Questions about Memory

1
Questions about Memory

• Do we learn only with intention or also
  without intention?
• We learn with and without intention.
• Is learning influenced by what we already know?
  And if so, how?

2
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
  whats there) or to construct a memory of what
  usually happens (not what actually happened).

3
What we already know (WWAK)

• 1. Knowledge effects at encoding
• Levels of Processing theory
• Criticisms of LoP
• 2. Knowledge effects at retrieval
• Transfer Appropriate Training
• Encoding Specificity

4
Knowledge effects at encoding

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

5
Craik Lockharts (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.

6
Craik Lockharts (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 gt Rhyme gt Spelling

7
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?

8
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

9
Levels of Processing the explanation

• Compare sound banana vs. meaning "banana" -
  what associations can be made?
•  
• Semantic associations are richer, more
  distinctive therefore more memorable.
• LoP effect reflects richer associations to
  stimuli processed for meaning.

10
Criticisms of LoP 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.

11
Criticisms of LoP Theory

• Baddeley result does not generalize to other
  tasks.
• E.g., Glenberg, Smith, Green (1972) LoP
  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.

12
Knowledge effects at retrieval

• Bransford Transfer Appropriate Training
• LoP memory performance depends upon conditions
  under which encoding occurs.
• Bransford memory performance also depends upon
  conditions under which retrieval occurs.

13
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.

14
Transfer Appropriate Training

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

15
Transfer Appropriate Training

• Morris et al. (1977) - Results
• Recall Rhyme Judgment
• SO gt RO RO gt SO
• Encode for the way you plan to use the
  information.
• SO Semantic orienting RO Rhyme orienting

16
Knowledge effects at retrieval

• Bransfords 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

17
Knowledge effects at retrieval

• If behaviour was controlled by pre-existing
  learning, it shouldnt matter whether retrieval
  conditions matched learning conditions.
• Match/mismatch could not alter the facts about
  pre-existing learning.
• Bransfords results questioned this idea. So did
  Tulvings (Tulving Osler, 1968).

18
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.

19
Encoding Specificity Theory

• In Thompson Tulvings (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 hot COLD

20
Encoding Specificity Theory

• Thompson Tulving (1970)
• Condition Learn Recall cue
• Strong 1 COLD hot hot
• Strong 2 COLD hot blow

21
Encoding Specificity Theory

• Thompson Tulving (1970)
• Condition Learn Recall cue
• Weak 1 COLD blow hot
• Weak 2 COLD blow blow

22
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.

23
Encoding Specificity Theory

• Thompson Tulvings 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.

24
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.

25
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.

26
Review Criticisms of LoP Theory

• Baddeley
• LoP theory is circular.
• LoP effects do not generalize well to tasks
  other than recall.

27
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

28
Review Thompson Tulving

• Encoding Specificity theory
• You encode aspects of context when you learn new
  information.
• Cues help most in retrieval if they re-establish
  the learning context.