In my research I deal primarily with the issue of cross-linguistic variation in syntactic phenomena. More specifically, my concern is how much of this variation stems from parametric differences in syntax proper and how much of it is shaped by the interaction of syntax with its interfaces: semantics (LF) and morphology/phonology (PF). Some of the phenomena I have looked at in the past are:
Below you can find descriptions of these phenomena, and my work in relation to them with suggested papers and presentations.
(... not all descriptions are there yet, but should be soon)
Person is the grammatical category that distinguishes "me" and "us" (1st person) from "you" (2nd person) and both from "him", "her", "it" and "them" (3rd person). In most current theories of syntax, person is assumed to be encoded on words through phi-features, which group person together with other grammatical categories like number and gender. However, in contrast to these other categories, the occurrence of specific person values is curiously restricted in some syntactic contexts.
An example of this is the Person-Case Constraint (or PCC), a restriction that affects co-occurring "weak" object pronouns. In French, one can say "il me l'envoya" ("he sent him to me") or "il la leur envoya" ("he sent her to them"), but not "*il me t'envoya" ("he sent you to me") or "*il me lui envoya" ("he sent me to him"). That is, when the two weak objects co-occur, the indirect object can have any person value (e.g. "to me/to you/to him"), but the direct object can only be 3rd person (e.g. "him/her"). In order to express the banned combinations, a weak pronoun can be replaced by a "strong" one, as in "il me envoya a lui" ("he sent me to him").
Another case of a syntactic person restriction are Person Hierarchy Effects. In English, the grammatical role of an argument (e.g. subject vs. object) determines its position in a clause ("I saw them" vs. "They saw me"), but there are also languages where an argument's position is determined by its person value. Thus, in Ojibwe/Nishnaabemwin (an Algonquian language) "n-waabm-aa-g" corresponds to "I see them" and "n-waabm-igoo-g" corresponds to "They see me". Notice that unlike in English, the marker corresponding to "I/me", "n-", always precedes the verb ("waabm"), whereas the other marker "-g" always follows it. Their respective grammatical roles are encoded via theme signs ("-aa" vs. "-igoo"), but their position is entirely determined by their person value.
In my work, I focus on broadening the cross-linguistic typology of syntactic person restrictions by looking at previously unnoticed patterns and typological gaps. The goal is to identify the core properties of all the different syntactic person restrictions in order to determine whether their existence can be attributed to a single underlying phenomenon. My claim is that it can.
The basic idea is that not all pronouns inherently have a person value: some only acquire one during syntactic derivation. Such pronouns must get their person value from a syntactic head that introduces valued person phi-features into the derivation. A syntactic person restriction then arises due to the locality restrictions on valuation that come into play when multiple pronouns must receive a person value from a single syntactic head - a pronoun that cannot receive a value in such configurations is spelled-out as 3rd person.
This approach can explain both the PCC and Person Hierarchy Effects. With the PCC, a pronoun is restricted to 3rd person because another pronoun intervenes between it and the source of valued person phi-features. In the case of Person Hierarchy Effects, the same kind of intervention is voided by pronoun movement that changes the position of pronouns in relation to the source of valued phi-features - resulting in person-sensitive word orders. The cross-linguistic variation in restriction patterns can then be derived by varying only the position of the source of valued phi-features in relation to the relevant pronouns and where and when the pronouns can move in the course of a derivation.
One of the previously unnoticed restriction patterns I discuss in my work is the Reverse PCC, which differs from the canonical PCC in that the person restriction applies to the indirect object (IO) as opposed to the direct object (DO). Because the Reverse PCC is sensitive to a very specific kind of pronoun order alternation, I propose that this kind of person restriction arises due to pronoun movement before the source of valued person phi-features enters the derivation. With the canonical PCC, the IO is the intervener for person valuation, so the DO is person restricted, but with the Reverse PCC, it is the moved DO which prevents the IO from getting a value.
The Reverse PCC is also important because it shows that case does not play a role in the PCC, as person restrictions can apply to either accusative (canonical PCC) or dative pronouns (Reverse PCC). Furthermore, the Reverse PCC is cross-linguistically only observed in languages where the pronoun movement that yields the pattern is optional and can be compared to scrambling.
Syntactic person restrictions come in different varieties. As seen above with the PCC in French and the Person Hierarchy Effect in Ojibwe/Nishnaabemwin, the former operates on object-object pairs and the latter on subject-object pairs. In addition to this kind of variation, syntactic person restrictions also differ cross-linguistically in terms of "strength", that is: how many pronoun combinations they exclude. But these two points of variation are not entirely independent of each other. Whenever restrictions of the subject-object kind and restrictions of the object-object kind co-exist in a language, the former is never "stronger" than the latter.
I offer an account of this in terms of my general approach to syntactic person restrictions. I propose that "weak" person restrictions occur when at least one of the relevant pronouns is person-valued in the specifier of the syntactic head imbued with valued person phi-features; this allows for more valuation options and hence the derivation of more pronoun combinations. "Strong" restrictions, on the other hand, occur when the relevant pronouns in are all valued in a long distance relation with the source of valued person phi-features; because of the intervention this results in, less pronoun combinations are possible.
The generalization follows from the universal order in which arguments enter the syntactic derivation (direct object, indirect object, subject) in conjunction with the limited possibilities for person valuation given a single source of valued person phi-features and three pronouns: (i) all pronouns valued in specifiers (only weak restrictions), (ii) only top pronoun valued in the specifier (weak restriction for subject-object, strong restriction for object-object), and finally (iii) all pronouns are valued long distance (only strong restrictions). Because this exhausts all the configurations in which three pronouns can be valued, it is impossible to have a language in which the person restriction for subject-object pairs will be stronger than the one for object-object pairs.
[ Description coming soon ... ]
[ Description coming soon ... ]
[ Description coming soon ... ]