본문 바로가기
  • Home

A New Approach to Socialization in Gaebyeok’s Tone: Focusing on Topic Network Analysis

  • Journal of Humanities, Seoul National University
  • 2021, 78(1), pp.221-262
  • DOI : 10.17326/jhsnu.78.1.202102.221
  • Publisher : Institute of Humanities, Seoul National University
  • Research Area : Humanities > Other Humanities
  • Received : February 9, 2021
  • Accepted : February 20, 2021
  • Published : February 28, 2021

HurSoo 1

1서울대학교

Accredited

ABSTRACT

This paper aims to investigate whether the phenomenon in which Gaebyeok (開闢) acquired a socialistic in tone in its late years impacted on the leadership of Gaebyeok. The analytical approach used to address this question includes three steps: topic modeling, network analysis, and statistical testing. The indicators used to define socialization are first identified and determined. 7 topics and 104 topic words are selected at the stage of topic modeling after the data pre-processing of 334 main editorials of Gaebyeok. Twenty of these topic words that are classified into topic 1 are the most important indicators for defining socialism. This paper then examines how the Gaebyeok became socialistic in tone. By inputting 26,060 ‘document-topic words-tfidf value’, a topic network map was computed, which reflects the relationships among the seven topics. The result reveals that the network centrality moved from Kaejoron (改造論) in the early Gaebyeok years to socialism in the late Gaebyeok periods. Lastly, this paper evaluates the influence of socialism on the leadership of Gaebyeok. One hundred twenty-one identified editorials published in the late Gaebyeok periods could be divided into two types: one written by the ‘Gaebyeok leaders’ and one written by the ‘ordinary authors’. Using these sources, this study examines whether there is a significant difference between two groups in terms of the proportions of topic 1 and topic 8 via the ‘T-test’, respectively. The result shows that socialism’s influence on the leaders of Gaebyeok was significantly weaker than that of other authors.

Citation status

* References for papers published after 2023 are currently being built.

This paper was written with support from the National Research Foundation of Korea.