[Prev][Next][Index]

CSLI Calendar, 3 May, vol 5:26



       C S L I   C A L E N D A R   O F   P U B L I C   E V E N T S
-----------------------------------------------------------------------------
3 May 1990                     Stanford                        Vol. 5, No. 26
_____________________________________________________________________________

    A weekly publication of the Center for the Study of Language and
Information (CSLI), Ventura Hall, Stanford University, Stanford, CA 94305-4115
			     ____________

	    CSLI ACTIVITIES FOR THIS THURSDAY, 3 MAY 1990

12:00 noon		TINLunch
      Cordura 100	Fuzzy Logic and its Applications
			Lotfi A. Zadeh
			Computer Science Division
			University of California, Berkeley
			(zadeh@ucbarpa.berkeley.edu)
			Abstract in last week's Calendar

 2:15 p.m.		CSLI Seminar
      Cordura 100	Controversies in Natural-Language Research 3
			led by Stanley Peters 
			(peters@csli.stanford.edu)
			Abstract below

 4:15 p.m.		CSLI Colloquium
      Cordura 100	Metaphor Comprehension by Neural Networks
			Kouichi Doi
			Department of Electrical Engineering
			University of Tokyo
			(doi@mtl.t.u-tokyo.ac.jp)
			Abstract in last week's Calendar
			     ____________

	    CSLI ACTIVITIES FOR NEXT THURSDAY, 10 MAY 1990

12:00 noon		TINLunch
      Cordura 100	Logical Consequence and Model-Theoretic
			Consequence, I: Etchemendy's Critique of the
			Tarskian Analysis
			Greg O'Hair
			Visiting Scholar from The Flinders University
			of South Australia
			Abstract below

 2:15 p.m.		CSLI Seminar
      Cordura 100	Controversies in Natural-Language Research 4
			led by Stanley Peters 
			(peters@csli.stanford.edu)
			Title: Derivation vs. Constraints in Syntax
			Speakers: Joan Bresnan, Ivan Sag, Peter Sells
			(bresnan@csli.stanford.edu, sag@csli.stanford.edu,
			sells@csli.stanford.edu)
			Abstract below
			     ____________
				   
			 THIS WEEK'S SEMINAR
	     Controversies in Natural-Language Research 3
			led by Stanley Peters

At this week's seminar, Ron Kaplan, Lauri Karttunen, Martin
Kay, Paul Kiparsky, and Bill Poser will discuss the following two
questions, among others raised by the previous two seminars.

(1) Finite-state algorithms for morpho-phonological analysis are very
    successful in performing rapid analysis of a large class of words
    using limited memory.  If they are essentially connected with
    Chomsky and Halle's SPE framework, how can they be successful over
    so broad a range of phenomena and languages? 

(2) Templatic morphology and autosegmental phonology are very
    successful in expressing the generalizations that characterize the 
    inventory of words of various languages, more so than the SPE
    framework.  Speaker-hearers can apparently analyze virtually any
    word of their language quickly using little short-term memory.
    How can such a computationally efficient algorithm exist for each
    language characterizable by these theories?
			     ____________
				   
			 NEXT WEEK'S TINLUNCH
       Logical Consequence and Model-Theoretic Consequence, I:
	    Etchemendy's Critique of the Tarskian Analysis
			     Greg O'Hair

This is the first of two talks about John Etchemendy's new book, _The
Concept of Logical Consequence_ (Harvard, 1990).

It is widely assumed that Tarski's model-theoretic account of logical
truth and logical consequence captures the corresponding intuitive
notions.

Etchemendy argues, I believe soundly, that this is not so: The
Tarskian analysis lacks theoretical or conceptual justification and is
not in general extensionally adequate.  In this talk, I examine his
arguments.
			     ____________
				   
		       NEXT WEEK'S CSLI SEMINAR
	     Controversies in Natural-Language Research 4
		 Derivation vs. Constraints in Syntax
		 Joan Bresnan, Ivan Sag, Peter Sells

The most widespread conception of syntactic levels is the derivational
one, which characterizes both Government-Binding Theory and Relational
Grammar, but not the constraint-based theories being developed here at
Stanford.  We will focus on some issues that can differentiate the two
conceptions.  Each week, all of the participants will briefly present
alternative views within a specified domain, and discuss them.

Next week, 10 May, we will look at the data analyzed within
Government-Binding Theory by verb-movement, concentrating on the
implications of this for generalizations about constituent order, and
the syntax/morphology relationship, and considering other ways of
accounting for the relevant variation.

The following week, 17 May, we will address the issue of how to
characterize the various levels of structure needed in syntax: Should
there be uniformity across levels, as in RG and GB?  Should the
underlying levels of structure of UG be characterized in the formal
categorial structure of familiar standard European languages (INFL,
VP, etc.), as in GB?  And generally, how abstract is/are underlying
levels of syntactic representation?

The week after that, 24 May, we will have a "panel discussion" of
issues arising out of the previous weeks' presentations.
			     ____________
				   
			SYMBOLIC SYSTEMS FORUM
	     The Neurophysiology of Conscious Experience
			      Ben Libet
		       Department of Physiology
	       University of California, San Francisco
		      Thursday, 3 May, 4:15 p.m.
			Building 60, Room 61G
         
The special requirements for cerebral activity to elicit a conscious
subjective experience have been experimentally addressed, with
intracranial studies in human subjects.  How the brain distinguishes
conscious from unconscious mental functions was also investigated.
The findings and their implications will be presented and discussed.
         
Next week, 10 May: The Problem of Idle Time, B. A. Huberman, Xerox
Palo Alto Research Center.
			     ____________
				   
		 LANGUAGE ACQUISITION INTEREST GROUP
			  Catherine Nichols
		      Friday, 4 May, 12:00 noon
		     Building 100, Greenberg Room

Catherine Nichols, a postdoctoral fellow in Psychology, will present
her dissertation in this workshop.
			     ____________
				   
		   PHILOSOPHY DEPARTMENT COLLOQUIUM
       The Rationality and Objectivity of Scientific Inference
			     John Earman
		       University of Pittsburgh
		       Friday, 4 May, 3:15 p.m.
			Building 90, Room 91A

No abstract available.
			     ____________
				   
		  LINGUISTICS DEPARTMENT COLLOQUIUM
		       Verb Agreement in Hindi
			     Tara Mohanan
		       (tara@csli.stanford.edu)
		       Friday, 4 May, 3:30 p.m.
			     Cordura 100

The verb in Hindi agrees with its nominative subject.  If the subject
is nonnominative, the verb agrees with its nominative object.  If the
object is also nonnominative, the verb is in the neutral or
nonagreeing form (third person masculine singular).  The
verb-agreement principle can be stated as: the verb agrees with its
highest nominative argument.
	
Hindi has an incorporation construction in which an uninflected object
or locative nominal forms a lexical compound with the verb.  Yet, when
the subject is not nominative, the verb agrees with this noun if it is
nominative.  Thus, a verb agrees with a noun inside the verb compound.
This situation raises questions about the notion "lexical."  A
related problem appears in the construction involving complex
predicates made up of a nominal host and a light verb: if the subject
is nonnominative, the light verb can agree with its host.  Given that
the light verb and the nominal host together form the predicate of the
clause, this amounts to saying that the predicate agrees with an
entity internal to the predicate.  In both these constructions, an
entity that is part of the predicate must at the same time be an
independent argument, so that it can control verb agreement, thus
creating an apparent contradiction.
	
In this talk, I will provide an analysis of agreement in the
incorporation and complex predicate constructions that yields a
solution to this apparent contradiction, in terms of a theory that
factors apart the representation of argument structure and grammatical
category structure, and allows for mismatches between the structures.
The notion "lexical item" may be expressed either as "categorial
word," which is a word-like unit in category structure, or as
"functional word," which is a word-like unit in argument structure
(and grammatical function structure).  These two notions need not
always coincide.  Thus, a categorial word may correspond to more than
one functional word, and a functional word may correspond to more than
one categorial word. This analysis derives from a conception in which
morpho-syntactically relevant semantic information, valency
information, grammatical function information, and grammatical
category information are factored apart into four copresent levels of
structure, simultaneously accessible to principles of grammar.
			     ____________
 
			  PHONOLOGY WORKSHOP
			   Jared Bernstein
			  SRI International
			 (jar@speech.sri.com)
		      Tuesday, 8 May, 7:30 p.m.
			      Ventura 17

No abstract available.
			     ____________
				   
		   PSYCHOLOGY DEPARTMENT COLLOQUIUM
	   Some Cognitive Aspects of Individual Differences
		      in Attachment Organization
			      Mary Main
		  University of California, Berkeley
		     Wednesday, 9 May, 3:45 p.m.
			Building 420, Room 050

No abstract available.
			     ____________