Cognitive Psyc
Terms
undefined, object
copy deck
- Alternative Names for Short-term memory
-
-Short term store
-Immediate memory
-Primary memory - Three Things Associated with Free Recall
-
1. Serial-position curve
2. Primacy effect
3. Recency effect - Serial-Position Curve Trends
-
-Contains Primacy and Recency effect
-Rehearse each early item until list becomes too long then shift to rehearse new items
-Usually, recent (last) items recalled first, then beginning and middle items - How is a serial-position curve evidence for 2 seperate memory systems?
-
-STM dumps last items out (recency effect)
-LTM for rehearsed items in first part of list - Primacy Effect
-
-Better memory for first few items on serial-position curve
-beginning items are rehearsed so higher probability of being encoded - Single Dissociation
-
-One cognitive system is dependent upon the other, but the other is not
-Damage to one, see if other works
-Example: (Radio's dependent upon car electrical system, but electrical system isn't dependent upon radio) - Example of Double Dissociation
-
-Selectively damage one and leave other intact
-Example: damage radio but windsheld wipers still work - What evidence is there for double dissociation with STM and LTM?
-
-Serial recall
-List-length effects
-Distractor Tasks
-Presentation rate (item spacing)
-STM as rehearsal buffer - Tests for Double Dissociation
-
-Have a variable that affects STM but not LTM
-Bring in a 3rd variable that affects LTM but not STM
-Used to show STM and LTM are seperate - What does serial recall affect?
- -Affects STM but not LTM
- List-Length Effects
-
-Always have recency effect
-Affects LTM but not STM
-Affects the probability of middle items in list going to LTM - Distractor Task
-
-Erase STM in a way so there is no recency effect
-Affects STM but not LTM
-Primacy portion is unaffected - Presentation Rate (item-spacing effect)
-
-Difference in items from middle of list
-Higher probability of moving middle list items from STM to LTM
-Affects LTM not STM
-Affect primacy portion but not recency portion of curve - STM as Rehearsal Buffer
-
-Limited capacity
-Keep info in active state by rehearsal:
Maintenance Rehearsal: STM->STM
Transfer: STM->LTM - Miller: "Magic Number"
-
-7 +/- 2 recall of items (average capacity)
-Chunking
-Recoding - Chunking
-
-Grouping items together into meaningful chunks
-7 +/- 2 is number of chunks that can be held in STM
-Enables breaking of informational "bottleneck" in STM - Brown, Peterson, & Peterson
-
-STM duration of non-rehearsed info
-Infamous Brown-Peterson task - Brown-Peterson Task
-
-Stimulus: simple three-letter triplet
-Distractor task: count backwards by three from number they'd been shown (variable length of counting)
-Task: report back the three-letter stimulus - Brown, Peterson, & Peterson: Results
-
-STM duration is short
-Increasing period of time, less and less info remained in STM
-Decay in STM (passive process)
-Distractor task prevents rehearsal of stimulus - Waugh and Norman
-
-Decay vs Interference
-Believed distractor task might be source of interference
-Longer counting intervals would provide more opportunity for interference - Waugh and Norman: Method
-
-List of 16 digits
-Read at rate of either 1 or 4 digits per second
-Final item in list was a repeat and served as probe or cue
-Task: write digit that followed the first presentation of repeated item - What was the important part of Waugh and Norman's experiment for the issue of decay vs interference?
-
-Time it took to present 16 digits btw groups
-Presenting entire list took 16 s for one group but only 4 s for the other group - Waugh and Norman: Results
-
-No difference in recall accuracy bwt 2 groups
-Suggesting forgetting was influenced by number of intervening items bwt critical digit and recall test - What results would Waugh and Norman have seen if decay theory was right?
-
-Worse performance when list presentation took 16 s
-More time intervenes during 16 s presentation so have more passive decay - What were the criticisms of Waugh and Norman's experiment?
-
-Failure to reject null hypothesis
-No difference bwt 2 conditions (rate doesn't matter)
-Poorly conducted experiment which is also support for the null hypothesis - Talland
-
-Brown-Peterson task with two different distractors
-One condition: count backward by 3's
-Second condition: read #'s from screen - Talland: Results & Conclusions
-
-Support for alternative hypothesis
-Group that used mental effort (counting backward by 3's) had worse recall
-Mental effort counting is causing interference - Peterson, Peterson, & Miller
-
-Brown-Peterson task with different kinds of stimulus material
-Stimulus: either nonsense syllables e.g.(MEQ) or meaningful syllables e.g.(NFL)
-NFL is single chunk whereas MEQ is 3 seperate chunks - Peterson, Peterson, & Miller: Results
-
-Higher recall for meaningful syllable condition vs. nonsense syllable condition after 6 s of backward counting
-Disagrees with decay theory
-Forgetting in STM is thus due to interference - Wickens
-
-Noticed that with subsequent trials on Brown-Peterson task, subjects did worse
-4 trials, each trial had 3 words (stimulus) - Wickens: Method
-
-2 conditions: control and experimental
-Control group: words in all trials were similiar (similiar stimuli)
-Experimental group: words in trial 4 were different (change stimuli in trial 4) - Wickens: Results
-
-Both groups did same on first 3 trials
-Build up of PI w/ first 3 trials
-Experimental group had 90% accuracy on trial 4
-Release from PI - How do we retrieve (locate) information from STM?
-
-Recognition: stimulus in front of you
and
-Recall: generate stimulus from LTM - Subtractive Method
-
-Measure RT1 for task involving mental processes A, B, and C
-Measure RT2 for comparision task only involving processes A and C
-Find duration of process B by subtracting RT2 from RT1 - Problems with subtractive method
-
-Difficult to find 2 task that differ in only one mental process
-Assumption of "pure assertion" that by subtracting out B, processes A and C are affected (all have to be independent) - Additive Factors Method
-
-Repeat mental process for some number of times during a single trial
-Still do same processes, but one process is done more than once
-Determine how long it takes to do that process by determining how much time was added to people's RT - Additive Factors Method Equation
-
B = RT2 - RT1 / n-1
B: process repeated
n: number of times process was repeated - Sternberg
-
-STM scanning task
-Get memory set (1-6 items)
-Single letter probe
-Respond "yes" or "no" depending upon whether probe item was in memory set
-Each trial memory set and probe varied - Why did Sternberg manipulate size of memory set?
-
-He wanted to influence the # of cycles through the search or scan process
-And by examing the slope of RT's, he could determine the additional time needed for each cycle in that process - 3 Possibilities for Sternberg's Results
-
-Serial self-terminating search
-Parallel search
-Serial exhaustive search - Serial Self-Terminating Search
-
-Search memory set items one-by-one and when find match is found
(how we look for car keys) - Parallel Search
-
-Scan each position in memory set simultaneously
-Equally fast with all set sizes (size doesn't matter)
-No increase in RT - Serial Exhaustive Search
-
-Compare probe to each item in memory set, even if find match still continue(exhaustive)
-Comparison is one at a time (serial) - Sternberg's Results
-
-Linear increases in RT's as set size increases
-Increase was nearly same for both "yes" and "no"
-Serial exhaustive search
-Scan contents of STM at a rate of about 38ms per item - Early Research on kinds of info held in STM
-
-Info was though to be based on verbal/acoustic cues or codes
-Conrad: verbal codes - Conrad
-
-Subject's studied list of words, recall in any order
-Specific errors based on acoustic errors ("E" say "G", "T" say "C")
-Even though letters were presented visually, they were stored in STM in an acoustic, sound-based fashion - Semantic Code
-
-Wickens found that release from PI due to semantic interference
-Info in STM is based on semantic code
-Evidence was switching categores resulted in release from PI - Shepard and Metzler
-
-Visual Codes
-See complex pattern in two forms
-Task: determine if same pattern just rotated or if it's rotated with mirror reversal (enantomers or diastereomers) :) - Shepard and Metzler Findings
-
-RT increased as angular rotation for second drawing increased
-Form a visual image that can rotate in mind, but further rotation takes more time - What in the world did Shand do?
-
-Congentially deaf subjects fluent ASL
-2 conditions
-Five item lists for serial recall - And what in the hell did Shand find...?
-
-More errors in 2nd condition than 1st condition
-Have kinesthetic code (STM can hold info in any format that can be sensed) - 2 Conditions in Shand's Experiment
-
1. sequence of ASL signs that were phonologically similiar in English
2. ASL signs that were cherologically related (similiar hand movements to form sign) - What were some theoretical problems w/ STM?
-
-Limited capacity (magic number)
-Largely based on verbal/acoustic codes
-Both wrong SO people began to change conception of STM - Brooks
-
-Burden/load on STM and switching bwt processing and remembering aren't addressed by simplie approached that emphasize 7 +/- 2 "slots"
-We instead have working memory - Brooks: Method
-
-See block letter; take away
-Primary Task: mentally draw letter and point thumb either up or down when come to certain corners
-Then combined with secondary tasks: counting and then laser pointer - Brooks: Results
-
-Counting (verbal task) while scanning visual image yield few errors
-Laser pointer task while scanning visual image was more difficult - Explaination for Brook's results
-
-Tracing letter: visuo-spatial
-Counting: phonological resource (doesn't overlap with visuo-spatial so can do simultaneously)
-Laser pointer: visuo-spatial (same as tracing letter so both consume resources from same visuo-spatial resource) - Baddeley's Model of Working Memory
-
-Instead of single buffer w/ limited capacity and verbal/acoustic codes, have working memory with 3 systems
-Central executive and slave systems - Slave Systems
-
-Articulatory (phonological) loop
-Visuo-spatial sketchpad
-Proceed w/o disrupting activities elsewhere in working memory if task isn't that demanding
-If task is demanding, drains resources from central executive - Characterisitcs of Slave Systems
-
-Responsible for low-level processing
-Domain specific
-Have own pool of attentional resources (pools are very limited)
-Don't overlap - Working Memory: Paired Central Executive task with secondary task that tapped one of the slave systems
- When slave system drains off attentional resources from central executive, the executive can no longer maintain its speed or accuracy
- Language comprehension in a dual-task setting
-
-Hold 6 digits in memory while doing comphrehension task
-Both comprehension scores and memory span performance disrupted -
Physiological evidence of areas of heightened activity in brain during verbal working memory tasks
(i.e Sternberg's task) -
Left hemisphere regions:
-Broca's area
-3 frontal sites
-Left supplementary motor area (SMA)
-Premotor areas -
Physiological evidence of task emphasizing executive control
(i.e. switching from one task to another) -
Heightened activity in:
-Brodmann's area 46
-Dorsolateral pre-frontal cortex - Physiological evidence of regions of heightened brain activity durning visual and spatial working memory task
-
Right hemisphere regions:
-Extrastriate occipital cortex
-Posterior parietal lobe
-Premotor area
-Dorsolateral prefrontal cortex (DLPFC) - Why is LTM important?
-
-Stores everthing we know
-Fundamental to nearly every mental process and almost every act of cognition - Squire: Taxomony of LTM
- -Distinction between declarative (explicit memory) and nondeclarative (implicit)
- Declarative Memory
-
-Explicit and with awareness
-Talk about these memories
-Can be episodic or semantic memories - Epsiodic Memory
-
-Autiobiographical
-Personal experiences
-Context specific - Semantic Memory
-
-Conceptual
-Encyclopedic
-Context free - Procedural Memory
-
-Implicit and w/o awareness
-How to do things
-Primings, conditioning, and other types of (non-associative) learning i.e. task-aversion learning - What is the debate about declarative memory and its sub-units?
- Debate about if episodic and semantic are distinct and seperate or a continum where some memories are more episodic or semantic but still part of same system
- Tulving: Memory and Consciousness
-
-Anoetic Consciousness
-Noetic Consciousness
-Auto-Noetic Consciousness - Anoetic Consciousness
-
-Procedural memory
-"Not aware"
-Animals with just procedural memory can do these things but aren't aware of themselves doing it
-Insects, fish, invertabrates - Noetic Consciousness
-
-Semantic memory
-Awareness about external world
-Mammals, birds, octopus - Auto-Noetic Consciousness
-
-Episodic memory
-"Self awareness"
-Humans, some apes, cetaceans (whales, dolphins) - Mnemonics
-
-Tool/trick
-"To help memory"
-Method of loci, Peg-word mnemonic, and acronyms - Formal mnemonics
- Rely on preestablished set of memory aids and considerable practice on the to-be-remembered information in connection with the preestablished set
- Informal mnemonics
-
-Less elaborate
-Suited for smaller amounts of info
-More idiosyncratic and personalized - What principles are important to the strength of mnemonics?
-
-Practice to-be-remembered material repeatedly
-Integrate material into existing framework
-Device provides excellent means of retrieving information to be remembered - Method of Loci
-
-First memorize physical locations you're familiar with
-Form mental image of first thing you want to memorize and mentally place it into first location
-Continue with second item, third, etc. - Recall Performance with method of loci
-
-38 out of 40 items in correct order
-One day later, averaged 34 correct, again in order - Mnemonic Devices facilitate
-
-Encoding (structure for learning)
-Retention (organization &/or imagery)
-Retrieval (cues) - Atkinson & Shiffrin: Two Effects of Rehearsal
-
-Maintenance: actively maintain info in STM
and
-Transfer: actively take info from STM and put into LTM - Hellyer: Frequency of Rehearsal
-
-Brown-Peterson task
-Modification: rehearse items before given an arithmetic distractor task
-Varied how many times one could rehearse
-Believe accuracy is a function of retention interval - Hellyer: Results
-
-The more frequently the items were rehearsed, the better retention across the distracting period
More rehearsal = Greater chance of getting into LTM - Rundus: Rehearsal and Serial Position Effects
-
-Subjects learned 20-item lists of unrelated words
-Presentation rate of 5 s per word
-Subjects rehearsed out loud whatever word from list they wanted during 5 s presentation - How did Rundus compute his results?
- Tabulated number of times each of the words had been rehearsed and compared this tally to the likelihood of recalling the word correctly in free recall (serial-position curve)
- Rundus' results
-
-Items at beginning of list are rehearsed frequently
-Middle items not really rehearsed
-End items not rehearsed at all - From Rundus' results what do primacy and recency effects depend on?
-
-Primacy effect was viewed entirely dependent on rehearsal
-Earlier items can be rehearsed more frequently and so are recalled better
-Recency effect was viewed as simply recall from STM - Craik & Lockhart: Depth (levels) of Processing
-
-Single memory system: sensory, STM, & LTM (not seperate systems)
-Perceived stimulus receives some amount of mental processing
-Processing can be shallow or deep and more meaningful
-Two types of rehearsal - Type I (maintaince)
-
-Low level, repetitive info recycling at same level
-Once stop cycling information, it's lost (no permanent record in memory) - Type II (elaborative)
-
-Complex rehearsal that uses the meaning of the info to help store and remember it
-Move info to deeper level in memory system - Craik & Watkins: Depth of processing and repetition priming (1975)
-
-Running sequence of words
-3 conditions
-Manipulate subject's depth of processing
-Free recall - What were the 3 conditions in Craik & Watkins experiment?
-
-Case decision: surface features of letters (shallow processing)
-Rhyme judgements: (medium processing)
-Pleasantness: (deeper processing) - Craik & Watkins: Results
-
-Processed words to different depths
-Case judgement was shallow
-Rhyme -> STM
-Pleasantness -> LTM
-Increase recognition accuracy w/ deeper level processing - Craik & Watson: Depth of processing (1973)
-
-Hypothesized that amount of time word is kept in Type I rehearsal will not help recall
-Heard long list of words
-Task: remember most recent word starting w/ "G" - Craik & Watson (1973): Results
-
-Hypothesis confirmed
-No recall difference for "G-words" held a long time in STM vs those maintained only briefly
-No benefit - Baddeley's Criticisms on levels of processing
-
-Circular definition of Type I and II rehearsal
-Task effects: different results with recall vs recognition
-Useful heuristic but not a real alternative today - Circular Definition of Type I & II rehearsal
-
-Have premise that's used to draw conclusion, then use conclusion to support premise
Thus,
-Items that are processed deeply are remembered better (premise)
-Items that are better remembered are processed deeply (conclusion) - Task Effects: Recognition
-
-Generally a higher accuracy found with recognition
-Influenced by recollection and familiarity
-Less retrieval effort than recall tasks
-Type I rehearsal had effect on LTM when recognition tasks were used
***not true with recall tasks - Glenberg, Smith, and Green: Against depth of processing position
-
-Brown-Peterson Task
-Remember 4 digit number as (supposed) primary task
-Distractor task: repeat 1 or 3 words out loud (assumed they only used Type I processing) - Glenberg, Smith, and Green: Results
-
-Replicated Craik/Watson: amount of rehearsal didn't help recall of 4 digit number
BUT
-recognition of 4 digit number was influenced by the amount of rehearsal - What is the significance of Glenberg, Smith, and Green's results?
-
-Disconfirmed levels of processing theory
-Mere repetition and time in STM did affect retention (better performance with recognition task)
-Shallow processing (i.e. repeating words) can result in equal or superior perfomance -
Bower: Organization of Information
Why should I care? -
-He proposed organizational schemas for improving storage of info into LTM
-Presented 112 words; 4 trials
-Words fit under categories (hierarchy)
-Control group: words were randomly assisgned to position
-Experimental group: organized in hierarchy - Bower: Results
-
-Hierarchical organization led to 100% accuracy on 3rd and 4th trial
-Control only got 62% accuracy by 4th trial - Visual Imagery and Storage of Info in LTM
-
-Paired-associate learning encoded by imagery were recalled at better than 80% accuracy
-Dual coding hypothesis - Tulving & Thompson: Encoding Specificity
-
-Recognition failure of recallable words
-Wierd pairing: Glue-Chair (learn chair)
-2 tests to see if can recognize chair - Tulving & Thompson: 2 Tests
-
Test 1: table, top, chair
-not able to recognize chair as target
Test 2: Glue-?
-able to recall chair - Tulving & Thompson: results
- Original context cues will give you the best access to the information during a recall attempt, whether those cues are based on verbal, visual, or other info
- Godden & Baddeley: Context Effects
-
-Had deep sea divers learn list of words underwater
-Couldn't recall words out of water, but were able to recall when back underwater - Eich: State-Dependent Learning
-
-Study high, take test high -> get higher scores vs. if one takes test sober
-Same with alcohol
-Increase access to information in memory by reinstating original learning context - Patient KC: Declarative Dissociation
-
-Frontal brain damage
-Retrograde and anterograde amnesia for episodic memory
-Intact semantic memory though
-Evidence of a dissociation bwt episodic and semantic systems - Patient HM: Hippocampus lesioned
-
-Almost complete anterograde amnesia (no new memories)
-Normal memory for motor learning
-Episodic and semantic memory may eventually be compromised since ability to encode new connections to existing knowledge was lost - What region of the brain is important for time-related aspects of memories?
- Frontal Lobes
- What region of the brain is important for semantic retrieval?
- Posterior regions (i.e. hippocampal regions)
- HM's and motor skill tasks
-
-Intact skill learning
-Normal learning curve for mirror drawing task w/ few errors by day 3
-However, HM has no memory of doing the repeated task before (no explicit memory record) - Role of Hippocampus in memory
- Critical pathway for successful transfer to LTM
- Typical Implicit Memory Results
-
-Measure with perceptual or word stem completion task
-Show significant priming or facilitation regardless of how info was studied
***Even with no conscious recollection of orginal event, there's facilitation when stimulus is repeated - Typical Explicit Memory Results
-
-Measure with a recall or yes/no recognition task
-Generally show strong effects depending on how the info was studied
-More elaborative processing leads to better explicit memory performance - Roedigor & McDermont: False Memory
-
-List of words in which they were all similiar (i.e. dream, night, nap, etc.)
-Reported false memory of hearing word "sleep" which wasn't on list
-Episodic and semanitc memory interacted b/c they said "sleep" - Aristotle: Classical View of Concept
-
-Concept is a set of neccessary and sufficient features
-Classical view works well with artificial concepts but not with natural concepts
-Doesn't tell much about how semantic memory works (i.e. organization of it) - Example of Artificial Concept
-
"Bachelor"
-Not married
-Male
Two neccessary, sufficient features - Example of Natural Concept
-
"Student"
-Fuzzier category
-Many different features -
Collins & Quillian: Semantic Network Model
Two fundamental assumptions are? -
- Structure of semantic memory
- Process of retrieving info from structure - What is the structure of semantic memory?
-
*Semantic Network Model*
-It is a network structure
-Each concept in the network is a node
-Concept nodes are linked by pathways
-Every concept is related to every other concept b/c some sets of pathway (however indirect and long) can be eventually traced bwt 2 nodes - What is the major process that operates on this is structure?
-
-Spreading activation
-Activation spreads through the network along stored pathways
-Once a concept is activated, the concept spreads activation to other concepts it is linked with - Two types of Propositions
-
-Property Statements
-Category-Membersip Statements - Property Statements
-
-Inherit properties from concept above "A robin has wings"
(i.e. has wings b/c bird)
-Can have specific property
"Robin has a red breast" - Category-Membership Statements
-
"A robin is a bird"
-Is a relationship
-Indicates category membership
-Reverse direction is not true: "All birds are robins" - What are the two orginzational principles of the semantic network model?
-
1. Cognitive economy
2. Inheritance - Cognitive Economy
-
-No redundance
-Economize number of concepts stored
-Store properties at highest level of generality as possible to save space - What do the mechanisms of inheritance and cognitive economy allow to happen?
- The mechanism allows inferences about higher-level properities rather than forcing the system to store each of them directly and repeatedly
- Collins & Quillian: Sentence Verification Task
-
-Decidide if sentence is T or F
-Used both category relationship and property statements
-Varied semantic distance bwt S and P (distance in hierarchial structure)
-DV was RT - Colling & Quillian: Results
-
-Longer to retrieve relationship bwt two concepts at more distant levels in the hierarchy vs, those stored closer together
-Fast at category relationships (is a)
-False statements were slowest - How did the subject access the information to make a "yes" decision?
-
-Spreading activiation
-Find intersection between two concepts
-Decision stage operates to make sure retrieved pathway is valid and repsonse of "true" can be given - What were the con's of Collins and Quillians experiment?
-
-Spreading activation doesn't explain for the "no" responses
-No explanation of how network is learning
-Silent on ad hoc categories
-Faster with typicality - Ad Hoc Categories
-
-Categories you come up with on the fly
-Example is "good things to take out of a burning builing" - Why does typicality matter in Collins and Quillians experiment?
-
-Faster with "yes" response to "A robin is a [bird]"
-Slower with "A penguin is a [bird]"
-Should be equally fast b/c both are one link apart from the concept of [bird]
-No explanation provided for this - Rosch: Typicallity Effects
- Collected category membership norms and found some items listed as members of category more frequently than others
- Rosch: Typicallity Effects Results
- Exemplars listed very frequently as category members yielded significantly faster judgements than those lose of lower frequency
- What did Rosch's results suggest?
-
-Pathways were not all equal in length
-Pathways to less frequent category members were longer,
-Those members were farther away in semantic network than frequent category members - Smith et al.: Feature List
-
-Most basic structural element was feature list
-Semantic memory is a collection of lists
-Each concept in semantic memory would be represented as a list of semantic features - How does feature list compare to Collin and Quillian's network model
-
-Simpler than Collins and Quillians network model
-More elaborate in assumptions about the process of retrieval - How are feature lists ordered?
-
-Defining features
-Characteristic features
-Feature lists are ordered in a priority ranking with defining features at top of list and least defining features at the bottoom - Defining Features
-
-Necessary features for concept
-Essential feature
-Placed at top of feature list - Characteristic Features
-
-Not necessary but generally true
-Common but not essential
-Lower features at the bottom of the feature list - How is information retrieved in the Smith model?
-
-Through feature comparison
-Two types of comparison:
1. Stage I
2. Stage II - What is Stage I Comparison?
-
-Some randomly selected subset of features on each of the two lists is compared to "compute" the similiarity btw the two concepts
-Comparison process yields a feature overlap score - Characteristics of Stage I Comparison
-
-Rapid, global comparison of the features
-High overlap score produces a rapid "yes"
-Little or no feature overlap produces a rapid "no"
-No need to continue search when overlap scores are very high or low - What is Stage II Comparison?
-
-Look at defining features of two concepts being compared
-Careful and slow comparison
-Respond with slow "yes" or "no"
-When overlap score is moderate, a second comparison is necessary - What are the pro's and con's of the feature list?
-
Pro: it handles what semantic network theory could not (i.e. slower comparison processing when have lower feature overlap)
Con: it is not clear what defining features exactly are - What three characteristics would a modified network representation have?
-
1. No rigid cognitive economy
2. Properties listed for a concept would be linked directly to concept rather than indirectly via pathways
3. Pathways would be different length reflecting semantic relatedness (which applies to both category and property statements) - How is the amount of one's knowledge connected to semantic relatedness?
-
-More knowledge and greater semantic relatedness go together
-More you know about something, the easier it is to integrate new related knowledge into memory
-Greater knowledge leads to more activation in memory and enhanced retrieval - What was Roch's argument about artificial concepts that led to the prototype theory?
- Roch argued that artificial categories bear little relationship to natural categories
- How did Roch view natural categories or real-world categories?
-
-Real-world category members do not belong to their categories in simple
yes/no, all-or-one fashion
-3 principles of natural categories - Three Principles of Natural Categories
-
1. Fuzzy definition
2. Family Resemblance
3. Prototypes - Fuzzy Definition
-
-Ill defined or uncertain membership for many categories
-No single feature seems absolutely necessary as a criterion of membership or classification - Prototype
-
-Has most characteristic feature of concept
-Central, core instance of a category
-Average or best example of a concept - Family Resemblance
-
-Degree of variation from prototype
-Typical members are stored close to prototype and peripheral members are stored farther away - Rosch-Heider: Reasons for studying Dani Tribe
-
-Dani language has two color terms, one for "dark" and one for "light"
-Used different colored chips as stimuli - Rosch-Heider: Recognition Experiment with Dani
-
-Administered short and long term memory tasks
-Task was to recognize the chips they had seen before
-Better recognition with prototypical colors even though they don't have words for specific color conepts - Rosch-Heider: Paired-Association Learning with Dani
-
-Paired chip with clan name
-Use chip as a cue to recall clan name
-Prototypical colored chips were most useful cues - What conclusions did Rosch-Heider draw from the results of the experiment?
-
-There are structured, mental categories of colors in the subjects' semantic memories
-Categories do not rely on spoken language
-Natural concepts have an internal structure corresponding the the idea of typical and atypical (semantic relatedness) - Rosch: Prototype Theory (Three Levels)
-
1. Basic level
2. Superordinate level
3. Subordinate level - Basic Level
-
-Highest level in category at which you can remember what generic concept looks like
-Level best able to use to think and talk about concepts
-Set of features that define concept at this level are specific - What does the Basic Level depend upon?
-
-Concepts at this level depend on your expertise with the area
-The more knowledge you have in an area, the more specific and distinct your basic level will be - Subordinate Level - How many features does it differ from others?
- -Differs by only one or a few features
- Priming is a.... (two things)
-
-Phenomenon
-Method - How is Priming a phenonmenon?
-
-Mental activation when think about or see concept
-Spreading activation - Why do we study Priming as a method?
- We study how things are mentally represented
- Key terms associated with priming
-
-Prime
-Target
-Task (lexical decision, naming)
-Facilitation vs cost
-Stimulus-onset asynchrony (SOA) - What is priming across trials?
-
-One trial is the prime for target trial
-Example: trial such as "fruit-P" is followed by another "fruit" trial
-The first was the prime and the second was the task - Lexical Decision Task
-
-Present targets consisting of words or nonwords
-Task is to decide if target is word or nonword
-Look at RT - How does semantic relatedness affect lexical decision task?
-
-Two related words are judged more quickly as words than two unrelated words
-Influence of the meanings of the words
-Words semantically related prime - Priming is Automatic
-
-Automatic spreading activation (ASA)
-Facilitation is rapid and w/o intent - Neely: Priming Summary
-
-Lexical decision task
-Each letter string was preceded by a prime, either a related word, an unrelated word, or a neutral prime (baseline condition)
-2 sources of priming: automatic spreading activation and expectancy - Expectancy Priming
-
-Controlled, deliberate, slower
-Neely told subjects that when see prime word "building" to expect target to be a member of the body part category - Neely: Results for semantic pairing
-
-Faciliation for prime-target trails such as "Bird-robin"
-Not dependent on SOA
*so semantic priming is automatic - Neely: Results for unrelated prime
-
-Inhibition
-Inhibition effect grew stronger across longer and longer SOA - Neely: Results for "building" prime and body part target
-
-Should expect switch to body part category: "building-leg"
-At short SOA there was no priming
-At long SOA there was significant priming b/c formed expectancy set due to instructions - Neely: Results for "building" prime when didn't switch to body part target
-
-"building-door" had faciliation at short SOA (tapped into normal semantic priming) BUT cost at long SOA
-"building-shark" had baseline at short SOA and cost at long SOA - Priming Experiment
-
-See prime, time passes, see targe
-Task is to name target as fast as you can
-Fastest at naming target when prime and target are semantically related (i.e. prime nurse, target doctor) - Marcel: Priming is implicit process
-
-Prime was immediately followed by visual mask (scrambled visual pattern)
-Purpose was to present mask soon after the prime so subjects were not consciously aware of the prime - Marcel's Results with Priming
- RT's still faster at target when prime was semantically related even though they didn't consciously register prime
- Implicit vs Explicit memory priming effect
-
-Implicit memory priming effect is rapid suggesting priming can operate automatically
-Explicit memory priming effect is deliberate (like in Neely experiment with "building" as expectancy prime)