My research regards different types of cross-clausal A-dependencies. The received wisdom is that finite and nonfinite embedded clauses have opposite behavior regarding their ability to license their subject. This difference is commonly stated in terms of phases: finite clauses are phases and, as such, not only do they have their own subject licensing resources, but they are also impervious to some phase-external operations. Nonfinite clauses, on the other hand, lack such resources and have to depend on clause-external operations in order to license their subject. I am captivated by data that challenge such generalizations. I have worked on both nonfinite clauses where a subject can be licensed with their own internal resources and finite clauses that are not impervious to phase-external operations. Broadening the empirical domain of cross-clausal dependencies allows us to deepen our understanding of the strategies needed to license a subject and of the variable degrees in which embedded clauses are permeable to or dependent on external resources and operations.
(Handbook overview chapter, with Dr. Claire Halpert)
Building on the Movement Theory of Control (Hornstein, 1999, et seq.), I propose that a lone determiner can be merged in a thematic position in the embedded clause, move into an additional thematic position in the matrix clause, upon which point its NP complement is Wholesale Late Merged (Takahashi & Hulsey, 2009). Along Takahashi (2019), I propose that Fox’s (2002) Trace Conversion consists in transforming the residue of movement into a set of features that are formally indistinguishable from a pronoun. Likewise, along wih Erlewine & Gould (2016), I propose that this interpretive procedure can apply at the Narrow Syntax, so that its result is visible for Vocabulary Insertion. The result is a theory of control that can account for why obligatory control PRO in a language like English has pronominal properties, a property that can be accounted for by a theory of control such as Chomsky’s (1981) PRO-based GB theory, but not by the original version of the MTC. Furthermore, this paper investigates languages where PRO is not phonologically null, but, rather, an overt pronoun, pointing out hitherto unnoticed generalizations about the nature of the clause where pronounced PRO’s occur, as well as about its status with respect to other pronouns that occur in languages that display an overt PRO. The theory proposed, coupled with a realizational framework of the grammar (Halle & Marantz, 1993, 1994) and with independently needed assumptions about phasehood, provides an account of such generalizations.
In Wolof, control clauses differ in whether or not the embedded PRO subject is pronounced. In some control clauses, the subject is phonologically null, as expected, while in others, it is an overt pronoun. The main questions that arise are then: why do control clauses in this language differ in the phonological realization of PRO? Which control theory is compatible with such realization? I suggest that control clauses where the subject is an overt pronoun project a ΣP which “impedes” movement. Assuming that control is derived by movement, I model the pronounced PRO as the partial residue of movement that has been impeded. Control clauses with a null subject, in turn, are restructured.
The semantic interpretation and syntactic distribution of BNs in Wolof differ from what is taken to characterize their counterparts in other languages. First, instead of being number neutral (i.e. compatible with either a singular or plural interpretation), BNs in Wolof are singular when unmodified. They can only receive a plural interpretation when the BN is accompanied by nominal morphology that can expone a plural feature. Importantly, if such morphological is number-invariant, the BN retains a singular interpretation. In order to account for these facts, I propose that Wolof instantiates a system of nominal licensing whereby interpretable features also require licensing. If correct, this analysis provides a window into the role that features play as driving forces of syntactic derivation.
BNs in Wolof have the syntactic distribution of pseudo noun incorporation. Alternatively said, these nominals have to be linearly adjacent to the verb. However, there are two seemingly unrelated circumstances under which these nominals can sidestep this adjacency requirement: they can be A-bar moved or there can be an additional argument inserted between the BN object and the subject of the sentence. Assuming dependent case, I propose that this set of circumstances is not accidental: BNs in Wolof do not have to be adjacent to the verb when a case competitor is provided to them – otherwise, they have to be licensed via linear adjacency with the verb. What A-bar movement and the introduction of an intermediate argument have in common is that they provide a case competitor to the BN. This research thus contributes to the long standing debate about the relationship between nominal licensing and case assignment.
In Mongolian and quite a few other languages, the subject of an embedded finite clause can receive case from a matrix assigner. They can also move into the matrix clause, i.e. they can hyperraise into it. Both of these operations challenge common assumptions about Phasehood (as well as other assumptions about Activity and case assignment. I propose that this phenomenon in Mongolian requires the movement of the embedded finite subject through Spec-CP. Being at the edge of a phase, it can be visible to a matrix case assigner, as well as to a movement trigger. What is striking about Mongolian hyperraising it can be demonstrated to be an instance of A-movement. This conclusion in turn poses a challenge to conventional assumptions about the A/A-bar distinctions, according to which a syntactic position like Spec-CP is inherently an A-bar position.
My M.A. thesis is a study of different types of gerund clauses in Brazilian Portuguese. I propose that their different properties can be captured if they have different sizes. I show that this proposal can also account for gerund clauses in adjunct position.