I’m a linguist.
I hold the position of Associate Professor in the Department of Linguistics at UMD.
My research interests span the linguistic sub-fields traditionally identified as syntax and morphology.
I work on phenomena that resist explanation in terms of sound and/or meaning. (If you are unfamiliar with linguistics, the very idea that such phenomena exist might strike you as a little bit counter-intuitive. But phenomena of this sort are surprisingly common in natural language!)
Modularity, and the difference between syntax and semantics
Lately, I have become particularly interested in issues of grammatical modularity. Specifically, I have come to believe that much of what passes for “syntax” these days is actually semantics, masquerading as syntax. (Or, if you prefer: semantics, obliquely described using the vocabulary of syntactic theory.) The idea of a transparent & reliable mapping between syntax and semantics is an indispensable methodological heuristic; but as a bona fide principle of grammar, it is obviously false. And yet much of contemporary syntactic theory is predicated on this very assumption.
I think we can do better; and I think agreement and case hold the key to this. That’s because agreement and case are hierarchy-sensitive phenomena that cannot be reduced to interpretation, and thus, provide an ideal window into that which is quintessentially syntactic.
A non-Saussurean approach to language
Relatedly, I’ve been thinking a lot about the nature of the atoms of linguistic computation. Traditional lexicalist notions (like word, or even morpheme) have proven inadequate for capturing the relation between syntax and morpho-phonology. While it is less often remarked upon, such notions are every bit as inadequate when it comes to the relation between syntax and semantics.
My conclusion is that we should reject once and for all the Saussurean, semiotic basis upon which linguistic theory has traditionally been built. Syntactic atoms certainly form the basis for the mapping to morpho-phonology, but that mapping is far from transparent. They also form the basis for the mapping to semantics, but that mapping is far from transparent, too. Most importantly, the two mappings are not always idiosyncratic in the same ways, or even in the same places.
To put it bluntly: ideas like “morphemes have meanings” or “meanings have spellouts” are illusory. They are on par with a claim like “the form of the English past-tense is fully regular”: it might be true for a large trunk of cases, but it is categorically false as it regards the underlying design of the system. (And the degree to which it is approximately true is likely an artifact of learning pressures.) In reality, morphemes are context-sensitive spellouts of bits of syntactic structure, and meanings are context-sensitive interpretations of bits of syntactic structure. And sometimes, those bits even align. But since that tendency is quite often counter-exemplified, we shouldn’t let it mislead us about the true nature of the system.
Some other topics I am interested in are:
- predicate-argument agreement
- nominal case
- the Person Case Constraint (PCC)
- clitic doubling
- head movement
I work on various (and often unrelated) languages, including: Basque, Icelandic, English, Hebrew, Kaqchikel, Q’anjob’al, Sakha, Kinyarwanda, Shi, Oromo, and Georgian.
See my research page for further details.
And now, a linguistic Dinosaur Comic:
(made using the blank Dinosaur Comic template, available here)