Questions about Memory 1 Do we learn only

Questions about Memory 1. Do we learn only with intention – or also without intention? We learn with and without intention. 2. Is learning influenced by what we already know? And if so, how?

The two-edged sword What we already know can help us perceive what is out there in the world (encoding) and help us recall it (retrieval). It can also cause us to see what we expect (not what’s there) or to construct a memory of “what usually happens” (not what actually happened).

What we already know (WWAK) 1. Knowledge effects at encoding Levels of Processing theory Criticisms of Lo. P 2. Knowledge effects at retrieval Transfer Appropriate Training Encoding Specificity

Knowledge effects at encoding Levels of Processing theory – Craik & Lockhart (1972) In 1950 s & 60 s, CP was most interested in questions about structure (e. g. , SM – STM – LTM). In early 1970’s, CP became more interested in process Craik & Lockhart said, ability to recall a stimulus depends upon how you processed it.

Craik & Lockhart’s (1972) tenets: 1. Ease of information retrieval depends upon type of code generated at learning. 2. Kind of code generated depends on your purpose when you first process the information. E. g. , are you looking for a red thing among green things? Then generate visual codes.

Craik & Lockhart’s (1972) evidence Orienting task - subjects read a list of words and answer one of three questions: Which words in list contain letter 'e'? Which words in list rhyme with CANE? Which words in list name animate objects? On surprise recall test, success varies with orienting task: Semantic > Rhyme > Spelling

Levels of Processing – the basic effect Dependence of recall on orienting task is the basic levels of processing effect Probability of recall varies with type of code generated when learning. See also studies described in text (pp. 153 - 156) Why does this effect happen?

Levels of Processing – the explanation Craik & Lockhart – 2 types of rehearsal: Maintenance Rehearsal uses articulatory loop simply saying words over and over Elaborative Rehearsal uses the meaning of the object or event requires establishing associations

Levels of Processing – the explanation Compare sound [banana] vs. meaning "banana" - what associations can be made? Semantic associations are richer, more distinctive – therefore memorable. Lo. P effect reflects richer associations to stimuli processed for meaning.

Criticisms of Lo. P Theory Baddeley – L. O. P Theory is circular Which levels produce best memory? Deepest Which levels are deepest? Those that produce best memory. No independent way of assessing ‘depth. ’

Criticisms of Lo. P Theory Baddeley – result does not generalize to other tasks. E. g. , Glenberg, Smith, & Green (1972) – Lo. P effect not found for recognition task. Recognition task: subject shown “old” and “new” stimuli, asked to say which is which. Recognition task generally easier than recall task.

Knowledge effects at retrieval Bransford – Transfer Appropriate Training Lo. P – memory performance depends upon conditions under which encoding occurs. Bransford – memory performance also depends upon conditions under which retrieval occurs.

Transfer Appropriate Training Morris, Bransford, & Franks (1977) used semantic & rhyme orienting tasks at retrieval, some subjects asked to recall words seen during orienting task. others asked to detect words that rhymed with words seen during orienting task.

Transfer Appropriate Training Morris et al. (1977) Group 1 2 3 4 Orientation Semantic Rhyme Task at Retrieval Recall Rhyme

Transfer Appropriate Training Morris et al. (1977) - Results Recall Rhyme Judgment SO > RO RO > SO Encode for the way you plan to use the information. SO: Semantic orienting RO: Rhyme orienting

Knowledge effects at retrieval Bransford’s idea was that retrieval success depends upon the match between what happens at retrieval and what happens at encoding. At the time, it was a radical idea, because most researchers still believed (as behaviourists argued) that behaviour was governed by pre-existing learning…

Knowledge effects at retrieval If behaviour was controlled by preexisting learning, it shouldn’t matter whether retrieval conditions matched learning conditions. Match/mismatch could not alter the facts about pre-existing learning. Bransford’s results questioned this idea. So did Tulving’s (Tulving & Osler, 1968).

Encoding Specificity Theory Thomson & Tulving (1970) Subjects learned a list of words for later recall. Some subjects got words without a context. Subjects who got words in a context, got either strong or weak contexts.

Encoding Specificity Theory In Thompson & Tulving’s (1970) study, subjects were given pairs of cues & targets in the learning phase. Then, in the test phase, they were given the cue and asked to recall the target. Learn Recall cue Response COLD – hot COLD

Encoding Specificity Theory Thompson & Tulving (1970) Condition Learn Recall cue Strong 1 COLD – hot Strong 2 COLD – hot blow

Encoding Specificity Theory Thompson & Tulving (1970) Condition Learn Recall cue Weak 1 COLD – blow hot Weak 2 COLD – blow

Encoding Specificity Theory Predictions: If long-term learning is most important, then hot should have been a better cue for COLD than blow, regardless of learning condition. If blow is a better cue (when it is presented at learning), that means that context matters.

Encoding Specificity Theory Thompson & Tulving’s results: Probability of recall with blow as cue was higher than with hot as cue, when blow was presented at learning. A cue is more effective if it re-establishes the learning context. This was a radical idea at the time.

Review – Craik & Lockhart Type of code you generate when you process a stimulus varies with your purpose. Ability to retrieve a stimulus later varies with type of code you generated. Maintenance Rehearsal involves simply repeating the stimulus, without creating new connnections.

Review – Craik & Lockhart Deeper processing gives access to meaning, which permits more elaboration. Elaborative Rehearsal involves working out connections between the new stimulus and what we already know. Greater elaboration usually leads to better memory.

Review – Criticisms of Lo. P Theory Baddeley: Lo. P theory is circular. Lo. P effects do not generalize well to tasks other than recall.

Review – Morris, Bransford, & Franks Transfer Appropriate Training: Conditions at retrieval are important as well as conditions at encoding. Encode for the way you plan to use the information

Review – Thompson & Tulving Encoding Specificity theory: You encode aspects of context when you learn new information. Cues help most in retrieval if they reestablish the learning context.