Research
I try to make all my work freely available online, including published papers, manuscripts and slides of my talks, plus the respective bibtex entries.
For Linguists
While my work spans two rather different areas --- computational phonology on the one hand, the formal study of linguistic constraints on the other --- I am guided in both cases by a profound curiosity to which degree linguistic proposals differ beyond mere notational quirks. That is to say, I am interested in what cognitive claims are embodied by linguistic theories. In the area of computational phonology, this has led to interesting insights concerning the relationship of SPE and Government Phonology and the generative capacity of phonology in general. My work on linguistic constraints has shown that various constraints of higher complexity (e.g. transderivational constraints) can be reencoded by constraints of lower complexity, which begs the question how the complexity of linguistic theories is to be determined.
For Non-linguists
Linguistics
Linguistics is not about languages, it is about language. By language, I mean the inborn ability of humans to acquire the language spoken around them without explicit instruction and use it to create new sentences that have never been uttered before but are nevertheless deemed correct by the speech community. For example, the following sentence should sound okay to all speakers of English, even though it is unlikely that anybody has ever used it before: "It was on a vanilla-colored comfort throw sporting a Bauhaus rendition of Baby Mario that Ymir's legs commenced their mating ritual." The study of the cognitive mechanisms that underly the acquisition and use of language is what makes up linguistics. Therefore, linguistics is way closer to psychology than to the humanities. Mathematical linguistics is a subfield of linguistics that investigates those cognitive mechanisms using concepts from mathematics and computer science. In particular, mathematical linguists try to determine what kind of computational resources our brain has to have at its disposal in order to carry out the computations required for the efficient use of a human language.
My Research
Linguists assume that a sentence is grammatical if it is possible to assign it a special kind of geometric structure that encodes the dependencies holding between the words in the sentence. Humans are endowed with a cognitive ability to determine in a principled way for their native language which structures are assigned to which sentences. Thinking in terms of chemistry, we may view words as atoms and sentences as molecules; a grammatical sentence, then, corresponds to a stable molecule. It is a contentious issue how the labor of creating these molecules is to be distributed between structure-building operations on the one hand and structure-constraining principles on the other. A conclusive answer to this question is currently not in sight, because little is known about the behavior of structure-constraining principles and their psychological reality. I'm trying to bridge this gap using techniques from mathematical logic and theoretical computer science and linking the results to the psychology of language and human cognition in general.
