Knowledge and Imagery Flashcards
knowledge
information or skills acquired through experience
concepts
mental representations of groups of things
categories
groups of things themselves, as they exist in the real world
type
specific mental representation
token
specific real life thing
concrete concepts
actual things
abstract concepts
your interpretation of these things
what occurs when we recognize an object?
we classify it
how is communication related to concepts?
communication is discussion of concepts we have
2 types of concepts
- formal
2. natural
formal concept
true by definition
example of a formal concept
a triangle has 3 sides
natural concept
defined by perceptions and interactions in the world
3 levels to concepts
- superordinate
- basic
- subordinate
superordinate level
super general
basic level
good balance between superordinate and subordinate
subordinate level
very specific
what is the basic level dependent on
your expertise in that category
age of acquisition
we learn names for basic level before superordinate and subordinate
classical theory
concepts have a definitional structure in the form of a list of attributes
according to classical theory, what determined an object?
if it matched the definition concept
necessary conditions
every member of a classification meets these
sufficient conditions
you don’t necessarily need this trait
classical theory category membership
you are either 100% in or 100% out
- everything is equally in
are classical theory properties transitive?
yes
transitive
properties carry over
- every that is X is also Y
5 problems with classical theory
- fuzziness
- intransitivity
- typicality effects
- ad-hoc categories
- culture
typicality effects
not all concepts in a category are equal
who created prototype theory?
Rosch
prototype
hypothetically typical or idealized item
how does prototype theory categorization occur
when you compare what you are looking at to the prototype
does prototypicality vary?
yes, it can vary from high to low
who coined family resemblance
Wittgenstein
family resemblance in prototype theory
one thing can look a bit like another, without necessarily sharing every common feature
4 problems with prototype theory
- defining prototypes
- prior knowledge effects
- ad-hoc categories
- context
prior knowledge effects
prior knowledge influences which features you pay attention to
exemplar theory
compare what you are looking at to an exemplar
exemplar
an individual item or token
what happens when you store each token?
you preserve variability of items
how does exemplar theory explain ad-hoc categories?
with all tokens in the mind we can make new groups
2 problems with exemplar theory
- storage of every exemplar ever encountered
2. definition of exemplar
why is definition of exemplar a problem?
is an exemplar every object or every encounter with that object
2 categories of models
- semantic networks
2. connectionist models
who proposed the semantic network model?
Collins and Quillian
how are nodes connected?
in meaningful ways
exceptions in the semantic network model
include the exception with notation in the model itself
spreading activation
method of searching the semantic network
what type of model is the semantic network model
hierarchical
3 problems with the semantic network model
- typicality effect
- cognitive economy
- sentence verification times
connectionism
an approach to modeling knowledge as an emergent process of interconnected units
PDP
concepts represented as activity distributed across a network
is connectionism a neural model
no
3 levels to connectionism
- input units
- output units
- hidden units
input units
receive info from senses
output units
give a response
hidden units
change information flow and allow models to emerge
what does the activation of one unit depend on? (2 things)
- signal that originates from the input
2. connections throughout the network
how are concepts represented
distributed pattern of activity across all units in the network
training networks
several iterations and simulations for the system to learn
what feature is critical for training networks
feedback
feedback
error signal going back into model saying a mistake is made
back propagation
when a mistake is made, the signal propagates backwards altering these connections
learning in connectionism
altering strength of connections
3 advantages of connectionism
- intuitive
- mimics brain damage
- generalization of learning
how is connectionism intuitive
looks like what we think is happening in the brain
how does connectionism mimic brain damage
if certain units are destroyed the model still outputs responses
how does connectionism show generalization of learning
we can generalize what we learn to other new circumstances
mental imagery
internal representation of perceptual experience without sensory input
main type of mental imagery
visual imagery
6 functions of imagery
- solve
- predict
- mental models
- memories
- navigate
- learn
4 properties of mental imagery
- scan images
- zoom in
- rotate images
- compose things
analogue
parts have many important qualities as picture but aren’t pictures
proposition
most basic unit of meaning that has a truth value
who said that imagery is analogue?
- Paivio
- Shepard & Metzler
- Kosslyn
who said imagery is propositional
- Pylysyn
who created the dual coding hypothesis
Paivio
Paivio experiment
paired associate task with abstract and concrete words
Paivio results
best at CC, then CA, then AA
- because people can create a visual representation of concrete items
dual coding hypothesis
concrete words are coded with verbal information and a mental image
Shepard and Metzler
rotation experiment thing
Kosslyn experiment
- mentally scan objects and imagine them after
- focus on one part and then answer questions of increasing distances from that part
- it takes longer to answer questions further away
Pylyshyn
representation is an epiphenomenon
how is representation an epiphenomenon
just an after effect of what the brain is doing
Pylyshyn argument about demand characteristics
people have tacit knowledge to give answers that they think are right
Kosslyn mental walk
walk up to an animal
- people were accurate when determining how close they were when the animal occupied the field of view
- proves analogue
Finke and Pinker experiment
dots appear on one screen
- arrow appears on another screen
- larger distance between dot and arrow = longer time to respond
- disproves tacit knowledge being an effect
Farah experiment
imagine a T and H on each screen
- T and H flash up on each screen
- if these coincide you are correct, if not you’re incorrect
what did Farah’s experiment show
shows that perception and visual imagery can interact
what do medial temporal lobe neurons respond to
balls but not faces
are MT neurons imagery neurons?
yes, because they respond when imagining balls too
why are less areas of the visual cortex active when imagining?
because imagining is less vivid than actually seeing
what other effect does imagining have on the brain?
some areas deactivated to tune out interference (auditory)
TMS experiment
inactivating visual areas also slows down imagery
- this shows imagery is not an epiphenomen
MGS mental walk
half of occipital lobe removed greatly impacted ability to do mental walk task
