Co-sponsored by Clarke Program in East Asian Law & Culture, Cornell Law School

Presenter：Yun-chien Chang [Cornell Law School]

Discussants：Masahiko Kinoshita [Kobe University], Hiroharu Saito [UTokyo]

Moderator：Takeshi Fujitani [UTokyo]

Date：July 14, 2023／15:00‐16:40 (JST)

Location：Zoom

Language：English

Title：Machine-learning human rights: an empirical study of constitutions around the world in 1900-2020

Target：Open to the public

Abstract：Drawing on a dataset of national constitutions, we introduce a new method for analyzing and categorizing legal documents. Whereas most approaches use unsupervised methods, we also draw on a supervised machine-learning method known as SEMMS. We apply both supervised and unsupervised machine- learning algorithms to the rights provisions in national constitutions. Our approach produces groupings of constitutions that confirm findings of other scholars, but also identifies some unconventional groupings. We are also able to analyze switchers, by which we mean those countries that move from one grouping to another, and thereby to identify moments of constitutional transition. Key variables that distinguish groupings and transitions are identified. We suggest that our technique and categorization can potentially produce gains for other areas in which large corpuses of legal materials are available.