[Prev][Next][Index]

CSLI Calendar, 10 November 1994, vol.10:7




        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
______________________________________________________________________________

10 November 1994                 Stanford                       Vol. 10, 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 10 -- 18 NOVEMBER 1994

  THURSDAY, 10 NOVEMBER
        10:00 - STASS Seminar
                Cordura Hall, Room 100
                Hyperproof: An Approach to Situated Inference
                David Barker-Plummer and Mark Greaves, CSLI
                Abstract below

         2:15 - CSLI Seminar
                Cordura Hall, Room 100
                What is the Digital Library?
                Terry Winograd, Stanford Computer Science
                Abstract below

         4:15 - Symbolic Systems Forum
                Building 60, Room 61-F
                Robots that Plan and Robots that React
                Illah Nourbakhsh, Stanford Computer Science
                Abstract below

         7:30 - Phonology Workshop
                Building 460, Seminar Room 146
                Vowel Length: *LONG-VOWEL in Eclipse
                Tim D. Sherer, UC Berkeley
                Abstract below

  FRIDAY, 11 NOVEMBER
        12:00 - Logic Lunch
                Building 380, Room 383-N
                Solving Set Constraints with Projections
                Leszek Pacholski, Wroclaw Institute of Computer Science
                Abstract below

        12:30 - HCI Seminar
                McCullough Building, Room 134
                How to Support Design with Research When Potential Users
                Can't Imagine Future Products
                Bonnie Johnson, Interval Research
                Abstract below

         3:15 - Philosophy Colloquium
                Building 90, Room 91-A
                Reasoning in Hume
                David Owen, University of Arizona Philosophy
                Abstract below

         3:30 - Linguistics Colloquium
                Cordura Hall, Room 100
                Underspecification by Optimization
                Sharon Inkelas, UC Berkeley Linguistics
                Abstract below
  
  MONDAY, 14 NOVEMBER
         8:30 - IAP Meeting
                Cordura Hall, Room 100
                Schedule below

  TUESDAY, 15 NOVEMBER
         8:30 - IAP Meeting
                Cordura Hall, Room 100
                Schedule below

         7:00 - Special Linguistics Talk
                Building 460, Room 146
                Core and Periphery in Loanword Adaptation: English
                Borrowings in Quebec French
                Carole Paradis, Laval University
                Abstract below

  WEDNESDAY, 16 NOVEMBER
         4:00 - Seminar on Computational Learning and Adaptation
                Cordura Hall, Room 100
                Self-Customizing Applications: Completing Patterns 
                and Constructing User Interfaces
                Jeffrey C. Schlimmer, Washington State University
                Abstract below

  THURSDAY, 17 NOVEMBER
        10:00 - STASS Seminar
                Cordura Hall, Room 100
                An Introduction to Situation Theory
                Keith Devlin, Saint Mary's College Mathematics
                Abstract below

         2:15 - CSLI Seminar
                Cordura Hall, Room 100
                Graphical Theorem Proving: An Approach to Reasoning
                with the Help of Diagrams
                David Barker-Plummer, CSLI
                Abstract below

         4:15 - SSP Forum
                Building 60, Room 61-F
                Genetic Programming: Programming Computers By Means
                of Natural Selection
                John Koza, Stanford Computer Science
                Abstract below

  FRIDAY, 18 NOVEMBER
        12:30 - HCI Seminar
                McCullough Building, Room 134
                Designing for Mathematical Understanding
                Rogers Hall, UC Berkeley Education and IRL
                Abstract below

         3:15 - Philosophy Colloquium
                Building 90, Room 91-A
                Tyler Burge, UCLA Philosophy

         3:30 - Linguistics Colloquium
                Building 460, Room 146
                On Maori `he' and the Uses of Indefinites
                Sandy Chung, UC Santa Cruz Linguistics          
                Abstract below

                               ____________

The CSLI Calendar appears on Thursday of each week throughout the academic
year.  Announcements, abstracts, and other information to appear in the
Calendar on a given Thursday should be submitted by e-mail to
incalendar@csli.stanford.edu by 5:00 p.m. on the previous Tuesday.

Past issues of the CSLI Calendar, a quarterly schedule of upcoming CSLI
events, and other information about CSLI are available on the World Wide Web:
<http://csli-www.stanford.edu/>.  The Calendar, with available abstracts, is
also posted each week to the csli.bboard newsgroup.

                               ____________

                              STASS SEMINAR
                         on Thursday, 10 November
                    10:00 a.m., Cordura Hall, Room 100
              Hyperproof: An Approach to Situated Inference
                  David Barker-Plummer and Mark Greaves
                       CSLI and Stanford Philosophy
                    <{dbp,greaves}@proof.stanford.edu>

This talk will focus on how situation theory can be used to account for
certain sorts of reasoning which are difficult to handle in traditional logic.
Many sorts of actual reasoning tasks involve extracting, manipulating, and
combining information in forms which are not, prima facie, sentences of a
language.  It turns out that this sort of reasoning, which we call
heterogeneous reasoning, can be modeled very nicely using the tools of
situation theory.  In our talk, we will present Hyperproof, a computer program
based on these ideas which addresses reasoning in which some of the
information is presented in a graphical form.  In addition to presenting the
program, though, we will informally describe its theoretical underpinnings:
how we model the information content of a diagram, how this relates to the
information given in a set of sentences, and how we combine these into a
precise and unified proof system.  We will also touch on issues of how the
choice of representation affects the computational complexity of reasoning,
and how situation theory accounts for this.

                               ____________

                               CSLI SEMINAR
                         on Thursday, 10 November
                    2:15 p.m., Cordura Hall, Room 100
                       What is the Digital Library?
                              Terry Winograd
                        Stanford Computer Science
                        <winograd@cs.stanford.edu>

The quest for the "digital library" has become a focus for work by a number of
people and organizations, including both commercial development and
university-based research.  Stanford is one of six universities recently
receiving large grants on this topic from a government funding consortium led
by the National Science Foundation.  The phrase "digital library" suggests a
thought-provoking encounter between our cultural images of "digital" (high
tech, change-making, efficient, automated), and "library" (old leatherbound
volumes, tradition-preserving, hands-on).  I will raise some questions about
what a digital library might be, what relationship it has to our current
institutions relating to both halves of this mix, and what research questions
we and other groups are posing in order to develop a new synthesis.

                               ____________

                          SYMBOLIC SYSTEMS FORUM
                         on Thursday, 10 November
                    4:15 p.m., Building 60, Room 61-F
                  Robots that Plan and Robots that React
                             Illah Nourbakhsh
                        Stanford Computer Science
                         <illah@cs.stanford.edu>

Robotics has been an area punctuated by extreme approaches for several years.
Many roboticists, most notably nearer the Atlantic Ocean, have constructed
systems that react to changes in their sensors' input from the environment.
Other roboticists have created planning systems, demonstrating flawless
simulations of robotic navigators.  We will take a short guided tour of these
two extremes and spend some time examining the middle ground, where
roboticists are very gradually converging.  The issues that, in the end, must
be combined to create robotic systems include reactivity, interleaving
planning and execution, and assumptive systems that defer planning for the
possibility of worst-case scenarios.

                               ____________

                   BERKELEY/STANFORD PHONOLOGY WORKSHOP
                         on Thursday, 10 November
               7:30, p.m., Building 460, Seminar Room 146
                   Vowel Length: *LONG-VOWEL in Eclipse
                              Tim D. Sherer
                         UC Berkeley Linguistics

Languages have two ways of dealing with vowel length.  Vowel length may be
redundant, dictated by prosody, such as foot and syllable structure, or it may
be distinctive, marked in the lexical representations of the words and mapped
to the surface as faithfully as possible.  In OT terms, the constraint against
long vowels is dominated by the relevant FAITH constraint.  Even within
languages of the latter type, the constraint that disfavors long vowels may
have effects.  In Dravidian languages, short vowels are preferred only when
preceding a geminate.  Certain languages, such as Fula and Woleaian, have
affixes with unspecified prosodic material and these affixes are filled with
consonantal material, in conflict with Alignment considerations.  In OT terms,
these are cases where *LONG-VOWEL, although dominated, is active on the
constraint hierarchy.

                               ____________

                               LOGIC LUNCH
                          on Friday, 11 November
                  12:00 noon, Building 380, Room 383-N 
                 Solving Set Constraints with Projections
                             Leszek Pacholski
                      Institute of Computer Science
                             Wroclaw, Poland

Systems of set constraints describe relations between sets of ground terms;
they have the form of inclusions between set expressions built over a set of
set-valued variables, constants and function symbols.  Set constraints have
been successfully used in program analysis and type inference algorithms for
functional, imperative and logic programming languages by a number of
researchers.  Solving a system of set constraints is the main part of these
algorithms.  However, until recently the satisfiability problem for such
constraints was open, except in some special cases.  Complexity of these
problems has also been extensively studied.  In my talk, I shall sketch the
history of set constraints, and provide the main ideas behind the result
establishing the NEXPTIME completeness of the full problem (with projections)
which I have recently obtained together with Witold Charatonik.

                               ____________

                  SEMINAR ON HUMAN-COMPUTER INTERACTION
                          on Friday, 11 November
                12:30 p.m., McCullough Building, Room 134
         How to Support Design with Research When Potential Users
                      Can't Imagine Future Products
                              Bonnie Johnson
                            Interval Research
                          <johnson@interval.com>

At one time, new product development was something done in a lab and presented
to stereotypical consumers for comment and refinement. "New and Improved" was
sufficient to generate interest. Now many new products are beyond the
understanding of their potential users. And many users are likewise beyond the
understanding of the marketers and developers who need to reach them. At a
time when thorough, accurate and perceptive market insights are most needed,
they have become increasingly difficult to get. This presentation will discuss
methods of "Applied Exploration" from observing "life contexts" of potential
users to synthesizing conclusions in the form of guidelines.

BONNIE JOHNSON is a member of the research staff at Interval Research, an
institute founded by Paul Allen to conduct basic research on future
applications of technology.  She heads a group exploring methods and
technologies for market research.

Prior to joining Interval, Bonnie spent over 10 years researching and managing
the introduction of office technology.  She held management positions at
Intel, Aetna Life & Casualty, Focus Systems and Humanware.  While at Aetna,
her group was the subject of a Harvard Business Case.  In a project funded by
the National Science Foundation, Bonnie led a two year investigation of 200
organizations to identify the management practices that distinguished
innovative users of office technology.

Bonnie has a Doctorate in Communication from State University of New York and
Post Doctoral study at Stanford University.  She is the author of several
books and articles on business organization and innovation including Managing
Organizational Innovation, The Evolution from Word Processing to Office
Information Systems (Columbia University Press, 1987; co-author Ronald E.
Rice)

                               ____________

                          PHILOSOPHY COLLOQUIUM
                          on Friday, 11 November
                    3:15 p.m., Building 90, Room 91-A
                            Reasoning in Hume
                                David Owen
                     University of Arizona Philosophy

Most discussion of Hume's views on reasoning, especially demonstrative
reasoning, have assumed that Hume is working with a concept of deduction,
formally construed either in the traditional syllogistic way or in the modern
sense.  Hume's own work, and the background of Descartes and Locke, makes it
clear that he rejected any formal conception of deduction.  This paper
attempts to reconstruct Hume's own views on demonstration, and puts forward
new solutions to several problems traditionally found in Hume: the relation
between conceivability of the contrary and demonstrability, the limitation of
demonstration to algebra and arithmetic, empirical measurement, what it is to
imply a contradiction, and the nature of demonstrative and probable reasoning.

                               ____________
                        
                          LINGUISTICS COLLOQUIUM
                          on Friday, 11 November
                    3:30 p.m., Cordura Hall, Room 100
                    Underspecification by Optimization
                              Sharon Inkelas
                         UC Berkeley Linguistics
                      <inkelas@cogsci.berkeley.edu>

Past theories of underspecification (e.g., Ringen 1975, Kiparsky 1982,
Archangeli 1984, Steriade 1987) have based underspecification on such
principles as markedness, redundancy in inventories, or predictability.
Critics of underspecification (e.g., Mohanan 1991, McCarthy and Taub 1992,
Steriade 1994) have, however, pointed out specific problem cases for each such
principle, concluding that underspecification is not a viable tool in the
analysis of various phenomena previously claimed to require it.

I argue in this paper that while it is true that past principles are
inadequate to fully characterize the distribution of underspecification, it is
not true that underspecification is dispensable or uncontrollable.  I propose
a new theory, called Archiphonemic Underspecification, in which
underspecification is regulated by simple, deterministic principles of
optimization.  Lexicon Optimization, adapted from work in Optimality Theory
(Prince and Smolensky 1993), selects underlying form so as to optimize the
phonological mapping from input to output.  Grammar Optimization, adopted from
Kiparsky 1993, ensures that the input-output mapping is maximally
structure-preserving (in the sense of not deleting structure).  Together,
these principles predict underspecification in exactly the case of predictable
alternations, and nowhere else.  Specifically, predictable but nonalternating
structure is fully specified underlyingly.

>From the perspective of this new account, the failure of past principles of
underspecification is understandable: they all attempt to regulate
underspecification without regard for the alternations in any given language.
Archiphonemic Underspecification underspecifies only in case of alternations,
enabling the description of ternary contrasts and numerous other phenomena
which past proposals could not account for.

                               ____________
                        
                             CSLI IAP MEETING
                  on Monday and Tuesday, 14--15 November
                    8:30 a.m., Cordura Hall, Room 100

The following is a schedule of talks to be given during this year's IAP
meeting.  These talks survey some of the research being carried out at CSLI
and are open to the public.

   MONDAY, 14 November
         9:15 - Keynote address: CSLI in the Age of Digital Convergence
                John Perry

        10:30 - Building a Usable English Grammar for Industry and Academia
                Dan Flickinger 

        11:30 - How People Respond to Communication Technology
                Cliff Nass and Byron Reeves

         2:15 - Text Understanding and Information Extraction
                Jerry Hobbs

         3:30 - Visual Reasoning and Hyperproof II
                John Etchemendy

   TUESDAY, 15 November
         9:15 - THE Disfluency Talk
                Herb Clark and Tom Wasow

        10:30 - Spatial Thinking
                Barbara Tversky

        11:30 - Applying NeuralNet Technology
                David Rumelhart

         1:30 - Building Intelligent Agents
                Barbara Hayes-Roth

Times have also been set aside on these two days for informal discussions with
these and other CSLI researchers.

                               ____________

                         SPECIAL LINGUISTICS TALK
                         on Tuesday, 15 November
               7:00, p.m., Building 460, Seminar Room 146
            Core and Periphery in Loanword Adaptation: English
                       Borrowings in Quebec French
                              Carole Paradis
                             Laval University

As shown in Paradis, Lebel, and LaCharite 1993 and Paradis, LaCharite, and
Lebel 1994, within the framework of TCRS (Theory of Constraints and Repair
Strategies), phonological malformations in loanwords are repaired
systematically.  Adaptations are minimal in order to preserve segmental
information maximally.  Segment deletion (loss of information) occurs only
when the distance between the input and the output is too large, i.e. when too
many steps are required to rescue an illformed segment (a generalization,
incidentally, that cannot be captured by a surface-only constraint framework).

However, although segmental adaptation and deletion in loanwords are
systematic and thus predictable from the point of view of TCRS, there are
facts which are puzzling for TCRS, as for any constraint-based theory.  Some
segments which are clearly non-native are nevertheless tolerated by speakers
of the borrowing language in numerous words.  This is the case of the English
liquids, diphthongs and affricates in Quebec French (QF).  These "tolerated"
segments contrast with the English interdentals and the laryngeal /h/, which
are totally prohibited and thus systematically adapted in QF.

Given this, one might be tempted to conclude that the so-called "tolerated"
segments are in fact now part of the QF inventory.  However, if these segments
were really allowed, one would thus expect them (i) not to undergo
modification in any borrowing (unless numerous exceptions or an optional
arbitrary rule is posited), (ii) to be found in native words, and (iii) to be
used productively, i.e. to form new words.  This is not the case.  The
so-called "tolerated" segments are clearly marked in QF.  A question arises
therefore: If they do not result from positive parameter-settings, how come
they are so often present in borrowed words, in contrast with the English
interdentals and the laryngeal, which are zero-tolerated?  I suggest that the
so-called "tolerated" segments are in fact part of the QF Periphery, as
opposed to the interdentals and the laryngeal which are not allowed by the
periphery or by the core.  Following Chomsky 1986, I view the periphery as the
domain where some constraints of the core (though obviously not all of them)
are eliminated or weakened. I will show that not only is the notion of
"periphery" able to contrast between "prohibited" and "tolerated" segments
but, if we maintain that the periphery is not homogeneous, as claimed by Ito &
Mester 1993, it also allows us to contrast AMONG "tolerated" segments, and
make important implicational predictions of adaptation.  This analysis is
based on a corpus of 1183 English borrowings in QF, containing a total of 3667
malformations.

                               ____________

             SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
                         on Wednesday, 16 November
                    4:00 p.m., Cordura Hall, Room 100
           Self-Customizing Applications: Completing Patterns 
                     and Constructing User Interfaces
                           Jeffrey C. Schlimmer
       Washington State Electrical Engineering and Computer Science
                         <schlimme@eecs.wsu.edu>

People like to record information.  Doing this on paper is initially efficient,
but lacks subsequent flexibility in information retrieval and organization.
Recording information on a computer is also efficient -- although more
powerful -- but it requires significant effort to pre-configure the applicable
software.  For instance, fields must be defined before data can be added to a
database.  Applications that customize themselves free the user from having to
explicitly command the computer.  The user can record information directly and
in a free-form manner.  Behind the interface, an autonomous process is acting
on behalf of the user, helping to capture and organize the information.

This talk will present a self-customizing application that learns the syntax
and semantics of the user's information.  To help the user capture
information, it provides default values by consulting learned semantics.  To
organize the user's information, it constructs a specialized user interface
from learned syntax.  The target of this research is to develop modules that
can assist users of highly portable or hand-held personal computing devices.
A video tape demonstration of an implemented system will be shown.

The goal of this seminar is to increase communication among local researchers
with interests in computational approaches to learning and adaptation.  If you
would like to be added to (or removed from) the mailing list, or if you are
interested in giving a talk in the seminar, please send email to
<langley@cs.stanford.edu>.

                               ____________

                              STASS SEMINAR
                          on Thursday, 17 November
                    10:00 a.m., Cordura Hall, Room 100
                   An Introduction to Situation Theory
                               Keith Devlin
                     Saint Mary's College Mathematics
                        <devlin@csli.stanford.edu>

I will continue my introductory lectures on situation theory.  Anyone wishing
to join the group should read the first few chapters of my book _Logic and
Information_ (1991).  The first half of the seminar will be devoted to a
discussion of work in progress, that tries to develop a framework that extends
the original ideas of situation theory to incorporate insights from speech
acts theory and conversation analysis.  The second half will have more of a
"lecture" format, and will consist of a continuation of my presentation of
"classical" situation theory and situation semantics.  I will provide links
between the two halves of the seminar whenever possible.

                               ____________

                               CSLI SEMINAR
                         on Thursday, 17 November
                    2:15 p.m., Cordura Hall, Room 100
                Graphical Theorem Proving: An Approach to
                   Reasoning with the Help of Diagrams
                           David Barker-Plummer
                                   CSLI
                         <dbp@proof.stanford.edu>

The role of diagrams in mathematics is something of a puzzle.  On the one hand,
mathematicians insist that diagrams have no role in formal mathematics, while
on the other mathematics texts frequently utilize diagrams with abandon.  I
believe that diagrams are often used to contain information about the strategy
that should be used to carry out a proof.  In this talk I will describe a
theorem proving system called GROVER which has been designed to test this
hypothesis.  GROVER is novel in that it may be guided in its search for a
proof by information contained in a diagram.  There are two parts to the
system: the underlying theorem prover, called &, and the graphical subsystem
which examines the diagram and makes calls to the underlying prover on the
basis of the information found there.  I will describe GROVER's proof of the
Diamond Lemma, a non-trivial proof in the theory of well-founded relations.

This is joint work with Sidney C. Bailin of CTA Incorporated.

                               ____________
                                     
                          SYMBOLIC SYSTEMS FORUM
                         on Thursday, 17 November
                    4:15 p.m., Building 60, Room 61-F
                Genetic Programming: Programming Computers
                      By Means of Natural Selection
                                John Koza
                        Stanford Computer Science
                          <koza@cs.stanford.edu>

Genetic algorithms are useful in solving many problems, including optimization
problems in non-linear multi-dimensional spaces.

Genetic programming extends the genetic algorithm to the domain of computer
programs.  In genetic programming, populations of program are genetically bred
to solve problems.  Genetic programming can solve problems of system
identification, classification, control, robotics, optimization, game-playing,
and pattern recognition.  Starting with a primordial ooze of hundreds or
thousands of randomly created programs composed of functions and terminals
appropriate to the problem, the population is progressively evolved over a
series of generations by applying the operations of Darwinian fitness
proportionate reproduction and crossover (sexual recombination).

Many problem environments have regularities, symmetries, and homogeneities
that can be exploited in solving the problem.  Automatic function definition
provides a way to dynamically decompose a problem into simpler subproblems,
solve the subproblems, and assemble the solutions to the subproblems into an
overall solution of the original problem.  Experimental evidence suggests that
automatic function definition reduces the computation effort needed to solve a
problem and produces a simpler and more understandable overall solution.

JOHN R. KOZA is consulting professor in the computer science department at
Stanford University.  He is author of the 1992 book Genetic Programming: On
the Programming of Computers by Means of Natural Selection from the MIT Press
and the 1994 book Genetic Programming II: Automatic Discovery of Reusable
Programs covering the subject of automatic function definition in the context
of genetic programming.

                               ____________

                  SEMINAR ON HUMAN-COMPUTER INTERACTION
                          on Friday, 18 November
                12:30 p.m., McCullough Building, Room 134
                 Designing for Mathematical Understanding
                               Rogers Hall
                      UC Berkeley Education and IRL
                       <rhall@garnet.berkeley.edu>

By some accounts there is a crisis in traditional mathematics instruction.
For example, the half life of the US study body in mathematics after the
eighth grade is currently estimated at about one year.  In response, there are
a variety of efforts underway to reorganize the learning and teaching of
mathematics, along both conceptual and technical lines.

I will talk about ongoing work in an NSF-sponsored curriculum development
project for middle school mathematics.  Our design goal has been to extend the
usual sense of an "applied" mathematics problem into a series of longer-term
(4 to 10 weeks), broadly accessible, and computer-supported "design projects"
for sixth, seventh, and eighth graders.  We are currently field testing a set
of curriculum units that borrow problem contexts from fields as diverse as
architectural design and population ecology.

The talk follows two comparisons.  First, I contrast student work from field
studies conducted early and late in our curriculum development process.
Decisions made over the course of developing each unit reflect the views of
very different design participants: educational researchers, interface
designers, teaching professionals, school children, and cognitive scientists.
What results (we hope) is curriculum materials that support alternative and
more authentic forms of mathematical work in school settings.

Second, on the question of what authentic mathematical work might be, I
contrast the problem-solving activities of middle school students with those
of adult professionals in design-oriented work places.  Since school
mathematics is usually intended to have some relation to adult mathematical
practices, we can examine activity in these settings to determine what form
that relation might take.  Field work in adult sites is drawn from a summer
practicum program for teachers that, in the words of one participant, helps to
"break the edges off" of traditional views of mathematics.

ROGERS HALL is an Assistant Professor in the Division for Education in
Mathematics, Science, and Technology at the University of California at
Berkeley.  He also works as a research scientist at the Institute for Research
on Learning.  His graduate training is in psychology and computer science, and
his research focuses on the development of discipline-specific
representational practices in and out of school.  He is not (yet) a www
entity.

                               ____________

                          PHILOSOPHY COLLOQUIUM
                          on Friday, 18 November
                    3:15 p.m., Building 90, Room 91-A
                               Tyler Burge
                             UCLA Philosophy

                               ____________

                          LINGUISTICS COLLOQUIUM
                          on Friday, 18 November
                    3:30 p.m., Building 460, Room 146
                On Maori `he' and the Uses of Indefinites
                               Sandy Chung
                        UC Santa Cruz Linguistics
                          <chung@ling.ucsc.edu>

At the heart of the Kamp/Heim theory of indefinites is the claim that
indefinites are not inherently quantificational, but instead acquire their
quantificational force from the larger structures in which they happen to
occur.  In this theory, the existential force of indefinites that lie within
the nuclear scope of an operator, or within the scope of no operator at all,
is attributed to one or more processes of default existential closure.

Heim (1982) explores two different routes for achieving this existential
closure: a structural route, in which an existential quantifier is adjoined to
the relevant domains in Logical Form (Chapter 2), and an interpretive route,
in which the existential reading falls out from the satisfaction conditions
that constrain the building of files in file-change semantics (Chapter 3).
Though Heim ultimately opts for the interpretive route, she observes that the
choice is rather weakly motivated.

In this talk I discuss some Maori facts which seem to offer further evidence
for an interpretive over an LF approach to existential closure.  Indefinite
DP's in Maori are headed by one of two determiners, `tetahi' (pl. `etahi') or
`he'.  The distribution of `tetahi' is roughly equivalent to English `a'; the
distribution of `he' is known to be severely restricted.  Drawing on recent
work by Chung, Mason, and Milroy, I show that the occurrence of `he' is
constrained by three conditions, one restricting its morphosyntactic
distribution; another, its role in discourse structure; and the third -- I
claim -- the structural configurations in which it can occur in Logical Form.
The statement of the third condition, I contend, requires us to assume that
default existential closure is not structurally implemented in Logical Form,
but rather occurs subsequently, in the interpretation.

                               ____________