Which individuals and corporate actors belong in a collective, and who decides? Sociology has not had good analytical tools for addressing these questions. Recent work that adapts probabilistic representations of concepts and probabilistic categorization to sociological research opens opportunities for making progress on the measurement of memberships. It turns out that the probabilistic cognitive-based reformulation reveals unexpected connections to language models and natural-language processing. In particular, the leading probabilistic classifier BERT provides new and powerful ways to measure core concepts.

中文翻译：

---

根据认知科学和自然语言处理的发展衡量集体成员资格

哪些个人和企业行为者属于一个集体，谁来决定？社会学还没有很好的分析工具来解决这些问题。最近将概念的概率表示和概率分类应用于社会学研究的工作为在成员测量方面取得进展提供了机会。事实证明，基于概率认知的重构揭示了与语言模型和自然语言处理的意想不到的联系。特别是，领先的概率分类器 BERT 提供了新的强大方法来衡量核心概念。