knowledge (concept, categories) Flashcards
1
Q
what is conceptual knowledge?
A
knowledge that enables us to:
- recognise objects/events
- make inferences about their properties
2
Q
concepts have been defined in a number of ways. name 2.
A
- mental rep of class or indivudal (smith, 1989)
- categories of objects/events/abstract ideas? (kiefer, 2012)
3
Q
knowledge exists in the form of _______ and one way we can organise that is using ______.
A
concepts; categories
4
Q
choose 1. a category includes all/some/a few of the possible examples of a particular concept.
A
all
5
Q
_________ provides the rules for creating _______.
A
concepts; categories
6
Q
categories have also been called _______ to knowledge. (yamauchi & markman, 2000)
A
pointers to knowledge
7
Q
when does categorization become difficult?
A
when you encounter something unfamiliar
when you have a brain injury and you cannot identify different objects or know what they are used for
8
Q
Things that closely resemble the prototype have high __________.
A
family resemblance
9
Q
what is family resemblance (Wittgenstein)?
A
- items in a category resemble one another in various ways, but there isn’t one essential feature that all of them share
- allows for variation within a category
10
Q
family resemblance suggests that membership is based on ________ rather than _________.
A
a network of overlapping similarities; strict definitions
11
Q
give an example of a concept
A
The concept of “dog” includes characteristics like four legs, fur, barks, friendly, etc
**A concept is a mental representation or idea of something. It’s how we mentally group things that share common features.
12
Q
give an example of a category
A
The category of “dog” includes real dogs like a Golden Retriever, a Chihuahua, and your neighbor’s pug.
**A category is a group of real-world items or instances that fall under a concept. (how we apply concepts to the real world)
13
Q
define cateogrisation
A
process by which things ar eplaced into categories
14
Q
concept or category more important?
A
Concepts are more important as the mental foundation.
Categories are important too, but they come AFTER you form concepts.
15
Q
what are the 3 main approaches in categorisation?
A
- definitional approach
- prototype approach
- exemplar approach
16
Q
what is the definitional approach in categorisation?
A
determining if object meets the definition of a category
- works well for some (eg. geometric objects)
- doesnt work well for natural objects (eg. birds, trees, human-made objects like chairs) because members of everyday categories don’t always have the same features
17
Q
why doesnt the definitional approach work for natural objects like birds, trees and human-made stuff like chairs? give examples.
A
1) members of everyday categories don’t always have the same features (real objects vary alot)
2) exceptions, borderline cases, and fuzzy boundaries
examples:
- “chair” can have 4 legs, 3 legs, or no legs at all (like a beanbag). yet they are still chairs.
- Birds: Penguins and ostriches don’t fly—but we still know they’re birds. Bats fly, have wings and lay eggs, yet they are not birds.
18
Q
what is the prototype approach?
A
membership in a category (eg. dog, bird, chair) is determined by comparing object to prototype that represents the category
19
Q
what does prototype mean?
A
a typical or average member of a category as a representation, not a real member
20
Q
high vs low typicality in the prototype approach of categorisation.
A
high typicality means category member closely resembles category prototype
low typicality means category member does not closely resemble prototype
21
Q
give examples of high and low typicality using the prototype approach of categorisation.
A
High typicality → Very similar to the prototype
Eg. A robin is a very typical bird.
Eg. A dog like a golden retriever is a typical pet mammal.
Low typicality → Not very similar to the prototype
Eg. A penguin is a bird, but it doesn’t fly or sing—so it has low typicality.
Eg. A bat is a mammal, but it flies and looks like a bird or a mouse—so it also has low typicality.
22
Q
Rosch’s experiment supports the ________ theory that people categorise things based on __________ and not _____________.
A
prototype; idealized prototype; strict definitions
23
Q
what is the method in Rosch’s experiment?
A
Participants were shown items from categories (like bird or furniture)
asked: “How good of an example is this of the category?”
They rated items on a typicality scale.
24
Q
what kind of family resemblance do prototypical objects have?
A
high
25
Which research shows that there is strong r/s btwn family resemblance and prototypicality? include method and result.
Rosch and Mervis (1975)
method: listing characteristics of common objects (chair, mirror, telephone, sofa)
result:
- chair and sofa shared many characteristics (high family resemblance)
- mirror and telephone had less overlap (low family resemblance)
26
statements about prototypical objects are verified _______.
rapidly
27
what is the sentence verification technique? what is the results in the study by smith et al., 1974? what effect does it demonstrate?
participants answer "yes" or "no" to statements about an object's category (eg. an apple is a fruit)
result: faster responses for high-prototypicality objects (apple) than low (promegranate)
typicality effect:
high-prototypicality members of a category are processed faster and more easily than low-prototypicality members
28
according to mervis et al., 1976, when listing objects in a cateogry, people tend to list the most _____________ members first.
prototypical
29
prototypical objects are affected more by priming.
1) which experiment showed this?
2) what is the method?
3) what is the result?
4) what does it support?
1) rosch (1975b) experiment
2) priming participants wth a color (green), then showing them a pair of colors, either same or differrent (one prototypical and one nonprototypical)
3) faster "same" judgements for prototypical colors (primary green) than nonprototypical colors (light green).
4) supports that people create images of prototypes
30
what is the exemplar approach to categorisation?
determining if an object belongs to a category, using many examples from that category rather than a single prototype (idealized, average example)
31
what is are exemplars?
actual objects/events of a category that we have encountered in the past
32
which of the 3 approaches to categorisation works the best?
- both exemplar and prototype approach
- averaging exemplars into prototype initially, then strengthening exemplar information later
- prototype provides general idea, exemplar provides specific knowledge.
33
what is the strength of exemplar approach?
can account for:
1) atypical cases (eg. penguin as a bird)
2) variable categories (different kinds of games eg. boardgames, nintendo switch, childhood games)
34
prototype vs exemplar
1) prototype is idealised, average idea, exemplar is actual object/event of a category that we have encountered in the past
2) prototype provides general idea, exemplar provides specific knowledge
35
define hierarchical organisation
larger, general categories divided into smaller, specific ones
36
Rosch's Approach consists of 3 Hierarchical Organization of Categories. what are they? give an example of each
1. global (superordinate)
- furniture
2. basic
- table
3. specific (subordinate)
- kitchen table
37
Rosch's demonstration of Hierarchical Organization of Categories required participants to do what? what is the result?
list common features of objects
result: participants listed few features for global, more for basic and slightly more for specific
38
why is the basic level in Rosch's Approach special? give an example.
- in describing an object, the basic level provides the most useful info with the least cognitive effort
ie. going to a more general level (global/superordinate) loses alot of info
ie. going to a more specific (subordinate) level gains only a little information
39
how does knowledge affect categorisation?
- rosch's experiment shows that basic level is preferred by undergraduates when describing things
- experts categorise differently (tanaka & taylor (1991) showed that bird experts used species name while nonexperts said "birds" [experts focus on more specific info)
- therefore, categorisation depends on experience
40
categorisation depends on ___________ and ___________.
experience/knowledge
which object characteristics we pay attn to
41
____________ networks (eg. Collins and Quillian's Hierarchical Model) focuses on how categories/concepts are organised in the mind.
semantic
42
_____ and ____ are key components in a __________ network which focuses on how categories/concepts are organised in the mind.
nodes; links; semantic
43
nodes represent what in a semantic network?
individual categories/concepts
- bird
- canary
- animal
- has wings
44
links represent what in a semantic network?
connections btwn nodes indicating r/s btwn concepts/categories
- is a
- has
eg.
Canary → is a → Bird
Bird → is a → Animal
Bird → has → Wings
45
the semantic network is structured in a __________ organisation.
- we organise ___________ from _______ categories to ______ examples
- more general concepts are at ________ levels
- more specific concepts are at ________ levels
arrange their levels: bird, animal, canary
hierarchical; knowledge; broad; specific; higher; lower
animal
bird
canary
46
3 key principles/implications of collins and quillian's hierarchical model
1) retrieving properties
- typically start at lowest node, move towards higher ones to gain more general info
2) cognitive economy
- crucial feature of model
- shared properties are stored only once at higher-level node
- eg. "can fly" node is stored at "bird" node, and not repeated at "canary" or "eagle" node
- efficient, avoids redundant storage
3) exceptions
- eg. ostriches cannot fly
- allows for addition of exceptions at lower nodes
- "ostrich" node will have "cannot fly" node even though general "bird" says they can
47
which is easier to verify? why?
a canary is an animal
a canary is a bird
a canary is a bird is easier to verify
- traverse through fewer links
48
activity spreads along links connected to activated node in the collin and quillian's hierarchical model. what does this do? and what is this process called?
activates related concepts, priming them for easier retrieval
spreading activation
49
the lexical decision task (meyer & schvaneveldt, 1971) demonstrates ___________.
method
results
reason
spreading activation
method:
- participants read stimuli and indicate if each entry is a word or non-word
- variation with pairs of words
result:
- faster rxn time when words are associated (bread/wheat)
reason:
- triggered spread of activation to nearby locations in network
50
3 criticisms of collins and quillian's model
1) cannot explain typicality effect
2) questioned cognitive economy - some ppl may store specific properties at nodes ("have wings" at "canary")
3) led researchers to look for alternative ways of using networks
51
why can't collin and quillian's model explain typicality effect?
1) equal distance, equal time
- assumes that all members of category are equally representative
- verifying "a canary is a bird" should take the same amount of time as verifying "an ostrich is a bird" because both canary and ostrich are 1 node away from bird in the semantic network
2) typicality effect
- in reality, ppl verify statements about typical members (canary) faster than those about atyical members (ostrich)
therefore,
- treats all members as equal
- focuses on hierarchy and distance
- doesn't account for the fact that some category members are more typical than others
- thus cannot explain why ppl respond faster to tpical examples than atypical ones
52
what is the connectionist approach aka Parallel Distributed Processing?
another theory that explains how knowledge/info is represented in the brain
it is like a computer model representing cognitive processes.
concepts are represented by activity distributed across network
53
connectionist approach consists of what?
units - circles in network, inspired by neurons
connections - lines transferring info btwn units
input units - activated by stimuli from envt
hidden units - receive signals from input units, send signals to output units
output units - receive signals from hidden units
connection weights - determine how signals either increase/decrease activity of the next unit, inspired by synapses
54
connectionist approach includes connection weights. what do they do?
determine how signals either increase/decrease activity of the next unit, inspired by synapses
55
in a connectionist network, concepts/categories/properties are represented by _____________ instead of _______ in collin's semantic network.
activity pattern; nodes
56
in a connectionist network, activation depends on what?
signal in input units and connection weights
57
stimulus presented to ______ is represented by activity across units.
input units
58
how are properties and concepts represented in a connectionist network?
model by mcclelland and rogers (2003)
- properties are indicated by activity in attribute units and by pattern of activity in representation and hidden units
- concepts are indicated by activity pattern in all units
59
does a network require training?
yes
60
what is back propagation?
error signal sent back through the network due to initial erroneous responses
61
in training a network, error signals adjust ________.
network is trained on multiple _____ with small _______ changes after each _____________.
connection weights
concepts; weight; presentation
62
learning process in training a network
begins with incorrect connection weights
modified in response to error signals
63
is semantic network model (collin and quillian) or representation/connectionist model (mcclelland and rogers) more complex and more like a brain? why?
representation/connectionist model (mcclelland and rogers) because it involves more units per concept
64
what 2 results from research support connectionism?
1) graceful degradation
- operation won't be completey disrupted by damage because info is distrubuted across many units
- disruption occurs gradually when systeem is damaged, just like brain damage resulting in partial loss of function
2) generalisation of learning
- training a system to recognise properties of one concept provides info about related concepts
65
4 neuroscientific/biological models explaining how concepts are represented in the mind are?
1. sensory-functional hypothesis
2. multiple-factor approach
3. semantic category approach
4. embodied approach
66
what does the sensory functional hypothesis (SF hypothesis) state?
- ability to differentiate depends on memory system distinguishing sensory attributes and a system distinguishing functions
- distinguishing living things depend on sensory features (5 senses) while artifacts (tools, furniture) depend on function
- research based on patients losing ability to understand types of concepts due to brain damage
- category-specific memory impairment - losing ability to identify one object type (eg. fruit), retaining ability to identify others (eg. furniture, animals)
67
2 limitations of SF hypothesis
1. unexplainable cases like presence of sensory deficits but better identification of animals than artifacts
2. artifacts isnt a homogenous category
- there are patients who can identify mechanical devices better than other types of artifacts
- there are patients who show poor comprehension of small artifacts (tools) but better knowledge of larger ones (vehicles)
68
what is the multiple-factor approach in explaining how we represent info in our brains?
We use many factors (not just sensory or function) to represent concepts
eg. Shape, motion, color, taste
69
experiment by hoffman and lambon ralph (2013) demonstrates the __________ approach.
method
results
multiple-factor approach - We use many factors (not just sensory or function) to represent concepts
method: participants rated large number of concepts (both living and non-living things) based on a wide set of features (visual properties, function, motion, color, etc.)
results:
- animals more associated with color/motion
- artifacts more associated with function
- some overlap (eg. musical instruments)
70
what is crowding?
some categories of concepts (especially living things) tend to share many features with each other — making them more similar and harder to tell apart
eg. animals share many properties, while artifacts share fewer, so artifacts are easier to distinguish
71
how does crowding affect recognition?
3 points
1. feature overlap
- highly crowded (many common features)
- features that define members overlap significantly
- difficult to distinguish
2. impaired discrimination
- difficulty processing/distinguishing btwn features makes it difficult to differentiate btwn items in a crowded category
3. apparent category-specific impairment
- due to impaired discrimination, it might seem as though the person has category-specific impairment
- eg. difficulty recognising animals but not artifacts, BUT actually they are just struggling to distinguish btwn objects that are highly crowded (many common features)
72
_____________ may manifest as apparent category-specific impairment. give an example.
impaired discrimination
eg. difficulty recognising animals but not artifacts, BUT actually they are just struggling to distinguish btwn objects that are highly crowded (many common features)
73
describe the semantic category approach
brain has innate, specialized systems for certain categories that are biologically important (e.g. faces, animals, tools)
not just one tiny spot in the brain for each category, activity soreads out across multiple brain areas
74
to support semantic category approach, huth et al., 2012 mapped brain using _____, which created what?
experiment repeated in 2016, using ____ this time. what 2 things did it show?
data is recorded using _______.
these maps are ___________ from person to person.
2012 - movies (created map showing whwere diff categories are processed in brain)
2016 - stories
1. brain can organise info about categories even with just listening to language
2. abstract ideas have specific locations of activation
fMRI
similiar
75
what is the embodied approach?
knowledge of concepts is based on reactivation of sensory and motor processes during interaction with object
76
give an example of an embodied approach.
when we use hammer, sensory areas (size, shape, color, feel) and motor areas (actions involved in using it) are activated.
later when we see or read the word "hammer", these same sensory and motor areas are reactivated, representing our knowledge of the hammer.
77
embodied approach - knowledge of concepts is based on reactivation of sensory and motor processes during interaction with object.
what is this reactivation faciliated by?
mirror neurons - fire/activate when performing an action and when observing someone else performing it
78
hauk et al (2004) experiment demonstrates the _________ approach in explaining how info is represented in our mind.
method
result
criticism
embodied
method: used fMRI to examine brain activity while participants verbally processed different types of action-related words
result: motor regions of the brain were activated when participants processed action words related to specific movements (Semantic somatotopy)
criticism: not well-suited for abstract concepts like democracy/truth
79
define Semantic somatotopy. which approach does this fall under and which experiment demonstrate this?
correspondence between words related to body parts and location of brain activity
falls under embodied approach
demonstrated in hauk et al (2004) - fMRI to examine brain activity while participants verbally processed different types of action-related words
80
hub and spoke model is based on patients with ________, which is the ____________.
hub and spoke model states that areas with __________ functions connect to __________ (ATL) which is the ____ that __________ information and holds ____ information. damage to _____ (sensorimotor areas that store more specific sensory and motor info) causes ____________, while damage to ATL (hub) causes ___________.
semantic dementia; general loss of knowledge for all concepts
specific; anterior temporal lobe; hub; integrates; general semantic; spokes; specific deficits; general deficits
81
in proving hub and spoke model, ____________ (TMS) studies can temporarily _______ brain activity in specific regions.
TMS to the _____ disrupts general knowledge about concepts, showing the _____'s role in conceptual integration.
TMS to sensorimotor areas affects the specific knowledge of objects or actions tied to those areas, demonstrating the role of the _______.
Transcranial magnetic stimulation; disrupt
ATL; hub
spokes
