读《Nature》第425期 (2003年) 论文：The nature of human altruism (人类利他主义的本质) Ernst Fehr & Urs Fischbacher
利维坦不需要自上而下地建立，它内化在每个“强互惠者”的基因与文化直觉中。
传统的生物学用汉密尔顿法则解释亲属间的利他行为，但无法解释陌生人之间的合作。论文指出，“利他性惩罚”和“利他性奖励”才是维系庞大人类社群的基石。当你为了惩罚一个不公平的分配者而宁愿自己也遭受经济损失时，你正在执行一种微观的、分布式的“社会契约”。
人类极度关注个人声誉的建立，声誉机制能极大提升合作率。但在今天，科技将“声誉”和“第三方惩罚”剥离了人类的生物学直觉，交由算法来量化。系统越是通过算法强制保证合作，人与人之间基于人性的真实信任感就越薄弱。
论文在探讨终极起源时，将答案落脚于基因-文化协同进化。文化规范，如通过社会化学习到的惩罚和奖励机制，重塑了我们的演化环境，使得具备合作倾向的基因得以保留。但是，如今主导文化传播的，不再是部落长者或村落领袖，而是推荐算法。如果我们的文化演化方向开始由追求用户粘性和商业转化的硅基逻辑所引导，人类的“强互惠”本能会被如何利用？
人类在实现相互合作时，多巴胺系统会被激活，但是科技正在劫持这个回路，社交媒体上的点赞就是一种廉价的“利他性奖励”，它用多巴胺的即时满足，替代了现实中建立深度合作的艰辛。长此以往，人类原本用于维系宏大共同体的利他惩罚与奖励机制，会不会退化为在虚拟世界中无休止的情绪发泄？
少数的强互惠者可以迫使多数自私者合作。但在现代科技放大了自私者的隐蔽性，比如匿名，同时又将惩罚权力交给平台算法。那么，当一个社会的合作不再依赖于公民内心的“不公平厌恶”，而是依赖于全天候的监控探头、人脸识别和智能合约的自动执行时，人类还算是一个“利他”的物种吗？如果作恶的成本被技术无限拔高以至于无人作恶，那么“善良”的道德价值是否也就随之消解了？
人类的利他主义跨越了冰河时期的严寒，走出了非洲大草原。这种力量极其强大，但也极其脆弱。
——
Reflections after reading: “The Nature of Human Altruism” by Ernst Fehr & Urs Fischbacher, Nature, Vol. 425 (2003).
In the history of thought, Thomas Hobbes pessimistically argued that humanity in its natural state is a “war of all against all,” necessitating an absolute Leviathan to maintain order. However, the paper ‘The Nature of Human Altruism’ offers a different answer through the Ultimatum Game and third-party punishment experiments: the Leviathan does not need to be established from the top down; it is internalized within the genetic and cultural intuition of every “strong reciprocator.”
Traditional biology uses Hamilton’s rule to explain altruistic behavior among kin, but it fails to explain cooperation among strangers. The paper points out that “altruistic punishment” and “altruistic rewarding” are the true cornerstones sustaining large-scale human communities. When you are willing to incur a personal economic loss to punish an unfair allocator, you are enforcing a microscopic, distributed “social contract.”
Humans are intensely focused on building personal reputation, and reputation mechanisms can dramatically increase cooperation rates. Today, however, technology has stripped “reputation” and “third-party punishment” of their human biological intuition, handing them over to algorithms for quantification. The more a system enforces cooperation through algorithms, the more the genuine sense of trust between people, grounded in human nature, is diluted.
In exploring ultimate origins, the paper roots its answer in gene-culture coevolution. Cultural norms, such as punishment and reward mechanisms learned through socialization, have reshaped our evolutionary environment, allowing genes with a propensity for cooperation to be preserved. Yet, the dominant force in cultural transmission today is no longer the tribal elder or the village leader, but the recommendation algorithm. If the trajectory of our cultural evolution is increasingly guided by the silicon-based logic of pursuing user stickiness and commercial conversion, how will humanity’s instinct for “strong reciprocity” be exploited?
When humans achieve mutual cooperation, the dopamine system is activated. But technology is hijacking this circuit; a “like” on social media is a cheap form of “altruistic rewarding.” It substitutes the immediate gratification of dopamine for the arduous process of building deep cooperation in reality. Over time, will the mechanisms of altruistic punishment and rewarding—originally utilized to sustain grand communities—degenerate into endless emotional venting in the virtual world?
A minority of strong reciprocators can force a majority of selfish individuals to cooperate. However, modern technology magnifies the concealment of selfish individuals—for example, through anonymity—while simultaneously relinquishing the power of punishment to platform algorithms. Then, when a society’s cooperation no longer relies on its citizens’ internal “inequity aversion,” but rather on round-the-clock surveillance cameras, facial recognition, and the automatic execution of smart contracts, can humanity still be considered an “altruistic” species? If the cost of committing evil is driven so infinitely high by technology that no one dares to do wrong, does the moral value of “goodness” dissolve along with it?
Human altruism survived the bitter cold of the Ice Age and walked out of the African savanna. This force is extraordinarily powerful, yet it is also extraordinarily fragile.