Lexical diversity as a predictor of genre in TV shows

Mary Akbary,Scott Jarvis
DOI: https://doi.org/10.1093/llc/fqad004
IF: 1.299
2023-03-25
Digital Scholarship in the Humanities
Abstract:Abstract Many studies have investigated the linguistic characteristics of television and have found important differences between categories of TV programs. Yet, little is known specifically about the lexical profiles of different genres of television discourse. The present study sought to address this gap by exploring the lexical diversity of 714 episodes representing four TV genres. The lexical diversity of each episode was measured using a six-dimensional model of lexical diversity. Multinomial logistic regression was used to determine whether the four TV genres in the present study have unique lexical diversity profiles and whether the genres of individual TV episodes can be predicted based on the adopted model. The results indicated that the four genres do indeed exhibit unique lexical diversity profiles; it was also found that the genres of individual TV episodes can be predicted with approximately 91% accuracy based on this model. These findings were interpreted as underscoring the relevance of lexical diversity to genre analysis of TV shows and the importance of using a theoretically grounded multivariate model of this construct.
linguistics
What problem does this paper attempt to address?