Joint Colloquium

The Department of Linguistics and the Department of Computer Science at The University of Arizona invite you to a colloquium presentation by
Speaker: Sandiway Fong
NEC Laboratories America
Princeton, NJ
Topic: Minimalist Parsing: Probes and Goals
Date:Tuesday, March 11th, 2003
Time:12:30 PM
Place:Gould-Simpson Bldg., Room 701
Refreshments will be served in the 7th floor lobby at 12:00 PM


Recently, there has been a shift in the structure of linguistic theories of narrow syntax from abstract systems of declarative rules and principles, e.g. Chomsky (1980), to systems where design specifications call for efficient computation within the human language faculty. In particular, recent work in the Minimalist Program, e.g. Chomsky (1998, 1999), has highlighted the role of locally deterministic computation in the construction of syntactic representation. Instead of a system involving Spec-Head agreement, Chomsky re-analyzes the Case-agreement system in terms of a system of probes, both phi-complete and phi-incomplete, targeting goals, i.e. referential and non-referential elements such as there-type expletives. Under this system, probe-goal agreement can be long-distance and in situ, divorced from considerations about movement (separately motivated by EPP properties).

Efficient assembly (i.e. locally deterministic computation) from a generative perspective with respect to (bottom-up) Merge does not guarantee that parsing with probes and goals will also be similarly efficient. This paper describe a implemented system that handles a range of examples described in Chomsky (1998) and (1999). In particular, it explores the computational and empirical properties of the probe-goal system from a left-to-right, incremental parsing perspective. Instead of Merge as the primitive combinatory operation, we describe a system driven by elementary trees composition with respect to a range of heads in the extended verb projection (v*, V, c and T). Elementary tree composition is an operation that is an important sub-component of Tree-Adjoining Grammars (TAG), Joshi & Schabes (1997).

The system is an on-line incremental parser. To minimize search, in such a system, there is not only no lookahead, but there can also be no "lookback" in the sense of being able to examine or search the derivational history. Instead, we make use of two novel devices with well-defined properties: a "move box" that encodes the residual properties of CHAINs and theta theory, and a single or current "probe box" to encode structural Case assignment mediated by Phase boundaries. Examples from Chomsky (1998) and (1999) will be used to illustrate the empirical properties of these computational elements.

Chomsky, Noam (1980) Lectures in Government and Binding, Foris.
Chomsky, Noam (1998) Minimalist Inquiries, MIT Working Papers in Linguistics.
Chomsky, Noam (1999) Derivation by Phase, MIT Working Papers in Linguistics.
Joshi, Aravind and Yves Schabes (1997) Tree-Adjoining Grammars, in Handbook of Formal Languages, Vol. 3, ed. by Grzegorz Rosenberg and Arto Salomaa, pp. 69-123, Springer-Verlag.

ABOUT THE SPEAKER: Sandiway Fong received his BSc(Eng) in Computer Science from the Imperial College of Science and Technology, University of London, in 1984. At MIT, he received his SM in 1986 and his PhD in 1991 in Artificial Intelligence under the supervision of Robert Berwick. He was employed from 1991 to 2002 as Research Scientist at NEC Research Institute, Princeton NJ, and since 2002 as Research Staff Member at NEC Laboratories America, Inc., Princeton NJ. He was a consultant on natural language parsing in Japanese for a Japanese Ministry of Education, Science and Culture project at Kanda University, Chiba, Japan from 1998 to 2000, and since 2000 has been a collaborator with Anna Maria di Sciullo at UQAM on an SSHRC Canada grant, the "Asymmetry Project". Sandiway has published widely on lexical semantics, including WordNet, morphology, syntax, parsing, and machine translation. He describes his work as "at the intersection of computer science, formal linguistics and cognitive science".