You wouldn't really be that surprised to find social identity encoded in microblogging, but then to find regional dialect and particularly from a corpus of US tweets is impressive. That's the work of Jacob Eisenstein and colleagues at Carnegie Melon which is carefully reported by Rupert Cornwell here. And the fact that they could accurately predict the regional origin of tweets to within 300 miles adds to the wow factor. I suppose that all underscores the fact that from a sociolinguistic perspective there's nothing radically new about digital literacies except for the very fact that old themes persist in what seem like radically new media.
No comments:
Post a Comment