Measuring Paradigmaticness of Disciplines Using Text

Abstract

In this paper, we describe new methods that use the text of publications to measure the paradigmaticness of disciplines. Drawing on the text of published articles in the Web of Science, we build samples of disciplinary discourse. Using these language samples, we measure the two core concepts of paradigmaticness—consensus and rapid discovery (Collins 1994)—and show the relative positioning of eight example disciplines on each of these measures. Our measures show consistent differences between the “hard” sciences and “soft” social sciences. Deviations in the expected ranking of disciplines within the sciences and social sciences suggest new interpretations of the hierarchy of disciplines, directions for future research, and further insight into the developments in disciplinary structure and discourse that shape paradigmaticness.

Keywords

• Computational Linguistics
• Disciplinary Paradigms
• Hierarchy of Sciences
• Sociology of Science