Intellectual history and text analytics

Abstract

This paper provides some early thoughts on a highly experimental approach to the study of early modern Korean intellectual history that takes advantage of modern computing power and advances in computational text analysis. The objective is to identify 'intellectual communities' on the basis of textual patterns and key expressions present across a digitized corpus of Korean Neo-Confucian writing. Thanks in large part to the National DB initiative that launched in 1998, South Korea boasts a massive repository of digitized sources. More than seventy National DBs that would be of primary interest to academics in the humanities and the social sciences have been made available online and accessible for free. Each National DB is a sizable project in itself. The online edition of the Han’guk munjip ch’onggan alone has so far digitized 1,199 collected works of medieval and early modern intellectuals who were active before the early twentieth century. The staggering amount of digitized sources made available in Korean Studies is the envy of other area studies disciplines. However, new modes of research enabled by digitization have received surprisingly limited attention in the field. The pilot project to be presented is an attempt to apply machine learning and natural language processing to the extant writings of Korean Neo-Confucians who were active in the mid and late 1500s. The corpus has been extracted from approximately 200 collected works available in the online edition of the Han’guk munjip ch’onggan, and the textual data have been deliberately restricted to expository genres such as classical exegesis, philosophical treatise, letter, memorial, and front and back matter.