Mathematical properties of natural language.
I am interested in mathematical characterizations of properties of natural languages. For instance, it is an open question whether English is in the class of context-free languages, though it is generally believed that some natural languages (e.g. Zurich Swiss German) are not context free. Geoff Pullum and I have a recent paper in Linguistics and Philosophy discussing a potential argument for non-context-freeness from the “X or no X” construction, an apparent case of syntactic reduplication. If there were a syntactic identity constraint on the two Xs, this would ensure that (the syntax of) English is not context free. However, we show that the relevant constraint is semantic, not syntactic.
- Pullum and Rawlins 2007: Argument or no argument?, Linguistics and Philosophy vol. 30 no. 2.
A long-standing interest is definite descriptions, especially “of”-possessives (the side of the cube, the sister of a famous linguist). This construction is unusual on both familiarity and uniqueness theories of definite descriptions — it seems to be exceptional on standard formulations of either. It can, in some cases, appear in existential constructions and antecede donkey pronouns. I have a paper in the WCCFL 25 proceedings, and several manuscripts.
- In preparation, “Of”-possessive pivots. (should appear here soon)
- Rawlins 2006a: Possessive antecedents to donkey pronouns, Proceedings of WCCFL 25.
- Rawlins 2005: Possessive Definites and the Definite Article, Qualifying paper, UCSC.
I have a continuing interest in the structure of Iroquoian languages, especially Mohawk; my work in this area has centered on the prosody of Lake Iroquoian languages. My qualifying paper explores the stress/epenthesis interaction in Mohawk, which is notoriously complex. The classical derivational analysis (due to Postal and Michelson) involves a significant amount of opacity, and this has been extremely challenging to deal with in constraint-based frameworks. Previous constraint-based analyses have tended to rely on highly non-standard assumptions: empty nuclei, discontinuous feet, and free moras. I show that the complicated set of data can follow straightforwardly if Mohawk is actually a moraic trochaic system, with feet alternating between a single heavy syllable and a light-light trochee.
While I have done little semantic work on Mohawk, this is part of my longer-term research plans.
- Rawlins 2006b: Stress and Epenthesis in Mohawk, Qualifying paper, UCSC. (I am currently revising this for submission.)