Computational Text Analysis for Studying the Evolution of Meaning of Categories

Abstract

Over the decades, organization scholars have studied the formation of the meaning using various theoretical concepts. Analyzing meanings, however, pose methodological challenges as meanings are dynamically formed through a complicated process, which requires the analysis of large-scale historical data. Recent advances in computational text analysis of large digitized corpora have provided an opportunity to improve our understanding of the evolution of meanings. In this study, I will focus on a stream of research that can expect a significant benefit from computational text analytics: category perspective. This article suggests a framework to carry out the research on the evolution of the meaning of a category using computational methods. Specifically, I propose a new computationally-supported grounded theory development method that combines manual approaches for the interpretations with computational text analysis. This method aims to scale traditional theory-building methods, incorporating more complexity without losing nuance to enrich extant theories. Using the case of blockchain, this paper illustrates how this technique can be applied to study the evolution of the meaning of emergent technology categories in society.