Analyzing Moravian Feelings Using Computational Methods to Ask Questions about Norms and Sentiments in Eighteenth-Century Moravian Lebensläufe

Abstract

ABSTRACT This article introduces a computational methodology of analysis to enhance and assist in understanding the occurrences of sentiment in a sample eighteenth-century corpus of German and English Moravian memoirs (Lebensläufe). The computational methods used include sentiment tagging and scoring to derive sentiment trendlines, part of speech (POS) tagging with lemmatization (grouping inflected forms together as a single base form), word frequency analysis, semantic tagging in XML-compliant TEI, and “key word” analysis. This analysis using DH (Digital Humanities) methods, sentiment trendlines, and contextual word usage can give a more complete picture of how sentiment is used in Moravian memoirs. The corpus of forty-eight transcribed German and English memoirs from the Moravian Lives transcription desk on the project website (https://moravian.bucknell.edu) is originally from both the Moravian Archives in Bethlehem, Pennsylvania, and the former Fetter Lane Archive, now kept at Church House Archives in Muswell Hill, London, UK.