Algorithmic Close Reading: Using Semantic Triplets to Index and Analyze Agency in Holocaust Testimonies

The following article presents a digital humanities exploration of indexing and analyzing expressions of agency in Holocaust testimonies. Using a set of text analysis methods to identify, classify, and visualize semantic triplets, we show how attention to agency complements and extends conventional approaches to indexing. Our examples come from two corpora of Holocaust oral histories: the first were conducted in Displaced Persons camps in 1946 by an interviewer named David Boder; the second were conducted by the USC Shoah Foundation Visual History Archive in the 1990s. We focus on two salient testimonies from each corpus in order to describe the methodology and what the analysis of agency can contribute to the writing of microhistories of the Holocaust. Building on semantic web analyses, the methods provide a groundwork for the development of a graph database to search testimonies by agency and thereby provide historical insights about what people report they did and what was done to them.