Sequence Comparison Via Alignment and Gibbs Sampling: A Formal Analysis of the Emergence of the Modern Sociological Article

Andrew Abbott * and Emily Barman †

Various substantive literatures in sociology seek small regularities in sequences: turning points in the life course, catalytic moments in organizational change, sharp turns in occupational trajectories, and the like. Commonly these are turning points, but they may also be simple local patterns. This paper reports a method for discovering such regularities even when they are quite faint, applying that method to rhetorical regularities in sociological articles. The paper begins by analyzing the overall sequence structure of such articles and then gives a basic introduction to Gibbs sampling, one member of the broader class of Markov Chain Monte Carlo (MCMC) methods. It then reports an algorithm employing Gibbs sampling to find local sequence regularities and applies that algorithm to demonstrate the subsequence regularities present in sociological articles. Substantively, the paper shows that the rhetorical structure of sociological articles changed from one pattern to another in the period 1895-1965 and that certain faint but standard rhetorical subsequences became characteristic of articles in the later period. Methodologically, it introduces a broad class of methods that provide effective approaches to a number of previously intractable statistical questions.

* University of Chicago
† University of Chicago