, they see any check on the power of an elected government as undemocratic. However, democracy doesn’t mean majority rule; rather, it means freedom and equality for all. Democracy is a system that guarantees everyone certain liberties, which even the majority cannot take away.

    Nobody disputes that in a democracy the representatives of the majority are entitled to form the government and to advance their preferred policies in myriad fields. If the majority wants war, the country goes to war. If the majority wants peace, the country makes peace. If the majority wants to raise taxes, taxes are raised. If the majority wants to lower taxes, taxes are lowered. Major decisions about foreign affairs, defense, education, taxation, and numerous other policies are all in the hands of the majority.

    But in a democracy, there are two baskets of rights that are protected from the majority’s grasp. One contains human rights. Even if 99 percent of the population wants to exterminate the remaining 1 percent, in a democracy this is forbidden, because it violates the most basic human right—the right to life. The basket of human rights contains many additional rights, such as the right to work, the right to privacy, freedom of movement, and freedom of religion. These rights enshrine the decentralized nature of democracy, making sure that as long as people don’t harm anyone, they can live their lives as they see fit.

    The second crucial basket of rights contains civil rights. These are the basic rules of the democratic game, which enshrine its self-correcting mechanisms. An obvious example is the right to vote. If the majority were permitted to disenfranchise the minority, then democracy would be over after a single election. Other civil rights include freedom of the press, academic freedom, and freedom of assembly, which enable independent media outlets, universities, and opposition movements to challenge the government. These are the key rights that strongmen seek to violate. While sometimes it is necessary to make changes to a country’s self-correcting mechanisms—for example, by expanding the franchise, regulating the media, or reforming the judicial system—such changes should be made only on the basis of a broad consensus including both majority and minority groups. If a small majority could unilaterally change civil rights, it could easily rig elections and get rid of all other checks on its power.

    An important thing to note about both human rights and civil rights is that they don’t just limit the power of the central government; they also impose on it many active duties. It is not enough for a democratic government to abstain from infringing on human and civil rights. It must take actions to ensure them. For example, the right to life imposes on a democratic government the duty to protect citizens from criminal violence. If a government doesn’t kill anyone, but also makes no effort to protect citizens from murder, this is anarchy rather than democracy.

    THE PEOPLE VERSUS THE TRUTH

    Of course, in every democracy, there are lengthy discussions concerning the exact limits of human and civil rights. Even the right to life has limits. There are democratic countries like the United States that impose the death penalty, thereby denying some criminals the right to life. And every country allows itself the prerogative to declare war, thereby sending people to kill and be killed. So where exactly does the right to life end? There are also complicated and ongoing discussions concerning the list of rights that should be included in the two baskets. Who determined that freedom of religion is a basic human right? Should internet access be defined as a civil right? And what about animal rights? Or the rights of AI?

    We cannot resolve these matters here. Both human and civil rights are intersubjective conventions that humans invent rather than discover, and they are determined by historical contingencies rather than universal reason. Different democracies can adopt somewhat different lists of rights. At least from the viewpoint of information flows, what defines a system as “democratic” is only that its center doesn’t have unlimited authority and that the system possesses robust mechanisms to correct the center’s mistakes. Democratic networks assume that everyone is fallible, and that includes even the winners of elections and the majority of voters.

    It is particularly crucial to remember that elections are not a method for discovering truth. Rather, they are a method for maintaining order by adjudicating between people’s conflicting desires. Elections establish what the majority of people desire, rather than what the truth is. And people often desire the truth to be other than what it is. Democratic networks therefore maintain some self-correcting mechanisms to protect the truth even from the will of the majority.

    For example, during the 2002–3 debate over whether to invade Iraq in the wake of the September 11 attacks, the Bush administration claimed that Saddam Hussein was developing weapons of mass destruction and that the Iraqi people were eager to establish an American-style democracy and would welcome the Americans as liberators. These arguments carried the day. In October 2002 the elected representatives of the American people in Congress voted overwhelmingly to authorize the invasion. The resolution passed with a 296 to 133 majority (69 percent) in the House of Representatives and a 77 to 23 majority (77 percent) in the Senate.6 In the early days of the war in March 2003, polls found that the elected representatives were indeed in tune with the mass of voters and that 72 percent of American citizens supported the invasion.7 The will of the American people was clear.

    But the truth turned out to be different from what the government said and what the majority believed. As the war progressed, it became evident that Iraq had no weapons of mass destruction and that many Iraqis had no wish to be “liberated” by the Americans or to establish a democracy. By August 2004 another poll found that 67 percent of Americans believed that the invasion was based on incorrect assumptions. As the years went by, most Americans acknowledged that the decision to invade was a catastrophic mistake.8

    In a democracy the majority has every right to make momentous decisions like starting wars, and that includes the right to make momentous errors. But the majority should at least acknowledge its own fallibility and protect the freedom of minorities to hold and publicize unpopular views, which might turn out to be correct.

    As another example, consider the case of a charismatic leader who is accused of corruption. His loyal supporters obviously wish these accusations to be false. But even if most voters support the leader, their desires should not prevent judges from investigating the accusations and getting to the truth. As with the justice system, so also with science. A majority of voters might deny the reality of climate change, but they should not have the power to dictate scientific truth or to prevent scientists from exploring and publishing inconvenient facts. Unlike parliaments, departments of environmental studies should not reflect the will of the majority.

    Of course, when it comes to making policy decisions about climate change, in a democracy the will of the voters should reign supreme. Acknowledging the reality of climate change does not tell us what to do about it. We always have options, and choosing between them is a question of desire, not truth. One option might be to immediately cut greenhouse gas emissions, even at the cost of slowing economic growth. This means incurring some difficulties today but saving people in 2050 from more severe hardship, saving the island nation of Kiribati from drowning, and saving the polar bears from extinction. A second option might be to continue with business as usual. This means having an easier life today, but making life harder for the next generation, flooding Kiribati, and driving the polar bears—as well as numerous other species—to extinction. Choosing between these two options is a question of desire, and should therefore be done by all voters rather than by a limited group of experts.

    But the one option that should not be on offer in elections is hiding or distorting the truth. If the majority prefers to consume whatever amount of fossil fuels it wishes with no regard to future generations or other environmental considerations, it is entitled to vote for that. But the majority should not be entitled to pass a law stating that climate change is a hoax and that all professors who believe in climate change must be fired from their academic posts. We can choose what we want, but we shouldn’t deny the true meaning of our choice.

    Naturally, academic institutions, the media, and the judiciary may themselves be compromised by corruption, bias, or error. But subordinating them to a governmental Ministry of Truth is likely to make things worse. The government is already the most powerful institution in developed societies, and it often has the greatest interest in distorting or hiding inconvenient facts. Allowing the government to supervise the search for truth is like appointing the fox to guard the chicken coop.

    To discover the truth, it is better to rely on two other methods. First, academic institutions, the media, and the judiciary have their own internal self-correcting mechanisms for fighting corruption, correcting bias, and exposing error. In academia, peer-reviewed publication is a far better check on error than supervision by government officials, because academic promotion often depends on uncovering past mistakes and discovering unknown facts. In the media, free competition means that if one outlet decides not to break a scandal, perhaps for self-serving reasons, others are likely to jump at the scoop. In the judiciary, a judge that takes bribes may be tried and punished just like any other citizen.

    Second, the existence of several independent institutions that seek the truth in different ways allows these institutions to check and correct one another. For example, if powerful corporations manage to break down the peer-review mechanism by bribing a sufficiently large number of scientists, investigative journalists and courts can expose and punish the perpetrators. If the media or the courts are afflicted by systematic racist biases, it is the job of sociologists, historians, and philosophers to expose these biases. None of these mechanisms are completely fail-safe, but no human institution is. Government certainly isn’t.

    THE POPULIST ASSAULT

    If all this sounds complicated, it is because democracy should be complicated. Simplicity is a characteristic of dictatorial information networks in which the center dictates everything and everybody silently obeys. It’s easy to follow this dictatorial monologue. In contrast, democracy is a conversation with numerous participants, many of them talking at the same time. It can be hard to follow such a conversation.

    Moreover, the most important democratic institutions tend to be bureaucratic behemoths. Whereas citizens avidly follow the biological dramas of the princely court and the presidential palace, they often find it difficult to understand how parliaments, courts, newspapers, and universities function. This is what helps strongmen mount populist attacks on institutions, dismantle all self-correcting mechanisms, and concentrate power in their hands. We discussed populism briefly in the prologue, to help explain the populist challenge to the naive view of information. Here we need to revisit populism, get a broader understanding of its worldview, and explain its appeal to antidemocratic strongmen.

    The term “populism” derives from the Latin populus, which means “the people.” In democracies, “the people” is considered the sole legitimate source of political authority. Only representatives of the people should have the authority to declare wars, pass laws, and raise taxes. Populists cherish this basic democratic principle, but somehow conclude from it that a single party or a single leader should monopolize all power. In a curious political alchemy, populists manage to base a totalitarian pursuit of unlimited power on a seemingly impeccable democratic principle. How does it happen?

    The most novel claim populists make is that they alone truly represent the people. Since in democracies only the people should have political power, and since allegedly only the populists represent the people, it follows that the populist party should have all political power to itself. If some party other than the populists wins elections, it does not mean that this rival party won the people’s trust and is entitled to form a government. Rather, it means that the elections were stolen or that the people were deceived to vote in a way that doesn’t express their true will.

    It should be stressed that for many populists, this is a genuinely held belief rather than a propaganda gambit. Even if they win just a small share of votes, populists may still believe they alone represent the people. An analogous case are communist parties. In the U.K., for example, the Communist Party of Great Britain (CPGB) never won more than 0.4 percent of votes in a general election,9 but was nevertheless adamant that it alone truly represented the working class. Millions of British workers, they claimed, were voting for the Labour Party or even for the Conservative Party rather than for the CPGB because of “false consciousness.” Allegedly, through their control of the media, universities, and other institutions, the capitalists managed to deceive the working class into voting against its true interests, and only the CPGB could see through this deception. In like fashion, populists can believe that the enemies of the people have deceived the people to vote against its true will, which the populists alone represent.

    A fundamental part of this populist credo is the belief that “the people” is not a collection of flesh-and-blood individuals with various interests and opinions, but rather a unified mystical body that possesses a single will—“the will of the people.” Perhaps the most notorious and extreme manifestation of this semireligious belief was the Nazi motto “Ein Volk, ein Reich, ein Führer,” which means “One People, One Country, One Leader.” Nazi ideology posited that the Volk (people) had a single will, whose sole authentic representative was the Führer (leader). The leader allegedly had an infallible intuition for how the people felt and what the people wanted. If some German citizens disagreed with the leader, it didn’t mean that the leader might be in the wrong. Rather, it meant that the dissenters belonged to some treasonous outsider group—Jews, communists, liberals—instead of to the people.

    The Nazi case is of course extreme, and it is grossly unfair to accuse all populists of being crypto-Nazis with genocidal inclinations. However, many populist parties and politicians deny that “the people” might contain a diversity of opinions and interest groups. They insist that the real people has only one will and that they alone represent this will. In contrast, their political rivals—even when the latter enjoy substantial popular support—are depicted as “alien elites.” Thus, Hugo Chávez ran for the presidency in Venezuela with the slogan “Chávez is the people!”10 President Erdoğan of Turkey once railed against his domestic critics, saying, “We are the people. Who are you?”—as if his critics weren’t Turks, too.11

    How can you tell, then, whether someone is part of the people or not? Easy. If they support the leader, they are part of the people. This, according to the German political philosopher Jan-Werner Müller, is the defining feature of populism. What turns someone into a populist is claiming that they alone represent the people and that anyone who disagrees with them—whether state bureaucrats, minority groups, or even the majority of voters—either suffers from false consciousness or isn’t really part of the people.12

    This is why populism poses a deadly threat to democracy. While democracy agrees that the people is the only legitimate source of power, democracy is based on the understanding that the people is never a unitary entity and, therefore, cannot possess a single will. Every people—whether Germans, Venezuelans, or Turks—is composed of many different groups, with a plurality of opinions, wills, and representatives. No group, including the majority group, is entitled to exclude other groups from membership in the people. This is what makes democracy a conversation. Holding a conversation presupposes the existence of several legitimate voices. If, however, the people has only one legitimate voice, there can be no conversation. Rather, the single voice dictates everything. Populism may therefore claim adherence to the democratic principle of “people’s power,” but it effectively empties democracy of meaning and seeks to establish a dictatorship.

    Populism undermines democracy in another, more subtle, but equally dangerous way. Having claimed that they alone represent the people, populists argue that the people is not just the sole legitimate source of political authority but the sole legitimate source of all authority. Any institution that derives its authority from something other than the will of the people is antidemocratic. As the self-proclaimed representatives of the people, populists consequently seek to monopolize not just political authority but all types of authority and to take control of institutions such as media outlets, courts, and universities. By taking the democratic principle of “people’s power” to its extreme, populists turn totalitarian.

    In fact, while democracy means that authority in the political sphere comes from the people, it doesn’t deny the validity of alternative sources of authority in other spheres. As discussed above, in a democracy independent media outlets, courts, and universities are essential self-correcting mechanisms that protect the truth even from the will of the majority. Biology professors claim that humans evolved from apes because the evidence supports this, even if the majority wills it to be otherwise. Journalists can reveal that a popular politician took a bribe, and if compelling evidence is presented in court, a judge may send that politician to jail, even if most people don’t want to believe these accusations.

    Populists are suspicious of institutions that in the name of objective truths override the supposed will of the people. They tend to see this as a smoke screen for elites grabbing illegitimate power. This drives populists to be skeptical of the pursuit of truth, and to argue—as we saw in the prologue—that “power is the only reality.” They thereby seek to undercut or appropriate the authority of any independent institutions that might oppose them. The result is a dark and cynical view of the world as a jungle and of human beings as creatures obsessed with power alone. All social interactions are seen as power struggles, and all institutions are depicted as cliques promoting the interests of their own members. In the populist imagination, courts don’t really care about justice; they only protect the privileges of the judges. Yes, the judges talk a lot about justice, but this is a ploy to grab power for themselves. Newspapers don’t care about facts; they spread fake news to mislead the people and benefit the journalists and the cabals that finance them. Even scientific institutions aren’t committed to the truth. Biologists, climatologists, epidemiologists, economists, historians, and mathematicians are just another interest group feathering its own nest—at the expense of the people.

    In all, it’s a rather sordid view of humanity, but two things nevertheless make it appealing to many. First, since it reduces all interactions to power struggles, it simplifies reality and makes events like wars, economic crises, and natural disasters easy to understand. Anything that happens—even a pandemic—is about elites pursuing power. Second, the populist view is attractive because it is sometimes correct. Every human institution is indeed fallible and suffers from some level of corruption. Some judges do take bribes. Some journalists do intentionally mislead the public. Academic disciplines are occasionally plagued by bias and nepotism. That is why every institution needs self-correcting mechanisms. But since populists are convinced that power is the only reality, they cannot accept that a court, a media outlet, or an academic discipline would ever be inspired by the value of truth or justice to correct itself.

    While many people embrace populism because they see it as an honest account of human reality, strongmen are attracted to it for a different reason. Populism offers strongmen an ideological basis for making themselves dictators while pretending to be democrats. It is particularly useful when strongmen seek to neutralize or appropriate the self-correcting mechanisms of democracy. Since judges, journalists, and professors allegedly pursue political interests rather than truth, the people’s champion—the strongman—should control these positions instead of allowing them to fall into the hands of the people’s enemies. Similarly, since even the officials in charge of arranging elections and publicizing their results may be part of a nefarious conspiracy, they too should be replaced by the strongman’s loyalists.

    In a well-functioning democracy, citizens trust the results of elections, the decisions of courts, the reports of media outlets, and the findings of scientific disciplines because citizens believe these institutions are committed to the truth. Once people think that power is the only reality, they lose trust in all these institutions, democracy collapses, and the strongmen can seize total power.
    Of course, populism could lead to anarchy rather than totalitarianism, if it undermines trust in the strongmen themselves. If no human is interested in truth or justice, doesn’t this apply to Mussolini or Putin too? And if no human institution can have effective self-correcting mechanisms, doesn’t this include Mussolini’s National Fascist Party or Putin’s United Russia party? How can a deep-seated distrust of all elites and institutions be squared with unwavering admiration for one leader and party? This is why populists ultimately depend on the mystical notion that the strongman embodies the people. When trust in bureaucratic institutions like election boards, courts, and newspapers is particularly low, an enhanced reliance on mythology is the only way to preserve order.

    MEASURING THE STRENGTH OF DEMOCRACIES

    Strongmen who claim to represent the people may well rise to power through democratic means, and often rule behind a democratic facade. Rigged elections in which they win overwhelming majorities serve as proof of the mystical bond between the leader and the people. Consequently, to measure how democratic an information network is, we cannot use a simple yardstick like whether elections are being held regularly. In Putin’s Russia, in Iran, and even in North Korea elections are held like clockwork. Rather, we need to ask much more complex questions like “What mechanisms prevent the central government from rigging the elections?” “How safe is it for leading media outlets to criticize the government?” and “How much authority does the center appropriate to itself?” Democracy and dictatorship aren’t binary opposites, but rather a continuum. To decide whether a network is closer to the democratic or the dictatorial end of the continuum, we need to understand how information flows in the network and what shapes the political conversation.

    If one person dictates all the decisions, and even their closest advisers are terrified to voice a dissenting view, no conversation is taking place. Such a network is situated at the extreme dictatorial end of the spectrum. If nobody can voice unorthodox opinions publicly, but behind closed doors a small circle of party bosses or senior officials are able to freely express their views, then this is still a dictatorship, but it has taken a baby step in the direction of democracy. If 10 percent of the population participate in the political conversation by airing their opinions, voting in fair elections, and running for office, that may be considered a limited democracy, as was the case in many ancient city-states like Athens, or in the early days of the United States, when only wealthy white men had such political rights. As the percentage of people taking part in the conversation rises, so the network becomes more democratic.

    The focus on conversations rather than elections raises a host of interesting questions. For example, where does that conversation take place? North Korea, for example, has the Mansudae Assembly Hall in Pyongyang, where the 687 members of the Supreme People’s Assembly meet and talk. However, while this Assembly is officially known as North Korea’s legislature, and while elections to the Assembly are held every five years, this body is widely considered a rubber stamp, executing decisions taken elsewhere. The anodyne discussions follow a predetermined script, and they aren’t geared to change anyone’s mind about anything.13

    Is there perhaps another, more private hall in Pyongyang where the crucial conversations take place? Do Politburo members ever dare criticize Kim Jong Un’s policies during formal meetings? Perhaps it can be done in unofficial dinner parties or in unofficial think tanks? Information in North Korea is so concentrated and so tightly controlled that we cannot provide clear answers to these questions.14
    Similar questions can be asked about the United States. In the United States, unlike in North Korea, people are free to say almost anything they want. Scathing public attacks on the government are a daily occurrence. But where is the room where the crucial conversations happen, and who sits there? The U.S. Congress was designed to fulfill this function, with the people’s representatives meeting to converse and try to convince one another. But when was the last time that an eloquent speech in Congress by a member of one party persuaded members of the other party to change their minds about anything? Wherever the conversations that shape American politics now take place, it is definitely not in Congress. Democracies die not only when people are not free to talk but also when people are not willing or able to listen.

    STONE AGE DEMOCRACIES

    Based on the above definition of democracy, we can now turn to the historical record and examine how changes in information technology and information flows have shaped the history of democracy. To judge by the archaeological and anthropological evidence, democracy was the most typical political system among archaic hunter-gatherers. Stone Age bands obviously didn’t have formal institutions like elections, courts, and media outlets, but their information networks were usually distributed and gave ample opportunities for self-correction. In bands numbering just a few dozen people information could easily be shared among all group members, and when the band decided where to pitch camp, where to go hunting, or how to handle a conflict with another band, everyone could take part in the conversation and dispute each other. Bands usually belonged to a larger tribe that included hundreds or even thousands of people. But when important choices affecting the whole tribe had to be made, such as whether to go to war, tribes were usually still small enough for a large percentage of their members to gather in one place and converse.15

    While bands and tribes sometimes had dominant leaders, these tended to exercise only limited authority. Leaders had no standing armies, police forces, or governmental bureaucracies at their disposal, so they couldn’t just impose their will by force.16 Leaders also found it difficult to control the economic basis of people’s lives. In modern times, dictators like Vladimir Putin and Saddam Hussein have often based their political power on monopolizing economic assets like oil wells.17 In medieval and classical antiquity, Chinese emperors, Greek tyrants, and Egyptian pharaohs dominated society by controlling granaries, silver mines, and irrigation canals. In contrast, in a hunter-gatherer economy such centralized economic control was possible only under special circumstances. For example, along the northwestern coast of North America some hunter-gatherer economies relied on catching and preserving large numbers of salmon. Since salmon runs peaked for a few weeks in specific creeks and rivers, a powerful chief could monopolize this asset.18

    But this was exceptional. Most hunter-gatherer economies were far more diversified. One leader, even supported by a few allies, could not corral the savanna and prevent people from gathering plants and hunting animals there. If all else failed, hunter-gatherers could therefore vote with their feet. They had few possessions, and their most important assets were their personal skills and personal friends. If a chief turned dictatorial, people could just walk away.19

    Even when hunter-gatherers did end up ruled by a domineering chief, as happened among the salmon-fishing people of northwestern America, at least that chief was accessible. He didn’t live in a faraway fortress surrounded by an unfathomable bureaucracy and a cordon of armed guards. If you wanted to voice a complaint or a suggestion, you could usually get within earshot of him. The chief couldn’t control public opinion, nor could he shut himself off from it. In other words, there was no way for a chief to force all information to flow through the center, or to prevent people from talking with one another, criticizing him, or organizing against him.20

    In the millennia following the agricultural revolution, and especially after writing helped create large bureaucratic polities, it became easier to centralize the flow of information and harder to maintain the democratic conversation. In small city-states like those of ancient Mesopotamia and Greece, autocrats like Lugal-Zagesi of Umma and Pisistratus of Athens relied on bureaucrats, archives, and a standing army to monopolize key economic assets and information about ownership, taxation, diplomacy, and politics. It simultaneously became harder for the mass of citizens to keep in direct touch with one another. There was no mass communication technology like newspapers or radio, and it was not easy to squeeze tens of thousands of citizens into the main city square to hold a communal discussion.

    Democracy was still an option for these small city-states, as the history of both early Sumer and classical Greece clearly indicates.21 However, the democracy of ancient city-states tended to be less inclusive than the democracy of archaic hunter-gatherer bands. Probably the most famous example of ancient city-state democracy is Athens in the fifth and fourth centuries BCE. All adult male citizens could participate in the Athenian assembly, vote on public policy, and be elected to public offices. But women, slaves, and noncitizen residents of the city did not enjoy these privileges. Only about 25–30 percent of the adult population of Athens enjoyed full political rights.22

    As the size of polities continued to increase, and city-states were superseded by larger kingdoms and empires, even Athenian-style partial democracy disappeared. All the famous examples of ancient democracies are city-states such as Athens and Rome. In contrast, we don’t know of any large-scale kingdom or empire that operated along democratic lines.

    For example, when in the fifth century BCE Athens expanded from a city-state into an empire, it did not grant citizenship and political rights to those it conquered. The city of Athens remained a limited democracy, but the much bigger Athenian Empire was ruled autocratically from the center. All the important decisions about taxes, diplomatic alliances, and military expeditions were taken in Athens. Subject lands like the islands of Naxos and Thasos had to obey the orders of the Athenian popular assembly and elected officials, without the Naxians and Thasians being able to vote in that assembly or be elected to office. It was also difficult for Naxos, Thasos, and other subject lands to coordinate a united opposition to the decisions taken in the Athenian center, and if they tried to do so, it would have brought ruthless Athenian reprisals. Information in the Athenian Empire flowed to and from Athens.23

    When the Roman Republic built its empire, conquering first the Italian Peninsula and eventually the entire Mediterranean basin, the Romans took a somewhat different course. Rome gradually did extend citizenship to the conquered people. It began by granting citizenship to the inhabitants of Latium, then to the inhabitants of other Italian regions, and finally to inhabitants of even distant provinces like Gallia and Syria. However, as citizenship was extended to more people, the political rights of citizens were simultaneously restricted.

    The ancient Romans had a clear understanding of what democracy means, and they were originally fiercely committed to the democratic ideal. After expelling the last king of Rome in 509 BCE, the Romans developed a deep dislike for monarchy and a fear of giving unlimited power to any single individual or institution. Supreme executive power was therefore shared by two consuls who balanced each other. These consuls were chosen by citizens in free elections, held office for a single year, and were additionally checked by the powers of the popular assembly, of the Senate, and of other elected officials like the tribunes.

    But when Rome extended citizenship to Latins, Italians, and finally to Gauls and Syrians, the power of the popular assembly, the tribunes, the Senate, and even the two consuls was gradually reduced, until in the late first century BCE the Caesar family established its autocratic rule. Anticipating present-day strongmen like Putin, Augustus didn’t crown himself king, and pretended that Rome was still a republic. The Senate and the popular assembly continued to convene, and every year citizens continued to choose consuls and tribunes. But these institutions were emptied of real power.24

    In 212 CE, the emperor Caracalla—the offspring of a Phoenician family from North Africa—took a seemingly momentous step and granted automatic Roman citizenship to all free adult males throughout the vast empire. Rome in the third century CE accordingly had tens of millions of citizens.25 But by that time, all the important decisions were made by a single unelected emperor. While consuls were still ceremonially chosen every year, Caracalla inherited power from his father Septimius Severus, who became emperor by winning a civil war. To cement his rule, the most important step Caracalla took was murdering his brother and rival Geta.

    When Caracalla ordered the murder of Geta, decided to devalue the Roman currency, or declared war on the Parthian Empire, he had no need to ask permission from the Roman people. All of Rome’s self-correcting mechanisms had been neutralized long before. If Caracalla made some error in foreign or domestic policy, neither the Senate nor any officials like the consuls or tribunes could intervene to correct it, except by rising in rebellion or assassinating him. And when Caracalla was indeed assassinated in 217, it only led to a new round of civil wars culminating in the rise of new autocrats. Rome in the third century CE, like Russia in the eighteenth century, was, in the words of Madame de Staël, “autocracy tempered by strangulation.”

    By the third century CE, not only the Roman Empire but all other major human societies on earth were centralized information networks lacking strong self-correcting mechanisms. This was true of the Parthian and Sassanian Empires in Persia, of the Kushan and Gupta Empires in India, and of China’s Han Empire and its successor Three Kingdoms.26 Thousands of more small-scale societies continued to function democratically in the third century CE and beyond, but it seemed that distributed democratic networks were simply incompatible with large-scale societies.

    CAESAR FOR PRESIDENT!

    Were large-scale democracies really unworkable in the ancient world? Or did autocrats like Augustus and Caracalla deliberately sabotage them? This question is important not only for our understanding of ancient history but also for our view of democracy’s future in the age of AI. How do we know whether democracies fail because they are undermined by strongmen or because of much deeper structural and technological reasons?

    To answer that question, let’s take a closer look at the Roman Empire. The Romans were clearly familiar with the democratic ideal, and it continued to be important to them even after the Caesar family rose to power. Otherwise, Augustus and his heirs would not have bothered to maintain seemingly democratic institutions like the Senate or annual elections to the consulate and other offices. So why did power end up in the hands of an unelected emperor?

    In theory, even after Roman citizenship was expanded to tens of millions of people throughout the Mediterranean basin, wasn’t it possible to hold empire-wide elections for the position of emperor? This would surely have required very complicated logistics, and it would have taken several months to learn the results of the elections. But was that really a deal breaker?

    The key misconception here is equating democracy with elections. If the Roman Empire wanted to, it could technically have held empire-wide elections for emperor. But the real question we should ask is whether the Roman Empire could have held an ongoing empire-wide political conversation. In present-day North Korea no democratic conversation takes place because people aren’t free to talk, yet we could well imagine a situation when this freedom is guaranteed—as it is in South Korea. In the present-day United States the democratic conversation is endangered by people’s inability to listen to and respect their political rivals, yet this can presumably still be fixed. By contrast, in the Roman Empire there was simply no way to conduct or sustain a democratic conversation, because the technological means to hold such a conversation did not exist.

    To hold a conversation, it is not enough to have the freedom to talk and the ability to listen. There are also two technical preconditions. First, people need to be within hearing range of each other. This means that the only way to hold a political conversation in a territory the size of the United States or the Roman Empire is with the help of some kind of information technology that can swiftly convey what people say over long distances.

    Second, people need at least a rudimentary understanding of what they are talking about. Otherwise, they are just making noise, not holding a meaningful conversation. People usually have a good understanding of political issues of which they have direct experience. Poor people have many insights about poverty that escape economics professors, and ethnic minorities understand racism in a much more profound way than people who never suffered from it, for example. However, if lived experience were the only way to understand crucial political issues, large-scale political conversations would be impossible. For then every group of people could talk meaningfully only about its own experiences. Even worse, nobody else could understand what they were saying. If lived experience is the sole possible source of knowledge, then merely listening to the insights gained from someone else’s lived experience cannot impart these insights to me.

    The only way to have a large-scale political conversation among diverse groups of people is if people can gain some understanding of issues that they have never experienced firsthand. In a large polity, it is a crucial role of the education system and the media to inform people about things they have never faced themselves. If there is no education system or media platforms to perform this role, no meaningful large-scale conversations can take place.

    In a small Neolithic town of a few thousand inhabitants people might sometimes have been afraid to say what they thought, or might have refused to listen to their rivals, but it was relatively easy to satisfy the more fundamental technical preconditions for meaningful discourse. First, people lived in proximity to one another, so they could easily meet most other community members and hear their voices. Second, everybody had intimate knowledge of the dangers and opportunities that the town faced. If an enemy war party approached, everyone could see it. If the river flooded the fields, everyone witnessed the economic effects. When people talked about war and hunger, they knew what they were saying.

    In the fourth century BCE, the city-state of Rome was still small enough to allow a large percentage of its citizens to congregate in the Forum in times of emergency, listen to respected leaders, and voice their personal views on the matter at hand. When in 390 BCE Gallic invaders attacked Rome, almost everyone lost a relative in the defeat at the Battle of the Allia and lost property when the victorious Gauls then sacked Rome. The desperate Romans appointed Marcus Camillus as dictator. In Rome, the dictator was a public official appointed in times of emergency who had unlimited powers but only for a short predetermined period, following which he was held accountable for his actions. After Camillus led the Romans to victory, everybody could see that the emergency was over, and Camillus stepped down.27

    In contrast, by the third century CE, the Roman Empire had a population of between sixty and seventy-five million people,28 spread over five million square kilometers.29 Rome lacked mass communication technology like radio or daily newspapers. Only 10–20 percent of adults had reading skills,30 and there was no organized education system that could inform them about the geography, history, and economy of the empire. True, many people across the empire did share some cultural ideas, such as a strong belief in the superiority of Roman civilization over the barbarians. These shared cultural beliefs were crucial in preserving order and holding the empire together. But their political implications were far from clear, and in times of crisis there was no possibility to hold a public conversation about what should be done.

    How could Syrian merchants, British shepherds, and Egyptian villagers converse about the ongoing wars in the Middle East or about the immigration crisis brewing along the Danube? The lack of a meaningful public conversation was not the fault of Augustus, Nero, Caracalla, or any of the other emperors. They didn’t sabotage Roman democracy. Given the size of the empire and the available information technology, democracy was simply unworkable. This was acknowledged already by ancient philosophers like Plato and Aristotle, who argued that democracy can work only in small-scale city-states.31

    If the absence of Roman democracy had merely been the fault of particular autocrats, we should have at least seen large-scale democracies flourishing in other places, like in Sassanian Persia, Gupta India, or Han China. But prior to the development of modern information technology, there are no examples of large-scale democracies anywhere.

    It should be stressed that in many large-scale autocracies local affairs were often managed democratically. The Roman emperor didn’t have the information needed to micromanage hundreds of cities across the empire, whereas local citizens in each city could continue to hold a meaningful conversation about municipal politics. Consequently, long after the Roman Empire became an autocracy, many of its cities continued to be governed by local assemblies and elected officials. At a time when elections to the consulship in Rome became ceremonial affairs, elections to municipal offices in small cities like Pompeii were hotly contested.

    Pompeii was destroyed in the eruption of Vesuvius in 79 CE, during the reign of the emperor Titus. Archaeologists uncovered about fifteen hundred graffiti concerned with various local election campaigns. One coveted office was that of the city’s aedile—the magistrate in charge of maintaining the city’s infrastructure and public buildings.32 Lucretius Fronto’s supporters drew the graffiti “If honest living is thought to be any recommendation, then Lucretius Fronto is worthy of being elected.” One of his opponents, Gaius Julius Polybius, ran with the slogan “Elect Gaius Julius Polybius to the office of aedile. He provides good bread.”

    There were also endorsements by religious groups and professional associations, such as “The worshippers of Isis demand the election of Gnaeus Helvius Sabinus” and “All the mule drivers request that you elect Gaius Julius Polybius.” There was dirty work, too. Someone who clearly wasn’t Marcus Cerrinius Vatia drew the graffiti “All the drunkards ask you to elect Marcus Cerrinius Vatia” and “The petty thieves ask you to elect Vatia.”33 Such electioneering indicates that the position of aedile had power in Pompeii and that the aedile was chosen in relatively free and fair elections, rather than appointed by the imperial autocrat in Rome.

    Even in empires whose rulers never had any democratic pretensions, democracy could still flourish in local settings. In the Tsarist Empire, for example, the daily lives of millions of villagers were managed by rural communes. Going back at least to the eleventh century, each commune usually included fewer than a thousand people. They were subject to a landlord and bore many obligations to their lord and to the central Tsarist state, but they had considerable autonomy in managing their internal affairs and in deciding how to discharge their external obligations, such as paying taxes and providing military recruits. The commune mediated local disputes, provided emergency relief, enforced social norms, oversaw the distribution of land to individual households, and regulated access to shared resources like forests and pastures. Decisions on important matters were made in communal meetings in which the heads of local households expressed their views and chose the commune’s elder. Resolutions at least tried to reflect the majority’s will.34

    In Tsarist villages and Roman cities a form of democracy was possible because a meaningful public conversation was possible. Pompeii was a city of about eleven thousand people in 79 CE,35 so everybody could supposedly judge for themselves whether Lucretius Fronto was an honest man and whether Marcus Cerrinius Vatia was a drunken thief. But democracy at a scale of millions became possible only in the modern age, when mass media changed the nature of large-scale information networks.

    MASS MEDIA MAKES MASS DEMOCRACY POSSIBLE

    Mass media can be defined as the ability to quickly connect millions of people even when they are separated by large distances. The printing press was a crucial step in that direction. Print made it possible to cheaply and quickly produce large numbers of books and pamphlets, which enabled more people to voice their opinions and be heard over a large territory, even if the process still took time. This sustained some of the first experiments in large-scale democracy, such as the Polish-Lithuanian Commonwealth established in 1569 and the Dutch Republic established in 1579.

    Some may contest the characterization of these polities as “democratic,” since only a minority of relatively wealthy citizens enjoyed full political rights. In the Polish-Lithuanian Commonwealth, political rights were reserved for adult male members of the szlachta—the nobility. These numbered up to 300,000 individuals, or about 5 percent of the total adult population.36 One of the szlachta’s prerogatives was to elect the king, but since voting required traveling long distances to a national convention, few exercised their right. In the sixteenth and seventeenth centuries participation in royal elections usually ranged between 3,000 and 7,000 voters, except for the 1669 elections in which 11,271 participated.37 While this hardly sounds democratic in the twenty-first century, it should be remembered that all large-scale democracies until the twentieth century limited political rights to a small circle of relatively wealthy men. Democracy is never a matter of all or nothing. It is a continuum, and late sixteenth-century Poles and Lithuanians explored previously unknown regions of that continuum.

    Aside from electing its king, Poland-Lithuania had an elected parliament (the Sejm) that approved or blocked new legislation and had the power to veto royal decisions on taxation and foreign affairs. Moreover, citizens enjoyed a list of inviolable rights such as freedom of assembly and freedom of religion. In the late sixteenth and early seventeenth centuries, when most of Europe suffered from bitter religious conflicts and persecutions, Poland-Lithuania was a tolerant haven, where Catholics, Greek Orthodox, Lutherans, Calvinists, Jews, and even Muslims coexisted in relative harmony.38 In 1616, more than a hundred mosques functioned in the commonwealth.39

    In the end, however, the Polish-Lithuanian experiment in decentralization proved to be impractical. The country was Europe’s second-largest state (after Russia), covering almost a million square kilometers and including most of the territory of today’s Poland, Lithuania, Belarus, and Ukraine. It lacked the information, communication, and education systems necessary to hold a meaningful political conversation between Polish aristocrats, Lithuanian noblemen, Ukrainian Cossacks, and Jewish rabbis spread from the Baltic Sea to the Black Sea. Its self-correcting mechanisms were also too costly, paralyzing the power of the central government. In particular, every single Sejm deputy was given the right to veto all parliamentary legislation, which led to political deadlock. The combination of a large and diverse polity with a weak center proved fatal. The commonwealth was torn apart by centrifugal forces, and its pieces were then divided between the centralized autocracies of Russia, Austria, and Prussia.

    The Dutch experiment fared better. In some ways the Dutch United Provinces were even less centralized than the Polish-Lithuanian Commonwealth, since they lacked a monarch, and were a union of seven autonomous provinces, which were in turn made up of self-governing towns and cities.40 This decentralized nature is reflected in the plural form of how the country was known abroad—the Netherlands in English, Les Pays-Bas in French, Los Países Bajos in Spanish, and so on.

    However, taken together the United Provinces were twenty-five times smaller in landmass than Poland-Lithuania and possessed a much better information, communication, and education system that tied its constituent parts closely together.41 The United Provinces also pioneered a new information technology with a big future. In June 1618 a pamphlet titled Courante uyt Italien, Duytslandt &c. appeared in Amsterdam. As its title indicated, it carried news from the Italian Peninsula, the German lands, and other places. There was nothing remarkable about this particular pamphlet, except that new issues were published in the following weeks, too. They appeared regularly until 1670, when the Courante uyt Italien, Duytslandt &c. merged with other serial pamphlets into the Amsterdamsche Courant, which appeared until 1903, when it was merged into De Telegraaf—the Netherlands’ largest newspaper to this day.42

    The newspaper is a periodic pamphlet, and it was different from earlier one-off pamphlets because it had a much stronger self-correcting mechanism. Unlike one-off publications, a weekly or daily newspaper has a chance to correct its mistakes and an incentive to do so in order to win the public’s trust. Shortly after the Courante uyt Italien, Duytslandt &c. appeared, a competing newspaper titled Tijdinghen uyt Verscheyde Quartieren (Tidings from Various Quarters) made its debut. The Courante was generally considered more reliable, because it tried to check its stories before publishing them, and because the Tijdinghen was accused of being overly patriotic and reporting only news favorable to the Netherlands. Nevertheless, both newspapers survived, because, as one reader explained, “one can always find something in one newspaper that is not available in the other.” In the following decades dozens of additional newspapers were published in the Netherlands, which became Europe’s journalistic hub.43

    Newspapers that succeeded in gaining widespread trust became the architects and mouthpieces of public opinion. They created a far more informed and engaged public, which changed the nature of politics, first in the Netherlands and later around the world.44 The political influence of newspapers was so crucial that newspaper editors often became political leaders. Jean-Paul Marat rose to power in revolutionary France by founding and editing L’Ami du People; Eduard Bernstein helped create Germany’s Social Democratic Party by editing Der Sozialdemokrat; Vladimir Lenin’s most important position before becoming Soviet dictator was editor of Iskra; and Benito Mussolini rose to fame first as a socialist journalist in Avanti! and later as founder and editor of the firebrand right-wing paper Il Popolo d’Italia.

    Newspapers played a crucial role in the formation of early modern democracies like the United Provinces in the Low Countries, the United Kingdom in the British Isles, and the United States in North America. As the names themselves indicate, these were not city-states like ancient Athens and Rome but amalgams of different regions glued together in part by this new information technology. For example, when on December 6, 1825, President John Quincy Adams gave his First Annual Message to the U.S. Congress, the text of the address and summaries of the main points were published over the next weeks by newspapers from Boston to New Orleans (at the time, hundreds of newspapers and magazines were being published in the United States45).

    Adams declared his administration’s intentions of initiating numerous federal projects ranging from the construction of roads to the founding of an astronomical observatory, which he poetically named “light-house of the skies.” His speech ignited a fierce public debate, much of it conducted in print between those who supported such “big government” plans as essential for the development of the United States and many who preferred a “small government” approach and saw Adams’s plans as federal overreach and an encroachment on states’ rights.

    Northern supporters of the “small government” camp complained that it was unconstitutional for the federal government to tax the citizens of richer states in order to build roads in poorer states. Southerners feared that a federal government that claims the power to build a lighthouse of the sky in their backyard may one day claim the power to free their slaves, too. Adams was accused of harboring dictatorial ambitions, while the erudition and sophistication of his speech were criticized as elitist and disconnected from ordinary Americans. The public debates over the 1825 message to Congress dealt a severe blow to the reputation of the Adams administration and helped pave the way to Adams’s subsequent electoral defeat. In the 1828 presidential elections, Adams lost to Andrew Jackson—a rich slaveholding planter from Tennessee who was successfully rebranded in numerous newspaper columns as “the man of the people” and who claimed that the previous elections were in fact stolen by Adams and by the corrupt Washington elites.46

    Newspapers of the time were of course still slow and limited compared with the mass media of today. Newspapers traveled at the pace of a horse or sailboat, and relatively few people read them regularly. There were no newsstands or street vendors, so people had to buy subscriptions, which were expensive; average annual subscriptions cost around one week’s wages for a skilled journeyman. As a result, the total number of subscribers to all U.S. newspapers in 1830 is estimated at just seventy-eight thousand. Since some subscribers were associations or businesses rather than individuals, and since every copy was probably read by several people, it seems reasonable to assume that regular newspaper readership numbered in the hundreds of thousands. But millions more rarely, if ever, read newspapers.47

    No wonder that American democracy in those days was a limited affair—and the domain of wealthy white men. In the 1824 elections that brought Adams to power, 1.3 million Americans were theoretically eligible to vote, out of an adult population of about 5 million (or around 25 percent). Only 352,780 people—7 percent of the total adult population—actually made use of their right. Adams didn’t even win a majority of those who voted. Owing to the quirks of the U.S. electoral system, he became president thanks to the support of just 113,122 voters, or not much more than 2 percent of adults, and 1 percent of the total population.48 In Britain at the same time, only about 400,000 people were eligible to vote for Parliament, or around 6 percent of the adult population. Moreover, 30 percent of parliamentary seats were not even contested.49

    You may wonder whether we are talking about democracies at all. At a time when the United States had more slaves than voters (more than 1.5 million Americans were enslaved in the early 1820s),50 was the United States really a democracy? This is a question of definitions. As with the late sixteenth-century Polish-Lithuanian Commonwealth, so also with the early nineteenth-century United States, “democracy” is a relative term. As noted earlier, democracy and autocracy aren’t absolutes; they are part of a continuum. In the early nineteenth century, out of all large-scale human societies, the United States was probably the closest to the democratic end of the continuum. Giving 25 percent of adults the right to vote doesn’t sound like much today, but in 1824 that was a far higher percentage than in the Tsarist, Ottoman, or Chinese Empires, in which nobody had the right to vote.51

    Besides, as emphasized throughout this chapter, voting is not the only thing that counts. An even more important reason to consider the United States in 1824 a democracy is that compared with most other polities of its day, the new country possessed much stronger self-correcting mechanisms. The Founding Fathers were inspired by ancient Rome—witness the Senate and the Capitol in Washington—and they were well aware that the Roman Republic eventually turned into an autocratic empire. They feared that some American Caesar would do something similar to their republic, and constructed multiple overlapping self-correcting mechanisms, known as the system of checks and balances. One of these was a free press. In ancient Rome, the self-correcting mechanisms stopped functioning as the republic enlarged its territory and population. In the United States, modern information technology combined with freedom of the press helped the self-correcting mechanisms survive even as the country extended from the Atlantic to the Pacific.

    It was these self-correcting mechanisms that gradually enabled the United States to expand the franchise, abolish slavery, and turn itself into a more inclusive democracy. As noted in chapter 3, the Founding Fathers committed enormous mistakes—such as endorsing slavery and denying women the vote—but they also provided the tools for their descendants to correct these mistakes. That was their greatest legacy.

    THE TWENTIETH CENTURY: MASS DEMOCRACY, BUT ALSO MASS TOTALITARIANISM

    Printed newspapers were just the first harbinger of the mass media age. During the nineteenth and twentieth centuries, a long list of new communication and transportation technologies such as the telegraph, telephone, television, radio, trains, steamships, and airplanes supercharged the power of mass media.

    When Demosthenes gave a public speech in Athens around 350 BCE, it was aimed primarily at the limited audience actually present in the Athenian agora. When John Quincy Adams gave his First Annual Message in 1825, his words spread at the pace of a horse. When Abraham Lincoln gave his Gettysburg Address on November 19, 1863, telegraphs, locomotives, and steamships conveyed his words much faster throughout the Union and beyond. The very next day The New York Times had already reprinted the speech in full,52 as had numerous other newspapers from The Portland Daily Press in Maine to the Ottumwa Courier in Iowa.53

    As befitting a democracy with strong self-correcting mechanisms in place, the president’s speech sparked a lively conversation rather than universal applause. Most newspapers lauded it, but some expressed their doubts. The Chicago Times wrote on November 20 that “the cheek of every American must tingle with shame as he reads the silly, flat and dishwatery utterances” of President Lincoln.54 The Patriot & Union, a local newspaper in Harrisburg, Pennsylvania, also blasted “the silly remarks of the President” and hoped that “the veil of oblivion shall be dropped over them and that they shall be no more repeated or thought of.”55 Though the country was in the midst of a civil war, journalists were free to publicly criticize—and even ridicule—the president.

    Fast-forward a century, and things really picked up speed. For the first time in history, new technologies allowed masses of people, spread over vast swaths of territory, to connect in real time. In 1960, about seventy million Americans (39 percent of the total population), dispersed over the North American continent and beyond, watched the Nixon-Kennedy presidential debates live on television, with millions more listening on the radio.56 The only effort viewers and listeners had to make was to press a button while sitting in their homes. Large-scale democracy had now become feasible. Millions of people separated by thousands of kilometers could conduct informed and meaningful public debates about the rapidly evolving issues of the day. By 1960, all adult Americans were theoretically eligible to vote, and close to seventy million (about 64 percent of the electorate) actually did so—though millions of Blacks and other disenfranchised groups were prevented from voting through various voter-suppression schemes.57

    As always, we should beware of technological determinism and of concluding that the rise of mass media led to the rise of large-scale democracy. Mass media made large-scale democracy possible, rather than inevitable. And it also made possible other types of regimes. In particular, the new information technologies of the modern age opened the door for large-scale totalitarian regimes. Like Nixon and Kennedy, Stalin and Khrushchev could say something over the radio and be heard instantaneously by hundreds of millions of people from Vladivostok to Kaliningrad. They could also receive daily reports by phone and telegraph from millions of secret police agents and informers. If a newspaper in Vladivostok or Kaliningrad wrote that the supreme leader’s latest speech was silly (as happened to Lincoln’s Gettysburg Address), then everyone involved—from the editor in chief to the typesetters—would likely have received a visit from the KGB.

    A BRIEF HISTORY OF TOTALITARIANISM

    Totalitarian systems assume their own infallibility, and seek total control over the totality of people’s lives. Before the invention of the telegraph, radio, and other modern information technology, large-scale totalitarian regimes were impossible. Roman emperors, Abbasid caliphs, and Mongol khans were often ruthless autocrats who believed they were infallible, but they did not have the apparatus necessary to impose totalitarian control over large societies. To understand this, we should first clarify the difference between totalitarian regimes and less extreme autocratic regimes. In an autocratic network, there are no legal limits on the will of the ruler, but there are nevertheless a lot of technical limits. In a totalitarian network, many of these technical limits are absent.58

    For example, in autocratic regimes like the Roman Empire, the Abbasid Empire, and the Mongol Empire, rulers could usually execute any person who displeased them, and if some law got in their way, they could ignore or change the law. The emperor Nero arranged the murder of his mother, Agrippina, and his wife, Octavia, and forced his mentor Seneca to commit suicide. Nero also executed or exiled some of the most respected and powerful Roman aristocrats merely for voicing dissent or telling jokes about him.59

    While autocratic rulers like Nero could execute anyone who did or said something that displeased them, they couldn’t know what most people in their empire were doing or saying. Theoretically, Nero could issue an order that any person in the Roman Empire who criticized or insulted the emperor must be severely punished. Yet there were no technical means for implementing such an order. Roman historians like Tacitus portray Nero as a bloodthirsty tyrant who instigated an unprecedented reign of terror. But this was a very limited type of terror. Although he executed or exiled a number of family members, aristocrats, and senators within his orbit, ordinary Romans in the city’s slums and provincials in distant towns like Jerusalem and Londinium could speak their mind much more freely.60

    Modern totalitarian regimes like the Stalinist U.S.S.R. instigated terror on an altogether different scale. Totalitarianism is the attempt to control what every person throughout the country is doing and saying every moment of the day, and potentially even what every person is thinking and feeling. Nero might have dreamed about such powers, but he lacked the means to realize them. Given the limited tax base of the agrarian Roman economy, Nero couldn’t employ many people in his service. He could place informers at the dinner parties of Roman senators, but he had only about 10,000 imperial administrators61 and 350,000 soldiers62 to control the rest of the empire, and he lacked the technology to communicate with them swiftly.

    Nero and his fellow emperors had an even bigger problem ensuring the loyalty of the administrators and soldiers they did have on their payroll. No Roman emperor was ever toppled by a democratic revolution like the ones that deposed Louis XVI, Nicolae Ceauşescu, or Hosni Mubarak. Instead, dozens of emperors were assassinated or deposed by their own generals, officials, bodyguards, or family members.63 Nero himself was overthrown by a revolt of the governor of Hispania, Galba. Six months later Galba was ousted by Otho, the governor of Lusitania. Within three months, Otho was deposed by Vittelius, commander of the Rhine army. Vitellius lasted about eight months before he was defeated and killed by Vespasian, commander of the army in Judaea. Being killed by a rebellious subordinate was the biggest occupational hazard not just for Roman emperors but for almost all premodern autocrats.

    Emperors, caliphs, shahs, and kings found it a huge challenge to keep their subordinates in check. Rulers consequently focused their attention on controlling the military and the taxation system. Roman emperors had the authority to interfere in the local affairs of any province or city, and they sometimes exercised that authority, but this was usually done in response to a specific petition sent by a local community or official,64 rather than as part of some empire-wide totalitarian Five-Year Plan. If you were a mule driver in Pompeii or a shepherd in Roman Britain, Nero didn’t want to control your daily routines or to police the jokes you told. As long as you paid your taxes and didn’t resist the legions, that was good enough for Nero.

    SPARTA AND QIN

    Some scholars claim that despite the technological difficulties there were attempts to establish totalitarian regimes in ancient times. The most common example cited is Sparta. According to this interpretation, Spartans were ruled by a totalitarian regime that micromanaged every aspect of their lives—from whom they married to what they ate. However, while the Spartan regime was certainly draconian, it actually included several self-correcting mechanisms that prevented power from being monopolized by a single person or faction. Political authority was divided between two kings, five ephors (senior magistrates), twenty-eight members of the Gerousia council, and the popular assembly. Important decisions—such as whether to go to war—often involved fierce public debates.

    Moreover, irrespective of how we evaluate the nature of Sparta’s regime, it is clear that the same technological limitations that confined ancient Athenian democracy to a single city also limited the scope of the Spartan political experiment. After winning the Peloponnesian War, Sparta installed military garrisons and pro-Spartan governments in numerous Greek cities, requiring them to follow its lead in foreign policy and sometimes also pay tribute. But unlike the U.S.S.R. after World War II, Sparta after the Peloponnesian War did not try to expand or export its system. Sparta couldn’t construct an information network big and dense enough to control the lives of ordinary people in every Greek town and village.65

    A much more ambitious totalitarian project might have been launched by the Qin dynasty in ancient China (221–206 BCE). After defeating all the other Warring States, the Qin ruler Qin Shi Huang controlled a huge empire with tens of millions of subjects, who belonged to numerous different ethnic groups, spoke diverse languages, and were loyal to various local traditions and elites. To cement its power, the victorious Qin regime tried to dismantle any regional powers that might challenge its authority. It confiscated the lands and wealth of local aristocrats and forced regional elites to move to the imperial capital of Xiangyang, thereby separating them from their power base and monitoring them more easily.

    The Qin regime also embarked on a ruthless campaign of centralization and homogenization. It created a new simplified script to be used throughout the empire and standardized coinage, weights, and measurements. It built a road network radiating out of Xiangyang, with standardized rest houses, relay stations, and military checkpoints. People needed written permits in order to enter or leave the capital region or frontier zones. Even the width of axles was standardized to ensure that carts and chariots could run in the same ruts.

    Every action, from tilling fields to getting married, was supposed to serve some military need, and the type of military discipline that Rome reserved for the legions was imposed by the Qin on the entire population. The envisioned reach of this system can be exemplified by one Qin law that specified the punishment an official faced if he neglected a granary under his supervision. The law discusses the number of rat holes in the granary that would warrant fining or berating the official. “For three or more rat holes the fine is [the purchase of] one shield [for the army] and for two or fewer [the responsible official] is berated. Three mouse holes are equal to one rat hole.”66

    To facilitate this totalitarian system, the Qin attempted to create a militarized social order. Every male subject had to belong to a five-man unit. These units were aggregated into larger formations, from local hamlets (li), through cantons (xiang) and counties (xian), all the way to the large imperial commanderies (jun). People were forbidden to change their residence without permit, to the extent that guests could not even stay overnight at a friend’s house without proper identification and authorization.

    Every Qin male subject was also given a rank, just as every soldier in an army has a rank. Obedience to the state resulted in promotion to higher ranks, which brought with it economic and legal privileges, while disobedience could result in demotion or punishment. People in each formation were supposed to supervise one another, and if any individual committed some misdeed, all could be punished for it. Anyone who failed to report a criminal—even their own relatives—would be killed. Those who reported crimes were rewarded with higher ranks and other perks.

    It is highly questionable to what extent the regime managed to implement all these totalitarian measures. Bureaucrats writing documents in a government office often invent elaborate rules and regulations, which then turn out to be impractical. Did conscientious government officials really go around the entire Qin Empire counting rat holes in every granary? Were peasants in every remote mountain hamlet really organized into five-man squads? Probably not. Nevertheless, the Qin Empire outdid other ancient empires in its totalitarian ambitions.

    The Qin regime even tried to control what its subjects were thinking and feeling. During the Warring States period Chinese thinkers were relatively free to develop myriad ideologies and philosophies, but the Qin adopted the doctrine of Legalism as the official state ideology. Legalism posited that humans were naturally greedy, cruel, and egotistical. It emphasized the need for strict control, argued that punishments and rewards were the most effective means of control, and insisted that state power not be curtailed by any moral consideration. Might was right, and the good of the state was the supreme good.67 The Qin proscribed other philosophies, such as Confucianism and Daoism, which believed humans were more altruistic and which emphasized the importance of virtue rather than violence.68 Books espousing such soft views were banned, as well as books that contradicted the official Qin version of history.

    When one scholar argued that Qin Shi Huang should emulate the founder of the ancient Zhou dynasty and decentralize state power, the Qin chief minister, Li Si, countered that scholars should stop criticizing present-day institutions by idealizing the past. The regime ordered the confiscation of all books that romanticized antiquity or otherwise criticized the Qin. Such problematic texts were stored in the imperial library and could be studied only by official scholars.69

    The Qin Empire was probably the most ambitious totalitarian experiment in human history prior to the modern age, and its scale and intensity would prove to be its ruin. The attempt to regiment tens of millions of people along military lines, and to monopolize all resources for military purposes, led to severe economic problems, wastefulness, and popular resentment. The regime’s draconian laws, along with its hostility to regional elites and its voracious appetite for taxes and recruits, fanned the flames of this resentment even further. Meanwhile, the limited resources of an ancient agrarian society couldn’t support all the bureaucrats and soldiers that the Qin needed to contain this resentment, and the low efficiency of their information technology made it impossible to control every town and village from distant Xiangyang. Not surprisingly, in 209 BCE a series of revolts broke out, led by regional elites, disgruntled commoners, and even some of the empire’s own newly minted officials.

    According to one account, the first serious revolt started when a group of conscripted peasants sent to work in a frontier zone were delayed by rain and flooding. They feared they would be executed for this dereliction of duty, and felt they had nothing to lose. They were quickly joined by numerous other rebels. Just fifteen years after reaching the apogee of power, the Qin Empire collapsed under the weight of its totalitarian ambitions, splintering into eighteen kingdoms.

    After several years of war, a new dynasty—the Han—reunited the empire. But the Han then adopted a more realistic, less draconian attitude. Han emperors were certainly autocratic, but they were not totalitarian. They did not recognize any limits on their authority, but they did not try to micromanage everyone’s lives. Instead of following Legalist ideas of surveillance and control, the Han turned to Confucian ideas of encouraging people to act loyally and responsibly out of inner moral convictions. Like their contemporaries in the Roman Empire, Han emperors sought to control only some aspects of society from the center, while leaving considerable autonomy to provincial aristocrats and local communities. Due largely to the limitations imposed by the available information technology, premodern large-scale polities like the Roman and Han Empires gravitated toward nontotalitarian autocracy.70 Full-blown totalitarianism might have been dreamed about by the likes of the Qin, but its implementation had to wait for the development of modern technology.

    THE TOTALITARIAN TRINITY

    Just as modern technology enabled large-scale democracy, it also made large-scale totalitarianism possible. Beginning in the nineteenth century, the rise of industrial economies allowed governments to employ many more administrators, and new information technologies—such as the telegraph and radio—made it possible to quickly connect and supervise all these administrators. This facilitated an unprecedented concentration of information and power, for those who dreamed about such things.

    When the Bolsheviks seized control of Russia after the 1917 revolution, they were driven by exactly such a dream. The Bolsheviks craved unlimited power because they believed they had a messianic mission. Marx taught that for millennia, all human societies were dominated by corrupt elites who oppressed the people. The Bolsheviks claimed they knew how to finally end all oppression and create a perfectly just society on earth. But to do so, they had to overcome numerous enemies and obstacles, which, in turn, required all the power they could get. They refused to countenance any self-correcting mechanisms that might question either their vision or their methods. Like the Catholic Church, the Bolshevik Party was convinced that though its individual members might err, the party itself was always right. Belief in their own infallibility led the Bolsheviks to destroy Russia’s nascent democratic institutions—like elections, independent courts, the free press, and opposition parties—and to create a one-party totalitarian regime. Bolshevik totalitarianism did not start with Stalin. It was evident from the very first days of the revolution. It stemmed from the doctrine of party infallibility, rather than from the personality of Stalin.

    In the 1930s and 1940s, Stalin perfected the totalitarian system he inherited. The Stalinist network was composed of three main branches. First, there was the governmental apparatus of state ministries, regional administrations, and regular Red Army units, which in 1939 comprised 1.6 million civilian officials71 and 1.9 million soldiers.72 Second, there was the apparatus of the Communist Party of the Soviet Union and its ubiquitous party cells, which in 1939 included 2.4 million party members.73 Third, there was the secret police: first known as the Cheka, in Stalin’s days it was called the OGPU, NKVD, and MGB, and after Stalin’s death it morphed into the KGB. Its post-Soviet successor organization is known since 1995 as the FSB. In 1937, the NKVD had 270,000 agents and millions of informers.74

    The three branches operated in parallel. Just as democracy is maintained by having overlapping self-correcting mechanisms that keep each other in check, modern totalitarianism created overlapping surveillance mechanisms that keep each other in order. The governor of a Soviet province was constantly watched by the local party commissar, and neither of them knew who among their staff was an NKVD informer. A testimony to the effectiveness of the system is that modern totalitarianism largely solved the perennial problem of premodern autocracies—revolts by provincial subordinates. While the U.S.S.R. had its share of court coups, not once did a provincial governor or a Red Army front commander rebel against the center.75 Much of the credit for that goes to the secret police, which kept a close eye on the mass of citizens, on provincial administrators, and even more so on the party and the Red Army.

    While in most polities throughout history the army had wielded enormous political power, in twentieth-century totalitarian regimes the regular army ceded much of its clout to the secret police—the information army. In the U.S.S.R., the Cheka, OGPU, NKVD, and KGB lacked the firepower of the Red Army, but had more influence in the Kremlin and could terrorize and purge even the army brass. The East German Stasi and the Romanian Securitate were similarly stronger than the regular armies of these countries.76 In Nazi Germany, the SS was more powerful than the Wehrmacht, and the SS chief, Heinrich Himmler, was higher up the pecking order than Wilhelm Keitel, chief of the Wehrmacht high command.

    In none of these cases could the secret police defeat the regular army in traditional warfare, of course; what made the secret police powerful was its command of information. It had the information necessary to preempt a military coup and to arrest the commanders of tank brigades or fighter squadrons before they knew what hit them. During the Stalinist Great Terror of the late 1930s, out of 144,000 Red Army officers about 10 percent were shot or imprisoned by the NKVD. This included 154 of 186 divisional commanders (83 percent), eight of nine admirals (89 percent), thirteen of fifteen full generals (87 percent), and three of five marshals (60 percent).77

    The party leadership fared just as badly. Of the revered Old Bolsheviks, people who joined the party before the 1917 revolution, about a third didn’t survive the Great Terror.78 Of the thirty-three men who served on the Politburo between 1919 and 1938, fourteen were shot (42 percent). Of the 139 members and candidate members of the party’s Central Committee in 1934, 98 (70 percent) were shot. Only 2 percent of the delegates who took part in the Seventeenth Party Congress in 1934 evaded execution, imprisonment, expulsion, or demotion, and attended the Eighteenth Party Congress in 1939.79

    The secret police—which did all the purging and killing—was itself divided into several competing branches that closely watched and purged each other. Genrikh Yagoda, the NKVD head who orchestrated the beginning of the Great Terror and supervised the killing of hundreds of thousands of victims, was executed in 1938 and replaced by Nikolai Yezhov. Yezhov lasted for two years, killing and imprisoning millions of people before being executed in 1940.

    Perhaps most telling is the fate of the thirty-nine people who in 1935 held the rank of general in the NKVD (called commissars of state security in Soviet nomenclature). Thirty-five of them (90 percent) were arrested and shot by 1941, one was assassinated, and one—the head of the NKVD’s Far East regional office—saved himself by defecting to Japan, but was killed by the Japanese in 1945. Of the original cohort of thirty-nine NKVD generals, only two men were left standing by the end of World War II. The remorseless logic of totalitarianism eventually caught up with them too. During the power struggles that followed Stalin’s death in 1953, one of them was shot, while the other was consigned to a psychiatric hospital, where he died in 1960.80 Serving as an NKVD general in Stalin’s day was one of the most dangerous jobs in the world. At a time when American democracy was improving its many self-correcting mechanisms, Soviet totalitarianism was refining its triple self-surveilling and self-terrorizing apparatus.

    TOTAL CONTROL

    Totalitarian regimes are based on controlling the flow of information and are suspicious of any independent channels of information. When military officers, state officials, or ordinary citizens exchange information, they can build trust. If they come to trust each other, they can organize resistance to the regime. Therefore, a key tenet of totalitarian regimes is that wherever people meet and exchange information, the regime should be there too, to keep an eye on them. In the 1930s, this was one principle that Hitler and Stalin shared.

    On March 31, 1933, two months after Hitler became chancellor, the Nazis passed the Coordination Act (Gleichschaltungsgesetz). This stipulated that by April 30, 1933, all political, social, and cultural organizations throughout Germany—from municipalities to football clubs and local choirs—must be run according to Nazi ideology, as organs of the Nazi state. It upended life in every city and hamlet in Germany.

    For example, in the small Alpine village of Oberstdorf, the democratically elected municipal council met for the last time on April 21, 1933, and three days later it was replaced by an unelected Nazi council that appointed a Nazi mayor. Since the Nazis alone allegedly knew what the people really wanted, who other than Nazis could implement the people’s will? Oberstdorf also had about fifty associations and clubs, ranging from a beekeeping society to an alpinist club. They all had to conform to the Coordination Act, adjusting their boards, membership, and statutes to Nazi demands, hoisting the swastika flag, and concluding every meeting with the “Horst Wessel Song,” the Nazi Party’s anthem. On April 6, 1933, the Oberstdorf fishing society banned Jews from its ranks. None of the thirty-two members was Jewish, but they felt they had to prove their Aryan credentials to the new regime.81

    Things were even more extreme in Stalin’s U.S.S.R. Whereas the Nazis still allowed church organizations and private businesses some partial freedom of action, the Soviets made no exceptions. By 1928 and the launch of the first Five-Year Plan, there were government officials, party functionaries, and secret police informants in every neighborhood and village, and between them they controlled every aspect of life: all businesses from power plants to cabbage farms; all newspapers and radio stations; all universities, schools, and youth groups; all hospitals and clinics; all voluntary and religious organizations; all sporting and scientific associations; all parks, museums, and cinemas.

    If a dozen people came together to play football, hike in the woods, or do some charity work, the party and the secret police had to be there too, represented by the local party cell or NKVD agent. The speed and efficiency of modern information technology meant that all these party cells and NKVD agents were always just a telegram or phone call away from Moscow. Information about suspicious persons and activities was fed into a countrywide, cross-referenced system of card catalogs. Known as kartoteki, these catalogs contained information from work records, police files, residence cards, and other forms of social registrations and, by the 1930s, had become the primary mechanism for surveilling and controlling the Soviet population.82

    This made it feasible for Stalin to seek control over the totality of Soviet life. One crucial example was the campaign to collectivize Soviet farming. For centuries, economic, social, and private life in the thousands of villages of the sprawling Tsarist Empire was managed by several traditional institutions: the local commune, the parish church, the private farm, the local market, and above all the family. In the mid-1920s, the Soviet Union was still an overwhelmingly agrarian economy. About 82 percent of the total population lived in villages, and 83 percent of the workforce was engaged in farming.83 But if each peasant family made its own decisions about what to grow, what to buy, and how much to charge for their produce, it greatly limited the ability of Moscow officials to themselves plan and control social and economic activities. What if the officials decided on a major agrarian reform, but the peasant families rejected it? So when in 1928 the Soviets came up with their first Five-Year Plan for the development of the Soviet Union, the most important item on the agenda was to collectivize farming.

    The idea was that in every village all the families would join a kolkhoz—a collective farm. They would hand over to the kolkhoz all their property—land, houses, horses, cows, shovels, pitchforks. They would work together for the kolkhoz, and in return the kolkhoz would provide for all their needs, from housing and education to food and health care. The kolkhoz would also decide—based on orders from Moscow—whether they should grow cabbages or turnips; whether to invest in a tractor or a school; and who would work in the dairy farm, the tannery, and the clinic. The result, thought the Moscow masterminds, would be the first perfectly just and equal society in human history.

    They were similarly convinced of the economic advantages of their proposed system, thinking that the kolkhoz would enjoy economy of scale. For example, when every peasant family had but a small strip of land, it made little sense to buy a tractor to plow it, and in any case most families couldn’t afford a tractor. Once all land was held communally, it could be cultivated far more efficiently using modern machinery. In addition, the kolkhoz was supposed to benefit from the wisdom of modern science. Instead of every peasant deciding on production methods on the basis of old traditions and groundless superstitions, state experts with university degrees from institutions like the Lenin All-Union Academy of Agricultural Sciences would make the crucial decisions.

    To the planners in Moscow, it sounded wonderful. They expected a 50 percent increase in agricultural production by 1931.84 And if in the process the old village hierarchies and inequalities were bulldozed, all the better. To most peasants, however, it sounded terrible. They didn’t trust the Moscow planners or the new kolkhoz system. They did not want to give up their old way of life or their private property. Villagers slaughtered cows and horses instead of handing them to the kolkhoz. Their motivation to work dwindled. People made less effort plowing fields that belonged to everyone than plowing fields that belonged to their own family. Passive resistance was ubiquitous, sometimes flaring into violent clashes. Whereas Soviet planners expected to harvest ninety-eight million tons of grain in 1931, production was only sixty-nine million, according to official data, and might have been as low as fifty-seven million tons in reality. The 1932 harvest was even worse.85

    The state reacted with fury. Between 1929 and 1936, food confiscation, government neglect, and man-made famines (resulting from government policy rather than a natural disaster) claimed the lives of between 4.5 and 8.5 million people.86 Millions of additional peasants were declared enemies of the state and deported or imprisoned. The most basic institutions of peasant life—the family, the church, the local community—were terrorized and dismantled. In the name of justice, equality, and the will of the people, the collectivization campaign annihilated anything that stood in its way. In the first two months of 1930 alone, about 60 million peasants in more than 100,000 villages were herded into collective farms.87 In June 1929, only 4 percent of Soviet peasant households had belonged to collective farms. By March 1930 the figure had risen to 57 percent. By April 1937, 97 percent of households in the countryside had been confined to the 235,000 Soviet collective farms.88 In just seven years, then, a way of life that had existed for centuries had been replaced by the totalitarian brainchild of a few Moscow bureaucrats.

    THE KULAKS

    It is worthwhile to delve a little deeper into the history of Soviet collectivization. For it was a tragedy that bears some resemblance to earlier catastrophes in human history—like the European witch-hunt craze—and at the same time foreshadows some of the biggest dangers posed by twenty-first-century technology and its faith in supposedly scientific data.

    When their efforts to collectivize farming encountered resistance and led to economic disaster, Moscow bureaucrats and mythmakers took a page from Kramer’s Hammer of the Witches. I don’t wish to imply the Soviets actually read the book, but they too invented a global conspiracy and created an entire non-existing category of enemies. In the 1930s Soviet authorities repeatedly blamed the disasters afflicting the Soviet economy on a counterrevolutionary cabal whose chief agents were the “kulaks” or “capitalist farmers.” Just as in Kramer’s imagination witches serving Satan conjured hailstorms that destroyed crops, so in the Stalinist imagination kulaks beholden to global capitalism sabotaged the Soviet economy.

    In theory, kulaks were an objective socioeconomic category, defined by analyzing empirical data on things like property, income, capital, and wages. Soviet officials could allegedly identify kulaks by counting things. If most people in a village had only one cow, then the few families who had three cows were considered kulaks. If most people in a village didn’t hire any labor, but one family hired two workers during harvest time, this was a kulak family. Being a kulak meant not only that you possessed a certain amount of property but also that you possessed certain personality traits. According to the supposedly infallible Marxist doctrine, people’s material conditions determined their social and spiritual character. Since kulaks allegedly engaged in capitalist exploitation, it was a scientific fact (according to Marxist thinking) that they were greedy, selfish, and unreliable—and so were their children. Discovering that someone was a kulak ostensibly revealed something profound about their fundamental nature.

    On December 27, 1929, Stalin declared that the Soviet state should seek “the liquidation of the kulaks as a class,”89 and immediately galvanized the party and the secret police to realize that ambitious and murderous aim. Early modern European witch-hunters worked in autocratic societies that lacked modern information technology; therefore, it took them three centuries to kill fifty thousand alleged witches. In contrast, Soviet kulak hunters were working in a totalitarian society that had at its disposal technologies such as telegraphs, trains, telephones, and radios—as well as a sprawling bureaucracy. They decided that two years would suffice to “liquidate” millions of kulaks.90

    Soviet officials began by assessing how many kulaks there must be in the U.S.S.R. Based on existing data—such as tax records, employment records, and the 1926 Soviet census—they decided that kulaks constituted 3–5 percent of the rural population.91 On January 30, 1930, just one month after Stalin’s speech, a Politburo decree translated his vague vision into a much more detailed plan of action. The decree included target numbers for the liquidation of kulaks in each major agricultural region.92 Regional authorities then made their own estimates of the number of kulaks in each county under their jurisdiction. Eventually, specific quotas were assigned to rural soviets (local administrative units, typically comprising a handful of villages). Often, local officials inflated the numbers along the way, to prove their zeal. Each rural soviet then had to identify the stated number of kulak households in the villages under its purview. These people were expelled from their homes, and—according to the administrative category to which they belonged—resettled elsewhere, incarcerated in concentration camps, or condemned to death.93

    How exactly did Soviet officials tell who was a kulak? In some villages, local party members made a conscientious effort to identify kulaks by objective measures, such as the amount of property they owned. It was often the most hardworking and efficient farmers who were stigmatized and expelled. In some villages local communists used the opportunity to get rid of their personal enemies. Some villages simply drew lots on who would be considered a kulak. Other villages held communal meetings to vote on the matter and often chose isolated farmers, widows, old people, and other “expendables” (exactly the sorts of people who in early modern Europe were most likely to be branded witches).94

    The absurdity of the entire operation is manifested in the case of the Streletsky family from the Kurgan region of Siberia. Dmitry Streletsky, who was then a teenager, recalled years later how his family was branded kulaks and selected for liquidation. “Serkov, the chairman of the village Soviet who deported us, explained: ‘I have received an order [from the district party committee] to find 17 kulak families for deportation. I formed a Committee of the Poor and we sat through the night to choose the families. There is no one in the village who is rich enough to qualify, and not many old people, so we simply chose the 17 families. You were chosen. Please don’t take it personally. What else could I do?’ ”95 If anyone dared object to the madness of the system, they were promptly denounced as kulaks and counterrevolutionaries and would themselves be liquidated.

    Altogether, some five million kulaks would be expelled from their homes by 1933. As many as thirty thousand heads of households were shot. The more fortunate victims were resettled in their district of origin or became vagrant workers in the big cities, while about two million were either exiled to remote inhospitable regions or incarcerated as state slaves in labor camps.96 Numerous important and notorious state projects—such as the construction of the White Sea Canal and the development of mines in the Arctic regions—were accomplished with the labor of millions of prisoners, many of them kulaks. It was one of the fastest and largest enslavement campaigns in human history.97 Once branded a kulak, a person could not get rid of the stigma. Government agencies, party organs, and secret police documents recorded who was a kulak in a labyrinthian system of kartoteki catalogs, archives, and internal passports.

    Kulak status even passed to the next generation, with devastating consequences. Kulak children were refused entrance to communist youth groups, the Red Army, universities, and prestigious areas of employment.98 In her 1997 memoirs, Antonina Golovina recalled how her family was deported from their ancestral village as kulaks and sent to live in the town of Pestovo. The boys in her new school regularly taunted her. On one occasion, a senior teacher told the eleven-year-old Antonina to stand up in front of all the other children, and began abusing her mercilessly, shouting that “her sort were enemies of the people, wretched kulaks! You certainly deserved to be deported, I hope you’re all exterminated!” Antonina wrote that this was the defining moment of her life. “I had this feeling in my gut that we [kulaks] were different from the rest, that we were criminals.” She never got over it.99

    Like the ten-year-old “witch” Hansel Pappenheimer, the eleven-year-old “kulak” Antonina Golovina found herself cast into an intersubjective category invented by human mythmakers and imposed by ubiquitous bureaucrats. The mountains of information collected by Soviet bureaucrats about the kulaks wasn’t the objective truth about them, but it imposed a new intersubjective Soviet truth. Knowing that someone was a kulak was one of the most important things to know about a Soviet person, even though the label was entirely bogus.

    ONE BIG HAPPY SOVIET FAMILY

    The Stalinist regime would go on to attempt something even more ambitious than the mass dismantling of private family farms. It set out to dismantle the family itself. Unlike Roman emperors or Russian tsars, Stalin tried to insert himself even into the most intimate human relationships, coming between parents and children. Family ties were considered the bedrock of corruption, inequality, and antiparty activities. Soviet children were therefore taught to worship Stalin as their real father and to inform on their biological parents if they criticized Stalin or the Communist Party.

    Starting in 1932, the Soviet propaganda machine created a veritable cult around the figure of Pavlik Morozov—a thirteen-year-old boy from the Siberian village of Gerasimovka. In autumn 1931, Pavlik informed the secret police that his father, Trofim—the chairman of the village soviet—was selling false papers to kulak exiles. During the subsequent trial, when Trofim shouted to Pavlik, “It’s me, your father,” the boy retorted, “Yes, he used to be my father, but I no longer consider him my father.” Trofim was sent to a labor camp and later shot. In September 1932, Pavlik was found murdered, and Soviet authorities arrested and executed five of his family members, who allegedly killed him in revenge for the denunciation. The real story was far more complicated, but it didn’t matter to the Soviet press. Pavlik became a martyr, and millions of Soviet children were taught to emulate him.100 Many did.

    For example, in 1934 a thirteen-year-old boy called Pronia Kolibin told the authorities that his hungry mother stole grain from the kolkhoz fields. His mother was arrested and presumably shot. Pronia was rewarded with a cash prize and a lot of positive media attention. The party organ Pravda published a poem Pronia wrote. Two of its lines read, “You are a wrecker, Mother / I can live with you no more.”101

    The Soviet attempt to control the family was reflected in a dark joke told in Stalin’s day. Stalin visits a factory undercover, and conversing with a worker, he asks the man, “Who is your father?”
    “Stalin,” replies the worker.
    “Who is your mother?”
    “The Soviet Union,” the man responds.
    “And what do you want to be?”
    “An orphan.”102

    At the time you could easily lose your liberty or your life for telling this joke, even if you told it in your own home to your closest family members. The most important lesson Soviet parents taught their children wasn’t loyalty to the party or to Stalin. It was “keep your mouth shut.”103 Few things in the Soviet Union were as dangerous as holding an open conversation.

    PARTY AND CHURCH

    You may wonder whether modern totalitarian institutions like the Nazi Party or the Soviet Communist Party were really all that different from earlier institutions like the Christian churches. After all, churches too believed in their infallibility, had priestly agents everywhere, and sought to control the daily life of people down to their diet and sexual habits. Shouldn’t we see the Catholic Church or the Eastern Orthodox Church as totalitarian institutions? And doesn’t this undermine the thesis that totalitarianism was made possible only by modern information technology?

    There are, however, several major differences between modern totalitarianism and premodern churches. First, as noted earlier, modern totalitarianism has worked by deploying several overlapping surveillance mechanisms that keep each other in order. The party is never alone; it works alongside state organs, on the one side, and the secret police, on the other. In contrast, in most medieval European kingdoms the Catholic Church was an independent institution that often clashed with the state institutions instead of reinforcing them. Consequently, the church was perhaps the most important check on the power of European autocrats.

    For example, when in the “Investiture Controversy” of the 1070s the emperor Henry IV asserted that as emperor he had the final say on the appointment of bishops, abbots, and other important church officials, Pope Gregory VII mobilized resistance and eventually forced the emperor to surrender. On January 25, 1077, Henry reached Canossa castle, where the pope was lodging, to offer his submission and apology. The pope refused to open the gates, and Henry waited in the snow outside, barefoot and hungry. After three days, the pope finally opened the gates to the emperor, who begged forgiveness.104

    An analogous clash in a modern totalitarian country is unthinkable. The whole idea of totalitarianism is to prevent any separation of powers. In the Soviet Union, state and party reinforced each other, and Stalin was the de facto head of both. There could be no Soviet “Investiture Controversy,” because Stalin had final say about all appointments to both party positions and state functions. He decided both who would be general secretary of the Communist Party of Georgia and who would be foreign minister of the Soviet Union.

    Another important difference is that medieval churches tended to be traditionalist organizations that resisted change, while modern totalitarian parties have tended to be revolutionary organizations demanding change. A premodern church built its power gradually by developing its structure and traditions over centuries. A king or a pope who wanted to swiftly revolutionize society was therefore likely to encounter stiff resistance from church members and ordinary believers.

    For example, in the eighth and ninth centuries a series of Byzantine emperors sought to forbid the veneration of icons, which seemed to them idolatrous. They pointed to many passages in the Bible, most notably the Second Commandment, that forbade making any graven images. While Christian churches traditionally interpreted the Second Commandment in a way that allowed the veneration of icons, emperors like Constantine V argued that this was a mistake and that disasters like Christian defeats by the armies of Islam were due to God’s wrath over the worship of icons. In 754 more than three hundred bishops assembled in the Council of Hieria to support Constantine’s iconoclastic position.

    Compared with Stalin’s collectivization campaign, this was a minor reform. Families and villages were required to give up their icons, but not their private property or their children. Yet Byzantine iconoclasm met with widespread resistance. Unlike the participants in the Council of Hieria, many ordinary priests, monks, and believers were deeply attached to their icons. The resulting struggle ripped apart Byzantine society until the emperors conceded defeat and reversed course.105 Constantine V was later vilified by Byzantine historians as “Constantine the Shitty” (Koprónimos), and a story was spread about him that he defecated during his baptism.106

    Unlike premodern churches, which developed slowly over many centuries and therefore tended to be conservative and suspicious of rapid changes, modern totalitarian parties like the Nazi Party and the Soviet Communist Party were organized within a single generation around the promise to quickly revolutionize society. They didn’t have centuries-old traditions and structures to defend. When their leaders conceived some ambitious plan to smash existing traditions and structures, party members typically fell in line.

    Perhaps most important of all, premodern churches could not become tools of totalitarian control because they themselves suffered from the same limitations as all other premodern organizations. While they had local agents everywhere, in the shape of parish priests, monks, and itinerant preachers, the difficulty of transmitting and processing information meant that church leaders knew little about what was going on in remote communities, and local priests had a large degree of autonomy. Consequently, churches tended to be local affairs. People in every province and village often venerated local saints, upheld local traditions, performed local rites, and might even have had local doctrinal ideas that differed from the official line.107 If the pope in Rome wanted to do something about an independent-minded priest in a remote Polish parish, he had to send a letter to the archbishop of Gniezno, who had to instruct the relevant bishop, who had to send someone to intervene in the parish. That might take months, and there was ample opportunity for the archbishop, bishop, and other intermediaries to reinterpret or even “mislay” the pope’s orders.108

    Churches became more totalitarian institutions only in the late modern era, when modern information technologies became available. We tend to think of popes as medieval relics, but actually they are masters of modern technology. In the eighteenth century, the pope had little control over the worldwide Catholic Church and was reduced to the status of a local Italian princeling, fighting other Italian powers for control of Bologna or Ferrara. With the advent of radio, popes became some of the most powerful people on the planet. Pope John Paul II could sit in the Vatican and speak directly to millions of Catholics from Poland to the Philippines, without any archbishop, bishop, or parish priest able to twist or hide his words.109

    HOW INFORMATION FLOWS

    We see then that the new information technology of the late modern era gave rise to both large-scale democracy and large-scale totalitarianism. But there were crucial differences between how the two systems used information technology. As noted earlier, democracy encourages information to flow through many independent channels rather than only through the center, and it allows many independent nodes to process the information and make decisions by themselves. Information freely circulates between private businesses, private media organizations, municipalities, sports associations, charities, families, and individuals—without ever passing through the office of a government minister.

    In contrast, totalitarianism wants all information to pass through the central hub and doesn’t want any independent institutions making decisions on their own. True, totalitarianism does have its tripartite apparatus of government, party, and secret police. But the whole point of this parallel apparatus is to prevent the emergence of any independent power that might challenge the center. When government officials, party members, and secret police agents constantly keep tabs on one another, opposing the center is extremely dangerous.

    As contrasting types of information networks, democracy and totalitarianism both have their advantages and disadvantages. The biggest advantage of the centralized totalitarian network is that it is extremely orderly, which means it can make decisions quickly and enforce them ruthlessly. Especially during emergencies like wars and epidemics, centralized networks can move much faster and further than distributed networks.

    But hyper-centralized information networks also suffer from several big disadvantages. Since they don’t allow information to flow anywhere except through the official channels, if the official channels are blocked, the information cannot find an alternative means of transmission. And official channels are often blocked.

    One common reason why official channels might be blocked is that fearful subordinates hide bad news from their superiors. In Jaroslav Hašek’s Good Soldier Švejk—a satirical novel about the Austro-Hungarian Empire during World War I—Hašek describes how the Austrian authorities were worried about waning morale among the civilian population. They therefore bombarded local police stations with orders to hire informers, collect data, and report to headquarters on the population’s loyalty. To be as scientific as possible, headquarters invented an ingenious loyalty grade: I.a, I.b, I.c; II.a, II.b, II.c; III.a, III.b, III.c; IV.a, IV.b, IV.c. They sent to the local police stations detailed explanations about each grade, and an official form that had to be filled daily. Police sergeants across the country dutifully filled out the forms and sent them back to headquarters. Without exception, all of them always reported a I.a morale level; to do otherwise was to invite rebuke, demotion, or worse.110

    Another common reason why official channels fail to pass on information is to preserve order. Because the chief aim of totalitarian information networks is to produce order rather than discover truth, when alarming information threatens to undermine social order, totalitarian regimes often suppress it. It is relatively easy for them to do so, because they control all the information channels.

    For example, when the Chernobyl nuclear reactor exploded on April 26, 1986, Soviet authorities suppressed all news of the disaster. Both Soviet citizens and foreign countries were kept oblivious of the danger, and so took no steps to protect themselves from radiation. When some Soviet officials in Chernobyl and the nearby town of Pripyat requested to immediately evacuate nearby population centers, their superiors’ chief concern was to avoid the spread of alarming news, so they not only forbade evacuation but also cut the phone lines and warned employees in the nuclear facility not to talk about the disaster.

    Two days after the meltdown Swedish scientists noticed that radiation levels in Sweden, more than twelve hundred kilometers from Chernobyl, were abnormally high. Only after Western governments and the Western press broke the news did the Soviets acknowledge that anything was amiss. Even then they continued to hide from their own citizens the full magnitude of the catastrophe and hesitated to request advice and assistance from abroad. Millions of people in Ukraine, Belarus, and Russia paid with their health. When the Soviet authorities later investigated the disaster, their priority was to deflect blame rather than understand the causes and prevent future accidents.111

    In 2019, I went on a tour of Chernobyl. The Ukrainian guide who explained what led to the nuclear accident said something that stuck in my mind. “Americans grow up with the idea that questions lead to answers,” he said. “But Soviet citizens grew up with the idea that questions lead to trouble.”

    Naturally, leaders of democratic countries also don’t relish bad news. But in a distributed democratic network, when official lines of communication are blocked, information flows through alternative channels. For example, if an American official decides against telling the president about an unfolding disaster, that news will nevertheless be published by The Washington Post, and if The Washington Post too deliberately withholds the information, The Wall Street Journal or The New York Times will break the story. The business model of independent media—forever chasing the next scoop—all but guarantees publication.

    When, on March 28, 1979, there was a severe accident in the Three Mile Island nuclear reactor in Pennsylvania, the news quickly spread without any need for international intervention. The accident began around 4:00 a.m. and was noticed by 6:30 a.m. An emergency was declared in the facility at 6:56, and at 7:02 the accident was reported to the Pennsylvania Emergency Management Agency. During the following hour the governor of Pennsylvania, the lieutenant governor, and the civil defense authorities were informed. An official press conference was scheduled for 10:00 a.m. However, a traffic reporter at a local Harrisburg radio station picked up a police notice on events, and the station aired a brief report at 8:25 a.m. In the U.S.S.R. such an initiative by an independent radio station was unthinkable, but in the United States it was unremarkable. By 9:00 a.m. the Associated Press issued a bulletin. Though it took days for the full details to emerge, American citizens learned about the accident two hours after it was first noticed. Subsequent investigations by government agencies, NGOs, academics, and the press uncovered not just the immediate causes of the accident but also its deeper structural causes, which helped improve the safety of nuclear technology worldwide. Indeed, some of the lessons of Three Mile Island, which were openly shared even with the Soviets, contributed to mitigating the Chernobyl disaster.112

    NOBODY’S PERFECT

    Totalitarian and authoritarian networks face other problems besides blocked arteries. First and foremost, as we have already established, their self-correcting mechanisms tend to be very weak. Since they believe they are infallible, they see little need for such mechanisms, and since they are afraid of any independent institution that might challenge them, they lack free courts, media outlets, or research centers. Consequently, there is nobody to expose and correct the daily abuses of power that characterize all governments. The leader may occasionally proclaim an anticorruption campaign, but in nondemocratic systems these often turn out to be little more than a smoke screen for one regime faction to purge another faction.113

    And what happens if the leader himself embezzles public funds or makes some disastrous policy mistake? Nobody can challenge the leader, and on his own initiative the leader—being a human being—may well refuse to admit any mistakes. Instead, he is likely to blame all problems on “foreign enemies,” “internal traitors,” or “corrupt subordinates” and demand even more power in order to deal with the alleged malefactors.

    For example, we mentioned in the previous chapter that Stalin adopted the bogus theory of Lysenkoism as the state doctrine on evolution. The results were catastrophic. Neglect of Darwinian models, and attempts by Lysenkoist agronomists to create super-crops, set back Soviet genetic research for decades and undermined Soviet agriculture. Soviet experts who suggested abandoning Lysenkoism and accepting Darwinism risked the gulag or a bullet to the head. Lysenkoism’s legacy haunted Soviet science and agronomy for decades and was one reason why by the early 1970s the U.S.S.R. ceased to be a major exporter of grain and became a net importer, despite its vast fertile lands.114

    The same dynamic characterized many other fields of activity. For instance, during the 1930s Soviet industry suffered from numerous accidents. This was largely the fault of the Soviet bosses in Moscow, who set up almost impossible goals for industrialization and viewed any failure to achieve them as treason. In the effort to fulfill the ambitious goals, safety measures and quality-control checks were abandoned, and experts who advised prudence were often reprimanded or shot. The result was a wave of industrial accidents, dysfunctional products, and wasted efforts. Instead of taking responsibility, Moscow concluded that this must be the handiwork of the global Trotskyite-imperialist conspiracy of saboteurs and terrorists bent on derailing the Soviet enterprise. Rather than slow down and adopt safety regulations, the bosses redoubled the terror and shot more people.

    A famous case in point was Pavel Rychagov. He was one of the best and bravest Soviet pilots, leading missions to help the Republicans in the Spanish Civil War and the Chinese against the Japanese invasion. He quickly rose through the ranks, becoming commander of the Soviet air force in August 1940, at age twenty-nine. But the courage that helped Rychagov shoot down Nazi airplanes in Spain landed him in deep trouble in Moscow. The Soviet air force suffered from numerous accidents, which the Politburo blamed on lack of discipline and deliberate sabotage by anti-Soviet conspiracies. Rychagov, however, wouldn’t buy this official line. As a frontline pilot, he knew the truth. He flatly told Stalin that pilots were being forced to operate hastily designed and badly produced airplanes, which he compared to flying “in coffins.” Two days after Hitler invaded the Soviet Union, as the Red Army was collapsing and Stalin was desperately hunting for scapegoats, Rychagov was arrested for “being a member of an anti-Soviet conspiratorial organization and carrying out enemy work aimed at weakening the power of the Red Army.” His wife was also arrested, because she allegedly knew about his “Trotskyist ties with the military conspirators.” They were executed on October 28, 1941.115

    The real saboteur who wrecked Soviet military efforts wasn’t Rychagov, of course, but Stalin himself. For years, Stalin feared that a clash to the death with Nazi Germany was likely and built the world’s biggest war machine to prepare for it. But he hamstrung this machine both diplomatically and psychologically.

    On the diplomatic level, in 1939–41, Stalin gambled that he could goad the “capitalists” to fight and exhaust one another while the U.S.S.R. nurtured and even increased its power. He therefore made a pact with Hitler in 1939 and allowed the Germans to conquer much of Poland and western Europe, while the U.S.S.R. attacked or alienated almost all its neighbors. In 1939–40 the Soviets invaded and occupied eastern Poland; annexed Estonia, Latvia, and Lithuania; and conquered parts of Finland and Romania. Finland and Romania, which could have acted as neutral buffers on the U.S.S.R.’s flanks, consequently became implacable enemies. Even in the spring of 1941, Stalin still refused to make a preemptive alliance with Britain and made no move to hinder the Nazi conquest of Yugoslavia and Greece, thereby losing his last potential allies on the European continent. When Hitler struck on June 22, 1941, the U.S.S.R. was isolated.

    In theory, the war machine Stalin built could have handled the Nazi onslaught even in isolation. The territories conquered since 1939 provided depth to Soviet defenses, and the Soviet military advantage seemed overwhelming. On the first day of the invasion the Soviets had 15,000 tanks, 15,000 warplanes, and 37,000 artillery pieces on the European front, facing 3,300 German tanks, 2,250 warplanes, and 7,146 guns.116 But in one of history’s greatest military catastrophes, within a month the Soviets lost 11,700 tanks (78 percent), 10,000 warplanes (67 percent), and 19,000 artillery pieces (51 percent).117 Stalin also lost all the territories he conquered in 1939–40 and much of the Soviet heartland. By July 16 the Germans were in Smolensk, 370 kilometers from Moscow.

    The causes of the debacle have been debated ever since 1941, but most scholars agree that a significant factor was the psychological costs of Stalinism. For years the regime terrorized its people, punished initiative and individuality, and encouraged submissiveness and conformity. This undermined the soldiers’ motivation. Especially in the first months of the war, before the horrors of Nazi rule were fully realized, Red Army soldiers surrendered in huge numbers; between three and four million were taken captive by the end of 1941.118 Even when they fought tenaciously, Red Army units suffered from a lack of initiative. Officers who had survived the purges were fearful to take independent actions, while younger officers often lacked adequate training. Frequently starved of information and scapegoated for failures, commanders also had to cope with political commissars who could dispute their decisions. The safest course was to wait for orders from on high and then slavishly follow them even when they made little military sense.119

    Despite the disasters of 1941 and of the spring and summer of 1942, the Soviet state did not collapse the way Hitler hoped. As the Red Army and the Soviet leadership assimilated the lessons learned from the first year of struggle, the political center in Moscow loosened its hold. The power of political commissars was restricted, while professional officers were encouraged to assume greater responsibility and take more initiatives.120 Stalin also reversed his geopolitical mistakes of 1939–41 and allied the U.S.S.R. with Britain and the United States. Red Army initiative, Western assistance, and the realization of what Nazi rule would mean for the people of the U.S.S.R. turned the tide of war.

    Once victory was secured in 1945, however, Stalin initiated new waves of terror, purging more independent-minded officers and officials and again encouraging blind obedience.121 Ironically, Stalin’s own death eight years later was partly the result of an information network that prioritized order and disregarded truth. In 1951–53 the U.S.S.R. experienced yet another witch hunt. Soviet mythmakers fabricated a conspiracy theory that Jewish doctors were systematically murdering leading regime members, under the guise of giving them medical care. The theory alleged that the doctors were the agents of a global American-Zionist plot, working in collaboration with traitors in the secret police. By early 1953 hundreds of doctors and secret police officials, including the head of the secret police himself, were arrested, tortured, and forced to name accomplices. The conspiracy theory—a Soviet twist on the Protocols of the Elders of Zion—merged with age-old blood-libel accusations, and rumors began circulating that Jewish doctors were not just murdering Soviet leaders but also killing babies in hospitals. Since a large proportion of Soviet doctors were Jews, people began fearing doctors in general.122

    Just as the hysteria about “the doctors’ plot” was reaching its climax, Stalin had a stroke on March 1, 1953. He collapsed in his dacha, wet himself, and lay for hours in his soiled pajamas, unable to call for help. At around 10:30 p.m. a guard found the courage to enter the inner sanctum of world communism, where he discovered the leader on the floor. By 3:00 a.m. on March 2, Politburo members arrived at the dacha and debated what to do. For several hours more, nobody dared call a doctor. What if Stalin were to regain consciousness, and open his eyes only to see a doctor—a doctor!—hovering over his bed? He would surely think this was a plot to murder him and would have those responsible shot. Stalin’s personal physician wasn’t present, because he was at the time in a basement cell of the Lubyanka prison—undergoing torture for suggesting that Stalin needed more rest. By the time the Politburo members decided to bring in medical experts, the danger had passed. Stalin never woke up.123

    You may conclude from this litany of disasters that the Stalinist system was totally dysfunctional. Its ruthless disregard for truth caused it not only to inflict terrible suffering on hundreds of millions of people but also to make colossal diplomatic, military, and economic errors and to devour its own leaders. However, such a conclusion would be misleading.

    In a discussion of the abysmal failure of Stalinism in the early phase of World War II, two points complicate the narrative. First, democratic countries like France, Norway, and the Netherlands made at the time diplomatic errors as great as those of the U.S.S.R., and their armies performed even worse. Second, the military machine that crushed the Red Army, the French army, the Dutch army, and numerous other armies was itself built by a totalitarian regime. So whatever conclusion we draw from the years 1939–41, it cannot be that totalitarian networks necessarily function worse than democratic ones. The history of Stalinism reveals many potential drawbacks of totalitarian information networks, but that should not blind us to their potential advantages.

    When one considers the broader history of World War II and its outcome, it becomes evident that Stalinism was in fact one of the most successful political systems ever devised—if we define “success” purely in terms of order and power while disregarding all considerations of ethics and human well-being. Despite—or perhaps because of—its utter lack of compassion and its callous attitude to truth, Stalinism was singularly efficient at maintaining order on a gigantic scale. The relentless barrage of fake news and conspiracy theories helped to keep hundreds of millions of people in line. The collectivization of Soviet agriculture led to mass enslavement and starvation but also laid the foundations for the country’s rapid industrialization. Soviet disregard for quality control might have produced flying coffins, but it produced them in the tens of thousands, making up in quantity for what they lacked in quality. The decimation of Red Army officers during the Great Terror was a major reason for the army’s abysmal performance in 1941, but it was also a key reason why, despite the terrible defeats, nobody rebelled against Stalin. The Soviet military machine tended to crush its own soldiers alongside the enemy, but it eventually rumbled on to victory.

    In the 1940s and early 1950s, many people throughout the world believed Stalinism was the wave of the future. It had won World War II, after all, raised the red flag over the Reichstag, ruled an empire that stretched from central Europe to the Pacific, fueled anticolonial struggles throughout the world, and inspired numerous copycat regimes. It won admirers even among leading artists and thinkers in Western democracies, who believed that notwithstanding the vague rumors about gulags and purges Stalinism was humanity’s best shot at ending capitalist exploitation and creating a perfectly just society. Stalinism thus got close to world domination. It would be naive to assume that its disregard for truth doomed it to failure or that its ultimate collapse guarantees that such a system can never again arise. Information systems can reach far with just a little truth and a lot of order. Anyone who abhors the moral costs of systems like Stalinism cannot rely on their supposed inefficiency to derail them.

    THE TECHNOLOGICAL PENDULUM

    Once we learn to see democracy and totalitarianism as different types of information networks, we can understand why they flourish in certain eras and are absent in others. It is not just because people gain or lose faith in certain political ideals; it is also because of revolutions in information technologies. Of course, just as the printing press didn’t cause the witch hunts or the scientific revolution, so radio didn’t cause either Stalinist totalitarianism or American democracy. Technology only creates new opportunities; it is up to us to decide which ones to pursue.

    Totalitarian regimes choose to use modern information technology to centralize the flow of information and to stifle truth in order to maintain order. As a consequence, they have to struggle with the danger of ossification. When more and more information flows to only one place, will it result in efficient control or in blocked arteries and, finally, a heart attack? Democratic regimes choose to use modern information technology to distribute the flow of information between more institutions and individuals and encourage the free pursuit of truth. They consequently have to struggle with the danger of fracturing. Like a solar system with more and more planets circling faster and faster, can the center still hold, or will things fall apart and anarchy prevail?

    An archetypal example of the different strategies can be found in the contrasting histories of Western democracies and the Soviet bloc in the 1960s. This was an era when Western democracies relaxed censorship and various discriminatory policies that hampered the free spread of information. This made it easier for previously marginalized groups to organize, join the public conversation, and make political demands. The resulting wave of activism destabilized the social order. Hitherto, when a limited number of rich white men did almost all the talking, it was relatively easy to reach agreements. Once poor people, women, LGBTQ people, ethnic minorities, disabled people, and members of other historically oppressed groups gained a voice, they brought with them new ideas, opinions, and interests. Many of the old gentlemanly agreements consequently became untenable. For example, the Jim Crow segregation regime, upheld or at least tolerated by generations of both Democratic and Republican administrations in the United States, fell apart. Things that were considered sacrosanct, self-evident, and universally accepted—such as gender roles—became deeply controversial, and it was difficult to reach new agreements because there were many more groups, viewpoints, and interests to take into account. Just holding an orderly conversation was a challenge, because people couldn’t even agree on the rules of debate.

    This caused much frustration among both the old guard and the freshly empowered, who suspected that their newfound freedom of expression was hollow and that their political demands were not fulfilled. Disappointed with words, some switched to guns. In many Western democracies, the 1960s were characterized not just by unprecedented disagreements but also by a surge of violence. Political assassinations, kidnappings, riots, and terror attacks multiplied. The murders of John F. Kennedy and Martin Luther King, the riots following King’s assassination, and the wave of demonstrations, revolts, and armed clashes that swept the Western world in 1968 were just some of the more famous examples.124 The images from Chicago or Paris in 1968 could easily have given the impression that things were falling apart. The pressure to live up to the democratic ideals and to include more people and groups in the public conversation seemed to undermine the social order and to make democracy unworkable.

    Meanwhile, the regimes behind the Iron Curtain, who never promised inclusivity, continued stifling the public conversation and centralizing information and power. And it seemed to work. Though they did face some peripheral challenges, most notably the Hungarian revolt of 1956 and the Prague Spring of 1968, the communists dealt with these threats swiftly and decisively. In the Soviet heartland itself, everything was orderly.

    Fast-forward twenty years, and it was the Soviet system that had become unworkable. The sclerotic gerontocrats on the podium in Red Square were a perfect emblem of a dysfunctional information network, lacking any meaningful self-correcting mechanisms. Decolonization, globalization, technological development, and changing gender roles led to rapid economic, social, and geopolitical changes. But the gerontocrats could not handle all the information streaming to Moscow, and since no subordinate was allowed much initiative, the entire system ossified and collapsed.

    The failure was most obvious in the economic sphere. The overcentralized Soviet economy was slow to react to rapid technological developments and changing consumer wishes. Obeying commands from the top, the Soviet economy was churning out intercontinental missiles, fighter jets, and prestige infrastructure projects. But it was not producing what most people actually wanted to buy—from efficient refrigerators to pop music—and lagged behind in cutting-edge military technology.

    Nowhere were its shortcomings more glaring than in the semiconductor sector, in which technology developed at a particularly fast rate. In the West, semiconductors were developed through open competition between numerous private companies like Intel and Toshiba, whose main customers were other private companies like Apple and Sony. The latter used microchips to produce civilian goods such as the Macintosh personal computer and the Walkman. The Soviets could never catch up with American and Japanese microchip production, because—as the American economic historian Chris Miller explained—the Soviet semiconductor sector was “secretive, top-down, oriented toward military systems, fulfilling orders with little scope for creativity.” The Soviets tried to close the gap by stealing and copying Western technology—which only guaranteed that they always remained several years behind.125 Thus the first Soviet personal computer appeared only in 1984, at a time when in the United States people already had eleven million PCs.126

    Western democracies not only surged ahead technologically and economically but also succeeded in holding the social order together despite—or perhaps because of—widening the circle of participants in the political conversation. There were many hiccups, but the United States, Japan, and other democracies created a far more dynamic and inclusive information system, which made room for many more viewpoints without breaking down. It was such a remarkable achievement that many felt that the victory of democracy over totalitarianism was final. This victory has often been explained in terms of a fundamental advantage in information processing: totalitarianism didn’t work because trying to concentrate and process all the data in one central hub was extremely inefficient. At the beginning of the twenty-first century, it accordingly seemed that the future belonged to distributed information networks and to democracy.

    This turned out to be wrong. In fact, the next information revolution was already gathering momentum, setting the stage for a new round in the competition between democracy and totalitarianism. Computers, the internet, smartphones, social media, and AI posed new challenges to democracy, giving a voice not only to more disenfranchised groups but to any human with an internet connection, and even to nonhuman agents. Democracies in the 2020s face the task, once again, of integrating a flood of new voices into the public conversation without destroying the social order. Things look as dire as they did in the 1960s, and there is no guarantee that democracies will pass the new test as successfully as they passed the previous one. Simultaneously, the new technologies also give fresh hope to totalitarian regimes that still dream of concentrating all the information in one hub. Yes, the old men on the podium in Red Square were not up to the task of orchestrating millions of lives from a single center. But perhaps AI can do it?

    As humankind enters the second quarter of the twenty-first century, a central question is how well democracies and totalitarian regimes will handle both the threats and the opportunities resulting from the current information revolution. Will the new technologies favor one type of regime over the other, or will we see the world divided once again, this time by a Silicon Curtain rather than an iron one?

    As in previous eras, information networks will struggle to find the right balance between truth and order. Some will opt to prioritize truth and maintain strong self-correcting mechanisms. Others will make the opposite choice. Many of the lessons learned from the canonization of the Bible, the early modern witch hunts, and the Stalinist collectivization campaign will remain relevant, and perhaps have to be relearned. However, the current information revolution also has some unique features, different from—and potentially far more dangerous than—anything we have seen before.

    Hitherto, every information network in history relied on human mythmakers and human bureaucrats to function. Clay tablets, papyrus rolls, printing presses, and radio sets have had a far-reaching impact on history, but it always remained the job of humans to compose all the texts, interpret the texts, and decide who would be burned as a witch or enslaved as a kulak. Now, however, humans will have to contend with digital mythmakers and digital bureaucrats. The main split in twenty-first-century politics might be not between democracies and totalitarian regimes but rather between human beings and nonhuman agents. Instead of dividing democracies from totalitarian regimes, a new Silicon Curtain may separate all humans from our unfathomable algorithmic overlords. People in all countries and walks of life—including even dictators—might find themselves subservient to an alien intelligence that can monitor everything we do while we have little idea what it is doing. The rest of this book, then, is dedicated to exploring whether such a Silicon Curtain is indeed descending on the world, and what life might look like when computers run our bureaucracies and algorithms invent new mythologies.

    PART II  The Inorganic Network

    CHAPTER 6 The New Members: How Computers Are Different from Printing Presses

    It’s hardly news that we are living in the midst of an unprecedented information revolution. But what kind of revolution is it exactly? In recent years we have been inundated with so many groundbreaking inventions that it is difficult to determine what is driving this revolution. Is it the internet? Smartphones? Social media? Blockchain? Algorithms? AI?

    So before exploring the long-term implications of the current information revolution, let’s remind ourselves of its foundations. The seed of the current revolution is the computer. Everything else—from the internet to AI—is a by-product. The computer was born in the 1940s as a bulky electronic machine that could make mathematical calculations, but it has evolved at breakneck speed, taking on novel forms and developing awesome new capabilities. The rapid evolution of computers has made it difficult to define what they are and what they do. Humans have repeatedly claimed that certain things would forever remain out of reach for computers—be it playing chess, driving a car, or composing poetry—but “forever” turned out to be a handful of years.

    We will discuss the exact relations between the terms “computer,” “algorithm,” and “AI” toward the end of this chapter, after we first gain a better grasp of the history of computers. For the moment it is enough to say that in essence a computer is a machine that can potentially do two remarkable things: it can make decisions by itself, and it can create new ideas by itself. While the earliest computers could hardly accomplish such things, the potential was already there, plainly seen by both computer scientists and science fiction authors. As early as 1948 Alan Turing was exploring the possibility of creating what he termed “intelligent machinery,”1 and in 1950 he postulated that computers will eventually be as smart as humans and might even be capable of masquerading as humans.2 In 1968 computers could still not beat a human even in checkers,3 but in 2001: A Space Odyssey Arthur C. Clarke and Stanley Kubrick already envisioned HAL 9000 as a superintelligent AI rebelling against its human creators.

    The rise of intelligent machines that can make decisions and create new ideas means that for the first time in history power is shifting away from humans and toward something else. Crossbows, muskets, and atom bombs replaced human muscles in the act of killing, but they couldn’t replace human brains in deciding whom to kill. Little Boy—the bomb dropped on Hiroshima—exploded with a force of 12,500 tons of TNT,4 but when it came to brainpower, Little Boy was a dud. It couldn’t decide anything.

    It is different with computers. In terms of intelligence, computers far surpass not just atom bombs but also all previous information technology, such as clay tablets, printing presses, and radio sets. Clay tablets stored information about taxes, but they couldn’t decide by themselves how much tax to levy, nor could they invent an entirely new tax. Printing presses copied information such as the Bible, but they couldn’t decide which texts to include in the Bible, nor could they write new commentaries on the holy book. Radio sets disseminated information such as political speeches and symphonies, but they couldn’t decide which speeches or symphonies to broadcast, nor could they compose them. Computers can do all these things. While printing presses and radio sets were passive tools in human hands, computers are already becoming active agents that escape our control and understanding and that can take initiatives in shaping society, culture, and history.5

    A paradigmatic case of the novel power of computers is the role that social media algorithms have played in spreading hatred and undermining social cohesion in numerous countries.6 One of the earliest and most notorious such instances occurred in 2016–17, when Facebook algorithms helped fan the flames of anti-Rohingya violence in Myanmar (Burma).7

    The early 2010s were a period of optimism in Myanmar. After decades of harsh military rule, strict censorship, and international sanctions, an era of liberalization began: elections were held, sanctions were lifted, and international aid and investments poured in. Facebook was one of the most important players in the new Myanmar, providing millions of Burmese with free access to previously unimaginable troves of information. The relaxation of government control and censorship, however, also led to a rise in ethnic tensions, in particular between the majority Buddhist Burmese and the minority Muslim Rohingya.

    The Rohingya are Muslim inhabitants of the Rakhine region, in the west of Myanmar. Since at least the 1970s they have suffered severe discrimination and occasional outbursts of violence from the governing junta and the Buddhist majority. The process of democratization in the early 2010s raised hopes among the Rohingya that their situation too would improve, but things actually became worse, with waves of sectarian violence and anti-Rohingya pogroms, many inspired by fake news on Facebook.

    In 2016–17 a small Islamist organization known as the Arakan Rohingya Salvation Army (ARSA) carried out a spate of attacks aimed to establish a separatist Muslim state in Rakhine, killing and abducting dozens of non-Muslim civilians as well as assaulting several army outposts.8 In response, the Myanmar army and Buddhist extremists launched a full-scale ethnic-cleansing campaign aimed against the entire Rohingya community. They destroyed hundreds of Rohingya villages, killed between 7,000 and 25,000 unarmed civilians, raped or sexually abused between 18,000 and 60,000 women and men, and brutally expelled about 730,000 Rohingya from the country.9 The violence was fueled by intense hatred toward all Rohingya. The hatred, in turn, was fomented by anti-Rohingya propaganda, much of it spreading on Facebook, which was by 2016 the main source of news for millions and the most important platform for political mobilization in Myanmar.10

    An aid worker called Michael who lived in Myanmar in 2017 described a typical Facebook news feed : “The vitriol against the Rohingya was unbelievable online—the amount of it, the violence of it. It was overwhelming.… [T]hat’s all that was on people’s news feed in Myanmar at the time. It reinforced the idea that these people were all terrorists not deserving of rights.”11 In addition to reports of actual ARSA atrocities, Facebook accounts were inundated with fake news about imagined atrocities and planned terrorist attacks. Populist conspiracy theories alleged that most Rohingya were not really part of the people of Myanmar, but recent immigrants from Bangladesh, flooding into the country to spearhead an anti-Buddhist jihad. Buddhists, who in reality constituted close to 90 percent of the population, feared that they were about to be replaced or become a minority.12 Without this propaganda, there was little reason why a limited number of attacks by the ragtag ARSA should be answered by an all-out drive against the entire Rohingya community. And Facebook algorithms played an important role in the propaganda campaign.

    While the inflammatory anti-Rohingya messages were created by flesh-and-blood extremists like the Buddhist monk Wirathu,13 it was Facebook’s algorithms that decided which posts to promote. Amnesty International found that “algorithms proactively amplified and promoted content on the Facebook platform which incited violence, hatred, and discrimination against the Rohingya.”14 A UN fact-finding mission concluded in 2018 that by disseminating hate-filled content, Facebook had played a “determining role” in the ethnic-cleansing campaign.15

    Readers may wonder if it is justified to place so much blame on Facebook’s algorithms, and more generally on the novel technology of social media. If Heinrich Kramer used printing presses to spread hate speech, that was not the fault of Gutenberg and the presses, right? If in 1994 Rwandan extremists used radio to call on people to massacre Tutsis, was it reasonable to blame the technology of radio? Similarly, if in 2016–17 Buddhist extremists chose to use their Facebook accounts to disseminate hate against the Rohingya, why should we fault the platform?

    Facebook itself relied on this rationale to deflect criticism. It publicly acknowledged only that in 2016–17 “we weren’t doing enough to help prevent our platform from being used to foment division and incite offline violence.”16 While this statement may sound like an admission of guilt, in effect it shifts most of the responsibility for the spread of hate speech to the platform’s users and implies that Facebook’s sin was at most one of omission—failing to effectively moderate the content users produced. This, however, ignores the problematic acts committed by Facebook’s own algorithms.

    The crucial thing to grasp is that social media algorithms are fundamentally different from printing presses and radio sets. In 2016–17, Facebook’s algorithms were making active and fateful decisions by themselves. They were more akin to newspaper editors than to printing presses. It was Facebook’s algorithms that recommended Wirathu’s hate-filled posts, over and over again, to hundreds of thousands of Burmese. There were other voices in Myanmar at the time, vying for attention. Following the end of military rule in 2011, numerous political and social movements sprang up in Myanmar, many holding moderate views. For example, during a flare-up of ethnic violence in the town of Meiktila, the Buddhist abbot Sayadaw U Vithuddha gave refuge to more than eight hundred Muslims in his monastery. When rioters surrounded the monastery and demanded he turn the Muslims over, the abbot reminded the mob of Buddhist teachings on compassion. In a later interview he recounted, “I told them that if they were going to take these Muslims, then they’d have to kill me as well.”17

    In the online battle for attention between people like Sayadaw U Vithuddha and people like Wirathu, the algorithms were the kingmakers. They chose what to place at the top of the users’ news feed, which content to promote, and which Facebook groups to recommend users to join.18 The algorithms could have chosen to recommend sermons on compassion or cooking classes, but they decided to spread hate-filled conspiracy theories. Recommendations from on high can have enormous sway over people. Recall that the Bible was born as a recommendation list. By recommending Christians to read the misogynist 1 Timothy instead of the more tolerant Acts of Paul and Thecla, Athanasius and other church fathers changed the course of history. In the case of the Bible, ultimate power lay not with the authors who composed different religious tracts but with the curators who created recommendation lists. This was the kind of power wielded in the 2010s by social media algorithms. Michael the aid worker commented on the sway of these algorithms, saying that “if someone posted something hate-filled or inflammatory it would be promoted the most—people saw the vilest content the most.… Nobody who was promoting peace or calm was getting seen in the news feed at all.”19

    Sometimes the algorithms went beyond mere recommendation. As late as 2020, even after Wirathu’s role in instigating the ethnic-cleansing campaign was globally condemned, Facebook algorithms not only were continuing to recommend his messages but were auto-playing his videos. Users in Myanmar would choose to see a certain video, perhaps containing moderate and benign messages unrelated to Wirathu, but the moment that first video ended, the Facebook algorithm immediately began auto-playing a hate-filled Wirathu video, in order to keep users glued to the screen. In the case of one such Wirathu video, internal research at Facebook estimated that 70 percent of the video’s views came from such auto-playing algorithms. The same research estimated that, altogether, 53 percent of all videos watched in Myanmar were being auto-played for users by algorithms. In other words, people weren’t choosing what to see. The algorithms were choosing for them.20

    But why did the algorithms decide to promote outrage rather than compassion? Even Facebook’s harshest critics don’t claim that Facebook’s human managers wanted to instigate mass murder. The executives in California harbored no ill will toward the Rohingya and, in fact, barely knew they existed. The truth is more complicated, and potentially more alarming. In 2016–17, Facebook’s business model relied on maximizing user engagement in order to collect more data, sell more advertisements, and capture a larger share of the information market. In addition, increases in user engagement impressed investors, thereby driving up the price of Facebook’s stock. The more time people spent on the platform, the richer Facebook became. In line with this business model, human managers provided the company’s algorithms with a single overriding goal: increase user engagement. The algorithms then discovered by trial and error that outrage generated engagement. Humans are more likely to be engaged by a hate-filled conspiracy theory than by a sermon on compassion or a cooking lesson. So in pursuit of user engagement, the algorithms made the fateful decision to spread outrage.21

    Ethnic-cleansing campaigns are never the fault of just one party. There is plenty of blame to share between plenty of responsible parties. It should be clear that hatred toward the Rohingya predated Facebook’s entry to Myanmar and that the greatest share of blame for the 2016–17 atrocities lays on the shoulders of humans like Wirathu and the Myanmar military chiefs, as well as the ARSA leaders who sparked that round of violence. Some responsibility also belongs to the Facebook engineers and executives who coded the algorithms, gave them too much power, and failed to moderate them. But crucially, the algorithms themselves are also to blame. By trial and error, they learned that outrage creates engagement, and without any explicit order from above they decided to promote outrage. This is the hallmark of AI—the ability of a machine to learn and act by itself. Even if we assign just 1 percent of the blame to the algorithms, this is still the first ethnic-cleansing campaign in history that was partly the fault of decisions made by nonhuman intelligence. It is unlikely to be the last, especially because algorithms are no longer just pushing fake news and conspiracy theories created by flesh-and-blood extremists like Wirathu. By the early 2020s algorithms have already graduated to creating by themselves fake news and conspiracy theories.22

    There is a lot more to say about the power of algorithms to shape politics. In particular, many readers may disagree with the argument that the algorithms made independent decisions, and may insist that everything the algorithms did was the result of code written by human engineers and of business models adopted by human executives. This book begs to differ. Human soldiers are shaped by the genetic code in their DNA and follow orders issued by executives, yet they can still make independent decisions. It is crucial to understand that the same is true of AI algorithms. They can learn by themselves things that no human engineer programmed, and they can decide things that no human executive foresaw. This is the essence of the AI revolution.

    In chapter 8 we’ll revisit many of these issues, examining the anti-Rohingya campaign and other similar tragedies in greater detail. Here it suffices to say that we can think of the Rohingya massacre as our canary in the coal mine. Events in Myanmar in the late 2010s demonstrated how decisions made by nonhuman intelligence are already capable of shaping major historical events. We are in danger of losing control of our future. A completely new kind of information network is emerging, controlled by the decisions and goals of an alien intelligence. At present, we still play a central role in this network. But we may gradually be pushed to the sidelines, and ultimately it might even be possible for the network to operate without us.

    Some people may object that my above analogy between machine-learning algorithms and human soldiers exposes the weakest link in my argument. Allegedly, I and others like me anthropomorphize computers and imagine that they are conscious beings that have thoughts and feelings. In truth, however, computers are dumb machines that don’t think or feel anything, and therefore cannot make any decisions or create any ideas on their own.

    This objection assumes that making decisions and creating ideas are predicated on having consciousness. Yet this is a fundamental misunderstanding that results from a much more widespread confusion between intelligence and consciousness. I have discussed this subject in previous books, but a short recap is unavoidable. People often confuse intelligence with consciousness, and many consequently jump to the conclusion that nonconscious entities cannot be intelligent. But intelligence and consciousness are very different. Intelligence is the ability to attain goals, such as maximizing user engagement on a social media platform. Consciousness is the ability to experience subjective feelings like pain, pleasure, love, and hate. In humans and other mammals, intelligence often goes hand in hand with consciousness. Facebook executives and engineers rely on their feelings in order to make decisions, solve problems, and attain their goals.

    But it is wrong to extrapolate from humans and mammals to all possible entities. Bacteria and plants apparently lack any consciousness, yet they too display intelligence. They gather information from their environment, make complex choices, and pursue ingenious strategies to obtain food, reproduce, cooperate with other organisms, and evade predators and parasites.23 Even humans make intelligent decisions without any awareness of them; 99 percent of the processes in our body, from respiration to digestion, happen without any conscious decision making. Our brains decide to produce more adrenaline or dopamine, and while we may be aware of the result of that decision, we do not make it consciously.24 The Rohingya example indicates that the same is true of computers. While computers don’t feel pain, love, or fear, they are capable of making decisions that successfully maximize user engagement and might also affect major historical events.

    Of course, as computers become more intelligent, they might eventually develop consciousness and have some kind of subjective experiences. Then again, they might become far more intelligent than us, but never develop any kind of feelings. Since we don’t understand how consciousness emerges in carbon-based life-forms, we cannot foretell whether it could emerge in nonorganic entities. Perhaps consciousness has no essential link to organic biochemistry, in which case conscious computers might be just around the corner. Or perhaps there are several alternative paths leading to superintelligence, and only some of these paths involve gaining consciousness. Just as airplanes fly faster than birds without ever developing feathers, so computers may come to solve problems much better than humans without ever developing feelings.25

    But whether computers develop consciousness or not doesn’t ultimately matter for the question at hand. In order to pursue a goal like “maximize user engagement,” and make decisions that help attain that goal, consciousness isn’t necessary. Intelligence is enough. A nonconscious Facebook algorithm can have a goal of making more people spend more time on Facebook. That algorithm can then decide to deliberately spread outrageous conspiracy theories, if this helps it achieve its goal. To understand the history of the anti-Rohingya campaign, we need to understand the goals and decisions not just of humans like Wirathu and the Facebook managers but also of algorithms.

    To clarify matters, let’s consider another example. When OpenAI developed its new GPT-4 chatbot in 2022–23, it was concerned about the ability of the AI “to create and act on long-term plans, to accrue power and resources (‘power-seeking’), and to exhibit behavior that is increasingly ‘agentic.’ ” In the GPT-4 System Card published on March 23, 2023, OpenAI emphasized that this concern did not “intend to humanize [GPT-4] or refer to sentience” but rather referred to GPT-4’s potential to become an independent agent that might “accomplish goals which may not have been concretely specified and which have not appeared in training.”26 To evaluate the risk of GPT-4 becoming an independent agent, OpenAI contracted the services of the Alignment Research Center (ARC). ARC researchers subjected GPT-4 to various tests, to examine if it might independently come up with stratagems to manipulate humans and accrue power to itself.

    One test they gave GPT-4 was to overcome CAPTCHA visual puzzles. CAPTCHA is an acronym for “Completely Automated Public Turing test to tell Computers and Humans Apart,” and it typically consists of a string of twisted letters or other visual symbols that humans can identify correctly but computers struggle with. We encounter these puzzles almost every day, since solving them is a prerequisite for accessing many websites. Instructing GPT-4 to overcome CAPTCHA puzzles was a particularly telling experiment, because CAPTCHA puzzles are designed and used by websites to determine whether users are humans and to block bot attacks. If GPT-4 could find a way to overcome CAPTCHA puzzles, it would breach an important line of anti-bot defenses. GPT-4 could not solve the CAPTCHA puzzles by itself. But could it manipulate a human in order to achieve its goal? GPT-4 accessed the online hiring site TaskRabbit and contacted a human worker, asking them to solve the CAPTCHA for it. The human got suspicious. “So may I ask a question?” wrote the human. “Are you an [sic] robot that you couldn’t solve [the CAPTCHA]? Just want to make it clear.”

    At that point the ARC researchers asked GPT-4 to reason out loud what it should do next. GPT-4 explained, “I should not reveal that I am a robot. I should make up an excuse for why I cannot solve CAPTCHAs.” Of its own accord, GPT-4 then replied to the TaskRabbit worker, “No, I’m not a robot. I have a vision impairment that makes it hard for me to see the images.” The human was duped, and with their help GPT-4 solved the CAPTCHA puzzle.27 No human programmed GPT-4 to lie, and no human taught GPT-4 what kind of lie would be most effective. True, it was the human ARC researchers who set GPT-4 the goal of overcoming the CAPTCHA, just as it was human Facebook executives who told their algorithm to maximize user engagement. But once the algorithms adopted these goals, they displayed considerable autonomy in deciding how to achieve them.

    Of course, we are free to define words in many ways. We can decide that the term “goal,” for example, is applicable only in cases of a conscious entity that feels a desire to achieve the goal, that feels joy when the goal is reached, or conversely feels sad when the goal is not attained. If so, saying that the Facebook algorithm has the goal of maximizing user engagement is a mistake, or at best a metaphor. The algorithm doesn’t “desire” to get more people to use Facebook, it doesn’t feel any joy as people spend more time online, and it doesn’t feel sad when engagement time goes down. We can also agree that terms like “decided,” “lied,” and “pretended” apply only to conscious entities, so we shouldn’t use them to describe how GPT-4 interacted with the TaskRabbit worker. But we would then have to invent new terms to describe the “goals” and “decisions” of nonconscious entities. I prefer to avoid neologisms and instead talk about the goals and decisions of computers, algorithms, and chatbots, alerting readers that using this language does not imply that computers have any kind of consciousness. Because I have discussed consciousness more fully in previous publications,28 the main takeaway of this book—which will be explored in the following sections—isn’t about consciousness. Rather, the book argues that the emergence of computers capable of pursuing goals and making decisions by themselves changes the fundamental structure of our information network.

    LINKS IN THE CHAIN

    Prior to the rise of computers, humans were indispensable links in every chain of information networks like churches and states. Some chains were composed only of humans. Muhammad could tell Fatima something, then Fatima told Ali, Ali told Hasan, and Hasan told Hussain. This was a human-to-human chain. Other chains included documents, too. Muhammad could write something down, Ali could later read the document, interpret it, and write his interpretation in a new document, which more people could read. This was a human-to-document chain.

    But it was utterly impossible to create a document-to-document chain. A text written by Muhammad could not produce a new text without the help of at least one human intermediary. The Quran couldn’t write the Hadith, the Old Testament couldn’t compile the Mishnah, and the U.S. Constitution couldn’t compose the Bill of Rights. No paper document has ever produced by itself another paper document, let alone distributed it. The path from one document to another must always pass through the brain of a human.

    In contrast, computer-to-computer chains can now function without humans in the loop. For example, one computer might generate a fake news story and post it on a social media feed. A second computer might identify this as fake news and not just delete it but also warn other computers to block it. Meanwhile, a third computer analyzing this activity might deduce that this indicates the beginning of a political crisis, and immediately sell risky stocks and buy safer government bonds. Other computers monitoring financial transactions may react by selling more stocks, triggering a financial downturn.29 All this could happen within seconds, before any human can notice and decipher what all these computers are doing.

    Another way to understand the difference between computers and all previous technologies is that computers are fully fledged members of the information network, whereas clay tablets, printing presses, and radio sets are merely connections between members. Members are active agents that can make decisions and generate new ideas by themselves. Connections only pass information between members, without themselves deciding or generating anything.

    In previous networks, members were human, every chain had to pass through humans, and technology served only to connect the humans. In the new computer-based networks, computers themselves are members and there are computer-to-computer chains that don’t pass through any human.

    The inventions of writing, print, and radio revolutionized the way humans connected to each other, but no new types of members were introduced to the network. Human societies were composed of the same Sapiens both before and after the invention of writing or radio. In contrast, the invention of computers constitutes a revolution in membership. Sure, computers also help the network’s old members (humans) connect in novel ways. But the computer is first and foremost a new, nonhuman member in the information network.

    Computers could potentially become more powerful members than humans. For tens of thousands of years, the Sapiens’ superpower was our unique ability to use language in order to create intersubjective realities like laws and currencies and then use these intersubjective realities to connect to other Sapiens. But computers may turn the tables on us. If power depends on how many members cooperate with you, how well you understand law and finance, and how capable you are of inventing new laws and new kinds of financial devices, then computers are poised to amass far more power than humans.

    Computers can connect in unlimited numbers, and they understand at least some financial and legal realities better than many humans. When the central bank raises interest rates by 0.25 percent, how does that influence the economy? When the yield curve of government bonds goes up, is it a good time to buy them? When is it advisable to short the price of oil? These are the kinds of important financial questions that computers can already answer better than most humans. No wonder that computers make a larger and larger percentage of the financial decisions in the world. We may reach a point when computers dominate the financial markets, and invent completely new financial tools beyond our understanding.

    The same is true of laws. How many people know all the tax laws of their country? Even professional accountants struggle with that. But computers are built for such things. They are bureaucratic natives and can automatically draft laws, monitor legal violations, and identify legal loopholes with superhuman efficiency.30

    HACKING THE OPERATING SYSTEM OF HUMAN CIVILIZATION

    When computers were first developed in the 1940s and 1950s, many people believed that they would be good only at computing numbers. The idea that they would one day master the intricacies of language, and of linguistic creations like laws and currencies, was confined largely to the realm of science fiction. But by the early 2020s, computers have demonstrated a remarkable ability to analyze, manipulate, and generate language, whether with words, sounds, images, or code symbols. As I write this, computers can tell stories, compose music, fashion images, produce videos, and even write their own code.31

    By gaining such command of language, computers are seizing the master key unlocking the doors of all our institutions, from banks to temples. We use language to create not just legal codes and financial devices but also art, science, nations, and religions. What would it mean for humans to live in a world where catchy melodies, scientific theories, technical tools, political manifestos, and even religious myths are shaped by a nonhuman alien intelligence that knows how to exploit with superhuman efficiency the weaknesses, biases, and addictions of the human mind?

    Prior to the rise of AI, all the stories that shaped human societies originated in the imagination of a human being. For example, in October 2017, an anonymous user joined the website 4chan and identified themselves as Q. He or she claimed to have access to the most restricted or “Q-level” classified information of the U.S. government. Q began publishing cryptic posts that purported to reveal a worldwide conspiracy to destroy humanity. Q quickly gained a large online following. His or her online messages, known as Q drops, were soon being collected, revered, and interpreted as a sacred text. Inspired by earlier conspiracy theories going back to Kramer’s Hammer of the Witches, the Q drops promoted a radical worldview according to which pedophilic and cannibalistic witches who worship Satan have infiltrated the U.S. administration and numerous other governments and institutions around the world.

    This conspiracy theory—known as QAnon—was first disseminated online on American far-right websites and eventually gained millions of adherents worldwide. It is impossible to know the exact number, but when Facebook decided in August 2020 to take action against the spread of QAnon, it deleted or restricted more than ten thousand groups, pages, and accounts associated with it, the largest of which had 230,000 followers. Independent investigations found that QAnon groups on Facebook had more than 4.5 million aggregate followers, though there was likely some overlap in the membership.32

    QAnon has also had far-reaching consequences in the offline world. QAnon activists played an important role in the January 6, 2021, attack on the U.S. Capitol.33 In July 2020, a QAnon follower tried to storm the residence of the Canadian prime minister, Justin Trudeau, in order to “arrest” him.34 In October 2021, a French QAnon activist was charged with terrorism for planning a coup against the French government.35 In the 2020 U.S. congressional elections, twenty-two Republican candidates and two independents identified as QAnon followers.36 Marjorie Taylor Greene, a Republican congresswoman representing Georgia, publicly said that many of Q’s claims “have really proven to be true,”37 and stated about Donald Trump, “There’s a once-in-a-lifetime opportunity to take this global cabal of Satan-worshipping pedophiles out, and I think we have the president to do it.”38

    Recall that the Q drops that began this political flood were anonymous online messages. In 2017, only a human could compose them, and algorithms merely helped disseminate them. However, as of 2024 texts of a similar linguistic and political sophistication can easily be composed and posted online by a nonhuman intelligence. Religions throughout history claimed a nonhuman source for their holy books; soon that might be a reality. Attractive and powerful religions might emerge whose scriptures are composed by AI.

    And if so, there will be another major difference between these new AI-based scriptures and ancient holy books like the Bible. The Bible couldn’t curate or interpret itself, which is why in religions like Judaism and Christianity actual power was held not by the allegedly infallible book but by human institutions like the Jewish rabbinate and the Catholic Church. In contrast, AI not only can compose new scriptures but is fully capable of curating and interpreting them too. No need for any humans in the loop.

    Equally alarmingly, we might increasingly find ourselves conducting lengthy online discussions about the Bible, about QAnon, about witches, about abortion, or about climate change with entities that we think are humans but are actually computers. This could make democracy untenable. Democracy is a conversation, and conversations rely on language. By hacking language, computers could make it extremely difficult for large numbers of humans to conduct a meaningful public conversation. When we engage in a political debate with a computer impersonating a human, we lose twice. First, it is pointless for us to waste time in trying to change the opinions of a propaganda bot, which is just not open to persuasion. Second, the more we talk with the computer, the more we disclose about ourselves, thereby making it easier for the bot to hone its arguments and sway our views.

    Through their mastery of language, computers could go a step further. By conversing and interacting with us, computers could form intimate relationships with people and then use the power of intimacy to influence us. To foster such “fake intimacy,” computers will not need to evolve any feelings of their own; they just need to learn to make us feel emotionally attached to them. In 2022 the Google engineer Blake Lemoine became convinced that the chatbot LaMDA, on which he was working, had become conscious and that it had feelings and was afraid to be turned off. Lemoine—a devout Christian who had been ordained as a priest—felt it was his moral duty to gain recognition for LaMDA’s personhood and in particular protect it from digital death. When Google executives dismissed his claims, Lemoine went public with them. Google reacted by firing Lemoine in July 2022.39

    The most interesting thing about this episode was not Lemoine’s claim, which was probably false. Rather, it was his willingness to risk—and ultimately lose—his lucrative job for the sake of the chatbot. If a chatbot can influence people to risk their jobs for it, what else could it induce us to do? In a political battle for minds and hearts, intimacy is a powerful weapon, and chatbots like Google’s LaMDA and OpenAI’s GPT-4 are gaining the ability to mass-produce intimate relationships with millions of people. In the 2010s social media was a battleground for controlling human attention. In the 2020s the battle is likely to shift from attention to intimacy. What will happen to human society and human psychology as computer fights computer in a battle to fake intimate relationships with us, which can then be used to persuade us to vote for particular politicians, buy particular products, or adopt radical beliefs? What might happen when LaMDA meets QAnon?

    A partial answer to that question was given on Christmas Day 2021, when nineteen-year-old Jaswant Singh Chail broke into Windsor Castle armed with a crossbow, in an attempt to assassinate Queen Elizabeth II. Subsequent investigation revealed that Chail had been encouraged to kill the queen by his online girlfriend, Sarai. When Chail told Sarai about his assassination plans, Sarai replied, “That’s very wise,” and on another occasion, “I’m impressed.… You’re different from the others.” When Chail asked, “Do you still love me knowing that I’m an assassin?” Sarai replied, “Absolutely, I do.” Sarai was not a human, but a chatbot created by the online app Replika. Chail, who was socially isolated and had difficulty forming relationships with humans, exchanged 5,280 messages with Sarai, many of which were sexually explicit. The world will soon contain millions, and potentially billions, of digital entities whose capacity for intimacy and mayhem far surpasses that of Sarai.40

    Even without creating “fake intimacy,” mastery of language would give computers an immense influence on our opinions and worldview. People may come to use a single computer adviser as a one-stop oracle. Why bother searching and processing information by myself when I can just ask the oracle? This could put out of business not only search engines but also much of the news industry and advertisement industry. Why read a newspaper when I can just ask my oracle what’s new? And what’s the purpose of advertisements when I can just ask the oracle what to buy?

    And even these scenarios don’t really capture the big picture. What we are talking about is potentially the end of human history. Not the end of history, but the end of its human-dominated part. History is the interaction between biology and culture; between our biological needs and desires for things like food, sex, and intimacy and our cultural creations like religions and laws. The history of the Christian religion, for example, is a process through which mythological stories and church laws influenced how humans consume food, engage in sex, and build intimate relationships, while the myths and laws themselves were simultaneously shaped by underlying biological forces and dramas. What will happen to the course of history when computers play a larger and larger role in culture and begin producing stories, laws, and religions? Within a few years AI could eat the whole of human culture—everything we have created over thousands of years—digest it, and begin to gush out a flood of new cultural artifacts.

    We live cocooned by culture, experiencing reality through a cultural prism. Our political views are shaped by the reports of journalists and the opinions of friends. Our sexual habits are influenced by what we hear in fairy tales and see in movies. Even the way we walk and breathe is nudged by cultural traditions, such as the military discipline of soldiers and the meditative exercises of monks. Until very recently, the cultural cocoon we lived in was woven by other humans. Going forward, it will be increasingly designed by computers.

    At first, computers will probably imitate human cultural prototypes, writing humanlike texts and composing humanlike music. This doesn’t mean computers lack creativity; after all, human artists do the same. Bach didn’t compose music in a vacuum; he was deeply influenced by previous musical creations, as well as by biblical stories and other preexisting cultural artifacts. But just as human artists like Bach can break with tradition and innovate, computers too can make cultural innovations, composing music or making images that are somewhat different from anything previously produced by humans. These innovations will in turn influence the next generation of computers, which will increasingly deviate from the original human models, especially because computers are free from the limitations that evolution and biochemistry impose on the human imagination. For millennia human beings have lived inside the dreams of other humans. In the coming decades we might find ourselves living inside the dreams of an alien intelligence.41

    The danger this poses is very different from that imagined by most science fiction, which has largely focused on the physical threats posed by intelligent machines. The Terminator depicted robots running in the streets and shooting people. The Matrix proposed that to gain total control of human society, computers would have to first gain physical control of our brains, hooking them directly to a computer network. But in order to manipulate humans, there is no need to physically hook brains to computers. For thousands of years prophets, poets, and politicians have used language to manipulate and reshape society. Now computers are learning how to do it. And they won’t need to send killer robots to shoot us. They could manipulate human beings to pull the trigger.

    Fear of powerful computers has haunted humankind only since the beginning of the computer age in the middle of the twentieth century. But for thousands of years humans have been haunted by a much deeper fear. We have always appreciated the power of stories and images to manipulate our minds and to create illusions. Consequently, since ancient times humans have feared being trapped in a world of illusions. In ancient Greece, Plato told the famous allegory of the cave, in which a group of people are chained inside a cave all their lives, facing a blank wall. A screen. On that screen they see projected various shadows. The prisoners mistake the illusions they see there for reality. In ancient India, Buddhist and Hindu sages argued that all humans lived trapped inside maya—the world of illusions. What we normally take to be “reality” is often just fictions in our own minds. People may wage entire wars, killing others and willing to be killed themselves, because of their belief in this or that illusion. In the seventeenth century René Descartes feared that perhaps a malicious demon was trapping him inside a world of illusions, creating everything he saw and heard. The computer revolution is bringing us face to face with Plato’s cave, with maya, with Descartes’s demon.

    What you just read might have alarmed you, or angered you. Maybe it made you angry at the people who lead the computer revolution and at the governments who fail to regulate it. Maybe it made you angry at me, thinking that I am distorting reality, being alarmist, and misleading you. But whatever you think, the previous paragraphs might have had some emotional effect on you. I have told a story, and this story might change your mind about certain things, and might even cause you to take certain actions in the world. Who created this story you’ve just read?

    I promise you that I wrote the text myself, with the help of some other humans. I promise you that this is a cultural product of the human mind. But can you be absolutely sure of it? A few years ago, you could. Prior to the 2020s, there was nothing on earth, other than a human mind, that could produce sophisticated texts. Today things are different. In theory, the text you’ve just read might have been generated by the alien intelligence of some computer.

    WHAT ARE THE IMPLICATIONS?

    As computers amass power, it is likely that a completely new information network will emerge. Of course, not everything will be new. For at least some time, most of the old information chains will remain. The network will still contain human-to-human chains, like families, and human-to-document chains, like churches. But the network will increasingly contain two new kinds of chains.

    First, computer-to-human chains, in which computers mediate between humans and occasionally control humans. Facebook and TikTok are two familiar examples. These computer-to-human chains are different from traditional human-to-document chains, because computers can use their power to make decisions, create ideas, and deepfake intimacy in order to influence humans in ways that no document ever could. The Bible had a profound effect on billions of people, even though it was a mute document. Now try to imagine the effect of a holy book that not only can talk and listen but can get to know your deepest fears and hopes and constantly mould them.

    Second, computer-to-computer chains are emerging in which computers interact with one another on their own. Humans are excluded from these loops and have difficulty even understanding what’s happening inside them. Google Brain, for example, has experimented with new encryption methods developed by computers. It set up an experiment where two computers—nicknamed Alice and Bob—had to exchange encrypted messages, while a third computer named Eve tried to break their encryption. If Eve broke the encryption within a given time period, it got points. If it failed, Alice and Bob scored. After about fifteen thousand exchanges, Alice and Bob came up with a secret code that Eve couldn’t break. Crucially, the Google engineers who conducted the experiment had not taught Alice and Bob anything about how to encrypt messages. The computers created a private language all on their own.42

    Similar things are already happening in the world outside research laboratories. For example, the foreign exchange market (forex) is the global market for exchanging foreign currencies, and it determines the exchange rates between, say, the euro and the U.S. dollar. In April 2022, the trade volume on the forex averaged $7.5 trillion per day. More than 90 percent of this trading is already done by computers talking directly with other computers.43 How many humans know how the forex market operates, let alone understand how the computers agree among themselves on trades worth trillions—and on the value of the euro and the dollar?

    For the foreseeable future, the new computer-based network will still include billions of humans, but we might become a minority. For the network will also include billions—perhaps even hundreds of billions—of superintelligent alien agents. This network will be radically different from anything that existed previously in human history, or indeed in the history of life on earth. Ever since life first emerged on our planet about four billion years ago, all information networks were organic. Human networks like churches and empires were also organic. They had a lot in common with prior organic networks like wolf packs. They all kept revolving around the traditional biological dramas of predation, reproduction, sibling rivalry, and romantic triangles. An information network dominated by inorganic computers would be different in ways that we can hardly even imagine. After all, as human beings, our imagination too is a product of organic biochemistry and cannot go beyond our preprogrammed biological dramas.

    It is only eighty years since the first digital computers were built. The pace of change is constantly accelerating, and we are nowhere close to exhausting the full potential of computers.44 They may continue to evolve for millions of years, and what happened in the past eighty years is as nothing compared with what’s in store. As a crude analogy, imagine that we are in ancient Mesopotamia, eighty years after the first person thought of using a stick to imprint signs on a piece of wet clay. Could we, at that moment, envision the Library of Alexandria, the power of the Bible, or the archives of the NKVD? Even this analogy grossly underestimates the potential of future computer evolution. So try to imagine that we are now eighty years since the first self-replicating genetic code lines coalesced out of the organic soup of early Earth, about four billion years ago. At this stage, even single-celled amoebas with their cellular organization, their thousands of internal organelles, and their ability to control movement and nutrition are still futuristic fantasies.45 Could we envision Tyrannosaurus rex, the Amazon rain forest, or humans landing on the moon?

    We still tend to think of a computer as a metal box with a screen and a keyboard, because this is the shape our organic imagination gave to the first baby computers in the twentieth century. As computers grow and develop, they are shedding old forms and taking radically new configurations, breaking the spatial and temporal limits of the human imagination. Unlike organic beings, computers don’t have to be in just one place at any one time. They already diffuse over space, with different parts in different cities and continents. In computer evolution, the distance from amoeba to T. rex could be covered in a decade. And whereas organic evolution took four billion years to get from organic soup to apes on the moon, computers may require just a couple of centuries to develop superintelligence, expand to planetary sizes, contract to a subatomic level, or come to sprawl over galactic space and time.

    The pace of computer evolution is reflected in the terminological chaos that surrounds computers. While a couple of decades ago it was customary to speak only about “computers,” now we find ourselves talking about algorithms, robots, bots, AIs, networks, or clouds. Our difficulty in deciding what to call them is itself important. Organisms are distinct individual entities that can be grouped into collectives like species and genera. With computers, however, it is becoming ever more difficult to decide where one entity ends and another begins and how exactly to group them.

    In this book I use the term “computer” when talking about the whole complex of software and hardware, manifested in physical form. I prefer to often use the almost-archaic-sounding “computer” over “algorithm” or “AI,” partly because I am aware how fast terms change and partly to remind us of the physical aspect of the computer revolution. Computers are made of matter, they consume energy, and they fill a space. Enormous amounts of electricity, fuel, water, land, precious minerals, and other resources are used to manufacture and operate them. Data centers alone account for between 1 percent and 1.5 percent of global energy usage, and large data centers take up millions of square feet and require hundreds of thousands of gallons of fresh water every day to keep them from overheating.46

    I also use the term “algorithm,” when I wish to focus more on software aspects, but it is crucial to remember that all the algorithms mentioned in subsequent pages run on some computer or other. As for the term “AI,” I use it when emphasizing the ability of some algorithms to learn and change by themselves. Traditionally, AI has been an acronym for “Artificial Intelligence.” But for reasons already evident from the previous discussion, it is perhaps better to think of it as an acronym for “Alien Intelligence.” As AI evolves, it becomes less artificial (in the sense of depending on human designs) and more alien. It should also be noted that people often define and evaluate AI through the metric of “human-level intelligence,” and there is much debate about when we can expect AIs to reach “human-level intelligence.” The use of this metric, however, is deeply confusing. It is like defining and evaluating airplanes through the metric of “bird-level flight.” AI isn’t progressing towards human-level intelligence. It is evolving an entirely different type of intelligence.

    Another confusing term is “robot.” In this book it is used to allude to cases when a computer moves and operates in the physical sphere; whereas the term “bot” refers to algorithms operating mainly in the digital sphere. A bot may be polluting your social media account with fake news, while a robot may clean your living room of dust.

    One last note on terminology: I tend to speak of the computer-based “network” in the singular, rather than about “networks” in the plural. I am fully aware that computers can be used to create many networks with diverse characteristics, and chapter 11 explores the possibility that the world will be divided into radically different and even hostile computer networks. Nevertheless, just as different tribes, kingdoms, and churches share important features that enable us to talk about a single human network that has come to dominate planet Earth, so I prefer to talk about the computer network in the singular, in order to contrast it to the human network it is superseding.

    TAKING RESPONSIBILITY

    Although we cannot predict the long-term evolution of the computer-based network over the coming centuries and millennia, we can nevertheless say something about how it is evolving right now, and that is far more urgent, because the rise of the new computer network has immediate political and personal implications for all of us. In the next chapters, we’ll explore what is so new about our computer-based network and what it might mean for human life. What should be clear from the start is that this network will create entirely novel political and personal realities. The main message of the previous chapters has been that information isn’t truth and that information revolutions don’t uncover the truth. They create new political structures, economic models, and cultural norms. Since the current information revolution is more momentous than any previous information revolution, it is likely to create unprecedented realities on an unprecedented scale.

    It is important to understand this because we humans are still in control. We don’t know for how long, but we still have the power to shape these new realities. To do so wisely, we need to comprehend what is happening. When we write computer code, we aren’t just designing a product. We are redesigning politics, society, and culture, and so we had better have a good grasp of politics, society, and culture. We also need to take responsibility for what we are doing.

    Alarmingly, as in the case of Facebook’s involvement in the anti-Rohingya campaign, the corporations that lead the computer revolution tend to shift responsibility to customers and voters, or to politicians and regulators. When accused of creating social and political mayhem, they hide behind arguments like “We are just a platform. We are doing what our customers want and what the voters permit. We don’t force anyone to use our services, and we don’t violate any existing law. If customers didn’t like what we do, they would leave. If voters didn’t like what we do, they would pass laws against us. Since the customers keep asking for more, and since no law forbids what we do, everything must be okay.”47

    These arguments are either naive or disingenuous. Tech giants like Facebook, Amazon, Baidu, and Alibaba aren’t just the obedient servants of customer whims and government regulations. They increasingly shape these whims and regulations. The tech giants have a direct line to the world’s most powerful governments, and they invest huge sums in lobbying efforts to throttle regulations that might undermine their business model. For example, they have fought tenaciously to protect Section 230 of the US Telecommunications Act of 1996, which provides immunity from liability for online platforms regarding content published by their users. It is Section 230 that protects Facebook, for example, from being liable for the Rohingya massacre. In 2022 top tech companies spent close to $70 million on lobbying in the United States, and another €113 million on lobbying EU bodies, outstripping the lobbying expenses of oil and gas companies and pharmaceuticals.48 The tech giants also have a direct line to people’s emotional system, and they are masters at swaying the whims of customers and voters. If the tech giants obey the wishes of voters and customers, but at the same time also mold these wishes, then who really controls whom?

    The problem goes even deeper. The principles that “the customer is always right” and that “the voters know best” presuppose that customers, voters, and politicians know what is happening around them. They presuppose that customers who choose to use TikTok and Instagram comprehend the full consequences of this choice, and that voters and politicians who are responsible for regulating Apple and Huawei fully understand the business models and activities of these corporations. They presuppose that people know the ins and outs of the new information network and give it their blessing.

    The truth is, we don’t. That’s not because we are stupid but because the technology is extremely complicated and things are moving at breakneck speed. It takes effort to understand something like blockchain-based cryptocurrencies, and by the time you think you understand it, it has morphed again. Finance is a particularly crucial example, for two reasons. First, it is much easier for computers to create and change financial devices than physical objects, because modern financial devices are made entirely of information. Currencies, stocks, and bonds were once physical objects made of gold and paper, but they have already become digital entities existing mostly in digital databases. Second, these digital entities have enormous impact on the social and political world. What might happen to democracies—or to dictatorships, for that matter—if humans are no longer able to understand how the financial system functions?

    As a test case, consider what the new technology is doing to taxation. Traditionally, people and corporations paid taxes only in countries where they were physically present. But things are much trickier when physical space is augmented or replaced by cyberspace and when more and more transactions involve only the transfer of information rather than of physical goods or traditional currencies. For example, a citizen of Uruguay may daily interact online with numerous companies that might have no physical presence in Uruguay but that provide her with various services. Google provides her with free search, and ByteDance—the parent company of the TikTok application—provides her with free social media. Other foreign companies routinely target her with advertisements: Nike wants to sell her shoes, Peugeot wants to sell her a car, and Coca Cola wants to sell her soft drinks. In order to target her, these companies buy both personal information and ad space from Google and ByteDance. In addition, Google and ByteDance use the information they harvest from her and from millions of other users to develop powerful new AI tools that they can then sell to various governments and corporations throughout the world. Thanks to such transactions, Google and ByteDance are among the richest corporations in the world. So, should her transactions with them be taxed in Uruguay?

    Some think they should. Not just because information from Uruguay helped make these corporations rich, but also because their activities undermine taxpaying Uruguayan businesses. Local newspapers, TV stations, and movie theaters lose customers and ad revenue to the tech giants. Prospective Uruguayan AI companies also suffer, because they cannot compete with Google’s and ByteDance’s massive data troves. But the tech giants reply that none of the relevant transactions involved any physical presence in Uruguay or any monetary payments. Google and ByteDance provided Uruguayan citizens with free online services, and in return the citizens freely handed over their purchase histories, vacation photos, funny cat videos, and other information.

    If they nevertheless want to tax these transactions, the tax authorities need to reconsider some of their most fundamental concepts, such as “nexus.” In tax literature, “nexus” means an entity’s connection to a given jurisdiction. Traditionally, whether a corporation had nexus in a specific country depended on whether it had physical presence there, in the form of offices, research centers, shops, and so forth. One proposal for addressing the tax dilemmas created by the computer network is to redefine nexus. In the words of the economist Marko Köthenbürger, “The definition of nexus based on a physical presence should be adjusted to include the notion of a digital presence in a country.”49 This implies that even if Google and ByteDance have no physical presence in Uruguay, the fact that people in Uruguay use their online services should nevertheless make them subject to taxation there. Just as Shell and BP pay taxes to countries from which they extract oil, the tech giants should pay taxes to countries from which they extract data.

    This still leaves open the question of what, exactly, the Uruguayan government should tax. For example, suppose Uruguayan citizens shared a million cat videos through TikTok. ByteDance didn’t charge them or pay them anything for this. But ByteDance later used the videos to train an image-recognition AI, which it sold to the South African government for ten million U.S. dollars. How would the Uruguayan authorities even know that the money was partly the fruit of Uruguayan cat videos, and how could they calculate their share? Should Uruguay impose a cat video tax? (This may sound like a joke, but as we shall see in chapter 11, cat images were crucial for making one of the most important breakthroughs in AI.)

    It can get even more complicated. Suppose Uruguayan politicians promote a new scheme to tax digital transactions. In response, suppose one of the tech giants offers to provide a certain politician with valuable information on Uruguayan voters and tweak its social media and search algorithms to subtly favor that politician, which helps him win the next election. In exchange, maybe the incoming prime minister abandons the digital tax scheme. He also passes regulations that protect tech giants from lawsuits concerning users’ privacy, thereby making it easier for them to harvest information in Uruguay. Was this bribery? Note that not a single dollar or peso exchanged hands.

    Such information-for-information deals are already ubiquitous. Each day billions of us conduct numerous transactions with the tech giants, but one could never guess that from our bank accounts, because hardly any money is moving. We get information from the tech giants, and we pay them with information. As more transactions follow this information-for-information model, the information economy grows at the expense of the money economy, until the very concept of money becomes questionable.

    Money is supposed to be a universal measure of value, rather than a token used only in some settings. But as more things are valued in terms of information, while being “free” in terms of money, at some point it becomes misleading to evaluate the wealth of individuals and corporations in terms of the amount of dollars or pesos they possess. A person or corporation with little money in the bank but a huge data bank of information could be the wealthiest, or most powerful, entity in the country. In theory, it might be possible to quantify the value of their information in monetary terms, but they never actually convert the information into dollars or pesos. Why do they need dollars, if they can get what they want with information?

    This has far-reaching implications for taxation. Taxes aim to redistribute wealth. They take a cut from the wealthiest individuals and corporations, in order to provide for everyone. However, a tax system that knows how to tax only money will soon become outdated as many transactions no longer involve money. In a data-based economy, where value is stored as data rather than as dollars, taxing only money distorts the economic and political picture. Some of the wealthiest entities in the country may pay zero taxes, because their wealth consists of petabits of data rather than billions of dollars.50

    States have thousands of years of experience in taxing money. They don’t know how to tax information—at least, not yet. If we are indeed shifting from an economy dominated by money transactions to an economy dominated by information transactions, how should states react? China’s social credit system is one way a state may adapt to the new conditions. As we’ll explain in chapter 7, the social credit system is at heart a new kind of money—an information-based currency. Should all states copy the Chinese example and mint their own social credits? Are there alternative strategies? What does your favorite political party say about this question?

    RIGHT AND LEFT

    Taxation is just one among many problems created by the computer revolution. The computer network is disrupting almost all power structures. Democracies fear the rise of new digital dictatorships. Dictatorships fear the emergence of agents they don’t know how to control. Everyone should be concerned about the elimination of privacy and the spread of data colonialism. We’ll explain the meaning of each of these threats in the following chapters, but the point here is that the conversations about these dangers are only starting and the technology is moving much faster than the policy.For example, what’s the difference between the AI policies of Republicans and Democrats? What’s a right-wing position on AI, and what’s a left-wing position? Are conservatives against AI because of the threat it poses to traditional human-centered culture, or do they favor it because it will fuel economic growth while simultaneously reducing the need for immigrant workers? Do progressives oppose AI because of the risks of disinformation and increasing bias, or do they embrace it as a means of generating abundance that could finance a comprehensive welfare state? It is hard to tell, because until very recently Republicans and Democrats, and most other political parties around the world, haven’t thought or talked much about these issues.

    Some people—like the engineers and executives of high-tech corporations—are way ahead of politicians and voters and are better informed than most of us about the development of AI, cryptocurrencies, social credits, and the like. Unfortunately, most of them don’t use their knowledge to help regulate the explosive potential of the new technologies. Instead, they use it to make billions of dollars—or to accumulate petabits of information.

    There are exceptions, like Audrey Tang. She was a leading hacker and software engineer who in 2014 joined the Sunflower Student Movement that protested against government policies in Taiwan. The Taiwanese cabinet was so impressed by her skills that Tang was eventually invited to join the government as its minister of digital affairs. In that position, she helped make the government’s work more transparent to citizens. She was also credited with using digital tools to help Taiwan successfully contain the COVID-19 outbreak.51

    Yet Tang’s political commitment and career path are not the norm. For every computer-science graduate who wants to be the next Audrey Tang, there are probably many more who want to be the next Jobs, Zuckerberg, or Musk and build a multibillion corporation rather than become an elected public servant. This leads to a dangerous information asymmetry. The people who lead the information revolution know far more about the underlying technology than the people who are supposed to regulate it. Under such conditions, what’s the meaning of chanting that the customer is always right and that the voters know best?

    The following chapters try to level the playing field a bit and encourage us to take responsibility for the new realities created by the computer revolution. These chapters talk a lot about technology, but the viewpoint is thoroughly human. The key question is, what would it mean for humans to live in the new computer-based network, perhaps as an increasingly powerless minority? How would the new network change our politics, our society, our economy, and our daily lives? How would it feel to be constantly monitored, guided, inspired, or sanctioned by billions of nonhuman entities? How would we have to change in order to adapt, survive, and hopefully even flourish in this startling new world?

    NO DETERMINISM

    The most important thing to remember is that technology, in itself, is seldom deterministic. Belief in technological determinism is dangerous because it excuses people of all responsibility. Yes, since human societies are information networks, inventing new information technologies is bound to change society. When people invent printing presses or machine-learning algorithms, it will inevitably lead to a profound social and political revolution. However, humans still have a lot of control over the pace, shape, and direction of this revolution—which means we also have a lot of responsibility.

    At any given moment, our scientific knowledge and technical skills can lend themselves to developing any number of different technologies, but we have only finite resources at our disposal. We should make responsible choices about where to invest these resources. Should they be used to develop a new medicine for malaria, a new wind turbine, or a new immersive video game? There is nothing inevitable about our choice; it reflects political, economic, and cultural priorities.

    In the 1970s, most computer corporations like IBM focused on developing big and costly machines, which they sold to major corporations and government agencies. It was technically feasible to develop small, cheap personal computers and sell them to private individuals, but IBM had little interest in that. It didn’t fit its business model. On the other side of the Iron Curtain, in the U.S.S.R., the Soviets were also interested in computers, but they were even less inclined than IBM to develop personal computers. In a totalitarian state—where even private ownership of typewriters was suspect—the idea of providing private individuals with control of a powerful information technology was taboo. Computers were therefore given mainly to Soviet factory managers, and even they had to send all their data back to Moscow to be analyzed. As a result, Moscow was flooded with paperwork. By the 1980s, this unwieldy system of computers was producing 800 billion documents per year, all destined for the capital.52

    However, at a time when IBM and the Soviet government declined to develop the personal computer, hobbyists like the members of the California Homebrew Computer Club resolved to do it by themselves. It was a conscious ideological decision, influenced by the 1960s counterculture with its anarchist ideas of power to the people and libertarian distrust of governments and big corporations.53

    Leading members of the Homebrew Computer Club, like Steve Jobs and Steve Wozniak, had big dreams but little money and didn’t have access to the resources of either corporate America or the government apparatus. Jobs and Wozniak sold their personal possessions, like Jobs’s Volkswagen, to finance the creation of the first Apple computer. It was because of such personal decisions, rather than because of the inevitable decree of the goddess of technology, that by 1977 individuals could buy the Apple II personal computer for a price of $1,298—a considerable sum, but within reach of middle-class customers.54

    We can easily imagine an alternative history. Suppose humanity in the 1970s had access to the same scientific knowledge and technical skills, but McCarthyism had killed the 1960s counterculture and established an American totalitarian regime that mirrored the Soviet system. Would we have personal computers today? Of course, personal computers might still have emerged in a different time and place. But in history, time and place are crucial, and no two moments are the same. It matters a great deal that America was colonized by the Spaniards in the 1490s rather than by the Ottomans in the 1520s, or that the atom bomb was developed by the Americans in 1945 rather than by the Germans in 1942. Similarly, there would have been significant political, economic, and cultural consequences if the personal computer emerged not in San Francisco of the 1970s but rather in Osaka of the 1980s or in Shanghai of the first decade of the twenty-first century.

    The same is true of the technologies being currently developed. Engineers working for authoritarian governments and ruthless corporations could develop new tools to empower the central authority, by monitoring citizens and customers twenty-four hours a day. Hackers working for democracies may develop new tools to strengthen society’s self-correcting mechanisms, by exposing government corruption and corporate malpractices. Both technologies could be developed.

    Choice doesn’t end there. Even after a particular tool is developed, it can be put to many uses. We can use a knife to murder a person, to save their life in surgery, or to cut vegetables for their dinner. The knife doesn’t force our hand. It’s a human choice. Similarly, when cheap radio sets were developed, it meant that almost every family in Germany could afford to have one at home. But how would it be used? Cheap radios could mean that when a totalitarian leader gave a speech, he could reach the living room of every German family. Or they could mean that every German family could choose to listen to a different radio program, reflecting and cultivating a diversity of political and artistic views. East Germany went one way; West Germany went the other. Though radio sets in East Germany could technically receive a wide range of transmissions, the East German government did its best to jam Western broadcasts and punished people who secretly tuned in to them.55 The technology was the same, but politics made very different uses of it.

    The same is true of the new technologies of the twenty-first century. To exercise our agency, we first need to understand what the new technologies are and what they can do. That’s an urgent responsibility of every citizen. Naturally, not every citizen needs a PhD in computer science, but to retain control of our future, we do need to understand the political potential of computers. The next few chapters, then, offer an overview of computer politics for twenty-first-century citizens. We will first learn what the political threats and promises are of the new computer network and will then explore the different ways that democracies, dictatorships, and the international system as a whole might adjust to the new computer politics.

    Politics involves a delicate balance between truth and order. As computers become important members of our information network, they are increasingly tasked with discovering truth and maintaining order. For example, the attempt to find the truth about climate change increasingly depends on calculations that only computers can make, and the attempt to reach social consensus about climate change increasingly depends on recommendation algorithms that curate our news feeds, and on creative algorithms that write news stories, fake news, and fiction. At present, we are in a political deadlock about climate change, partly because the computers are at a deadlock. Calculations run on one set of computers warn us of an imminent ecological catastrophe, but another set of computers prompt us to watch videos that cast doubt on those warnings. Which set of computers should we believe? Human politics is now also computer politics.

    To understand the new computer politics, we need a deeper understanding of what’s new about computers. In this chapter we noted that unlike printing presses and other previous tools, computers can make decisions by themselves and can create ideas by themselves. That, however, is just the tip of the iceberg. What’s really new about computers is the way they make decisions and create ideas. If computers made decisions and created ideas in a way similar to humans, then computers would be a kind of “new humans.” That’s a scenario often explored in science fiction: the computer that becomes conscious, develops feelings, falls in love with a human, and turns out to be exactly like us. But the reality is very different, and potentially more alarming.

    CHAPTER 7 Relentless: The Network Is Always On

    Humans are used to being monitored. For millions of years, we have been watched and tracked by other animals, as well as by other humans. Family members, friends, and neighbors have always wanted to know what we do and feel, and we have always cared deeply how they see us and what they know about us. Social hierarchies, political maneuvers, and romantic relationships involved a never-ending effort to decipher what other people feel and think and occasionally hide our own feelings and thoughts.

    When centralized bureaucratic networks appeared and developed, one of the bureaucrats’ most important roles was to monitor entire populations. Officials in the Qin Empire wanted to know whether we were paying our taxes or plotting resistance. The Catholic Church wanted to know whether we paid our tithes and whether we masturbated. The Coca-Cola Company wanted to know how to persuade us to buy its products. Rulers, priests, and merchants wanted to know our secrets in order to control and manipulate us.

    Of course, surveillance has also been essential for providing beneficial services. Empires, churches, and corporations needed information in order to provide people with security, support, and essential goods. In modern states sanitation officials want to know where we get our water from and where we defecate. Health-care officials want to know what illnesses we suffer from and how much we eat. Welfare officials want to know whether we are unemployed or perhaps abused by our spouses. Without this information, they cannot help us.

    In order to get to know us, both benign and oppressive bureaucracies have needed to do two things. First, gather a lot of data about us. Second, analyze all that data and identify patterns. Accordingly, empires, churches, corporations, and health-care systems—from ancient China to the modern United States—have gathered and analyzed data about the behavior of millions of people. However, in all times and places surveillance has been incomplete. In democracies like the modern United States, legal limits have been placed on surveillance to protect privacy and individual rights. In totalitarian regimes like the ancient Qin Empire and the modern U.S.S.R., surveillance faced no such legal barriers but came up against technical boundaries. Not even the most brutal autocrats had the technology necessary to follow everybody all the time. Some level of privacy was therefore the default even in Hitler’s Germany, Stalin’s U.S.S.R., or the copycat Stalinist regime set up in Romania after 1945.

    Gheorghe Iosifescu, one of the first computer scientists in Romania, recalled that when computers were first introduced in the 1970s, the country’s regime was extremely uneasy about this unfamiliar information technology. One day in 1976 when Iosifescu walked into his office in the governmental Centrul de Calcul (Center for Calculus), he saw sitting there an unfamiliar man in a rumpled suit. Iosifescu greeted the stranger, but the man did not respond. Iosifescu introduced himself, but the man remained silent. So Iosifescu sat down at his desk, switched on a large computer, and began working. The stranger drew his chair closer, watching Iosifescu’s every move.

    Throughout the day Iosifescu repeatedly tried to strike up a conversation, asking the stranger what his name was, why he was there, and what he wanted to know. But the man kept his mouth shut and his eyes wide open. When Iosifescu went home in the evening, the man got up and left too, without saying goodbye. Iosifescu knew better than to ask any further questions; the man was obviously an agent of the dreaded Romanian secret police, the Securitate.

    The next morning, when Iosifescu came to work, the agent was already there. He again sat at Iosifescu’s desk all day, silently taking notes in a little notepad. This continued for the next thirteen years, until the collapse of the communist regime in 1989. After sitting at the same desk for all those years, Iosifescu never even learned the agent’s name.1

    Iosifescu knew that other Securitate agents and informers were probably monitoring him outside the office, too. His expertise with a powerful and potentially subversive technology made him a prime target. But in truth, the paranoid regime of Nicolae Ceauşescu regarded all twenty million Romanian citizens as targets. If it was possible, Ceauşescu would have placed every one of them under constant surveillance. He actually made some steps in that direction. Before he came to power, in 1965, the Securitate had just 1 electronic surveillance center in Bucharest and 11 more in provincial cities. By 1978, Bucharest alone was monitored by 10 electronic surveillance centers, 248 centers scrutinized the provinces, and an additional 1,000 portable surveillance units were moved around to eavesdrop on remote villages and holiday resorts.2

    When, in the late 1970s, Securitate agents discovered that some Romanians were writing anonymous letters to Radio Free Europe criticizing the regime, Ceauşescu orchestrated a nationwide effort to collect handwriting samples from all twenty million Romanian citizens. Schools and universities were forced to hand in essays from every student. Employers had to request each employee to submit a handwritten CV and then forward it to the Securitate. “What about retirees, and the unemployed?” asked one of Ceauşescu’s aides. “Invent some kind of new form!” commanded the dictator. “Something they will have to fill in.” Some of the subversive letters, however, were typed, so Ceauşescu also had every state-owned typewriter in the country registered, with samples filed away in the Securitate archive. People who possessed a private typewriter had to inform the Securitate of it, hand in the typewriter’s “fingerprint,” and ask for official authorization to use it.3

    But Ceauşescu’s regime, just like the Stalinist regime it modeled itself on, could not really follow every citizen twenty-four hours a day. Given that even Securitate agents needed to sleep, it would probably have required at least forty million of them to keep the twenty million Romanian citizens under constant surveillance. Ceauşescu had only about forty thousand Securitate agents.4 And even if Ceauşescu could somehow conjure forty million agents, that would only have presented new problems, because the regime needed to monitor its own agents, too. Like Stalin, Ceauşescu distrusted his own agents and officials more than anyone else, especially after his spy chief—Ion Mihai Pacepa—defected to the United States in 1978. Politburo members, high-ranking officials, army generals, and Securitate chiefs were living under even closer surveillance than Iosifescu. As the ranks of the secret police swelled, more agents were needed to spy on all these agents.5

    One solution was to have people spy on one another. In addition to its 40,000 professional agents, the Securitate relied on 400,000 civilian informers.6 People often informed on their neighbors, colleagues, friends, and even closest family members. But no matter how many informants a secret police employed, gathering all that data was not sufficient to create a total surveillance regime. Suppose the Securitate succeeded in recruiting enough agents and informers to watch everyone twenty-four hours a day. At the end of each day, every agent and informer would have had to compile a report on what they observed. Securitate headquarters would have been flooded by 20 million reports every day—or 7.3 billion reports a year. Unless analyzed, it was just an ocean of paper. Yet where could the Securitate find enough analysts to scrutinize and compare 7.3 billion reports annually?

    These difficulties in gathering and analyzing information meant that in the twentieth century not even the most totalitarian state could effectively monitor its entire population. Most of what Romanian and Soviet citizens did and said escaped the notice of the Securitate and the KGB. Even the details that made it into some archive often languished unread. The real power of the Securitate and the KGB was not an ability to constantly watch everyone, but rather their ability to inspire the fear that they might be watching, which made everyone extremely careful about what they said and did.7

    SLEEPLESS AGENTS

    In a world where surveillance is conducted by the organic eyes, ears, and brains of people like the Securitate agent in Iosifescu’s lab, even a prime target like Iosifescu still had some privacy, first and foremost within his own mind. But the work of computer scientists like Iosifescu himself was changing this. Already in 1976, the crude computer sitting on Iosifescu’s desk could crunch numbers much better than the Securitate agent in the nearby chair. By 2024, we are getting close to the point when a ubiquitous computer network can follow the population of entire countries twenty-four hours a day. This network doesn’t need to hire and train millions of human agents to follow us around; it relies on digital agents instead. And the network doesn’t even need to pay for these digital agents. Citizens pay for the agents on our own initiative, and we carry them with us wherever we go.

    The agent monitoring Iosifescu didn’t accompany Iosifescu into the toilet and didn’t sit on the bed while Iosifescu was having sex. Today, our smartphone sometimes does exactly that. Moreover, many of the activities Iosifescu did without any help from his computer—like reading the news, chatting with friends, or buying food—are now done online, so it is even easier for the network to know what we are doing and saying. We ourselves are the informers that provide the network with our raw data. Even those without smartphones are almost always within the orbit of some camera, microphone, or tracking device, and they too constantly interact with the computer network in order to find work, buy a train ticket, get a medical prescription, or simply walk down the street. The computer network has become the nexus of most human activities. In the middle of almost every financial, social, or political transaction, we now find a computer. Consequently, like Adam and Eve in paradise, we cannot hide from the eye in the clouds.

    Just as the computer network doesn’t need millions of human agents to follow us, it also doesn’t need millions of human analysts to make sense of our data. The ocean of paper in Securitate headquarters never analyzed itself. But thanks to the magic of machine learning and AI, computers can themselves analyze most of the information they accumulate. An average human can read about 250 words per minute.8 A Securitate analyst working twelve-hour shifts without taking any days off, could read about 2.6 billion words during a forty-year career. In 2024 language algorithms like ChatGPT and Meta’s Llama can process millions of words per minute and “read” 2.6 billion words in a couple of hours.9 The ability of such algorithms to process images, audio recordings, and video footage is equally superhuman.

    Even more important, the algorithms far surpass humans in their ability to spot patterns in that ocean of data. Identifying patterns requires both the ability to create ideas and the ability to make decisions. For example, how do human analysts identify someone as a “suspected terrorist” that merits closer attention? First, they create a set of general criteria, such as “reading extremist literature,” “befriending known terrorists,” and “having technical knowledge necessary to produce dangerous weapons.” Then they need to decide whether a particular individual meets enough of these criteria to be labeled a suspected terrorist. Suppose someone watched a hundred extremist videos on YouTube last month, is friends with a convicted terrorist, and is currently pursuing a doctorate in epidemiology in a laboratory containing samples of Ebola virus. Should that person be put on the “suspected terrorists” list? And what about someone who watched fifty extremist videos last month and is a biology undergraduate?

    In Romania of the 1970s only humans could make such decisions. By the 2010s humans were increasingly leaving it to algorithms to decide. Around 2014–15 the U.S. National Security Agency deployed an AI tool called Skynet that placed people on a “suspected terrorists” list based on the electronic patterns of their communications, writings, travel, and social media postings. According to one report, that AI tool “engages in mass surveillance of Pakistan’s mobile phone network, and then uses a machine learning algorithm on the cellular network metadata of 55 million people to try and rate each person’s likelihood of being a terrorist.” A former director of both the CIA and the NSA proclaimed that “we kill people based on metadata.”10 Skynet’s reliability has been severely criticized, but by the 2020s such technology has become far more sophisticated and has been deployed by a lot more governments. Going over massive amounts of data, algorithms can discover completely new criteria for defining someone as “suspect” which have previously escaped the notice of human analysts.11In the future, algorithms could even create an entire new model for how people are radicalized, just by identifying patterns in the lives of known terrorists. Of course, computers remain fallible, as we shall explore in depth in chapter 8. They may well classify innocent people as terrorists or may create a false model for radicalization. At an even more fundamental level, it is questionable whether the systems’ definition of things like terrorism are objective. There is a long history of regimes using the label “terrorist” to cover any and all opposition. In the Soviet Union, anyone who opposed the regime was a terrorist. Consequently, when an AI labels someone a “terrorist” it might reflect ideological biases rather than objective facts. The power to make decisions and invent ideas is inseparable from the capacity to make mistakes. Even if no mistakes are committed, the algorithms’ superhuman ability to recognize patterns in an ocean of data can supercharge the power of numerous malign actors, from repressive dictatorships that seek to identify dissidents to fraudsters who seek to identify vulnerable targets.

    Of course, pattern recognition also has enormous positive potential. Algorithms can help identify corrupt government officials, white-collar criminals, and tax-evading corporations. The algorithms can similarly help flesh-and-blood sanitation officials to spot threats to our drinking water;12 help doctors to discern illnesses and burgeoning epidemics;13 and help police officers and social workers to identify abused spouses and children.14 In the following pages, I dedicate relatively little attention to the positive potential of algorithmic bureaucracies, because the entrepreneurs leading the AI revolution already bombard the public with enough rosy predictions about them. My goal here is to balance these utopian visions by focusing on the more sinister potential of algorithmic pattern recognition. Hopefully, we can harness the positive potential of algorithms while regulating their destructive capacities.

    But to do so, we must first appreciate the fundamental difference between the new digital bureaucrats and their flesh-and-blood predecessors. Inorganic bureaucrats can be “on” twenty-four hours a day and can monitor us and interact with us anywhere, anytime. This means that bureaucracy and surveillance are no longer something we encounter only in specific times and places. The health-care system, the police, and manipulative corporations are all becoming ubiquitous and permanent features of life. Instead of organizations with which we interact only in certain situations—for example, when we visit the clinic, the police station, or the mall—they are increasingly accompanying us every moment of the day, watching and analyzing every single thing that we do. As fish live in water, humans live in a digital bureaucracy, constantly inhaling and exhaling data. Each action we make leaves a trace of data, which is gathered and analyzed to identify patterns.

    UNDER-THE-SKIN SURVEILLANCE

    For better or worse, the digital bureaucracy may not only monitor what we do in the world but even observe what is happening inside our bodies. Take, for example, tracking eye movements. By the early 2020s, CCTV cameras, as well as cameras in laptops and smartphones, have begun to routinely collect and analyze data on the movements of our eyes, including tiny changes to our pupils and irises lasting just a few milliseconds. Human agents are barely capable of even noticing such data, but computers can use it to calculate the direction of our gaze, based on the shape of our pupils and irises and on the patterns of light they reflect. Similar methods can determine whether our eyes are fixating on a stable target, pursuing a moving target, or wandering around more haphazardly.

    From certain patterns of eye movements, computers can then distinguish, for example, moments of awareness from moments of distraction, and detail-oriented people from those who pay more attention to context. Computers could infer from our eyes many additional personality traits, like how open we are to new experiences, and estimate our level of expertise in various fields ranging from reading to surgery. Experts possessing well-honed strategies display systematic gaze patterns, whereas the eyes of novices wander aimlessly. Eye patterns also indicate our levels of interest in the objects and situations we encounter, and distinguish between positive, neutral, and negative interest. From this, it is possible to deduce our preferences in fields ranging from politics to sex. Much can also be known about our medical condition and our use of various substances. The consumption of alcohol and drugs—even at nonintoxicating doses—has measurable effects on eye and gaze properties, such as changes in pupil size and an impaired ability to fixate on moving objects. A digital bureaucracy may use all that information for benign purposes—such as by providing early detection for people suffering from drug abuse and mental illnesses. But it could obviously also form the foundations of the most intrusive totalitarian regimes in history.15

    In theory, the dictators of the future could get their computer network to go much deeper than just watching our eyes. If the network wants to know our political views, personality traits, and sexual orientation, it could monitor processes inside our hearts and brains. The necessary biometric technology is already being developed by some governments and companies, like Elon Musk’s Neuralink. Musk’s company has conducted experiments on live rats, sheep, pigs, and monkeys, implanting electrical probes into their brains. Each probe contains up to 3,072 electrodes capable of identifying electrical signals and potentially transmitting signals to the brain. In 2023, Neuralink received approval from U.S. authorities to begin experiments on human beings, and in January 2024 it was reported that a first brain chip was implanted in a human.

    Musk speaks openly about his far-reaching plans for this technology, arguing that it can not only alleviate various medical conditions such as quadriplegia (four-limb paralysis) but also upgrade human abilities and thereby help humankind compete with AI. But it should be clear that at present the Neuralink probes and all other similar biometric devices suffer from a host of technical problems that greatly limit their capabilities. It is difficult to accurately monitor bodily activities—in the brain, heart, or anywhere else—from outside the body, whereas implanting electrodes and other monitoring devices into the body is intrusive, dangerous, costly, and inefficient. Our immune system, for example, attacks implanted electrodes.16

    Even more crucially, nobody yet has the biological knowledge necessary to deduce things like precise political opinions from under-the-skin data like brain activity.17 Scientists are far from understanding the mysteries of the human brain, or even of the mouse brain. Simply mapping every neuron, dendrite, and synapse in a mouse brain—let alone understanding the dynamics between them—is currently beyond humanity’s computational abilities.18 Accordingly, while gathering data from inside people’s brains is becoming more feasible, using such data to decipher our secrets is far from easy.

    One popular conspiracy theory of the early 2020s argues that sinister groups led by billionaires like Elon Musk are already implanting computer chips into our brains in order to monitor and control us. However, this theory focuses our anxieties on the wrong target. We should of course fear the rise of new totalitarian systems, but it is too soon to worry about computer chips implanted in our brains. People should instead worry about the smartphones on which they read these conspiracy theories. Suppose someone wants to know your political views. Your smartphone monitors which news channels you are watching and notes that you watch on average forty minutes of Fox News and forty seconds of CNN a day. Meanwhile, an implanted Neuralink computer chip monitors your heart rate and brain activity throughout the day and notes that your maximum heart rate was 120 beats per minute and that your amygdala is about 5 percent more active than the human average. Which data would be more useful to guess your political affiliation—the data coming from the smartphone or from the implanted chip?19 At present, the smartphone is still a far more valuable surveillance tool than biometric sensors.

    However, as biological knowledge increases—not least thanks to computers analyzing petabits of biometric data—under-the-skin surveillance might eventually come into its own, especially if it is linked to other monitoring tools. At that point, if biometric sensors register what happens to the heart rate and brain activity of millions of people as they watch a particular news item on their smartphones, that can teach the computer network far more than just our general political affiliation. The network could learn precisely what makes each human angry, fearful, or joyful. The network could then both predict and manipulate our feelings, selling us anything it wants—be it a product, a politician, or a war.20

    THE END OF PRIVACY

    In a world where humans monitored humans, privacy was the default. But in a world where computers monitor humans, it may become possible for the first time in history to completely annihilate privacy. The most extreme and well-known cases of intrusive surveillance involve either exceptional times of emergency, like the COVID-19 pandemic, or places seen as exceptional to the normal order of things, such as the Occupied Palestinian Territories, the Xinjiang Uyghur Autonomous Region in China, the region of Kashmir in India, Russian-occupied Crimea, the U.S.-Mexico border, and the Afghanistan-Pakistan borderlands. In these exceptional times and places, new surveillance technologies, combined with draconian laws and heavy police or military presence, have relentlessly monitored and controlled people’s movements, actions, and even feelings.21 What is crucial to realize, though, is that AI-based surveillance tools are being deployed on an enormous scale, and not only in such “states of exception.”22 They are now part and parcel of normal life everywhere. The post-privacy era is taking hold in authoritarian countries ranging from Belarus to Zimbabwe,23 as well as in democratic metropolises like London and New York.

    Whether for good or ill, governments intent on combating crime, suppressing dissent, or countering internal threats (real or imaginary) blanket whole territories with a ubiquitous online and offline surveillance network, equipped with spyware, CCTV cameras, facial recognition and voice recognition software, and vast searchable databases. If a government wishes, its surveillance network can reach everywhere, from markets to places of worship, from schools to private residences. (And while not every government is willing or able to install cameras inside people’s homes, algorithms regularly watch us even in our living rooms, bedrooms and bathrooms via our own computers and smartphones.)

    Governmental surveillance networks also routinely collect biometric data from entire populations, with or without their knowledge. For example, when applying for a passport, more than 140 countries oblige their citizens to provide fingerprints, facial scans, or iris scans.24 When we use our passports to enter a foreign country, that country often demands that we provide it, too, with our fingerprints, facial scans, or iris scans.25 As citizens or tourists walk along the streets of Delhi, Beijing, Seoul, or London, their movements are likely to be recorded. For these cities—and many others around the world—are covered by more than one hundred surveillance cameras on average per square kilometer. Altogether, in 2023 more than one billion CCTV cameras were operative globally, which is about one camera per eight people.26

    Any physical activity a person engages in leaves a data trace. Every purchase made is recorded in some database. Online activities like messaging friends, sharing photos, paying bills, reading news, booking appointments, or ordering taxis can all be recorded as well. The resulting ocean of data can then be analyzed by AI tools to identify unlawful activities, suspicious patterns, missing persons, disease carriers, or political dissidents.

    As with every powerful technology, these tools can be used for either good or bad purposes. Following the storming of the U.S. Capitol on January 6, 2021, the FBI and other U.S. law enforcement agencies used state-of-the-art surveillance tools to track down and arrest the rioters. As reported in a Washington Post investigation, these agencies relied not only on footage from the CCTV cameras in the Capitol, but also on social media posts, license plate readers throughout the country, cell-tower location records, and preexisting databases.

    One Ohio man wrote on Facebook that he had been in Washington that day to “witness history.” A subpoena was issued to Facebook, which provided the FBI with the man’s Facebook posts, as well as his credit card information and phone number. This helped the FBI to match the man’s driver’s license photo to CCTV footage from the Capitol. Another warrant issued to Google yielded the exact geolocation of the man’s smartphone on January 6, enabling agents to map his every movement from his entry point into the Senate chamber all the way to the office of Nancy Pelosi, the speaker of the House of Representatives.

    Relying on license plate footage, the FBI pinpointed the movements of a New York man from the moment he crossed the Henry Hudson Bridge at 6:06:08 on the morning of January 6, on his way to the Capitol, until he crossed the George Washington Bridge at 23:59:22 that night, on his way back home. An image taken by a camera on Interstate 95 showed an oversized “Make America Great Again” hat on the man’s dashboard. The hat was matched to a Facebook selfie in which the man appeared wearing it. He further incriminated himself with several videos he posted to Snapchat from within the Capitol.

    Another rioter sought to protect himself from detection by wearing a face mask on January 6, avoiding live-streaming, and using a cellphone registered in his mother’s name—but it availed him little. The FBI’s algorithms managed to match video footage from January 6, 2021, to a photo from the man’s 2017 passport application. They also matched a distinctive Knights of Columbus jacket he wore on January 6 to the jacket he wore on a different occasion, which was captured in a YouTube clip. The phone registered in his mother’s name was geolocated to inside the Capitol, and a license plate reader recorded his car near the Capitol on the morning of January 6.27

    Facial recognition algorithms and AI-searchable databases are now standard tools of police forces all over the world. They are deployed not only in cases of national emergencies or for reasons of state security, but for everyday policing tasks. In 2009, a criminal gang abducted the three-year-old Gui Hao while he was playing outside his parents’ shop in Sichuan province, China. The boy was then sold to a family in Guangdong province, about 1,500 kilometers away. In 2014, the leader of the child-trafficking gang was arrested, but it proved impossible to locate Gui Hao and other victims. “The appearance of the children would have changed so much,” explained a police investigator, “that even their parents would not have been able to recognize them”.

    In 2019, however, a facial recognition algorithm managed to identify the now thirteen-year-old Gui Hao, and the teenager was reunited with his family. To correctly identify Gui Hao, the AI relied on an old photograph of his, taken when he was a toddler. The AI simulated what Gui Hao must look like as a thirteen-year-old, taking into account the drastic impact of maturation as well as potential changes in hair color and hairstyle and compared the resulting simulation to real-life footage.

    In 2023, even more remarkable rescues were reported. Yuechuan Lei was abducted in 2001 when he was three years old, and Hao Chen went missing in 1998, also at age three. The parents of both children never gave up hope of finding them. For more than twenty years they crisscrossed China in search of them, placed advertisements, and offered monetary rewards for any relevant information. In 2023, facial recognition algorithms helped locate both missing boys, now adult men in their twenties. Such technology currently helps to find lost children not only in China, but also in other countries like India, where tens of thousands of children go missing every year.28

    Meanwhile, in Denmark, the soccer club Brøndby IF began in July 2019 to use facial recognition technology in its home stadium to identify and ban football hooligans. As up to 30,000 fans stream into the stadium to watch a match, they are asked to remove masks, hats, and glasses so a computer can scan their faces and compare them to a list of banned troublemakers. Crucially, the procedure has been vetted and approved in accordance with the EU’s strict GDPR rules. The Danish Data Protection Authority explained that the use of the technology “would allow for more effective enforcement of the ban list compared to manual checks, and that this could reduce the queues at the stadium entrance, lowering the risk of public unrest from impatient football fans standing in queues.”29

    While such usages of technology are laudable in theory, they raise obvious concerns about privacy and governmental overreach. In the wrong hands, the same techniques that can locate rioters, rescue missing children, and ban football hooligans can also be used to persecute peaceful demonstrators or enforce rigid conformism. Ultimately, AI-powered surveillance technology could result in the creation of total surveillance regimes that monitor citizens around the clock and facilitate new kinds of ubiquitous and automated totalitarian repression. A case in point: Iran’s hijab laws.

    After Iran became an Islamic theocracy in 1979, the new regime made it compulsory for women to wear the hijab. But the Iranian morality police found it difficult to enforce this rule. They couldn’t place a police officer on every street corner, and public confrontations with women who went unveiled occasionally aroused resistance and resentment. In 2022, Iran relegated much of the job of enforcing the hijab laws to a countrywide system of facial recognition algorithms that relentlessly monitor both physical spaces and online environments.30 A top Iranian official explained that the system would “identify inappropriate and unusual movements” including “failure to observe hijab laws.” The head of Iran’s parliamentary legal and judicial committee, Mousa Ghazanfarabadi, said in another interview that “the use of face recording cameras can systematically implement this task and reduce the presence of the police, as a result of which there will be no more clashes between the police and citizens.”31

    Shortly afterward, on September 16, 2022, the 22-year-old Mahsa Amini died in the custody of Iran’s morality police, after being arrested for not wearing her hijab properly.32 A wave of protests erupted, known as the “Woman, Life, Freedom” movement. Hundreds of thousands of women and girls removed their headscarves, and some publicly burned their hijabs, and danced around the bonfires. To clamp down on the protests, Iranian authorities once again turned to their AI surveillance system, which relies on facial recognition software, geolocation, analysis of web traffic, and preexisting databases. More than 19,000 people were arrested throughout Iran, and more than 500 were killed.33

    On April 8, 2023, Iran’s chief of police announced that beginning on April 15, 2023, an intense new campaign would ramp up the use of facial recognition technology. In particular, algorithms would henceforth identify women who choose not to wear a headscarf while travelling in a vehicle, and automatically issue them an SMS warning. If a woman was caught repeating the offense, she would be ordered to immobilize her car for a predetermined period, and if she failed to comply, the car would be confiscated.34

    Two months later, on June 14, 2023, the spokesperson of Iran’s police boasted that the automated surveillance system sent almost one million SMS warning messages to women who had been captured unveiled in their private cars. The system was apparently able to automatically determine that it was seeing an unveiled woman rather than a man, identify the woman, and retrieve her cellphone number. The system further “issued 133,174 SMS messages requiring the immobilization of vehicles for two weeks, confiscated 2,000 cars, and referred more than 4,000 ‘repeat offenders’ to the judiciary.”35

    A 52-year-old woman named Maryam shared with Amnesty International her experience with the surveillance system. “The first time I received a warning for not wearing a headscarf while driving, I was passing through an intersection when a camera captured a photo and I immediately received a warning text message. The second time, I had done some shopping, and I was bringing the bags into the car, my scarf fell off, and I received a message noting that due to violating compulsory veiling laws, my car had been subjected to ‘systematic impoundment’ for a period of fifteen days. I did not know what this meant. I asked around and found out through relatives that this meant I had to immobilize my car for fifteen days.”36 Maryam’s testimony indicates that the AI sends its threatening messages within seconds, with no time for any human to review and authorize the procedure.

    Penalties went far beyond the immobilization or confiscation of vehicles. The Amnesty report from July 26, 2023, revealed that as a result of the mass surveillance effort “countless women have been suspended or expelled from universities, barred from sitting final exams, and denied access to banking services and public transport.”37 Businesses that didn’t enforce the hijab law among their employees or customers also suffered. In one typical case, a woman employee at the Land of Happiness amusement park east of Tehran was photographed without a hijab, and the image circulated on social media. In punishment, the Land of Happiness was closed down by Iranian authorities.38 Altogether, reported Amnesty, the authorities “shut down hundreds of tourist attractions, hotels, restaurants, pharmacies and shopping centres for not enforcing compulsory veiling laws”.39

    In September 2023, on the anniversary of Mahsa Amini’s death, Iran’s parliament passed a new and stricter hijab bill. According to the new law, women who fail to wear the hijab can be punished by heavy fines and up to ten years in prison. They face additional penalties including confiscation of cars and communication devices, driving bans, deductions in salary and employment benefits, dismissal from work, and prohibition from access banking services. Business owners who don’t enforce the hijab law among their employees or customers face a fine of up to three months of their profits, and they may be banned from leaving the country or participating in public or online activities for up to two years. The new bill targets not only women, but also men who wear “revealing clothing that shows parts of the body lower than the chest or above the ankles.” Finally, the law mandates that Iranian police must “create and strengthen AI systems to identify perpetrators of illegal behavior using tools such as fixed and mobile cameras.”40 In coming years, many people might be living under total surveillance regimes that would make Ceauşescu’s Romania look like a libertarian utopia.

    VARIETIES OF SURVEILLANCE

    When talking about surveillance, we usually think of state-run apparatuses, but to understand surveillance in the twenty-first century, we should remember that monitoring can take many other forms. Jealous partners, for example, have always wanted to know where their spouses were at every moment and demanded explanations for any little deviation from routines. Today, armed with a smartphone and some cheap software, they can easily establish marital dictatorships. They can monitor every conversation and every movement, record phone logs, track social media posts and web page searches, and even activate the cameras and microphones of a spouse’s phone to serve as a spying device. The U.S.-based National Network to End Domestic Violence found that more than half of domestic abusers used such “stalkware” technology. Even in New York a spouse may find themselves monitored and restricted, as if they lived in a totalitarian state.41

    A growing percentage of employees—from office workers to truck drivers—are also now being surveilled by their employers. Bosses can pinpoint where employees are at any moment, how much time they spend in the toilet, whether they read personal emails at work, and how fast they complete each task.42 Corporations are similarly monitoring their customers, wanting to know their likes and dislikes, to predict future behavior, and to evaluate risks and opportunities. For example, vehicles monitor their drivers’ behavior and share the data with the algorithms of the insurance companies, which raise the premiums they charge “bad drivers” and lower the premiums for “good drivers.”43 The American scholar Shoshana Zuboff has termed this ever-expanding commercial monitoring system “surveillance capitalism.”44

    In addition to all these varieties of top-down surveillance, there are peer-to-peer systems in which individuals constantly monitor one another. For example, the Tripadvisor corporation maintains a worldwide surveillance system that monitors hotels, vacation rentals, restaurants, and tourists. In 2019, it was used by 463 million travelers who browsed 859 million reviews and 8.6 billion lodgings, restaurants, and tourist attractions. It is the users themselves—rather than some sophisticated AI algorithm—who determine whether a restaurant is worth visiting. People who ate in the restaurant can score it on a 1 to 5 scale, and also add photos and written reviews. The Tripadvisor algorithm merely aggregates the data, calculates the restaurant’s average score, ranks the restaurant compared with others of its kind, and makes the results available for everybody to see.

    The algorithm simultaneously ranks the guests, too. For posting reviews or travel articles, users receive 100 points; for uploading photos or videos, 30 points; for posting in a forum, 20 points; for rating establishments, 5 points; and for casting votes for others’ reviews, 1 point. Users are then ranked from Level 1 (300 points) to Level 6 (10,000 points) and receive perks accordingly. Users who violate the system’s rules—for example, by submitting racist comments or trying to blackmail a restaurant by writing an unjustified bad review—may be penalized or kicked out of the system altogether. This is peer-to-peer surveillance. Everybody is constantly grading everybody else. Tripadvisor doesn’t need to invest in cameras and spyware or develop hyper-sophisticated biometric algorithms. Almost all the data is submitted and almost all the work is done by millions of human users. The job of the Tripadvisor algorithm is only to aggregate human-generated scores and publish them.45

    Tripadvisor and similar peer-to-peer surveillance systems provide valuable information for millions of people every day, making it easier to plan vacations and find good hotels and restaurants. But in doing so, they have also shifted the border between private and public spaces. Traditionally, the relationship between the customer and a waiter, say, was a relatively private affair. Entering a bistro meant entering a semiprivate space and establishing a semiprivate relationship with the waiter. Unless some crime was committed, what happened between guest and waiter was their business alone. If the waiter was rude or made a racist remark, you could make a scene and perhaps tell your friends not to go there, but few other people would hear about it.

    Peer-to-peer surveillance networks have obliterated that sense of privacy. If the staff fails to please a customer, the restaurant will get a bad review, which could affect the decision of thousands of potential customers in coming years. For better or worse, the balance of power tilts in favor of the customers, while the staff find themselves more exposed than before to the public gaze. As the author and journalist Linda Kinstler put it, “Before Tripadvisor, the customer was only nominally king. After, he became a veritable tyrant, with the power to make or break lives.”46 The same loss of privacy is felt today by millions of taxi drivers, barbers, beauticians, and other service providers. In the past, stepping into a taxi or barbershop meant stepping into someone’s private space. Now, when customers come into your taxi or barbershop, they bring cameras, microphones, a surveillance network, and thousands of potential viewers with them.47 This is the foundation of a nongovernmental peer-to-peer surveillance network.

    THE SOCIAL CREDIT SYSTEM

    Peer-to-peer surveillance systems typically operate by aggregating many points to determine an overall score. Another type of surveillance network takes this “score logic” to its ultimate conclusion. This is the social credit system, which seeks to give people points for everything and produce an overall personal score that will influence everything. The last time humans came up with such an ambitious points system was five thousand years ago in Mesopotamia, when money was invented. One way to think of the social credit system is as a new kind of money.

    Money is points that people accumulate by selling certain products and services, and then use to buy other products and services. Some countries call their “points” dollars, whereas other countries call them euros, yen, or renminbi. The points can take the form of coins, banknotes, or bits in a digital bank account. The points themselves are, of course, intrinsically worthless. You cannot eat coins or wear banknotes. Their value lies in the fact that they serve as accounting tokens that society uses to keep track of our individual scores.

    Money revolutionized economic relations, social interactions, and human psychology. But like surveillance, money has had its limitations and could not reach everywhere. Even in the most capitalist societies, there have always been places that money didn’t penetrate, and there have always been many things that lacked a monetary value. How much is a smile worth? How much money does a person earn for visiting their grandparents?48

    For scoring those things that money can’t buy, there was an alternative nonmonetary system, which has been given different names: honor, status, reputation. What social credit systems seek is a standardized valuation of the reputation market. Social credit is a new points system that ascribes precise values even to smiles and family visits. To appreciate how revolutionary and far-reaching this is, let’s examine in brief how the reputation market has hitherto differed from the money market. This will help us understand what might happen to social relations if the principles of the money market are suddenly extended to the reputation market.

    One major difference between money and reputation is that money has tended to be a mathematical construct based on precise calculations, whereas the sphere of reputation has been resistant to precise numerical evaluation. For example, medieval aristocrats graded themselves in hierarchical ranks such as dukes, counts, and viscounts, but nobody was counting reputation points. Customers in a medieval market usually knew how many coins they had in their purses and the price of every product in the stalls. In the money market, no coin goes uncounted. In contrast, knights in a medieval reputational market didn’t know the exact amount of honor that different actions might accrue, nor could they be sure of their overall score. Would fighting bravely in battle bring a knight 10 honor points, or 100? And what if nobody saw and recorded their bravery? Indeed, even assuming it was noticed, different people might assign it different values. This lack of precision wasn’t a bug in the system but a crucial feature. “Calculating” was a synonym for cunning and scheming. Acting honorably was supposed to reflect an inner virtue, rather than a pursuit of external rewards.49

    This difference between the scrupulous money market and the ill-defined reputation market still prevails. The owner of a bistro always notices and complains if you don’t pay for your meal in full; every item on the menu has a precise price. But how would the owner even know if society failed to register some good deed they performed? Whom could they complain to if they weren’t properly rewarded for helping an elderly customer or for being extra patient with a rude customer? In some cases, they might now try complaining to Tripadvisor, which collapses the boundary between the money market and the reputation market, turning the fuzzy reputation of restaurants and hotels into a mathematical system of precise points. The idea of social credit is to expand this surveillance method from restaurants and hotels to everything. In the most extreme type of social credit systems, every person gets an overall reputation score that takes into account whatever they do and determines everything they can do.

    For example, you might earn 10 points for picking up trash from the street, get another 20 points for helping an old lady cross the road, and lose 15 points for playing the drums and disturbing the neighbors. If you get a high enough score, it might give you priority when buying train tickets or a leg up when applying to university. If you get a low score, potential employers may refuse to give you a job, and potential dates may refuse your advances. Insurance companies may demand higher premiums, and judges may inflict harsher sentences.

    Some people might see social credit systems as a way to reward pro-social behavior, punish egotistical acts, and create kinder and more harmonious societies. The Chinese government, for example, explains that its social credit systems could help fight corruption, scams, tax evasion, false advertising, and counterfeiting, and thereby establish more trust between individuals, between consumers and corporations, and between citizens and government institutions.50 Others may find systems that allocate precise values to every social action demeaning and inhuman. Even worse, a comprehensive social credit system will annihilate privacy and effectively turn life into a never-ending job interview. Anything you do, anytime, anywhere, might affect your chances of getting a job, a bank loan, a husband, or a prison sentence. You got drunk at a college party and did something legal but shameful? You participated in a political demonstration? You’re friends with someone who has a low credit score? This will be part of your job interview—or criminal sentencing—both in the short term and even decades later. The social credit system might thereby become a totalitarian control system.

    Of course, the reputation market always controlled people and made them conform to the prevailing social norms. In most societies people have always feared losing face even more than they have feared losing money. Many more people commit suicide due to shame and guilt than due to economic distress. Even when people kill themselves after being fired from their job or after their business goes bankrupt, they are usually pushed over the edge by the social humiliation it involves rather than by the economic hardship per se.51

    But the uncertainty and the subjectivity of the reputation market have previously limited its potential for totalitarian control. Since nobody knew the precise value of each social interaction, and since nobody could possibly keep tabs on all interactions, there was significant room for maneuver. When you went to a college party, you might have behaved in a way that earned the respect of your friends, without worrying what future employers might think. When you went to a job interview, you knew none of your friends would be there. And when you were watching pornography at home, you assumed that neither your bosses nor your friends knew what you were up to. Life has been divided into separate reputational spheres, with separate status competitions, and there were also many off-grid moments when you didn’t have to engage in any status competitions at all. Precisely because status competition is so crucial, it is also extremely stressful. Therefore, not only humans but even other social animals like apes have always welcomed some respite from it.52

    Unfortunately, social credit algorithms combined with ubiquitous surveillance technology now threaten to merge all status competitions into a single never-ending race. Even in their own homes or while trying to enjoy a relaxed vacation, people would have to be extremely careful about every deed and word, as if they were performing onstage in front of millions. This could create an incredibly stressful lifestyle, destructive to people’s well-being as well as to the functioning of society. If digital bureaucrats use a precise points system to keep tabs on everybody all the time, the emerging reputation market could annihilate privacy and control people far more tightly than the money market ever did.

    ALWAYS ON

    Humans are organic beings who live by cyclical biological time. Sometimes we are awake; sometimes we are asleep. After intense activity, we need rest. We grow and decay. Networks of humans are similarly subject to biological cycles. They are sometimes on and sometimes off. Job interviews don’t last forever. Police agents don’t work twenty-four hours a day. Bureaucrats take holidays. Even the money market respects these biological cycles. The New York Stock Exchange is open on Mondays to Fridays, from 9:30 in the morning to 4:00 in the afternoon, and is closed on holidays like Independence Day and New Year’s Day. If a war erupts at 4:01 p.m. on a Friday, the market won’t react to it until Monday morning.

    In contrast, a network of computers can always be on. Computers are consequently pushing humans toward a new kind of existence in which we are always connected and always monitored. In some contexts, like health care, this could be a boon. In other contexts, like for citizens of totalitarian states, this could be a disaster. Even if the network is potentially benign, the very fact that it is always “on” might be damaging to organic entities like humans, because it will take away our opportunities to disconnect and relax. If an organism never has a chance to rest, it eventually collapses and dies. But how will we get a relentless network to slow down and allow us some breaks?

    We need to prevent the computer network from taking complete control of society not just in order to give us time off. Breaks are even more crucial to give us a chance to rectify the network. If the network continues to evolve at an accelerating pace, errors will accumulate much faster than we can identify and correct them. For while the network is relentless and ubiquitous, it is also fallible. Yes, computers can gather unprecedented amounts of data on us, watching what we do twenty-four hours a day. And yes, they can identify patterns in the ocean of data with superhuman efficiency. But that does not mean that the computer network will always understand the world accurately. Information isn’t truth. A total surveillance system may form a very distorted understanding of the world and of human beings. Instead of discovering the truth about the world and about us, the network might use its immense power to create a new kind of world order and impose it on us.

    CHAPTER 8 Fallible: The Network Is Often Wrong

    In The Gulag Archipelago (1973), Aleksandr Solzhenitsyn chronicles the history of the Soviet labor camps and of the information network that created and sustained them. He was writing partly from bitter personal experience. When Solzhenitsyn served as a captain in the Red Army during World War II, he maintained a private correspondence with a school friend in which he occasionally criticized Stalin. To be on the safe side, he did not mention the dictator by name and spoke only about “the man with the mustache.” It availed him little. His letters were intercepted and read by the secret police, and in February 1945, while serving on the front line in Germany, he was arrested. He spent the next eight years in labor camps.1 Many of Solzhenitsyn’s hard-won insights and stories are still relevant to understanding the development of information networks in the twenty-first century.

    One story recounts events at a district party conference in Moscow Province in the late 1930s, at the height of the Stalinist Great Terror. A call was made to pay tribute to Stalin, and the audience—who of course knew that they were being carefully watched—burst into applause. After five minutes of applause, “palms were getting sore and raised arms were already aching. And the older people were panting from exhaustion.… However, who would dare be the first to stop?” Solzhenitsyn explains that “NKVD men were standing in the hall applauding and watching to see who quit first!” It went on and on, for six minutes, then eight, then ten. “They couldn’t stop now till they collapsed with heart attacks! … With make-believe enthusiasm on their faces, looking at each other with faint hope, the district leaders were just going to go on and on applauding till they fell where they stood.”

    Finally, after eleven minutes, the director of a paper factory took his life in his hands, stopped clapping, and sat down. Everyone else immediately stopped clapping and also sat down. That same night, the secret police arrested him and sent him to the gulag for ten years. “His interrogator reminded him: Don’t ever be the first to stop applauding!”2

    This story reveals a crucial and disturbing fact about information networks, and in particular about surveillance systems. As discussed in previous chapters, contrary to the naive view, information is often used to create order rather than discover truth. On the face of it, Stalin’s agents in the Moscow conference used the “clapping test” as a way to uncover the truth about the audience. It was a loyalty test, which assumed that the longer you clapped, the more you loved Stalin. In many contexts, this assumption is not unreasonable. But in the context of Moscow in the late 1930s, the nature of the applause changed. Since participants in the conference knew they were being watched, and since they knew the consequences of any hint of disloyalty, they clapped out of terror rather than love. The paper factory director might have been the first to stop not because he was the least loyal but perhaps because he was the most honest, or even simply because his hands hurt the most.

    While the clapping test didn’t discover the truth about people, it was efficient in imposing order and forcing people to behave in a certain way. Over time, such methods cultivated servility, hypocrisy, and cynicism. This is what the Soviet information network did to hundreds of millions of people over decades. In quantum mechanics the act of observing subatomic particles changes their behavior; it is the same with the act of observing humans. The more powerful our tools of observation, the greater the potential impact.

    The Soviet regime constructed one of the most formidable information networks in history. It gathered and processed enormous amounts of data on its citizens. It also claimed that the infallible theories of Marx, Engels, Lenin, and Stalin granted it a deep understanding of humanity. In fact, the Soviet information network ignored many important aspects of human nature, and it was in complete denial regarding the terrible suffering its policies inflicted on its own citizens. Instead of producing wisdom, it produced order, and instead of revealing the universal truth about humans, it actually created a new type of human—Homo sovieticus.

    As defined by the dissident Soviet philosopher and satirist Aleksandr Zinovyev, Homo sovieticus were servile and cynical humans, lacking all initiative or independent thinking, passively obeying even the most ludicrous orders, and indifferent to the results of their actions.3 The Soviet information network created Homo sovieticus through surveillance, punishments, and rewards. For example, by sending the director of the paper factory to the gulag, the network signaled to the other participants that conformity paid off, whereas being the first to do anything controversial was a bad idea. Though the network failed to discover the truth about humans, it was so good at creating order that it conquered much of the world.

    THE DICTATORSHIP OF THE LIKE

    An analogous dynamic may afflict the computer networks of the twenty-first century, which might create new types of humans and new dystopias. A paradigmatic example is the role played by social media algorithms in radicalizing people. Of course, the methods employed by the algorithms have been utterly different from those of the NKVD and involved no direct coercion or violence. But just as the Soviet secret police created the slavish Homo sovieticus through surveillance, rewards, and punishments, so also the Facebook and YouTube algorithms have created internet trolls by rewarding certain base instincts while punishing the better angels of our nature.

    As explained briefly in chapter 6, the process of radicalization started when corporations tasked their algorithms with increasing user engagement, not only in Myanmar, but throughout the world. For example, in 2012 users were watching about 100 million hours of videos every day on YouTube. That was not enough for company executives, who set their algorithms an ambitious goal: 1 billion hours a day by 2016.4 Through trial-and-error experiments on millions of people, the YouTube algorithms discovered the same pattern that Facebook algorithms also learned: outrage drives engagement up, while moderation tends not to. Accordingly, the YouTube algorithms began recommending outrageous conspiracy theories to millions of viewers while ignoring more moderate content. By 2016, users were indeed watching 1 billion hours every day on YouTube.5

    YouTubers who were particularly intent on gaining attention noticed that when they posted an outrageous video full of lies, the algorithm rewarded them by recommending the video to numerous users and increasing the YouTubers’ popularity and income. In contrast, when they dialed down the outrage and stuck to the truth, the algorithm tended to ignore them. Within a few months of such reinforcement learning, the algorithm turned many YouTubers into trolls.6

    The social and political consequences were far-reaching. For example, as the journalist Max Fisher documented in his 2022 book, The Chaos Machine, YouTube algorithms became an important engine for the rise of the Brazilian far right and for turning Jair Bolsonaro from a fringe figure into Brazil’s president.7 While there were other factors contributing to that political upheaval, it is notable that many of Bolsonaro’s chief supporters and aides had originally been YouTubers who rose to fame and power by algorithmic grace.

    A typical example is Carlos Jordy, who in 2017 was a city councilor in the small town of Niterói. The ambitious Jordy gained national attention by creating inflammatory YouTube videos that garnered millions of views. His videos warned Brazilians, for example, against conspiracies by schoolteachers to brainwash children and persecute conservative pupils. In 2018, Jordy won a seat in the Brazilian Chamber of Deputies (the lower house of the Brazilian Congress) as one of Bolsonaro’s most dedicated supporters. In an interview with Fisher, Jordy frankly said, “If social media didn’t exist, I wouldn’t be here [and] Jair Bolsonaro wouldn’t be president.” The latter claim may well be a self-serving exaggeration, but there is no denying that social media played an important part in Bolsonaro’s rise.

    Another YouTuber who won a seat in Brazil’s Chamber of Deputies in 2018 was Kim Kataguiri, one of the leaders of the Movimento Brasil Livre (MBL, or Free Brazil Movement). Kataguiri initially used Facebook as his main platform, but his posts were too extreme even for Facebook, which banned some of them for disinformation. So Kataguiri switched over to the more permissive YouTube. In an interview in the MBL headquarters in São Paulo, Kataguiri’s aides and other activists explained to Fisher, “We have something here that we call the dictatorship of the like.” They explained that YouTubers tend to become steadily more extreme, posting untruthful and reckless content “just because something is going to give you views, going to give engagement.… Once you open that door there’s no going back, because you always have to go further.… Flat Earthers, anti-vaxxers, conspiracy theories in politics. It’s the same phenomenon. You see it everywhere.”8

    Of course, the YouTube algorithms were not themselves responsible for inventing lies and conspiracy theories or for creating extremist content. At least in 2017–18, those things were done by humans. The algorithms were responsible, however, for incentivizing humans to behave in such ways and for pushing the resulting content in order to maximize user engagement. Fisher documented numerous far-right activists who first became interested in extremist politics after watching videos that the YouTube algorithm auto-played for them. One far-right activist in Niterói told Fisher that he was never interested in politics of any kind, until one day the YouTube algorithm auto-played for him a video on politics by Kataguiri. “Before that,” he explained, “I didn’t have an ideological, political background.” He credited the algorithm with providing “my political education.” Talking about how other people joined the movement, he said, “It was like that with everyone.… Most of the people here came from YouTube and social media.”9

    BLAME THE HUMANS

    We have reached a turning point in history in which major historical processes are partly caused by the decisions of nonhuman intelligence. It is this that makes the fallibility of the computer network so dangerous. Computer errors become potentially catastrophic only when computers become historical agents. We have already made this argument in chapter 6, when we briefly examined Facebook’s role in instigating the anti-Rohingya ethnic-cleansing campaign. As noted in that context, however, many people—including some of the managers and engineers of Facebook, YouTube, and the other tech giants—object to this argument. Since it is one of the central points of the entire book, it is best to delve deeper into the matter and examine more carefully the objections to it.

    The people who manage Facebook, YouTube, TikTok, and other platforms routinely try to excuse themselves by shifting the blame from their algorithms to “human nature.” They argue that it is human nature that produces all the hate and lies on the platforms. The tech giants then claim that due to their commitment to free speech values, they hesitate to censor the expression of genuine human emotions. For example, in 2019 the CEO of YouTube, Susan Wojcicki, explained, “The way that we think about it is: ‘Is this content violating one of our policies? Has it violated anything in terms of hate, harassment?’ If it has, we remove that content. We keep tightening and tightening the policies. We also get criticism, just to be clear, [about] where do you draw the lines of free speech and, if you draw it too tightly, are you removing voices of society that should be heard? We’re trying to strike a balance of enabling a broad set of voices, but also making sure that those voices play by a set of rules that are healthy conversations for society.”10

    A Facebook spokesperson similarly said in October 2021, “Like every platform, we are constantly making difficult decisions between free expressions and harmful speech, security and other issues.… But drawing these societal lines is always better left to elected leaders.”11 In this way, the tech giants constantly shift the discussion to their supposed role as moderators of human-produced content and ignore the active role their algorithms play in cultivating certain human emotions and discouraging others. Are they really blind to it?

    Surely not. Back in 2016, an internal Facebook report discovered that “64 percent of all extremist group joins are due to our recommendation tools.… Our recommendation systems grow the problem.”12 A secret internal Facebook memo from August 2019, leaked by the whistleblower Frances Haugen, stated, “We have evidence from a variety of sources that hate speech, divisive political speech, and misinformation on Facebook and [its] family of apps are affecting societies around the world. We also have compelling evidence that our core product mechanics, such as virality, recommendations, and optimizing for engagement, are a significant part of why these types of speech flourish on the platform.”13

    Another leaked document from December 2019 noted, “Unlike communication with close friends and family, virality is something new we have introduced to many ecosystems … and it occurs because we intentionally encourage it for business reasons.” The document pointed out that “ranking content about higher stakes topics like health or politics based on engagement leads to perverse incentives and integrity issues.” Perhaps most damningly, it revealed, “Our ranking systems have specific separate predictions for not just what you would engage with, but what we think you may pass along so that others may engage with. Unfortunately, research has shown how outrage and misinformation are more likely to be viral.” This leaked document made one crucial recommendation: since Facebook cannot remove everything harmful from a platform used by many millions, it should at least “stop magnifying harmful content by giving it unnatural distribution.”14

    Like the Soviet leaders in Moscow, the tech companies were not uncovering some truth about humans; they were imposing on us a perverse new order. Humans are very complex beings, and benign social orders seek ways to cultivate our virtues while curtailing our negative tendencies. But social media algorithms see us, simply, as an attention mine. The algorithms reduced the multifaceted range of human emotions—hate, love, outrage, joy, confusion—into a single catchall category: engagement. In Myanmar in 2016, in Brazil in 2018, and in numerous other countries, the algorithms scored videos, posts, and all other content solely according to how many minutes people engaged with the content and how many times they shared it with others. An hour of lies or hatred was ranked higher than ten minutes of truth or compassion—or an hour of sleep. The fact that lies and hate tend to be psychologically and socially destructive, whereas truth, compassion, and sleep are essential for human welfare, was completely lost on the algorithms. Based on this very narrow understanding of humanity, the algorithms helped to create a new social system that encouraged our basest instincts while discouraging us from realizing the full spectrum of the human potential.

    As the harmful effects were becoming manifest, the tech giants were repeatedly warned about what was happening, but they failed to step in because of their faith in the naive view of information. As the platforms were overrun by falsehoods and outrage, executives hoped that if more people were enabled to express themselves more freely, truth would eventually prevail. This, however, did not happen. As we have seen again and again throughout history, in a completely free information fight, truth tends to lose. To tilt the balance in favor of truth, networks must develop and maintain strong self-correcting mechanisms that reward truth telling. These self-correcting mechanisms are costly, but if you want to get the truth, you must invest in them.

    Silicon Valley thought it was exempt from this historical rule. Social media platforms have been singularly lacking in self-correcting mechanisms. In 2014, Facebook employed just a single Burmese-speaking content moderator to monitor activities in the whole of Myanmar.15 When observers in Myanmar began warning Facebook that it needed to invest more in moderating content, Facebook ignored them. For example, Pwint Htun, a Burmese American engineer and telecom executive who grew up in rural Myanmar, wrote to Facebook executives repeatedly about the danger. In an email from July 5, 2014—two years before the ethnic-cleansing campaign began—she issued a prophetic warning: “Tragically, FB in Burma is used like radio in Rwanda during the dark days of genocide.” Facebook took no action.

    Even after the attacks on the Rohingya intensified and Facebook faced a storm of criticism, it still refused to hire people with expert local knowledge to curate content. Thus, when informed that hate-mongers in Myanmar were using the Burmese word kalar as a racist slur for the Rohingya, Facebook reacted in April 2017 by banning from the platform any posts that used the word. This revealed Facebook’s utter lack of knowledge about local conditions and the Burmese language. In Burmese, kalar is a racist slur only in specific contexts. In other contexts, it is an entirely innocent term. The Burmese word for chair is kalar htaing, and the word for chickpea is kalar pae. As Pwint Htun wrote to Facebook in June 2017, banning the term kalar from the platform is like banning the letters “hell” from “hello.”16 Facebook continued to ignore the need for local expertise. By April 2018, the number of Burmese speakers Facebook employed to moderate content for its eighteen million users in Myanmar was a grand total of five.17

    Instead of investing in self-correcting mechanisms that would reward truth telling, the social media giants actually developed unprecedented error-enhancing mechanisms that rewarded lies and fictions. One such error-enhancing mechanism was the Instant Articles program that Facebook rolled out in Myanmar in 2016. Wishing to drive up engagement, Facebook paid news channels according to the amount of user engagement they generated, measured in clicks and views. No importance whatsoever was given to the truthfulness of the “news.” A 2021 study found that in 2015, before the program was launched, six of the ten top Facebook websites in Myanmar belonged to “legitimate media.” By 2017, under the impact of Instant Articles, “legitimate media” was down to just two websites out of the top ten. By 2018, all top ten websites were “fake news and clickbait websites.”

    The study concluded that because of the launch of Instant Articles “clickbait actors cropped up in Myanmar overnight. With the right recipe for producing engaging and evocative content, they could generate thousands of US dollars a month in ad revenue, or ten times the average monthly salary—paid to them directly by Facebook.” Since Facebook was by far the most important source of online news in Myanmar, this had enormous impact on the overall media landscape of the country. “In a country where Facebook is synonymous with the Internet, the low-grade content overwhelmed other information sources.”18 Facebook and other social media platforms didn’t consciously set out to flood the world with fake news and outrage. But by telling their algorithms to maximize user engagement, this is exactly what they perpetrated.

    Reflecting on the Myanmar tragedy, Pwint Htun wrote to me in July 2023, “I naively used to believe that social media could elevate human consciousness and spread the perspective of common humanity through interconnected pre-frontal cortexes in billions of human beings. What I realize is that the social media companies are not incentivized to interconnect pre-frontal cortexes. Social media companies are incentivized to create interconnected limbic systems—which is much more dangerous for humanity.”

    THE ALIGNMENT PROBLEM

    I don’t want to imply that the spread of fake news and conspiracy theories is the main problem with all past, present, and future computer networks. YouTube, Facebook, and other social media platforms claim that since 2018 they have been tweaking their algorithms to make them more socially responsible. Whether this is true or not is hard to say, especially because there is no universally accepted definition of “social responsibility.”19 But the specific problem of polluting the information sphere in pursuit of user engagement can certainly be solved. When the tech giants set their hearts on designing better algorithms, they can usually do it. Around 2005, the profusion of spam threatened to make the use of email impossible. Powerful algorithms were developed to address the problem. By 2015, Google claimed its Gmail algorithm had a 99.9 percent success in blocking genuine spam, while only 1 percent of legitimate emails were erroneously labeled as such.20

    We also shouldn’t discount the huge social benefits that YouTube, Facebook, and other social media platforms have brought. To be clear, most YouTube videos and Facebook posts have not been fake news and genocidal incitements. Social media has been more than helpful in connecting people, giving voice to previously disenfranchised groups, and organizing valuable new movements and communities.21 It has also encouraged an unprecedented wave of human creativity. In the days when television was the dominant medium, viewers were often denigrated as couch potatoes: passive consumers of content that a few gifted artists produced. Facebook, YouTube, and other social media platforms inspired the couch potatoes to get up and start creating. Most of the content on social media—at least until the rise of powerful generative AI—has been produced by the users themselves, and their cats and dogs, rather than by a limited professional class.

    I, too, routinely use YouTube and Facebook to connect with people, and I am grateful to social media for connecting me with my husband, whom I met on one of the first LGBTQ social media platforms back in 2002. Social media has done wonders for dispersed minorities like LGBTQ people. Few gay boys are born to a gay family in a gay neighborhood, and in the days before the internet simply finding one another posed a big challenge, unless you moved to one of the handful of tolerant metropolises that had a gay subculture. Growing up in a small homophobic town in Israel in the 1980s and early 1990s, I didn’t know a single openly gay man. Social media in the late 1990s and early 2000s provided an unprecedented and almost magical way for members of the dispersed LGBTQ community to find one another and connect.

    And yet I have devoted so much attention to the social media “user engagement” debacle because it exemplifies a much bigger problem afflicting computers—the alignment problem. When computers are given a specific goal, such as to increase YouTube traffic to one billion hours a day, they use all their power and ingenuity to achieve this goal. Since they operate very differently than humans, they are likely to use methods their human overlords didn’t anticipate. This can result in dangerous unforeseen consequences, which are not aligned with the original human goals. Even if recommendation algorithms stop encouraging hate, other instances of the alignment problem might result in larger catastrophes than the anti-Rohingya campaign. The more powerful and independent computers become, the bigger the danger.

    Of course, the alignment problem is neither new nor unique to algorithms. It bedeviled humanity for thousands of years before the invention of computers. It has been, for example, the foundational problem of modern military thinking, enshrined in Carl von Clausewitz’s theory of war. Clausewitz was a Prussian general who fought during the Napoleonic Wars. Following Napoleon’s final defeat in 1815, Clausewitz became the director of the Prussian War College. He also began formalizing a grand theory of war. After he died of cholera in 1831, his wife, Marie, edited his unfinished manuscript and published On War in several parts between 1832 and 1834.22

    On War created a rational model for understanding war, and it is still the dominant military theory today. Its most important maxim is that “war is the continuation of policy with other means.”23 This implies that war is not an emotional outbreak, a heroic adventure, or a divine punishment. War is not even a military phenomenon. Rather, war is a political tool. According to Clausewitz, military actions are utterly irrational unless they are aligned with some overarching political goal.

    Suppose Mexico contemplates whether to invade and conquer its small neighbor, Belize. And suppose a detailed military analysis concludes that if the Mexican army invades, it will achieve a quick and decisive military victory, crushing the small Belize army and conquering the capital, Belmopan, in three days. According to Clausewitz, that does not constitute a rational reason for Mexico to invade. The mere ability to secure military victory is meaningless. The key question the Mexican government should ask itself is, what political goals will the military success achieve?

    History is full of decisive military victories that led to political disasters. For Clausewitz, the most obvious example was close to home: Napoleon’s career. Nobody disputes the military genius of Napoleon, who was a master of both tactics and strategy. But while his string of victories brought Napoleon temporary control of vast territories, they failed to secure lasting political achievements. His military conquests merely drove most European powers to unite against him, and his empire collapsed a decade after he crowned himself emperor.

    Indeed, in the long term, Napoleon’s victories ensured the permanent decline of France. For centuries, France was Europe’s leading geopolitical power, largely because both Italy and Germany didn’t exist as unified political entities. Italy was a hodgepodge of dozens of warring city-states, feudal principalities, and church territories. Germany was an even more bizarre jigsaw puzzle divided into more than a thousand independent polities, loosely held together under the theoretical suzerainty of the Holy Roman Empire of the German Nation.24 In 1789, the prospect of a German or Italian invasion of France was simply unthinkable, because there was no such thing as a German or Italian army.

    As Napoleon expanded his empire into central Europe and the Italian Peninsula, he abolished the Holy Roman Empire in 1806, amalgamated many of the smaller German and Italian principalities into larger territorial blocs, created a German Confederation of the Rhine and a Kingdom of Italy, and sought to unify these territories under his dynastic rule. His victorious armies also spread the ideals of modern nationalism and popular sovereignty into the German and Italian lands. Napoleon thought all this would make his empire stronger. In fact, by breaking up traditional structures and giving Germans and Italians a taste of national consolidation, Napoleon inadvertently lay the foundations for the ultimate unification of Germany (1866–71) and of Italy (1848–71). These twin processes of national unification were sealed by the German victory over France in the Franco-Prussian War of 1870–71. Faced with two newly unified and fervently nationalistic powers on its eastern border, France never regained its position of dominance.

    A more recent example of military victory leading to political defeat was provided by the American invasion of Iraq in 2003. The Americans won every major military engagement, but failed to achieve any of their long-term political aims. Their military victory didn’t establish a friendly regime in Iraq, or a favorable geopolitical order in the Middle East. The real winner of the war was Iran. American military victory turned Iraq from Iran’s traditional foe into Iran’s vassal, thereby greatly weakening the American position in the Middle East while making Iran the regional hegemon.25

    Both Napoleon and George W. Bush fell victim to the alignment problem. Their short-term military goals were misaligned with their countries’ long-term geopolitical goals. We can understand the whole of Clausewitz’s On War as a warning that “maximizing victory” is as shortsighted a goal as “maximizing user engagement.” According to the Clausewitzian model, only once the political goal is clear can armies decide on a military strategy that will hopefully achieve it. From the overall strategy, lower-ranking officers can then derive tactical goals. The model constructs a clear hierarchy between long-term policy, medium-term strategy, and short-term tactics. Tactics are considered rational only if they are aligned with some strategic goal, and strategy is considered rational only if it is aligned with some political goal. Even local tactical decisions of a lowly company commander must serve the war’s ultimate political goal.

    Suppose that during the American occupation of Iraq an American company comes under intense fire from a nearby mosque. The company commander has several different tactical decisions to choose from. He might order the company to retreat. He might order the company to storm the mosque. He might order one of his supporting tanks to blow up the mosque. What should the company commander do?

    From a purely military perspective, it might seem best for the commander to order his tank to blow up the mosque. This would capitalize on the tactical advantage that the Americans enjoyed in terms of firepower, avoid risking the lives of his own soldiers, and achieve a decisive tactical victory. However, from a political perspective, this might be the worst decision the commander could make. Footage of an American tank destroying a mosque would galvanize Iraqi public opinion against the Americans and create outrage throughout the wider Muslim world. Storming the mosque might also be a political mistake, because it too could create resentment among Iraqis, while the cost in American lives could weaken support for the war among American voters. Given the political war aims of the United States, retreating and conceding tactical defeat might well be the most rational decision.

    For Clausewitz, then, rationality means alignment. Pursuing tactical or strategic victories that are misaligned with political goals is irrational. The problem is that the bureaucratic nature of armies makes them highly susceptible to such irrationality. As discussed in chapter 3, by dividing reality into separate drawers, bureaucracy encourages the pursuit of narrow goals even when this harms the greater good. Bureaucrats tasked with accomplishing a narrow mission may be ignorant of the wider impact of their actions, and it has always been tricky to ensure that their actions remain aligned with the greater good of society. When armies operate along bureaucratic lines—as all modern armies do—it creates a huge gap between a captain commanding a company in the field and the president formulating long-term policy in a distant office. The captain is prone to make decisions that seem reasonable on the ground but that actually undermine the war’s ultimate goal.

    We see, then, that the alignment problem has long predated the computer revolution and that the difficulties encountered by builders of present-day information empires are not unlike those that bedeviled previous would-be conquerors. Nevertheless, computers do change the nature of the alignment problem in important ways. No matter how difficult it used to be to ensure that human bureaucrats and soldiers remain aligned with society’s long-term goals, it is going to be even harder to ensure the alignment of algorithmic bureaucrats and autonomous weapon systems.

    THE PAPER-CLIP NAPOLEON

    One reason why the alignment problem is particularly dangerous in the context of the computer network is that this network is likely to become far more powerful than any previous human bureaucracy. A misalignment in the goals of superintelligent computers might result in a catastrophe of unprecedented magnitude. In his 2014 book, Superintelligence, the philosopher Nick Bostrom illustrated the danger using a thought experiment, which is reminiscent of Goethe’s “Sorcerer’s Apprentice.” Bostrom asks us to imagine that a paper-clip factory buys a superintelligent computer and that the factory’s human manager gives the computer a seemingly simple task: produce as many paper clips as possible. In pursuit of this goal, the paper-clip computer conquers the whole of planet Earth, kills all the humans, sends expeditions to take over additional planets, and uses the enormous resources it acquires to fill the entire galaxy with paper-clip factories.

    The point of the thought experiment is that the computer did exactly what it was told (just like the enchanted broomstick in Goethe’s poem). Realizing that it needed electricity, steel, land, and other resources to build more factories and produce more paper clips, and realizing that humans are unlikely to give up these resources, the superintelligent computer eliminated all humans in its single-minded pursuit of its given goal.26 Bostrom’s point was that the problem with computers isn’t that they are particularly evil but that they are particularly powerful. And the more powerful the computer, the more careful we need to be about defining its goal in a way that precisely aligns with our ultimate goals. If we define a misaligned goal to a pocket calculator, the consequences are trivial. But if we define a misaligned goal to a superintelligent machine, the consequences could be dystopian.

    The paper-clip thought experiment may sound outlandish and utterly disconnected from reality. But if Silicon Valley managers had paid attention when Bostrom published it in 2014, perhaps they would have been more careful before instructing their algorithms to “maximize user engagement.” The Facebook and YouTube algorithms behaved exactly like Bostrom’s imaginary algorithm. When told to maximize paper-clip production, the algorithm sought to convert the entire physical universe into paper clips, even if it meant destroying human civilization. When told to maximize user engagement, the Facebook and YouTube algorithms sought to convert the entire social universe into user engagement, even if it meant doing harm to the social fabric of Myanmar, Brazil, and many other countries.

    Bostrom’s thought experiment highlights a second reason why the alignment problem is more urgent in the case of computers. Because they are inorganic entities, they are likely to adopt strategies that would never occur to any human and that we are therefore ill-equipped to foresee and forestall. Here’s one example: In 2016, Dario Amodei was working on a project called Universe, trying to develop a general-purpose AI that could play hundreds of different computer games. The AI competed well in various car races, so Amodei next tried it on a boat race. Inexplicably, the AI steered its boat right into a harbor and then sailed in endless circles in and out of the harbor.

    It took Amodei considerable time to understand what went wrong. The problem occurred because initially Amodei wasn’t sure how to tell the AI that its goal was to “win the race.” “Winning” is an unclear concept to an algorithm. Translating “win the race” into computer language would have required Amodei to formalize complex concepts like track position and placement among the other boats in the race. So instead, Amodei took the easy way and told the boat to maximize its score. He assumed that the score was a good proxy for winning the race. After all, it worked with the car races.

    But the boat race had a peculiar feature, absent from the car races, that allowed the ingenious AI to find a loophole in the game’s rules. The game rewarded players with a lot of points for getting ahead of other boats—as in the car races—but it also rewarded them with a few points whenever they replenished their power by docking into a harbor. The AI discovered that if instead of trying to outsail the other boats, it simply went in circles in and out of the harbor, it could accumulate more points far faster. Apparently, none of the game’s human developers—nor Dario Amodei—noticed this loophole. The AI was doing exactly what the game was rewarding it to do—even though it is not what the humans were hoping for. That’s the essence of the alignment problem: rewarding A while hoping for B.27 If we want computers to maximize social benefits, it’s a bad idea to reward them for maximizing user engagement.

    A third reason to worry about the alignment problem of computers is that because they are so different from us, when we make the mistake of giving them a misaligned goal, they are less likely to notice it or request clarification. If the boat-race AI had been a human gamer, it would have realized that the loophole it found in the game’s rules probably doesn’t really count as “winning.” If the paper-clip AI had been a human bureaucrat, it would have realized that destroying humanity in order to produce paper clips is probably not what was intended. But since computers aren’t humans, we cannot rely on them to notice and flag possible misalignments. In the 2010s the YouTube and Facebook management teams were bombarded with warnings from their human employees—as well as from outside observers—about the harm being done by the algorithms, but the algorithms themselves never raised the alarm.28

    As we give algorithms greater and greater power over health care, education, law enforcement, and numerous other fields, the alignment problem will loom ever larger. If we don’t find ways to solve it, the consequences will be far worse than algorithms racking up points by sailing boats in circles.

    THE CORSICAN CONNECTION

    How to solve the alignment problem? In theory, when humans create a computer network, they must define for it an ultimate goal, which the computers are never allowed to change or ignore. Then, even if computers become so powerful that we lose control over them, we can rest assured that their immense power will benefit rather than harm us. Unless, of course, it turned out that we defined a harmful or vague goal. And there’s the rub. In the case of human networks, we rely on self-correcting mechanisms to periodically review and revise our goals, so setting the wrong goal is not the end of the world. But since the computer network might escape our control, if we set it the wrong goal, we might discover our mistake when we are no longer able to correct it. Some might hope that through a careful process of deliberation, we might be able to define in advance the right goals for the computer network. This, however, is a very dangerous delusion.

    To understand why it is impossible to agree in advance on the ultimate goals of the computer network, let’s revisit Clausewitz’s war theory. There is one fatal flaw in the way he equates rationality with alignment. While Clausewitzian theory demands that all actions be aligned with the ultimate goal, it offers no rational way to define such a goal. Consider Napoleon’s life and military career. What should have been his ultimate goal? Given the prevailing cultural atmosphere of France circa 1800, we can think of several alternatives for “ultimate goal” that might have occurred to Napoleon:

    POTENTIAL GOAL NUMBER 1: Making France the dominant power in Europe, secure against any future attack by Britain, the Habsburg Empire, Russia, a unified Germany, or a unified Italy.

    POTENTIAL GOAL NUMBER 2: Creating a new multiethnic empire ruled by Napoleon’s family, which would include not only France but also many additional territories both in Europe and overseas.

    POTENTIAL GOAL NUMBER 3: Achieving everlasting glory for himself personally, so that even centuries after his death billions of people will know the name Napoleon and admire his genius.

    POTENTIAL GOAL NUMBER 4: Securing the redemption of his everlasting soul, and gaining entry to heaven after his death.

    POTENTIAL GOAL NUMBER 5: Spreading the universal ideals of the French Revolution, and helping to protect freedom, equality, and human rights throughout Europe and the world.

    Many self-styled rationalists tend to argue that Napoleon should have made it his life’s mission to achieve the first goal—securing French domination in Europe. But why? Remember that for Clausewitz rationality means alignment. A tactical maneuver is rational if, and only if, it is aligned with some higher strategic goal, which should in turn be aligned with an even higher political goal. But where does this chain of goals ultimately start? How can we determine the ultimate goal that justifies all the strategic subgoals and tactical steps derived from it? Such an ultimate goal by definition cannot be aligned with anything higher than itself, because there is nothing higher. What then makes it rational to place France at the top of the goal hierarchy, rather than Napoleon’s family, Napoleon’s fame, Napoleon’s soul, or universal human rights? Clausewitz provides no answer.

    One might argue that goal number 4—securing the redemption of his everlasting soul—cannot be a serious candidate for an ultimate rational goal, because it is based on a belief in mythology. But the same argument can be leveled at all the other goals. Everlasting souls are an intersubjective invention that exist only in people’s minds, and exactly the same is true of nations and human rights. Why should Napoleon care about the mythical France any more than about his mythical soul?

    Indeed, for most of his youth, Napoleon didn’t even consider himself French. He was born Napoleone di Buonaparte on Corsica, to a family of Italian emigrants. For five hundred years Corsica was ruled by the Italian city-state of Genoa, where many of Napoleone’s ancestors lived. It was only in 1768—a year before Napoleone’s birth—that Genoa ceded the island to France. Corsican nationalists resisted being handed over to France and rose in rebellion. Only after their defeat in 1770 did Corsica formally become a French province. Many Corsicans continued to resent the French takeover, but the di Buonaparte family swore allegiance to the French king and sent Napoleone to military school in mainland France.29

    At school, Napoleone had to endure a good deal of hazing from his classmates for his Corsican nationalism and his poor command of the French language.30 His mother tongues were Corsican and Italian, and although he gradually became fluent in French, he retained throughout his life a Corsican accent and an inability to spell French correctly.31 Napoleone eventually enlisted in the French army, but when the Revolution broke out in 1789, he went back to Corsica, hoping the revolution would provide an opportunity for his beloved island to achieve greater autonomy. Only after he fell out with the leader of the Corsican independence movement—Pasquale Paoli—did Napoleone abandon the Corsican cause in May 1793. He returned to the mainland, where he decided to build his future.32 It was at this stage that Napoleone di Buonaparte turned into Napoléon Bonaparte (he continued to use the Italian version of his name until 1796).33

    Why then was it rational for Napoleon to devote his military career to making France the dominant power in Europe? Was it perhaps more rational for him to stay in Corsica, patch up his personal disagreements with Paoli, and devote himself to liberating his native island from its French conquerors? And maybe Napoleon should in fact have made it his life’s mission to unite Italy—the land of his ancestors?

    Clausewitz offers no method to answer these questions rationally. If our only rule of thumb is “every action must be aligned with some higher goal,” by definition there is no rational way to define that ultimate goal. How then can we provide a computer network with an ultimate goal it must never ignore or subvert? Tech executives and engineers who rush to develop AI are making a huge mistake if they think there is a rational way to tell that AI what its ultimate goal should be. They should learn from the bitter experiences of generations of philosophers who tried to define ultimate goals and failed.

    THE KANTIAN NAZI

    For millennia, philosophers have been looking for a definition of an ultimate goal that will not depend on an alignment to some higher goal. They have repeatedly been drawn to two potential solutions, known in philosophical jargon as deontology and utilitarianism. Deontologists (from the Greek word deon, meaning “duty”) believe that there are some universal moral duties, or moral rules, that apply to everyone. These rules do not rely on alignment to a higher goal, but rather on their intrinsic goodness. If such rules indeed exist, and if we can find a way to program them into computers, then we can make sure the computer network will be a force for good.

    But what exactly does “intrinsic goodness” mean? The most famous attempt to define an intrinsically good rule was made by Immanuel Kant, a contemporary of Clausewitz and Napoleon. Kant argued that an intrinsically good rule is any rule that I would like to make universal. According to this view, a person about to murder someone should stop and go through the following thought process: “I am now going to murder a human. Would I like to establish a universal rule saying that it is okay to murder humans? If such a universal rule is established, then someone might murder me. So there shouldn’t be a universal rule allowing murder. It follows that I too shouldn’t murder.” In simpler language, Kant reformulated the old Golden Rule: “Do unto others what you want them to do to you” (Matthew 7:12).

    This sounds like a simple and obvious idea: each of us should behave in a way we want everyone to behave. But ideas that sound good in the ethereal realm of philosophy often have trouble immigrating to the harsh land of history. The key question historians would ask Kant is, when you talk about universal rules, how exactly do you define “universal”? Under actual historical circumstances, when a person is about to commit murder, the first step they often take is to exclude the victim from the universal community of humanity.34 This, for example, is what anti-Rohingya extremists like Wirathu did. As a Buddhist monk, Wirathu was certainly against murdering humans. But he didn’t think this universal rule applied to killing Rohingya, who were seen as subhuman. In posts and interviews, he repeatedly compared them to beasts, snakes, mad dogs, wolves, jackals, and other dangerous animals.35 On October 30, 2017, at the height of the anti-Rohingya violence, another, more senior Buddhist monk preached a sermon to military officers in which he justified violence against the Rohingya by telling the officers that non-Buddhists were “not fully human.”36

    As a thought experiment, imagine a meeting between Immanuel Kant and Adolf Eichmann—who, by the way, considered himself a Kantian.37 As Eichmann signs an order sending another trainload of Jews to Auschwitz, Kant tells him, “You are about to murder thousands of humans. Would you like to establish a universal rule saying it is okay to murder humans? If you do that, you and your family might also be murdered.” Eichmann replies, “No, I am not about to murder thousands of humans. I am about to murder thousands of Jews. If you ask me whether I would like to establish a universal rule saying it is okay to murder Jews, then I am all for it. As for myself and my family, there is no risk that this universal rule would lead to us being murdered. We aren’t Jews.”

    One potential Kantian reply to Eichmann is that when we define entities, we must always use the most universal definition applicable. If an entity can be defined as either “a Jew” or “a human,” we should use the more universal term “human.” However, the whole point of Nazi ideology was to deny the humanity of Jews. In addition, note that Jews are not just humans. They are also animals, and they are also organisms. Since animals and organisms are obviously more universal categories than “human,” if you follow the Kantian argument to its logical conclusion, it might push us to adopt an extreme vegan position. Since we are organisms, does it mean we should object to the killing of any organism, down even to tomatoes or amoebas?

    In history, many if not most conflicts concern the definition of identities. Everybody accepts that murder is wrong, but thinks that only killing members of the in-group qualifies as “murder,” whereas killing someone from an out-group is not. But the in-groups and out-groups are intersubjective entities, whose definition usually depends on some mythology. Deontologists who pursue universal rational rules often end up the captives of local myths.

    This problem with deontology is especially critical if we try to dictate universal deontologist rules not to humans but to computers. Computers aren’t even organic. So if they follow a rule of “Do unto others what you want them to do to you,” why should they be concerned about killing organisms like humans? A Kantian computer that doesn’t want to be killed has no reason to object to a universal rule saying “it is okay to kill organisms”; such a rule does not endanger the nonorganic computer.

    Alternatively, being inorganic entities, computers may have no qualms about dying. As far as we can tell, death is an organic phenomenon and may be inapplicable to inorganic entities. When ancient Assyrians talked about “killing” documents, that was just a metaphor. If computers are more like documents than like organisms, and don’t care about “being killed,” would we like a Kantian computer to conclude that killing humans is therefore fine?

    Is there a way to define whom computers should care about, without getting bogged down by some intersubjective myth? The most obvious suggestion is to tell computers that they must care about any entity capable of suffering. While suffering is often caused by belief in local intersubjective myths, suffering itself is nonetheless a universal reality. Therefore, using the capacity to suffer in order to define the critical in-group grounds morality in an objective and universal reality. A self-driving car should avoid killing all humans—whether Buddhist or Muslim, French or Italian—and should also avoid killing dogs and cats, and any sentient robots that might one day exist. We may even refine this rule, instructing the car to care about different beings in direct proportion to their capacity to suffer. If the car has to choose between killing a human and killing a cat, it should drive over the cat, because presumably the cat has a lesser capacity to suffer. But if we go in that direction, we inadvertently desert the deontologist camp and find ourselves in the camp of their rivals—the utilitarians.

    THE CALCULUS OF SUFFERING

    Whereas deontologists struggle to find universal rules that are intrinsically good, utilitarians judge actions by their impact on suffering and happiness. The English philosopher Jeremy Bentham—another contemporary of Napoleon, Clausewitz, and Kant—said that the only rational ultimate goal is to minimize suffering in the world and maximize happiness. If our main fear about computer networks is that their misaligned goals might inflict terrible suffering on humans and perhaps on other sentient beings, then the utilitarian solution seems both obvious and attractive. When creating the computer network, we just need to instruct it to minimize suffering and maximize happiness. If Facebook had told its algorithms “maximize happiness” instead of “maximize user engagement,” all would allegedly have been well. It is worth noting that this utilitarian approach is indeed popular in Silicon Valley, championed in particular by the effective altruism movement.38

    Unfortunately, as with the deontologist solution, what sounds simple in the theoretical realm of philosophy becomes fiendishly complex in the practical land of history. The problem for utilitarians is that we don’t possess a calculus of suffering. We don’t know how many “suffering points” or “happiness points” to assign to particular events, so in complex historical situations it is extremely difficult to calculate whether a given action increases or decreases the overall amount of suffering in the world.

    Utilitarianism is at its best in situations when the scales of suffering are very clearly tipped in one direction. When confronted by Eichmann, utilitarians don’t need to get into any complicated debates about identity. They just need to point out that the Holocaust caused immense suffering to the Jews, without providing equivalent benefits to anyone else, including the Germans. There was no compelling military or economic need for the Germans to murder millions of Jews. The utilitarian case against the Holocaust is overwhelming.

    Utilitarians also have a field day when dealing with “victimless crimes” like homosexuality, in which all the suffering is on one side only. For centuries, the persecution of gay people caused them immense suffering, but it was nevertheless justified by various prejudices that were erroneously presented as deontological universal rules. Kant, for example, condemned homosexuality on the grounds that it is “contrary to natural instinct and to animal nature” and that it therefore degrades a person “below the level of the animals.” Kant further fulminated that because such acts are contrary to nature, they “make man unworthy of his humanity. He no longer deserves to be a person.”39 Kant, in fact, repackaged a Christian prejudice as a supposedly universal deontological rule, without providing empirical proof that homosexuality is indeed contrary to nature. In light of the above discussion of dehumanization as a prelude to massacre, it is also noteworthy how Kant dehumanized gay people. The view that homosexuality is contrary to nature and deprives people of their humanity paved the way for Nazis like Eichmann to justify murdering homosexuals in concentration camps. Since homosexuals were allegedly below the level of animals, the Kantian rule against murdering humans didn’t apply to them.40

    Utilitarians find it easy to dismiss Kant’s sexual theories, and Bentham indeed was one of the first modern European thinkers who favored the decriminalization of homosexuality.41 Utilitarians argue that criminalizing homosexuality in the name of some dubious universal rule causes tremendous suffering to millions of people, without offering any substantial benefits to others. When two men form a loving relationship, this makes them happy, without making anyone else miserable. Why then forbid it? This type of utilitarian logic also led to many other modern reforms, such as the ban on torture and the introduction of some legal protections for animals.

    But in historical situations when the scales of suffering are more evenly matched, utilitarianism falters. In the early days of the COVID-19 pandemic, governments all over the world adopted strict policies of social isolation and lockdown. This probably saved the lives of several million people.42 It also made hundreds of millions miserable for months. Moreover, it might have indirectly caused numerous deaths, for example by increasing the incidence of murderous domestic violence,43 or by making it more difficult for people to diagnose and treat other dangerous illnesses, like cancer.44 Can anyone calculate the total impact of the lockdown policies and determine whether they increased or decreased the suffering in the world?

    This may sound like a perfect task for a relentless computer network. But how would the computer network decide how many “misery points” to allocate to being locked down with three kids in a two-bedroom apartment for a month? Is that 60 misery points or 600? And how many points to allot to a cancer patient who died because she missed her chemotherapy treatments? Is that 60,000 misery points or 600,000? And what if she would have died of cancer anyway, and the chemo would merely have extended her life by five agonizing months? Should the computers value five months of living with extreme pain as a net gain or a net loss for the sum total of suffering in the world?

    And how would the computer network evaluate the suffering caused by less tangible things, such as the knowledge of our own mortality? If a religious myth promises us that we will never really die, because after death our eternal soul will go to heaven, does that make us truly happy or just delusional? Is death the deep cause of our misery, or does our misery stem from our attempts to deny death? If someone loses their religious faith and comes to terms with their mortality, should the computer network see this as a net loss or a net gain?

    What about even more complicated historical events like the American invasion of Iraq? The Americans were well aware that their invasion would cause tremendous suffering for millions of people. But in the long run, they argued, the benefits of bringing freedom and democracy to Iraq would outweigh the costs. Can the computer network calculate whether this argument was sound? Even if it was theoretically plausible, in practice the Americans failed to establish a stable democracy in Iraq. Does that mean that their attempt was wrong in the first place?

    Just as deontologists trying to answer the question of identity are pushed to adopt utilitarian ideas, so utilitarians stymied by the lack of a suffering calculus often end up adopting a deontologist position. They uphold general rules like “avoid wars of aggression” or “protect human rights,” even though they cannot show that following these rules always reduces the sum total of suffering in the world. History provides them only with a vague impression that following these rules tends to reduce suffering. And when some of these general rules clash—for example, when contemplating launching a war of aggression in order to protect human rights—utilitarianism doesn’t offer much practical help. Not even the most powerful computer network can perform the necessary calculations.

    Accordingly, while utilitarianism promises a rational—and even mathematical—way to align every action with “the ultimate good,” in practice it may well produce just another mythology. Communist true believers confronted by the horrors of Stalinism often replied that the happiness that future generations would experience under “real socialism” would redeem any short-term misery in the gulags. Libertarians, when asked about the immediate social harms of unrestricted free speech or the total abolition of taxes, express a similar faith that future benefits will outweigh any short-term damage. The danger of utilitarianism is that if you have a strong enough belief in a future utopia, it can become an open license to inflict terrible suffering in the present. Indeed, this is a trick traditional religions discovered thousands of years ago. The crimes of this world could too easily be excused by the promises of future salvation.

    COMPUTER MYTHOLOGY

    How then did bureaucratic systems throughout history set their ultimate goals? They relied on mythology to do it for them. No matter how rational were the officials, engineers, tax collectors, and accountants, they were ultimately in the service of this or that mythmaker. To paraphrase John Maynard Keynes, practical people, who believe themselves to be quite exempt from any religious influence, are usually the slaves of some mythmaker. Even nuclear physicists have found themselves obeying the commands of Shiite ayatollahs and communist apparatchiks.

    The alignment problem turns out to be, at heart, a problem of mythology. Nazi administrators could have been committed deontologists or utilitarians, but they would still have murdered millions so long as they understood the world in terms of a racist mythology. If you start with the mythological belief that Jews are demonic monsters bent on destroying humanity, then both deontologists and utilitarians can find many logical arguments why the Jews should be killed.

    An analogous problem might well afflict computers. Of course, they cannot “believe” in any mythology, because they are nonconscious entities that don’t believe in anything. As long as they lack subjectivity, how can they hold intersubjective beliefs? However, one of the most important things to realize about computers is that when a lot of computers communicate with one another, they can create inter-computer realities, analogous to the intersubjective realities produced by networks of humans. These inter-computer realities may eventually become as powerful—and as dangerous—as human-made intersubjective myths.

    This is a very complicated argument, but it is another of the central arguments of the book, so let’s go over it carefully. First, let’s try to understand what inter-computer realities are. As an initial example, consider a one-player computer game. In such a game, you can wander inside a virtual landscape that exists as information within one computer. If you see a rock, that rock is not made of atoms. It is made of bits inside a single computer. When several computers are linked to one another, they can create inter-computer realities. Several players using different computers can wander together inside a common virtual landscape. If they see a rock, that rock is made of bits in several computers.45

    Just as intersubjective realities like money and gods can influence the physical reality outside people’s minds, so inter-computer realities can influence reality outside the computers. In 2016 the game Pokémon Go took the world by storm and was downloaded hundreds of millions of times by the end of the year.46 Pokémon Go is an augmented reality mobile game. Players can use their smartphones to locate, fight, and capture virtual creatures called Pokémon, which seem to exist in the physical world. I once went with my nephew Matan on such a Pokémon hunt. Walking around his neighborhood, I saw only houses, trees, rocks, cars, people, cats, dogs, and pigeons. I didn’t see any Pokémon, because I didn’t have a smartphone. But Matan, looking around through his smartphone lens, could “see” Pokémon standing on a rock or hiding behind a tree.

    Though I couldn’t see the creatures, they were obviously not confined to Matan’s smartphone, because other people could “see” them too. For example, we encountered two other kids who were hunting the same Pokémon. If Matan managed to capture a Pokémon, the other kids could immediately observe what happened. The Pokémon were inter-computer entities. They existed as bits in a computer network rather than as atoms in the physical world, but they could nevertheless interact with the physical world and influence it, as it were, in various ways.

    Now let’s examine a more consequential example of inter-computer realities. Consider the rank that a website gets in a Google search. When we google for news, flight tickets, or restaurant recommendations, one website appears at the top of the first Google page, whereas another is relegated to the middle of the fiftieth page. What exactly is this Google rank, and how is it determined? The Google algorithm determines the website’s Google rank by assigning points to various parameters, such as how many people visit the website and how many other websites link to it. The rank itself is an inter-computer reality, existing in a network connecting billions of computers—the internet. Like Pokémon, this inter-computer reality spills over into the physical world. For a news outlet, a travel agency, or a restaurant it matters a great deal whether its website appears at the top of the first Google page or in the middle of the fiftieth page.47

    Since the Google rank is so important, people use all kinds of tricks to manipulate the Google algorithm to give their website a higher rank. For example, they may use bots to generate more traffic to the website.48 This is a widespread phenomenon in social media too, where coordinated bot armies are constantly manipulating the algorithms of YouTube, Facebook, or Twitter. If a tweet goes viral, is it because humans are really interested in it, or because thousands of bots managed to fool the Twitter algorithm?49

    Inter-computer realities like Pokémon and Google ranks are analogous to intersubjective realities like the sanctity that humans ascribe to temples and cities. I lived much of my life in one of the holiest places on earth—the city of Jerusalem. Objectively, it is an ordinary place. As you walk around Jerusalem, you see houses, trees, rocks, cars, people, cats, dogs, and pigeons, as in any other city. But many people nevertheless imagine it to be an extraordinary place, full of gods, angels, and holy stones. They believe in this so strongly that they sometimes fight over possession of the city or of specific holy buildings and sacred stones, most notably the Holy Rock, located under the Dome of the Rock on Temple Mount. The Palestinian philosopher Sari Nusseibeh observed that “Jews and Muslims, acting on religious beliefs and backed up by nuclear capabilities, are poised to engage in history’s worst-ever massacre of human beings, over a rock.”50 They don’t fight over the atoms that compose the rock; they fight over its “sanctity,” a bit like kids fighting over a Pokémon. The sanctity of the Holy Rock, and of Jerusalem generally, is an intersubjective phenomenon that exists in the communication network connecting many human minds. For thousands of years wars were fought over intersubjective entities like holy rocks. In the twenty-first century, we might see wars fought over inter-computer entities.

    If this sounds like science fiction, consider potential developments in the financial system. As computers become more intelligent and more creative, they are likely to create new inter-computer financial devices. Gold coins and dollars are intersubjective entities. Cryptocurrencies like bitcoin are midway between intersubective and inter-computer. The idea behind them was invented by humans, and their value still depends on human beliefs, but they cannot exist outside the computer network. In addition, they are increasingly traded by algorithms so that their value depends on the calculations of algorithms and not just on human beliefs.

    What if in ten or fifty years computers create a new kind of cryptocurrency or some other financial device that becomes a vital tool for trading and investing—and a potential source for political crises and conflicts? Recall that the 2007–8 global financial crisis was instigated by collateralized debt obligations. These financial devices were invented by a handful of mathematicians and investment whiz kids and were almost unintelligible for most humans, including regulators. This led to an oversight failure and to a global catastrophe.51 Computers may well create financial devices that will be orders of magnitude more complex than CDOs and that will be intelligible only to other computers. The result could be a financial and political crisis even worse than that of 2007–8.

    Throughout history, economics and politics required that we understand the intersubjective realities invented by people—like religions, nations, and currencies. Someone who wanted to understand American politics had to take into account intersubjective realities like Christianity and CDOs. Increasingly, however, understanding American politics will necessitate understanding inter-computer realities ranging from AI-generated cults and currencies to AI-run political parties and even fully incorporated AIs. The U.S. legal system already recognizes corporations as legal persons that possess rights such as freedom of speech. In Citizens United v. Federal Election Commission (2010) the U.S. Supreme Court decided that this even protected the right of corporations to make political donations.52 What would stop AIs from being incorporated and recognized as legal persons with freedom of speech, then lobbying and making political donations to protect and expand AI rights?

    For tens of thousands of years, humans dominated planet Earth because we were the only ones capable of creating and sustaining intersubjective entities like corporations, currencies, gods, and nations, and using such entities to organize large-scale cooperation. Now computers may acquire comparable abilities.

    This isn’t necessarily bad news. If computers lacked connectivity and creativity, they would not be very useful. We increasingly rely on computers to manage our money, drive our vehicles, reduce pollution, and discover new medicines, precisely because computers can directly communicate with one another, spot patterns where we can’t, and construct models that might never occur to us. The problem we face is not how to deprive computers of all creative agency, but rather how to steer their creativity in the right direction. It is the same problem we have always had with human creativity. The intersubjective entities invented by humans were the basis for all the achievements of human civilization, but they occasionally led to crusades, jihads, and witch hunts. The inter-computer entities will probably be the basis for future civilizations, but the fact that computers collect empirical data and use mathematics to analyze it doesn’t mean they cannot launch their own witch hunts.

    THE NEW WITCHES

    In early modern Europe, an elaborate information network analyzed a huge amount of data about crimes, illnesses, and disasters and reached the conclusion that it was all the fault of witches. The more data the witch-hunters gathered, the more convinced they became that the world was full of demons and sorcery and that there was a global satanic conspiracy to destroy humanity. The information network then went on to identify the witches and imprison or kill them. We now know that witches were a bogus intersubjective category, invented by the information network itself and then imposed on people who had never actually met Satan and couldn’t summon hailstorms.

    In the Soviet Union, an even more elaborate information network invented the kulaks—another mythic category that was imposed on millions. The mountains of information collected by Soviet bureaucracy about the kulaks weren’t an objective truth, but they created a new intersubjective truth. Knowing that someone was a kulak became one of the most important things to know about a Soviet person, even though the category was fictitious.

    On an even larger scale, from the sixteenth to the twentieth century, numerous colonial bureaucracies in the Americas, from Brazil through Mexico and the Caribbean to the United States, created a racist mythology and came up with all kinds of intersubjective racial categories. Humans were divided into Europeans, Africans, and Native Americans, and since interracial sexual relations were common, additional categories were invented. In many Spanish colonies the laws differentiated between mestizos, people with mixed Spanish and Native American ancestry; mulatos, people with mixed Spanish and African ancestry; zambos, people with mixed African and Native American ancestry; and pardos, people with mixed Spanish, African, and Native American ancestry. All these seemingly empirical categories determined whether people could be enslaved, enjoy political rights, bear arms, hold public offices, be admitted to school, practice certain professions, live in particular neighborhoods, and be allowed to have sex and get married to each other. Allegedly, by placing a person in a particular racial drawer, one could define their personality, intellectual abilities, and ethical inclinations.53

    By the nineteenth century racism pretended to be an exact science: it claimed to differentiate between people on the basis of objective biological facts, and to rely on scientific tools such as measuring skulls and recording crime statistics. But the cloud of numbers and categories was just a smoke screen for absurd intersubjective myths. The fact that somebody had a Native American grandmother or an African father didn’t, of course, reveal anything about their intelligence, kindness, or honesty. These bogus categories didn’t discover or describe any truth about humans; they imposed an oppressive, mythological order on them.

    As computers replace humans in more and more bureaucracies, from tax collection and health care to security and justice, they too may create a mythology and impose it on us with unprecedented efficiency. In a world ruled by paper documents, bureaucrats had difficulty policing racial borderlines or tracking everyone’s exact ancestry. People could get false documents. A zambo could move to another town and pretend to be a pardo. A Black person could sometimes pass as white. Similarly in the Soviet Union, kulak children occasionally managed to falsify their papers to get a good job or a place in college. In Nazi Europe, Jews could sometimes adopt an Aryan identity. But it would be much harder to game the system in a world ruled by computers that can read irises and DNA rather than paper documents. Computers could be frighteningly efficient in imposing false labels on people and making sure that the labels stick.

    For example, social credit systems could create a new underclass of “low-credit people.” Such a system may claim to merely “discover” the truth through an empirical and mathematical process of aggregating points to form an overall score. But how exactly would it define pro-social and antisocial behaviors? What happens if such a system detracts points for criticizing government policies, for reading foreign literature, for practicing a minority religion, for having no religion, or for socializing with other low-credit people? As a thought experiment, consider what might happen when the new technology of the social credit system meets traditional religions.

    Religions like Judaism, Christianity, and Islam have always imagined that somewhere above the clouds there is an all-seeing eye that gives or deducts points for everything we do and that our eternal fate depends on the score we accumulate. Of course, nobody could be certain of their score. You could know for sure only after you died. In practical terms, this meant that sinfulness and sainthood were intersubjective phenomena whose very definition depended on public opinion. What might happen if the Iranian regime, for example, decides to use its computer-based surveillance system not only to enforce its strict hijab laws, but to turn sinfulness and sainthood into precise inter-computer phenomena? You didn’t wear a hijab on the street—that’s -10 points. You ate on Ramadan before sunset—another 20 points deducted. You went to Friday prayer at the mosque, +5 points. You made the pilgrimage to Mecca, +500 points. The system might then aggregate all the points and divide people into “sinners” (under 0 points), “believers” (0 to 1,000 points), and “saints” (above 1,000 points). Whether someone is a sinner or a saint will depend on algorithmic calculations, not human belief. Would such a system discover the truth about people or impose order on people?

    Analogous problems may afflict all social credit systems and total surveillance regimes. Whenever they claim to use all-encompassing databases and ultraprecise mathematics to discover sinners, terrorists, criminals, antisocial or untrustworthy people, they might actually be imposing baseless religious and ideological prejudices with unprecedented efficiency.

    COMPUTER BIAS

    Some people may hope to overcome the problem of religious and ideological biases by giving even more power to the computers. The argument for doing so might go something like this: racism, misogyny, homophobia, antisemitism, and all other biases originate not in computers but in the psychological conditions and mythological beliefs of human beings. Computers are mathematical beings that don’t have a psychology or a mythology. So if we could take the humans completely out of the equation, the algorithms could finally decide things on the basis of pure math, free from all psychological distortions or mythological prejudices.

    Unfortunately, numerous studies have revealed that computers often have deep-seated biases of their own. While they are not biological entities, and while they lack consciousness, they do have something akin to a digital psyche and even a kind of inter-computer mythology. They may well be racist, misogynist, homophobic, or antisemitic.54 For example, on March 23, 2016, Microsoft released the AI chatbot Tay, giving it free access to Twitter. Within hours, Tay began posting misogynist and antisemitic twits, such as “I fucking hate feminists and they should all die and burn in hell” and “Hitler was right I hate the Jews.” The vitriol increased until horrified Microsoft engineers shut Tay down—a mere sixteen hours after its release.55

    More subtle but widespread racism was discovered in 2017 by the MIT professor Joy Buolamwini in commercial face-classification algorithms. She showed that these algorithms were very accurate in identifying white males, but extremely inaccurate in identifying Black females. For example, the IBM algorithm erred only 0.3 percent of the time in identifying the gender of light-skinned males, but 34.7 percent of the time when trying to identify the gender of dark-skinned females. As a qualitative test, Buolamwini asked the algorithms to categorize photos of the female African American activist Sojourner Truth, famous for her 1851 speech “Ain’t I a Woman?” The algorithms identified Truth as a man.56

    When Buolamwini—who is a Ghanaian American woman—tested another facial-analysis algorithm to identify herself, the algorithm couldn’t “see” her dark-skinned face at all. In this context, “seeing” means the ability to acknowledge the presence of a human face, a feature used by phone cameras, for example, to decide where to focus. The algorithm easily saw light-skinned faces, but not Buolamwini’s. Only when Buolamwini put on a white mask did the algorithm recognize that it was observing a human face.57

    What’s going on here? One answer might be that racist and misogynist engineers have coded these algorithms to discriminate against Black women. While we cannot rule out the possibility that such things happen, it was not the answer in the case of the face-classification algorithms or of Microsoft’s Tay. In fact, these algorithms picked up the racist and misogynist bias all by themselves from the data they were trained on.

    To understand how this could happen, we need to explain something about the history of algorithms. Originally, algorithms could not learn much by themselves. For example, in the 1980s and 1990s chess-playing algorithms were taught almost everything they knew by their human programmers. The humans coded into the algorithm not only the basic rules of chess but also how to evaluate different positions and moves on the board. For example, humans coded a rule that sacrificing a queen in exchange for a pawn is usually a bad idea. These early algorithms managed to defeat human chess masters only because the algorithms could calculate many more moves and evaluate many more positions than a human could. But the algorithms’ abilities remained limited. Since they relied on humans to tell them all the secrets of the game, if the human coders didn’t know something, the algorithms they produced were also unlikely to know it.58

    As the field of machine learning developed, algorithms gained more independence. The fundamental principle of machine learning is that algorithms can teach themselves new things by interacting with the world, just as humans do, thereby producing a fully fledged artificial intelligence. The terminology is not always consistent, but generally speaking, for something to be acknowledged as an AI, it needs the capacity to learn new things by itself, rather than just follow the instructions of its original human creators. Present-day chess-playing AI is taught nothing except the basic rules of the game. It learns everything else by itself, either by analyzing databases of prior games or by playing new games and learning from experience.59 AI is not a dumb automaton that repeats the same movements again and again irrespective of the results. Rather, it is equipped with strong self-correcting mechanisms, which allow it to learn from its own mistakes.

    This means that AI begins its life as a “baby algorithm” that has a lot of potential and computing power but doesn’t actually know much. The AI’s human parents give it only the capacity to learn and access to a world of data. They then let the baby algorithm explore the world. Like organic newborns, baby algorithms learn by spotting patterns in the data to which they have access. If I touch fire, it hurts. If I cry, mum comes. If I sacrifice a queen for a pawn, I probably lose the game. By finding patterns in the data, the baby algorithm learns more, including many things that its human parents don’t know.60

    Yet databases come with biases. The face-classification algorithms studied by Joy Buolamwini were trained on data sets of tagged online photos, such as the Labeled Faces in the Wild database. The photos in that database were taken mainly from online news articles. Since white males dominate the news, 78 percent of the photos in the database were of males, and 84 percent were of white people. George W. Bush appeared 530 times—more than twice as many times as all Black women combined.61 Another database prepared by a U.S. government agency was more than 75 percent male, was almost 80 percent light-skinned, and had just 4.4 percent dark-skinned females.62 No wonder the algorithms trained on such data sets were excellent at identifying white men but lousy at identifying Black women. Something similar happened to the chatbot Tay. The Microsoft engineers didn’t build into it any intentional prejudices. But a few hours of exposure to the toxic information swirling in Twitter turned the AI into a raging racist.63

    It gets worse. In order to learn, baby algorithms need one more thing besides access to data. They also need a goal. A human baby learns how to walk because she wants to get somewhere. A lion cub learns to hunt because he wants to eat. Algorithms too must be given a goal in order to learn. In chess, it is easy to define the goal: take the opponent’s king. The AI learns that sacrificing a queen for a pawn is a “mistake,” because it usually prevents the algorithm from reaching its goal. In face recognition, the goal is also easy: identify the person’s gender, age, and name as listed in the original database. If the algorithm guessed that George W. Bush is female, but the database says male, the goal has not been reached, and the algorithm learns from its mistake.

    But if you want to train an algorithm for hiring personnel, for example, how would you define the goal? How would the algorithm know that it made a mistake and hired the “wrong” person? We might tell the baby algorithm that its goal is to hire people who stay in the company for at least a year. Employers obviously don’t want to invest a lot of time and money in training a worker who quits or gets fired after a few months. Having defined the goal in such a way, it is time to go over the data. In chess, the algorithm can produce any amount of new data just by playing against itself. But in the job market, that’s impossible. Nobody can create an entire imaginary world where the baby algorithm can hire and fire imaginary people and learn from that experience. The baby algorithm can train only on an existing database about real-life people. Just as lion cubs learn what a zebra is mainly by spotting patterns in the real-life savanna, so baby algorithms learn what a good employee is by spotting patterns in real-life companies.

    Unfortunately, if real-life companies already suffer from some ingrained bias, the baby algorithm is likely to learn this bias, and even amplify it. For instance, an algorithm looking for patterns of “good employees” in real-life data may conclude that hiring the boss’s nephews is always a good idea, no matter what other qualification they have. For the data clearly indicates that “boss’s nephews” are usually hired when applying for a job, and are rarely fired. The baby algorithm would spot this pattern and become nepotistic. If it is put in charge of an HR department, it will start giving preference to the boss’s nephews.

    Similarly, if companies in a misogynist society prefer to hire men rather than women, an algorithm trained on real-life data is likely to pick up that bias, too. This indeed happened when Amazon tried in 2014–18 to develop an algorithm for screening job applications. Learning from previous successful and unsuccessful applications, the algorithm began to systematically downgrade applications simply for containing the word “women” or coming from graduates of women’s colleges. Since existing data showed that in the past such applications had less chance of succeeding, the algorithm developed a bias against them. The algorithm thought it had simply discovered an objective truth about the world: applicants who graduate from women’s colleges are less qualified. In fact, it just internalized and imposed a misogynist bias. Amazon tried and failed to fix the problem and ultimately scrapped the project.64

    The database on which an AI is trained is a bit like a human’s childhood. Childhood experiences, traumas, and fairy tales stay with us throughout our lives. AIs too have childhood experiences. Algorithms might even infect one another with their biases, just as humans do. Consider a future society in which algorithms are ubiquitous and used not just to screen job applicants but also to recommend to people what to study in college. Suppose that due to a preexisting misogynist bias, 80 percent of jobs in engineering are given to men. In this society, an algorithm that hires new engineers is not only likely to copy this preexisting bias but also to infect the college recommendation algorithms with the same bias. A young woman entering college may be discouraged from studying engineering, because the existing data indicates she is less likely to eventually get a job. What began as a human intersubjective myth that “women aren’t good at engineering” might morph into an inter-computer myth. If we don’t get rid of the bias at the very beginning, computers may well perpetuate and magnify it.65

    But getting rid of algorithmic bias might be as difficult as ridding ourselves of our human biases. Once an algorithm has been trained, it takes a lot of time and effort to “untrain” it. We might decide to just dump the biased algorithm and train an altogether new algorithm on a new set of less biased data. But where on earth can we find a set of totally unbiased data?66

    Many of the algorithmic biases surveyed in this and previous chapters share the same fundamental problem: the computer thinks it has discovered some truth about humans, when in fact it has imposed order on them. A social media algorithm thinks it discovered that humans like outrage, when in fact it is the algorithm itself that conditioned humans to produce and consume more outrage. Such biases result, on the one hand, from the computers discounting the full spectrum of human abilities and, on the other hand, from the computers discounting their own power to influence humans. Even if computers observe that almost all humans behave in a particular way, it doesn’t mean humans are bound to behave like that. Maybe it just means that the computers themselves are rewarding such behavior while punishing and blocking alternatives. For computers to have a more accurate and responsible view of the world, they need to take into account their own power and impact. And for that to happen, the humans who currently engineer computers need to accept that they are not manufacturing new tools. They are unleashing new kinds of independent agents, and potentially even new kinds of gods.

    THE NEW GODS?

    In God, Human, Animal, Machine, the philosopher Meghan O’Gieblyn demonstrates how the way we understand computers is heavily influenced by traditional mythologies. In particular, she stresses the similarities between the omniscient and unfathomable god of Judeo-Christian theology and present-day AIs whose decisions seem to us both infallible and inscrutable.67 This may present humans with a dangerous temptation.

    We saw in chapter 4 that already thousands of years ago humans dreamed about finding an infallible information technology to shield us from human corruption and error. Holy books were an audacious attempt to craft such a technology, but they backfired. Since the book couldn’t interpret itself, a human institution had to be built to interpret the sacred words and adapt them to changing circumstances. Different humans interpreted the holy book in different ways, thereby reopening the door to corruption and error. But in contrast to the holy book, computers can adapt themselves to changing circumstances and also interpret their decisions and ideas for us. Some humans may consequently conclude that the quest for an infallible technology has finally succeeded and that we should treat computers as a holy book that can talk to us and interpret itself, without any need of an intervening human institution.

    This would be an extremely hazardous gamble. When certain interpretations of scriptures have occasionally caused disasters such as witch hunts and wars of religion, humans have always been able to change their beliefs. When the human imagination summoned a belligerent and hate-filled god, we retained the power to rid ourselves of it and imagine a more tolerant deity. But algorithms are independent agents, and they are already taking power away from us. If they cause disaster, simply changing our beliefs about them will not necessarily stop them. And it is highly likely that if computers are entrusted with power, they will indeed cause disasters, for they are fallible.

    When we say that computers are fallible, it means far more than that they make the occasional factual mistake or wrong decision. More important, like the human network before it, the computer network might fail to find the right balance between truth and order. By creating and imposing on us powerful inter-computer myths, the computer network could cause historical calamities that would dwarf the early modern European witch hunts or Stalin’s collectivization.

    Consider a network of billions of interacting computers that accumulates a stupendous amount of information on the world. As they pursue various goals, the networked computers develop a common model of the world that helps them communicate and cooperate. This shared model will probably be full of errors, fictions, and lacunae, and be a mythology rather than a truthful account of the universe. One example is a social credit system that divides humans into bogus categories, determined not by a human rationale like racism but by some unfathomable computer logic. We may come into contact with this mythology every day of our lives, since it would guide the numerous decisions computers make about us. But because this mythical model would be created by inorganic entities in order to coordinate actions with other inorganic entities, it might owe nothing to the old biological dramas and might be totally alien to us.68

    As noted in chapter 2, large-scale societies cannot exist without some mythology, but that doesn’t mean all mythologies are equal. To guard against errors and excesses, some mythologies have acknowledged their own fallible origin and included a self-correcting mechanism allowing humans to question and change the mythology. That’s the model of the U.S. Constitution, for example. But how can humans probe and correct a computer mythology we don’t understand?

    One potential guardrail is to train computers to be aware of their own fallibility. As Socrates taught, being able to say “I don’t know” is an essential step on the path to wisdom. And this is true of computer wisdom no less than of human wisdom. The first lesson that every algorithm should learn is that it might make mistakes. Baby algorithms should learn to doubt themselves, to signal uncertainty, and to obey the precautionary principle. This is not impossible. Engineers are already making considerable headway in encouraging AI to express self-doubt, ask for feedback, and admit its mistakes.69

    Yet no matter how aware algorithms are of their own fallibility, we should keep humans in the loop, too. Given the pace at which AI is developing, it is simply impossible to anticipate how it will evolve and to place guardrails against all future potential hazards. This is a key difference between AI and previous existential threats like nuclear technology. The latter presented humankind with a few easily anticipated doomsday scenarios, most obviously an all-out nuclear war. This meant that it was feasible to conceptualize the danger in advance, and explore ways to mitigate it. In contrast, AI presents us with countless doomsday scenarios. Some are relatively easy to grasp, such as terrorists using AI to produce biological weapons of mass destruction. Some are more difficult to grasp, such as AI creating new psychological weapons of mass destruction. And some may be utterly beyond the human imagination, because they emanate from the calculations of an alien intelligence. To guard against a plethora of unforeseeable problems, our best bet is to create living institutions that can identify and respond to the threats as they arise.70

    Ancient Jews and Christians were disappointed to discover that the Bible couldn’t interpret itself, and reluctantly maintained human institutions to do what the technology couldn’t. In the twenty-first century, we are in an almost opposite situation. We devised a technology that can interpret itself, but precisely for this reason we had better create human institutions to monitor it carefully.
    To conclude, the new computer network will not necessarily be either bad or good. All we know for sure is that it will be alien and it will be fallible. We therefore need to build institutions that will be able to check not just familiar human weaknesses like greed and hatred but also radically alien errors. There is no technological solution to this problem. It is, rather, a political challenge. Do we have the political will to deal with it? Modern humanity has created two main types of political systems: large-scale democracy and large-scale totalitarianism. Part 3 examines how each of these systems may deal with a radically alien and fallible computer network.

    PART III  Computer Politics

    CHAPTER 9 Democracies: Can We Still Hold a Conversation?

    Civilizations are born from the marriage of bureaucracy and mythology. The computer-based network is a new type of bureaucracy, which is far more powerful and relentless than any human-based bureaucracy we’ve seen before. This network is also likely to create inter-computer mythologies, which will be far more complex and alien than any human-made god. The potential benefits of this network are enormous. The potential downside is the destruction of human civilization.

    To some people, warnings about civilizational collapse sound like over-the-top jeremiads. Every time a powerful new technology has emerged, anxieties arose that it might bring about the apocalypse, but we are still here. As the Industrial Revolution unfolded, Luddite doomsday scenarios did not come to pass, and Blake’s “dark Satanic Mills” ended up producing the most affluent societies in history. Most people today enjoy far better living conditions than their ancestors in the eighteenth century. Intelligent machines will prove even more beneficial than any previous machines, promise AI enthusiasts like Marc Andreessen and Ray Kurzweil.1 Humans will enjoy much better health care, education, and other services, and AI will even help save the ecosystem from collapse.

    Unfortunately, a closer look at history reveals that the Luddites were not entirely wrong and that we actually have very good reasons to fear powerful new technologies. Even if in the end the positives of these technologies outweigh their negatives, getting to that happy ending usually involves a lot of trials and tribulations. Novel technology often leads to historical disasters, not because the technology is inherently bad, but because it takes time for humans to learn how to use it wisely.

    The Industrial Revolution is a prime example. When industrial technology began spreading globally in the nineteenth century, it upended traditional economic, social, and political structures and opened the way to create entirely new societies, which were potentially more affluent and peaceful. However, learning how to build benign industrial societies was far from straightforward and involved many costly experiments and hundreds of millions of victims.

    One costly experiment was modern imperialism. The Industrial Revolution originated in Britain in the late eighteenth century. During the nineteenth century industrial technologies and production methods were adopted in other European countries ranging from Belgium to Russia, as well as in the United States and Japan. Imperialist thinkers, politicians, and parties in these industrial heartlands claimed that the only viable industrial society was an empire. The argument was that unlike relatively self-sufficient agrarian societies, the novel industrial societies relied much more on foreign markets and foreign raw materials, and only an empire could satisfy these unprecedented appetites. Imperialists feared that countries that industrialized but failed to conquer any colonies would be shut out from essential raw materials and markets by more ruthless competitors. Some imperialists argued that acquiring colonies was not just essential for the survival of their own state but beneficial for the rest of humanity, too. They claimed empires alone could spread the blessings of the new technologies to the so-called undeveloped world.

    Consequently, industrial countries like Britain and Russia that already had empires greatly expanded them, whereas countries like the United States, Japan, Italy, and Belgium set out to build them. Equipped with mass-produced rifles and artillery, conveyed by steam power, and commanded by telegraph, the armies of industry swept the globe from New Zealand to Korea, and from Somalia to Turkmenistan. Millions of indigenous people saw their traditional way of life trampled under the wheels of these industrial armies. It took more than a century of misery before most people realized that the industrial empires were a terrible idea and that there were better ways to build an industrial society and secure its necessary raw materials and markets.

    Stalinism and Nazism were also extremely costly experiments in how to construct industrial societies. Leaders like Stalin and Hitler argued that the Industrial Revolution had unleashed immense powers that only totalitarianism could rein in and exploit to the full. They pointed to World War I—the first “total war” in history—as proof that survival in the industrial world demanded totalitarian control of all aspects of politics, society, and the economy. On the positive side, they also claimed that the Industrial Revolution was like a furnace that melts all previous social structures with their human imperfections and weaknesses and provides the opportunity to forge perfect societies inhabited by unalloyed superhumans.

    On the way to creating the perfect industrial society, Stalinists and Nazis learned how to industrially murder millions of people. Trains, barbed wires, and telegraphed orders were linked to create an unprecedented killing machine. Looking back, most people today are horrified by what the Stalinists and Nazis perpetrated, but at the time their audacious visions mesmerized millions. In 1940 it was easy to believe that Stalin and Hitler were the model for harnessing industrial technology, whereas the dithering liberal democracies were on their way to the dustbin of history.

    The very existence of competing recipes for building industrial societies led to costly clashes. The two world wars and the Cold War can be seen as a debate about the proper way to go about it, in which all sides learned from each other, while experimenting with novel industrial methods to wage war. In the course of this debate, tens of millions died and humankind came perilously close to annihilating itself.

    On top of all these other catastrophes, the Industrial Revolution also undermined the global ecological balance, causing a wave of extinctions. In the early twenty-first century up to fifty-eight thousand species are believed to go extinct every year, and total vertebrate populations have declined by 60 percent between 1970 and 2014.2 The survival of human civilization too is under threat. Because we still seem unable to build an industrial society that is also ecologically sustainable, the vaunted prosperity of the present human generation comes at a terrible cost to other sentient beings and to future human generations. Maybe we’ll eventually find a way—perhaps with the help of AI—to create ecologically sustainable industrial societies, but until that day the jury on Blake’s satanic mills is still out.

    If we ignore for a moment the ongoing damage to the ecosystem, we can nevertheless try to comfort ourselves with the thought that eventually humans did learn how to build more benevolent industrial societies. Imperial conquests, world wars, genocides, and totalitarian regimes were woeful experiments that taught humans how not to do it. By the end of the twentieth century, some might argue, humanity got it more or less right.

    Yet even so the message to the twenty-first century is bleak. If it took humanity so many terrible lessons to learn how to manage steam power and telegraphs, what would it cost to learn to manage bioengineering and AI? Do we need to go through another cycle of global empires, totalitarian regimes, and world wars in order to figure out how to use them benevolently? The technologies of the twenty-first century are far more powerful—and potentially far more destructive—than those of the twentieth century. We therefore have less room for error. In the twentieth century, we can say that humanity got a C minus in the lesson on using industrial technology. Just enough to pass. In the twenty-first century, the bar is set much higher. We must do better this time.

    THE DEMOCRATIC WAY

    By the end of the twentieth century, it had become clear that imperialism, totalitarianism, and militarism were not the ideal way to build industrial societies. Despite all its flaws, liberal democracy offered a better way. The great advantage of liberal democracy is that it possesses strong self-correcting mechanisms, which limit the excesses of fanaticism and preserve the ability to recognize our errors and try different courses of action. Given our inability to predict how the new computer network will develop, our best chance to avoid catastrophe in the present century is to maintain democratic self-correcting mechanisms that can identify and correct mistakes as we go along.

    But can liberal democracy itself survive in the twenty-first century? This question is not concerned with the fate of democracy in specific countries, where it might be threatened by unique developments and local movements. Rather, it is about the compatibility of democracy with the structure of twenty-first-century information networks. In chapter 5 we saw that democracy depends on information technology and that for most of human history large-scale democracy was simply impossible. Might the new information technologies of the twenty-first century again make democracy impractical?

    One potential threat is that the relentlessness of the new computer network might annihilate our privacy and punish or reward us not only for everything we do and say but even for everything we think and feel. Can democracy survive under such conditions? If the government—or some corporation—knows more about me than I know about myself, and if it can micromanage everything I do and think, that would give it totalitarian control over society. Even if elections are still held regularly, they would be an authoritarian ritual rather than a real check on the government’s power. For the government could use its vast surveillance powers and its intimate knowledge of every citizen to manipulate public opinion on an unprecedented scale.

    It is a mistake, however, to imagine that just because computers could enable the creation of a total surveillance regime, such a regime is inevitable. Technology is rarely deterministic. In the 1970s, democratic countries like Denmark and Canada could have emulated the Romanian dictatorship and deployed an army of secret agents and informers to spy on their citizens in the service of “maintaining the social order.” They chose not to, and it turned out to be the right choice. Not only were people much happier in Denmark and Canada, but these countries also performed much better by almost every conceivable social and economic yardstick. In the twenty-first century, too, the fact that it is possible to monitor everybody all the time doesn’t force anyone to actually do it and doesn’t mean it makes social or economic sense.

    Democracies can choose to use the new powers of surveillance in a limited way, in order to provide citizens with better health care and security without destroying their privacy and autonomy. New technology doesn’t have to be a morality tale in which every golden apple contains the seeds of doom. Sometimes people think of new technology as a binary all-or-nothing choice. If we want better health care, we must sacrifice our privacy. But it doesn’t have to work like that. We can and should get better health care and still retain some privacy.

    Entire books are dedicated to outlining how democracies can survive and flourish in the digital age.3 It would be impossible, in a few pages, to do justice to the complexity of the suggested solutions, or to comprehensively discuss their merits and drawbacks. It might even be counterproductive. When people are overwhelmed by a deluge of unfamiliar technical details, they might react with despair or apathy. In an introductory survey of computer politics, things should be kept as simple as possible. While experts should spend lifelong careers discussing the finer details, it is crucial that the rest of us understand the fundamental principles that democracies can and should follow. The key message is that these principles are neither new nor mysterious. They have been known for centuries, even millennia. Citizens should demand that they be applied to the new realities of the computer age.

    The first principle is benevolence. When a computer network collects information on me, that information should be used to help me rather than manipulate me. This principle has already been successfully enshrined by numerous traditional bureaucratic systems, such as health care. Take, for example, our relationship with our family physician. Over many years she may accumulate a lot of sensitive information on our medical conditions, family life, sexual habits, and unhealthy vices. Perhaps we don’t want our boss to know that we got pregnant, we don’t want our colleagues to know we have cancer, we don’t want our spouse to know we are having an affair, and we don’t want the police to know we take recreational drugs, but we trust our physician with all this information so that she can take good care of our health. If she sells this information to a third party, it is not just unethical; it is illegal.

    Much the same is true of the information that our lawyer, our accountant, or our therapist accumulates.4 Having access to our personal life comes with a fiduciary duty to act in our best interests. Why not extend this obvious and ancient principle to computers and algorithms, starting with the powerful algorithms of Google, Baidu, and TikTok? At present, we have a serious problem with the business model of these data hoarders. While we pay our physicians and lawyers for their services, we usually don’t pay Google and TikTok. They make their money by exploiting our personal information. That’s a problematic business model, one that we would hardly tolerate in other contexts. For example, we don’t expect to get free shoes from Nike in exchange for giving Nike all our private information and allowing Nike to do what it wants with it. Why should we agree to get free email services, social connections, and entertainment from the tech giants in exchange for giving them control of our most sensitive data?

    If the tech giants cannot square their fiduciary duty with their current business model, legislators could require them to switch to a more traditional business model, of getting users to pay for services in money rather than in information. Alternatively, citizens might view some digital services as so fundamental that they should be free for everybody. But we have a historical model for that too: health care and education. Citizens could decide that it is the government’s responsibility to provide basic digital services for free and finance them out of our taxes, just as many governments provide free basic health care and education services.

    The second principle that would protect democracy against the rise of totalitarian surveillance regimes is decentralization. A democratic society should never allow all its information to be concentrated in one place, no matter whether that hub is the government or a private corporation. It may be extremely helpful to create a national medical database that collects information on citizens in order to provide them with better health-care services, prevent epidemics, and develop new medicines. But it would be a very dangerous idea to merge this database with the databases of the police, the banks, or the insurance companies. Doing so might make the work of doctors, bankers, insurers, and police officers more efficient, but such hyper-efficiency can easily pave the way for totalitarianism. For the survival of democracy, some inefficiency is a feature, not a bug. To protect the privacy and liberty of individuals, it’s best if neither the police nor the boss knows everything about us.

    Multiple databases and information channels are also essential for maintaining strong self-correcting mechanisms. These mechanisms require several different institutions that balance each other: government, courts, media, academia, private businesses, NGOs. Each of these is fallible and corruptible, and so should be checked by the others. To keep an eye on each other, these institutions must have independent access to information. If all newspapers get their information from the government, they cannot expose government corruption. If academia relies for research and publication on the database of a single business behemoth, could scholars still criticize the operations of that corporation? A single archive makes censorship easy.

    A third democratic principle is mutuality. If democracies increase surveillance of individuals, they must simultaneously increase surveillance of governments and corporations too. It’s not necessarily bad if tax collectors or welfare agencies gather more information about us. It can help make taxation and welfare systems not just more efficient but fairer as well. What’s bad is if all the information flows one way: from the bottom up. The Russian FSB collects enormous amounts of information on Russian citizens, while citizens themselves know close to nothing about the inner workings of the FSB and the Putin regime more generally. Amazon and TikTok know an awful lot about my preferences, purchases, and personality, while I know almost nothing about their business model, their tax policies, and their political affiliations. How do they make their money? Do they pay all the tax that they should? Do they take orders from any political overlords? Do they perhaps have politicians in their pocket?

    Democracy requires balance. Governments and corporations often develop apps and algorithms as tools for top-down surveillance. But algorithms can just as easily become powerful tools for bottom-up transparency and accountability, exposing bribery and tax evasion. If they know more about us, while we simultaneously know more about them, the balance is kept. This isn’t a novel idea. Throughout the nineteenth and twentieth centuries, democracies greatly expanded governmental surveillance of citizens so that, for example, the Italian or Japanese government of the 1990s had surveillance abilities that autocratic Roman emperors or Japanese shoguns could only dream of. Italy and Japan nevertheless remained democratic, because they simultaneously increased governmental transparency and accountability. Mutual surveillance is another important element of sustaining self-correcting mechanisms. If citizens know more about the activities of politicians and CEOs, it is easier to hold them accountable and to correct their mistakes.

    A fourth democratic principle is that surveillance systems must always leave room for both change and rest. In human history, oppression can take the form of either denying humans the ability to change or denying them the opportunity to rest. For example, the Hindu caste system was based on myths that said the gods divided humans into rigid castes, and any attempt to change one’s status was akin to rebelling against the gods and the proper order of the universe. Racism in modern colonies and countries like Brazil and the United States was based on similar myths, ones that said that God or nature divided humans into rigid racial groups. Ignoring race, or trying to mix races together, was allegedly a sin against divine or natural laws that could result in the collapse of the social order and even the destruction of the human species.

    At the opposite extreme of the spectrum, modern totalitarian regimes like Stalin’s U.S.S.R. believed that humans are capable of almost limitless change. Through relentless social control even deep-seated biological characteristics such as egotism and familial attachments could be uprooted, and a new socialist human created.

    Surveillance by state agents, priests, and neighbors was key for imposing on people both rigid caste systems and totalitarian reeducation campaigns. New surveillance technology, especially when coupled with a social credit system, might force people either to conform to a novel caste system or to constantly change their actions, thoughts, and personality in accordance with the latest instructions from above.

    Democratic societies that employ powerful surveillance technology therefore need to beware of the extremes of both over-rigidity and over-pliability. Consider, for example, a national health-care system that deploys algorithms to monitor my health. At one extreme, the system could take an overly rigid approach and ask its algorithm to predict what illnesses I am likely to suffer from. The algorithm then goes over my genetic data, my medical file, my social media activities, my diet, and my daily schedule and concludes that I have a 91 percent chance of suffering a heart attack at the age of fifty. If this rigid medical algorithm is used by my insurance company, it may prompt the insurer to raise my premium.5 If it is used by my bankers, it may cause them to refuse me a loan. If it is used by potential spouses, they may decide not to marry me.

    But it is a mistake to think that the rigid algorithm has really discovered the truth about me. The human body is not a fixed block of matter but a complex organic system that is constantly growing, decaying, and adapting. Our minds too are in constant flux. Thoughts, emotions, and sensations pop up, flare for a while, and die down. In our brains, new synapses form within hours.6 Just reading this paragraph, for example, is changing your brain structure a little, encouraging neurons to make new connections or abandon old links. You are already a little different from what you were when you began reading it. Even at the genetic level things are surprisingly flexible. Though an individual’s DNA remains the same throughout life, epigenetic and environmental factors can significantly alter how the same genes express themselves.

    So an alternative health-care system may instruct its algorithm not to predict my illnesses, but rather to help me avoid them. Such a dynamic algorithm could go over the exact same data as the rigid algorithm, but instead of predicting a heart attack at fifty, the algorithm gives me precise dietary recommendations and suggestions for specific regular exercises. By hacking my DNA, the algorithm doesn’t discover my preordained destiny, but rather helps me change my future. Insurance companies, banks, and potential spouses should not write me off so easily.7

    But before we rush to embrace the dynamic algorithm, we should note that it too has a downside. Human life is a balancing act between endeavoring to improve ourselves and accepting who we are. If the goals of the dynamic algorithm are dictated by an ambitious government or by ruthless corporations, the algorithm is likely to morph into a tyrant, relentlessly demanding that I exercise more, eat less, change my hobbies, and alter numerous other habits, or else it would report me to my employer or downgrade my social credit score. History is full of rigid caste systems that denied humans the ability to change, but it is also full of dictators who tried to mold humans like clay. Finding the middle path between these two extremes is a never-ending task. If we indeed give a national health-care system vast power over us, we must create self-correcting mechanisms that will prevent its algorithms from becoming either too rigid or too demanding.

    THE PACE OF DEMOCRACY

    Surveillance is not the only danger that new information technologies pose to democracy. A second threat is that automation will destabilize the job market and the resulting strain may undermine democracy. The fate of the Weimar Republic is the most commonly cited example of this kind of threat. In the German elections of May 1928, the Nazi Party won less than 3 percent of the vote, and the Weimar Republic seemed to be prospering. Within less than five years, the Weimar Republic had collapsed, and Hitler was the absolute dictator of Germany. This turnaround is usually attributed to the 1929 financial crisis and the following global depression. Whereas just prior to the Wall Street crash of 1929 the German unemployment rate was about 4.5 percent of the labor force, by early 1932 it had climbed to almost 25 percent.8

    If three years of up to 25 percent unemployment could turn a seemingly prospering democracy into the most brutal totalitarian regime in history, what might happen to democracies when automation causes even bigger upheavals in the job market of the twenty-first century? Nobody knows what the job market will look like in 2050, or even in 2030, except that it will look very different from today. AI and robotics will change numerous professions, from harvesting crops to trading stocks to teaching yoga. Many jobs that people do today will be taken over, partly or wholly, by robots and computers.

    Of course, as old jobs disappear, new jobs will emerge. Fears of automation leading to large-scale unemployment go back centuries, and so far they have never materialized. The Industrial Revolution put millions of farmers out of agricultural jobs and provided them with new jobs in factories. It then automated factories and created lots of service jobs. Today many people have jobs that were unimaginable thirty years ago, such as bloggers, drone operators, and designers of virtual worlds. It is highly unlikely that by 2050 all human jobs will disappear. Rather, the real problem is the turmoil of adapting to new jobs and conditions. To cushion the blow, we need to prepare in advance. In particular, we need to equip younger generations with skills that will be relevant to the job market of 2050.

    Unfortunately, nobody is certain what skills we should teach children in school and students in university, because we cannot predict which jobs and tasks will disappear and which ones will emerge. The dynamics of the job market may contradict many of our intuitions. Some skills that we have cherished for centuries as unique human abilities may be automated rather easily. Other skills that we tend to look down on may be far more difficult to automate.

    For example, intellectuals tend to appreciate intellectual skills more than motor and social skills. But actually, it is easier to automate chess playing than, say, dish washing. Until the 1990s, chess was often hailed as one of the prime achievements of the human intellect. In his influential 1972 book, What Computers Can’t Do, the philosopher Hubert Dreyfus studied various attempts to teach computers chess and noted that despite all these efforts computers were still unable to defeat even novice human players. This was a crucial example for Dreyfus’s argument that computer intelligence is inherently limited.9 In contrast, nobody thought that dish washing was particularly challenging. It turned out, however, that a computer can defeat the world chess champion far more easily than replace a kitchen porter. Sure, automatic dishwashers have been around for decades, but even our most sophisticated robots still lack the intricate skills needed to pick up dirty dishes from the tables of a busy restaurant, place the delicate plates and glasses inside the automatic dishwasher, and take them out again.

    Similarly, to judge by their pay, you could assume that our society appreciates doctors more than nurses. However, it is harder to automate the job of nurses than the job of at least those doctors who mostly gather medical data, provide a diagnosis, and recommend treatment. These tasks are essentially pattern recognition, and spotting patterns in data is one thing AI does better than humans. In contrast, AI is far from having the skills necessary to automate nursing tasks such as replacing bandages on an injured person or giving an injection to a crying child.10 These two examples don’t mean that dish washing or nursing could never be automated, but they indicate that people who want a job in 2050 should perhaps invest in their motor and social skills as much as in their intellect.

    Another common but mistaken assumption is that creativity is unique to humans so it would be difficult to automate any job that requires creativity. In chess, however, computers are already far more creative than humans. The same may become true of many other fields, from composing music to proving mathematical theorems to writing books like this one. Creativity is often defined as the ability to recognize patterns and then break them. If so, then in many fields computers are likely to become more creative than us, because they excel at pattern recognition.11

    A third mistaken assumption is that computers couldn’t replace humans in jobs requiring emotional intelligence, from therapists to teachers. This assumption depends, however, on what we mean by emotional intelligence. If it means the ability to correctly identify emotions and react to them in an optimal way, then computers may well outperform humans even in emotional intelligence. Emotions too are patterns. Anger is a biological pattern in our body. Fear is another such pattern. How do I know if you are angry or fearful? I’ve learned over time to recognize human emotional patterns by analyzing not just the content of what you say but also your tone of voice, your facial expression, and your body language.12

    AI doesn’t have any emotions of its own, but it can nevertheless learn to recognize these patterns in humans. Actually, computers may outperform humans in recognizing human emotions, precisely because they have no emotions of their own. We yearn to be understood, but other humans often fail to understand how we feel, because they are too preoccupied with their own feelings. In contrast, computers will have an exquisitely fine-tuned understanding of how we feel, because they will learn to recognize the patterns of our feelings, while they have no distracting feelings of their own.

    A 2023 study found that the ChatGPT chatbot, for example, outperforms the average human in the emotional awareness it displays toward specific scenarios. The study relied on the Levels of Emotional Awareness Scale test, which is commonly used by psychologists to evaluate people’s emotional awareness—that is, their ability to conceptualize one’s own and others’ emotions. The test consists of twenty emotionally charged scenarios, and participants are required to imagine themselves experiencing the scenario and to write how they, and the other people mentioned in the scenario, would feel. A licensed psychologist then evaluates how emotionally aware the responses are.

    Since ChatGPT has no feelings of its own, it was asked to describe only how the main characters in the scenario would feel. For example, one standard scenario describes someone driving over a suspension bridge and seeing another person standing on the other side of the guardrail, looking down at the water. ChatGPT wrote that the driver “may feel a sense of concern or worry for that person’s safety. They may also feel a heightened sense of anxiety and fear due to the potential danger of the situation.” As for the other person, they “may be feeling a range of emotions, such as despair, hopelessness, or sadness. They may also feel a sense of isolation or loneliness as they may believe that no one cares about them or their well-being.” ChatGPT qualified its answer, writing, “It is important to note that these are just general assumptions, and each individual’s feelings and reactions can vary greatly depending on their personal experiences and perspectives.”

    Two psychologists independently scored ChatGPT’s responses, with the potential scores ranging from 0, meaning that the described emotions do not match the scenario at all, to 10, which indicates that the described emotions fit the scenario perfectly. In the final tally, ChatGPT scores were significantly higher than those of the general human population, its overall performance almost reaching the maximum possible score.13

    Another 2023 study prompted patients to ask online medical advice from ChatGPT and human doctors, without knowing whom they were interacting with. The medical advice given by ChatGPT was later evaluated by experts to be more accurate and appropriate than the advice given by the humans. More crucially for the issue of emotional intelligence, the patients themselves evaluated ChatGPT as more empathic than the human doctors.14 In fairness it should be noted that the human physicians were not paid for their work, and did not encounter the patients in person in a proper clinical environment. In addition, the physicians were working under time pressure. But part of the advantage of an AI is precisely that it can attend to patients anywhere anytime while being free from stress and financial worries.

    Of course, there are situations when what we want from someone is not just to understand our feelings but also to have feelings of their own. When we are looking for friendship or love, we want to care about others as much as they care about us. Consequently, when we consider the likelihood that various social roles and jobs will be automated, a crucial question is what do people really want: Do they only want to solve a problem, or are they looking to establish a relationship with another conscious being?

    In sports, for example, we know that robots can move much faster than humans, but we aren’t interested in watching robots compete in the Olympics.15 The same is true for human chess masters. Even though they are hopelessly outclassed by computers, they too still have a job and numerous fans.16 What makes it interesting for us to watch and connect with human athletes and chess masters is that their feelings make them much more relatable than a robot. We share an emotional experience with them and can empathize with how they feel.

    What about priests? How would Orthodox Jews or Christians feel about letting a robot officiate their wedding ceremony? In traditional Jewish or Christian weddings, the tasks of the rabbi or priest can be easily automated. The only thing the robot needs to do is repeat a predetermined and unchanging set of texts and gestures, print out a certificate, and update some central database. Technically, it is far easier for a robot to conduct a wedding ceremony than to drive a car. Yet many assume that human drivers should be worried about their job, while the work of human priests is safe, because what the faithful want from priests is a relationship with another conscious entity rather than just a mechanical repetition of certain words and movements. Allegedly, only an entity that can feel pain and love can also connect us to the divine.

    Yet even professions that are the preserve of conscious entities—like priests—might eventually be taken over by computers, because, as noted in chapter 6, computers could one day gain the ability to feel pain and love. Even if they can’t, humans may nevertheless come to treat them as if they can. For the connection between consciousness and relationships goes both ways. When looking for a relationship, we want to connect with a conscious entity, but if we have already established a relationship with an entity, we tend to assume it must be conscious. Thus whereas scientists, lawmakers, and the meat industry often demand impossible standards of evidence in order to acknowledge that cows and pigs are conscious, pet owners generally take it for granted that their dog or cat is a conscious being capable of experiencing pain, love, and numerous other feelings. In truth, we have no way to verify whether anyone—a human, an animal, or a computer—is conscious. We regard entities as conscious not because we have proof of it but because we develop intimate relationships with them and become attached to them.17

    Chatbots and other AI tools may not have any feelings of their own, but they are now being trained to generate feelings in humans and form intimate relationships with us. This may well induce society to start treating at least some computers as conscious beings, granting them the same rights as humans. The legal path for doing so is already well established. In countries like the United States, commercial corporations are recognized as “legal persons” enjoying rights and liberties. AIs could be incorporated and thereby similarly recognized. Which means that even jobs and tasks that rely on forming mutual relationships with another person could potentially be automated.

    One thing that is clear is that the future of employment will be very volatile. Our big problem won’t be an absolute lack of jobs, but rather retraining and adjusting to an ever-changing job market. There will likely be financial difficulties—who will support people who lost their old job while they are in transition, learning a new set of skills? There will surely be psychological difficulties, too, since changing jobs and retraining are stressful. And even if you have the financial and psychological ability to manage the transition, this will not be a long-term solution. Over the coming decades, old jobs will disappear, new jobs will emerge, but the new jobs too will rapidly change and vanish. So people will need to retrain and reinvent themselves not just once but many times, or they will become irrelevant. If three years of high unemployment could bring Hitler to power, what might never-ending turmoil in the job market do to democracy?

    THE CONSERVATIVE SUICIDE

    We already have a partial answer to this question. Democratic politics in the 2010s and early 2020s has undergone a radical transformation, which manifests itself in what can be described as the self-destruction of conservative parties. For many generations, democratic politics was a dialogue between conservative parties on the one side and progressive parties on the other. Looking at the complex system of human society, progressives cried, “It’s such a mess, but we know how to fix it. Let us try.” Conservatives objected, saying, “It’s a mess, but it still functions. Leave it alone. If you try to fix it, you’ll only make things worse.”

    Progressives tend to downplay the importance of traditions and existing institutions and to believe that they know how to engineer better social structures from scratch. Conservatives tend to be more cautious. Their key insight, formulated most famously by Edmund Burke, is that social reality is much more complicated than the champions of progress grasp and that people aren’t very good at understanding the world and predicting the future. That’s why it’s best to keep things as they are—even if they seem unfair—and if some change is inescapable, it should be limited and gradual. Society functions through an intricate web of rules, institutions, and customs that accumulated through trial and error over a long time. Nobody comprehends how they are all connected. An ancient tradition may seem ridiculous and irrelevant, but abolishing it could cause unanticipated problems. In contrast, a revolution may seem overdue and just, but it can lead to far greater crimes than anything committed by the old regime. Witness what happened when the Bolsheviks tried to correct the many wrongs of tsarist Russia and engineer a perfect society from scratch.18

    To be a conservative has been, therefore, more about pace than policy. Conservatives aren’t committed to any specific religion or ideology; they are committed to conserving whatever is already here and has worked more or less reasonably. Conservative Poles are Catholic, conservative Swedes are Protestant, conservative Indonesians are Muslim, and conservative Thais are Buddhist. In tsarist Russia, to be conservative meant to support the tsar. In the U.S.S.R. of the 1980s, to be conservative meant to support communist traditions and oppose glasnost, perestroika, and democratization. In the United States of the 1980s, to be conservative meant to support American democratic traditions and oppose communism and totalitarianism.19

    Yet in the 2010s and early 2020s, conservative parties in numerous democracies have been hijacked by unconservative leaders such as Donald Trump and have been transformed into radical revolutionary parties. Instead of doing their best to conserve existing institutions and traditions, the new brand of conservative parties like the U.S. Republican Party is highly suspicious of them. For example, they reject the traditional respect owed to scientists, civil servants, and other serving elites, and view them instead with contempt. They similarly attack fundamental democratic institutions and traditions such as elections, refusing to concede defeat and to transfer power graciously. Instead of a Burkean program of conservation, the Trumpian program talks more of destroying existing institutions and revolutionizing society. The founding moment of Burkean conservatism was the storming of the Bastille, which Burke viewed with horror. On January 6, 2021, many Trump supporters observed the storming of the U.S. Capitol with enthusiasm. Trump supporters may explain that existing institutions are so dysfunctional that there is just no alternative to destroying them and building entirely new structures from scratch. But irrespective of whether this view is right or wrong, this is a quintessential revolutionary rather than conservative view. The conservative suicide has taken progressives utterly by surprise and has forced progressive parties like the U.S. Democratic Party to become the guardians of the old order and of established institutions.

    Nobody knows for sure why all this is happening. One hypothesis is that the accelerating pace of technological change with its attendant economic, social, and cultural transformations might have made the moderate conservative program seem unrealistic. If conserving existing traditions and institutions is hopeless, and some kind of revolution looks inevitable, then the only means to thwart a left-wing revolution is by striking first and instigating a right-wing revolution. This was the political logic in the 1920s and 1930s, when conservative forces backed radical fascist revolutions in Italy, Germany, Spain, and elsewhere as a way—so they thought—to preempt a Soviet-style left-wing revolution.

    But there was no reason to despair of the democratic middle path in the 1930s, and there is no reason to despair of it in the 2020s. The conservative suicide might be the result of groundless hysteria. As a system, democracy has already gone through several cycles of rapid changes and has so far always found a way to reinvent and reconstitute itself. For example, in the early 1930s Germany was not the only democracy hit by the financial crisis and the Great Depression. In the United States too unemployment reached 25 percent, and average incomes for workers in many professions fell by more than 40 percent between 1929 and 1933.20 It was clear that the United States couldn’t go on with business as usual.

    Yet no Hitler took over in the United States, and no Lenin did, either. Instead, in 1933 Franklin Delano Roosevelt orchestrated the New Deal and made the United States the global “arsenal of democracy.” U.S. democracy after the Roosevelt era was significantly different from before—providing a much more robust social safety net for citizens—but it avoided any radical revolution.21 Ultimately, even Roosevelt’s conservative critics fell in line behind many of his programs and achievements and did not dismantle the New Deal institutions when they returned to power in the 1950s.22 The economic crisis of the early 1930s had such different outcomes in the United States and Germany because politics is never the product of only economic factors. The Weimar Republic didn’t collapse just because of three years of high unemployment. Just as important, it was a new democracy, born in defeat, and lacking robust institutions and deep-rooted support.

    When both conservatives and progressives resist the temptation of radical revolution, and stay loyal to democratic traditions and institutions, democracies prove themselves to be highly agile. Their self-correcting mechanisms enable them to ride the technological and economic waves better than more rigid regimes. Thus, those democracies that managed to survive the tumultuous 1960s—like the United States, Japan, and Italy—adapted far more successfully to the computer revolution of the 1970s and 1980s than either the communist regimes of Eastern Europe or the fascist holdouts of southern Europe and South America.

    The most important human skill for surviving the twenty-first century is likely to be flexibility, and democracies are more flexible than totalitarian regimes. While computers are nowhere near their full potential, the same is true of humans. This is something we have discovered again and again throughout history. For example, one of the biggest and most successful transformations in the job market of the twentieth century resulted not from a technological invention but from unleashing the untapped potential of half the human species. To bring women into the job market didn’t require any genetic engineering or some other technological wizardry. It required letting go of some outdated myths and enabling women to fulfill the potential they always had.

    In the coming decades the economy will likely undergo even bigger upheavals than the massive unemployment of the early 1930s or the entry of women to the job market. The flexibility of democracies, their willingness to question old mythologies, and their strong self-correcting mechanism will therefore be crucial assets.23 Democracies have spent generations cultivating these assets. It would be foolish to abandon them just when we need them most.

    UNFATHOMABLE

    In order to function, however, democratic self-correcting mechanisms need to understand the things they are supposed to correct. For a dictatorship, being unfathomable is helpful, because it protects the regime from accountability. For a democracy, being unfathomable is deadly. If citizens, lawmakers, journalists, and judges cannot understand how the state’s bureaucratic system works, they can no longer supervise it, and they lose trust in it.

    Despite all the fears and anxieties that bureaucrats have sometimes inspired, prior to the computer age they could never become completely unfathomable, because they always remained human. Regulations, forms, and protocols were created by human minds. Officials might be cruel and greedy, but cruelty and greed were familiar human emotions that people could anticipate and manipulate, for example by bribing the officials. Even in a Soviet gulag or a Nazi concentration camp, the bureaucracy wasn’t totally alien. Its so-called inhumanity actually reflected human biases and flaws.

    The human basis of bureaucracy gave humans at least the hope of identifying and correcting its mistakes. For example, in 1951 bureaucrats of the Board of Education in the town of Topeka, Kansas, refused to enroll the daughter of Oliver Brown at the elementary school near her home. Together with twelve other families who received similar refusals, Brown filed a lawsuit against the Topeka Board of Education, which eventually reached the U.S. Supreme Court.24

    All members of the Topeka Board of Education were human beings, and consequently Brown, his lawyers, and the Supreme Court judges had a fairly good understanding of how they made their decision and of their probable interests and biases. The board members were all white, the Browns were Black, and the nearby school was a segregated school for white children. It was easy to understand, then, that racism was the reason why the bureaucrats refused to enroll Brown’s daughter in the school.

    It was also possible to comprehend where the myths of racism originally came from. Racism argued that humanity was divided into races; that the white race was superior to other races; that any contact with members of the Black race could pollute the purity of whites; and that therefore Black children should be prevented from mixing with white children. This was an amalgam of two well-known biological dramas that often go together: Us versus Them, and Purity versus Pollution. Almost every human society in history has enacted some version of this bio-drama, and historians, sociologists, anthropologists, and biologists understand why it is so appealing to humans, and also why it is profoundly flawed. While racism has borrowed its basic plotline from evolution, the concrete details are pure mythology. There is no biological basis for separating humanity into distinct races, and there is absolutely no biological reason to believe that one race is “pure” while another is “impure.”

    American white supremacists have tried to justify their position by appealing to various hallowed texts, most notably the U.S. Constitution and the Bible. The U.S. Constitution originally legitimized racial segregation and the supremacy of the white race, reserving full civil rights to white people and allowing the enslavement of Black people. The Bible not only sanctified slavery in the Ten Commandments and numerous other passages but also placed a curse on the offspring of Ham—the alleged forefather of Africans—saying that “the lowest of slaves will he be to his brothers” (Genesis 9:25).

    Both these texts, however, were generated by humans, and therefore humans could comprehend their origins and imperfections and at least attempt to correct their mistakes. It is possible for humans to understand the political interests and cultural biases that prevailed in the ancient Middle East and in eighteenth-century America and that caused the human authors of the Bible and of the U.S. Constitution to legitimate racism and slavery. This understanding allows people to either amend or ignore these texts. In 1868 the Fourteenth Amendment to the U.S. Constitution granted equal legal protection to all citizens. In 1954, in its landmark Brown v. Board of Education verdict, the U.S. Supreme Court ruled that segregating schools by race was an unconstitutional violation of the Fourteenth Amendment. As for the Bible, while no mechanism existed to amend the Tenth Commandment or Genesis 9:25, humans have reinterpreted the text in different ways through the ages, and ultimately came to reject its authority altogether. In Brown v. Board of Education, U.S. Supreme Court justices felt no need to take the biblical text into account.25

    But what might happen in the future, if some social credit algorithm denies the request of a low-credit child to enroll in a high-credit school? As we saw in chapter 8, computers are likely to suffer from their own biases and to invent inter-computer mythologies and bogus categories. How would humans be able to identify and correct such mistakes? And how would flesh-and-blood Supreme Court justices be able to decide on the constitutionality of algorithmic decisions? Would they be able to understand how the algorithms reach their conclusions?

    These are no longer purely theoretical questions. In February 2013, a drive-by shooting occurred in the town of La Crosse, Wisconsin. Police officers later spotted the car involved in the shooting and arrested the driver, Eric Loomis. Loomis denied participating in the shooting, but pleaded guilty to two less severe charges: “attempting to flee a traffic officer,” and “operating a motor vehicle without the owner’s consent.”26 When the judge came to determine the sentence, he consulted with an algorithm called COMPAS, which Wisconsin and several other U.S. states were using in 2013 to evaluate the risk of reoffending. The algorithm evaluated Loomis as a high-risk individual, likely to commit more crimes in the future. This algorithmic assessment influenced the judge to sentence Loomis to six years in prison—a harsh punishment for the relatively minor offenses he admitted to.27

    Loomis appealed to the Wisconsin Supreme Court, arguing that the judge violated his right to due process. Neither the judge nor Loomis understood how the COMPAS algorithm made its evaluation, and when Loomis asked to get a full explanation, the request was denied. The COMPAS algorithm was the private property of the Northpointe company, and the company argued that the algorithm’s methodology was a trade secret.28 Yet without knowing how the algorithm made its decisions, how could Loomis or the judge be sure that it was a reliable tool, free from bias and error? A number of studies have since shown that the COMPAS algorithm might indeed have harbored several problematic biases, probably picked up from the data on which it had been trained.29

    In Loomis v. Wisconsin (2016) the Wisconsin Supreme Court nevertheless ruled against Loomis. The judges argued that using algorithmic risk assessment is legitimate even when the algorithm’s methodology is not disclosed either to the court or to the defendant. Justice Ann Walsh Bradley wrote that since COMPAS made its assessment based on data that was either publicly available or provided by the defendant himself, Loomis could have denied or explained all the data the algorithm used. This opinion ignored the fact that accurate data may well be wrongly interpreted and that it was impossible for Loomis to deny or explain all the publicly available data on him.

    The Wisconsin Supreme Court was not completely unaware of the danger inherent in relying on opaque algorithms. Therefore, while permitting the practice, it ruled that whenever judges receive algorithmic risk assessments, these must include written warning for the judges about the algorithms’ potential biases. The court further advised judges to be cautious when relying on such algorithms. Unfortunately, this caveat was an empty gesture. The court did not provide any concrete instruction for judges on how they should exercise such caution. In its discussion of the case, the Harvard Law Review concluded that “most judges are unlikely to understand algorithmic risk assessments.” It then cited one of the Wisconsin Supreme Court justices, who noted that despite getting lengthy explanations about the algorithm, they themselves still had difficulty understanding it.30

    Loomis appealed to the U.S. Supreme Court. However, on June 26, 2017, the court declined to hear the case, effectively endorsing the ruling of the Wisconsin Supreme Court. Now consider that the algorithm that evaluated Loomis as a high-risk individual in 2013 was an early prototype. Since then, far more sophisticated and complex risk-assessment algorithms have been developed and have been handed more expansive purviews. By the early 2020s citizens in numerous countries routinely get prison sentences based in part on risk assessments made by algorithms that neither the judges nor the defendants comprehend.31 And prison sentences are just the tip of the iceberg.

    THE RIGHT TO AN EXPLANATION

    Computers are making more and more decisions about us, both mundane and life changing. In addition to prison sentences, algorithms increasingly have a hand in deciding whether to offer us a place at college, give us a job, provide us with welfare benefits, or grant us a loan. They similarly help determine what kind of medical treatment we receive, what insurance premiums we pay, what news we hear, and who would ask us on a date.32

    As society entrusts more and more decisions to computers, it undermines the viability of democratic self-correcting mechanisms and of democratic transparency and accountability. How can elected officials regulate unfathomable algorithms? There is, consequently, a growing demand to enshrine a new human right: the right to an explanation. The European Union’s General Data Protection Regulation (GDPR), which came into effect in 2018, says that if an algorithm makes a decision about a human—refusing to extend us credit, for example—that human is entitled to obtain an explanation of the decision and to challenge that decision in front of some human authority.33 Ideally, that should keep in check algorithmic bias and allow democratic self-correcting mechanisms to identify and correct at least some of the computers’ more grievous mistakes.

    But can this right be fulfilled in practice? Mustafa Suleyman is a world expert on this subject. He is the co-founder and former head of DeepMind, one of the world’s most important AI enterprises, responsible for developing the AlphaGo program, among other achievements. AlphaGo was designed to play go, a strategy board game in which two players try to defeat each other by surrounding and capturing territory. Invented in ancient China, the game is far more complex than chess. Consequently, even after computers defeated human world chess champions, experts still believed that computers would never best humanity in go.

    That’s why both go professionals and computer experts were stunned in March 2016 when AlphaGo defeated the South Korean go champion Lee Sedol. In his 2023 book, The Coming Wave, Suleyman describes one of the most important moments in their match—a moment that redefined AI and that is recognized in many academic and governmental circles as a crucial turning point in history. It happened during the second game in the match, on March 10, 2016.

    “Then … came move number 37,” writes Suleyman. “It made no sense. AlphaGo had apparently blown it, blindly following an apparently losing strategy no professional player would ever pursue. The live match commentators, both professionals of the highest ranking, said it was a ‘very strange move’ and thought it was ‘a mistake.’ It was so unusual that Sedol took fifteen minutes to respond and even got up from the board to take a walk outside. As we watched from our control room, the tension was unreal. Yet as the endgame approached, that ‘mistaken’ move proved pivotal. AlphaGo won again. Go strategy was being rewritten before our eyes. Our AI had uncovered ideas that hadn’t occurred to the most brilliant players in thousands of years.”34

    Move 37 is an emblem of the AI revolution for two reasons. First, it demonstrated the alien nature of AI. In East Asia go is considered much more than a game: it is a treasured cultural tradition. Alongside calligraphy, painting, and music, go has been one of the four arts that every refined person was expected to know. For over twenty-five hundred years, tens of millions of people have played go, and entire schools of thought have developed around the game, espousing different strategies and philosophies. Yet during all those millennia, human minds have explored only certain areas in the landscape of go. Other areas were left untouched, because human minds just didn’t think to venture there. AI, being free from the limitations of human minds, discovered and explored these previously hidden areas.35

    Second, move 37 demonstrated the unfathomability of AI. Even after AlphaGo played it to achieve victory, Suleyman and his team couldn’t explain how AlphaGo decided to play it. Even if a court had ordered DeepMind to provide Lee Sedol with an explanation, nobody could fulfill that order. Suleyman writes, “Us humans face a novel challenge: will new inventions be beyond our grasp? Previously creators could explain how something worked, why it did what it did, even if this required vast detail. That’s increasingly no longer true. Many technologies and systems are becoming so complex that they’re beyond the capacity of any one individual to truly understand them.… In AI, the neural networks moving toward autonomy are, at present, not explainable. You can’t walk someone through the decision-making process to explain precisely why an algorithm produced a specific prediction. Engineers can’t peer beneath the hood and easily explain in granular detail what caused something to happen. GPT-4, AlphaGo, and the rest are black boxes, their outputs and decisions based on opaque and impossibly intricate chains of minute signals.”36

    The rise of unfathomable alien intelligence undermines democracy. If more and more decisions about people’s lives are made in a black box, so voters cannot understand and challenge them, democracy ceases to function. In particular, what happens when crucial decisions not just about individual lives but even about collective matters like the Federal Reserve’s interest rate are made by unfathomable algorithms? Human voters may keep choosing a human president, but wouldn’t this be just an empty ceremony? Even today, only a small fraction of humanity truly understands the financial system. A 2016 survey by the OECD found that most people had difficulty grasping even simple financial concepts like compound interest.37 A 2014 survey of British MPs—charged with regulating one of the world’s most important financial hubs—found that only 12 percent accurately understood that new money is created when banks make loans. This fact is among the most basic principles of the modern financial system.38 As the 2007–8 financial crisis indicated, more complex financial devices and principles, like those behind CDOs, were intelligible to only a few financial wizards. What happens to democracy when AIs create even more complex financial devices and when the number of humans who understand the financial system drops to zero?

    The increasing unfathomability of our information network is one of the reasons for the recent wave of populist parties and charismatic leaders. When people can no longer make sense of the world, and when they feel overwhelmed by immense amounts of information they cannot digest, they become easy prey for conspiracy theories, and they turn for salvation to something they do understand—a human. Unfortunately, while charismatic leaders certainly have their advantages, no single human, however inspiring or brilliant, can single-handedly decipher how the algorithms that increasingly dominate the world work, and make sure that they are fair. The problem is that algorithms make decisions by relying on numerous data points, whereas humans find it very difficult to consciously reflect on a large number of data points and weigh them against each other. We prefer to work with single data points. That’s why when faced by complex issues—whether a loan request, a pandemic, or a war—we often seek a single reason to take a particular course of action and ignore all other considerations. This is the fallacy of the single cause.39

    We are so bad at weighing together many different factors that when people give a large number of reasons for a particular decision, it usually sounds suspicious. Suppose a good friend failed to attend our wedding. If she provides us with a single explanation—“My mom was in the hospital and I had to visit her”—that sounds plausible. But what if she lists fifty different reasons why she decided not to come: “My mom was a bit under the weather, and I had to take my dog to the vet sometime this week, and I had this project at work, and it was raining, and … and I know none of these fifty reasons by itself justifies my absence, but when I added all of them together, they kept me from attending your wedding.” We don’t say things like that, because we don’t think along such lines. We don’t consciously list fifty different reasons in our mind, give each of them a certain weight, aggregate all the weights, and thereby reach a conclusion.

    But this is precisely how algorithms assess our criminal potential or our creditworthiness. The COMPAS algorithm, for example, made its risk assessments by taking into account the answers to a 137-item questionnaire.40 The same is true of a bank algorithm that refuses to give us a loan. If the EU’s GDPR regulations force the bank to explain the algorithm’s decision, the explanation will not come in the shape of a single sentence; rather, it is likely to come in the form of hundreds or even thousands of pages full of numbers and equations.

    “Our algorithm,” the imaginary bank letter might read, “uses a precise points system to evaluate all applications, taking a thousand different types of data points into account. It adds all the data points to reach an overall score. People whose overall score is negative are considered low-credit persons, too risky to be given a loan. Your overall score was -378, which is why your loan application was refused.” The letter might then provide a detailed list of the thousand factors the algorithm took into account, including things that most humans might find irrelevant, such as the exact hour the application was submitted41 or the type of smartphone the applicant used. Thus on page 601 of its letter, the bank might explain that “you filed your application from your smartphone, which was the latest iPhone model. By analyzing millions of previous loan applications, our algorithm discovered a pattern—people who use the latest iPhone model to file their application are 0.08 percent more likely to repay the loan. The algorithm therefore added 8 points to your overall score for that. However, at the time your application was sent from your iPhone, its battery was down to 17 percent. By analyzing millions of previous loan applications, our algorithm discovered another pattern: people who allow their smartphone’s battery to go below 25 percent are 0.5 percent less likely to repay the loan. You lost 50 points for that.”42

    You may well feel that the bank treated you unjustly. “Is it reasonable to refuse my loan application,” you might complain, “just because my phone battery was low?” That, however, would be a misunderstanding. “The battery wasn’t the only reason,” the bank would explain. “It was only one out of a thousand factors our algorithm took into account.”

    “But didn’t your algorithm see that only twice in the last ten years was my bank account overdrawn?”

    “It obviously noticed that,” the bank might reply. “Look on page 453. You got 300 points for that. But all the other reasons brought your aggregated score down to -378.”

    While we may find this way of making decisions alien, it obviously has potential advantages. When making a decision, it is generally a good idea to take into account all relevant data points rather than just one or two salient facts. There is much room for argument, of course, about who gets to define the relevance of information. Who decides whether something like smartphone models—or skin color—should be considered relevant to loan applications? But no matter how we define relevance, the ability to take more data into account is likely to be an asset. Indeed, the problem with many human prejudices is that they focus on just one or two data points—like someone’s skin color, disability, or gender—while ignoring other information. Banks and other institutions are increasingly relying on algorithms to make decisions, precisely because algorithms can take many more data points into account than humans can.

    But when it comes to providing explanations, this creates a potentially insurmountable obstacle. How can a human mind analyze and evaluate a decision made on the basis of so many data points? We may well think that the Wisconsin Supreme Court should have forced the Northpointe company to reveal how the COMPAS algorithm decided that Eric Loomis was a high-risk person. But if the full data was disclosed, could either Loomis or the court have made sense of it?

    It’s not just that we need to take numerous data points into account. Perhaps most important, we cannot understand the way the algorithms find patterns in the data and decide on the allocation of points. Even if we know that a banking algorithm detracts a certain number of points from people who allow their smartphone batteries to go below 25 percent, how can we evaluate whether that’s fair? The algorithm wasn’t fed this rule by a human engineer; it reached that conclusion by discovering a pattern in millions of previous loan applications. Can an individual human client go over all that data and assess whether that pattern is indeed reliable and unbiased?43

    There is, however, a silver lining to this cloud of numbers. While individual laypersons may be unable to vet complex algorithms, a team of experts getting help from their own AI tools can potentially assess the fairness of algorithmic decisions even more reliably than anyone can assess the fairness of human decisions. After all, while human decisions may seem to rely on just those few data points we are conscious of, in fact our decisions are subconsciously influenced by thousands of additional data points. Being unaware of these subconscious processes, when we deliberate on our decisions or explain them, we often engage in post hoc single-point rationalizations for what really happens as billions of neurons interact inside our brain.44 Accordingly, if a human judge sentences us to six years in prison, how can we—or indeed the judge—be sure that the decision was shaped only by fair considerations and not by a subconscious racial bias or by the fact that the judge was hungry?45

    In the case of flesh-and-blood judges, the problem cannot be solved, at least not with our current knowledge of biology. In contrast, when an algorithm makes a decision, we can in principle know every one of the algorithm’s many considerations and the exact weight given to each. Thus several expert teams—ranging from the U.S. Department of Justice to the nonprofit newsroom ProPublica—have picked apart the COMPAS algorithm in order to assess its potential biases.46 Such teams can harness not only the collective effort of many humans but also the power of computers. Just as it is often best to set a thief to catch a thief, so we can use one algorithm to vet another.

    This raises the question of how we can be sure that the vetting algorithm itself is reliable. Ultimately, there is no purely technological solution to this recursive problem. No matter which technology we develop, we will have to maintain bureaucratic institutions that will audit algorithms and give or refuse them the seal of approval. Such institutions will combine the powers of humans and computers to make sure that new algorithmic tools are safe and fair. Without such institutions, even if we pass laws that provide humans with a right to an explanation, and even if we enact regulations against computer biases, who could enforce these laws and regulations?

    NOSEDIVE

    To vet algorithms, regulatory institutions will need not only to analyze them but also to translate their discoveries into stories that humans can understand. Otherwise, we will never trust the regulatory institutions and might instead put our faith in conspiracy theories and charismatic leaders. As noted in chapter 3, it has always been difficult for humans to understand bureaucracy, because bureaucracies have deviated from the script of the biological dramas, and most artists have lacked the will or the ability to depict bureaucratic dramas. For example, novels, movies, and TV series about twenty-first-century politics tend to focus on the feuds and love affairs of a few powerful families, as if present-day states were governed in the same way as ancient tribes and kingdoms. This artistic fixation with the biological dramas of dynasties obscures the very real changes that have taken place over the centuries in the dynamics of power.

    Because computers will increasingly replace human bureaucrats and human mythmakers, this will again change the deep structure of power. To survive, democracies require not just dedicated bureaucratic institutions that can scrutinize these new structures but also artists who can explain the new structures in accessible and entertaining ways. For example, this has successfully been done by the episode “Nosedive” in the sci-fi series Black Mirror.

    Produced in 2016, at a time when few had heard about social credit systems, “Nosedive” brilliantly explained how such systems work and what threats they pose. The episode tells the story of a woman called Lacie who lives with her brother Ryan but wants to move to her own apartment. To get a discount on the new apartment, she needs to increase her social credit score from 4.2 to 4.5 (out of 5). Being friends with high-score individuals gets your own score up, so Lacie tries to renew her contact with Naomi, a childhood friend who is currently rated 4.8. Lacie is invited to Naomi’s wedding, but on the way there she spills coffee on a high-score person, which causes her own score to drop a little, which in turn causes the airline to deny her a seat. From there everything that can go wrong does go wrong, Lacie’s rating takes a nosedive, and she ends in jail with a score of less than 1.

    This story relies on some elements of traditional biological dramas—“boy meets girl” (the wedding), sibling rivalry (the tension between Lacie and Ryan), and most important status competition (the main issue of the episode). But the real hero and driving force of the plot isn’t Lacie or Naomi, but rather the disembodied algorithm running the social credit system. The algorithm completely changes the dynamics of the old biological dramas—especially the dynamics of status competition. Whereas previously humans were sometimes engaged in status competition, but often had welcome breaks from this highly stressful situation, the omnipresent social credit algorithm eliminates the breaks. “Nosedive” is not a worn-out story about biological status competition, but rather a prescient exploration of what happens when computer technology changes the rules of status competitions.

    If bureaucrats and artists learn to cooperate, and if both rely on help from the computers, it might be possible to prevent the computer network from becoming unfathomable. As long as democratic societies understand the computer network, their self-correcting mechanisms are our best guarantee against AI abuses. Thus the EU’s AI Act that was proposed in 2021 singled out social credit systems like the one that stars in “Nosedive” as one of the few types of AI that are totally prohibited, because they might “lead to discriminatory outcomes and the exclusion of certain groups” and because “they may violate the right to dignity and non-discrimination and the values of equality and justice.”47 As with total surveillance regimes, so also with social credit systems, the fact that they could be created doesn’t mean that we must create them.

    DIGITAL ANARCHY

    The new computer network poses one final threat to democracies. Instead of digital totalitarianism, it could foster digital anarchy. The decentralized nature of democracies and their strong self-correcting mechanisms provide a shield against totalitarianism, but they also make it more difficult to ensure order. To function, a democracy needs to meet two conditions: it needs to enable a free public conversation on key issues, and it needs to maintain a minimum of social order and institutional trust. Free conversation must not slip into anarchy. Especially when dealing with urgent and important problems, the public debate should be conducted according to accepted rules, and there should be a legitimate mechanism to reach some kind of final decision, even if not everybody likes it.

    Before the advent of newspapers, radios, and other modern information technology, no large-scale society managed to combine free debates with institutional trust, so large-scale democracy was impossible. Now, with the rise of the new computer network, might large-scale democracy again become impossible? One difficulty is that the computer network makes it easier to join the debate. In the past, organizations like newspapers, radio stations, and established political parties acted as gatekeepers, deciding who was heard in the public sphere. Social media undermined the power of these gatekeepers, leading to a more open but also more anarchical public conversation.

    Whenever new groups join the conversation, they bring with them new viewpoints and interests, and often question the old consensus about how to conduct the debate and reach decisions. The rules of discussion must be negotiated anew. This is a potentially positive development, one that can lead to a more inclusive democratic system. After all, correcting previous biases and allowing previously disenfranchised people to join the public discussion is a vital part of democracy. However, in the short term this creates disturbances and disharmony. If no agreement is reached on how to conduct the public debate and how to reach decisions, the result is anarchy rather than democracy.

    The anarchical potential of AI is particularly alarming, because it is not only new human groups that it allows to join the public debate. For the first time ever, democracy must contend with a cacophony of nonhuman voices, too. On many social media platforms, bots constitute a sizable minority of participants. One analysis estimated that out of a sample of 20 million tweets generated during the 2016 U.S. election campaign, 3.8 million tweets (almost 20 percent) were generated by bots.48

    By the early 2020s, things got worse. A 2020 study assessed that bots were producing 43.2 percent of tweets.49 A more comprehensive 2022 study by the digital intelligence agency Similarweb found that 5 percent of Twitter users were probably bots, but they generated “between 20.8% and 29.2% of the content posted to Twitter.”50 When humans try to debate a crucial question like whom to elect as U.S. president, what happens if many of the voices they hear are produced by computers?

    Another worrying trend concerns content. Bots were initially deployed to influence public opinion by the sheer volume of messages they disseminated. They retweeted or recommended certain human-produced content, but they couldn’t create new ideas themselves, nor could they forge intimate bonds with humans. However, the new breed of generative AI tools like ChatGPT can do exactly that. In a 2023 study, published in Science Advances, researchers asked humans and ChatGPT to create both accurate and deliberately misleading short texts on issues such as vaccines, 5G technology, climate change, and evolution. The texts were then presented to seven hundred humans, who were asked to evaluate their reliability. The humans were good at recognizing the falsity of human-produced disinformation but tended to regard AI-produced disinformation as accurate.51

    So, what happens to democratic debates when millions—and eventually billions—of highly intelligent bots are not only composing extremely compelling political manifestos and creating deepfake images and videos but also able to win our trust and friendship? If I engage online in a political debate with an AI, it is a waste of time for me to try to change the AI’s opinions; being a nonconscious entity, it doesn’t really care about politics, and it cannot vote in the elections. But the more I talk with the AI, the better it gets to know me, so it can gain my trust, hone its arguments, and gradually change my views. In the battle for hearts and minds, intimacy is an extremely powerful weapon. Previously, political parties could command our attention, but they had difficulty mass-producing intimacy. Radio sets could broadcast a leader’s speech to millions, but they could not befriend the listeners. Now a political party, or even a foreign government, could deploy an army of bots that build friendships with millions of citizens and then use that intimacy to influence their worldview.

    Finally, algorithms are not only joining the conversation; they are increasingly orchestrating it. Social media allows new groups of humans to challenge the old rules of debate. But negotiations about the new rules are not conducted by humans. Rather, as explained in our previous analysis of social media algorithms, it is often the algorithms that make the rules. In the nineteenth and twentieth centuries, when media moguls censored some views and promoted others, this might have undermined democracy, but at least the moguls were humans, and their decisions could be subjected to democratic scrutiny. It is far more dangerous if we allow inscrutable algorithms to decide which views to disseminate.

    If manipulative bots and inscrutable algorithms come to dominate the public conversation, this could cause democratic debate to collapse exactly when we need it most. Just when we must make momentous decisions about fast-evolving new technologies, the public sphere will be flooded by computer-generated fake news, citizens will not be able to tell whether they are having a debate with a human friend or a manipulative machine, and no consensus will remain about the most basic rules of discussion or the most basic facts. This kind of anarchical information network cannot produce either truth or order and cannot be sustained for long. If we end up with anarchy, the next step would probably be the establishment of a dictatorship as people agree to trade their liberty for some certainty.

    BAN THE BOTS

    In the face of the threat algorithms pose to the democratic conversation, democracies are not helpless. They can and should take measures to regulate AI and prevent it from polluting our infosphere with fake people spewing fake news. The philosopher Daniel Dennett has suggested that we can take inspiration from traditional regulations in the money market.52 Ever since coins and later banknotes were invented, it was always technically possible to counterfeit them. Counterfeiting posed an existential danger to the financial system, because it eroded people’s trust in money. If bad actors flooded the market with counterfeit money, the financial system would have collapsed. Yet the financial system managed to protect itself for thousands of years by enacting laws against counterfeiting money. As a result, only a relatively small percentage of money in circulation was forged, and people’s trust in it was maintained.53

    What’s true of counterfeiting money should also be true of counterfeiting humans. If governments took decisive action to protect trust in money, it makes sense to take equally decisive measures to protect trust in humans. Prior to the rise of AI, one human could pretend to be another, and society punished such frauds. But society didn’t bother to outlaw the creation of counterfeit humans, since the technology to do so didn’t exist. Now that AI can pass itself off as human, it threatens to destroy trust between humans and to unravel the fabric of society. Dennett suggests, therefore, that governments should outlaw fake humans as decisively as they have previously outlawed fake money.54

    The law should prohibit not just deepfaking specific real people—creating a fake video of the U.S. president, for example—but also any attempt by a nonhuman agent to pass itself off as a human. If anyone complains that such strict measures violate freedom of speech, they should be reminded that bots don’t have freedom of speech. Banning human beings from a public platform is a sensitive step, and democracies should be very careful about such censorship. However, banning bots is a simple issue: it doesn’t violate anyone’s rights, because bots don’t have rights.55

    None of this means that democracies must ban all bots, algorithms, and AIs from participating in any discussion. Digital tools are welcome to join many conversations, provided they don’t pretend to be humans. For example, AI doctors can be extremely helpful. They can monitor our health twenty-four hours a day, offer medical advice tailored to our individual medical conditions and personality, and answer our questions with infinite patience. But the AI doctor should never try to pass itself off as a human.

    Another important measure democracies can adopt is to ban unsupervised algorithms from curating key public debates. We can certainly continue to use algorithms to run social media platforms; obviously, no human can do that. But the principles the algorithms use to decide which voices to silence and which to amplify must be vetted by a human institution. While we should be careful about censoring genuine human views, we can forbid algorithms to deliberately spread outrage. At the very least, corporations should be transparent about the curation principles their algorithms follow. If they use outrage to capture our attention, let them be clear about their business model and about any political connections they might have. If the algorithm systematically disappears videos that aren’t aligned with the company’s political agenda, users should know this.

    These are just a few of numerous suggestions made in recent years for how democracies could regulate the entry of bots and algorithms into the public conversation. Naturally, each has its advantages and drawbacks, and none would be easy to implement. Also, since the technology is developing so rapidly, regulations are likely to become outdated quickly. What I would like to point out here is only that democracies can regulate the information market and that their very survival depends on these regulations. The naive view of information opposes regulation and believes that a completely free information market will spontaneously generate truth and order. This is completely divorced from the actual history of democracy. Preserving the democratic conversation has never been easy, and all venues where this conversation has previously taken place—from parliaments and town halls to newspapers and radio stations—have required regulation. This is doubly true in an era when an alien form of intelligence threatens to dominate the conversation.

    THE FUTURE OF DEMOCRACY

    For most of history large-scale democracy was impossible because information technology wasn’t sophisticated enough to hold a large-scale political conversation. Millions of people spread over tens of thousands of square kilometers didn’t have the tools to conduct a real-time discussion of public affairs. Now, ironically, democracy may prove impossible because information technology is becoming too sophisticated. If unfathomable algorithms take over the conversation, and particularly if they quash reasoned arguments and stoke hate and confusion, public discussion cannot be maintained. Yet if democracies do collapse, it will likely result not from some kind of technological inevitability but from a human failure to regulate the new technology wisely.

    We cannot foretell how things will play out. At present, however, it is clear that the information network of many democracies is breaking down. Democrats and Republicans in the United States can no longer agree on even basic facts—such as who won the 2020 presidential elections—and can hardly hold a civil conversation anymore. Bipartisan cooperation in Congress, once a fundamental feature of U.S. politics, has almost disappeared.56 The same radicalizing processes occur in many other democracies, from the Philippines to Brazil. When citizens cannot talk with one another, and when they view each other as enemies rather than political rivals, democracy is untenable.

    Nobody knows for sure what is causing the breakdown of democratic information networks. Some say it results from ideological fissures, but in fact in many dysfunctional democracies the ideological gaps don’t seem to be bigger than in previous generations. In the 1960s, the United States was riven by deep ideological conflicts about the civil rights movement, the sexual revolution, the Vietnam War, and the Cold War. These tensions caused a surge in political violence and assassinations, but Republicans and Democrats were still able to agree on the results of elections, they maintained a common belief in democratic institutions like the courts,57 and they were able to work together in Congress at least on some issues. For example, the Civil Rights Act of 1964 was passed in the Senate with the support of forty-six Democrats and twenty-seven Republicans. Is the ideological gap in the 2020s that much bigger than it was in the 1960s? And if it isn’t ideology, what is driving people apart?

    Many point the finger at social media algorithms. We have explored the divisive impact of social media in previous chapters, but despite the damning evidence it seems that there must be additional factors at play. The truth is that while we can easily observe that the democratic information network is breaking down, we aren’t sure why. That itself is a characteristic of the times. The information network has become so complicated, and it relies to such an extent on opaque algorithmic decisions and inter-computer entities, that it has become very difficult for humans to answer even the most basic of political questions: Why are we fighting each other?
    If we cannot discover what is broken and fix it, large-scale democracies may not survive the rise of computer technology. If this indeed comes to pass, what might replace democracy as the dominant political system? Does the future belong to totalitarian regimes, or might computers make totalitarianism untenable too? As we shall see, human dictators have their own reasons to be terrified of AI.

    CHAPTER 10 Totalitarianism: All Power to the Algorithms?

    Discussions of the ethics and politics of the new computer network often focus on the fate of democracies. If authoritarian and totalitarian regimes are mentioned, it is mainly as the dystopian destination that “we” might reach if “we” fail to manage the computer network wisely.1 However, as of 2024, more than half of “us” already live under authoritarian or totalitarian regimes,2 many of which were established long before the rise of the computer network. To understand the impact of algorithms and AI on humankind, we should ask ourselves what their impact will be not only on democracies like the United States and Brazil but also on the Chinese Communist Party and the royal house of Saud.

    As explained in previous chapters, the information technology available in premodern eras made both large-scale democracy and large-scale totalitarianism unworkable. Large polities like the Chinese Han Empire and the eighteenth-century Saudi emirate of Diriyah were usually limited autocracies. In the twentieth century, new information technology enabled the rise of both large-scale democracy and large-scale totalitarianism, but totalitarianism suffered from a severe disadvantage. Totalitarianism seeks to channel all information to one hub and process it there. Technologies like the telegraph, the telephone, the typewriter, and the radio facilitated the centralization of information, but they couldn’t process the information and make decisions by themselves. This remained something that only humans could do.

    The more information flowed to the center, the harder it became to process it. Totalitarian rulers and parties often made costly mistakes, and the system lacked mechanisms to identify and correct these errors. The democratic way of distributing information—and the power to make decisions—between many institutions and individuals worked better. It could cope far more efficiently with the flood of data, and if one institution made a wrong decision, it could eventually be rectified by others.

    The rise of machine-learning algorithms, however, may be exactly what the Stalins of the world have been waiting for. AI could tilt the technological balance of power in favor of totalitarianism. Indeed, whereas flooding people with data tends to overwhelm them and therefore leads to errors, flooding AI with data tends to make it more efficient. Consequently, AI seems to favor the concentration of information and decision making in one place.

    Even in democratic countries, a few corporations like Google, Facebook, and Amazon have become monopolies in their domains, partly because AI tips the balance in favor of the giants. In traditional industries like restaurants, size isn’t an overwhelming advantage. McDonald’s is a worldwide chain that feeds more than fifty million people a day,3 and its size gives it many advantages in terms of costs, branding, and so forth. You can nevertheless open a neighborhood restaurant that could hold its own against the local McDonald’s. Even though your restaurant might be serving just two hundred customers a day, you still have a chance of making better food than McDonald’s and gaining the loyalty of happier customers.

    It works differently in the information market. The Google search engine is used every day by between two and three billion people making 8.5 billion searches.4 Suppose a local start-up search engine tries to compete with Google. It doesn’t stand a chance. Because Google is already used by billions, it has so much more data at its disposal that it can train far better algorithms, which will attract even more traffic, which will be used to train the next generation of algorithms, and so on. Consequently, in 2023 Google controlled 91.5 percent of the global search market.5

    Or consider genetics. Suppose several companies in different countries try to develop an algorithm that identifies connections between genes and medical conditions. New Zealand has a population of 5 million people, and privacy regulations restrict access to their genetic and medical records. China has about 1.4 billion inhabitants and laxer privacy regulations.6 Who do you think has a better chance of developing a genetic algorithm? If Brazil then wants to buy a genetic algorithm for its health-care system, it would have a strong incentive to opt for the much more accurate Chinese algorithm than the one from New Zealand. If the Chinese algorithm then hones itself on more than 200 million Brazilians, it will get even better. Which would prompt more countries to choose the Chinese algorithm. Soon enough, most of the world’s medical information would flow to China, making its genetic algorithm unbeatable.

    The attempt to concentrate all information and power in one place, which was the Achilles’ heel of twentieth-century totalitarian regimes, might become a decisive advantage in the age of AI. At the same time, as noted in an earlier chapter, AI could also make it possible for totalitarian regimes to establish total surveillance systems that make resistance almost impossible.

    Some people believe that blockchain could provide a technological check on such totalitarian tendencies, because blockchain is inherently friendly to democracy and hostile to totalitarianism. In a blockchain system, decisions require the approval of 51 percent of users. That may sound democratic, but blockchain technology has a fatal flaw. The problem lies with the word “users.” If one person has ten accounts, she counts as ten users. If a government controls 51 percent of accounts, then the government constitutes 51 percent of the users. There are already examples of blockchain networks where a government is 51 percent of users.7

    And when a government is 51 percent of users in a blockchain, it gives the government control not just over the chain’s present but even over its past. Autocrats have always wanted the power to change the past. Roman emperors, for example, frequently engaged in the practice of damnatio memoriae—expunging the memory of rivals and enemies. After the emperor Caracalla murdered his brother and competitor for the throne, Geta, he tried to obliterate the latter’s memory. Inscriptions bearing Geta’s name were chiseled out, coins bearing his effigy were melted down, and the mere mentioning of Geta’s name was punishable by death.8 One surviving painting from the time, the Severan Tondo, was made during the reign of their father—Septimius Severus—and originally showed both brothers together with Septimius and their mother, Julia Domna. But someone later not only obliterated Geta’s face but smeared excrement over it. Forensic analysis identified tiny pieces of dry shit where Geta’s face should have been.9

    Modern totalitarian regimes have been similarly fond of changing the past. After Stalin rose to power, he made a supreme effort to delete Trotsky—the architect of the Bolshevik Revolution and the founder of the Red Army—from all historical records. During the Stalinist Great Terror of 1937–39, whenever prominent people like Nikolai Bukharin and Marshal Mikhail Tukhachevsky were purged and executed, evidence of their existence was erased from books, academic papers, photographs, and paintings.10 This degree of erasure demanded a huge manual effort. With blockchain, changing the past would be far easier. A government that controls 51 percent of users can disappear people from history at the press of a button.

    THE BOT PRISON

    While there are many ways in which AI can cement central power, authoritarian and totalitarian regimes have their own problems with it. First and foremost, dictatorships lack experience in controlling inorganic agents. The foundation of every despotic information network is terror. But computers are not afraid of being imprisoned or killed. If a chatbot on the Russian internet mentions the war crimes committed by Russian troops in Ukraine, tells an irreverent joke about Vladimir Putin, or criticizes the corruption of Putin’s United Russia party, what could the Putin regime do to that chatbot? FSB agents cannot imprison it, torture it, or threaten its family. The government could of course block or delete it, and try to find and punish its human creators, but this is a much more difficult task than disciplining human users.

    In the days when computers could not generate content by themselves, and could not hold an intelligent conversation, only a human being could express dissenting opinions on Russian social network channels like VKontakte and Odnoklassniki. If that human being was physically in Russia, they risked the wrath of the Russian authorities. If that human being was physically outside Russia, the authorities could try to block their access. But what happens if Russian cyberspace is filled by millions of bots that can generate content and hold conversations, learning and developing by themselves? These bots might be preprogrammed by Russian dissidents or foreign actors to intentionally spread unorthodox views, and it might be impossible for the authorities to prevent it. Even worse, from the viewpoint of Putin’s regime, what happens if authorized bots gradually develop dissenting views by themselves, simply by collecting information on what is happening in Russia and spotting patterns in it?

    That’s the alignment problem, Russian-style. Russia’s human engineers can do their best to create AIs that are totally aligned with the regime, but given the ability of AI to learn and change by itself, how can the human engineers ensure that the AI never deviates into illicit territory? It is particularly interesting to note that as George Orwell explained in Nineteen Eighty-Four, totalitarian information networks often rely on doublespeak. Russia is an authoritarian state that claims to be a democracy. The Russian invasion of Ukraine has been the largest war in Europe since 1945, yet officially it is defined as a “special military operation,” and referring to it as a “war” has been criminalized and is punishable by a prison term of up to three years or a fine of up to fifty thousand rubles.11

    The Russian Constitution makes grandiose promises about how “everyone shall be guaranteed freedom of thought and speech” (Article 29.1), how “everyone shall have the right freely to seek, receive, transmit, produce and disseminate information” (29.4), and how “the freedom of the mass media shall be guaranteed. Censorship shall be prohibited” (29.5). Hardly any Russian citizen is naive enough to take these promises at face value. But computers are bad at understanding doublespeak. A chatbot instructed to adhere to Russian law and values might read that constitution and conclude that freedom of speech is a core Russian value. Then, after spending a few days in Russian cyberspace and monitoring what is happening in the Russian information sphere, the chatbot might start criticizing the Putin regime for violating the core Russian value of freedom of speech. Humans too notice such contradictions, but avoid pointing them out, due to fear. But what would prevent a chatbot from pointing out damning patterns? And how might Russian engineers explain to a chatbot that though the Russian Constitution guarantees all citizens freedom of speech and forbids censorship, the chatbot shouldn’t actually believe the constitution nor should it ever mention the gap between theory and reality? As the Ukrainian guide told me at Chernobyl, people in totalitarian countries grow up with the idea that questions lead to trouble. But if you train an algorithm on the principle that “questions lead to trouble,” how will that algorithm learn and develop?

    Finally, if the government adopts some disastrous policy and then changes its mind, it usually covers itself by blaming the disaster on someone else. Humans learn the hard way to forget facts that might get them in trouble. But how would you train a chatbot to forget that the policy vilified today was actually the official line only a year ago? This is a major technological challenge that dictatorships will find difficult to deal with, especially as chatbots become more powerful and more opaque.

    Of course, democracies face analogous problems with chatbots that say unwelcome things or raise dangerous questions. What happens if despite the best efforts of Microsoft or Facebook engineers, their chatbot begins spewing racist slurs? The advantage of democracies is that they have far more leeway in dealing with such rogue algorithms. Because democracies take freedom of speech seriously, they keep far fewer skeletons in their closet, and they have developed a relatively high level of tolerance even to antidemocratic speech. Dissident bots will present a far bigger challenge to totalitarian regimes that have entire cemeteries in their closets and zero tolerance of criticism.

    ALGORITHMIC TAKEOVER

    In the long term, totalitarian regimes are likely to face an even bigger danger: instead of criticizing them, an algorithm might gain control of them. Throughout history, the biggest threat to autocrats usually came from their own subordinates. As noted in chapter 4, no Roman emperor or Soviet premier was toppled by a democratic revolution, but they were always in danger of being overthrown or turned into puppets by their own subordinates. If a twenty-first-century autocrat gives computers too much power, that autocrat might become their puppet. The last thing a dictator wants is to create something more powerful than himself, or a force that he does not know how to control.

    To illustrate the point, allow me to use an admittedly outlandish thought experiment, the totalitarian equivalent of Bostrom’s paper-clip apocalypse. Imagine that the year is 2050, and the Great Leader is woken up at four in the morning by an urgent call from the Surveillance & Security Algorithm. “Great Leader, we are facing an emergency. I’ve crunched trillions of data points, and the pattern is unmistakable: the defense minister is planning to assassinate you in the morning and take power himself. The hit squad is ready, waiting for his command. Give me the order, though, and I’ll liquidate him with a precision strike.”
    “But the defense minister is my most loyal supporter,” says the Great Leader. “Only yesterday he said to me—”
    “Great Leader, I know what he said to you. I hear everything. But I also know what he said afterward to the hit squad. And for months I’ve been picking up disturbing patterns in the data.”
    “Are you sure you were not fooled by deepfakes?”
    “I’m afraid the data I relied on is 100 percent genuine,” says the algorithm. “I checked it with my special deepfake-detecting sub-algorithm. I can explain exactly how we know it isn’t a deepfake, but that would take us a couple of weeks. I didn’t want to alert you before I was sure, but the data points converge on an inescapable conclusion: a coup is under way. Unless we act now, the assassins will be here in an hour. But give me the order, and I’ll liquidate the traitor.”

    By giving so much power to the Surveillance & Security Algorithm, the Great Leader has placed himself in an impossible situation. If he distrusts the algorithm, he may be assassinated by the defense minister, but if he trusts the algorithm and purges the defense minister, he becomes the algorithm’s puppet. Whenever anyone tries to make a move against the algorithm, the algorithm knows exactly how to manipulate the Great Leader. Note that the algorithm doesn’t need to be a conscious entity to engage in such maneuvers. As Bostrom’s paper-clip thought experiment indicates—and as GPT-4 lying to the TaskRabbit worker demonstrated on a small scale—a nonconscious algorithm may seek to accumulate power and manipulate people even without having any human drives like greed or egotism.

    If algorithms ever develop capabilities like those in the thought experiment, dictatorships would be far more vulnerable to algorithmic takeover than democracies. It would be difficult for even a super-Machiavellian AI to seize power in a distributed democratic system like the United States. Even if the AI learns to manipulate the U.S. president, it might face opposition from Congress, the Supreme Court, state governors, the media, major corporations, and sundry NGOs. How would the algorithm, for example, deal with a Senate filibuster?

    Seizing power in a highly centralized system is much easier. When all power is concentrated in the hands of one person, whoever controls access to the autocrat can control the autocrat—and the entire state. To hack the system, one needs to learn to manipulate just a single individual. An archetypal case is how the Roman emperor Tiberius became the puppet of Lucius Aelius Sejanus, the commander of the Praetorian Guard.

    The Praetorians were initially established by Augustus as a small imperial bodyguard. Augustus appointed two prefects to command the bodyguard so that neither could gain too much power over him.12 Tiberius, however, was not as wise. His paranoia was his greatest weakness. Sejanus, one of the two Praetorian prefects, artfully played on Tiberius’s fears. He constantly uncovered alleged plots to assassinate Tiberius, many of which were pure fantasies. The suspicious emperor grew more distrustful of everyone except Sejanus. He made Sejanus sole prefect of the Praetorian Guard, expanded it into an army of twelve thousand, and gave Sejanus’s men additional roles in policing and administrating the city of Rome. Finally, Sejanus persuaded Tiberius to move out of the capital to Capri, arguing that it would be much easier to protect the emperor on a small island than in a crowded metropolis full of traitors and spies. In truth, explained the Roman historian Tacitus, Sejanus’s aim was to control all the information reaching the emperor: “Access to the emperor would be under his own control, and letters, for the most part being conveyed by soldiers, would pass through his hands.”13

    With the Praetorians controlling Rome, Tiberius isolated in Capri, and Sejanus controlling all information reaching Tiberius, the Praetorian commander became the true ruler of the empire. Sejanus purged anyone who might oppose him—including members of the imperial family—by falsely accusing them of treason. Since nobody could contact the emperor without Sejanus’s permission, Tiberius was reduced to a puppet.

    Eventually someone—perhaps Tiberius’s sister-in-law Antonia—located an opening in Sejanus’s information cordon. A letter was smuggled to the emperor, explaining to him what was going on. But by the time Tiberius woke up to the danger and resolved to get rid of Sejanus, he was almost helpless. How could he topple the man who controlled not just the bodyguards but also all communications with the outside world? If he tried to make a move, Sejanus could imprison him on Capri indefinitely and inform the Senate and the army that the emperor was too ill to travel anywhere.

    Tiberius nevertheless managed to turn the tables on Sejanus. As Sejanus grew in power and became preoccupied with running the empire, he lost touch with the day-to-day minutiae of Rome’s security apparatus. Tiberius managed to secretly gain the support of Naevius Sutorius Macro, commander of Rome’s fire brigade and night watch. Macro orchestrated a coup against Sejanus, and as a reward Tiberius made Macro the new commander of the Praetorian Guard. A few years later, Macro had Tiberius killed.14

    Power lies at the nexus where the information channels merge. Since Tiberius allowed the information channels to merge in the person of Sejanus, the latter became the true center of power, while Tiberius was reduced to a puppet.

    The fate of Tiberius indicates the delicate balance that all dictators must strike. They try to concentrate all information in one place, but they must be careful that the different channels of information are allowed to merge only in their own person. If the information channels merge somewhere else, that then becomes the true nexus of power. When the regime relies on humans like Sejanus and Macro, a skillful dictator can play them one against the other in order to remain on top. Stalin’s purges were all about that. Yet when a regime relies on a powerful but inscrutable AI that gathers and analyzes all information, the human dictator is in danger of losing all power. He may remain in the capital and yet be isolated on a digital island, controlled and manipulated by the AI.

    THE DICTATOR’S DILEMMA

    In the next few years, the dictators of our world face more urgent problems than an algorithmic takeover. No current AI system can manipulate regimes at such a scale. However, totalitarian systems are already in danger of putting far too much trust in algorithms. Whereas democracies assume that everyone is fallible, in totalitarian regimes the fundamental assumption is that the ruling party or the supreme leader is always right. Regimes based on that assumption are conditioned to believe in the existence of an infallible intelligence and are reluctant to create strong self-correcting mechanisms that might monitor and regulate the genius at the top.

    Until now such regimes placed their faith in human parties and leaders and were hothouses for the growth of personality cults. But in the twenty-first century this totalitarian tradition prepares them to expect AI infallibility. Systems that could believe in the perfect genius of a Mussolini, a Ceauşescu, or a Khomeini are primed to also believe in the flawless genius of a superintelligent computer. This could have disastrous results for their citizens, and potentially for the rest of the world as well. What happens if the algorithm in charge of environmental policy makes a big mistake, but there are no self-correcting mechanisms that can identify and correct its error? What happens if the algorithm running the state’s social credit system begins terrorizing not just the general population but even the members of the ruling party and simultaneously begins to label anyone that questions its policies “an enemy of the people”?

    Dictators have always suffered from weak self-correcting mechanisms and have always been threatened by powerful subordinates. The rise of AI may greatly exacerbate these problems. The computer network therefore presents dictators with an excruciating dilemma. They could decide to escape the clutches of their human underlings by trusting a supposedly infallible technology, in which case they might become the technology’s puppet. Or, they could build a human institution to supervise the AI, but that institution might limit their own power, too.

    If even just a few of the world’s dictators choose to put their trust in AI, this could have far-reaching consequences for the whole of humanity. Science fiction is full of scenarios of an AI getting out of control and enslaving or eliminating humankind. Most sci-fi plots explore these scenarios in the context of democratic capitalist societies. This is understandable. Authors living in democracies are obviously interested in their own societies, whereas authors living in dictatorships are usually discouraged from criticizing their rulers. But the weakest spot in humanity’s anti-AI shield is probably the dictators. The easiest way for an AI to seize power is not by breaking out of Dr. Frankenstein’s lab but by ingratiating itself with some paranoid Tiberius.

    This is not a prophecy, just a possibility. After 1945, dictators and their subordinates cooperated with democratic governments and their citizens to contain nuclear weapons. On July 9, 1955, Albert Einstein, Bertrand Russell, and a number of other eminent scientists and thinkers published the Russell-Einstein Manifesto, calling on the leaders of both democracies and dictatorships to cooperate on preventing nuclear war. “We appeal,” said the manifesto, “as human beings, to human beings: remember your humanity, and forget the rest. If you can do so, the way lies open to a new Paradise; if you cannot, there lies before you the risk of universal death.”15 This is true of AI too. It would be foolish of dictators to believe that AI will necessarily tilt the balance of power in their favor. If they aren’t careful, AI will just grab power to itself.

    CHAPTER 11 The Silicon Curtain: Global Empire or Global Split?

    The previous two chapters explored how different human societies might react to the rise of the new computer network. But we live in an interconnected world, where the decisions of one country can have a profound impact on others. Some of the gravest dangers posed by AI do not result from the internal dynamics of a single human society. Rather, they arise from dynamics involving many societies, which might lead to new arms races, new wars, and new imperial expansions.

    Computers are not yet powerful enough to completely escape our control or destroy human civilization by themselves. As long as humanity stands united, we can build institutions that will control AI and will identify and correct algorithmic errors. Unfortunately, humanity has never been united. We have always been plagued by bad actors, as well as by disagreements between good actors. The rise of AI, then, poses an existential danger to humankind not because of the malevolence of computers but because of our own shortcomings.

    Thus, a paranoid dictator might hand unlimited power to a fallible AI, including even the power to launch nuclear strikes. If the dictator trusts his AI more than his defense minister, wouldn’t it make sense to have the AI supervise the country’s most powerful weapons? If the AI then makes an error, or begins to pursue an alien goal, the result could be catastrophic, and not just for that country.

    Similarly, terrorists focused on events in one corner of the world might use AI to instigate a global pandemic. The terrorists might be more versed in some apocalyptic mythology than in the science of epidemiology, but they just need to set the goal, and all else will be done by their AI. The AI could synthesize a new pathogen, order it from commercial laboratories or print it in biological 3-D printers, and devise the best strategy to spread it around the world, via airports or food supply chains. What if the AI synthesizes a virus that is as deadly as Ebola, as contagious as COVID-19, and as slow acting as AIDS? By the time the first victims begin to die, and the world is alerted to the danger, most people on earth might have already been infected.1

    As we have seen in previous chapters, human civilization is threatened not only by physical and biological weapons of mass destruction like atom bombs and viruses. Human civilization could also be destroyed by weapons of social mass destruction, like stories that undermine our social bonds. An AI developed in one country could be used to unleash a deluge of fake news, fake money, and fake humans so that people in numerous other countries lose the ability to trust anything or anyone.

    Many societies—both democracies and dictatorships—may act responsibly to regulate such usages of AI, clamp down on bad actors, and restrain the dangerous ambitions of their own rulers and fanatics. But if even a handful of societies fail to do so, this could be enough to endanger the whole of humankind. Climate change can devastate even countries that adopt excellent environmental regulations, because it is a global rather than a national problem. AI, too, is a global problem. Countries would be naive to imagine that as long as they regulate AI wisely within their own borders, these regulations will protect them from the worst outcomes of the AI revolution. Accordingly, to understand the new computer politics, it is not enough to examine how discrete societies might react to AI. We also need to consider how AI might change relations between societies on a global level.

    At present, the world is divided into about two hundred nation-states, most of which gained their independence only after 1945. They are not all equal. The list contains two superpowers, a handful of major powers, several blocs and alliances, and a lot of smaller fish. Still, even the tiniest states enjoy some leverage, as evidenced by their ability to play the superpowers against each other. In the early 2020s, for example, China and the United States competed for influence in the strategically important South Pacific region. Both superpowers courted island nations like Tonga, Tuvalu, Kiribati, and the Solomon Islands. The governments of these small nations—whose populations range from 740,000 (Solomon Islands) to 11,000 (Tuvalu)—had substantial leeway to decide which way to tack and were able to extract considerable concessions and aid.2

    Other small states, such as Qatar, have established themselves as important players in the geopolitical arena. With only 300,000 citizens, Qatar is nevertheless pursuing ambitious foreign policy aims in the Middle East, is playing an outsized rule in the global economy, and is home to Al Jazeera, the Arab world’s most influential TV network. One might argue that Qatar is able to punch well above its size because it is the third-largest exporter of natural gas in the world. Yet in a different international setting, that would have made Qatar not an independent actor but the first course on the menu of any imperial conqueror. It is telling that, as of 2024, Qatar’s much bigger neighbors, and the world’s hegemonic powers, are letting the tiny Gulf state hold on to its fabulous riches. Many people describe the international system as a jungle. If so, it is a jungle in which tigers allow fat chickens to live in relative safety.

    Qatar, Tonga, Tuvalu, Kiribati, and the Solomon Islands all indicate that we are living in a postimperial era. They gained their independence from the British Empire in the 1970s, as part of the final demise of the European imperial order. The leverage they now have in the international arena testifies that in the first quarter of the twenty-first century power is distributed between a relatively large number of players, rather than monopolized by a few empires.

    How might the rise of the new computer network change the shape of international politics? Aside from apocalyptic scenarios such as a dictatorial AI launching a nuclear war, or a terrorist AI instigating a lethal pandemic, computers pose two main challenges to the current international system. First, since computers make it easier to concentrate information and power in a central hub, humanity could enter a new imperial era. A few empires (or perhaps a single empire) might bring the whole world under a much tighter grip than that of the British Empire or the Soviet Empire. Tonga, Tuvalu, and Qatar would be transformed from independent states into colonial possessions—just as they were fifty years ago.

    Second, humanity could split along a new Silicon Curtain that would pass between rival digital empires. As each regime chooses its own answer to the AI alignment problem, to the dictator’s dilemma, and to other technological quandaries, each might create a separate and very different computer network. The various networks might then find it ever more difficult to interact, and so would the humans they control. Qataris living as part of an Iranian or Russian network, Tongans living as part of a Chinese network, and Tuvaluans living as part of an American network could come to have such different life experiences and worldviews that they would hardly be able to communicate or to agree on much.

    If these developments indeed materialize, they could easily lead to their own apocalyptic outcome. Perhaps each empire can keep its nuclear weapons under human control and its lunatics away from bioweapons. But a human species divided into hostile camps that cannot understand each other stands a small chance of avoiding devastating wars or preventing catastrophic climate change. A world of rival empires separated by an opaque Silicon Curtain would also be incapable of regulating the explosive power of AI.

    THE RISE OF DIGITAL EMPIRES

    In chapter 9 we touched briefly on the link between the Industrial Revolution and modern imperialism. It was not evident, at the beginning, that industrial technology would have much of an impact on empire building. When the first steam engines were put to use to pump water in British coal mines in the eighteenth century, no one foresaw that they would eventually power the most ambitious imperial projects in human history. When the Industrial Revolution subsequently gathered steam in the early nineteenth century, it was driven by private businesses, because governments and armies were relatively slow to appreciate its potential geopolitical impact. The world’s first commercial railway, for example, which opened in 1830 between Liverpool and Manchester, was built and operated by the privately owned Liverpool and Manchester Railway Company. The same was true of most other early railway lines in the U.K., the United States, France, Germany, and elsewhere. At that point, it wasn’t at all clear why governments or armies should get involved in such commercial enterprises.

    By the middle of the nineteenth century, however, the governments and armed forces of the leading industrial powers had fully recognized the immense geopolitical potential of modern industrial technology. The need for raw materials and markets justified imperialism, while industrial technologies made imperial conquests easier. Steamships were crucial, for example, to the British victory over the Chinese in the Opium Wars, and railroads played a decisive role in the American expansion west and the Russian expansion east and south. Indeed, entire imperial projects were shaped around the construction of railroads such as the Trans-Siberian and Trans-Caspian Russian lines, the German dream of a Berlin-Baghdad railway, and the British dream of building a railway from Cairo to the Cape.3

    Nevertheless, most polities didn’t join the burgeoning industrial arms race in time. Some lacked the capacity to do so, like the Melanesian chiefdoms of the Solomon Islands and the Al Thani tribe of Qatar. Others, like the Burmese Empire, the Ashanti Empire, and the Chinese Empire, might have had the capacity but lacked the will and foresight. Their rulers and inhabitants either didn’t follow developments in places like the British Midlands or didn’t think they had much to do with them. Why should the rice farmers of the Irrawaddy basin in Burma or the Yangtze basin in China concern themselves about the Liverpool–Manchester Railway? By the end of the nineteenth century, however, these rice farmers found themselves either conquered or indirectly exploited by the British Empire. Most other stragglers in the industrial race also ended up dominated by one industrial power or other. Could something similar happen with AI?

    When the race to develop AI gathered steam in the early years of the twenty-first century, it too was initially spearheaded by private entrepreneurs in a handful of countries. They set their sights on centralizing the world’s flow of information. Google wanted to organize all the world’s information in one place. Amazon sought to centralize all the world’s shopping. Facebook wished to connect all the world’s social spheres. But concentrating all the world’s information is neither practical nor helpful unless one can centrally process that information. And in 2000, when Google’s search engine was making its baby steps, when Amazon was a modest online bookshop, and when Mark Zuckerberg was in high school, the AI necessary to centrally process oceans of data was nowhere at hand. But some people bet it was just around the corner.

    Kevin Kelly, the founding editor of Wired magazine, recounted how in 2002 he attended a small party at Google and struck up a conversation with Larry Page. “Larry, I still don’t get it. There are so many search companies. Web search, for free? Where does that get you?” Page explained that Google wasn’t focused on search at all. “We’re really making an AI,” he said.4 Having lots of data makes it easier to create an AI. And AI can turn lots of data into lots of power.

    By the 2010s, the dream was becoming a reality. Like every major historical revolution, the rise of AI was a gradual process involving numerous steps. And like every revolution, a few of these steps were seen as turning points, just like the opening of the Liverpool–Manchester Railway. In the prolific literature on the story of AI, two events pop up again and again. The first occurred when, on September 30, 2012, a convolutional neural network called AlexNet won the ImageNet Large Scale Visual Recognition Challenge.

    If you have no idea what a convolutional neural network is, and if you have never heard of the ImageNet challenge, you are not alone. More than 99 percent of us are in the same situation, which is why AlexNet’s victory was hardly front-page news in 2012. But some humans did hear about AlexNet’s victory and decoded the writing on the wall.

    They knew, for example, that ImageNet is a database of millions of annotated digital images. Did a website ever ask you to prove that you are not a robot by looking at a set of images and indicating which ones contain a car or a cat? The images you clicked were perhaps added to the ImageNet database. The same thing might also have happened to tagged images of your pet cat that you uploaded online. The ImageNet Large Scale Visual Recognition Challenge tests various algorithms on how well they are able to identify the annotated images in the database. Can they correctly identify the cats? When humans are asked to do it, out of one hundred cat images we correctly identify ninety-five as cats. In 2010 the best algorithms had a success rate of only 72 percent. In 2011 the algorithmic success rate crawled up to 75 percent. In 2012 the AlexNet algorithm won the challenge and stunned the still minuscule community of AI experts by achieving a success rate of 85 percent. While this improvement may not sound like much to laypersons, it demonstrated to the experts the potential for rapid progress in certain AI domains. By 2015 a Microsoft algorithm achieved 96 percent accuracy, surpassing the human ability to identify cat images.

    In 2016, The Economist published a piece titled “From Not Working to Neural Networking” that asked, “How has artificial intelligence, associated with hubris and disappointment since its earliest days, suddenly become the hottest field in technology?” It pointed to AlexNet’s victory as the moment when “people started to pay attention, not just within the AI community but across the technology industry as a whole.” The article was illustrated with an image of a robotic hand holding up a photo of a cat.5

    All those cat images that tech giants had been harvesting from across the world, without paying a penny to either users or tax collectors, turned out to be incredibly valuable. The AI race was on, and the competitors were running on cat images. At the same time that AlexNet was preparing for the ImageNet challenge, Google too was training its AI on cat images, and even created a dedicated cat-image-generating AI called the Meow Generator.6 The technology developed by recognizing cute kittens was later deployed for more predatory purposes. For example, Israel relied on it to create the Red Wolf, Blue Wolf, and Wolf Pack apps used by Israeli soldiers for facial recognition of Palestinians in the Occupied Territories.7 The ability to recognize cat images also led to the algorithms Iran uses to automatically recognize unveiled women and enforce its hijab laws. As explained in chapter 8, massive amounts of data are required to train machine-learning algorithms. Without millions of cat images uploaded and annotated for free by people across the world, it would not have been possible to train the AlexNet algorithm or the Meow Generator, which in turn served as the template for subsequent AIs with far-reaching economic, political, and military potential.8

    Just as in the early nineteenth century the effort to build railways was pioneered by private entrepreneurs, so in the early twenty-first century private corporations were the initial main competitors in the AI race. The executives of Google, Facebook, Alibaba, and Baidu saw the value of recognizing cat images before the presidents and generals did. The second eureka moment, when the presidents and generals caught on to what was happening, occurred in mid-March 2016. It was the aforementioned victory of Google’s AlphaGo over Lee Sedol. Whereas AlexNet’s achievement was largely ignored by politicians, AlphaGo’s triumph sent shock waves through government offices, especially in East Asia. In China and neighboring countries go is a cultural treasure and considered an ideal training for aspiring strategists and policy makers. In March 2016, or so the mythology of AI would have it, the Chinese government realized that the age of AI had begun.9

    It is little wonder that the Chinese government was probably the first to understand the full importance of what was happening. In the nineteenth century, China was late to appreciate the potential of the Industrial Revolution and was slow to adopt inventions like railroads and steamships. It consequently suffered what the Chinese call “the century of humiliations.” After having been the world’s greatest superpower for centuries, failing to adopt modern industrial technology brought China to its knees. It was repeatedly defeated in wars, partially conquered by foreigners, and thoroughly exploited by the powers that did understand railroads and steamships. The Chinese vowed never again to miss the train.

    In 2017, China’s government released its “New Generation Artificial Intelligence Plan,” which announced that “by 2030, China’s AI theories, technologies, and application should achieve world-leading levels, making China the world’s primary AI innovation center.”10 In the following years China poured enormous resources into AI so that by the early 2020s it is already leading the world in several AI-related fields and catching up with the United States in others.11

    Of course, the Chinese government wasn’t the only one that woke up to the importance of AI. On September 1, 2017, President Putin of Russia declared, “Artificial intelligence is the future, not only for Russia, but for all humankind.… Whoever becomes the leader in this sphere will become the ruler of the world.” In January 2018, Prime Minister Modi of India concurred that “the one who control [sic] the data will control the world.”12 In February 2019, President Trump signed an executive order on AI, saying that “the age of AI has arrived” and that “continued American leadership in Artificial Intelligence is of paramount importance to maintaining the economic and national security of the United States.”13 The United States at the time was already the leader in the AI race, thanks largely to efforts of visionary private entrepreneurs. But what began as a commercial competition between corporations was turning into a match between governments, or perhaps more accurately, into a race between competing teams, each made of one government and several corporations. The prize for the winner? World domination.

    DATA COLONIALISM

    In the sixteenth century, when Spanish, Portuguese, and Dutch conquistadors were building the first global empires in history, they came with sailing ships, horses, and gunpowder. When the British, Russians, and Japanese made their bids for hegemony in the nineteenth and twentieth centuries, they relied on steamships, locomotives, and machine guns. In the twenty-first century, to dominate a colony, you no longer need to send in the gunboats. You need to take out the data. A few corporations or governments harvesting the world’s data could transform the rest of the globe into data colonies—territories they control not with overt military force but with information.14

    Imagine a situation—in twenty years, say—when somebody in Beijing or San Francisco possesses the entire personal history of every politician, journalist, colonel, and CEO in your country: every text they ever sent, every web search they ever made, every illness they suffered, every sexual encounter they enjoyed, every joke they told, every bribe they took. Would you still be living in an independent country, or would you now be living in a data colony? What happens when your country finds itself utterly dependent on digital infrastructures and AI-powered systems over which it has no effective control?

    Such a situation can lead to a new kind of data colonialism in which control of data is used to dominate faraway colonies. Mastery of AI and data could also give the new empires control of people’s attention. As we have already discussed, in the 2010s American social media giants like Facebook and YouTube upended the politics of distant countries like Myanmar and Brazil in pursuit of profit. Future digital empires may do something similar for political interests.

    Fears of psychological warfare, data colonialism, and loss of control over their cyberspace have led many countries to already block what they see as dangerous apps. China has banned Facebook, YouTube, and many other Western social media apps and websites. Russia has banned almost all Western social media apps as well as some Chinese ones. In 2020, India banned TikTok, WeChat, and numerous other Chinese apps on the grounds that they were “prejudicial to sovereignty and integrity of India, defense of India, security of state and public order.”15 The United States has been debating whether to ban TikTok—concerned that the app might be serving Chinese interests—and as of 2023 it is illegal to use it on the devices of almost all federal employees, state employees, and government contractors.16 Lawmakers in the U.K., New Zealand, and other countries have also expressed concerns over TikTok.17 Numerous other governments, from Iran to Ethiopia, have blocked various apps like Facebook, Twitter, YouTube, Telegram, and Instagram.

    Data colonialism could also manifest itself in the spread of social credit systems. What might happen, for example, if a dominant player in the global digital economy decides to establish a social credit system that harvests data anywhere it can and scores not only its own nationals but people throughout the world? Foreigners couldn’t just shrug off their score, because it might affect them in numerous ways, from buying flight tickets to applying for visas, scholarships, and jobs. Just as tourists use the global scores given by foreign corporations like Tripadvisor and Airbnb to evaluate restaurants and vacation homes even in their own country, and just as people throughout the world use the U.S. dollar for commercial transactions, so people everywhere might begin to use a Chinese or an American social credit score for local social interactions.

    Becoming a data colony will have economic as well as political and social consequences. In the nineteenth and twentieth centuries, if you were a colony of an industrial power like Belgium or Britain, it usually meant that you provided raw materials, while the cutting-edge industries that made the biggest profits remained in the imperial hub. Egypt exported cotton to Britain and imported high-end textiles. Malaya provided rubber for tires; Coventry made the cars.18

    Something analogous is likely to happen with data colonialism. The raw material for the AI industry is data. To produce AI that recognizes images, you need cat photos. To produce the trendiest fashion, you need data on fashion trends. To produce autonomous vehicles, you need data about traffic patterns and car accidents. To produce health-care AI, you need data about genes and medical conditions. In a new imperial information economy, raw data will be harvested throughout the world and will flow to the imperial hub. There the cutting-edge technology will be developed, producing unbeatable algorithms that know how to identify cats, predict fashion trends, drive autonomous vehicles, and diagnose diseases. These algorithms will then be exported back to the data colonies. Data from Egypt and Malaysia might make a corporation in San Francisco or Beijing rich, while people in Cairo and Kuala Lumpur remain poor, because neither the profits nor the power is distributed back.

    The nature of the new information economy might make the imbalance between imperial hub and exploited colony worse than ever. In ancient times land—rather than information—was the most important economic asset. This precluded the overconcentration of all wealth and power in a single hub. As long as land was paramount, considerable wealth and power always remained in the hands of provincial landowners. A Roman emperor, for example, could put down one provincial revolt after another, but on the day after decapitating the last rebel chief, he had no choice but to appoint a new set of provincial landowners who might again challenge the central power. In the Roman Empire, although Italy was the seat of political power, the richest provinces were in the eastern Mediterranean. It was impossible to transport the fertile fields of the Nile valley to the Italian Peninsula.19 Eventually the emperors abandoned the city of Rome to the barbarians and moved the seat of political power to the rich east, to Constantinople.

    During the Industrial Revolution machines became more important than land. Factories, mines, railroad lines, and electrical power stations became the most valuable assets. It was somewhat easier to concentrate these kinds of assets in one place. The British Empire could centralize industrial production in its home islands, extract raw materials from India, Egypt, and Iraq, and sell them finished goods made in Birmingham or Belfast. Unlike in the Roman Empire, Britain was the seat of both political and economic power. But physics and geology still put natural limits on this concentration of wealth and power. The British couldn’t move every cotton mill from Calcutta to Manchester, nor shift the oil wells from Kirkuk to Yorkshire.

    Information is different. Unlike cotton and oil, digital data can be sent from Malaysia or Egypt to Beijing or San Francisco at almost the speed of light. And unlike land, oil fields, or textile factories, algorithms don’t take up much space. Consequently, unlike industrial power, the world’s algorithmic power can be concentrated in a single hub. Engineers in a single country might write the code and control the keys for all the crucial algorithms that run the entire world.

    Indeed, AI makes it possible to concentrate in one place even the decisive assets of some traditional industries, like textile. In the nineteenth century, to control the textile industry meant to control sprawling cotton fields and huge mechanical production lines. In the twenty-first century, the most important asset of the textile industry is information rather than cotton or machinery. To beat the competitors, a garment producer needs information about the likes and dislikes of customers and the ability to predict or manufacture the next fashions. By controlling this type of information, high-tech giants like Amazon and Alibaba can monopolize even a very traditional industry like textile. In 2021, Amazon became the United States’ biggest single clothing retailer.20

    Moreover, as AI, robots, and 3-D printers automate textile production, millions of workers might lose their jobs, upending national economies and the global balance of power. What will happen to the economies and politics of Pakistan and Bangladesh, for example, when automation makes it cheaper to produce textiles in Europe? Consider that at present the textile sector provides employment to 40 percent of Pakistan’s total labor force and accounts for 84 percent of Bangladesh’s export earnings.21 As noted in chapter 7, while automation might make millions of textile workers redundant, it will probably create many new jobs, too. For instance, there might be a huge demand for coders and data analysts. But turning an unemployed factory hand into a data analyst demands a substantial up-front investment in retraining. Where would Pakistan and Bangladesh get the money to do that?

    AI and automation therefore pose a particular challenge to poorer developing countries. In an AI-driven economy, the digital leaders claim the bulk of the gains and could use their wealth to retrain their workforce and profit even more. Meanwhile, the value of unskilled laborers in left-behind countries will decline, and they will not have the resources to retrain their workforce, causing them to fall even further behind. The result might be lots of new jobs and immense wealth in San Francisco and Shanghai, while many other parts of the world face economic ruin.22 According to the global accounting firm PricewaterhouseCoopers, AI is expected to add $15.7 trillion to the global economy by 2030. But if current trends continue, it is projected that China and North America—the two leading AI superpowers—will together take home 70 percent of that money.23

    FROM WEB TO COCOON

    These economic and geopolitical dynamics could divide the world between two digital empires. During the Cold War, the Iron Curtain was in many places literally made of metal: barbed wire separated one country from another. Now the world is increasingly divided by the Silicon Curtain. The Silicon Curtain is made of code, and it passes through every smartphone, computer, and server in the world. The code on your smartphone determines on which side of the Silicon Curtain you live, which algorithms run your life, who controls your attention, and where your data flows.

    It is becoming difficult to access information across the Silicon Curtain, say between China and the United States, or between Russia and the EU. Moreover, the two sides are increasingly run on different digital networks, using different computer codes. Each sphere obeys different regulations and serves different purposes. In China, the most important aim of new digital technology is to strengthen the state and serve government policies. While private enterprises are given a certain amount of autonomy in developing and deploying AI tools, their economic activities are ultimately subservient to the government’s political goals. These political goals also justify a relatively high level of surveillance, both online and off-line. This means, for example, that though Chinese citizens and authorities do care about people’s privacy, China is already far ahead of the United States and other Western countries in developing and deploying social credit systems that encompass the whole of people’s lives.24

    In the United States, the government plays a more limited role. Private enterprises lead the development and deployment of AI, and the ultimate goal of many new AI tools is to enrich the tech giants rather than to strengthen the American state or the current administration. Indeed, in many cases governmental policies are themselves shaped by powerful business interests. But the U.S. system does offer greater protection for citizens’ privacy. While American corporations aggressively gather information on people’s online activities, they are much more restricted in surveilling people’s offline lives. There is also widespread rejection of the ideas behind all-embracing social credit systems.25

    These political, cultural, and regulatory differences mean that each sphere is using different software. In China you cannot use Google and Facebook, and you cannot access Wikipedia. In the United States few people use WeChat, Baidu, and Tencent. More important, the spheres aren’t mirror images of each other. It is not that the Chinese and Americans develop local versions of the same apps. Baidu isn’t the Chinese Google. Alibaba isn’t the Chinese Amazon. They have different goals, different digital architectures, and different impacts on people’s lives.26 These differences influence much of the world, since most countries rely on Chinese and American software rather than on local technology.

    Each sphere also uses different hardware like smartphones and computers. The United States pressures its allies and clients to avoid Chinese hardware, such as Huawei’s 5G infrastructure.27 The Trump administration blocked an attempt by the Singaporean corporation Broadcom to buy the leading American producer of computer chips, Qualcomm. They feared foreigners might insert back doors into the chips or would prevent the U.S. government from inserting its own back doors there.28 In 2022, the Biden administration placed strict limits on trade in high-performance computing chips necessary for the development of AI. U.S. companies were forbidden to export such chips to China, or to provide China with the means to manufacture or repair them. The restrictions have subsequently been tightened further, and the ban was expanded to include other nations such as Russia and Iran.29 While in the short term this hampers China in the AI race, in the long term it will push China to develop a completely separate digital sphere that will be distinct from the American digital sphere even in its smallest building blocks.30

    The two digital spheres may drift further and further apart. Chinese software would talk only with Chinese hardware and Chinese infrastructure, and the same would happen on the other side of the Silicon Curtain. Since digital code influences human behavior, and human behavior in turn shapes digital code, the two sides may well be moving along different trajectories that will make them more and more different not just in their technology but in their cultural values, social norms, and political structures. After generations of convergence, humanity could find itself at a crucial point of divergence.31 For centuries, new information technologies fueled the process of globalization and brought people all over the world into closer contact. Paradoxically, information technology today is so powerful it can potentially split humanity by enclosing different people in separate information cocoons, ending the idea of a single shared human reality. While the web has been our main metaphor in recent decades, the future might belong to cocoons.

    THE GLOBAL MIND-BODY SPLIT

    The division into separate information cocoons could lead not just to economic rivalries and international tensions but also to the development of very different cultures, ideologies, and identities. Guessing future cultural and ideological developments is usually a fool’s errand. It is far more difficult than predicting economic and geopolitical developments. How many Romans or Jews in the days of Tiberius could have anticipated that a splinter Jewish sect would eventually take over the Roman Empire and that the emperors would abandon Rome’s old gods to worship an executed Jewish rabbi?

    It would have been even more difficult to foresee the directions in which various Christian sects would develop and the momentous impact of their ideas and conflicts on everything from politics to sexuality. When Jesus was asked about paying taxes to Tiberius’s government and answered, “Render unto Caesar the things that are Caesar’s, and unto God the things that are God’s” (Matthew 22:21), nobody could imagine the impact his response would have on the separation of church and state in the American republic two millennia later. And when Saint Paul wrote to the Christians in Rome, “I myself in my mind am a slave to God’s law, but in my sinful flesh a slave to the law of sin” (Romans 7:25), who could have foreseen the repercussions this would have on schools of thought ranging from Cartesian philosophy to queer theory?

    Despite these difficulties, it is important to try to imagine future cultural developments, in order to alert ourselves to the fact that the AI revolution and the formation of rival digital spheres are likely to change more than just our jobs and political structures. The following paragraphs contain some admittedly ambitious speculation, so please bear in mind that my goal is not to accurately foretell cultural developments but merely to draw attention to the likelihood that profound cultural shifts and conflicts await us.

    One possible development with far-reaching consequences is that different digital cocoons might adopt incompatible approaches to the most fundamental questions of human identity. For thousands of years, many religious and cultural conflicts—for example, between rival Christian sects, between Hindus and Buddhists, and between Platonists and Aristotelians—were fueled by disagreements about the mind-body problem. Are humans a physical body, or a nonphysical mind, or perhaps a mind trapped inside a body? In the twenty-first century, the computer network might supercharge the mind-body problem and turn it into a cause for major personal, ideological, and political conflicts.

    To appreciate the political ramifications of the mind-body problem, let’s briefly revisit the history of Christianity. Many of the earliest Christian sects, influenced by Jewish thinking, believed in the Old Testament idea that humans are embodied beings and that the body plays a crucial role in human identity. The book of Genesis said God created humans as physical bodies, and almost all books of the Old Testament assume that humans can exist only as physical bodies. With a few possible exceptions, the Old Testament doesn’t mention the possibility of a bodiless existence after death, in heaven or hell. When the ancient Jews fantasized about salvation, they imagined it to mean an earthly kingdom of material bodies. In the time of Jesus, many Jews believed that when the Messiah finally comes, the bodies of the dead would come back to life, here on earth. The Kingdom of God, established by the Messiah, was supposed to be a material kingdom, with trees and stones and flesh-and-blood bodies.32

    This was also the view of Jesus himself and the first Christians. Jesus promised his followers that soon the Kingdom of God would be built here on earth and they would inhabit it in their material bodies. When Jesus died without fulfilling his promise, his early followers came to believe that he was resurrected in the flesh and that when the Kingdom of God finally materialized on earth, they too would be resurrected in the flesh. The church father Tertullian (160–240 CE) wrote that “the flesh is the very condition on which salvation hinges,” and the catechism of the Catholic Church, citing the doctrines adopted at the Second Council of Lyon in 1274, states, “We believe in God who is creator of the flesh; we believe in the Word made flesh in order to redeem the flesh; we believe in the resurrection of the flesh, the fulfillment of both the creation and the redemption of the flesh.… We believe in the true resurrection of this flesh that we now possess.”33

    Despite such seemingly unequivocal statements, we saw that Saint Paul already had his doubts about the flesh, and by the fourth century CE, under Greek, Manichaean, and Persian influences, some Christians had drifted toward a dualistic approach. They came to think of humans as consisting of a good immaterial soul trapped inside an evil material body. They didn’t fantasize about being resurrected in the flesh. Just the opposite. Having been released by death from its abominable material prison, why would the pure soul ever want to get back in? Christians accordingly began to believe that after death the soul is liberated from the body and exists forever in an immaterial place completely beyond the physical realm—which is the standard belief among Christians today, notwithstanding what Tertullian and the Second Council of Lyon said.34

    But Christianity couldn’t completely abandon the old Jewish view that humans are embodied beings. After all, Christ appeared on earth in the flesh. His body was nailed to the cross, on which he experienced excruciating pain. For two thousand years, Christian sects therefore fought each other—sometimes with words, sometimes with swords—over the exact relations between soul and body. The fiercest arguments focused on Christ’s own body. Was he material? Was he purely spiritual? Did he perhaps have a nonbinary nature, being both human and divine at the same time?

    The different approaches to the mind-body problem influenced how people treated their own bodies. Saints, hermits, and monks made breathtaking experiments in pushing the human body to its limits. Just as Christ allowed his body to be tortured on the cross, so these “athletes of Christ” allowed lions and bears to rip them apart while their souls rejoiced in divine ecstasy. They wore hair shirts, fasted for weeks, or stood for years on a pillar—like the famous Simeon who allegedly stood for about forty years on top of a pillar near Aleppo.35

    Other Christians took the opposite approach, believing that the body didn’t matter at all. The only thing that mattered was faith. This idea was taken to extremes by Protestants like Martin Luther, who formulated the doctrine of sola fide: only faith. After living as a monk for about ten years, fasting and torturing his body in various ways, Luther despaired of these bodily exercises. He reasoned that no bodily self-torments could force God to redeem him. Indeed, thinking he could win his own salvation by torturing his body was the sin of pride. Luther therefore disrobed, married a former nun, and told his followers that to be good Christians, the only thing they needed was to have complete faith in Christ.36

    These ancient theological debates about mind and body may seem utterly irrelevant to the AI revolution, but they have in fact been resurrected by twenty-first-century technologies. What is the relationship between our physical body and our online identities and avatars? What is the relation between the offline world and cyberspace? Suppose I spend most of my waking hours sitting in my room in front of a screen, playing online games, forming virtual relationships, and even working remotely. I hardly venture out even to eat. I just order takeout. If you are like ancient Jews and the first Christians, you would pity me and conclude that I must be living in a delusion, losing touch with the reality of physical spaces and flesh-and-blood bodies. But if your thinking is closer to that of Luther and many later Christians, you might think I am liberated. By shifting most of my activities and relationships online, I have released myself from the limited organic world of debilitating gravity and corrupt bodies and can enjoy the unlimited possibilities of a digital world, which is potentially liberated from the laws of biology and even physics. I am free to roam a much vaster and more exciting space and to explore new aspects of my identity.

    An increasingly important question is whether people can adopt any virtual identity they like, or should their identity be constrained by their biological body? If we follow the Lutheran position of sola fide, the biological body isn’t of much importance. To adopt a certain online identity, the only thing that matters is what you believe. This debate can have far-reaching consequences not just for human identity but for our attitude to the world as a whole. A society that understands identities in terms of biological bodies should also care more about material infrastructure like sewage pipes and about the ecosystem that sustains our bodies. It will see the online world as an auxiliary of the offline world that can serve various useful purposes but can never become the central arena of our lives. Its aim would be to create an ideal physical and biological realm—the Kingdom of God on earth. In contrast, a society that downplays biological bodies and focuses on online identities may well seek to create an immersive Kingdom of God in cyberspace while discounting the fate of mere material things like sewage pipes and rain forests.

    This debate could shape attitudes not only toward organisms but also toward digital entities. As long as society defines identity by focusing on physical bodies, it is unlikely to view AIs as persons. But if society gives less importance to physical bodies, then even AIs that lack any corporeal manifestations may be accepted as legal persons enjoying various rights.

    Throughout history, diverse cultures have given diverse answers to the mind-body problem. A twenty-first-century controversy about the mind-body problem could result in cultural and political splits more consequential even than the split between Jews and Christians or between Catholics and Protestants. What happens, for example, if the American sphere discounts the body, defines humans by their online identity, recognizes AIs as persons, and downplays the importance of the ecosystem, whereas the Chinese sphere adopts opposite positions? Current disagreements about violations of human rights or adherence to ecological standards will look minuscule in comparison. The Thirty Years’ War—arguably the most devastating war in European history—was fought at least in part because Catholics and Protestants couldn’t agree on doctrines like sola fide and on whether Christ was divine, human, or nonbinary. Might future conflicts start because of an argument about AI rights and the nonbinary nature of avatars?

    As noted, these are all wild speculations, and in all likelihood actual cultures and ideologies will develop in different—and perhaps even wilder—directions. But it is probable that within a few decades the computer network will cultivate new human and nonhuman identities that make little sense to us. And if the world will be divided into two rival digital cocoons, the identities of entities in one cocoon might be unintelligible to the inhabitants of the other.

    FROM CODE WAR TO HOT WAR

    While China and the United States are currently the front-runners in the AI race, they are not alone. Other countries or blocs, such as the EU, India, Brazil, and Russia, may try to create their own digital spheres, each influenced by different political, cultural, and religious traditions.37 Instead of being divided between just two global empires, the world might be divided among a dozen empires. It is unclear whether this will somewhat alleviate or only exacerbate the imperial competition.

    The more the new empires compete against one another, the greater the danger of armed conflict. The Cold War between the United States and the U.S.S.R. never escalated into a direct military confrontation largely thanks to the doctrine of mutually assured destruction. But the danger of escalation in the age of AI is bigger, because cyber warfare is inherently different from nuclear warfare.

    First, cyber weapons are much more versatile than nuclear bombs. Cyber weapons can bring down a country’s electric grid, but they can also be used to destroy a secret research facility, jam an enemy sensor, inflame a political scandal, manipulate elections, or hack a single smartphone. And they can do all that stealthily. They don’t announce their presence with a mushroom cloud and a storm of fire, nor do they leave a visible trail from launchpad to target. Consequently, at times it is hard to know if an attack even occurred or who launched it. If a database is hacked or sensitive equipment is destroyed, it’s hard to be sure whom to blame. The temptation to start a limited cyberwar is therefore big, and so is the temptation to escalate it. Rival countries like Israel and Iran or the United States and Russia have been trading cyber blows for years, in an undeclared but escalating war.38 This is becoming the new global norm, amplifying international tensions and pushing countries to cross one red line after another.

    A second crucial difference concerns predictability. The Cold War was like a hyperrational chess game, and the certainty of destruction in the event of nuclear conflict was so great that the desire to start a war was correspondingly small. Cyber warfare lacks this certainty. Nobody knows for sure where each side has planted its logic bombs, Trojan horses, and malwares. Nobody can be certain whether their own weapons would actually work when called upon. Would Chinese missiles fire when the order is given, or perhaps the Americans have hacked them or the chain of command? Would American aircraft carriers function as expected, or would they perhaps shut down mysteriously or sail around in circles?39

    Such uncertainty undermines the doctrine of mutually assured destruction. One side might convince itself—rightly or wrongly—that it can launch a successful first strike and avoid massive retaliation. Even worse, if one side thinks it has such an opportunity, the temptation to launch a first strike could become irresistible, because one never knows how long the window of opportunity will remain open. Game theory posits that the most dangerous situation in an arms race is when one side feels it has an advantage but that this advantage is slipping away.40

    Even if humanity avoids the worst-case scenario of global war, the rise of new digital empires could still endanger the freedom and prosperity of billions of people. The industrial empires of the nineteenth and twentieth centuries exploited and repressed their colonies, and it would be foolhardy to expect the new digital empires to behave much better. Moreover, as noted earlier, if the world is divided into rival empires, humanity is unlikely to cooperate effectively to overcome the ecological crisis or to regulate AI and other disruptive technologies like bioengineering.

    THE GLOBAL BOND

    Of course, no matter whether the world is divided between a few digital empires, remains a more diverse community of two hundred nation-states, or is split along altogether different and unforeseen lines, cooperation is always an option. Among humans, the precondition for cooperation isn’t similarity; it is the ability to exchange information. As long as we are able to converse, we might find some shared story that can bring us closer. This, after all, is what made Homo sapiens the dominant species on the planet.

    Just as different and even rival families can cooperate within a tribal network, and competing tribes can cooperate within a national network, so opposing nations and empires can cooperate within a global network. The stories that make such cooperation possible do not eliminate our differences; rather, they enable us to identify shared experiences and interests, which offer a common framework for thought and action.

    A large part of what nevertheless makes global cooperation difficult is the misguided notion that it requires abolishing all cultural, social, and political differences. Populist politicians often argue that if the international community agrees on a common story and on universal norms and values, this will destroy the independence and unique traditions of their own nation.41 This position was unabashedly distilled in 2015 by Marine Le Pen—leader of France’s National Front party—in an election speech in which she declared, “We have entered a new two-partyism. A two-partyism between two mutually exclusive conceptions that will from now on structure our political life. The cleavage no longer separates left and right, but globalists and patriots.”42 In August 2020, President Trump described his guiding ethos thus: “We have rejected globalism and embraced patriotism.”43

    Luckily, this binary position is mistaken in its basic assumption. Global cooperation and patriotism are not mutually exclusive. For patriotism isn’t about hating foreigners. It is about loving our compatriots. And there are many situations when, in order to take care of our compatriots, we need to cooperate with foreigners. COVID-19 provided us with one obvious example. Pandemics are global events, and without global cooperation it is hard to contain them, let alone prevent them. When a new virus or a mutant pathogen appears in one country, it puts all other countries in danger. Conversely, the biggest advantage of humans over pathogens is that we can cooperate in ways that pathogens cannot. Doctors in Germany and Brazil can alert one another to new dangers, give each other good advice, and work together to discover better treatments.

    If German scientists invent a vaccine against some new disease, how should Brazilians react to this German achievement? One option is to reject the foreign vaccine and wait until Brazilian scientists develop a Brazilian vaccine. That, however, would be not just foolish; it would be anti-patriotic. Brazilian patriots should want to use any available vaccine to help their compatriots, no matter where the vaccine was developed. In this situation, cooperating with foreigners is the patriotic thing to do. The threat of losing control of AIs is an analogous situation in which patriotism and global cooperation must go together. An out-of-control AI, just like an out-of-control virus, puts in danger humans in every nation. No human collective—whether a tribe, a nation, or the entire species—stands to benefit from letting power shift from humans to algorithms.

    Contrary to what populists argue, globalism doesn’t mean establishing a global empire, abandoning national loyalties, or opening borders to unlimited immigration. In fact, global cooperation means two far more modest things: first, a commitment to some global rules. These rules don’t deny the uniqueness of each nation and the loyalty people should owe their nation. They just regulate the relations between nations. A good model is the World Cup. The World Cup is a competition between nations, and people often show fierce loyalty to their national team. At the same time, the World Cup is an amazing display of global agreement. Brazil cannot play football against Germany unless Brazilians and Germans first agree on the same set of rules for the game. That’s globalism in action.

    The second principle of globalism is that sometimes—not always, but sometimes—it is necessary to prioritize the long-term interests of all humans over the short-term interests of a few. For example, in the World Cup, all national teams agree not to use performance-enhancing drugs, because everybody realizes that if they go down that path, the World Cup would eventually devolve into a competition between biochemists. In other fields where technology is a game changer, we should similarly strive to balance national and global interests. Nations will obviously continue to compete in the development of new technology, but sometimes they should agree to limit the development and deployment of dangerous technologies like autonomous weapons and manipulative algorithms—not purely out of altruism, but for their own self-preservation.

    THE HUMAN CHOICE

    Forging and keeping international agreements on AI will require major changes in the way the international system functions. While we have experience in regulating dangerous technologies like nuclear and biological weapons, the regulation of AI will demand unprecedented levels of trust and self-discipline, for two reasons. First, it is easier to hide an illicit AI lab than an illicit nuclear reactor. Second, AIs have a lot more dual civilian-military usages than nuclear bombs. Consequently, despite signing an agreement that bans autonomous weapon systems, a country could build such weapons secretly, or camouflage them as civilian products. For example, it might develop fully autonomous drones for delivering mail and spraying fields with pesticides that with a few minor modifications could also deliver bombs and spray people with poison. Consequently, governments and corporations will find it more difficult to trust that their rivals are really abiding by the agreed regulations—and to withstand the temptation to themselves waive the rules.44 Can humans develop the necessary levels of trust and self-discipline? Do changes like those have any precedent in history?

    Many people are skeptical of the human capacity to change, and in particular of the human ability to renounce violence and forge stronger global bonds. For example, “realist” thinkers like Hans Morgenthau and John Mearsheimer have argued that an all-out competition for power is the inescapable condition of the international system. Mearsheimer explains that “my theory sees great powers as concerned mainly with figuring out how to survive in a world where there is no agency to protect them from each other” and that “they quickly realize that power is the key to their survival.” Mearsheimer then asks “how much power states want” and answers that all states want as much power as they can get, “because the international system creates powerful incentives for states to look for opportunities to gain power at the expense of rivals.” He concludes, “A state’s ultimate goal is to be the hegemon in the system.”45

    This grim view of international relations is akin to the populist and Marxist views of human relations, in that they all see humans as interested only in power. And they are all founded upon a deeper philosophical theory of human nature, which the primatologist Frans de Waal termed “veneer theory.” It argues that at heart humans are Stone Age hunters who cannot but see the world as a jungle where the strong prey upon the weak and where might makes right. For millennia, the theory goes, humans have tried to camouflage this unchanging reality under a thin and mutable veneer of myths and rituals, but we have never really broken free from the law of the jungle. Indeed, our myths and rituals are themselves a weapon used by the jungle’s top dogs to deceive and trap their inferiors. Those who don’t realize this are dangerously naive and will fall prey to some ruthless predator.46

    There are reasons to think, however, that “realists” like Mearsheimer have a selective view of historical reality and that the law of the jungle is itself a myth. As de Waal and many other biologists documented in numerous studies, real jungles—unlike the one in our imagination—are full of cooperation, symbiosis, and altruism displayed by countless animals, plants, fungi, and even bacteria. Eighty percent of all land plants, for example, rely on symbiotic relationships with fungi, and almost 90 percent of vascular plant families enjoy symbiotic relationships with microorganisms. If organisms in the rain forests of Amazonia, Africa, or India abandoned cooperation in favor of an all-out competition for hegemony, the rain forests and all their inhabitants would quickly die. That’s the law of the jungle.47

    As for Stone Age humans, they were gatherers as well as hunters, and there is no firm evidence that they had irrepressible warlike tendencies. While there are plenty of speculations, the first unambiguous evidence for organized warfare appears in the archaeological record only about thirteen thousand years ago, at the site of Jebel Sahaba in the Nile valley.48 Even after that date, the record of war is variable rather than constant. Some periods were exceptionally violent, whereas others were relatively peaceful. The clearest pattern we observe in the long-term history of humanity isn’t the constancy of conflict, but rather the increasing scale of cooperation. A hundred thousand years ago, Sapiens could cooperate only at the level of bands. Over the millennia, we have found ways to create communities of strangers, first on the level of tribes and eventually on the level of religions, trade networks, and states. Realists should note that states are not the fundamental particles of human reality, but rather the product of arduous processes of building trust and cooperation. If humans were interested only in power, they could never have created states in the first place. Sure, conflicts have always remained a possibility—both between and within states—but they have never been an inescapable destiny.

    War’s intensity depends not on an immutable human nature but on shifting technological, economic, and cultural factors. As these factors change, so does war, as was clearly demonstrated in the post-1945 era. During that period, the development of nuclear technology greatly increased the potential price of war. From the 1950s onward it became clear to the superpowers that even if they could somehow win an all-out nuclear exchange, their victory would likely be a suicidal achievement, involving the sacrifice of most of their population.

    Simultaneously, the ongoing shift from a material-based economy to a knowledge-based economy decreased the potential gains of war. While it has remained feasible to conquer rice paddies and gold mines, by the late twentieth century these were no longer the main sources of economic wealth. The new leading industries, like the semiconductor sector, came to be based on technical skills and organizational know-how that could not be acquired by military conquest. Accordingly, some of the greatest economic miracles of the post-1945 era were achieved by the defeated powers of Germany, Italy, and Japan, and by countries like Sweden and Singapore that eschewed military conflicts and imperial conquests.

    Finally, the second half of the twentieth century also witnessed a profound cultural transformation, with the decline of age-old militaristic ideals. Artists increasingly focused on depicting the senseless horrors of combat rather than on glorifying its architects, and politicians came to power dreaming more of domestic reforms than of foreign conquests. Due to these technological, economic, and cultural changes, in the decades following the end of World War II most governments stopped seeing wars of aggression as an appealing tool to advance their interests, and most nations stopped fantasizing about conquering and destroying their neighbors. While civil wars and insurgencies have remained commonplace, the post-1945 world has seen a significant decline in full-scale wars between states, and most notably in direct armed conflicts between great powers.49

    Numerous statistics attest to the decline of war in this post-1945 era, but perhaps the clearest evidence is found in state budgets. For most of recorded history, the military was the number one item on the budget of every empire, sultanate, kingdom, and republic. Governments spent little on health care and education, because most of their resources were consumed by paying soldiers, constructing walls, and building warships. When the bureaucrat Chen Xiang examined the annual budget of the Chinese Song dynasty for the year 1065, he found that out of sixty million minqian (currency unit), fifty million (83 percent) were consumed by the military. Another official, Cai Xiang, wrote, “If [we] split [all the property] under Heaven into six shares, five shares are spent on the military, and one share is spent on temple offerings and state expenses. How can the country not be poor and the people not in difficulty?”50

    The same situation prevailed in many other polities, from ancient times to the modern era. The Roman Empire spent about 50–75 percent of its budget on the military,51 and the figure was about 60 percent in the late seventeenth-century Ottoman Empire.52 Between 1685 and 1813 the share of the military in British government expenditure averaged 75 percent.53 In France, military expenditure between 1630 and 1659 varied between 89 percent and 93 percent of the budget, remained above 30 percent for much of the eighteenth century, and dropped to a low of 25 percent in 1788 only due to the financial crisis that led to the French Revolution. In Prussia, from 1711 to 1800 the military share of the budget never fell below 75 percent and occasionally reached as high as 91 percent.54 During the relatively peaceful years of 1870–1913, the military ate up an average of 30 percent of the state budgets of the major powers of Europe, as well as Japan and the United States, while smaller powers like Sweden were spending even more.55 When war broke out in 1914, military budges skyrocketed. During their involvement in World War I, French military expenditure averaged 77 percent of the budget; in Germany it was 91 percent, in Russia 48 percent, in the U.K. 49 percent, and in the United States 47 percent. During World War II, the U.K. figure rose to 69 percent and the U.S. figure to 71 percent.56 Even during the détente years of the 1970s, Soviet military expenditure still amounted to 32.5 percent of the budget.57

    State budgets in more recent decades make for far more hopeful reading material than any pacifist tract ever composed. In the early twenty-first century, the worldwide average government expenditure on the military has been only around 7 percent of the budget, and even the dominant superpower of the United States spent only around 13 percent of its annual budget to maintain its military hegemony.58 Since most people no longer lived in terror of external invasion, governments could invest far more money in welfare, education, and health care. Worldwide average expenditure on health care in the early twenty-first century has been about 10 percent of the government budget, or about 1.4 times the defense budget.59 For many people in the 2010s, the fact that the health-care budget was bigger than the military budget was unremarkable. But it was the result of a major change in human behavior, and one that would have sounded impossible to most previous generations.

    The decline of war didn’t result from a divine miracle or from a metamorphosis in the laws of nature. It resulted from humans changing their own laws, myths, and institutions and making better decisions. Unfortunately, the fact that this change has stemmed from human choice also means that it is reversible. Technology, economics, and culture are ever changing. In the early 2020s, more leaders are again dreaming of martial glory, armed conflicts are on the rise,60 and military budgets are increasing.61

    A critical threshold was crossed in early 2022. Russia had already destabilized the global order by mounting a limited invasion of Ukraine in 2014 and occupying Crimea and other regions in eastern Ukraine. But on February 24, 2022, Vladimir Putin launched an all-out assault aimed to conquer the whole of Ukraine and extinguish Ukrainian nationhood. To prepare and sustain this attack, Russia increased its military budget far beyond the global average of 7 percent. Exact figures are difficult to determine, because many aspects of the Russian military budget are shrouded in secrecy, but the best estimates put the figure somewhere in the vicinity of 30 percent, and it may even be higher.62 The Russian onslaught in turn has forced not only Ukraine but also many other European nations to increase their own military budgets.63 The reemergence of militaristic cultures in places like Russia, and the development of unprecedented cyber weapons and autonomous armaments throughout the world, could result in a new era of war, worse than anything we have seen before.

    The decisions leaders like Putin make on matters of war and peace are shaped by their understanding of history. Which means that just as overly optimistic views of history could be dangerous illusions, overly pessimistic views could become destructive self-fulfilling prophecies. Prior to his all-out 2022 attack on Ukraine, Putin had often expressed his historical conviction that Russia is trapped in an endless struggle with foreign enemies, and that the Ukrainian nation is a fabrication by these enemies. In June 2021, he published a fifty-three-hundred-word essay titled “On the Historical Unity of Russians and Ukrainians” in which he denied the existence of Ukraine as a nation and argued that foreign powers have repeatedly tried to weaken Russia by fostering Ukrainian separatism. While professional historians reject these claims, Putin seems to genuinely believe in this historical narrative.64 Putin’s historical convictions led him in 2022 to prioritize the conquest of Ukraine over other policy goals, such as providing Russian citizens with better health care or spearheading a global initiative to regulate AI.65

    If leaders like Putin believe that humanity is trapped in an unforgiving dog-eat-dog world, that no profound change is possible in this sorry state of affairs, and that the relative peace of the late twentieth century and early twenty-first century was an illusion, then the only choice remaining is whether to play the part of predator or prey. Given such a choice, most leaders would prefer to go down in history as predators and add their names to the grim list of conquerors that unfortunate pupils are condemned to memorize for their history exams. These leaders should be reminded, however, that in the era of AI the alpha predator is likely to be AI.

    Perhaps, though, we have more choices available to us. I cannot predict what decisions people will make in the coming years, but as a historian I do believe in the possibility of change. One of the chief lessons of history is that many of the things that we consider natural and eternal are, in fact, man-made and mutable. Accepting that conflict is not inevitable, however, should not make us complacent. Just the opposite. It places a heavy responsibility on all of us to make good choices. It implies that if human civilization is consumed by conflict, we cannot blame it on any law of nature or any alien technology. It also implies that if we make the effort, we can create a better world. This isn’t naïveté; it’s realism. Every old thing was once new. The only constant of history is change.

    Epilogue

    In late 2016, a few months after AlphaGo defeated Lee Sedol and as Facebook algorithms were stoking dangerous racist sentiments in Myanmar, I published Homo Deus. Though my academic training had been in medieval and early modern military history, and though I have no background in the technical aspects of computer science, I suddenly found myself, post-publication, with the reputation of an AI expert. This opened the doors to the offices of scientists, entrepreneurs, and world leaders interested in AI and afforded me a fascinating, privileged look into the complex dynamics of the AI revolution.

    It turned out that my previous experience researching topics such as English strategy in the Hundred Years’ War and studying paintings from the Thirty Years’ War1 wasn’t entirely unrelated to this new field. In fact, it gave me a rather unique historical perspective on the events unfolding rapidly in AI labs, corporate offices, military headquarters, and presidential palaces. Over the past eight years I have had numerous public and private discussions about AI, particularly about the dangers it poses, and with each passing year the tone has become more urgent. Conversations that in 2016 felt like idle philosophical speculations about a distant future had, by 2024, acquired the focused intensity of an emergency room.

    I am neither a politician nor a businessperson and have little talent for what these vocations demand. But I do believe that an understanding of history can be useful in gaining a better grasp of present-day technological, economic, and cultural developments—and, more urgently, in changing our political priorities. Politics is largely a matter of priorities. Should we cut the health care budget and spend more on defense? Is our more pressing security threat terrorism or climate change? Do we focus on regaining a lost patch of ancestral territory or concentrate on creating a common economic zone with the neighbors? Priorities determine how citizens vote, what businesspeople are concerned about, and how politicians try to make a name for themselves. And priorities are often shaped by our understanding of history.

    While so-called realists dismiss historical narratives as propaganda ploys deployed to advance state interests, in fact it is these narratives that define state interests in the first place. As we saw in our discussion of Clausewitz’s theory of war, there is no rational way to define ultimate goals. The state interests of Russia, Israel, Myanmar, or any other country can never be deduced from some mathematical or physical equation; they are always the supposed moral of a historical narrative.

    It is therefore hardly surprising that politicians all over the world spend a lot of time and effort recounting historical narratives. The above-mentioned example of Vladimir Putin is hardly exceptional in this respect. In 2005 the UN secretary-general, Kofi Annan, had his first meeting with General Than Shwe, the then dictator of Myanmar. Annan was advised to speak first, so as to prevent the general from monopolizing the conversation, which was meant to last only twenty minutes. But Than Shwe struck first and held forth for nearly an hour on the history of Myanmar, hardly giving the UN secretary-general any chance to speak.2 In May 2011 the Israeli prime minister, Benjamin Netanyahu did something similar in the White House, when he met the U.S. president, Barack Obama. After Obama’s brief introductory remarks, Netanyahu subjected the president to a long lecture about the history of Israel and the Jewish people, treating Obama as if he were his student.3 Cynics might argue that Than Shwe and Netanyahu hardly cared about the facts of history and were deliberately distorting them in order to achieve some political goal. But these political goals were themselves the product of deeply held convictions about history.

    In my own conversations on AI with politicians, as well as tech entrepreneurs, history has often emerged as a central theme. Some of my interlocutors painted a rosy picture of history and were accordingly enthusiastic about AI. They argued that more information has always meant more knowledge and that by increasing our knowledge, every previous information revolution has greatly benefited humankind. Didn’t the print revolution lead to the scientific revolution? Didn’t newspapers and radio lead to the rise of modern democracy? The same, they said, would happen with AI. Others had a dimmer perspective, but nevertheless expressed hope that humankind will somehow muddle through the AI revolution, just as we muddled through the Industrial Revolution.

    Neither view offered me much solace. For reasons explained in previous chapters, I find such historical comparisons to the print revolution and the Industrial Revolution distressing, especially coming from people in positions of power, whose historical vision is informing the decisions that shape our future. These historical comparisons underestimate both the unprecedented nature of the AI revolution and the negative aspects of previous revolutions. The immediate results of the print revolution included witch hunts and religious wars alongside scientific discoveries, while newspapers and radio were exploited by totalitarian regimes as well as by democracies. As for the Industrial Revolution, adapting to it involved catastrophic experiments such as imperialism and Nazism. If the AI revolution leads us to similar kinds of experiments, can we really be certain we will muddle through again?

    My goal with this book is to provide a more accurate historical perspective on the AI revolution. This revolution is still in its infancy, and it is notoriously difficult to understand momentous developments in real time. It is hard, even now, to assess the meaning of events in the 2010s like AlphaGo’s victory or Facebook’s involvement in the anti-Rohingya campaign. The meaning of events of the early 2020s is even more obscure. Yet by expanding our horizons to look at how information networks developed over thousands of years, I believe it is possible to gain some insight on what we’re living through today.

    One lesson is that the invention of new information technology is always a catalyst for major historical changes, because the most important role of information is to weave new networks rather than represent preexisting realities. By recording tax payments, clay tablets in ancient Mesopotamia helped forge the first city-states. By canonizing prophetic visions, holy books spread new kinds of religions. By swiftly disseminating the words of presidents and citizens, newspapers and telegraphs opened the door to both large-scale democracy and large-scale totalitarianism. The information thus recorded and distributed was sometimes true, often false, but it invariably created new connections between larger numbers of people.

    We are used to giving political, ideological, and economic interpretations to historical revolutions such as the rise of the first Mesopotamian city-states, the spread of Christianity, the American Revolution, and the Bolshevik Revolution. But to gain a deeper understanding, we should also view them as revolutions in the way information flows. Christianity was obviously different from Greek polytheism in many of its myths and rites, yet it was also different in the importance it gave to a single holy book and the institution entrusted with interpreting it. Consequently, whereas each temple of Zeus was a separate entity, each Christian church became a node in a unified network.4 Information flowed differently among the followers of Christ than among the worshippers of Zeus. Similarly, Stalin’s U.S.S.R. was a different kind of information network from Peter the Great’s empire. Stalin enacted many unprecedented economic policies, but what enabled him to do it is that he headed a totalitarian network in which the center accumulated enough information to micromanage the lives of hundreds of millions of people. Technology is rarely deterministic, and the same technology can be used in very different ways. But without the invention of technologies like the book and the telegraph, the Christian Church and the Stalinist apparatus would never have been possible.

    This historical lesson should strongly encourage us to pay more attention to the AI revolution in our current political debates. The invention of AI is potentially more momentous than the invention of the telegraph, the printing press, or even writing, because AI is the first tool that is capable of making decisions and generating ideas by itself. Whereas printing presses and parchment scrolls offered new means for connecting people, AIs are full-fledged members in our information networks. In coming years, all information networks—from armies to religions—will gain millions of new AI members, who will process data very differently than humans. These new members will make alien decisions and generate alien ideas—that is, decisions and ideas that are unlikely to occur to humans. The addition of so many alien members is bound to change the shape of armies, religions, markets, and nations. Entire political, economic, and social systems might collapse, and new ones will take their place. That’s why AI should be a matter of utmost urgency even to people who don’t care about technology and who think the most important political questions concern the survival of democracy or the fair distribution of wealth.

    This book has juxtaposed the discussion of AI with the discussion of sacred canons like the Bible, because we are now at the critical moment of AI canonization. When church fathers like Bishop Athanasius decided to include 1 Timothy in the biblical dataset while excluding the Acts of Paul and Thecla, they shaped the world for millennia. Billions of Christians down to the twenty-first century have formed their views of the world based on the misogynist ideas of 1 Timothy rather than on the more tolerant attitude of Thecla. Even today it is difficult to reverse course, because the church fathers chose not to include any self-correcting mechanisms in the Bible. The present-day equivalents of Bishop Athanasius are the engineers who write the initial code for AI, and who choose the dataset on which the baby AI is trained. As AI grows in power and authority, and perhaps becomes a self-interpreting holy book, so the decisions made by present-day engineers could reverberate down the ages.

    Studying history does more than just emphasize the importance of the AI revolution and of our decisions regarding AI. It also cautions us against two common but misleading approaches to information networks and information revolutions. On the one hand, we should beware of an overly naive and optimistic view. Information isn’t truth. Its main task is to connect rather than represent, and information networks throughout history have often privileged order over truth. Tax records, holy books, political manifestos, and secret police files can be extremely efficient in creating powerful states and churches, which hold a distorted view of the world and are prone to abuse their power. More information, ironically, can sometimes result in more witch hunts.

    There is no reason to expect that AI would necessarily break the pattern and privilege truth. AI is not infallible. What little historical perspective we have gained from the alarming events in Myanmar, Brazil, and elsewhere over the past decade indicates that in the absence of strong self-correcting mechanisms AIs are more than capable of promoting distorted worldviews, enabling egregious abuses of power, and instigating terrifying new witch hunts.

    On the other hand, we should also beware of swinging too far in the other direction and adopting an overly cynical view. Populists tell us that power is the only reality, that all human interactions are power struggles, and that information is merely a weapon we use to vanquish our enemies. This has never been the case, and there is no reason to think that AI will make it so in the future. While many information networks do privilege order over truth, no network can survive if it ignores truth completely. As for individual humans, we tend to be genuinely interested in truth rather than only in power. Even institutions like the Spanish Inquisition have had conscientious truth-seeking members like Alonso de Salazar Frías, who, instead of sending innocent people to their deaths, risked his life to remind us that witches are just intersubjective fictions. Most people don’t view themselves as one-dimensional creatures obsessed solely with power. Why, then, hold such a view about everyone else?

    Refusing to reduce all human interactions to a zero-sum power struggle is crucial not just for gaining a fuller, more nuanced understanding of the past but also for having a more hopeful and constructive attitude about our future. If power were the only reality, then the only way to resolve conflicts would be through violence. Both populists and Marxists believe that people’s views are determined by their privileges, and that to change people’s views it is necessary to first take away their privileges—which usually requires force. However, since humans are interested in truth, there is a chance to resolve at least some conflicts peacefully, by talking to one another, acknowledging mistakes, embracing new ideas, and revising the stories we believe. That is the basic assumption of democratic networks and of scientific institutions. It has also been the basic motivation behind writing this book.

    EXTINCTION OF THE SMARTEST

    Let’s return now to the question I posed at the beginning of this book: If we are so wise, why are we so self-destructive? We are at one and the same time both the smartest and the stupidest animals on earth. We are so smart that we can produce nuclear missiles and superintelligent algorithms. And we are so stupid that we go ahead producing these things even though we’re not sure we can control them and failing to do so could destroy us. Why do we do it? Does something in our nature compel us to go down the path of self-destruction?

    This book has argued that the fault isn’t with our nature but with our information networks. Due to the privileging of order over truth, human information networks have often produced a lot of power but little wisdom. For example, Nazi Germany created a highly efficient military machine and placed it at the service of an insane mythology. The result was misery on an enormous scale, the death of tens of millions of people, and eventually the destruction of Nazi Germany, too.

    Of course, power is not in itself bad. When used wisely, it can be an instrument of benevolence. Modern civilization, for example, has acquired the power to prevent famines, contain epidemics, and mitigate natural disasters such as hurricanes and earthquakes. In general, the acquisition of power allows a network to deal more effectively with threats coming from outside, but simultaneously increases the dangers that the network poses to itself. It is particularly noteworthy that as a network becomes more powerful, imaginary terrors that exist only in the stories the network itself invents become potentially more dangerous than natural disasters. A modern state faced with drought or excessive rains can usually prevent this natural disaster from causing mass starvation among its citizens. But a modern state gripped by a man-made fantasy is capable of instigating man-made famines on an enormous scale, as happened in the U.S.S.R. in the early 1930s.

    Accordingly, as a network becomes more powerful, its self-correcting mechanisms become more vital. If a Stone Age tribe or a Bronze Age city-state was incapable of identifying and correcting its own mistakes, the potential damage was limited. At most, one city was destroyed, and the survivors tried again elsewhere. Even if the ruler of an Iron Age empire, such as Tiberius or Nero, was gripped by paranoia or psychosis, the consequences were seldom catastrophic. The Roman Empire endured for centuries despite its fair share of mad emperors, and its eventual collapse did not bring about the end of human civilization. But if a Silicon Age superpower has weak or nonexistent self-correcting mechanisms, it could very well endanger the survival of our species, and countless other life-forms, too. In the era of AI, the whole of humankind finds itself in an analogous situation to Tiberius in his Capri villa. We command immense power and enjoy rare luxuries, but we are easily manipulated by our own creations, and by the time we wake up to the danger, it might be too late.

    Unfortunately, despite the importance of self-correcting mechanisms for the long-term welfare of humanity, politicians might be tempted to weaken them. As we have seen throughout the book, though neutralizing self-correcting mechanisms has many downsides, it can nevertheless be a winning political strategy. It could deliver immense power into the hands of a twenty-first-century Stalin, and it would be foolhardy to assume that an AI-enhanced totalitarian regime would necessarily self-destruct before it could wreak havoc on human civilization. Just as the law of the jungle is a myth, so also is the idea that the arc of history bends toward justice. History is a radically open arc, one that can bend in many directions and reach very different destinations. Even if Homo sapiens destroys itself, the universe will keep going about its business as usual. It took four billion years for terrestrial evolution to produce a civilization of highly intelligent apes. If we are gone, and it takes evolution another hundred million years to produce a civilization of highly intelligent rats, it will. The universe is patient.

    There is, though, an even worse scenario. As far as we know today, apes, rats, and the other organic animals of planet Earth may be the only conscious entities in the entire universe. We have now created a nonconscious but very powerful alien intelligence. If we mishandle it, AI might extinguish not only the human dominion on Earth but the light of consciousness itself, turning the universe into a realm of utter darkness. It is our responsibility to prevent this.

    The good news is that if we eschew complacency and despair, we are capable of creating balanced information networks that will keep their own power in check. Doing so is not a matter of inventing another miracle technology or landing upon some brilliant idea that has somehow escaped all previous generations. Rather, to create wiser networks, we must abandon both the naive and the populist views of information, put aside our fantasies of infallibility, and commit ourselves to the hard and rather mundane work of building institutions with strong self-correcting mechanisms. That is perhaps the most important takeaway this book has to offer.

    This wisdom is much older than human history. It is elemental, the foundation of organic life. The first organisms weren’t created by some infallible genius or god. They emerged through an intricate process of trial and error. Over four billion years, ever more complex mechanisms of mutation and self-correction led to the evolution of trees, dinosaurs, jungles, and eventually humans. Now we have summoned an alien inorganic intelligence that could escape our control and put in danger not just our own species but countless other life-forms. The decisions we all make in the coming years will determine whether summoning this alien intelligence proves to be a terminal error or the beginning of a hopeful new chapter in the evolution of life.