Toward a Computational Social Scientific Analysis of Meaning

Abstract

This study discusses the significance of the topic model, which is a method of natural language processing in sociology. First, according to the results of an examination of the previous research, we demonstrated that the topic model can be applied to analyze meaning; moreover, it is necessary for examining the relationship between the construction of reality and language to consider the “topics” in the model. Thereafter, we focused on Schutz's theory of relevance, which explains how people make the world meaningful in terms of linguistic knowledge. Finally, theoretical examination and empirical analysis of free descriptive answers to a questionnaire survey revealed that the model can clarify the system of relevance. The results suggest that the topic model enables mathematical or quantitative development of the phenomenological sociology of knowledge; furthermore, it is expected that a computational social scientific approach would be useful for analyzing meaning.