7. Concepts and Categorization Flashcards
What is a concept
Mental representation of some object, event, or pattern + stores knowledge relevant to it
allows us to classify stuff
What is a category
Class of similar things that share:
1. an essential core
or
2. Similarity in perceptual, biological, or functional properties
What are 5 theoretical views of the nature of concepts
classical view
prototype view
exemplar view
schemata view
knowledge based view
What are 4 functions of categorization
- Understand individual cases we have not seen before and make inferences on them
- Reduces complexity of the environment
- Requires less learning and memorization
- Guides the appropriate action (etc. dog vs. wolf)
What are concepts in the Classical View?
- membership is determined by a set of defining characteristics
- defining properties are NECESSARY and SUFFICIENT
ie. a set of features is enough to classify anything as an instance of a concept
What are the assumptions of the classical view of concepts
- concepts are not representations of specific examples, but as lists of features
- membership is clear cut
- there is no better or worse examples of a category
What are some problems with the classical view of concepts
- no defining features for categories (etc. games)
- defining features do not always fit
- no typicality (graded membership)
What are concepts in the Prototype view?
There is an idealized representation of a class of objects
- includes features that are typical (averaged over past experiences of category members)
What is the family resemblance structure of concepts
Each member has a number of features that it shares with others
The more commonly shared features, the more typical a member is
police + doctor experiment
What are some advantages of prototype view of concepts
- explains why people have a hard time defining concepts (strict definitions do not exist)
- explains how members in a category may seem more typical than others
What are some problems with the prototype view of concepts
- There are no clear cut boundaries for category membership
- Typicality of an instance can depend on context (not fixed)
What are 2 views of categorization that have to do with the abstraction of information
Classical view and Prototype view
What are concepts in the Exemplar view?
- concepts include actual representations of real instances we’ve experienced in the past
- categorization based on comparison to previously stored instances (exemplars)
- assumes no defining characteristics with specific categories
Why might concepts be hard to categorize in the exemplar view vs. prototype view
exemplar view:
to-be categorized instance is similar to exemplars from many different categories
prototype view:
to-be categorized instance shares many features with multiple categories
Explain the typicality effect in the exemplar view vs prototype view
Exemplar view:
Typical instances are more likely to be stored than less typical instances
Prototype view:
More overlap in features with the prototype allows someone to respond faster to it
What are some problems with the exemplar views
unconstrained
- hard to define where category boundaries are and why some objects/instances are stored as exemplars and others aren’t
- requires storage of a lot of exemplars
What are concepts in the Schemata view?
- Concepts are a form of schemata
- Schemata: frameworks of knowledge that have roles, slots, variables
What are problems with the schemata view
- ill defined and cannot be empirically tested
What are concepts in the Knowledge based view?
- Category only becomes meaningful once you know the purpose of the category
- Individuality of categorization (depends on individual’s purpose for categorization)
What limitation of prototype and exemplar view does the knowledge based view address
- did not explain how categories are formed
- in a knowledge based view, the purpose of the category is to define what the category membership is
What does Rosch’s experiment, where participants were asked to respond to various items from same + different categories, tell us about categorization
Categorization is graded; may be based on prototypes or examples rather than a set of defining features
Experiment: asked participants to list common characteristics of instances of different categories (etc. chair, table from furniture) + compared the overlap
What are Concept Attainment strategies as found by Bruner and colleagues
Simultaneous scanning
Successive scanning
What is Simultaneous scanning
- testing multiple hypotheses at the same time
- heavy demands on working memory
What is Successive scanning
- testing one hypothesis at a time
- slow but less demand on working memory