Research

Documentation, description, and grammatical analysis of Uab Meto

The main focus of my current research is the documentation, description, and grammatical analysis of Uab Meto (ISO code: aoz), an understudied Austronesian language of Timor in Indonesia and Timor-Leste.  I am particularly interested in the morphology and syntax of case, agreement, and valence-changing operations in the language.  This has turned into a more general project to figure out the syntax of the clausal spine and the place of the various functional items in it, all the way from the verb root through valence-changing affixes, negation, TAM markers, and A'-elements.  Soon to appear in the Proceedings of WCCFL 39 is a paper discussing the low placement of agreement in the clausal spine, in an AgrP projection directly above Voice and thus below TAM marking and negation.  Uab Meto has nominative-accusative alignment in its pronominal case marking and in its verbal agreement, and the language shows that such an agreement pattern does not need to derive from a ɸ-probe on T; the probe just needs to be above the Merge site of external arguments.  Furthermore, the fact that nominative subjects surface in a pre-verbal and pre TAM position suggests that case and agreement can be handled separately.  The low placement of agreement on an Agr head present in every verbal complex leads to some interesting effects like the possibility of verbs agreeing with accusative-marked DPs as well as different verbs in a clause agreeing with different arguments.  These finding are expanded upon in a 2023 manuscript that also puts them in a typological context.

Uab Meto is also a very interesting language from a morphophonological standpoint.  The language has allomorphy between C- and CV- sets of verbal agreement prefixes conditioned by phonological, morphological, and lexical factors.  This allomorphy raises questions about the types of structural relationships that are necessary for allomorphy to occur, how to resolve competition between different conditioning factors, as well as the kinds of information present before and after Vocabulary Insertion.  The language also has synchronic metathesis in a number of contexts, and this interacts with various processes of vowel assimilation, vowel epenthesis, consonant deletion, encliticization, and syllabification (see Edwards 2020 for details).  Notably, the language has a dispreference for tautosyllabic consonant clusters, but many words start with CC clusters, including verbs with agreement prefixes.  The usual repair is to resyllabify the first consonant of the cluster to act as a coda of the preceding V-final word (which often involves blocking CV → VC metathesis), but this repair is seen in variation with a few others, including vowel epenthesis, deleting the first consonant of the cluster, or simply not repairing the cluster and pronouncing it as is.  These latter repairs are especially common when there is no preceding word to resyllabify into, and all of this raises intriguing questions about how speakers choose repair strategies and handle conflict between the faithful representation of CC clusters and the language's phonotactic constraints.

I started working on Uab Meto because in summer 2019.  I had the opportunity to participate in a linguistic fieldwork and documentation training program headed by Peter Cole and Gabriella Hermon, working in conjunction with Yanti, Jermy Balukh, and Asako Shiohara.  This program placed early-stage US grad students into teams with two Indonesian college students interested in documenting their native language.  Each team traveled to a community in eastern Indonesia, in my case the Uab Meto-speaking village of Oelneke in West Timor, which is one teammate's home village.  While there, our team worked together to record naturalistic narratives about a variety of topics including personal narratives, folk tales, historical accounts (like the history of Oelneke), discussions of farming or weaving techniques, among others.  We devised a practical orthography for ourselves, transcribed and glossed the texts to the best of our ability, and prepared the various files for archiving in PARADISEC.

The archived materials are in the PARADISEC collection AOZ2019, which can be accessed here!

Comparatives and other degree constructions in Vietnamese

My first qualifying paper, published in the Proceedings of Sinn und Bedeutung 24, analyzes comparatives and other degree constructions in Vietnamese from a typological perspective.  Beck et al. (2009) conducted a cross-linguistic survey of degree constructions and proposed three parameters to classify languages according to the constructions they allow and their available interpretations: 1. whether a language has degrees in its semantics; 2. whether a language has degree abstraction; and 3. whether a language allows the degree argument position of a gradable predicate to be overtly filled.  This paper provides novel data from Vietnamese that test the predictions of these parameters.  Languages with clausal comparatives (e.g. Phoebe sang louder than Tyler whistled.) and a positive setting for these parameters should allow subcomparatives (e.g. John is taller than the car is long.).  This paper shows that Vietnamese has clausal comparatives and a positive setting for all the parameters, but despite this, many subcomparatives are ungrammatical.  Further examination of the data reveals a crucial generalization: a predicate’s ability to remain in the standard of a subcomparative is linked to its ability to interact with nhiều ‘much, many’.  I propose that this generalization can be captured by positing that degrees combine directly with some Vietnamese predicates, while in other cases degrees combine with nhiều or its silent counterpart μ before combining with predicates, an idea inspired by Bresnan (1973), Grano and Kennedy (2012) and Wellwood (2012).  I also propose a mandatory deletion operation that occurs in the standards of Vietnamese comparatives, forcing predicates to elide when they combine directly with degrees but allowing them to remain overt when degrees must first combine with nhiều/μ.

Pronoun case variation in English

My undergraduate thesis and the LSA proceedings paper derived from it examine patterns of English pronoun case marking in multiple environments where variation is observed.  Coordinate structures are a well-known environment that exhibits variation (e.g. {She/Her} and Sandy went to the store yesterday.), and these are given a thorough treatment is work like Grano (2006), so my own work focuses on case variation in non-coordinated pronouns in environments including pronouns modified by numerals (e.g. {We/Us} three have to be leaving now.), PPs (e.g. How much would {?we/us} with insurance have to pay?), and NPs (e.g. {We/Us} linguists are a crazy bunch.), it-clefts (e.g. It was {?she/her} that won the contest.), comparatives with than (e.g. Sarah is taller than {?I/me}.) and as (e.g. Sarah is as tall as {?I/me}.), and in fragment answers after not (e.g. Who took the cookies? - Not {?I/me}.).  A Mechanical Turk study is conducted to collect acceptability judgments for different pronouns of different cases in these constructions, and the data from this study is bolstered by a corpus study of the Corpus of Contemporary American English.  These studies find that pronouns modified by numerals, PPs, and NPs are preferred with nominative case in finite subject position and accusative case in object position, though accusatives are relatively more acceptable in finite subject position than nominatives in object position.  Pronouns in the other environments mentioned were preferred with accusative case.  The corpus study also finds an increased usage of nominative pronouns in more formal contexts, especially following than and not.  These findings lead to a model of English pronoun case variation based on OT syntax that has three competing case assignment mechanisms: structural case (assigned by an external syntactic head), default case (always accusative in English), and prescriptive case (nominative in English).  Variation results from less optimal candidates surfacing or from a reranking of constraints caused by the influence of prescriptivism.  This is one formalization of the idea of interacting constraints discussed in works like Quinn 2002, Schütze 2001, Parrott 2006, Grano 2006, and Sobin 1997, among others.