ABSTRACT This study presents a dialect identification task in which 191 listeners drew on a digital map around the area(s) they thought 99 speakers were from and provided evaluative responses based on speech excerpts. This study is the first to demonstrate the importance of uniting five strands of investigation in dialect identification tasks: (1) listeners’ accuracy compared to chance; (2) listener‐group and speaker‐group accuracy; (3) precision; (4) misidentifications; (5) the relationship between attitudes and patterns of (mis)identification. This approach provides detailed insights into non‐specialists’ perceptual representation of linguistic variation, comprising the first perceptual dialectology study in South East England. Results show that rather than identifying regional linguistic features, listeners were closely attuned to variation by class and ethnicity, which predicted how they geographically assigned speakers. In addition, the method provides insights into perceptual processes, including the important role of social meaning in the perception and representation of dialects.
A. H. Cole (Mon,) studied this question.