MEMORY

Alan Baddeley, Michael W. Eysenck, And Michael C. Anderson

Third Edition, 2020
 CONTENTS
1. Why do we need memory?
2. One memory or many?
3. Theories, maps, and models
4. How can we study memory?
5. How many kinds of memory?
6. Sensory memory
7. Short-term and working memory
8. Long-term memory
9. Memory: Beyond the laboratory
10. Summary
11. Points for discussion
12. Further reading
13. References
 INTRODUCTION
• Why are we quite happy to claim “I have a terrible memory!”
• But, not to assert that “I am amazingly stupid”?
• Schacter (2001) who, having described what he refers to as the seven
sins of memory, accepts that the sins are in fact the necessary
consequences of the virtues that make our memories so rich and flexible.
• For these reasons, computer scientists are beginning to be interested in
learning from human memory and importantly forgetting, with a view to
potentially building some of these characteristics into computer memory
(Mezaris, Niederee, & Logie, 2018).
• Perhaps the most dramatic evidence for the usefulness of human memory
comes from the plight of patients who have lost these capacities as in the
case of Clive Wearing who has the misfortune to have had much of his
memory capacity destroyed by disease (Wilson, Baddeley, & Kapur, 1995).
 WHY DO WE NEED MEMORY?
[Link]
[Link]
[Link]
[Link]
 ONE MEMORY OR MANY?
MEMORY

TIME
Retrospective / Prospective /
Past Future

NO. OF Individual Intrapersonal / Social /

PERSONS Transactive Collective

DURATION Sensory Short-Term Long-Term

Memory Memory Memory
VERBAL Non-Declarative Declarative
EXPRESSIONS Memory Memory
Implicit Memory Explicit
Memory
CONTENTS Episodic Semantic
Memory Memory
Primin Procedural Associative Learning Non-associative
g Memory Classical & Learning
Skills & Habits Instrumental
 THEORIES, MAPS, MODELS
• A good theory will help us to ask new questions and that in turn will help
us find out more about the topic we are mapping.
• The nature of the theory will depend on the questions we want to answer,
just as in the case of maps of a city.
• In the case of psychological theories, different theories operate at
different levels of explanation and focus on different issues.
 THEORIES, MAPS, MODELS –
• ADDITIONAL
Reductionism: MATERIAL
The view that all scientific explanations should aim to be
based on a lower level of analysis: Psychology in terms of physiology,
physiology in terms of chemistry, and chemistry in terms of physics.
• Sciences tend to evolve in a direction that introduces greater emphasis
on formal theorizing.
• Ideal way of formal theorizing is to model.
• A model is a deliberately simplified and usually idealized and imaginary
representation of a phenomenon, with fundamental properties that are
explicitly defined and from which other properties can be deduced by
logical reasoning or by empirical observations.
• Models can take three form viz. verbal models of memory, mathematical
models of memory and computer simulation models of memory.
• Verbal models of memory includes empirical investigations, and
descriptions such as levels of processing framework of Craik & Lockhart,
and Multiple systems theory of Schacter & Tulving.
• Many psychologists are not satisfied with mere verbal models, they would
like to go beyond identification of lawful patterns of memory and -
HOW CAN WE STUDY MEMORY?
• The case of Clive Wearing demonstrates how important memory is, and
how complex, but leaves open the question of how it can best be studied.
• The attempt to understand human memory extends at least as far back
as Aristotle, and forms one of the classic questions within the philosophy
of mind, although without reaching any firm conclusions.
• This was vividly illustrated by a lecture on memory by the eminent
philosopher A. J. Ayer that I attended as a student.
• He began rather unpromisingly, by declaring that memory was not a very
interesting philosophical question.
• He seems to have demonstrated this pretty effectively as I can remember
none of the lecture, apart from his statement that his memory was totally
devoid of imagery, prompting a skeptical questioner to ask “If I tell you
that the band of the grenadier guards is marching past the end of the
street, banners flying and trumpets sounding, do you not hear or see
anything?”
• “No,” replied the philosopher.
• “I don’t believe you!” said the questioner and sat down crossly.
HOW CAN WE STUDY MEMORY?
• This point illustrates a limitation of a purely philosophical approach to
the understanding of memory in particular, and to mind in general,
namely its reliance on introspection, the capacity to reflect and report
our ongoing thoughts.
• These are not unimportant, but are not a reliable indication of the way
our minds work, for two principal reasons.
• The first of these, as our example shows, is that people differ in what
they appear to experience in a given situation; does memory depend on
visual imagery, and if not, why do some of us experience it?
• Second and even more importantly, we are only consciously aware of a
relatively small proportion of the mechanisms underpinning our mental
life, and as we will see, the tip of the mental iceberg that is available to
conscious awareness is not necessarily a good guide to what lies beneath.
• While there are still important issues addressed by the philosophy of
mind, it is now generally acknowledged these can best be pursued in
collaboration with a scientific approach based on empirical evidence.
• To return to the question of imagery, as I suspect Ayer knew, in the late
19th century, Sir Francis Galton had asked a number of “eminent men”
HOW CAN WE STUDY MEMORY?
• What was not known by Galton is that these huge differences are not
reflected in how accurate our memories are, suggesting that accuracy
depends on some nonconscious process.
• Could it be that different people have the same experience but just
describe it differently?
• Or do they have different memory systems?
• Or perhaps they have the same basic systems but have a different way of
using them?
• So how can we move beyond introspection?
• An answer to this started to develop in Germany in the latter half of the
19th century.
• It was concerned initially with the discipline of psychophysics, an
attempt to systematically map the relationship between physical stimuli
such as brightness and loudness onto their perceived magnitude.
• Despite success in linking physical stimuli to the psychological
experience of participants, capacities such as learning and memory were
initially regarded as unsuitable for experimental study.
HOW CAN WE STUDY MEMORY?
• This view was dramatically overturned by a German philosopher Herman
Ebbinghaus who conducted an intensive series of experiments on himself
over a two-year period, showing that it was indeed possible to plot
systematic relationships between the conditions of learning and the
amount learned.
• Having published this, the first classic book on the science of memory
(Ebbinghaus, 1885), he moved on to study color vision, intelligence, and
a range of other questions in the newly developing field of experimental
psychology.
• So what did Ebbinghaus do?
• He began by simplifying the experimental situation, attempting to
develop material that was devoid of meaning but was verbally learnable
and reportable, inventing what has become known as the nonsense
syllable, a consonant-vowel-consonant nonword such as zug pij and tev.
• He served as his own subject, always holding constant the room in which
he learned, the time of day, and the rate of presentation which was rapid,
so as to avoid any temptation to attempt to find meaning in the stimuli.
• Ebbinghaus established some of the basic principles of learning that will
HOW CAN WE STUDY MEMORY?
• The Ebbinghaus tradition was subsequently most strongly developed in
the US, focusing particularly on the factors and conditions surrounding
the important question of how new learning interacted with what was
already known.
• Results were interpreted in terms of associations that were assumed to
be formed between stimuli and responses, using a limited range of
methods that typically involved remembering lists of nonsense syllables
or words (McGeoch & Irion, 1952).
HOW CAN WE STUDY MEMORY?
• This is often referred to as the verbal learning approach.
• It developed from the 1930s to the 1960s, particularly in US mid-Western
laboratories, and emphasized the careful mapping of phenomena rather
than the ambitious building of grand theories such as that proposed by
Clark Hull’s general theory of learning based largely on the behavior of
rats in mazes which was presented in an elaborate style based directly on
that used by Isaac Newton in presenting his classic work, the Principia.
• However the various disputes between such theories appeared to reach
deadlock in the late 1950s leading to their general abandonment.
• This allowed the more staid verbal learning approach, previously
disparagingly discounted by its critics as “dust bowl empiricism,” to
attract a broader range of investigators interested in studying learning
and memory.
• This in turn led to the founding of a new journal, The Journal of Verbal
Learning and Verbal Behavior, which, when the term “verbal learning”
later became unfashionable, became The Journal of Memory and
Language.
• A second development that occurred at this point had its roots in both
HOW CAN WE STUDY MEMORY?
• In the 1930s, a German approach known as Gestalt psychology began
attempting to apply ideas developed in the study of perception to the
understanding of human memory.
• Unlike the behaviorist approaches, Gestalt psychologists tended to
emphasize the importance of internal representations rather than
observable stimuli and responses, and to stress the active role of the
rememberer.
• Gestalt psychology suffered badly from Nazi persecution, but enough
Gestalt psychologists moved to North America to sow the seeds of an
alternative approach to verbal learning; an approach that placed much
more emphasis on the activity of the learner in organizing material.
• This approach was typified by two investigators who had grown up in
Europe but had then emigrated and been trained in North America:
George Mandler and Endel Tulving.
• In Britain, a third approach to memory had developed, based on Frederic
Bartlett’s (1932) book Remembering.
• Bartlett explicitly rejected the learning of meaningless material as an
appropriate way to study memory, using instead complex material such
HOW CAN WE STUDY MEMORY?
• This approach emphasized the study of the memory errors that people
made, explaining them in terms of the participants’ cultural assumptions
about the world.
• Bartlett proposed that these depended on internal representations that
he referred to as schemas.
• His approach differed radically from the Ebbinghaus tradition that
influenced the verbal learning approach, but left Bartlett with the
problem of how to study these elusive inner representations of the world.
• A possible answer to this problem evolved gradually during World War II
with the development of computers.
• Mathematicians such as Weiner (1950) in the US, and physiologists such
as Gray Walter (1953) in the UK described machines that were able to
demonstrate a degree of control that resembled purposive behavior.
• During the 1940s, a Scottish psychologist, Kenneth Craik (1943),
working with Bartlett in Cambridge produced a brief but influential book
entitled The Nature of Explanation.
• Here he proposed the idea of representing theories as models, and using
the computer to develop such models.
HOW CAN WE STUDY MEMORY?
• He carried out what were probably the first psychological experiments
based on this idea, using analog computers (digital computers were still
being invented) and applying his computer-based theoretical model to
the practical problem of gun-aiming in tanks.
• Tragically, in 1945 he was killed in a traffic accident while still a young
man.
• Fortunately, the new approach to psychology, based on the computer
metaphor, was being taken up by a range of young investigators, and in
the years following the war, this information-processing approach to
psychology became increasingly influential.
• Two books were particularly important.
• Donald Broadbent’s Perception and Communication (1958) developed
and applied Craik’s seminal ideas to a range of work carried out at the
Medical Research Council Applied Psychology Unit in Cambridge,
England, much of it stimulated by practical problems originating during
the war.
• Some nine years later, this growing field was then brilliantly synthesized
and summarized by Ulric Neisser (1967) in a book whose title provided a
HOW CAN WE STUDY MEMORY?
• Any memory system - whether physical, electronic, or human - requires
three things, the capacity to encode, or enter information into the
system, the capacity to store it, and - subsequently - the capacity to find
and retrieve it.
• However, although these three stages serve different functions, they
interact: The method of registering material or encoding determines
what and how the information is stored, which in turn will limit what can
subsequently be retrieved.
• Consider a simple physical memory device, a shopping list.
• If it is to work, you need to write legibly in a language the recipient
shopper understands.
• If it were to get wet, the ink would blur (impaired storage) making it less
distinct and harder to read (retrieval).
• Retrieval would be harder if your handwriting was poor (an encoding-
retrieval interaction), and if the writing was smudged (a storage-retrieval
interaction).
• The situation is further complicated by the discovery that our memories
comprise not one, but several interrelated memory systems.
HOW CAN WE STUDY MEMORY?
• Verbal learning: A term applied to an approach to memory that relies
principally on the learning of lists of words and nonsense syllables.
• Gestalt psychology: An approach to psychology that was strong in
Germany in the 1930s and that attempted to use perceptual principles to
understand memory and reasoning.
• Schema: Proposed by Bartlett to explain how our knowledge of the world
is structured and influences the way in which new information is stored
and subsequently recalled.
• Model: A method of expressing a theory more precisely, allowing
predictions to be made and tested.
HOW CAN WE STUDY MEMORY?
• The
HOW CAN WE STUDY MEMORY?
• Information processing approach / computational approach

• Memory became memory systems - physical, electrical or human

 SENSORY MEMORY
• If you wave your hand while holding a sparkler in a dark room, it leaves a
trail, which rapidly fades.
• The fact that the image persists long enough to draw an apparent line
suggests that it is being stored in some way, and the fact that the line
rapidly fades implies some simple form of forgetting.
• This phenomenon forms the basis for movies; a sequence of static images
is presented rapidly, with blank intervals in between, but is perceived as
a continuous moving image.
• This occurs because the perceptual system stores the visual information
long enough to bridge the gap between the static images, integrating
each one with the next, very slightly different, image.
• Neisser (1967) referred to this brief visual memory system as iconic
memory, referring to its auditory counterpart as “echoic memory.”
• In the early 1960s, a number of investigators at Bell Laboratories in the
US used the new information-processing approach to analyze this
fleeting visual memory system (Averbach & Sperling, 1961; Sperling,
1960, 1963).
• Sperling (1960) briefly presented a visual array of 12 letters in three
 SENSORY MEMORY
• People could typically remember four or five items correctly.
• If you try this task, however, you will have the sensation that you have
seen more than four or five, but that they have gone before you can
report them.
• One way of avoiding the problem of forgetting during reporting is to
present the same array but ask for only one row to be reported, but not
telling the participant in advance which row will be tested, hence
requiring the whole set to be encoded.
• The row to be recalled is then specified by presenting a tone; a high tone
for the top line, a medium tone for line two, and a low tone for line three.
• As Sperling did not tell the participant in advance which line would be
cued, the report could be treated as representative of the whole array;
multiplying the score by three will thus give an estimate of the total
number of letters stored.
• However, as shown in Figure 1.4, amount recalled depends on when the
recall tone is presented.
SENSORY MEMORY
 SENSORY MEMORY
• When recall is tested immediately, it should provide an estimate of the
total capacity of the memory store, with the fall-off in performance as the
tone is delayed representing the loss of information.
• Note that Figure 1.4 shows two curves, one with a bright field before and
after the letters, and the other with the letters preceded and followed by
a dark visual field.
• A subsequent experiment (Sperling, 1963) found that the brighter the
light during the interval, the poorer the performance, suggesting that
the light is interfering with the memory trace in some way, a process
known as masking.
• Later work by Michael Turvey (1973) investigated two separate types of
masking operating at different stages.
• The first of these involves brightness masking, with the degree of
masking increasing when the mask becomes brighter, or is presented
closer in time to the stimulus.
• This effect only occurs if the mask and the stimulus are presented to the
same eye, suggesting that it is operating at a peripheral retinal level.
• If you were a subject in such an experiment, this type of masking would
 SENSORY MEMORY
• This is distinct from pattern masking, the second type studied which
occurs when targets are followed by a mask comprising broadly similar
features to the target, for example jumbled fragments of letters.
• This type of mask operates even when the target is presented to one eye
and the mask to the other.
• This suggests that it influences a later stage of visual processing that
occurs after information from the two eyes has been combined into a
single percept.
• It is relatively insensitive to brightness and subjectively feels as if a clear
image has been disrupted before the information could adequately be
read off from it.
• What function does iconic memory serve other than that of keeping
psychologists busy, or as Haber concluded in desperation, reading at
night in a thunderstorm?
• The answer is that its function is probably indirect, forming part of the
process of perceiving the world.
• As we scan the visual world, stimuli of huge complexity will fall on our
retina, comprising far more information than it is useful for us to
 SENSORY MEMORY
• It seems likely that iconic memory represents two early stages of a
process whereby information is read off from the retina, and some of it
then fed through to a more durable short-term visual store.
• It is this that allows us to build up a coherent representation of the
visual world and that allows a movie to be perceived, not a series of static
frames with gaps in between, but as a continuous and realistic visual
experience.
• The early stages of iconic memory are probably best regarded as aspects
of perception, while the subsequent more stable stage will be discussed
in the chapter on short-term memory.
• The auditory system also involves a brief sensory memory component
that Neisser named echoic memory.
• If you are asked to remember a long telephone number, then your
pattern of errors will differ depending on whether the number is heard or
read.
• With visual presentation, the likelihood of an error increases
systematically from the beginning to the end of the sequence, whereas,
as shown in Figure 1.5, with auditory presentation the last one or two
 SENSORY MEMORY
• This recency advantage can be removed by
interposing another spoken item between
presentation and recall, even when this
item itself does not need to be processed,
and is always the same, for example, the
instruction “recall.”
• In an extensive series of experiments,
Crowder and Morton (1969; Crowder, 1971)
showed that the nature of this suffix is
critical.
• A visual or nonspeech-like auditory suffix,
such as a buzzer, does not disrupt
performance, whereas a spoken suffix does,
regardless of its meaning.
• Crowder and Morton postulated what they
term a precategorical acoustic store as the
basis for the auditory recency effect.
• However, the question of whether the
process responsible for the enhanced
 SENSORY MEMORY
• Regardless of its interpretation, the auditory recency component is
sufficiently large and robust to play a potentially significant role in
studies of verbal short-term memory, and has even been proposed as an
alternative to more conventional views of performance on short-term
verbal memory tasks (Jones et al., 2006).
• We will return to this issue when discussing short-term memory.
• In the meantime, it seems likely that an adequate explanation of echoic
memory will need to be fully integrated with a broader theory of speech
perception.
SHORT-TERM & WORKING MEMORY
• R
 LONG TERM MEMORY
• R
 MEMORY BEYOND THE
• We haveLABORATORY
so far discussed the question of how to
develop a theoretical understanding of human memory:
How it encodes, stores, and retrieves information.
• However, if our theory is to be useful as well as
informative, then it needs to be applicable beyond the
confines of the laboratory, to tell about how our
memories will work in the world.
• It must aim to extend beyond the student population,
on which much of the research is based, and tell us
about how memory functions in children and the
elderly, across different cultures, and in health and
disease.
• In general, attempts to generalize our theories have
worked well, and have in turn enriched theory.
• One important application of theory is to the memory
performance of particular groups such as children, the
elderly, and patients with memory problems.
• These not only demonstrate the robustness and
MEMORY BEYOND THE LABORATORY
• A good case in point is the study of patients with a very dense but pure
amnesia, which has told us about the everyday importance of episodic
memory, has helped develop tests and rehabilitation techniques for clinical
neuropsychologists, and has, at the same time, had a major impact on our
theories of memory.
• A second major benefit from moving beyond the laboratory comes from a
realization that certain very important aspects of memory were not being
directly covered by existing theories.
• This is the case with the study of semantic memory which, as mentioned
earlier, was initially prompted by the attempt of computer scientists to
develop programs that could understand language.
• Another area of very active research that was driven by a practical need is
that of eyewitness testimony, where it became clear that the failures of the
judiciary to understand the limitations of human memory were often leading
to potentially very serious miscarriages of justice (Loftus, 1979).
• Other areas have developed as a result of identifying practical problems that
have failed to be addressed by theory.
• A good example of this is prospective memory, remembering to do things.
• This use of memory is of great practical importance, but for many years was
 THE CONTRIBUTION OF
• NEUROSCIENCE
Both the Ebbinghaus and Bartlett approaches to the study of memory
were based on the psychological study of memory performance in normal
individuals.
• In recent years, however, this approach has increasingly been enriched
by data from neuroscience, looking at the contribution of the brain to
our capacity to learn and remember.
• We will come across cases in which the study of memory disorders in
patients has thrown light on the normal functioning of human memory.
• In particular, the problems faced by patients with memory problems can
often tell us about the function that our memories serve, and how they
can be further investigated.
• Recent years have seen a rapid development of methods that allow the
neuroscientist to observe and record the operation of the brain in
healthy people both at rest and while performing complex activities,
including those involved in learning and remembering.
 SUMMARY
• Although we complain about our memories, they are remarkably efficient
and flexible in storing the information we need and discarding what is
less important.
• Many of our memory lapses result from this important need to forget
nonessentials, if we are to remember efficiently.
• The study of memory began with Ebbinghaus, who greatly simplified the
experimental situation, creating a carefully constrained approach that
continued in North America into the 20th century.
• Alternative traditions developed in Germany, where the study of
perception influenced the way in which Gestalt psychologists thought
about memory, and in Britain, where Bartlett used a richer and more
open approach to memory.
• During the 1950s and 1960s, these ideas, influenced further by the
development of the computer, resulted in an approach that became
known as cognitive psychology.
• In the case of memory, this emphasized the need to distinguish between
encoding or input into memory, memory storage, and memory retrieval,
and to the proposal to divide memory into three broad types, sensory
 SUMMARY
• These were assumed to lead into a temporary short-term or working
memory. This was initially thought to be largely verbal in nature but
other modalities were subsequently shown to be capable of temporary
storage.
• The short-term memory system was assumed to feed information into
and out of long-term memory.
• Long-term memory was further subdivided into explicit or declarative
memory, and implicit or nondeclarative memory.
• Explicit memory was further divided into two types: The capacity to
recollect individual experiences, allowing “mental time travel,” became
known as episodic memory, whereas our stored knowledge of the world
was termed semantic memory.
• A range of implicit or nondeclarative learning and memory systems were
identified, including classical conditioning, the acquisition of motor
skills, and various types of priming.
• An important development in recent years has been the increased
interest in extending theory beyond the laboratory.
• However, this has led to controversy - it is clear that we need the