[Prev][Next][Index]

CSLI Calendar, 09 Nov 1995, vol.11:07




        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
______________________________________________________________________________

9 November 1995                   Stanford                      Vol. 11, No. 7
______________________________________________________________________________

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

               CSLI ACTIVITIES DURING 8 -- 17 NOVEMBER 1995

  WEDNESDAY, 8 NOVEMBER
         3:15 - Philosophy of Computation Seminar
                Ventura Hall, Room 17
                Turing Machines Revisited (II)
                Brian Cantwell Smith, Xerox PARC
                Abstract below

  THURSDAY, 9 NOVEMBER
         4:15 - SSP Forum
                Building 60, Room 61-F
                Designing Applications that Help People Collaborate
                Ellen Isaacs, Sun Microsystems
                Abstract below

  FRIDAY, 10 NOVEMBER
        12:30 - HCI Seminar
                Skilling Auditorium, SITN Channel E1
                The Establishment of Identity in Virtual Communities
                Judith Donath, MIT Media Lab
                Abstract below

         3:15 - Philosophy Colloquium
                Encina Hall, Room 423
                Reason and Actions
                Philip Clark, UC Davis Philosophy

         3:30 - Linguistics Colloquium
                Building 460, Room 146
                On the Need to Recognize Constructions
                Adele Goldberg, UC San Diego Linguistics
                Abstract below

  SATURDAY, 11 NOVEMBER
        10:00 - Phonology Workshop
                287 Dwinelle Hall, UC Berkeley
                Trilateral Phonology Weekend (TREND) V:
                UC Berkeley / Stanford / UC Santa Cruz
                Program below

  MONDAY, 13 NOVEMBER
         2:00 - Semantics Workshop
                Cordura Hall, Room 100
                Anchoring Definites in Discourse
                Kjetil Strand, Oslo Linguistics
                Abstract below

  TUESDAY, 14 NOVEMBER
         7:00 - SSP Film Series
                Cubberley Hall, Room 128
                Daniel Dennett (A Glorious Accident, Part 3) (1993)
                Abstract below

  WEDNESDAY, 15 NOVEMBER
         3:15 - Philosophy of Computation Seminar
                Ventura Hall, Room 17
                Brian Cantwell Smith, Xerox PARC

  THURSDAY, 16 NOVEMBER
        10:45 - STASS Seminar  (NOTE LATER-THAN-USUAL TIME)
                Cordura Hall, Room 100
                Mathematical Water in the Refrigerator
                Keith Devlin, CSLI & St. Mary's College
                Abstract below

        12:00 - Cognitive Science Lunch
                Cordura Hall, Room 100
                Consciousness as an Information Bearing Medium
                Bruce Mangan, UC Berkeley Institute of Cognitive Studies
                Abstract below

         4:15 - SSP Forum
                Building 60, Room 61-F
                Interpreting Discourse
                Henriette de Swart, Stanford Linguistics
                Abstract below

         4:15 - Seminar on Computational Learning and Adaptation
                Cordura Hall, Room 100
                PAGODA: A Model for Autonomous Learning in
                Probabilistic Domains
                Marie desJardins, SRI International
                Abstract below

  FRIDAY, 17 NOVEMBER
        12:30 - HCI Seminar
                Skilling Auditorium, SITN Channel E1
                New Voices, New Visions
                Sally Rosenthal, Interval Research
                Abstract below

         3:30 - Linguistics Colloquium
                Building 460, Room 146
                Reciprocals in Bantu: Quantification in Morphology?
                Sam Mchombo, UC Berkeley Linguistics

                               ____________

The CSLI Calendar appears on Wednesday of each week throughout the academic
year.  Announcements, abstracts, and other information to appear in the
Calendar can be submitted by e-mail to <incalendar@csli.stanford.edu>.

Further information about CSLI and past issues of the CSLI Calendar
are available on the Internet at URL <http://www-csli.stanford.edu/csli/>.
The Calendar is also posted each week to the <csli.bboard> newsgroup.

                               ____________
     
                    PHILOSOPHY OF COMPUTATION SEMINAR
                         on Wednesday, 8 November
                     3:15 p.m., Ventura Hall, Room 17
                      Turing Machines Revisited (II)
                           Brian Cantwell Smith
                    Xerox PARC and Stanford Philosophy
                         [bcsmith@parc.xerox.com]

The reading for today's seminar is Volume III Chapter 1 of _The Middle
Distance_.  A course description can be found at [http://shr.stanford.edu/
BCSmith/phil395a.html]

                               ____________

                          SYMBOLIC SYSTEMS FORUM
                         on Thursday, 9 November
                    4:15 p.m., Building 60, Room 61-F
           Designing Applications that Help People Collaborate
                               Ellen Isaacs
                             Sun Microsystems
                        [Ellen.Isaacs@eng.sun.com]

In this talk, I will describe two multimedia-based prototype applications
developed by our group to help people collaborate when they are distributed
across different locations. I will also discuss the process we used to build
these applications to make sure they are usable and useful for people. The two
applications are Montage and Forum. Montage is a "next-generation" desktop
video conferencing system that focuses on helping people find opportune times
to interact.  Forum lets speakers give live, interactive talks over the
network to participants who watch and ask questions from their desktops. To
build these tools, user interface designers and developers collaborated to use
a highly iterative process of studying use -> designing -> implementing ->
studying use, etc.  We found that the process allowed us to create
easy-to-use, well-architected applications that were gratifying to the
developers, UI designers and users alike -- all without costing extra time or
expense.

ELLEN ISAACS works in SunSoft's Collaborative Computing group, an advanced
development group that studies the use of technology to help people who are in
different locations work together more closely.  Her job involves designing
user interfaces for tools to support group work and studying how groups
incorporate the technology into their work.  Before coming to Sun, Isaacs
worked at Stanford's Knowledge Systems Lab on a project to develop a speech
interface to a medical expert system.  She received her PhD in cognitive
psychology from Stanford, where she studied language use and collaboration in
conversation. She received her BS from Brown University in psychology and
semiotics.

                              _____________

                  SEMINAR ON HUMAN-COMPUTER INTERACTION
                          on Friday, 10 November
                     12:30 p.m., Skilling Auditorium
           The Establishment of Identity in Virtual Communities
                              Judith Donath
                              MIT Media Lab
                          [judith@media.mit.edu]

When you first meet someone, a multitude of details -- many gleaned
subconsciously -- go into creating your impression of the new acquaintance. 
The same is true when you meet someone online, except that there are far fewer
details to observe and the resulting impression is quite sketchy.  With more
and more of our social interactions occuring online, the question of how to
make this virtual impression more vibrant and meaningful becomes increasingly
important.

I will begin this talk by examining how identity is established online today
in Usenet news and Web home pages.  In these technologically constrained
environments, some very interesting ways of establishing (or hiding) one's
personal identity have evolved.  And, for the designer, it is the constraints
which are most interesting -- which ones have had the most impact on the
resulting culture? And which are the most important to remove? I will then
discuss designs for more sociable interfaces, focusing on the design and
implementation of the recent Media Lab / Art Technology Group project: A Day
in the Life of Cyberspace.

JUDITH DONATH is a doctoral student at the MIT Media Lab.  Her research
focuses on the social side of computing -- on the online communities and
virtual identities and computer-mediated collaborations that have emerged with
the convergence of computing and communication.  Prior to entering the doctoral
program she taught interface design, produced multimedia applications and
designed educational software.

                               ____________

                          PHILOSOPHY COLLOQUIUM
                          on Friday, 10 November
                     3:15 p.m., Encina Hall, Room 423
                            Reason and Actions
                               Philip Clark
                           UC Davis Philosophy
                        [pclark@csli.stanford.edu]

                               ____________

                          LINGUISTICS COLLOQUIUM
                          on Friday, 10 November
                    3:30 p.m., Building 460, Room 146
                  On the Need to Recognize Constructions
                              Adele Goldberg
                         UC San Diego Linguistics
                          [aegoldberg@ucsd.edu]

Basic sentence patterns of a language are traditionally taken to be determined
by semantic or syntactic information specified by the main verb in the
sentence. Thus, the sentence patterns given in (1) and (2) appear to be
determined by the specifications of give and put respectively:

        1. Chris GAVE Pat a ball.
        2. Pat PUT the ball on the table.

In this talk I will argue that while (1) and (2) represent perhaps the
prototypical case, sentence patterns of a language are not reliably determined
by independent specifications of the main verb. For example, it is implausible
to claim that sneeze has a three argument sense in (3):

        3. Pat SNEEZED the foam off the cappuccino.

The following attested examples similarly involve sentential patterns that do
not seem to be determined by independent specification of the main verb:

        4. "My father FROWNED away the compliment."
        5. "We LAUGHED our conversation to an end."
        6. "Pauline SMILED her thanks."
        7. "The Miami quarterback was BOO-ED to the bench."

It is argued, on the basis of linguistic and psycholinguistic evidence, that
lexically unfilled phrasal constructions exist and contribute significantly to
the overall semantic interpretation. In addition, a way to capture linking
generalizations across constructions is discussed.

It is suggested that the recognition of phrasal constructions corresponding to
basic sentence patterns leads one toward a view of grammar in which the
construction (or sign): any non-predictable form-meaning pairing, plays a
central role.

Note: Linguistics Colloquium abstracts are available at
[http://www-linguistics.stanford.edu/~kessler/colloq/].

                               ____________

                            PHONOLOGY WORKSHOP
                         on Saturday, 11 November
           starting 10:00 a.m., 287 Dwinelle Hall, UC Berkeley
                  Trilateral Phonology Weekend (TREND) V
                  UC Berkeley / Stanford / UC Santa Cruz
                  [contact: inkelas@cogsci.berkeley.edu]

Program:
   10:00   Coffee and pastries
   10:30   Ronald Sprouse, UC Berkeley
           "Vowels that borrow moras: Geminates and weight in OT"
   11:05   Edward Flemming, Stanford University
           "Vowels undergo consonant harmony"
   11:40   Philip Spaelti, UC Santa Cruz
           TBA (re: reduplication)
   12:15   Andrew Dolbey, UC Berkeley
           "Cyclic allomorph optimization in Sami"
   12:50   Lunch
   2:20    Arman Maghbouleh, Stanford University
           "Detecting prominences using a vowel duration model (building
            blocks for a numerical model of duration)"
   2:55    Chris Albert, UC Santa Cruz
           TBA (re: Turkana vowel harmony)
   3:30    Susanne Gahl, UC Berkeley
           "Mathimathi stress is not onset-sensitive"
   4:05    Happy hour

                               ____________

                            SEMANTICS SEMINAR
                          on Monday, 13 November
                    2:00 p.m., Cordura Hall, Room 100
                   Anchoring of Definites in Discourse
                              Kjetil Strand
                             Oslo Linguistics
                       [kstrand@csli.stanford.edu]

A number of theories on definiteness agree that interpretation of
"the"-phrases relies on one ore more contextual elements to identify the
reference unambiguously.  Spelled out in the framework of DRT, there should be
at least one discourse referent A available in the DRS which the discourse
referent D corresponding to the definite noun phrase should be related to by a
condition L(D,A).  I will refer to the familiar discourse referent A by the
term "anchor", and to the relation L that obtains between D and A by the term
"link".  According to this view, the interpretation process involves both
location of anchors and identification of links, yielding an unambiguous
reading of the definite noun phrase.

In this talk I will focus on the taxonomy of linking relations, as the
computability of the account will depend on the possibility to give precise
descriptions of the different kinds of links involved.  I want to propose a set
of 5 main classes and 20 subclasses of linking relations, and I show how they
distribute in a Norwegian newspaper corpus comprising 1808 "the"-phrases.  I
also present the results regarding the location of anchors.  To my own
surprise, I found as many as 86% of the anchors provided by the text itself,
that is, the anchor(s) were expressed by linguistic means.

                               ____________

                             SSP FILM SERIES
                         on Tuesday, 14 November
                   7:00 p.m., Cubberley Hall, Room 128
           Daniel Dennett (A Glorious Accident, Part 3) (1993)

This 100 minute video is from the 8-part Dutch series, "A Glorious Accident:
Understanding Our Place in the Cosmic Puzzle," produced by Wim Kayzer.  This
series brings together some of the greatest scientific thinkers of our time,
recording their thoughts, their speculations, their questions on man's place
in the cosmic puzzle.  Part 3 features Daniel Dennett, a philosophy professor
>From Tufts University, who is well known for his work in the philosophy of
mind and is, most recently, the author of _Darwin's Dangerous Idea_.

The Symbolic Systems Film Series showcases videos and films of general
cognitive science interest.  Attendance at the films can substitute for
attendance at the Symbolic Systems Forum for students enrolled in SSP 10 for
one unit.  All are welcome at these events.  The showing of the videos is
followed by a discussion, and researchers who are knowledgeable about the
program's topic are urged to join us in evaluating it.

                               ____________

                    PHILOSOPHY OF COMPUTATION SEMINAR
                        on Wednesday, 15 November
                     3:15 p.m., Ventura Hall, Room 17
                           Brian Cantwell Smith
                    Xerox PARC and Stanford Philosophy
                         [bcsmith@parc.xerox.com]

                               ____________

                              STASS SEMINAR
                         on Thursday, 16 November
                    10:45 a.m., Cordura Hall, Room 100
                       (NOTE LATER-THAN-USUAL TIME)
                  Mathematical Water in the Refrigerator
                               Keith Devlin
                         CSLI & St.Mary's College
                        [devlin@csli.stanford.edu]

A common objection to mathematical models of meaning is that they assume words
have fixed meanings, or at most a small finite range of meanings.  In their
book _Understanding Computers and Cognition_, Winograd and Flores use the
example of "water" in the refrigerator to illustrate how a seemingly simple
noun can have an indefinite number of "meanings", depending on the context of
use.  Likewise, in his book _Arenas of Language Use_, Herb Clark provides
numerous examples of "ad hocing", whereby the meaning of a particular word (or
phrase) is determined by the speaker and listener in a given context.
Objections such as these seem to restrict the applicability of relational
models of meaning such as the situation-theoretic models presented in Barwise
and Perry's _Situations and Attitudes_ and my own _Logic and Information_.
However, I recently discovered a relational model of meaning that does capture
ad hocing.  (The model uses standard situation-theoretic apparatus.)  The talk
will describe the model, and illustrate its range with examples from
Winograd-Flores, Clark, and Harvey Sacks.

PLEASE NOTE: The seminar will start 45 minutes later than usual, at 10:45am,
as the speaker is flying back from Reno that morning.  (Any additional delay
should be blamed on Southwest Airlines.)

                               ____________

                              CSLI COGLUNCH
                         on Thursday, 16 November
                    12:00 noon, Cordura Hall, Room 100
              Consciousness as an Information Bearing Medium
                               Bruce Mangan
                UC Berkeley Institute of Cognitive Studies
                       [mangan@cogsci.berkeley.edu]

I propose to formalize the notion of consciousness in a slightly new way:
consciousness is simply one information bearing medium, among many others, at
work in our organism.  In general, scientific analysis of a biological
information bearing medium (e.g., of DNA or the fluid in the cochlea) aim to
answer at least two related but different questions: (1) what information does
the medium bear? and (2) in what specific *way* does the medium bear its
information?  Answering the first question lets us assert that something
belongs to the *genus* information bearing media, answering the second lets us
assert its uniqueness as a subordinate *species*.

Common-sense notions about consciousness usually emphasize its unique aspect,
while cognitive science usually does the opposite, treating consciousness as
if its generic status as an information bearing medium is all we need to
consider.  But each stance tacitly recognizes the other in many cases, and I
believe there is no necessary conflict between them.  In particular I will
claim that (1) consciousness tends to bear information that is relevant to
novel evaluations either expected or at hand; (2) consciousness bears its
information as experience, or qualia.  In doing so, I hope to sharpen the core
dispute between functionalism and its enemies, and try to show a way out of
one research dead around the notion of qualia.

A description and schedule for the CogLunch series on consciousness can be
found at [http://www-csli.stanford.edu/csli/9495reps/coglunch.html].

 Fall Calendar:
 October  12:  Owen Flanagan (Philosophy, Duke U.)
          19:  Ernest Hilgard (Psychology, Stanford U.)
          26:  John McCarthy (Computer Science, Stanford U.)
 November  2:  Fred Dretske (Philosophy, Stanford U.)
           9:  IAP EVENTS (No CogLunch)
          16:  Bruce Mangan (Cognitive Science, UC Berkeley)
          23:  THANKSGIVING (No CogLunch)
          30:  Henry Stapp (Physics, Berkeley Livermore Labs)

                               ____________

                          SYMBOLIC SYSTEMS FORUM
                         on Thursday, 16 November
                    4:15 p.m., Building 60, Room 61-F
                          Interpreting Discourse
                            Henriette de Swart
                           Stanford Linguistics
                       [deswart@csli.stanford.edu]

Many syntactic and semantic theories focus on the structure and meaning of
sentences.  But communication typically involves a sequence of sentences which
hang together in a certain way.  This means that we have to look beyond the
meaning of individual sentences and determine the way they are pieced together
to make a coherent discourse.  In recent years, a number of linguistic
theories have been developed which take the discourse as their primary
syntactic and semantic unit.  In this talk, I give an overview of some of the
general issues involved in the analysis of discourse, such as structure and
use, the role of context and situation, information structuring, etc.  I
illustrate some of the similarities and differences between various approaches
to discourse with the way they treat reference to individuals.  Questions
which arise are: which form does reference to individuals take? how are
discourse anaphora licensed? how do we find the antecedent of an anaphor? The
answers given to these questions reflect the different perspectives adopted in
the study of natural language in human communication.

HENRIETTE DE SWART is an assistant professor in the Linguistics Department at
Stanford University.  She got her PhD from the University of Groningen (the
Netherlands) and was one of the coordinators of the cognitive science program
in Groningen for a year.  She continued to be affiliated with the program when
she obtained a research fellowship granted by the Royal Dutch Academy of
Sciences.  At Stanford, she teaches classes in semantics and pragmatics.  Her
current research focuses on phenomena both at the sentential and at the
discourse level, such as negation, quantification, anaphoric relations, focus,
tense and aspect.

                              _____________

             SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
                         on Thursday, 16 November
                    4:15 p.m., Cordura Hall, Room 100
     PAGODA: A Model for Autonomous Learning in Probabilistic Domains
                             Marie desJardins
                            SRI International
                    [contact: langley@cs.stanford.edu]

PAGODA (Probabilistic Autonomous GOal-Directed Agent) is a model for
autonomous learning in probabilistic domains that incorporates innovative
techniques for using the agent's existing knowledge to guide and constrain the
learning process and for representing, reasoning with, and learning
probabilistic knowledge.  This talk presents an overview of the PAGODA
learning model and describes in detail the probabilistic representation and
inference mechanism used in the system.  PAGODA forms theories about the
effects of its actions and the world state on the environment over time.
These theories are represented as conditional probability distributions.  A
restriction is imposed on the structure of the theories that allows the
inference mechanism to find a unique predicted distribution for any action and
world state description.  These restricted theories are called "uniquely
predictive theories."  The inference mechanism, Probability Combination using
Independence (PCI), uses minimal independence assumptions to combine the
probabilities in a theory to make probabilistic predictions.

                              _____________

                  SEMINAR ON HUMAN-COMPUTER INTERACTION
                          on Friday, 17 November
                     12:30 p.m., Skilling Auditorium
                         New Voices, New Visions
                             Sally Rosenthal
                            Interval Research
                         [rosenthal@interval.com]

As the personal computer comes of age, it has evolved into an essential tool
for film makers, writers, musicians, animators, photographers, and designers.
As the artistic community becomes more intrigued and integrated with the
available technology, we are witnessing the birth of a new genre: works
created expressly for the computer.

Last year more than 2,000 people from around the world responded to our call
for entries and 550 of them sent their work for review.  The winning pieces
were shown to a sell-out audience at the New York Video Festival and were
exhibited for two weeks at Stanford University.  We are again seeking digital
works from the creative community.  See [http://www.nvnv.org/] for entry
requirements, contact information, and other details.

                               ____________

                          LINGUISTICS COLLOQUIUM
                          on Friday, 17 November
                    3:30 p.m., Building 460, Room 146
           Reciprocals in Bantu: Quantification in Morphology?
                               Sam Mchombo
                         UC Berkeley Linguistics
                      [mchombo@garnet.berkeley.edu]

                               ____________