[Prev][Next][Index]

CSLI Calendar, 22 February, vol. 5:18


       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
-----------------------------------------------------------------------------
22 February 1990                    Stanford                   Vol. 5, No. 18
_____________________________________________________________________________

    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 NEXT THURSDAY, 1 MARCH 1990

12:00 noon		TINLunch
      Cordura 100	Reading: Underspecification in Phonetics
			by Patricia Keating
			Discussion led by Bill Poser
			(poser@csli.stanford.edu)
			Abstract below
			     ____________

			     ANNOUNCEMENT

There will be no TINLunch, CSLI Seminar, or CSLI Colloquium this
Thursday, 22 February.  
			     ____________
				   
			 NEXT WEEK'S TINLUNCH
	       Reading: Underspecification in Phonetics
			 by Patricia Keating
		    (_Phonology_ 5 (1988): 275-92)
		     Discussion led by Bill Poser

Over the past twenty years, there has been an evolution in thinking
about the extent to which information in phonological representations
is specified and at what levels of representation, from classical
generative phonology, in which every feature was fully specified at
every level of representation, through various versions of
autosegmental phonology, in which underlying representations need not
be fully specified, to recent work, which suggests that the input to
the phonetic rules need not be fully specified.  I will discuss
Keating's paper as an overview of
these issues.
			     ____________
				   
	    COMMONSENSE AND NONMONOTONIC REASONING SEMINAR
	     Using Autoepistemic Expansions for Diagnosis
			     Kave Eshghi
	    Hewlett-Packard Laboratories, Bristol, England
		   Thursday, 22 February, 3:15 p.m.
		       Margaret Jacks Hall 352
				   
An extension of the DeKleer and Reiter's theory of diagnosis is
discussed that allows nonmonotonic constructs in the system
description.  This extension uses the negation as failure rule of
logic programming.  The diagnoses generated correspond to stable
expansion of the corresponding autoepistemic theories.
			     ____________

		 SEMINAR ON ISSUES IN LOGICAL THEORY
			    Philosophy 396
			     David Israel
		      (israel@csli.stanford.edu)
		   Thursday, 22 February, 3:45 p.m.
			     Cordura 100
			     ____________

David Israel will present materials on proof theory and meaning,
inspired by chapter III.8 of the _Handbook of Philosophical Logic_.
			     ____________

			SYMBOLIC SYSTEMS FORUM
	       Intention, Commitment, and Communication
			      Phil Cohen
	   SRI International and Department of Linguistics
			 (pcohen@ai.sri.com)
		   Thursday, 22 February, 4:15 p.m.
			Building 60, Room 61G
         
It is often said that language is a special case of human action;
theories of individual action should have something interesting to say
about how language is used.  This talk will sketch an analysis of
intention and action, will apply that analysis to something as simple
as a request, and then show why it fails to capture some common
properties of dialogue.  I will conclude with a brief theory of joint
(i.e., collective or collaborative) activity, and show why it is a
better basis for describing dialogue than speech act theory.

Next, 1 March, Pat Hayes, Xerox PARC and Department of Computer
Science, will talk.  Title: Time Points or Time Intervals? Problems in
Axiomatizing Common Sense.
			     ____________

	PHILOSOPHY DEPARTMENT AND ETHICS IN SOCIETY COLLOQUIUM
		  Inequalities of Income and Wealth
			   Joseph Stiglitz
		       Department of Economics
		    Friday, 23 February, 3:15 p.m.
			Building 90, Room 92A

No abstract available.
			     ____________
				   
		COMPUTER SCIENCE DEPARTMENT COLLOQUIUM
		      Program Synthesis in Large
			      Enn Tyugu
	     Estonian Academy of Sciences, Tallinn, USSR
		    Friday, 23 February, 3:15 p.m.
		       Margaret Jacks Hall 252

Some years ago, a program synthesizer was implemented in which
data-flow schemas with higher-order objects were used as program
specifications.  It turned out that a class of higher-order data-flow
schemas was equivalent to a fragment of intuitionistic propositional
logic.  Furthermore, several other classes of constructive
propositional theories appeared to be useful for automatic program
synthesis due the existence of fast proof-searching algorithms for
them.  Recently, a fine classification of such theories, together with
derivation complexity estimation, was given by M. Kanovich.

In this talk, logic and architecture of several programming
environments with automatic program synthesis based on propositional
logic is considered.  Application examples, including CAD, data logic,
and compiler construction are discussed.
			     ____________

		  LINGUISTICS DEPARTMENT COLLOQUIUM
	   Nominal Comparatives and Generalized Quantifiers
			    John Nerbonne
		     Hewlett-Packard Laboratories
		       (nerbonne@hplabs.hp.com)
		    Friday, 23 February, 3:30 p.m.
			     Cordura 100

There's an extensive literature on the semantics of _adjectival_
comparison, but much less on _nominal_ comparatives (i.e., of the sort
illustrated below).  Keenan and Stavi investigate these fairly
abstractly from a generalized-quantifier perspective, demonstrating,
e.g., conservativity (and adducing an interesting class of ternary
determiners).  But their approach is broad and systematic; comparative
determiners are syncategorematic.

In the work presented here, we likewise adopt the GQ perspective, but
focus on microstructure of comparative determiners.  We begin with a
domain of discourse upon which a lattice structure (Link's) is
imposed, and which maps, first to a domain of abstract dimensions
(cardinalities, volumes), and then to concrete measures (in N,R+).
The mapping must be Archimedean.  Comparisons begin as simple
predicates on dimensions or measures; from these we derive relations
(classes) of predicates on the domain, i.e., generalized determiners
(quantifiers), and show, e.g., how monotonicity properties follow in
the derivation.  This results in a proposal for a logical language,
which includes _derived_ determiners, and which is an attractive
target for semantics interpretation.

   More (fewer) than seven children sang.
   How many children sang?
   A trained seven more (fewer) children than B saw (dogs).
   A trained twice as many (few) children as B saw (dogs).
   A trained at least twice as many (fewer) children as B saw (dogs).

   More (less) than two liters of water spilled.
   How much water spilled?
   How many liters of water spilled?
   A spilled two liters more (less) beer than B drank (water).
   A spilled twice as much (little?) beer as B drank (water).
   A spilled at least twice as much (little?) beer as B drank (water).
			     ____________
				   
	    COMMONSENSE AND NONMONOTONIC REASONING SEMINAR
	      General Patterns of Nonmonotonic Reasoning
			    Daniel Lehmann
			  Hebrew University
		    Monday, 26 February, 2:30 p.m.
		       Margaret Jacks Hall 252

Recent developments in the area of nonmonotonic reasoning have
converged onto a number of specific properties of nonmonotonic
inference operations.  Nonmonotonic inference operations appear as a
natural generalization of Tarski's consequence operations.  The need
and justification for considering such a generalization will be
discussed and some of the recent results reviewed.
			     ____________
				   
		SEMINAR ON COMPUTERS, DESIGN, AND WORK
	      Participatory Design and User Programming
		    Jeff Johnson and Bonnie Nardi
		Human-Computer Interaction Department
		     Hewlett-Packard Laboratories
		  Wednesday, 28 February, 12:15 p.m.
			      Ventura 17

Participatory Design is a new approach to designing computer systems
that places great emphasis on getting the eventual users of a system
involved in its design.  It goes beyond the traditional design wisdom
of constructing prototypes and conducting user-tests, i.e., "test
early and often."  Instead of being regarded as mere information
sources for designers, users are considered their _peers_, as experts
in the task domain, the expertise that matters most, rather than in
the technology.  Participatory Design embodies methodologies that
allow technical experts to work together with task-domain experts to
produce designs.  Without Participatory Design, systems result that
are satisfactory neither to users, who must use the system every day,
nor to their employers, who are presumably adding computer systems to
realize productivity gains.

Participatory Design has been more successful in Europe, particularly
in Scandinavia, than it has been in the United States, where the
techniques are less well known.  It has also been applied, thus far,
more to the design of custom systems for particular work sites than to
the design of "off-the-shelf" applications and computer-based
appliances sold by vendors.

To promote greater understanding of Participatory Design in the U.S.
and to explore its applicability to the design of "off-the-shelf"
applications, Computer Professionals for Social Responsibility (CPSR)
is sponsoring a conference, PDC '90.  CPSR is also sponsoring a panel
on Participatory Design at the upcoming ACM SIGCHI '90 conference,
which immediately follows PDC '90 in Seattle.  Jeff Johnson is a
cochair of PDC '90 and the organizer of the panel.  Bonnie Nardi is
one of the panelists.

Jeff will introduce Participatory Design and the notions of
"task-specific applications" and "end-user programming," and Bonnie
will discuss end-user programming as a possible approach to
Participatory Design.
			     ____________
				   
		   PSYCHOLOGY DEPARTMENT COLLOQUIUM
	       Self-Affirmation: Applications to Social
		   Theory and Minority Achievement
			    Claude Steele
			University of Michigan
		  Wednesday, 28 February, 3:45 p.m.
			Building 420, Room 050

No abstract available.
			     ____________
				   
		     SITUATION SEMANTICS SEMINAR
			      Questions
			  Jonathan Ginzburg
		     (ginzburg@csli.stanford.edu)
		  Wednesday, 28 February, 4:00 p.m.
			     Cordura 100

This will be the start of a series of meetings in which we discuss a
new topic: Questions.  In the first meeting, I will present a review
of some previous approaches to questions, within other semantic
frameworks.
			     ____________

	     ANNUAL MEETING OF THE WEST COAST CONFERENCE
			ON FORMAL LINGUISTICS
			 Stanford University
		  Friday, 2 March - Sunday, 4 March

The 9th Annual Meeting of the West Coast Conference on Formal
Linguistics (WCCFL) will be held at Stanford University, hosted by the
Department of Linguistics, and coordinated by Peter Sells.  

WCCFL was founded at Stanford in 1982.  This year, forty-two papers on
all aspects of formal linguistics will be presented.  There will be
speakers from all over the U.S., Canada, and Western Europe, as well
as five speakers from the Stanford community.  Meetings will be held
at CSLI during Friday, 2 March, and in Jordan Hall during Saturday and
Sunday, 3 and 4 March.

The schedule for the conference is readable online in
/user/wccfl/schedule; for registration information, and so on, please
send mail to wccfl@csli.stanford.edu (registration fees for the
Stanford community are $10 for students and $15 for nonstudents).
			     ____________