1
Q
What is conceptual knowledge?
A
- knowledge that enables us to recognize objects and events
- to make inferences about their properties
2
Q
Define concepts
A
- mental representation of a class
- or individual categories of objects, events and abstract ideas
3
Q
Define categories and categorization
A
categories:
- all possible egs. of a particular concept
- they provide rules for sorting objects into categories
categorization:
- process by which things are placed in categories
- once we assign an object to a cat, we know alot about it
concepts and categories are often discussed tgt.
4
Q
What is the Definitional approach to categorization?
A
- we decide whether smth is a member of a cat
- if it meets the definition, it does
- (-) prob is that not all the members of everyday cats have the same features
- there will be some objects which do not meet the definition
5
Q
What is Wittgenstein’s family resemblance?
A
- deals with the prob that not all object may fit a fixed definition
- refers to the idea that things in a particular cat resemble one another in a no. of ways
6
Q
Describe the Prototype approach to categorization (Rosch, 1973)
A
Prototypicality:
- membership in a cat is determined by comparing the object to a prototype that represents the cat
- the typical prototype is based on an avg member of the cat that are commonly experienced
eg. the prototype for birds may be based on those you usually see: sparrows, robins, but dont necessarily look like any one of them
7
Q
Explain Typicality and the Typicality effect, which are commonly associated with the Prototype approach to categorization
A
Typicality:
- variations within cats that represent differences
- high typicality -> cat member closely resembles the cat prototype
- low typicality -> cat member does not closely resemble a typical member of the cat
- eg. not all birds are like robins or sparrows
The Typicality effect:
- ability to judge highly prototypical objects more rapidly
- when partis are asked to list as many objs in a cat, they tend to list the most prototypical members first
8
Q
Explain the Sentence Verification Technique.
A
- to determine how rapidly people can answer questions about an obj’s cat
- partis are presented with statements
- they are asked to ans yes if they think the statement is true
- or ans no if they dont think so
9
Q
Explain Rosch (1985b)’s experiment on prototypical members being affected by priming.
A
priming: occurs when the presentation of one stimulus facilitates the response to another stimulus that usually follows closely in time
- partis first hear the prime (name of color green)
- they then see a pair of colors side by side
- they need to press a key as quickly as possible if colors are same or different
- RESULTS: priming results in faster ‘same’ judgements for prototypical colors vs non protoypical colors
- prototypical colors: good repres of the colors
- non prototypical colors: poor repres of the colors
- hence when partis hear green, they judged 2 patches of primary green as being the same MORE RAPIDLY than two patches of lighter green
- rosch explains that when partis hear ‘green’ they imagine a good or highly prototypical green
- the prime will facilitate the participants’ response to a stimulus if it contains some of the info needed to respond to the stimulus
10
Q
Name the 4 effects that come under Prototypicality.
A
- family resemblance- things in a cat resemble each other in a no. of way s
- typicality- ppl react rapidly to members of a cat that are ‘typical’ of the cat
- naming- ppl are more likely to list some objs when asked to name objs in a cat
- priming- presentation of one stimulus affects responses to a stimulus that follows
11
Q
Describe the Exemplar approach to categorization
A
- involves determining whether an obj is similar to other objs
- the standard for this approach here involves many exemplar, each one is called an exemplar
exemplar:
- actual members of the cat that a person has encountered in the past
- hence if a person encountered sparrow and robins, each of these would be an exemplar for the cat of birds
12
Q
What are the advantages of using Exemplar Approach as compared to the Prototype Approach?
A
- The exemplar approach explains the typicality effect by proposing that objs that are like more of the exemplars are classified faster.
- eg. because a sparrow is similar to many bird exemplars, it is classified faster than a penguin which is similar to few bird exemplars. - By using real egs. it can more easily take into account atypical cases
- eg. rather than comparing a penguin to an avg bird we can rmb that there are some birds which do not fly - It deals more easily with variable categories
- eg. it is diff to imagine what the prototype might be for a cat that contains football, computer frames etc. the exemplar approach requires that we only rmb some of these varying egs.
13
Q
Describe Rosch’s approach to hierarchical organization
A
Hierarchical organization- larger, more general cats are divided into smaller, more specific cats -> creating a no. of level of cats
Superordinate level/ global level:
- the most general or broad
- eg. furniture
Basic level:
- more specific
- eg. table
Subordinate/ specific level:
- even more specific
- eg. kitchen table
Rosch proposed that the basic level is psychologically special cause going above it goes to global -> loss of info while going below it -> little gain of info
14
Q
What are the 2 main approaches to representing relationships among categories?
A
- Semantic network approach
- proposes that concepts are arranged in networks
- also known as Collins and Quillians netwiork
- network consists of nodes that are connected by links
- each node represents a cat or concept
- concepts are placed in the network so that related concepts are connected
- Connectionist approach
- creating computer models for representing cognitive processes
- also called Parallel Distributed Processing (PDP)
- as they propose that concepts are represented by activity that is distributed across a network
15
Q
What are the concepts related to Collins and Quillians Semantic Network Approach?
A
- Hierarchical model
- consists of levels arranged so that more specific concepts
- eg. ‘canary’ and ‘salmon’ are at the bottom
- more general concepts are at higher levels
- eg. at node of canary we obtain info that it is yellow
- to access more info we move up and learn that a canary is a bird
- the more we move up, the more general the links get - Cognitive Economy
- storing shared properties just once at a higher level node
- this makes the network more efficient
- but it has a problem as for eg. not all birds may fly
- to deal with this, they add exceptions at lower nodes
- eg. the node for ostrich would be ‘cant fly’ at the bottom
16
Q
What are the properties of Collins and Quillians Semantic Network approach?
A
Semantic network approach- proposes that concepts are arranged in networks
- The model uses the sentence verification technique
- partis need to answer ‘yes’ or ‘no’ to statements about the concepts
- it should take longer to ans yes to more statements which are further away
- remember the model goes upward, hence if the sequence upwards for robin: canary-> bird-> animal
- it would take longer to say yes to canary-> bird-> animal than canary ->bird
- this is cause canary has to travel 2 nodes to animal rather than 1 node to bird
- Spreading activation leads to further predictions
- spreading activation: is activity that spreads out along any link that is connected to an activated node
- eg. both robin and canary lead to the upward node of bird
- when we think of robin->bird, we might also think of bird-> canary
- HENCE, the result of this activation is that additional concepts may become ‘primed’ and thus more easily retrieved from memory
17
Q
Describe Meyer and Schvaneveldt’s (1971) experiment that spreading activation can influence priming
A
- they used the lexical decision task to study this
- partis’ task to press as quickly as possible ‘yes’ or ‘no’ if there was at least one item in the pair which was a nonword
- the key variable here is the association betw the pairs of real words
- in some trials the words were closely associated (bread and wheat)
- in other trials the words were weakly associated (chair and monkey)
- RESULTS: reaction time was faster when 2 words were associated
- they proposed that this might have occurred as retrieving one word from memory triggered a spread of activation to other nearby locations in a network
- AS more activation would spread to words that were related, the response to the related words was FASTER then to unrelated words
18
Q
What are some criticisms of the Collins and Quillian Semantic Network Approach model?
A
- Cannot explain the typicality effect
- reaction times for statements about an obj are faster for more typical members than for less typical members
- the typicality effect states that ‘a canary is a bird’ should be verified more quickly than ‘an ostrich is a bird’
- BUT this approach predicts equally fast reaction times as ‘canary’ and ‘ostrich’ are both one node away from ‘bird’. - The concept of Cognitive Economy is questionable
- people may store specific properties of concepts ‘has wings for canary’ right at the node for the concept - There is opposing evidence for their theory
- they have found that in some experis where pig -> mammal -> animal
- accord to the model pig -> mammal shd be verified faster
- but in other experis, pig -> animal is verified faster
19
Q
Describe the Connectionist Approach for representing relationships among categories
A
McClelland and Rumelhart’s connectionism:
- creating computer models for representing cognitive processes
- called parallel distributed processing (PDP)
- there are different units
- these units are inspired by neurons found in the brain
- the lines are connections that transfer information between units
- they are roughly equivalent to axons in the brain
- some units can be activated by stimuli from the environment while some can be activated by signals received from other units
Input units -> hidden units -> output units:
- units activated by stimuli from the environment
- they send signals to hidden units
- hidden units then send signals to output units
Connection weight:
- determines how signals are sent from one unit
- can increase or decrease activity of the next unit
- this corresponds to synapses that transmit signals from one neuron to the other
- high connection weight -> strong tendency to excite the next unit
- lower connection weight -> cause less excitation
- neg weight -> decrease excitation or inhibit activation of receiving unit
20
Q
Describe in further detail how exactly the Connectionist Approach works.
A
- Represent an object (eg. canary)
- eg. a canary is a __
- a canary is __
- a canary can __
- a canary has __
- the object is tied to 4 relation statements.
- but in the semantic network approach which utilises hierarchy, those properties are at each node instead - Training a network
- a network has to be trained before it activates the appropriate property units/ pattern of activation
- bef trg: connection weights are all 1
- this is cause they are all the same
- then activity spreads throughout the network
- irrelevant nodes are activated
- for the network to operate properly, the connection weights have to be adjusted so that only appropriate nods are activated
- this adjustment is achieved through “learning process”
- error signals and the representative units provide info about how the connection weights should be adjusted so that correct property units will be activated
learning process: occurs when the erroneous responses in the property units cause an error signal to be sent back, by a process called back propagation
21
Q
Explain the pros of using the Connectionist Approach for representing relationships among categories
A
- The operation of connectionist network is not totally disrupted by damage
- because info is distributed across many units, damage to the system does not completely disrupt its operation
- graceful degradation: disruption of performance occurs only gradually as parts of the system is damaged
- it is similar to what happens when brains are damaged, and only causes a partial loss of functioning to particular areas of the brain - Connectionist networks can explain generalisation of learning
- similar concepts can similar patterns
- training a system to recognise the properties of one concept (canary) also provides info about other related concepts (sparrow or robin)
22
Q
What are the four proposals about how concepts are represented in the brain?
A
- The Sensory Functional hypothesis
- The Multiple Factor approach
- The Semantic Category approach
- The Embodied approach
23
Q
Describe the Sensory Functional approach, about how concepts are represented in the brain.
A
Category specific memory impairment:
- impairment where lost the ability to identify one type of object
- but retained the ability to identify other types of objects
Sensory functional (S-F) hypothesis:
- our ability to differentiate living things and artifacts depends on:
i) a memory system that distinguishes sensory attributes
ii) and a system that distinguishes function
24
Q
Explain an eg. that can illustrate the Sensory Functional Hypothesis.
A
Warrington and Shallice (1984)
- reported 4 patients who suffered memory loss from encephalitis
- they had a category specific memory impairment
- they were able to identify nonanimals (eg. furniture)
- BUT couldnt identify animals
- warrington n shallice noted that distinguishing living things depends on perceiving their sensory features
- eg. distinguishing betw a tiger and leopard depends on perceiving stripes and spots. artifacts in constrast, are more likely to be distinguished by their function
- THIS OBSERVATION that living things are distinguished by sensory properties and artifacts by functions led to the sensory function (S-F) hypothesis
25
What is a criticism of the Sensory-functional hypothesis?
While the hypo explained the behaviour of warrington and shallice's patients plus a dozen of other patients, researchers began describing cases that couldnt be explained by this hypothesis
- eg. Ralph (1998) had a patient who had a sensory deficit where she performed poorly on perceptual tests
- yet she was better at identifying animals than artifacts
- this is opposite of the S-F hypothesis
26
Describe the Multiple Factor approach, about how concepts are represented in the brain.
Multiple factor approach:
- the idea of distributed representation
- led to searching for factors beyond sensory and functional that determined how concepts are divided within a cat
- members of specific cats do share similar perceptual attributed, but we need to use more than just 1 or 2 features when grouping objects in terms of similarity
27
Explain an eg. that can illustrate the Multiple Factor Approach
Hoffman and Ralph (2013)
- used 160 items, asked partis to rate each item on the features
- in terms of color, taste, smell and form etc
- assigned a rating of 7= very strongly to 1= not at all
- RESULTS: indicated that animals were more highly assoc with motion
- color was assoc with artifacts
- artifacts were assoc with performed actions (actions assoc with using an obj)
- the results conform to the s-f hypo, BUT when looked closer...
- mechanical devices (machines and vehicles) overlapped with artifacts and animals
- music instruments and some mechanical devices occupy a middle ground betw artifacts and living things as they both involve ACTION and SENSORY ATTRIBUTES
- another factor that has been proposed to differentiate animals and artifacts is crowing- animals tend to share many properties (eyes, legs...).
- in contrast artifacts share fewer properties than that they are both vehicles
- THIS HAS LED TO researchers to propose that patients who appear to have category specific impairments (like difficulty to recognize living things but not artifacts) TO DONT REALLY HAVE A CAT SPECIFIC IMPAIRMENT AFTER ALL LOL
28
Describe the Semantic category approach, about how concepts are represented in the brain.
- proposes that there are specific neural circuits for some specific cats
- Majon and Caramazza (2011): there are limited no. of cats that are innately determined because of their importance for survival
- this is based on the idea that identified areas of the brain respond to specific types of stimuli such as faces, places and bodies
29
Explain an eg. that can illustrate the Semantic Category Approach
Huth et al. (2016)
- had partis view filmm
- listen to more than 2h of stories from moth radio hour braodcast while in brain scanner
- the light area in the back of the brain was activated by words assoc with violence
- results show that the maps were similar for each of the partis
while the semantic cat approach focuses on areas of the brain that are specialized to respond to specific types of stimuli,
- it also emphasizes that the brain's response to items from a particular cat is distributed over a no. of different cortical areas
30
Describe the Embodied approach, about how concepts are represented in the brain.
- states that our knowledge of concepts is based on reactivation of sensory and motor processes that occur when we interact with the object
- eg. when people use a hammer, sensory areas are activated in response to the hammer's size, shape and color
- in addition, motor areas are activated that are involved in carrying out actions involved in using a hammer
31
Explain an eg. that can illustrate the Embodied approach
The link between perception (neuron fires when watching experi pick up food) and motor responses (the same neuron fires when monkey picks up food) is important in the EA.
- is about how the thinking of concepts cause activation of perceptual and motor areas assoc with these concepts
Hauk et al. (2004)
- measured brain acitivity in fmri when
a. parti move their right or left foot/ left or right index finger/ tongue
b. partis read action words like kick
- RESULTS: show areas of cortex activated by actual movements and by reading the action words
- activation is more extensive for actual movements, but
- activation caused by reading words also occur approx in same areas of the brain
- this correspondence betw words related to specific parts of body and location of brain = semantic somatopy
- HENCE, they concluded that the ability to represent motor activity assoc with actions is not necessary for recog objs, as the EA would predict
32
What is Semantic dementia?
- a general loss of knowledge for all concepts
- patients with SD tend to be equally deficient in identifying living things and artifacts
the general experienced deficits also come along with the finding that the anterior temporal lobe (ATL) is generally damaged in these patients. this led to the hub and spoke model of semantic knowledge
33
Explain the Hub and Spoke model of semantic knowledge.
- Spokes: areas of the brain that are assoc with specific functions, which are are connected to ATL
- Hub: ATL serves as a hub that integrates the info from these areas
- these functions include valence, praxis, functionality and visual
- valence= weak vs strong
- praxis= involving manipulation
- these diff functions are related to different colors that light up in transcranial magnetic stimulation (TMS)
- evi supporting the idea of A HUB is that damage to one of the specialized brain areas (the spokes) can cause specific deficits (such as inability to identify artifacts).
- but damage to ATL (the hub) causes general deficits such in semantic dementia
- these differences in hub and spoke functions has been demonstrated in non-brain damaged patients using TMS
34
Explain an eg. that demonstrates or describes the Hub and Spoke model.
Pobric et al. (2010)
- presented images of living things and artifacts to partis
- measured their response time for naming each pic
- repeated this while TMS was applied to either ATL or an area in parietal lobe that was normally activated when person is manipulating obj
- RESULTS: parietal inactivation slowed reaction times for manmade objs but not for living things
- whereas ATL inactivation caused the same effect for both manmade and living objects
- Parietal activation caused a large slowing effect in reaction time for highly manipulatable objs (tools) but not manipulatable objs (furniture)
- whereas ATL inactivation affected both types of objects equally
- this is a result of the general effect of stimulating the hub (ATL)
- but a more specific effect of stimulating an area that would be assoc with one of the spokes (parietal cortex) would support the idea that a hub has general functions, and spokes have more specific functions
