LONG TERM MEMORY Flashcards
(18 cards)
EYSENCK & EYSENCK (1980)
Distinctive processing manipulated independently from level of processing (distinctive processing results from greater depth of processing and semantic elaboration). Part. had to process nouns in semantically distinct fashion (i.e. by adding an unusual adj to it) and non-distinct fashion (with common adj.) & phoneticaly distinct fashion (by pronuncing irregular words in a way that their spelling suggests) and phonetically non-distinct. There was little difference between recognition performance after semantic and phonetic distinctive processing, and semantic non-dist. There was a big drop though after phonetic non-dist. Semantic processing enhances memory perf., but distinctive processing, even with phonetic (shallow) processing can lift memory performance.
CRAIK & TULVING 1975
Experiment manipulating levels of processing and tested recognition memory. 4 conditions. 1&2 required perceptual processes (1 - graphemic - is the word in capital letters, 2 - phonetic - does it rhyme with), 3&4 semantically orientated processing, with more elaborative processing required for q4 (does the word fit into a sentence). Participants memory performance increased with deeper levels of processing. HOWEVER problems with this theory include: lack of definition of levels of processing - it may end up being defined in a circular fashion; does not account for all memory phenomena - maintenance rehearsal has been found to improve recall. Depth of processing account only focuses on encoding, not on retrieval.
DISTINCTIVE & SIMILARITY-BASED PROCESSING
Distinctive processing helps lift memory performance (Eysenck&Eysenck, 1980), but categorizing items based on their similarities also helps organize memory. Hunt and McDaniel proposed there are different forms of processing: relational processing underlies similarities, item-specific processing - distinctiveness.
MANDLER (1980)
Dual process model of recognition. When an item is encountered, two processes take place.1) familiarity - influenced by the degree of integration at encoding (such as maintenance rehearsal ,which employs item specific processing) - fast process, facilitates simple recognition (have I seen this item before?). 2) search and retrieval - slower process used in recall - to check if the item was presented before - benefits from elaboration such as semantic processing, facilitates identification (where have I seen it before?). That’s why we know if we’ve seen sth before, before we know where was it. This theory explain the findings that maintenance rehearsal improves recognition, because it promotes integration and is item-specific.
HUNT & EINSTEIN 1981
Effects of item-specific and relational processing on free recall and recognition. Part. presented with a list of 36 categorized words (6 words in 6 categories), or 36 unrelated words. The assumption was part. would process the first lot spontaneously in a relational fashion, and the second lot - in item specific way. Part. were asked to sort the words into specified categories (a relational processing task), or rate their pleasantness (item-specific task). Then Recall and recognition were tested. It was found that free recall benefited from both relational and item-specific processing.Recognition of categorized list items was greater after item-specific processing, than after relational processing. Unlike in free recall, recognition of unrelated list items was the same irrespective of relational or item-specific processing. Additional item-specific processing may become redundant for free recall, but it continues to benefit recognition.
TULVING/SCHACTER
Episodic and semantic memory. Episodic memory is a record of person’s experience. Semantic memory is the general knowledge store. PROBLEM: these two systems need to communicate with eachother. For example full comprehension of info in episodic memory must rely on the knowledge of the world, in semantic memory. Tulvig suggested ep. is embedded in sem. Sentence verification (relies on semantic memory) & recognition (relies on episodic) were used to test this. It was measured how long part. took to verify a sentence. In 4 conditions part. were first presented with episodic info about the categories and exemplars, in one no info was provided. The results showed that the time necessary to verify the sentence was affected by the type of episodic info provided beforehand. Episodic info affected retrieval from semantic memory - they are not distinct entities, and info travels both ways.
MULTIPLE MEMORY SYSTEMS CRITICISM
Experiments such as Anderson & Ross (sentence verification, with episodic memory affecting semantic mem.)
Lack of theoretical development: fan effect - sentence recognition RT increase as more info about the concept is acquired.Recognition relies on episodic memory & the fan effect is only observed in tests of episodic memory. These observations could be interpreted as supportive of distinction b. ep & sem memory, but there is nothing in the theory of multiple memory systems that would explain why the fan effect only affects episodic and not semantic memory.
MILNER (1966)
Amnesic patient. HM underwent brain surgery which resulted in him becoming amnesic at age of 27. He retained memories up to shortly before the procedure, but was not able to form new memories. Could read the same book over and over again, didn’t remember seeing people the same morning. He retained his language skills and intellect from before the damage. Tulvig interpreted it as his episodic was gone, but semantic was largely unaffected. However it was noticed that HM used vocabulary common at the time of his operation and couldn’t learn any new words that entered the language since then. It cannot be used to support the multiple memory system.
SQUIRE
Proposes 2 LTM systems - declarative & procedural knowledge. Declarative corresponds to ‘knowing that’. Reponses to episodic and semantic mem tasks provide declarative info. Procedural knowledge corresponds to ‘knowing how’ - performance related. Amnesics can normally learn a range of different tasks, but have no memory of learning them. According to Squire it’s a failure in declarative mem that manifests itself in amnesics.
TULVIG & OSLER Encoding specificity
The role of cues in memory retrieval. Part. presented with a word to remember plus either zero, one, or two weak cues to help them remember the word. Part. were told the words in lower case might help them remember the target word & to think how it might help them remember it. It was found that a single weak cue helped retrieval, but only if it was present at learning. Retrieval cues facilitate recall only if the info about them and how they relate to target is stored at the same time as the target.
TAP HYPOTHESIS
Transfer Appropriate Processing - focuses on the overlap between the processes employed at encoding and the processes employed at retrieval. It predicts that the best memory performance will be observed when the processes engaged at encoding will transfer appropriately to retrieval.
MORRIS et al (1977)
Experiment into TAP hypothesis. Participants were presented with list of words, such as CAT or TABLE. Half of part. received orienting question that required phonetic processing (does the word rhyme with…), the other half - semantic (do you sit on it?). Next day, half of part. in the first group received a semantically oriented recognition test (which words were presented previously?), the other half - phonetically orientated recognition test - identification (which words on the list rhyme with previously presented words). The same was applied to the other group. As TAP predicts, there was better performance when there was a match between the processes engaged at encoding and retrieval.
TULVING et al (1982) IMPLICIT MEMORY
Participants asked to learn 96 words. after an hour they were asked to carry out a recognition test with 24 presented words (targets) and 24 similar words (distractors), and a word-fragment completion test, with 24 word-fragments based on words presented before & 24 not presented before. For the 24 presented before, those words were the only possible solutions to the task. 7 days later they received another recognition test for the remaining 48 words.
It was assumed part. will carry out the word completion taks without realising the previously presented words were the solutions for half of the cases - word completion task is assumed to be an implicit memory test. It was found that more word fragments were completed correctly when the word had been previously presented - repetition priming effect. Explicit memory recognition test declined dramatically between 1 hr and week interval, there was no significant decline in word completion task in that time . It could have been that part. realised that the words presented before completed the word-completion task & tried to remember them. If that was the case however, there would have been a decline in the time frame, in accordance with explicit recognition test performance.
TAP ACCOUNT OF IMPLICIT MEMORY
TAP can explain the differences in performance in implicit and explicit memory tests. It is not important whether the info is retrieved implicitly or explicitly, but it is important that the method of encoding matches the method of retrieval. It is believed that performance on all implicit conceptual tests should match that observed with free recall (explicit test), as there are no retrieval cues in free recall so it relies on top-down processing.
However, it has been found that processing words which are conceptually related to each target word (conceptual repetition) enhanced free recall, it did not enhance performance in the category exemplar generation test (implicit mem. test). This contradict TAP prediction of equivalent memory performance.
CERMAK et al (1995)
Offered explanation to why amnesic patients perform well in implicit memory test. This finding is against the TAP account. TAP says amnesics perform well because they retained their perceptual processing capabilities. Thus they lost their conceptual processing. BUT they perform as well as healthy controls in conceptual implicit tests. Cermak explained this finding with dual memory processes. Amnesics can perform well if the task can be accomplished by relying on item familiarity/ processing fluency (perceptual). Explicit tasks require more context-based processing (conceptual).
JACOBY et al (1993)
Process-dissociation procedure. Two conditions - full attention, where participants had to read a list of words; divided attention - they had to also listen to a recording of a string of numbers and indicate when 3 odd ones are said. This was to reduce the influence of recollective processes, but to leave automatic processes unaffected. A stem task inclusion and exclusion tests followed. Jacoby worked out the probabilities of responding with a previously presented word in line with the view that automatic processing are unaffected by changes in attentional resources available at encoding, whereas recollective processes suffer dramatically.
TULVING (1985) Remember and know judgements
Experiment 1: part. presented with lists of words. First word introduced a category, second was an exemplar of that category. 3 memory tests followed - free recall of exemplars, cued recall - cat. name the cue, cued recall - cat. name + first letter of exemplar. Part. had to judge if they remember the exemplar, or know it was that. All the free-recalled items were scored, in the other 2 tests only those not-recalled were scored. Feelings of remembering arise from the representational richness of episodic memory, remember judgments should be most prevalent in free-recall category (must be the richest representation if recalled without any cues), less prevalent with a cat. cue & least prevalent in cue and first letter. Data analysis confirmed these predictions.
In experiment 2, half of the stimulus was tested immediately, half after 8 days. Remember judgements decreased after 8 days interval, in accordance with Tulving’s view that remember judgements reflect richness of info held in episodic memory, which diminishes over longer retention intervals.
BUTLER & BERRY 2001 Offprint
Criteria for implicit memory: 1) performance reflects unintentional retrieval strategy 2) there is no conscious recollection.
Contamination problem - there are often insufficient safeguards to ensure these criteria are met. With these problems, there is a lack of strong evidence for implicit memory in normal partcipants. Performance often represents an unintentional but conscious use of memory, particularly for conceptual tasks. There is some evidence for implicit memory in amnesic patients, but can it be generalised to healthy population?
