Main area of research
Natural language & AI
Scott Martin works on linguistically informed approaches to natural language processing and understanding in the Linguistic Technologies group. His research interest is in the area of Natural Language & AI. In the past, he has done research in the fields of syntactic and semantic parsing, paraphrase alignment, and generation, with funding from DARPA and the National Science Foundation. His broader research interests are in computational, formal, and mathematical approaches to modeling natural language, especially discourse semantics and pragmatics. He holds both an M.A. and Ph.D. in Linguistics from Ohio State University, where he was a presidential fellow. Scott is also an award-winning software engineer and is one of the main authors of OpenCCG, an open source natural language parser and realizer.
I extend the recent unidimensional semantics of supplements due to Martin to a full syntax-semantics interface. The grammar formalism employs a two-component syntax, with one …
Read more
We present a compositional, dynamic categorial grammar for discourse analysis that captures the core insights of dynamic semantics: indefinites do not quantify but introduce discourse …
Read more
This thesis is a both a descriptive and theoretical examination of implicatures, parts of the contextual meanings of utterances that are separate from their sense, …
Read more
Monolingual alignment is frequently required for natural language tasks that involve similar or comparable sentences. We present a new model for monolingual alignment in which …
Read more
The accessibility constraints imposed on anaphora by dynamic theories of discourse are too strong because they rule out many perfectly felicitous cases. Several attempts have …
Read more