Schedule 73. StuTS

Yue is a Sinitic language spoken in Southern China which consists of numerous dialects including Cantonese. While we know that Yue is typologically and genetically different from European languages, do we find any similarities between these groups of languages when we compare the way dialects vary (e.g. do we see a dialect continuum in Yue too?)?

Dialectology is a discipline with a long history which is interested in the linguistic variation across space. For over half of the 20th century, dialectologists have been engaged with issues such as finding out whether dialect areas exist and seeking dialect classification. To address these issues, dialectologists had been conducting dialect surveys and plotting dialect features on maps. Findings such as dialects form a continuum, dialectal variation correlates to geography have been found in Europe, America and Japan, what about other parts in the world?

Yue, one of the ten branches in the Sinitic family, is spoken in southern China. Like western dialectology, there were several dialect surveys conducted in the Yue-speaking area and between the 1980s and 2010s, more than 100 dialects were surveyed. These data, however, were mainly used for language-specified studies with a historical orientation, implying they are relatively untapped for the purpose of dialect geography. Does Yue, a non-western language which was developed in a completely different sociolinguistic landscape, and social history, and from a completely different language family, show similar patterns as European languages?

In this presentation, I will explore the dialect geographical patterns found in Yue. Using over 120 words from 104 dialects, I aim to explore three aspects of dialect geography: 1) Does Yue show a dialect continuum just like western languages?, 2) Is the traditional classification within Yue justified? and 3) Do we find geographical and/ or historical correlates in the Yue data?

To answer these questions, I take an aggregate, computational approach to dialect data (Nerbonne 2010), which applies computational techniques to process and visualize a large quantity of dialect data automatically without feature selection (Nerbonne 2010), namely calculating dialect distances using Levenshtein distance (LD hereafter) and visualizing data with Multidimensional Scaling (MDS hereafter) and maps.

The results show the possibility that a typologically, genetically different language can still be analysed like European languages, and in fact, they show similar patterns. The shared methodology removes barriers from pursuing a cross-linguistic dialectology.

Info

Day: 2023-05-27
Start time: 11:40
Duration: 00:30
Room: SH 0.106
Track: Computational Linguistics

Links:

• iCalendar

Feedback

Click here to let us know how you liked this event.