Topic Modeling on Second Texts

The article analyzes the thematic structure in the issues of the scholarly journal Second Texts, in 1990-2012, using the topic modeling methodology. Based on 1923 articles from this period, the authors obtained and interpreted 120 automatically inferred topics from subsequent issues. The analysis allowed for the identification of permanent and seasonal topics, which appeared with a growing or declining frequency over time. The application of quantitative methods enabled the authors to capture the history of the discipline not so much as a series of distinct phases but rather as a continuous process built from different configurations of co-existing research interests.