Yosiane White will be defending her dissertation proposal, tentatively entitled: The Mental Representation of Variable Morphology, on Friday, May 22nd, at 11:00 a.m. 


Title: The Mental Representation of Variable Morphology

Supervisor: Meredith Tamminga

Proposal Committee:  Anna Papafragou (chair), David Embick & Kathryn Schuler

Time: 11:00 a.m. - 12:30 p.m.



Variationist sociolinguists are interested in questions concerning the grammatical locus of variation, which can be particularly complex at the phonology/morphology interface. Simultaneously, psycholinguistics offers tools for probing the mental representation of morphologically complex words, but these are rarely directly applied to sociolinguistic variables. This dissertation uses psycholinguistic methods to investigate the mental representation and processing of sociolinguistic variation at the phonology/morphology interface. I compare two sociolinguistic variables: the ING variable (thinking / thinkin’), and word final rhoticity which I will call ER (for baker : [bɛikɚ] - [bɛikə]). Both of these variables involve phonological differences that occur both in suffixes and monomorphemic words, giving rise to a debate of the morphological or phonological nature of the variables. Experimental priming methods allow for the disentangling of shared aspects of mental representation by showing how specific factors require similar or different types of processing during speech recognition, and can therefore stand to answer questions regarding the locus of variation of these variables, and potential differences between them. 


There are two core theoretical questions at the heart of this dissertation. First, how are the two socially meaningful variables ING and ER mentally represented? Are there distinct morphological pieces involved in their variability, or can we identify their processing as phonological? Second, how do the morphological or phonological similarities and/or differences between ING and ER, as found through experimental priming, compare to the variable production patterns found in naturalistic data, which for ING are relatively well established, and for ER will be examined through a corpus study. 


The work in this proposal begins to tackle the first of these questions for ING. In order to pinpoint fine-grained distinctions in morphological and phonological representation I design a structured set of auditory priming experiments that systematically isolate aspects of shared representation. The data I show here serves three purposes. First, it demonstrates successful affix priming for the regular suffix -ing. Second, it shows that this experimental paradigm is sensitive to processing differences for the socially meaningful variants involved in the variable ING. Third, and most importantly, it shows that not only do -ing and -in’ show priming in such a paradigm, there are differences in the priming effects between the variants. Crucially, the data shows an interesting difference in priming effect between -ing and -in’ variants in a phonological rhyming stem condition (drinking-thinking vs drinkin’-thinking), but equivalent priming for both variants in a stem repetition condition (thinking-thinking vs thinkin’-thinking) and in an unrelated stem condition (working-thinking vs workin’-thinking). This effect, along with a parallel study on word-final rhoticity, will be further explored in the dissertation.