Whamit!

We are happy to announce that next week Stephanie Shih will give not only a colloquium talk, but also a minicourse!

Speaker: Stephanie S. Shih (University of Southern California)
Time: April 7th and 8th, 12:30-2pm 
Location: Fourth floor Seminar Room (32-D461)

EVALuating the relationship between phonology and statistics

In this minicourse, I take up two seemingly parallel threads in current phonological research. The first: the surface optimizing grammars that have become the mainstay theoretical approach in phonology. And the second: the staple methodological statistical tools—namely, regression modeling—that the field has increasingly embraced with the rise of experimental and corpus-based evidence as the empirical material that we analyze. We will talk about how the design and goals of these two approaches, which are usually cast as theory vs. methodology, actually dovetail, both in current day work as well as in the history of the field. The goals will be to reconcile the conceptual parallels between optimizing grammars and statistical models, and to consider what the convergence of the two mean for phonological theory.

A big welcome to Bruna Pereira (Universidade Federal dos Vales do Jequitinhonha e Mucuri), who joins us as a visitor this month on her sabbatical!  

Bruna Karla Pereira carried out her Ph.D. (2011) at the Universidade Federal de Minas Gerais (UFMG) with a full year as a visiting graduate student (2010) at the University of Cambridge (UK), under Ian Roberts’ supervision. In addition, she developed her post-doctoral research (2016), as a visiting scholar, at the Massachusetts Institute of Technology (MIT, USA), under Shigeru Miyagawa’s supervision. In her Ph.D., she was interested in the Minimalist Program, especially in the cartography of syntactic structures and its implications for the analysis of light adverbs, such as ‘lá’ in Brazilian Portuguese (BP). In her postdoc, she investigated universals in nominal agreement that determine the DP-internal distribution of the plural morpheme in order to account for structures of BP with apparent “mismatch” of agreement including possessives, wh-determiners, cardinals, and silent nouns. During her education, she was awarded funding from CNPq, FAPEMIG, and CAPES. Concerning her teaching experience, after having worked at the Universidade Federal de Lavras (2011-2013), with a temporary contract, she is currently a permanent professor at the Universidade Federal dos Vales do Jequitinhonha e Mucuri (2013 onwards) where she has been conducting research on Syntax with emphasis on Generative Grammar. Her CV is available both in Portuguese and English, respectively, at the following links: http://lattes.cnpq.br/2671430917722911 and https://orcid.org/0000-0002-4958-8621.

Speaker: Giovanni Roversi (MIT)
Title: Definite Spans and Blocking in Classifier Languages
Time: Tuesday, April 5th, 1pm - 2pm

Where: 32-D461, or on zoom, ask organizers for details 

Abstract: Giovanni will be leading a discussion of Peter Jenks’ (2018) “Definite Spans and Blocking in Classifier Languages.” https://escholarship.org/uc/item/5tt1j4pj

Abstract of paper: This paper presents a novel analysis of definite noun phrases in numeral classifier languages without definite articles. The motivation for this analysis comes from the classifier-modifier construction (CMC) in Thai, in which a predicative modifier can license a bare classifier, resulting in a definite interpretation. I argue that the definite readings are due to a null choice-functional determiner (Reinhart 1997, Winter 1997), which takes the modifier as its complement (Kayne 1994). I propose that the modifier licenses the bare classifier, otherwise prohibited, because head raising relative clause structures bleed the environment for a D-Clf-N span to be realized as a bare noun (Brody 2000, Svenonius 2012, a.o.). I argue that this coalescence-based account of definite noun phrases, specifically definite bare nouns, is an improvement on accounts based on head movement (Cheng and Sybesma 1999) or semantic type-shifting (Chierchia 1998). This analysis correctly derives the generalization that languages allowing definite bare classifiers do not allow definite bare nouns in most cases, captures Chierchia’s nominal typology without resorting to semantic variation, and opens up new accounts for the apparent optionality of functional morphology in analytic languages.

Speaker: Canaan Breiss (MIT)
Title: When bases compete: experimental and computational studies of Lexical Conservatism
Time: Monday, April 4th, 5pm - 6:30pm

Abstract: In this talk I examine the interaction of the phonological grammar and the lexicon through the lens of Lexical Conservatism (Steriade, 1997). This is a theory that addresses how the distribution of bases (existing stem allomorphs in a morphological paradigm) influence the way those paradigms accommodate novel members. The idea is that a phonological alternation only applies to novel words if there is an existing base form present elsewhere in the paradigm that offers the needed phonological material. Thus compénsable, for “able to be compensated”, undergoes stress shift (that is, *cómpensable) because the existing word compénsatory contains the compéns- allomorph. In contrast, *inúndable, for “able to be inundated” is judged worse than ínundable, since there is no existing base that can provide the stressed vowel (there is no form in inúd-). Using experimental data from English and Mexican Spanish, I demonstrate that this dependency between paradigm structure and phonological process application generalizes to entirely novel words in a probabilistic manner. Further, contrary to previous assumptions, I find that all stem allomorphs in a paradigm play a role in determining the form of the novel word, rather than only those that could reduce the markedness of the novel form (per Steriade (1997), Steriade & Stanton (2020)). I propose a novel grammatical model where allomorphs in the lexicon exert analogical pressures on novel words, which are cross-cut by phonological markedness constraints.

Speaker: Stephanie S. Shih (University of Southern California)
Title: Lexically-conditioned phonology as multilevel grammar
Time: Friday, April 8th, 3:30pm - 5pm

Abstract: This talk takes up two interrelated issues for lexically-conditioned phonological patterns: (1) how the grammar captures the range of phonological variation that stems from lexical conditioning, and (2) whether the relevant lexical categories needed by the grammar can be learned from surface patterns. Previous approaches to category-sensitive phonology have focused largely on constraining it; however, only a limited understanding currently exists of the quantitative space of variation possible (i.e., entropy) within a coherent grammar. In this talk, I present an approach that models lexically-conditioned phonology as cophonology subgrammars of indexed constraint weight adjustments (i.e., ‘varying slopes’) in multilevel Maximum Entropy Harmonic Grammar. This approach leverages the structure of multilevel statistical models to quantify the space of lexically-conditioned variation in natural language data. Moreover, the approach allows for the deployment of information-theoretic model comparison to assess competing hypotheses of what the phonologically-relevant lexical categories are. Two case studies are examined: part of speech-conditioned tone patterns in Mende (joint work with Sharon Inkelas, UCB), and lexical versus grammatical word prosodification in English. Both case studies bring to bear new quantitative evidence to classic category-sensitive phenomena. The results illustrate how the multilevel approach developed here can capture the probabilistic heterogeneity and learnability of lexical conditioning in a phonological system.