Linguistic Circle

Speaker: David Willis (University of Cambridge)

Title: Tweetolectology: Issues in mapping morphosyntactic variation using social-media data

Abstract: Widespread use of social-media platforms like Twitter generates unprecedented quantities of data on real-time language use. In principle, this allows us to examine the distribution of linguistic variants across the population in finer detail than is possible using traditional variationist and dialectological methods. This talk will examine some of the issues involved in realizing this potential, showing that patterns of variation obtained via traditional methods are replicable, if appropriate methods of localization (using a range of user metadata as well as the GPS geolocation data automatically generated for around 2% of tweets) are adopted. Two case studies will be used to demonstrate this: for Welsh (second-person pronouns; deletion of present-tense auxiliary ‘be’; and the form of the embedded focus particle) and for British English (dative alternation). In each case, clear regional patterns of variation emerge that are remarkably close to those achieved by traditional methods.

Contact details

Dr Chris Cummins