conceptual knowledge
helps with recognition and generating inferences
concepts
mental representations of an object, event, or idea
category
examples of concepts that are grouped together
categorization
process of building a category
approaches to categorization
1) definition approach
2) family resemblance
3) prototype approach
4) exemplar approach
definitional approach
- categorization based on definition of the category
- each member of a category needs to meet the same criteria
- i.e. “a square is a plane figure having four equal sides…”
family resemblance
- categorization of objects based on ways they resemble each other
- each feature of objects within a category do not need to match
prototype approach
- categorization based on similarity with a prototype
- typicality effect: faster to verify prototypical members as belonging to a category than non-prototypical members
exemplar approach
- categorization based on the greatest similarity between an item in a category and the novel item (see if it matches an item within that category)
- drawback: too many exemplars in a category makes the comparison very time consuming
rosch’s hierarchical organization
1) global/superordinate category
2) basic-level category
3) subordinate category
superordinate category
- top of the hierarchy
- too broad
- can’t generate a prototype at this level
basic-level category
- middle of the hierarchy
- often first words learned
- categories are most differentiated at this level
- items within the basic-level category are similar (i.e. guitar and drums)
- items in different basic-level categories are dissimilar (i.e. guitar vs. apple)
subordinate category
- bottom of the hierarchy
which level is categorization the fastest?
- at the basic-level category
- there is an exception; i.e. if you are a dog trainer and know all breeds of dogs, you can categorize in the subordinate categories quicker, like different dog breeds
collins and quillian’s semantic network model
- hierarchical model
- explains the organization of semantic knowledge but goes one step further
- concepts are represented as “nodes”
- concepts are connected to related concepts through “links”
- retrieval of concepts involves spreading activation
activation
- anything that is similar to the original concept is activated with it
characteristics of semantic network models
- direction of movement is from the bottom to the top
- semantic facilitation occurs when the semantic path was just used
- accounts for the category-size effect
category-size effect
- faster to classify items that are part of a small category than a large category
- items at the bottom of the hierarchy
limitations of semantic networks
- can’t explain the typicality effect: typical members of a category are classified
more quickly than less typical members ( i.e. “a penguin is a bird” vs. “a canary is a bird”) - classification speed is not always dependent on amount of level changes (i.e. “a chimpanzee is a primate” is only one level change = slow RT vs. “a chimpanzee is an animal” is two level changes = faster RT)
connectionist model
- concepts are represented as “units”
- connection weight of each unit determines the activation or inhibition of a signal
- back propagation: you can move forwards and backwards with these nodes
representing concepts in the brain
- sensory-functional hypothesis
- multiple-factor approach
- semantic category approach
- embodied approach
sensory-functional hypothesis
we identify living things based on sensory features and nonliving things based on their function
multiple-factor approach
we identify living and nonliving things based on sensory features such as features, function, motion, colour
semantic category approach
neural circuits are associated with specific categories
