Upload
stephen-houser
View
243
Download
0
Tags:
Embed Size (px)
Citation preview
General KnowledgeGeneral Knowledge Structure of Semantic MemoryStructure of Semantic Memory
– BackgroundBackground
– Feature Comparison ModelFeature Comparison Model
– Prototype ApproachPrototype Approach
– Exemplar ApproachExemplar Approach
– Network ModelsNetwork Models
Schemas & ScriptsSchemas & Scripts– BackgroundBackground
– Recall of ScriptsRecall of Scripts
– Schemas & Memory SelectionSchemas & Memory Selection
– Schemas & Boundary ExtensionSchemas & Boundary Extension
– Schemas & Memory AbstractionSchemas & Memory Abstraction
– Schemas & Memory InferencesSchemas & Memory Inferences
– Schemas & Integration in MemorySchemas & Integration in Memory
– ConclusionsConclusions
Semantic MemorySemantic Memory General Conceptual KnowledgeGeneral Conceptual Knowledge
Lexical Knowledge (e.g., “apple” and )Lexical Knowledge (e.g., “apple” and )
Organized - (e.g., ‘pencil’ related to ‘pen’; think Organized - (e.g., ‘pencil’ related to ‘pen’; think
of ‘apple’ ----> ‘banana’of ‘apple’ ----> ‘banana’
Categories and ConceptsCategories and ConceptsCategoryCategory - a class of objects that belong together (e.g., - a class of objects that belong together (e.g.,
variety of objects: ‘fruits’ or ‘apple’)variety of objects: ‘fruits’ or ‘apple’)
ConceptConcept - mental representation of a category - mental representation of a category
Concepts allow us to make inferences when we Concepts allow us to make inferences when we encounter new instances (e.g. read ‘chair)encounter new instances (e.g. read ‘chair)
NaturalNatural concepts vs. concepts vs. ArtifactsArtifacts
QuestionsQuestions– Organization and Structure?Organization and Structure?– Storage?Storage?– Inferences?Inferences?– Cognitive Economy?Cognitive Economy?– Relatedness and Similarity?Relatedness and Similarity?
apple tabledog pentree chair
Feature Comparison ModelsFeature Comparison Models Concepts = list of features or attributes (e.g., Smith, Concepts = list of features or attributes (e.g., Smith,
Shoben, and Rips 1974)Shoben, and Rips 1974) DefiningDefining vs. vs. CharacteristicCharacteristic Features Features Decision Process - 2 StagesDecision Process - 2 Stages
– Stage 1Stage 1 = global comparison = global comparison
– Stage 2 = compare defining featuresStage 2 = compare defining features Research Research
– Typicality EffectTypicality Effect
– Category Size Effect (faster RTs for membership in Category Size Effect (faster RTs for membership in small category) NOT explainedsmall category) NOT explained
Other ProblemsOther Problems
The Sentence Verification TechniqueThe Sentence Verification Technique
For each of the items below, answer as For each of the items below, answer as quickly quickly as possibleas possible either true or false. either true or false.
1.1. A poodle is a dog.A poodle is a dog.2.2. A squirrel is an animal.A squirrel is an animal.3.3. A flower is a rock.A flower is a rock.4.4. A carrot is a vegetable.A carrot is a vegetable.5.5. A mango is a fruit.A mango is a fruit.6.6. A petunia is a tree.A petunia is a tree.7.7. A robin is a bird.A robin is a bird.8.8. A rutabaga is a vegetable.A rutabaga is a vegetable.
Prototype ApproachPrototype Approach
Classical View vs. ProtoypeClassical View vs. Protoype
RoschRosch
Idealized version of category (example)Idealized version of category (example)
Graded membership - not all memebersGraded membership - not all memebers
Bachelor = Unmarried, maleBachelor = Unmarried, male
But which of the following are really bachelors?But which of the following are really bachelors?
1.1. My 32-year old cousin, John, who works at a My 32-year old cousin, John, who works at a
bank in Chicagobank in Chicago
2.2. My 6 month old son TimMy 6 month old son Tim
3.3. An elderly Catholic PriestAn elderly Catholic Priest
Characteristics of PrototypesCharacteristics of Prototypes
1.1. Prototypes are supplied as examples of a category.Prototypes are supplied as examples of a category.
2.2. Prototypes serve as reference points.Prototypes serve as reference points.
3.3. Prototypes are judged more quickly after priming.Prototypes are judged more quickly after priming.
4.4. Prototypes can substitute for a category name in a Prototypes can substitute for a category name in a sentence.sentence.
5.5. Prototypes share common attributes in a family Prototypes share common attributes in a family
resemblance categoryresemblance category..• No one attribute shared by all membersNo one attribute shared by all members
• In / out phenomenonIn / out phenomenon
Mervis, Catlin, & Rosch (1976)Mervis, Catlin, & Rosch (1976)
Group 1Group 1: generated examples for 8 different categories: generated examples for 8 different categories– Birds? … robin, sparrow …Birds? … robin, sparrow …– Fruits?Fruits?– Sports?Sports?
– EtcEtc.. Group 2Group 2: provided prototype : provided prototype ratingsratings (low to high) for each (low to high) for each
exampleexamplee.g.,e.g., sparrow 7 - highsparrow 7 - high
penguin 2 - lowpenguin 2 - low
Strong correlation between frequency and ratingStrong correlation between frequency and rating Typicality EffectTypicality Effect
Lexical Decision TaskLexical Decision Task
apple
table
tadjld
mountain
pudor
The following items: Decide whether each item is a word (‘yes’) or not a word (‘no’). Respond by pressing the ‘yes’ button or the ‘no’ button:
What Is a What Is a PrimingPriming Effect? Effect?
doctor
Y/N
450 ms
Lexical Decision Task
doctor
Y/N
400 ms
hospital
doctor
Y/N
450 ms
automobile
Robin Penguin
Robin480 ms.
Penguin660 ms.
Bird Bird
550 ms. 670 ms.
Group 1:Group 1: Prototype Ratings Prototype Ratingse.g., vehicles: car, truck, tractor, slede.g., vehicles: car, truck, tractor, sled
vegetable: carrots, beets, eggplantvegetable: carrots, beets, eggplant
clothing: shirt, sweater, vestclothing: shirt, sweater, vest
Group 2:Group 2: List attributes possessed by each List attributes possessed by each item: item:
e.g., car: wheels, steering wheel, doors, e.g., car: wheels, steering wheel, doors, etc. etc.
Score:Score: What proportion of an item’s attributes were What proportion of an item’s attributes were shared by other category member’sshared by other category member’s
Strong correlation between score and prototype Strong correlation between score and prototype rating.rating.
Prototype Ratings for Words in Prototype Ratings for Words in Three CategoriesThree Categories
Levels of Categorization 1Levels of Categorization 1
Superordinate LevelSuperordinate Level
– furniture, animals, toolsfurniture, animals, tools
Basic LevelBasic Level
– chair, cat, screwdriverchair, cat, screwdriver
Subordinate LevelSubordinate Level
– desk chair, persian cat, phillips screwdriverdesk chair, persian cat, phillips screwdriver
Levels of Categorization 2Levels of Categorization 2
Superordinate levelSuperordinate levelBasic-levelBasic-levelSubordinate levelSubordinate level
1.1. Basic-level names are used to identify objectsBasic-level names are used to identify objects
2.2. Members of basic-level categories have more Members of basic-level categories have more
attributes in commonattributes in common
3.3. Basic-level names produce the priming effectBasic-level names produce the priming effect
4.4. Experts use subordinate categories differentlyExperts use subordinate categories differently
Carrot Vegetable
Same / Different Same / Different
Priming Effect No Priming Effect
Expert Novice
Superordinate
Basic 10 + 11 +
Subordinate 10 + 6 +
Exemplar ApproachExemplar Approach Store specific Store specific instancesinstances or examples ( or examples (exemplars)exemplars) Decision process = comparison of new item to stored Decision process = comparison of new item to stored exemplars.exemplars. Comparison to prototype approach = Comparison to prototype approach =
Stored Representation
Prototype Approach Exemplar
Typical or idealizedrepresentation
Stored representation = Specific members / instances
•Absence of features -- (characteristic vs. defining)
Picture of DogPicture of Dog
Exemplar ApproachExemplar Approach
No abstraction - no summary representation.No abstraction - no summary representation.
Storage requirements.Storage requirements.
May be more suitable for smaller categories.May be more suitable for smaller categories.
Evidence from Social Psychology - stereotypesEvidence from Social Psychology - stereotypes
Individual differencesIndividual differences
Co-existence: prototypes and exemplarsCo-existence: prototypes and exemplars
Strategic differencesStrategic differences
Explaining Explaining concept learningconcept learning! !
Network ModelsNetwork Models Semantic networksSemantic networks
– (concepts and connections ----> nodes and (concepts and connections ----> nodes and links)links)
Collins & LoftusCollins & Loftus– Node = conceptNode = concept
Link = relation or connectionLink = relation or connection– Spreading activationSpreading activation
Sentence verification ----> intersectionsSentence verification ----> intersections Explaining ‘Typicality Effect’Explaining ‘Typicality Effect’ Anderson’s ACT* TheoryAnderson’s ACT* Theory
Activation SpreadActivation Spread
Does a robin breathe?
AndersonAnderson
ACT = Adaptive Control of ThoughtACT = Adaptive Control of Thought
Declarative vs. Procedural KnowledgeDeclarative vs. Procedural Knowledge
Propositional NetworksPropositional Networks
PropositionProposition - the smallest unit of knowledge with - the smallest unit of knowledge with
a truth valuea truth value
Proposition = node + linkProposition = node + link
Working Memory - active part of Long Term Working Memory - active part of Long Term
MemoryMemory
Susan gave a white cat to Maria, who is the president of the club.
1. Susan gave a cat to Maria.
2. The cat was white.
3. Maria is the president of the club.
Susan gave a white cat to…Susan gave a white cat to…
SchemasSchemas
1.1. Larger cognitive unitsLarger cognitive units
2.2. Packages of interrelated unitsPackages of interrelated units
3.3. Used to interpret, encode, understand, and Used to interpret, encode, understand, and remember new instancesremember new instances
4.4. Provide Provide expectationsexpectations about what should occur about what should occur (top - down)(top - down)
5.5. Default values / parts - filled in when schema Default values / parts - filled in when schema activatedactivated
6.6. Sometimes - errorsSometimes - errors
““When Lisa was on her way back from When Lisa was on her way back from the store with the balloon, she fell and the store with the balloon, she fell and the balloon floated away.”the balloon floated away.”
ScriptsScripts
Simple, well- structured sequence of events Simple, well- structured sequence of events associated with a highly familiar activityassociated with a highly familiar activity
Schema vs. scriptSchema vs. script Recall of scriptsRecall of scripts
– Different from conceptual categories (Barsalow Different from conceptual categories (Barsalow & Sewell, 1985)& Sewell, 1985)
– Script Identification - early vs. late (Trafimow Script Identification - early vs. late (Trafimow and Wyer, 1993)and Wyer, 1993)
Appreciating the similarity of scriptsAppreciating the similarity of scripts
Trafimow & Wyer (1993)Trafimow & Wyer (1993) 4 different scripts4 different scripts
– Photocopying a piece of paper Photocopying a piece of paper
– Cashing a checkCashing a check
– Making teaMaking tea
– Taking the subwayTaking the subway
Irrelevant details added (e.g., taking candy out of pocket)Irrelevant details added (e.g., taking candy out of pocket) Script - identification information presented first or lastScript - identification information presented first or last FillerFiller Recall: of script - related eventsRecall: of script - related events
23%23% vs.vs. 10%10%
(script identified first)(script identified first) (script identified last)(script identified last)
Schemas and Memory SelectionSchemas and Memory Selection
Remember best info Remember best info consistentconsistent with schema with schema
or or inconsistentinconsistent
Brewer & Treyons (1981)Brewer & Treyons (1981)
Rojahn & Pettigrew (1992)Rojahn & Pettigrew (1992)
Incidental vs. Intentional learningIncidental vs. Intentional learning
Schemas and Boundary ExtensionSchemas and Boundary Extension
Schemas and Memory AbstractionSchemas and Memory Abstraction
Abstraction Abstraction Verbatim vs. GistVerbatim vs. Gist Constructive ApproachConstructive Approach
– Bransford & Franks (1971)Bransford & Franks (1971)– Holmes & Colleagues (1998)Holmes & Colleagues (1998)
Pragmatic ApproachPragmatic Approach– Murphy & Shapiro (1994)Murphy & Shapiro (1994)– Attention Allocation / ControlAttention Allocation / Control
C & P compatibleC & P compatible
Schemas and Inferences in MemorySchemas and Inferences in Memory
Bartlett (1932)Bartlett (1932)
Ebb vs. BartlettEbb vs. Bartlett
Interaction of prior knowledge and experience and Interaction of prior knowledge and experience and
formation of new memoriesformation of new memories
““War of the Ghosts” storyWar of the Ghosts” story
Initial vs. Delayed RecallInitial vs. Delayed Recall
Bransford, et al (1972)Bransford, et al (1972)
Implications - e.g., advertising Implications - e.g., advertising
Schemas and Integration in MemorySchemas and Integration in Memory
Final process in memory formationFinal process in memory formation
Result of Result of selection, abstraction,selection, abstraction, and and inferenceinference
Important!!Important!!
Integration and Delayed RecallIntegration and Delayed Recall
Integration and Limited Memory CapacityIntegration and Limited Memory Capacity