here - Linguistic Society of America

Meredith Tamminga (University of Pennsylvania) Session 21All aLIKE: item frequency and /aɪ/-raising in PhiladelphiaUsage-based models of phonology predict that different words should undergo phonetic change at different rates. The multipleroles of LIKE show systematic phonetic differences (Drager 2011) and thus should be susceptible to divergence during soundchanges such as /aɪ/-raising in Philadelphia. We fit linear mixed-effects regressions to data from 37 Philadelphia speakers topredict /aɪ/ height from birth year, vowel duration, and LIKE role (verb, preposition, conjunction, adverb, discourse marker). Ishow that despite large frequency differences, these roles undergo /aɪ/-raising as a unit (pace duration differences). Thus,phonological theories that can both allow and inhibit frequency effects are needed.Kevin Tang (University College London) Session 34Andrew Nevins (University College London)Michael Becker (Indiana University)Prosody drives alternations: evidence from a 51 million word corpus of Brazilian PortugueseToken frequency is widely claimed to promote phonological alternations (Bybee 1995, 2001). We examine plural alternations inBrazilian Portuguese (e.g. jorna[w] ~ jorna[j]s ‘newspapers’), and show that the correlation between token frequency andalternations is epiphenomenal, and actually depends on prosodic shape. We constructed a 51 million word corpus of spokenBrazilian Portuguese based on movie subtitles. Statistical analyses revealed that while a variable encoding monosyllables vs.polysyllables made significant improvement in predicting alternations, token frequency makes no significant improvement. Thisstudy suggests that frequency-tracking alone is unlikely to condition learners' generalizations about the patterns governingmorphophonological alternations.Anne-Michelle Tessier (University of Alberta) Session 39Modeling L1 transfer and L2 development in child phonologyThis paper studies the interaction between universal biases and language-specific experience in young children’s L2 phonology.It simulates L2 English cluster acquisition using an existing model (Becker & Tessier, 2011), varying its initial states andacquisition rates. Simulation results are then compared with spontaneous productions from ten children of L1 Chinese and SouthAsian (Hindi or Punjabi) backgrounds, tested after four to eight months exposure to English (data from J. Paradis.) The resultsprovide some quantified evidence of how early L2 phonology can arise via normal learning mechanisms, when L1 grammars andtheir potential for variability are taken into account.Margaret Thomas (Boston College) Session 75Otto Jespersen and “The Woman,” then and nowChapter 13 of Jespersen’s 1923 Language (“The Woman”) is often the starting point for feminist critique of the representation ofwomen’s language. Jespersen asserts (providing virtually no evidence) that women avoid vulgarity; use more emphatics; cedelinguistic innovation to men; have smaller vocabularies and less phonetic sensitivity; speak more, and faster; and leave sentencesunfinished. This presentation discusses Jespersen’s text in the intellectual context in which it was written, and in which it is nowread, showing how its representation of women’s speech has influenced the ways in which modern language scholars conceive ofearlier discussion of language and gender.Anie Thompson (University of California, Santa Cruz) Session 17Japanese argument ellipsis as (modified) LF-copyingI examine argument ellipsis (AE) in Japanese (Takahashi 2006, a.o.) and show that AE has properties of both deep and surfaceanaphors (Hankamer and Sag 1976). I show that AE should be analyzed as a syntactically simplex proform interpreted via amodified form of LF copying (Chung, Ladusaw and McCloskey 1995). The analysis claims that both PF deletion and (modified)LF copying are necessary to account for the broad range of surface anaphors, and raises questions about the status of other 'mixedstatus' anaphors, such as English do so and Scandinavian det anaphora.Jennifer Thorburn (Memorial University of Newfoundland) Session 17On English in Aboriginal communities: a variationist analysis of Inuit EnglishThis paper adopts a variationist sociolinguistic lens and new-dialect formation theory to examine English in Nain, an Inuit community innorthern Canada experiencing rapid language shift from Inuktitut to English as a first language. I discuss three variables salient in thregion’s input variety, Newfoundland English—interdental stopping (this thing as dis ting), verbal -s (I loves it), and adjectiva209