Bridging Community and Individual: Dual Approaches to Variation and Change in Hong Kong Cantonese

Variability and change are fundamental features of human language. Language isalways changing and no two people speak in exactly the same way. However, variability and change are not random. They follow systematic rules. For instance, in Hong Kong Cantonese, the two sounds /n/ and /l/ are merging so that the word for YOU can be pronounced as either nei5 or lei5. But the choice of nei5 or lei5 is not random and can be shaped by many social and linguistic factors. A Cantonese speaker in casual conversations with their local friends may prefer lei5. The use of /l/ may also be preferred before one vowel over another. However, a major mystery remains: Where do these systematic rules come from? What knowledge do individual speakers need to have for these patterns to manifest at the community level and how do their own production patterns and perceptual abilities contribute?To tackle these questions, this project adopts a dual approach, utilizing both corpus analysis and behavioral experiments to understand ongoing sound changes in Hong Kong Cantonese, bridging community-level patterns and individual-level processes.⁠⁠⁠⁠⁠⁠⁠Hong Kong Cantonese is a perfect living laboratory because it is currently undergoing several rapid sound changes within a vibrant multilingual environment, offering a fresh perspective beyond studies focused only on English or other Indo-European languages.This project uses the aforementioned n-l merger as a test case. First, a large-scale corpus of natural speech from 120 native speakers of Hong Kong Cantonese will be compiled to map production patterns at both the community and individual levels.Subsequently, three perception experiments will be conducted with the same 120 participants to probe how individuals leverage their sociolinguistic knowledge of language contact to resolve ambiguity. This approach bridges the gap between community-level patterns and individual cognitive processes, and critically, forges a direct link between production and perception in sound change.The results of this project will document ongoing sound changes in Hong Kong Cantonese amidst rapid socio-cultural shifts, informing theories of language change by bridging patterns observed at community and individual levels. Additionally, this project will have the long-term potential to train Large Language Models to better accommodate the linguistic variability used by diverse speakers and to select socially appropriate variants to meet the communicative needs of users in different social situations. Ultimately, this work will contribute to scientific and cultural understanding of how humans communicate across a diverse range of language backgrounds.