Ch. 9 Quiz

What are the 4 ways of explaining how concepts are represented in the brain?

1. Sensory-Functional Hypothesis Good starting point, but not complex enough 2. Semantic Category Approach Emphasizes specialized areas of brain and networks connections 3. Multiple-Factor Approach Emphasizes role of many different features and properties 4. Embodied Approach Emphasizes activity caused by sensory & motor properties of objects

If you say that "a Labrador retriever is my idea of a typical dog," you would be using the _______ approach to categorization. A. exemplar B. definitional C. family resemblance D. prototype

A

According to the S-F hypothesis, our ability to differentiate living things and artifacts depends on a semantic memory system that distinguishes ________ and one that distinguishes ________. A. serial nodes; familiar concepts B. sensory attributes; function C. sequential networks; familial resemblance D. sensations; facts

B

According to the typicality effect, A. objects in a category have a family resemblance to one another. B. items that are high in prototypicality are judged more rapidly as being in a group. C. objects that are not typical stand out and so are more easily remembered. D. we remember typical objects better than non-typical objects.

B

According to Rosch, the ______ level of categories is the psychologically "privileged" level of category that reflects people's everyday experience. A. superordinate B. prototypical C. basic D. subordinate

C

The definitional approach to categorization A. sets definite criteria called family resemblances that all category members must have. B. is not well suited for geometrical objects but works for familiar everyday objects. C. doesn't work well for most natural objects like birds, trees, and plants. D. was proposed to replace the prototype approach.

C

One of the key properties of the _____ approach is that a specific concept is represented by activity that is distributed over many units in the network. A. semantic network B. hierarchical C. spreading activation D. connectionist

D

The _______ model includes associations between concepts and the property of spreading activation. A. parallel distributed processing B. connectionist network C. neural network D. semantic network

D

Which approach to categorization involves forming a standard representation based on an average of category members that a person has encountered in the past? A. Exemplar B. Network C. Typicality D. Prototype

D

What is the evidence for the Collins and Quillian model?

FOR cognitive economy says having common features stored at higher lvl nodes is more efficient. sentence verification technique supports this bc it takes longer to verify sentence that requires traveling more nodes/lvls spreading activation says nodes connected to activated concept will also be activated (primed) and more easily accessed from mem. lexical decision task supports this bc part. had faster RT to decide if pair of grps of letters were words if pairs were related

What arguments did Rosch present to support the idea that one of these levels is "privileged"? How has research on categorization by experts led to modifications of Rosch's ideas about which category is "basic" or "privileged"?

Rosch had people list features of global, basic, and specific categories. they listed 3, 9.3, and 10.3 features respectively. conclusion: basic = privileged bc it's a big loss of info from global to basic but small gain from basic to specific privileged category isn't same for everyone. experts prefer specific category (exp. w/ bird experts vs. non experts identifying pics of objects. experts used specific categories but non experts used basic)

What is a connectionist network? Describe how a connectionist network learns, considering specifically how connection weights are adjusted. Also consider how the way information is represented in a connectionist network differs from the way it is represented in a semantic network.

a network of input, hidden, and output units that transfer info b/w each other. a specific category/concept is repped by activity that is distributed over many units in the network (parallel distributed processing/PDP) training: uses back propagation, a learning process in which erroneous responses in property units send error signals BACK through network to tell hidden/representation units to adjust connection weights (all weights start at 0) diff than semantic network: activity for a concept is distr. throughout network; uses back propagation to properly adjust weights so correct representation occurs

What is the evidence against the Collins and Quillian model?

can't explain some typicality effects: high-typ. objects verified faster than low-typ. even when both are same # of links away (i.e. ostrich/canary is a bird--canary verified faster but both 1 node away from bird) can't explain some examples of cog. economy: faster RT to verify "A pig is an animal" than "A pig is a mammal" even tho mammal requires traveling less links-->ppl may store specific properties at specific node, not at higher node

What is the prototype approach? What experiments (2) did Rosch do that demonstrated connections between prototypicality and behavior? What other experiments were done that demonstrated this (2)?

compare objects to a prototype that represents the category (typical/avg. rep of all category members) 1. Rosch had part. rate objects ("sparrow" "penguin") from 1-7 on how well they fit a category ("birds"). participants rated highly prototypical objects higher 2. sentence verification technique: faster RT for sentences w/ high-prototypicality objects (i.e. an apple vs. pomegranate is a fruit) 3. prototypical items are named first when listing objects in a category 4. Rosch: highly prototypical objects more affected by priming. part. heard a color (prime) then saw 2 colors quickly then had to decide if colors were same/diff. faster "same" judgement for prototypical colors (i.e. good example of that color). shows you imagine prototypical color when you hear the prime

What is the basic idea behind the semantic network approach? What is the goal of this approach, and how did the network created by Collins and Quillian accomplish this goal?

concepts are arranged in networks goal: dev. a computer model of human memory made a hierarchical model for how concepts/properties are assoc. in mind (nodes for categories/concepts connected by links)

How did Collins and Loftus deal with the criticisms of the semantic network w/ their modified model? What are its advantages and disadvantages?

created non-hierarchical model based on an individual's XP where shorter links connect closely related concepts and longer links connect less closely related concepts. adv: explains familiarity, typicality, and priming effect disadv: cannot make predictions until an individual's network is mapped out

What is a solution to the problem with the definitional approach that not all members have the same defining features? How does this relate to the prototype approach?

family resemblance: idea items in a category resemble each other in many ways, doesn't have to fit a strict definition (allows for some variation) prototypical objects (closely resemble prototype) have high family resemblance. aka these items have large overlap w/ characteristics of other items in the category

What does it mean to say that there are different levels within a category?

hierarchical organization: we have broader categories higher up, then get more specific as you go down

Why is the use of categories so important for our day-to-day functioning?

it is how we organize concepts. categories help us understand individual cases not previously encountered ("pointers to knowledge"). they give us lots of general info about an item, allow us to identify special characteristics of an item, and understand behavior we find baffling

What are the advantages of the connectionist approach?

operation of connectionist networks NOT totally disrupted by damage (bc info is distr. across units). graceful degradation says disruption of performance occurs only gradually w/ damage but operation isn't totally disrupted connectionist networks can explain generalization of learning bc as network is trained to recognize properties of a concept (i.e. canary), it also provides info abt related concepts (robin, sparrow)

When is the prototype approach best to use? What about the exemplar approach?

prototype: larger categories, i.e. furniture exemplar: smaller categories, i.e. US presidents

What is the exemplar approach to categorization? How does it differ from the prototype approach, and how might the two approaches work together?

put items in category if they are exemplars (examples of that cat.) diff. than prototype approach bc exemplars are actual cat. members you have XPed, not an average member. this way, it can take into acc. atypical members and deal w/ variable categories (i.e. games) you can start w/ prototype approach then exemplars play a larger role (better at taking exceptions into acc.)

Describe the definitional approach to categories. Why does it initially seem like a good way of thinking about categories, but then become troublesome when we consider the kinds of objects that can make up a category?

putting concepts in categories based on if that concept fits definition of category works well for some things but doesn't work well for all things bc there are many subtle differences and not all cat. members have same defining features (i.e. many types of chairs that don't fit strict definition of 4 legs, a seat, and a back)

Describe the sensory functional (S-F) approach to explaining how concepts are represented in the brain. Indicate the basic idea behind this hypothesis and the evidence for and against it.

says categories are broken down by how you think about them. we need systems that distinguish sensory features to ID living things (i.e. stripes vs. spots) and functional features to ID artifacts (i.e. purpose of tools) FOR: patients KC/EW who don't recognize specific categories (i.e. living animals) but do recognize others (artifacts)--category specific impairment AGAINST: patient who couldn't ID living things and had impaired sensory (expected) AND functional abilities (unexpected); patient who couldn't recognize artifacts but had impaired sensory ability (we'd expect opposite)

Describe the multiple factor approach to explaining how concepts are represented in the brain. Indicate the basic idea behind this hypothesis and the evidence for and against it.

says that multiple factors determine how concepts are divided up w/i a category (idea of distr. rep!) FOR: part. rated items on certain features. animals had more assoc. w/ features of color and motion. artifacts had more assoc. w/ performed action. mechanical devices (i.e. vehicles, musical instruments, etc.) overlapped for both-->widely distr. semantic rep. explains the patients w/ category specific mem. impairments (S-F hyp.) by saying these patients have trouble distinguishing b/w items that share similar features (i.e. animals share many features but artifacts don't--this is crowding)

Describe the embodied approach to explaining how concepts are represented in the brain. Indicate the basic idea behind this hypothesis and the evidence for and against it.

says that our k of concepts is based on reactivation of sensory and motor processes we use when we interact w/ that object FOR: -mirror neurons-activated when we do action and see others do it -hauk exp where part read action words (i.e. kick) and did actions (i.e. moved foot) and had similar brain activation for both (semantic somatotopy) AGAINST -patient AA couldn't do actions assoc. w/ objects but could ID pics of these objects (motor ability NOT necessary to ID objects) -bad at explaining k of abstract concepts

Describe the semantic category approach to explaining how concepts are represented in the brain. Indicate the basic idea behind the hypothesis and the evidence for and against it.

says there are specific neural circuits in brain for some specific categories (based on innate properties of survival importance). also says recog. depends on info distr. across many brain areas (i.e. when recognizing a face, areas for face recog., emotion, gaze, etc. activated) FOR: identical twins more highly correlated in facial recog. ability than fraternal-->face recog. mechanisms have genetic basis