Categories and Concepts
posted 3/07/06
Reading assignment: Chapter 10
Why is it important to group items into categories?
-
Organize the world
-
Apply knowledge to ____________________
-
Recognize repeated encounters with one object as the same thing
-
Distinguish between categories
Category vs. Concept
-
___________ = mental representation
-
___________ = set of examples picked out by the concept.
Write down as many examples of ________ as you can.
1. Classical View
-
Category membership is defined by a list of ____________and ______________
features
-
E.g. triangles are plane figures with 3 corners and 3 sides
-
Concepts have clear-cut boundaries
-
All members are equally ___________________
-
When concepts are organized heirarchically, the concept includes the defining
features of its _______________
-
Since chickens are birds, chickens have the necessary and sufficient features
of birds
Problems with Classical View
-
Failure to specify _______________ (What's the definition of a game? )
-
Unclear cases or fuzzy categories (Is a rug furniture? )
-
Use of ______________ features (Birds fly )
-
Nested concepts (Chicken is a good animal but a bad bird example )
-
____________ effects (Is the Pope a bachelor?)
2. Probabilistic View
-
Concepts are summary representations of characteristic features
-
Features need only be _______________, not necessary, to be included
-
_______________ is a category member (real or hypothetical) that shares
the largest number of features with all category members
-
Category membership is defined by _________________ - similarity to the
prototype
-
No defining features
-
Category boundaries are fuzzy or unclear
-
Category members can range in their typicality
Evidence for typicality
-
Prototypicality ratings
-
______________ times
-
More prototypical instances categorized ___________
-
Prototypical members listed first
-
Prototypical members learned first
-
Prototypical members are used as a reference point (en ellipse is almost
a circle)
-
Family resemblance is maxmized at the ___________ level
How the Probabilisitic view accounts for phenomena that Classical view
can't
-
Categories don't have a definition, just ______________________
-
Items more similar to the prototype will be judged more typical
-
Items that are equally similar to prototypes from two different categories
will be ________________
-
Since there are no necessary and sufficient features, the prototype can
contain unnecessary features.
Problems with Prototype view
-
Is prototype the average or the ________?
-
Experts tend to categorize using the ideal
-
____-derived categories based on an ideal
-
E.g., things to wear in the snow
-
People know more than the central tendency
-
Also know ________ among features
-
E.g., small birds more likely to sing
-
Context effects
-
E.g., harmonica is a typical musical instrument at campfire but not at
a concert hall
-
Prototypes should obey _____________
-
Each item is more similar to its own prototype than other prototypes
-
But, categorization decisions don't always follow typicality
-
E.g., A bat is more similar to the bird prototype than to the mammal prototype,
but we know its a mammal.
-
People are perfectly happy to apply typicality ratings to well-defined
categories
-
E.g., 3 is a better odd number than ___, even though both are definitely
odd
-
Typicality ratings might be flawed task
3. Exemplar View
-
Categories are not stored as defining features or as prototypes.
-
Instead, they are stored as the ____________ themselves
-
For example, the bird category is stored as representations of particular
birds
-
New instances are categorized by ____________________________
How the Exemplar view accounts for phenomena that Classical view can't
-
Categories don't have definitions; they have lists of exemplars
-
Items more similar to many of the exemplars are judged more __________
than other items
-
Items equally similar to the exemplars from two different categories will
be unclear cases
-
Exemplars are stored by all of their features, even _____________ ones
Advantages of Examplar View over Prototype View
-
________ categories: e.g., things to save during a fire
-
Easy to generate exemplars
-
Hard to generate a prototype
-
____________ of features (as well as central tendency) is stored (Rips,
1989)
-
Subjects asked about average size of quarters and pizzas.
-
Then shown a 5" circle size exactly in between the two
-
Subjects judged it to be a pizza because pizza size ________, but quarter
size doesn't
Problems with the Exemplar View
-
Doesn't explain how concept was ___________ in the first place.
-
Explains _______________ of new instances, but not prediction, reasoning,
etc. so well.
Are there any classical categories?
-
Some categories have definitions
-
Even so, some instances are judged more prototypical than others
-
Core features (definition) vs. Identifying features
-
Identifying features used for fast indentification
-
Sometimes conflict with core features, which explains unclear cases
Similarity
Medin, Goldstone, & Genter (1993) study, see figure 10.7
Psychological _____________________
-
Objects have an essence (underlying nature) that makes them what they are
-
These underlying core features differ for different types of categories:
-
Nominal kind (clear definitions)
-
Natural kind (naturally occurring)
-
___________ (constructed to serve a function)
Natural Kind vs. Artifacts
-
Rips (1989)
-
Animal that looks like a bird suffers an accident that causes it to look
like an insect. It can mate with a bird, though, and produce bird offspring
-
Subjects judged it as more similar to ________t, but classified it as really
a _________
-
Carey (1982)
-
If found out that the a raccoon had skunk insides, subjects judge it to
be a skunk (biological definition)
-
A birdfeeder that was made out of a coffeepot is still really a birdfeeder
(___________ definition)