Time: 4:30 - 6:00pm Venue: ArtsTwo 2.17
Janet B. Pierrehumbert
University of Oxford
People learn words from experiences with language. A surprising amount of statistical and social information has been shown to leave traces in the mental lexicon. Variation amongst speakers naturally results from social groupings and variation in linguistic experience. Some, but not all, of this variation becomes conventionally associated with different social characteristics, thereby acquiring social-indexical meaning. This phenomenon that has been very extensively investigated in the realm of sociophonetics. However, social- indexical meanings in morphology have been much less explored, and some researchers indeed suggest that morphology is too abstract for such meanings to arise. Here, I will present experimental results addressing this issue. We find that words have social associations that are learned and can be used in making judgments. Furthermore — and contrary to some expectations — people generalize these associations to novel pseudowords that have an apparent morphological decomposition, such as “thrafium” or “pelpcase”. However the generalizations are not mediated by the morphological parse (which is independently known to play a role in phonological and syntactic processing). The results are consistent with the theory that general analogical processes are active in social adaptation and judgment, while more structured morphophonological parsing has limited interactions with the domain of social-indexical knowledge.