concepts and general knowledge Flashcards
define concept
mental representation used for a variety of cognitive functions
how to study concepts
categorization
categorization is the process by which things are placed into groups called categories
why are categories useful
help to understand individual cases not previously encountered
pointers to knowledge
provide wealth of general info about an item
-allows us to indentify the special characteristics of a particular item
defining categories according to philosopher ludwig wittdenstein
simple concepts have no definition
eg games
defintional approach to categorization
determine category membershup based on whether the object meets the definition of the category
doesnt work
not all members of everyday categories have the same defining features
the moe characterstic features an object has,….
the more likely we are to believe it is part of the category
family resemblance
category members may not be defined, but rather resemble one another
example category = games
defintional models work?
instead
nope not well
prototype approach
exemplar approach
prototype approach
prototype = typical
an abstract resemblance of the typical member of the cateogry
characteristic features that describe what members of that concept are like
an average of category members in the past
contains the most salient features
true of most instances of the category
differs individual to individual
graded membership = some members are closer to the prototype than others
fuzzy boundaries = no clear dividing line for membership
certain members are considered better than other (best red)
high vs low prototypicality
high = category member closely resembles category phenotypes = a typical member lo = category member does not closely resemble category prototype = unusual member
typicality effect
prototypical objects are processed preferentially
high = judged more rapidly
naming effect
prototypical objects are named first
depends on culure!`
priming effect
prototypical category members are more affected by a priming stimulus
eg hearing the word green primes a highly prototypical green more than a less typical green
prototypicality and attractiveness
affects judgements = most typical more attractive
explain the exemplar approach
concept is represented by multiple exemplars (rather than a single prototype)
exemplars are actual category members (not abstract averages)
to categorize, compare the new item to each of the individually stored exemplars (one by one)
more similar to a specific exemplar is to a known category and the more instances of that category we have, the faster it will be categorized
explains typicality effect, easily takes into account how atypical cases and variable categories
explain how both exemplar and prototype explains ….
typicality
graded memberships
illustration
typicality - average of a category vs encountered more often
graded membership - less similar to average vs how often it is encountered
illustration - ideal to less ideal member vs often to not as often encountered
every concept is…
a mix of exemplar and prototype
- early learning involves exemplars
- experience involves averaging exemplars to get prototypes
- with more experienec, we can use both
flexibility of exemplars vs prototypes
prototypes are economical but less flexible
exemplars are flexible but less economical
so do we use exemplars or prototypes
probably both depending on the category
exemplars = best for small categories, also highly variable ones
prototypes = best for large fairly homogenous categories, highly bariable categories (like games) create strange, unrepresentative prototypes
difficulties with categorizing via resemblance (typicality)
prototype and exemplar appraoch are based on typicality of membership - but cannot explain everything
so as number gets bigger we rate numbers as less typically odd or even
atypical features do no exclude category members (we still call red painted stripy squashed sweet lemon a lemon)
all the typical features but not category members (so perfect counterfeit bill = still not a bill)
hierarchical organization of categories
just as certiain category members seem to be priviledged, so are certain types of category
global (superordinate) level = very general
basic level = somewhat specific category
specific (subordinate) level - very specific category
evidence = basic level is psychologically priviledged
why do we use the basic level of categorization most
is the best / econonmical middle ground
evidence to show basic level is “special”
people almost exclusively use basic-level names in free-naming tasks
quicker to identify basic-level category member as a category member
children learn basic names first ie doggie
basic level is much more common in adult discpurse than names for superordinate categories
different cultures tend to use the same basic-level categories at least for living things
to understand how people categorize objects we must consider…
properties of objects
learning and experience of perceivers
knowledge network
model of the representation of knowledge
suggests knowledge is represented via a vast network of connection and associations between all of the information you know
-early models = higherarchical structure (collins and quillian)
sentence verification task
participants must quikcly decide whether sentences like robins are … birds, animals, have hearts etc are true
requires links along network tree
reaction time goes up for longer associative paths
spreading activation in semantic networks
activation is the arousal level of a node
when a node is activated, activity spreads out along all connected lnks
concepts that receive activation are primed and more easily accessed from memory
criticism of knowledge networks
cannot explain typicality effects
cognitive economy?
some sentence-verification results are problematic for the model (eg mammal vs animal)
semantic networks modification by collins and loftus
shorter links to connect closely related concepts
longer links for less closely related concepts
no hierarchical structure based on individual persons experience
assessment of semantic networks modification
is predictive and explanatory of some results but not all
generated multiple experiments
lack of falisifiability
-no rules for determining link length or how long activation will spread
-therefore there is no way to prove it wrong
-circular reasoning
propositional knowledge networks
localist representations - each node is equivalent to one concept
connectionist knowledge networks
distributed processing - information involves a pattern of activation
parallel processing of information occurs at the same time
-a netowrk of nodes and links but operates very differently from semantic networks
= neuron like units
knowledge represented in the distributed activity of many units
weights determines at each connection how strongly an incoming signal will activate the next unit
3 neuron like units explained in connectionist approach
input units - activated by stimulation from environmant
hidden units - receive input from input units
output units - receive inputs from hidden units
how learning occurs in the connectionist approach
network responds to stimulus
provided with correct response
modifies responding (based on correct response) to minimize error signal
learning can be generalized
error signal in the connectionist approach
difference between actual activity of each output unit and the correct activity
difference between answer given by neural network and correct answer
error signals cause a node to decrease its connections to input nodes that led to the error = back propogation
back propogation
error signal transmitted back through the circuit
indicates how weights should be changed to allow the output signal to match the correct signal
the process repeats until the error signal is zero
graceful degradation
disruption of performance occurs gradually as part of the system are damaged
case studies KC and EW
trouble with categories that represent living things but no trouble with non living things
different areas of the brain may be specialized to process information about different categories
other specific deficits eg recognizing the categories of persons, animals and tools
how are categories represented in the brain
by distributed activity, but in similar areas across all people
more similar patterns of brain activity for categories with similar features
category specific neurons have been found in humans and aninals (stimuli made up of part dog and cat showed specific neurons fire for dog and specific for cat)
are infants capable of categorization
familiarization/ novelty preference procedure
2 months can categorize global
3-4 months can categorise basic
6-7 months can categorize specific
categorization becomes fine-tuned with age
