CONCEPTS

1
CONCEPTS READING GROUP
Edited by Eric Margolis and Stephen Laurence
2

• Roadmap for today
• General information about Reading group
• Ideals for Reading group
• Introductions
• Chapter 1 Presentation
• Discussion
• Sign-up/email confirmation for group

3
FAQs
When and where does this meet? Tuesdays
5(05)PM-7PM NE20-461 (if we can fit, we can
move to fishbowl) Who do I contact if I want to
email my talk, have general questions, or wanna
remind myself what the schedule is? Liz Baraff
liz_b_at_mit.edu (cell) 617-893-0125
web.mit.edu/liz_b/www/concepts How come I
havent gotten any emails about meetings,
etc? Emails are sent out via a mailing list
called concepts03_at_mit.edu . MIT users should be
able to add themselves to the list. Others that
wish to be added to the list should sign up on
the sign-up sheet at the end of todays
meeting. Whats the deal with dinner? Email or
call Liz by 400PM on the day of reading group.
Let her know how many slices you want to eat and
if you have any topping preferences/dislikes.
Orders will most likely be from Bertuccis, so
expect to pay about 1.50 per slice. Please bring
your own drink/dessert. (There is a water
fountain on the floor if youre happy with H2O.)
4
Tentative Syllabus   September 9th General
Info, first class meeting Chapter 1 Concepts and
cognitive Science (Laurence and Margolis)
Presented By Liz   September 16th Classical
Theory Chapters 2, 3, 4 Presented By TBA (Please
volunteer!)   September 23 Philosophical
Skepticism about the classical theory Chapters 5,
6   September 30th The probabilistic turn
Stereotypes, Prototypes, Exemplars Chapters 7, 8,
9   October 7th Critical Reactions to the
Probabilistic turn Chapters 10, 11, 12   October
14th Neoclassical theories Intro to Current
Theories and Research, Chapters 13, 14   October
21st Neoclassical theories cont. Chapter 15,
16   November 28th Rethinking Prototypes Chapters
17, 18   November 3rd Theory-theory Chapters
19, 20   November 10th NO MEETING
SCHEDULED   November 17th Conceptual
Atomism Chapters 21, 22, 23, 24   November 24th
Concept Possession in Infants and
children Chapters 25, 26   December 2nd TBA
(General discussion of book, make-up session,
debate, additional readings)
5

• Roadmap for today
• General information about Reading group
• Ideals for Reading group
• Introductions
• Chapter 1 Presentation
• Discussion
• Sign-up/email confirmation for group

6
Objectives To gain comprehensive knowledge of
differing theories about Concepts To encourage
group discussion/reading on Concepts To bring
together individuals/ideas from different
disciplines to encourage interdisciplinary
research Participant expectations Participants
are expected to attend all meetings. (Occasional
conflicts not-withstanding). Participants
should also expect to lead two or three chapter
discussions at some point during the
semester. Participants should expect about 2hours
of outside preparation each week. Participants
should feel comfortable bringing up thoughtful
questions and discussion points. Participants are
strongly encouraged to purchase Concepts. If
this is not possible for some individuals, please
contact liz_b_at_mit.edu to procure readings.
7
DISCLAMER Josh Tenenbaum (MIT) and Susan Carey
(Harvard) will be co-teaching an interesting and
thought provoking course this Fall on concepts as
well. Individuals interested in this topic are
strongly encouraged to take this course instead
of/or in addition to this reading group. The
class and reading group are unrelated and your
decision to participate in one or the other or
both is completely voluntary. (I just strongly
recommend Joshs course because I think it will
be really interested, and plan to take it in
addition to leading this reading group.)
8

• Roadmap for today
• General information about Reading group
• Ideals for Reading group
• Introductions
• Chapter 1 Presentation
• Discussion
• Sign-up/email confirmation for group

9
(No Transcript)
10

• Roadmap for today
• General information about Reading group
• Ideals for Reading group
• Introductions
• Chapter 1 Presentation
• Discussion
• Sign-up/email confirmation for group

11
Introduction Some Preliminaries

• "Should concepts be thought of as bundles of
  features, or do they embody mental theories?
• "Are concepts mental representations, or might
  they be abstract entities?
• "Are concepts objects or some cognitive or
  behavioral abilities?"

12
Primitive, Complex, Lexical Concepts

• subpropositional mental representations
• Primitive lack structure, (atomic/features)
• Complex Arent primitive
• Lexical Concepts that correspond to lexical
  items, like Bachelor, Bird, Bite.

13
Two Models of conceptual structure

• Containment Model One concept is a structured
  complex of other concepts
• C composed of X, Y, Z C necessarily involves
  X,Y,Z
• C - Dropped the Accordion cant be tokened w/o
  Accordion tokened.
• 2. Inferential Model Concepts are in privileged
  relations to one another via inferential
  disposition.
• red may implicate color, but you can token
  red w/o color
• Knowing these models will allow thoughtful
  evaluation of theories in later chapters.

14
Concepts as Abstracta vs. Concepts as Mental
Representations
Frege (sense-mode of presentation and reference)
e.g. Mark Twain/Samuel Clemens 22 vs.
161/2 "Senses are the cognitive content of
linguistic expressions (expressions have senses
differences in cognitive content correspond to
difference between senses). "Senses determine
reference "Senses are the indirect referents of
expressions in intentional contexts Problems of
subjective vs. objective (senses cannot be mental
entities) Can have same type of mental
representation, but not same token representation.
15
Classical Theory of Concepts
Concepts and Definitions "Most concepts (esp.
lexical concepts) are structured mental
representations that encode a set of necessary
and sufficient conditions for their application,
if possible in sensory or perceptual terms." (p.
10) Bachelor is not married, is male, is an
adult Most concepts are complex composed of
structurally simpler concepts. (containment
model.)
16
Classical Theory cont.
Concept acquisition simply assemble
features/primitives to form concepts
Categorization just requires checking to see if
concept required features are present. Analytic
Inferences Smith is a bachelor, so Smith is a
man. Reference Determination concepts are
semantically evaluable. (binds concept
acquisition, categorization, analytic inferences,
etc.)
17
Problems with Classical Theory
Plato's problem There are few examples of
defined concepts. Psychological Reality Lexical
concepts show no effects of definitional
structure in psychological experiments. Analytici
ty Philosophical arguments against analyticity
also work against the claim that concepts have
definitions. Ignorance and Error It is possible
to have a concept in spite of massive ignorance
and/or error, so concept possession can't be a
matter of knowing a definition. Conceptual
Fuzziness (are carpets furniture?) Typicality
effects some items seem to be better examples of
a category.
18
The Prototype Theory
"Most concepts (esp. lexical concepts) are
structured mental representations that encode the
properties that objects in their extension tend
to possess." (p.31) Tokens satisfy a sufficient
number of properties. Bird Robin vs. Ostrich
19
The Prototype Theory cont.
Lacks definitions (avoids specification problem
that classical faced). Rejection of necessary
condition Concepts acquired via statistical
procedure of feature identification
(trends). Categorization via similarity of
target to prototype. explains (graded judgments
of exemplariness, typicality correlates with
property list, graded speed of quick
categorization judgments, categorization errors
inversly correlated with typicality, fuzziness.)
20
Problems with the Prototype Theory
Problems with Prototype Primes Typicality
effects don't argue for prototype structure,
since even well-defined concepts exhibit
typicality effects. (8 vs. 34) Problems of
Ignorance and Error Concepts with prototype
structure fail to cover highly atypical instances
and incorrectly include non-instances. (toy
tigers vs. toothless albino tigers) "Missing
Prototype Problem Many concepts lack prototypes
(not a wolf.) "Problem of Compositionality
The prototypes of complex concepts aren't
generally a function of the prototypes of their
constituent concepts. (Smith et al --gt later in
book)
21
Theory-Theory of Concepts
"Concepts are representations whose structure
consists in their relations to other concepts as
specified by a mental theory." Theories,
Explanations, Conceptual Structure Concepts are
mental theories (domain knowledge) vs. Concepts
are like theoretical terms (constituents of
thought) Explain constituents in terms of their
domains (concepts identity is determined by its
role within a theory.)
22
Problems for Theory-Theory
Ignorance and Error It is possible to have a
concept in spite of its being tied up with a
deficient or erroneous mental theory. Problems
of Stability The content of a concept can't
remain invariant across changes in its mental
theory. "Mysteries of Science Problem" The
mechanisms that are responsible for the emergence
of new scientific theories and for the shift from
one theory to another are poorly understood."
23
The Neoclassical Theory of Concepts
Updating the classical theory "Most concepts
(esp. lexical concepts) are structured mental
representations that encode partial definitions,
i.e., necessary conditions for their
application. Used for explanatory reasons Red
must be a color
24
Problems for Neoclassical Theory
Problem of Completers If partial definitions
are turned into full definitions, then the
Neoclassical Theory has all the problems that are
associated with the Classical Theory. If,
instead, they are left incomplete, then the
Neoclassical Theory has no account of reference
determination. Problem of Ignorance of Error
Supplementing neoclassical structure with 3-D
models won't help in accounting for reference
determination. The Regress Problem for Semantic
Fields Neoclassical structure can't explain how
a word retains aspects of its meaning across
different semantic fields. Either its conceptual
constituents mush themselves have neoclassical
structure, and so on, or else no structure is
needed at all. Harry kept the bird in the cage.
Harry kept the money. Harry kept the crowd happy.
25
Conceptual Atomism
Concepts without structure Lexical concepts are
primitive they have no structure. Asymmetric
Dependence Theory (Foder) The content of a
primitive concepts is determined by the concepts
standing in an appropriate causal relation to
things in the world. Mental content requires
that a concept carry information about the
property it expresses. (information- reliable
correlation) Disjunction Problem (later in book)
26
Problems for Conceptual Atomism
Radical Nativism Under conceptual atomism, most
lexical concepts turn out to be innate, including
such unlikely candidates as xylophone and
carburetor. Explanatory Impotence If lexical
concepts are primitive, they can't explain
psychological phenomena such as
categorization. Analytic Data Conceptual
Atomism lacks an adequate explanation of why
people have intuitions of analyticity. Compositio
nality Much difficulty with conceptual
combination for non lexical concepts. Empty and
Coextensive Concepts If concepts are atoms and
the content of a concept is just its reference,
then coextensive concepts can't be distinguished.
As a result, all empty concepts have the same
content.
27
Concluding remarks
Different theories are differently good at
handling Fast categorization, considered acts of
categorization, semantic application, the
licensing of inductive inference, analytic
inference, concept acquisition, compositionally,
stability.
28
Four Types of Conceptual Structure
Compositional Reference Determining Structure
structure that contributes to the content and
reference of a concept via a compositional
semantics. Nonsemantic Structure Structure that
doesn't contribute to the content of a concept,
but does contribute significantly to some other
theoretically important explanatory function of
concepts. Nonreferential Semantic Structure
Structure that contributes to the content of a
concept but is isolated from referential
consequences. Sustaining Mechanism Structure
Structure that contributes to the content of a
concept indirectly by figuring in a theoretically
significant sustaining mechanism, i.e., a
mechanism that supports a relation such as
asymmetric dependence.
29

• Roadmap for today
• General information about Reading group
• Ideals for Reading group
• Introductions
• Chapter 1 Presentation
• Discussion
• Sign-up/email confirmation for group

30
Other thoughts, questions, points?
31

• Roadmap for today
• General information about Reading group
• Ideals for Reading group
• Introductions
• Chapter 1 Presentation
• Discussion
• Questions Sign-up/email confirmation for group