Semantics research is about how the meaning of a sentence is determined from its parts and the way the parts are put together. Semantics at Penn focuses on several new approaches to the field, including LTAG semantics and underspecification as well as the application of game theory.
Florian Schwarz and Robin Clark lead Penn's research in formal semantics, mathematical linguistics, and computational semantics. Schwarz's main research interests are in formal semantics and pragmatics of natural language, as well as their relationship to each other, in particular with respect to the role of context in interpretation. He has worked, most recently, on the analysis of different types of definite articles in German, couched within a situation semantics. This work also bears on our understanding of various types of implicit content, e.g., the issue of domain restriction and implicit arguments, and on theories of quantification and covariation. Other topics he has worked on include the syntactic encoding of information structure in Kikuyu and the analysis of intensional transitive verbs (e.g., need in I need a beer). In addition to his theoretical work, he is engaged in several types of research that provide new empirical perspectives on semantics and pragmatics, and is in the process of setting up a lab for the experimental study of meaning-related phenomena. His work in this area has looked at, among other things, the processing of pragmatic content, such as presuppositions and implicatures. Furthermore he has begun to develop, in collaboration with Chris Potts (Stanford), an approach for investigating expressive aspects of meaning and their relation to context using large online corpora.
Clark has been pursuing a line of research that began as an attempt to address the issue of language learnability. About a decade ago his work on Kolmogorov complexity led him to an automata-theoretic characterization of quantifiers. He then needed to confront the fact that, while first-order quantifiers can be simulated with finite-state machines, more complex quantifiers require more complex machines and more memory. Since it seems likely that processing different types of quantifiers might involve different parts of the brain, Clark is collaborating with Murray Grossman of the School of Medicine to test that hypothesis experimentally.
His unique perspective on linguistics has led him to pursue a new game-theoretic approach to the semantics of quantifiers and more generally a model of meaning rooted in the cooperative interaction between social agents. He is now collaborating with Prashant Parikh on the analysis of discourse anaphora using classic game theory; they have extended the game-theoretic analysis to other semantic and pragmatic problems. Clark introduces his students to important alternative approaches to semantics, including game theory and Categorial Grammar, which has been advanced by former Penn professor Mark Steedman, now at the University of Edinburgh.