A quantitative analysis of Cantopop lyrics with R

Abstract

 If language is derived from life, then it follows so does music. Popular music reflects the core interest and value of people. Thus, analysing the lyrics and the development of Cantopop can reveal insights into the sentiment of people of different generations, insofar as expressed by music. This dissertation presents a novel quantitative analysis of a Cantopop corpus to integrate analyses from both linguistic and sociolinguistic perspectives, unlike any other qualitative studies available in the literature. Also, while previous research has largely ignored the subsequent development of contemporary Cantopop after 1997, this dissertation updates the literature by covering the current development and changes of contemporary Cantopop for two more decades, i.e. 2000s and 2010s. Quantitative features and changes in the concept of love throughout the development of Cantopop are the main focus for discussion in this dissertation. Lyrics of 200 Cantopop songs from the 1970s to the 2010s are covered. Quantitative and qualitative features in songs are then analysed using the R programming language and discussed. Results show that love songs, especially songs with the theme of romantic love, have accounted for the majority of the popular Cantopop market. An in-depth qualitative analysis is then presented to depict how the concept of love has changed throughout the past five decades.