Mar 1, 2007:

American English dialect perception experiment

Keelan Evanini, Department of Linguistics

I am going to replicate the experiment done by Clopper and Pisoni, 2004 ("Some acoustic cues for the perceptual categorization of American English regional dialects", JPhon). The experiment involves playing 3 sentences each from 66 speakers in the TIMIT corpus, and conducting a forced-choice categorization of their dialect (from among six dialect regions). The key difference is that Clopper and Pisoni did the experiment with naive subjects, whereas I am specifically using linguists as subjects. I want to see how much difference this makes in the results (this is relevant, because I want to see what is a reasonable upper bound on the accuracy that can be expected from the best automatic dialect identification system).