; Concepts _ Categorization
Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Concepts _ Categorization

VIEWS: 15 PAGES: 35

  • pg 1
									  Concepts &
 Categorization
Internal Structure of Concepts
                           Outline
   What is a concept?
   Similarity-based Views
       The Classical View
       Probabilistic Views
            The Exemplar View
            The Prototype View
            What is a Concept?

   Concept: mental representation of a class
    (cars, cats, oaks, chairs) and associated
    knowledge and beliefs.

   List all of the kinds of BIRDS you can think of.
                        Living
                        Thing                    Breathes

          Plant                         Animal


                  Feathers
                                 Bird            Mammal


Ostrich
                       Barks
                                   Dog            Bat
            Canary
Yellow
                       Living
                       Thing               Breathes

            Plant                 Animal


Feathers

                       Bird                Mammal


  Ostrich
                       Barks
                                Dog         Bat
              Canary
  Yellow
What do Concepts do for us?

   Cognitive Economy

   Inferences

   Combine to form Complex Thoughts

   Communication
            The Classical View
   Features are singly necessary and jointly
    sufficient to define the concept.
       Singly necessary: every member must have that
        feature.
       Jointly sufficient: everything that has all features
        must be a member.


   Categorization => comparing properties of
    instance to definition
        Classical View: Problems
   Definitions are difficult to specify
   Typicality Effects
            Typicality Effects
   Some things are better examples of a
    concept than others.
   Categorization
   Retrieval
   Acquisition
   Reasoning
              Probabilistic Views
   How can the Classical View explain typicality
    effects?

   Probabilistic Views: Category membership is
    graded.
       Prototype View
       Exemplar View
          The Prototype View
   Prototype: abstracted representation of a
    category containing salient features that are
    true of most instances.
   Bird
   Categorization => comparison to prototype
          Prototype v. Classical
   Features aren’t necessary or sufficient.

   Features can be weighted.

   Category membership can be graded.

   Categories can have internal structure.
              The Exemplar View
   Similar to Prototype View
       Representation is not a definition
   Different: Representation is not abstract
       Descriptions of specific examples
   Bird
   Categorization => comparison to stored
    exemplars
            Typicality Effects
   How do Probabilistic Views explain
    typicality effects?
   Categorization
   Retrieval
   Acquisition
   Reasoning
            Probabilistic Views
   Thought question: How did you come up with
    the features weights for the prototype of bird?
      Distribution Exam 1
16
14
12
10
 8
 6
 4
 2
 0
     22 24 26 28 30 32 34 36 38 40 42 44 46 48 50
                        Score
        Within-Category Structure
   Similarity-Based Views
       Classical
       Probabilistic
            Prototype
            Exemplar
   Similarity-Based Views fail to explain...
       Relations among features
       Feature centrality
       Which features are part of representation
 Similarity Depends on Context

White Cloud
Gray Cloud
Black Cloud
 Similarity Depends on Context

White Cloud
Gray Cloud
Black Cloud
 Similarity Depends on Context

White Cloud     White Hair
Gray Cloud      Gray Hair
Black Cloud     Black Hair
 Similarity Depends on Context

White Cloud     White Hair
Gray Cloud      Gray Hair
Black Cloud     Black Hair
             Is Similarity Enough?
   Is similarity sufficient
    for category                                               70




                                    Percent Choosing Quarter
    membership?                                                60

   Imagine an object 3                                        50
    inches in diameter.                                        40
    (Rips, 1989).                                              30
       Is it more similar to a                                20
        pizza or a quarter?
                                                               10
       Is it more likely to be a
        pizza or a quarter?                                     0
                                                                    Sim   Cat
             Is Similarity Enough?
   Is similarity necessary                                                  10
    for category




                                    Insect----------------------------Bird
                                                                             9
    membership?                                                              8
   Sorp/Doon Story. (Rips,                                                  7
                                                                             6
    1989).                                                                                    Accident
                                                                             5                Essence
       Accident v Essence                                                   4                Control
       Is it more similar to a                                              3
        bird or an insect?                                                   2
       Is it more likely to be a                                            1
        bird or an insect?                                                   0
                                                                                  Sim   Cat
          Theory-based Approach
   Similarity is not enough.
   Knowledge of the world informs and shapes our
    concepts… folk theories
   Features embedded in complex network of causal and
    explanatory links. Folk theories point out
       Relative importance of features
       Relations among features
   Objects classified into concept that best explains their
    pattern of attributes.
        Flies                          Small

                  Cares for Young

Has Wings                                Two Legs



   Has Feathers                         Eats Bugs



 Lays Eggs                                 Tasty
                     Lives in a Nest
         Sings
                                        Colorful
          Theory-based Approach
   Similarity is not enough.
   Knowledge of the world informs and shapes our
    concepts… folk theories
   Features embedded in complex network of causal and
    explanatory links. Folk theories point out
       Relative importance of features
       Relations among features
   Objects classified into concept that best explains their
    pattern of attributes.
        Psychological Essentialism
   Previous accounts emphasize perceptually salient
    surface features.
   Non-obvious underlying features related to folk
    theories are often crucial for understanding concepts.
       A Dolphin is more similar to a shark than a deer.
       Dolphins and deer are both mammals because of internal
        properties.
   People’s representations of concepts include an
    essence.
        Flies                          Small

                  Cares for Young

Has Wings                                Two Legs



   Has Feathers                         Eats Bugs



 Lays Eggs                                 Tasty
                     Lives in a Nest
         Sings
                                        Colorful
        Flies                          Small

                  Cares for Young

Has Wings                                Two Legs



   Has Feathers                         Eats Bugs



 Lays Eggs                                 Tasty
                     Lives in a Nest
         Sings
                                        Colorful
           What is an Essence?
   Theory-related
   Causes observable properties
   Ultimately determines category membership
          Definition?
   Could be very vaguely understood or even
    taken on faith.
        Psychological Essentialism
   People act as though things have essences
    (underlying natures) that make them the thing
    they are.
   Essences constrains/generates properties that
    vary in their centrality.
   Essence is tied to Informal Theories
       Essentialist Heuristic: Things that look alike tend
        to share deeper properties.
     Psych. Ess’lism: Implications
   Transformations
       Change in surface features don’t change category
        membership
            Removing pigeon’s wings or feathers
       Change in essential features DO change category
        membership
            Changing a pigeon’s DNA structure
   Categorization ultimately based on essence
       Expertise
       Category membership all-or-none
           Is Category Membersip
              REALLY graded?
   Coley & Luhmann
   Three kinds of concepts
       Natural Kinds (bird, fish, tree, chemical element)
       Social Kinds (occupation, relative)
       Artifacts (furniture, vehicle, clothing)
   Ratings
       Typicality
       Absolute Category Membership (yes, no, sorta)
       Expert Judgments
Results: Does Typicality Predict
   Domain/        Graded Cat. Expert
   Index          Membership Judgments
   Natural
   Kinds             YES         NO

   Social Kinds
                     YES        YES

   Artifacts
                     YES        YES
                         Summary
   Similarity-Based Views
       Classical
       Probabilistic
            Prototype
            Exemplar
   Similarity alone is not enough
   Theory-Based Views
       Psychological Essentialism

								
To top