When Do Institutional Divisions Become Epistemic Boundaries? A Computational Framework for Measuring Multidimensional Boundary Permeability

• In Event: Methods and Methodologies in the History of Sociology and Social Thought

Abstract

This research asks whether and when institutional divisions produce genuine epistemic differences, and along which dimensions they emerge. Drawing on a large corpus of publications in criminology and criminal justice, this paper introduces a computational framework that uses unsupervised and supervised machine learning to detect boundaries across multiple dimensions and to measure their permeability over time. Boundaries may be bright on some dimensions and blurred on others, following independent logics that existing methods, treating boundaries as predetermined or unidimensional, cannot detect. This framework offers sociologists of knowledge and historians of social thought a replicable and scalable method for tracing how institutional separation may or may not produce intellectual divergence in the textual traces of editorial and authorial decisions.

Author

• Kaylee de la Motte, The University of Arizona