从众陷阱与独立判断的形成
Conformity Traps and the Formation of Independent Judgment
浏览论文内容
中文总结 AI 辅助
该研究探讨群体陷入从众陷阱或维持独立判断的原因，构建动态模型分析判断存量的继承与恢复，明确相关临界值、状态分类及人员流动对恢复的影响，通过示例验证核心结论。
中文摘要 AI 辅助
我研究为何部分群体——无论是政治实体还是组织——能维持独立判断，而另一些则陷入从众陷阱。社会认可不仅会抑制独立判断的表达，还会抑制维持此类判断的上游实践：维持判断会产生可见的质疑，而观察者无法区分勤勉与反对。由于质疑普遍时怀疑会降低，相同的认可渴望既可能维持质疑文化，也可能维持从众陷阱。该模型将维持判断与新情况出现时的判断行动相分离。当行动决定唯一确定时，我给出了共存的基础条件；在logit比例基准中，单一不等式是必要且充分的。主要贡献是动态层面的：在交错修订下，判断存量会被继承：保护异议会立即释放现有判断——在盈亏平衡能力阈值以下的稳态文化中，这具有负的预期适应价值——而重建存量则受人员流动的约束。单调算子从每个继承状态构建极值完美预见路径。当在所有可允许的延续下都能实现恢复时，两个临界值对继承状态进行分类——恢复不可能、依赖预期或确定——且不稳定文化处的聚焦条件严格区分它们，形成了预期选择长期文化的状态区间。人员流动随后在临时直接形成支持内，为使恢复可能或稳健提供了明确的实际持续时间界限。一个完全耦合的区间验证示例证实了共存、该区间以及不同的产出和福利影响。
英文摘要
I study why some groups---polities and organizations alike---sustain independent judgment while others fall into conformity traps. Social approval can suppress not only the expression of independent judgment but also the upstream practice that keeps such judgment available: maintaining judgment generates visible questioning, and observers cannot distinguish diligence from opposition. Because suspicion falls when questioning is common, the same desire for approval can sustain a questioning culture or a conformity trap. The model separates maintaining judgment from acting on it when novelty arrives. When the decision to act is uniquely determined, I give primitive conditions for coexistence; in a logit--proportional benchmark, a single inequality is necessary and sufficient. The principal contribution is dynamic. Under staggered revision the stock of judgment is inherited: protecting dissent releases existing judgment at once---with negative expected adaptive value on impact in a stationary culture below a break-even competence threshold---whereas rebuilding the stock is constrained by turnover. A monotone operator constructs the extremal perfect-foresight paths from every inherited state. When recovery under every admissible continuation is attainable, two cutoffs classify inherited states---recovery impossible, expectation-dependent, or assured---and a focus condition at the unstable culture strictly separates them, opening a band of states where expectations select the long-run culture. Turnover then yields sharp physical duration bounds, within temporary direct formation support, for making recovery possible or robust. A fully coupled interval-certified example verifies coexistence, the band, and the distinct output and welfare implications.
