Cog Psych CH 9
The Approaches to Categorization (e.g., define, characteristics, limits)
- classical approach - prototype theory - exemplar theory - knowledge based theories - embodied cognition - neural representation of knowledge
psychological essentialism
‐ The idea is that all category members possess a fundamental essence that is unique to that category and determines membership.
murphy and allopena 1994
found participants have difficulty learning about things that do not "make sense". When we learn about categories, we try to make meaningful connections from our past knowledge. We rely on categories to teach us about the world and use our knowledge about the world to help explain category membership.
Embodied Cognition
no general theory some suggest body influences cognition some think body has causal role in thought knowledge uses similar sensorimotor neurons as perception and action do
embodied cognition studies
FRMI showed predictable pattern of brain activity in motor areas found that doing actions and silently reading action words activated similar motor areas of the brain supports claim that knowledge is stored as modality specific neural activity knowledge is goal driven, flexible, and context dependent
Knowledge Based Theories
Ideas about category membership are implicit. membership not based on features; it is based on broad theories about essentialism. address the problem of context It is difficult to categorize the features because there are very few apparent similarities between them. once you have knowledge of the category, you can understand them as they relate to a specific goal
Symbol Grounding Problem
Related to the problem of how words (symbols) get their meanings, and hence to the problem of what meaning itself really is. • Analogous to trying to learn Chinese using a Chinese dictionary alone. explaining to an alien what the apple is
knowledge based approach to categorization
We rely on our broad knowledge base to explain the reasons for category membership.
concept
a mental representation of objects, ideas or events. explains WHY items belong together
semantic dementia
a progressive neurodegenerative disease characterized by an inability to name objects ATL damaged
category
a set / list of items that are perceptually, biologically, or functionally similar. these items are considered equivalent, and have same function
family resemblance
all category members share at least one feature with another member of the category, few features shared between all members
Probic, Jefferies, and Ralph (2010) and Neural Representation of Knowledge
applied TMS to the ATL and inferior parietal lobule (IPL) of healthy participants while they named pictures of both living and non‐living things. IPL is a cortical region involved in visuallyguided hand movements and corresponds to a spoke in the hub‐and‐spoke model.
ad hoc context categories
are an example of the flexibility of categories. define what categories activates belong
Rosch 1975 typicality
asked participants to rate items of how good an example they were of a category participants gave typicality ratings easily and participants ratings correlated nearly perfectly.
Berntsen and rubin 2004
asked students to imagine a newborn baby and make predictions about what would happen in babies life as they got older, large amount of overlap across life events to people in the same cultures.
typicality effects
behavior directed differently toward typical items compared to atypical ones. when listing categories, we name typical ones first, faster to put them into categories, and show priming effects accounting for atypical category members
commonsense knowledge problem
cannot give computers all the common sense we have
Classical Approach to Categorization
categories are defined by sets of features that are both necessary and sufficient for category membership if exemplar possesses the required defining features, it belongs to that category if it doesn't posses the defining features, it does not belong to the category limitation: impossible to decide absolute, final features (like 3 legged dog) also it claims items either do or dont belong to a category
barsalou 1993
categories flexible depending on the context.
Prototype Theory of Categorization
compare to a very specific item categorize items using characteristic features to compare prototype stored in memory limitations: categories have fuzzy boundaries like tomato, typical members of a category have more characteristic features than do atypical category members limitation: context effects, typicality effects
Neural Representation of Knowledge
damage to ATL means inability to name objects healthy brains barley demonstrate ATL activation during tasks with semantic memory, shows hub and spoke model supports ATL as general semantic hub supports semantic knowledge is stored in a localized and distributed way in our brains , more flexible than earlier theories
superordinate categories
distinctive, but not informative knowing something is an animal provides little information compared to knowing it is a dog (not informative) can distinguish easily from other categories at this level (distinctive) animal, fruit, professional
characteristic features
features that are likely to belong to category members but are not required for category membership
hub and spoke model
generalized and abstract semantic knowledge is stored in a semantic memory hub in the ATL, where general knowledge was stored
hierarchies
how prototype theory states individual items can belong to multiple levels basic level, subordinate, superordinate
classical approach to AI and cognitive psychology
idea that humans are information processes that receive input, use rule based strategies to manipulate information, and produce a behavioral output, much in the same way as a computer
reconstructive nature of memory
if we dont have memory for a specific item, we can use schematic knowledge to fill in the blanks
What are exemplars?
individual items in a category (different chairs)
subordinate categories
informative but not distinctive provide a lot of information about the items (informative) share many features in common (not distinctive) categories within a subordinate category are very similar. German Shepard, granny smith, cardiologist
Lexical Decision Tasks
measure how quickly people classify something as a word or nonword. took longer to put atypical members into the category than typical Only typical category members show these facilitatory priming effects. Collectively, these observations that we behave differently toward typical items compared to atypical ones, are called typicality effects.
Collins and Loftus (1975) Spreading activation model of semantic memory
nodes are connected to each other via semantic relatedness rather than hierarchical structure nodes connected to each other via pathways
schemata
our organized knowledge based about a particular topic
sentence verification task collins and quiliam
participants presented with sentences and asked if true or false results confirmed model
semantic network models
problem w hierarchal model, they dont account for typicality effects use lexical decision task to demonstrate semantic priming between two related words
Bartlett 1932
proposed that what we remember is influenced by our past experiences and knowledge includes everything we know about a particular thing, event, person, or situation broad and not well defined
Exemplar Theory of Categorization
proposes we store in memory examples of items we have encountered in the past compare new items to ones in memory, look for similarity between characteristic features explains context effects explains typicality effects an atypical member is more difficult to identify because there are fewer similar examples Differences show we have implicit ideas about category membership. Neither the prototype nor the exemplar theory specifies how we decide which features are important.
Zwaan, Stanfield, and Yaxley (2002) and Embodied Cognition
provided evidence that knowledge access depends on context read senates than saw picture, they would agree the object is in there if it was in the same context context dependent
Hauk, Johnsrude and Pulvermüller (2004) and fMRI of Embodied Cognition
provided support for the idea of distributed, modality‐specific knowledge representation. Using fMRI, they observed brain activity while participants were moving their tongues, their fingers and their feet. These motor actions elicited a predictable pattern of brain activity in motor areas of the brain.
basic level categories
provides the right amount of information about the category to provide useful information (informative) and can be used to distinguish members from embers of other categories (distinctive) most cognitively efficient dog, apple, doctor
information processing approach
sensory input > operation > behavioral output
armstrong, gleitmal, and gleitmal 1983
showed that people apply similarity ratings to categories that clearly are rule based participants agreed there are no "more off" numbers, and then gave odd number typicality ratings This suggests that perhaps the typicality ratings of categories, such as birds and fruit, are an artifact of the experimental method used and not indicative of actual fuzzy category borders.
cognitive economy
storing property only once at the highest level in the hierarchy
property inheritance
subordinate categories inherit the properties of the superordinate categories they are connected to node becomes active as a result of input from the environment nodes farther away from each other take longer to active
Collins and Quillian (1969) Semantic Network Models
suggested knowledge is stored as concepts within a network of interconnected units called nodes
prototype
the most typical member of the category, would receive the highest typicality rating (robin) has all the characteristic features of a category
cyc knowledge base
to input all of human knowledge into a computer knowledge base contains 1.5 million concepts and 20 million rules
method of repeated reproduction
to investigate schemata and the role of knowledge on memory, reproduce a drawing multiple times, as it goes on it looks less similar and starts to look more like familiar object over time details are lost from memory bur we can use the information from our schemata to guide memory retrieval used schematic knowledge about a face
neural representation of knowledge/ IPL stimulated
took more time to name nonliving things no change for living things support the role of IPL as a modality specific spoke
allen and brooks 1991
trained participants to identify drawings as diggers or builders based on physical features when shown a new feature, they can categorize when physically similar to the same category, not similar to a different category Instead of using the rule to categorize the new creatures, participants categorized the new item according to how similar they were to previously seen items. Made more than double the number of errors when the new items were physically similar to the wrong category.
limitations to BOTH theories
typicality ratings, theories are similarly- based Researchers agree we are flexible in our categorization strategies and may use prototype matching or exemplar retrieval in different circumstances.
context effects
typicality should depend on the number of shared features between category members. it doesn't, typicality depends on context
Schema View
use schemata to remember things, people with similar experiences will have similar schemata, can use shared cultural knowledge to make inferences and predict behavior
Meyer and Schvaneveldt (1971) and Semantic Network Models
used lexical decision task to demonstrate semantic priming between related words (butter and bread faster when together) the more similar the concepts, the more connections between them and the shorter the distance explains typicality effects
