Jin Liu, an associate professor of Chinese in the School of Modern Languages, created an algorithm to analyze tone use in Chinese rap songs. She also worked with students in the College of Computing to integrate linguistics and cultural studies with computer science and statistics.
"Technology provides empirical and quantitative evidence or data to verify human intuitive perceptions of art," Liu explained. "This is an interdisciplinary project that extends beyond the range of any single researcher's knowledge and expertise,"
Her co-authors on the paper included Amanda She (CS 2023), Haosong Ma (MS CS 2022), Jiahong Yuan, a professor at the University of Science and Technology of China, and Hongyuan Dong, an associate professor of Chinese language and linguistics at George Washington University.
"My field is linguistics and cultural studies, and we needed to collaborate with people in computer science to develop an algorithm and write a program to compute the Tonal Congruence Index," Liu said. "People in computer science need our expertise in linguistics and computational phonetics to develop different rules and criteria to distinguish the tones and learn the related pitch software."
Unlike English, Mandarin Chinese is a tonal language, meaning using the same word with different pitches can change the meaning of the words.
In Liu's study, she found that while Chinese rappers used standard tones in the earlier rap styles, artists following more recent trends — such as trap music and mumble rap — are more likely to change or ignore Chinese tones to better fit the hip hop style in their work.
Liu and her co-authors also report that Chinese rappers now use more English words in their work to achieve musical flow.
"As hip-hop music continues to diversify, the correlation between tone and rap gradually decreases," Liu said.
The researcher's next step is to develop an open-source tool to further automate the audio-processing pipeline.
Liu's work also has educational applications. For example, she recently presented "AI and Tones in Chinese Songs" at a symposium on education in the age of artificial intelligence.
"Language instructors should teach songs with a high score on the Tonal Congruence Index because tonal shapes and contrasts are better accommodated in them," she said.
This research is partially supported by the Ivan Allen College of Liberal Arts Small Grants for Research funding at Georgia Tech. "Linguistic Tone in Chinese Rap: An Interdisciplinary Approach" was published in the Journal of New Music in March 2024. It is available at: https://www.tandfonline.com/doi/full/10.1080/09298215.2024.2329075