LEXICAL DIVERSITY AND EMOTIONAL LANGUAGE IN POP, ROCK, AND RAP LYRICS

Abstract

Song lyrics constitute a rich and underexplored domain of linguistic inquiry, offering a window into the intersection of popular culture, emotional expression, and vocabulary use. The present study employs a corpus-based computational approach to examine lexical diversity, sentiment polarity, and thematic structures in contemporary English song lyrics drawn from three dominant musical genres: Pop, Rock, and Rap. A balanced corpus of 3,000 songs—1,000 per genre—was compiled from a publicly available Kaggle dataset. Lexical diversity was operationalized through the Type-Token Ratio (TTR), while sentiment polarity was measured using the TextBlob framework. Keyword frequency analysis and Latent Dirichlet Allocation (LDA) topic modeling were applied to uncover dominant vocabularies and latent semantic themes. One-way Analysis of Variance (ANOVA) and Tukey's Honest Significant Difference (HSD) post-hoc tests were used to determine the statistical significance of inter-genre differences. Results indicate that Rap exhibits significantly greater lexical diversity (M = 0.5070) than both Pop (M = 0.4707) and Rock (M = 0.4724), while Pop demonstrates significantly more positive sentiment polarity (M = 0.0668) than the other genres. Topic modeling identified four major thematic clusters—Narrative Action, Romance, Reflective Existentialism, and Socio-Cultural Realism—which vary in prominence across genres. These findings demonstrate how computational linguistic methods can systematically uncover meaningful differences in language use across contemporary musical genres.