This is a work of delightfully rampant interdisciplinarity. It draws “on evolutionary analysis, primatology, anthropology, moral psychology, analyses of complex systems, experimental economics, studies of norms, economic development, policy studies, analyses of governance and collective action, randomized control trials and much more” (ix). I am not expert in any of those fields, much less all of them. I said as much at a conference commenting on earlier versions of this material before Gaus’s untimely death. I went on to add, “Of course, neither is Gaus—but he has delved more deeply into them than most of us.” Even those who reject his conclusions will return time and again to Gaus’s book for its wonderful introductions to all of those literatures.Officially, the book is organized around “three unsettling theses” from Hayek. But those—and frequent asides like “Hayek’s prescience is striking” (123), “about this Hayek was undoubtedly correct” (132), and “I … dispute Hayek’s claim” (14)—are ignorable distractions. As Gaus himself says in the preface, “It ultimately does not much matter what Hayek said” (ix). Readers who wish he had hewn to that line more closely and abandoned more thoroughly his “original exegetical aims” (xi) can read around the recurring hommage à Hayek without loss.More awkwardly, the social order the book advocates—the Open Society (always capitalized, as if the proper name of a religion or political party)—is never formally defined. It is loosely characterized variously as an evolved system of “increasingly impartial and wide-ranging … rules and norms” (215) and “an ever-increasing and relentless engine of diversity and inclusivity” (248) in which “niche creation and innovation manifest themselves everywhere” (248). Most fundamentally, it is a “form of self-organization growing out of our evolved cooperative and egalitarian natures” (14). The Open Society resists attempts at intentional guidance or control, except occasionally by innovators just at the margins—Gaus’s “adjacent possible” (113–16). The Open Society rests on a remarkably thin set of social norms and rules, and inevitably so because its stability depends on literally everyone in it being able to endorse those from their own diverse perspectives.Is the Open Society an ideal to be sought or an inevitability to be accommodated? Gaus’s celebratory language might suggest the former. But Gaus explicitly eschews the exploration of “moral ideals and principles”: “I have nothing to say here about [those]” (89). Instead he provides an evolutionary account showing that the Open Society has advantages over alternative arrangements that will cause it to prevail. Gaus’s long-standing advocacy of “public justification” reappears here, not as a normative aspiration but rather as a stability condition for evolved social systems (48–52, 162). It is part of the Complex Society Moral Package that, sociologically, people must internalize if the system is to be stable—whatever its merits from the perspective of critical morality.The features of the Open Society that give it an edge in cultural evolution have to do with the complexity and diversity that it fosters. Complexity is associated with increased specialization and division of labor and the efficiency gains that come with that (think of Adam Smith’s pin factory). Diversity, in both producers’ skill sets and consumers’ preferences, increases the scope for mutually advantageous gains from trade, economically and otherwise.Those evolutionary advantages do not make the Open Society literally inevitable, however. Gaus acknowledges (against social planners, he insists) that there can and typically will be multiple equilibria in evolutionary dynamics (120). Once we settle into one of those equilibria it will take a major exogenous shock to get us out of it. Whether we then shift to the Open Society equilibrium remains indeterminate, even if that equilibrium is the best for all concerned. If we do happen into the Open Society equilibrium, rather more of a shock may be required to get us out of it (124–28). Still, getting there, and staying there, is far from guaranteed. Moreover, as Gaus himself acknowledges (123), group-level macro-selection—the prime mechanism of cultural evolution—is unlikely in developed societies, where it is rare for one group to push another aside completely.At this point, Gaus’s argument comes to rest on his argument for the inevitability of an evolutionarily hardwired cooperative norm of “strong reciprocity” being internalized by social actors (37–43). He vacillates on whether the evolution in question is genetic (38) or cultural (65). But virtually all his discussion treats it as cultural, and he gives no grounds for thinking there is any genetic component.The Folk Theorem to which Gaus alludes (38) shows there to be multiple equilibria in repeated games. Maybe conditional cooperation is the best equilibrium for all concerned. But again, there is no guarantee that we will necessarily get there. There is a lot of evidence in today’s experimental economics, as well as a long line of work in anthropology, prehistory, and so on, that people will bear personal costs to punish norm breakers in some (even many) societies—but not in others (86). There are communities where, shown a Prisoner’s Dilemma, nonkin will defect every time (Turnbull 1972; Banfield 1958). That may not be the best way for the group as a whole to prosper. But once they have fallen into that equilibrium, it is hard for them to get out of it.Gaus assumes that people who are “strong reciprocators” will, ipso facto, be “constantly on the lookout for cooperative possibilities” for mutually beneficial interactions (171). That is why they will seek out new market niches, thus fostering the ever-increasing complexity and diversity characterizing the Open Society. But being a reciprocator, even a “strong reciprocator,” just commits you to reciprocating when provoked; it does not commit you to seeking out provocations. “Strong reciprocators,” by definition, engage in “negative reciprocation” whenever the situation calls for it. But since punishing others’ bad behavior is costly, even they would more naturally strive to avoid such situations if they can. Maybe they would be more likely to seek out situations calling for “positive reciprocation,” for in those situations they or those they care about would have been done a good turn by someone else. But it is the good turn that they seek out, not the opportunity-cum-duty to reciprocate as such.People can and arguably do practice reciprocity for purely strategic reasons rather than as the reflection of Gaus’s strongly internalized social norms. In recurring interactions with others who are reciprocators, reciprocating is the strictly dominant strategy. But there are then big implications—and ones contrary to Gaus’s hyper-inclusive Open Society—for people’s selection of with whom to interact (on reciprocal terms, anyway). First, strategic reciprocators would choose to interact (reciprocally, anyway) only with others who are themselves reciprocators (Spiekermann 2009). People would exclude from the circle of reciprocity those who do not play by those rules. Second, strategic reciprocators would exclude from that circle those from whom they have nothing either to gain or to fear. If I have nothing to gain from your reciprocating a favor that I do you, then I have no strategic reason to do you the favor. If I have nothing to fear from your reciprocating any wrong I do you, then I have no strategic reason not to do you that wrong. Even “strong reciprocators” would have nothing to reciprocate, in those cases (unless what they are reciprocating are dispositions rather than actions [43]). Thus, certainly with strategic replicators and perhaps even with strong reciprocators, we end up with “clubbish justice” (Goodin 2008). The strong club together and exclude the weak. The benefits that Gaus traces to the Open Society flow to those inside the club, but not beyond. If that sounds like a left-wing caricature, recall that “club goods”—goods that are public goods for people inside the club but from whom outsiders are excluded—were the invention of James Buchanan (1965), a staunchly right-wing Public Choice Nobel Laureate economist.There is a glimmer of recognition of that prospect in Gaus’s remark that “a critical task of the democratic order is to ensure the equality and fairness on which large-scale human cooperation depends. Without these rules of the game, self-organization can lead to oppressive hierarchies” (245). But in addressing governance issues in the last part of the book, Gaus comes down strongly in favor of very localized “self-organization” with minimal central control (172). “Of course … legal and political institutions are necessary for innumerable aspects of large-scale cooperation” (173); but since any rules must be justifiable to all, “there will not be society-wide convergence on many moral [or legal] rules” (160). Gaus’s Open Society needs Nozick’s minimal state—property, contract, and tort law (128–39, 195). But it is not clear that it can stretch to much more, given the public justification rule needed to stabilize it and the autocatalytic processes that drives it.Everything therefore hinges on Gaus’s “evolutionary” story about human psychology, that internalized social norms inevitably induce us to be “strong dispositional reciprocators.” But insofar as the “evolution“ in question is cultural, not biological, the key transmission mechanism is not genetic but rather the social learning of handed-down social norms (65). The evolution is Lamarckian rather than Darwinian. That matters, because what has been learned can be lost or repudiated in a single generation rather than, as with genetic evolution, over the course of many. Reflect, for example, on the precipitous decline of norms of comity in the US Senate (Uslaner 1993; Shapiro 2018). But now recall what, on Gaus’s account, underwrites the stable Lamarckian transmission of the relevant acquired characteristics. That was supposed to be the cooperative social norm—which is itself precisely the acquired characteristic whose stable transmission is supposed to be being thereby underwritten. Gaus’s model thus swallows its own tail.Hence, I do not believe that Gaus’s ambitious positive program works in the end. But like Roman villas, its building blocks are endlessly recyclable by others wanting to put them to very different uses. Gaus’s astute summaries of so many disparate literatures are truly a major gift that he has left to the philosophical community.

中文翻译：

---

开放社会及其复杂性

这是一部令人愉快的跨学科作品。它借鉴了“进化分析、灵长类动物学、人类学、道德心理学、复杂系统分析、实验经济学、规范研究、经济发展、政策研究、治理和集体行动分析、随机对照试验等等”(ix)。我不是这些领域的专家，更不用说所有领域了。在 Gaus 英年早逝之前，我在一次评论该材料早期版本的会议上也说过同样的话。我继续补充说，“当然，高斯也不是——但他比我们大多数人更深入地研究了它们。” 即使那些拒绝接受他的结论的人也会一次又一次地回到高斯的书中，因为它对所有这些文献进行了精彩的介绍。正式地说，这本书是围绕哈耶克的“三个令人不安的论点”组织的。但是那些——以及像“哈耶克的先见之明是惊人的”（123）、“关于这一点哈耶克无疑是正确的”（132）和“我……质疑哈耶克的主张”（14）这样的频繁旁白——是可以忽略的干扰。正如高斯自己在序言中所说，“哈耶克说什么最终并不重要” (ix)。希望他更严格地遵循那条线并更彻底地放弃他的“原始解释目标”（xi）的读者可以毫无损失地阅读哈耶克反复出现的敬意。更尴尬的是，这本书倡导的社会秩序——开放社会（总是大写，就好像宗教或政党的专有名称一样）——从未被正式定义过。它被粗略地描述为一个“越来越公正和广泛的......规则和规范”（215）和“不断增长和不懈的多样性和包容性引擎”（248）的进化系统，其中“利基创造和创新体现自己无处不在”（248）。最根本的是，它是一种“从我们进化的合作和平等本性中产生的自组织形式”(14)。开放社会抵制有意引导或控制的尝试，除了偶尔处于边缘的创新者——高斯的“相邻可能”（113-16）。开放社会建立在一套非常薄的社会规范和规则之上，这是不可避免的，因为它的稳定性实际上取决于其中的每个人都能够从自己不同的角度认可这些规范和规则。开放社会是需要寻求的理想还是需要适应的必然性？Gaus 的庆祝语言可能暗示前者。但高斯明确回避对“道德理想和原则”的探索：“我在这里对[那些]无话可说”（89）。相反，他提供了一个演化说明，表明开放社会比其他安排具有优势，这将使其占上风。高斯长期以来对“公共辩护”的提倡在这里再次出现，不是作为规范的愿望，而是作为进化社会系统的稳定条件 (48–52, 162)。它是复杂社会道德包的一部分，从社会学上讲，如果系统要稳定，人们就必须内化——无论从批判道德的角度来看它有什么优点。开放社会的特征使其在文化进化中具有优势，这与它所培养的复杂性和多样性有关。复杂性与专业化和劳动分工的增加以及随之而来的效率提升有关（想想亚当·斯密的别针工厂）。生产者技能组合和消费者偏好的多样性增加了从贸易中获得互利收益的范围，无论是在经济上还是在其他方面。然而，这些进化优势并没有使开放社会真正不可避免。Gaus 承认（反对社会计划者，他坚持认为）进化动力学中可以而且通常会存在多重均衡 (120)。一旦我们进入其中一个均衡状态，就需要一个重大的外生冲击才能让我们摆脱它。我们是否会转向开放社会均衡仍不确定，即使该均衡对所有相关方而言都是最佳的。如果我们确实碰巧进入了开放社会的均衡状态，那么可能需要更多的冲击才能让我们摆脱它 (124-28)。尽管如此，到达那里并留在那里还远未得到保证。此外，正如 Gaus 自己承认的那样 (123)，群体层面的宏观选择——文化进化的主要机制——在发达社会中不太可能发生，在发达社会中，一个群体很少会完全排斥另一个群体。在这一点上，Gaus 的论点来了以他的论点为基础，即“强互惠”的进化硬连线合作规范不可避免地被社会行动者内化 (37-43)。他对所讨论的进化是遗传的 (38) 还是文化的 (65) 摇摆不定。但实际上他的所有讨论都将其视为文化因素，他没有给出任何理由认为存在任何遗传因素。Gaus 提到的大众定理 (38) 表明在重复博弈中存在多重均衡。也许有条件的合作是对所有相关方最好的平衡。但同样，不能保证我们一定会到达那里。当今的实验经济学以及人类学、史前史等领域的大量研究表明，在某些（甚至许多）社会中，人们将承担惩罚违反规范者的个人成本——但在其他社会中则不会。 (86)。在某些社区中，表现出囚徒困境，nonkin 每次都会背叛（Turnbull 1972；Banfield 1958）。这可能不是整个集团繁荣的最佳方式。但一旦他们陷入那种平衡，他们很难摆脱困境。Gaus 假设“强互惠者”的人事实上会“不断寻找合作的可能性”以实现互惠互利的互动 (171)。这就是为什么他们将寻求新的市场利基，从而促进不断增加的复杂性和多样性，这是开放社会的特征。但是，作为一个互惠者，即使是“强烈的互惠者”，也只会让你在被激怒时做出回报；它不会让您承诺寻求挑衅。根据定义，“强互惠者”会在情况需要时进行“负互惠”。但由于惩罚他人的不良行为代价高昂，如果可以的话，即使是他们也会更自然地努力避免这种情况。也许他们更有可能寻找需要“积极的回报，” 因为在那些情况下，他们或他们关心的人本来会被其他人做好事。但他们寻求的是良好的转变，而不是机会和义务的回报。人们可以而且可以说是出于纯粹的战略原因而不是作为高斯强烈内化的社会规范的反映。在与其他互惠者的反复互动中，互惠是严格占优的策略。但是，对于人们选择与谁互动（无论如何，在互惠条件下），这有很大的影响——并且与高斯的高度包容的开放社会相反。首先，战略互惠者会选择只与本身是互惠者的其他人互动（无论如何互惠）（Spiekermann 2009）。人们会将那些不遵守这些规则的人排除在互惠圈之外。其次，战略互惠者会将那些他们既无所获也无所畏惧的人排除在这个圈子之外。如果我从你回报我对你的帮助中没有任何好处，那么我就没有战略理由来帮助你。如果我不害怕你回报我对你的任何错误，那么我就没有战略理由不对你做错事。在那些情况下，即使是“强互惠者”也没有什么可以互惠的（除非他们互惠的是倾向而不是行动 [43]）。因此，当然有战略复制者，甚至可能有强大的互惠者，我们最终会得到“俱乐部正义”（Goodin 2008）。强者齐聚，弱者排斥。Gaus 追溯到开放社会的好处流向了俱乐部内部的人，但没有流向俱乐部以外的人。如果这听起来像是左翼漫画，请回想一下“俱乐部商品”——俱乐部内部人员的公共物品，但外部人员被排除在外的物品——是詹姆斯·布坎南 (James Buchanan) (1965) 的发明，他是坚定的右翼公众人物精选诺贝尔经济学奖获得者。高斯的评论中有一丝对这一前景的认可，“民主秩序的一项关键任务是确保大规模人类合作所依赖的平等和公平。没有这些游戏规则，自组织会导致压迫性的等级制度”（245）。但在本书最后一部分解决治理问题时，Gaus 强烈支持具有最小中央控制的非常本地化的“自组织”（172）。“当然……法律和政治制度对于大规模合作的无数​​方面都是必要的”（173）；但由于任何规则都必须对所有人合理，“许多道德 [或法律] 规则不会在整个社会范围内趋同”（160）。高斯的开放社会需要诺齐克的最小状态——财产法、合同法和侵权法 (128–39, 195)。但目前尚不清楚它是否可以扩展到更多，因为它需要稳定它的公共理由规则和驱动它的自动催化过程。因此，一切都取决于高斯关于人类心理学的“进化”故事，即内化的社会规范不可避免地诱导我们成为“强烈的性格互惠者”。但只要所讨论的“进化”是文化的，而不是生物的，关键的传播机制不是遗传，而是世代相传的社会规范的社会学习 (65)。进化是拉马克式的而不是达尔文式的。这很重要，因为所学到的东西可能会在一代人的时间内丢失或被否定，而不是像遗传进化那样，在许多人的过程中。例如，反思一下美国参议院礼让规范的急剧下降（Uslaner 1993；Shapiro 2018）。但现在回想一下，根据高斯的说法，是什么保证了相关后天特征的稳定拉马克传递。这应该是合作的社会规范——它本身恰恰是后天获得的特征，其稳定的传播应该因此得到保证。高斯的模型因此吞下了自己的尾巴。因此，我不相信 Gaus 雄心勃勃的积极计划最终不会奏效。但就像罗马别墅一样，它的建筑模块可以被其他想要将它们用于不同用途的人无休止地回收利用。高斯对众多不同文献的敏锐总结确实是他留给哲学界的一份重要礼物。