History of Memory

Theory and Research

Tend to think of it as some
kind of place where things
or pieces of information
are stored
 Plato. Memory is a wax tablet
 Plato. Memory is an aviary
 Locke. Memory is a cabinet
 Pear (1922) memory is a grammaphone
 broadbent (1950’s) memory is like a
telephone system
 Simon and Figenbaum (1960’s) memory is
like a computer.
 Romans (Cicero, 106-43 bce and
quintillian (40-96 ce)
 350-ish CE. – Augustine ---Two kinds
of memory
 Middle ages. (1200s)
 the renaissance Vives 1492-1540
 bacon. (1561-1626).
 Hobbes (worked for Bacon)
 Descartes (1596-1650)
 the 1700s Locke
 the 1800’s. Gustav Fechner
 ebbinghaus
 After ebbinghaus – the 1900s
what does memory research
have to do with the real
world?
 Neisser (1978) asked this question. in
1976, he wrote Cognition and Reality
 critical of the methodology of much
cognitive psychology, faulting it for
being "ecologically invalid."
What’s the typical way to
study memory?
 The Experiment
 IV – vary:
 Task or Processing strategies
 Classes of stimuli

 Quality of stimuli

 Conditions of study

 Type of test
 How do we test memory?
Recall
 Free recall: “tell me what you remember from what you
studied”
 Cued recall (recall with some hint): “Tell me the furniture
words you studied” or “Tell me the words you studied that
start with “b” “, etc
 Serial Recall: “tell me what you remember from what you
studied, but report them in order”

Recognition
 o Yes/no recognition (also called old/new
recognition): “For each item in the test, indicate whether it is
OLD (studied) or NEW (unstudied)
 o Forced-choice recognition: “From these two (or
three or four or six, etc) choices, pick the item you studied”
 How do we test memory?
Indirect Memory tests
Stem Completion CHA----
Partial cuing C-A-I--
The DV - WHAT gets
measured?
 Accuracy of response
 Speed of response
 Response bias.
 Number of intrusions
 Response confidence
 Nature of the intrusions
 Order of responses
Accuracy measures

 An accuracy score is meaningless

unless you know the accuracy one
could expect from chance
performance – what a person could
get from pure guessing with no
memory whatsoever.
Consider the following

 The subject studies 20 words.

 The test includes the 20 studied
words, and 10 new words.
Three response strategies

 Our subject responds “old”,”new”,

“old”,”new”, etc in perfect alternation
because he/she has no idea.
 he or she responds “old” to every item
on the list.
 he or she responds “new” to
everything
All three Ss had no
memory for the items
 Our subject responds “old”,”new”,
“old”,”new”, etc in perfect alternation
because he/she has no idea. (50%
accuracy)
 he or she responds “old” to every item
on the list. (67% accuracy)
 he or she responds “new” to
everything (33% accuracy)
 Solution #1. Average the percent-
correct for old and for new.
 Solution #2. Subtract incorrect
answers from correct answers
 Stimulus Response Label
 Old Old Hit (correct)
 New New Correct rejection (correct)
 Old New Miss (incorrect)
 New Old False alarm (incorrect

If we take the Hit rate, and subtract the FA rate, we have a

corrected calculation of what they’ve remembered.
 Scenario 1: 50%-50% = 0
 Scenario 2: 100% - 100% = 0
 Scenario 3: 0% - 0% = 0
 Accuracy in a free recall test
 Chance in a cued-recall test
New Lecture: Sensory
Memory
1. something that looks or sounds like
the stimulus continues to be present
for a little while after it goes away –
stimulus persistence
2. information can be extracted from a
stimulus for a short time after it has
been taken away – information
persistence
 when the persistence is visual – iconic
memory,
 and when its auditory, its echoic
memory.
 Javon – 1871 span of apprehension
with beans
 Sperling (1960) more complicated
stimuli. Letters in 3x4 matrix
 Sperling’s Experiment
 Presented matrix of letters for 1/20
seconds K Z R
 Report as many letters as possible
Q B T
 Subjects recall only half of the
letters S G N
 Was this because subjects didn’t
have enough time to view entire
matrix? No
 How did Sperling know this?
 Sperling showed people can see and recall ALL the
letters momentarily
 Sounded low, medium or high tone immediately
after matrix disappeared
– tone signaled 1 row to report
– recall was almost perfect
 Memory for image fades after 1/3 seconds or so,
making report of entire display hard to do
 Sensory Memory Store

 Sensory memory forms

Sensory automatically, without
Sensory attention or
Input
Memory interpretation
 Attention is needed to
transfer information to
working memory
 We mentioned stimulus and
information persistence earlier. Are
both of them the same thing?
 Sperling’s task Ss identified numbers and
letters – were interested in info persistence,
and their data suggested that the
information was available for about 200 ms,
maybe up to almost a second.

 Haber and Standing (1969)– flashing circle.

Increase speed until it looks like its not
flashing anymore. About the same amount
of time.
 http://
www.johnsadowski.com/big_spanish_c
astle.php
 if iconic memory is an afterimage,
people should make mistakes when
they are asked to report the colours of
letters.
 Banks and Barber (1977) it doesn’t
happen, the original colour is
maintained.
So for example:

ARPY
LPKQ
BELP
 Turvey (1973) used brightness and
pattern masking.
– Brightness masking – after letter is
presented, its followed by flash of light
– Pattern masking – after letter, followed
by features – look like Chinese writing.
 Present letter and brightness mask to
same eye, get mask, if different eyes,
no mask.
 When you use pattern mask, you get it
when letter presented to one eye and
mask to the other.
Bar probe task (Averbach
& Coriell, 1963)
Show string of letters for 200 ms
TYPOIUGB
Mask it
&&&&&&&
Probe a position and ask S to report
letter
&&&&&&&
Problems with iconic
memory.
 iconic memory has a precategorical
view of the original stimulus.
 item vs location information
 Accuracy in Bar Probe task

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
1 2 3 4 5 6 7 8
 Accuracy in Bar Probe task

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
1 2 3 4 5 6 7 8
Echoic memory
 Similar kind of phenomena with auditory
stimuli. Echoic memory. Neisser (1967)
 Similar to iconic memory, echoic memory is
a system that receives auditory stimuli and
preserves them for a short amount of time
 Usually noticed as the cocktail party effect
People have tried to study
it the same way that
iconic memory is studied
How many letters can the
subject report?

0 sec 1 sec 4 sec 5 sec

Partial 4.86 4.75 4.4

Whole 4.2
Other studies: tended to div up echoic
memory into two study of two types of
persistence – stimulus and information
persistence

1. Stimulus persistence: Efron (1970) tones. Range in

duration from 30 to 100 ms. S adjusts a light to turn off
when they perceive the tone to have ended. Found that
regardless of tone duration, Ss subjective duration was
about 130 ms.

1. Information persistence. The suffix and modality effects

Modality effect.

 Conrad and Hull. Ss read a list of 7

digits. One group reads them aloud,
the other reads them silently. Then,
have to recall them in order.
 The curves are similar, but find that
the performance for the aloud group is
better for last couple items than for
the silent group.
Modality Effect
0.9

0.8
Read aloud

0.7 Read silently

0.6

0.5

0.4

0.3

0.2

0.1

1 2 3 4 5 6 7
Suffix Effect
0.9

0.8
buzzer

0.7
speech

0.6

0.5

0.4

0.3

0.2

0.1

1 2 3 4 5 6 7
Morton, Crowder, Prussin (1991)
the suffix effect not affected by :

1. practice : you keep getting it no

matter how many trials you have the S
run.
2. meaning or predictability of suffix
3. whether it was /uh/ or /zero/
4. semantic relatedness to rest of list
5. word frequency of suffix
6. emotionality of suffix
Was REDUCED when

1. acoustic info changed. E.g., different

spatial location or pitch of suffix
2. when the sex differed of speaker from
items to suffix
3. when speaker differed

4. suffix was visually presented.

 Probs with the PAS?
1. Is it really precategorical?
 Ayres et al (1979)
 Neath, Suprenant & Crowder (1993)
2. Spoehr & Corin (1978) found that you get the suffix
effect if the suff is mouthed—shouldn’t happen if the
PAS stores acoustic info
3. You get the suffix effect during articulatory suppression.
Saying /the/ while the list is being presented.
4. You get a suffix effect when followed by distractor
activity. Shouldn’t happen.
Next Time:

 Where does the information go when

it leaves Sensory Memory?

SHORT TERM MEMORY