concepts and categories Flashcards
what is a concept?
what does it allow us to do?
mental representation of category
enable us to generalise from past experiences of instances to predict behaviours etc. of new instance
what are the 3 theories of how concepts are represented in memory?
- classical view
- features - abstractive theories:
- feature lists and/or prototypes - exemplar/instance theories
what is the purpose of conceptual hierarchies?
provide economy of representation in memory:
1) properties common to all members stored just once
2) subordinate categories inherit properties of superordinate e.g to ‘have wings’ is a property of superordinate ‘bird’ and so applies to all subordinates like ‘canary’
how does evidence support the economy principle of conceptual hierarchies?
longer to retrieve when ask for property stored at highest level
search down many links to determine if category and property linked
link between familiarity and verification time?
more familiar concepts verified quicker
which kinds of knowledge are hierarchical tree structures appropriate for?
what are some types of knowledge that tree structures aren’t appropriate for?
object concepts
not appropriate for:
semantic memory e.g adjectival properties and schemas
how do we grow conceptual hierarchies?
from basic level of abstraction
e.g single words
what is the classic account of conceptual representation?
why does this theory fit with conceptual hierarchies?
defining features - aristotle
classify by reference to mental definition listing features of category’s members
e.g sufficient and necessary features
ideal member (most features) is prototype at centre and marginal members on outside near marginal members of other concepts
fits with conceptual hierarchies as share some features with superordinate category but differ from coordiantes
what is a problem with the classical view?
hard to come up with defining features or any at all for some concepts e.g furniture and games
examples of why not all category members are equal?
- category boundaries fuzzy and context-dependent
- typicality effects - typical members reacted to earlier
- similarity effects
what is the modified feature-list theory/abstractive theory?
category membership share resemblance not essential list of features
categorisation determined by which category’s features produce the highest weighted sum of matching features (need some features but none are necessary)
prototype = statistcial average
e.g has more features to be in one category than another
what is a prototype?
average member -
central tendency of a distribution if continuous
feature list and dimensional representation if discrete
what assumptions are included in prototype theories (abstract and classical view theories)?
- properties of categories abstracted from encounters with many members
- assign category whose prototype has highest similarity to instance considered
what is involved in instance/exemplar theories?
no abstract representation of category is formed (prototype) merely hold instances in memory
assign stimulus to category whose instances have the greatest summed similarity to the stimulus
evidence from the ‘prototype effect’ supporting both prototype theories and exemplar theoy also?
prototype not seen before classified faster than old or new instances (bigger difference after big delay)
evidence for abstraction of prototype (prototype theories)
also evidence for exemplar theory as prototypical instance more similar to instances in memory than new instance
