Jacob Habinek, Institute for Analytical Sociology in Norrköping, Sweden
This paper takes stock of the increasing application of techniques from computational social science in the realm of social science history. Drawing on examples of current research situated within this scholarly borderland, it advances three interlocking claims. First, computational text analysis offers no escape from promises and pitfalls of working with archival materials. If anything, it highlights the special challenges pertaining to “found data” both within and beyond the archive. Second, at least for now computational text analysis is not well suited to replace the humanistic, interpretive work of reading. Efforts to use computational text analysis to extend or replace reading continue to encounter the same problems as earlier efforts to formalize or quantify the contents of textual materials. Third, computational text analysis holds the most promise when it attempts to inform our reading of historical sources. Computational text analysis can help address problems that stubbornly resist humanistic inquiry, such as those concerning prosopography, periodization, and cultural shifts.
No extended abstract or paper available
Presented in Session 109. The Future of Comparative-Historical Social Science I: Scholarly Borderlands