What Is Working Memory Good For?

By

Douglas B. Potts

© 1996 Douglas Potts. All rights reserved.

No portion of this document may be reproduced without the expressed written consent of the author.
What is Working Memory Good For?

It has only been within the last 30 years that the label Working Memory has been applied to what
was previously called Short-term Memory (STM). In Atkinson and Shiffrin’s 1968 model, they
treated the short-term store as a working memory system. As A. Baddeley i referred to their
definition, it is “…a system for temporarily holding and manipulating information as part of a
wide range of essential cognitive tasks such as learning, reasoning and comprehending.”

Since then, researchers such as Baddeley, R. Logie and G. Hitch have attempted to define the
structure of working memory, in an attempt to explain its function. This paper will show the
results of some of their work, as well as the work of others, in the search for an explanation as to
the role of working memory. The assumption that a working memory system exists, however,
must obviously be made in order to answer the question.

A presentation on the role of working memory would be incomplete without a discussion of

Baddeley’s model, shown below. Understanding the components of the model allows us to
understand the role of working memory. This model was developed from several experiments

Se e ing He aring

W orking Memory

Ce ntral Exe cutive

Vis uo-s patial s ke tch pad Pho no logical lo op

Knowle dge Bas e

that tested the idea of a working memory.

Page 2
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
In developing the model, one consideration was whether working memory performs a single
function or several functions. If working memory is a single unit, capable of performing only a
single function, then dual-tasks could not be performed. Baddeley addressed this question with a
test designed 10 years earlier by L. R. Brooksiiiii, which has come to be known as the Brooks
verbal task.

The response times of the spatial task (identifying corners of a letter as either inside or outside
corners) compared to those of the verbal task (read a proverb and respond ‘yes’ or ‘no’ if the
word being read is a noun) shows a double dissociation between the two tasks. The spatial task
shows better reaction time when the response is verbal and the verbal task shows a better reaction
time when the response is non-verbal.

These results support the idea that a working memory system is actually made up of at least two
separate sub-systems: Baddeley’s model refers to these as the visuo-spatial sketch pad and the
phonological loop. Attention to these sub-systems and transfer of information between them, as
well as to and from long-term memory must be executed by a third sub-system: the central
executive. In fact, Baddeley’s model shows that the central executive controls both the
phonological loop and visuo-spatial sketch pad. Hence they are referred to as slaves to the
central executive. Now that its components have been defined, we can examine the role of
working memory by looking at the role of each of its components in cognitive tasks.

The Central Executive

Arguably the least understood, the central executive is theorized to be involved in a variety of
functions. S. Gathercole and A. Baddeleyiv believe “Some of its primary functions are regulatory
in nature: It coordinates activity within working memory and controls the transmission of
information between other parts of the cognitive system.” They suggest that the central
executive has a limited capacity, and thus tasks that seem to deal specifically with either of the
slave systems require processing by the central executive.

Additionally, they propose that some cognitive tasks suggested to involve the central executive
include mental arithmetic and the recollection of events from long-term memory v, logical
reasoning and recall of lengthy lists of digits vi, random letter generationvii, and semantic
verificationviii.

Page 3
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
Norman and Shallice’s worksix,x,xi on the central executive include a model of the attentional
control of human action, referred to as the Supervisory Attentional System. This model shows
that a lot of automatic type processes are handled with schema or “templates” of how the process
should occur. Some examples may be eating at a restaurant, being stopped by a police car, or
buying a suit.

i
Baddeley, A., Human Memory: theory and practice, 1990, Boston: Allyn and Bacon
ii
Brooks, L., (1967), The suppression of visualisation by reading., Quarterly Journal of Experimental Psychology,
19, 289-299
iii
Brooks, L., (1968), Spatial and verbal components in the act of recall., Canadian Journal of Psychology, 22, 349-
368
iv
Gathercole, S., & Baddeley, A., Working Memory and Language, Essays in Cognitive Psychology 1993, Hove
(UK): Lawrence Erlbaum Associates
v
Hitch, G., (1980), Developing the concept of working memory. In G. Claxton (Ed.), Cognitive psychology: New
directions (pp. 156-196), London: Routledge & Kegan Paul
vi
Baddeley, A., & Hitch, G., (1974), Working memory., In G. Bower (Ed.), The psychology of learning and
motivation, (Vol. 8, pp. 47-90), New York: Academic Press
vii?
Baddeley, A., (1966), The capacity for generating information by randomisation, Quarterly Journal of
Experimental Psychology, 18, 119-129
viii
Baddeley, A., Lewis, V., Eldridge, M., & Thomson, N., (1984), Attention and retrieval from long-term memory.
Journal of Experimental Psychology: General, 113, 518-530
ix
Norman, D., & Shallice, T., (1980), Attention to action: Willed and automatic control of behavior. University of
California, San Diego, CHIP Report 99
x
Norman, D., & Shallice, T., (1986), Attention to action: Willed and automatic control of behavior. In R.J.
Davidson, G.E. Schwarts, & D. Shapiro (Eds.), Consciousness and self-regulation. Advances in research and theory,
Vol. 4, pp. 1-18, New Your: Plenum Press.
xi
Shallice, T., (1988), From neuropsychology to mental structure. Cambridge: Cambridge University Press

Page 4
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
The contention scheduler handles conflicts between schemas, based upon environmental triggers
(stimuli). The supervisory attentional system deals with new activities or items that require non-
schema-based changes in current processing loads. Baddeley indicates the possibility that the
supervisory attentional system may be the central executive, as it fits well with his current model.
However he points out that “It would be misleading to suggest that the central executive is as yet
uniquely identified with a single mechanism or model such as the Supervisory Attentional
System.”xii

The Role of the Central Executive

What role does the central executive play in working memory? Baddeley’s work, combined with
that of Norman and Shallice assign many roles to the central executive. They include:

1. Allocation of attention.

2. Controlling communication between the phonological loop and the visuo-spatial sketch
pad.

3. Controlling communication between working memory and other memory stores.

4. Generation of random numbers and letters.

5. Mental arithmetic.

6. Logical reasoning.

Phonological Loop

xii
Gathercole, S., & Baddeley, A., Working Memory and Language, Essays in Cognitive Psychology 1993, Hove
(UK): Lawrence Erlbaum Associates

Page 5
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
This portion of working memory is far better understood than the central executive. Gathercole
and Baddeley12 indicate that 20 years worth of collected evidence supports the idea that the
phonological loop is actually two sub-systems, each with a specialized function. Speech inputs
go
directly
into the
P h o n o lo g ic a l s h o rt- Subvo c a l
te rm s to re re he a rs a l No n-s pe e c h
Inputs

Spe e c h Inputs

phonol
ogical store, while non-speech inputs go into the subvocal rehearsal process. The rehearsal
process is used to restore decaying information in the phonological store, and to change non-
speech inputs (such as drawings, pictures or printed words) into their “phonological form.” The
results of

1. articulatory suppression,

2. word length,

3. phonological similarity,

4. irrelevant speech experiments

all support the two-component theory of the phonological loop. Although simple in design, the
model can explain several experimental findings, indicated below.

Articulatory suppression is the method of having a subject continually speak irrelevant words
while being presented with words from the memory list xiii. By speaking the irrelevant word(s),
the subvocal rehearsal process is interrupted, the words in the phonological store cannot be
refreshed, and are thus forgotten. This supports the idea of the subvocal rehearsal process.

Word length experiments test subjects ability to recall lists of words, varying the length of the
words in the list. Baddeley, Thomson and Buchanan xiv show that subjects remember fewer
xiii
Estes, W., (1973), Phonemic coding and rehearsal in short-term memory. Journal of Verbal Learning and Verbal
Behavior, 12, 360-372

Page 6
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
longer words than shorter words. This supports the idea of storage of phonemes in phonological
store, and that the processing of words is time-dependent. This is not part of the 7 2 chunks of
information which can be retained in working memory.

Subjects in irrelevant speech experiments constantly hear (or see) irrelevant words during
presentation of memory lists. Subjects recall is disrupted xv by the irrelevant words. Baddeley’s
model shows that at some point, all spoken or seen words end up in the phonological store.
Thus, the spoken and seen words exceed the store’s capacity causing the disruption of recall.
Salamé and Baddeleyxvi demonstrated that the degree to which recall was affected is directly
proportional to “the degree of phonological similarity between the irrelevant material and the
memory items.” In both cases, these experiments support the idea of a phonological store

In free recall tests, recall of words that sound the same (phonological similarity) is much poorer
than if the words are dissimilar in sound. Baddeley xvii showed that this effect does not occur for
semantically similar words. Gathercole and Baddeley indicate that “The phenomenon has been
found over many studies to be robust.” It appears that the phonological store, stores phonemes,
and traces for similar sounding words are interfered with.

What role does the phonological loop play in working memory? Baddeley’s model, and
subsequent research relating to the phonological loop, suggest at least 3 functions:

1. Providing temporary storage of phonological information for about two seconds

2. Refreshing material in the phonological store as needed

3. Processing non-verbal material into verbal form through subvocal rehearsal.

Visuo-spatial sketch pad

xiv
Baddeley, A., Thomson, N., & Buchanan, M., (1975), Word length and the structure of short-term memory.
Journal of Verbal Learning and Verbal Behavior, 14, 575-589
xv
Colle, H., & Welsh, A., (1976), Acoustic masking in primary memory. Journal of Verbal Learning and Verbal
Behavior, 15, 17-32
xvi
Salamé, P., & Baddeley, A., (1986), Phonological factors in STM: Similarity and the unattended speech effect.
Bulletin of the Psychonomic Society, 24, 263-265
xvii
Baddeley, A., (1966), Short-term memory for word sequences as a function of acoustic, semantic and formal
similarity. Quarterly Journal of Experimental Psychology, 18, 362-365

Page 7
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
Baddeley et al.xviii tested this aspect of the working memory model using the Brooks matrix task
with both spatial and nonsense instructions. During memorization of the instructions, he had half
of the subjects perform visuo-spatial tracking tasks. Without the tracking tasks, subjects did
equally well with both the nonsense and spatial instructions. However, when performing the
visuo-spatial tracking task, subjects’ performance with the spatial instructions was extremely
poor. They showed a slight degradation in performance with the nonsense instructions. This
supports the idea that there is a component in working memory dedicated to processing and
refreshing visuo-spatial material.

But is it also involved in the generation of images? Baddeley and Lieberman xix used pegword
mnemonics (association of a number with a phonologically similar word; one - gun, two - stew,
etc.) for generation of visual imagery. Given a list of words, the subjects had to visualize an
image which contained the word (in object form) and the pegword representing the number of the
word in the list. If the first word is chair, then the subject was to imagine a chair with a gun, etc.
Other subjects were simply asked to rote memorize the list.

When subjects performed spatial tracking tasks during recall, those subjects using the pegword
mnemonics did much poorer than those who did rote learning of the list. This supports the
notion that the visuo-spatial sketch pad is used in the generation of images, as well as storage and
refreshing.

Robert Logiexx suggests that the visuo-spatial sketch pad has sub-systems similar to the
phonological loop. These are a visual store or cache, and an inner scribe. Just as the
phonological store is subject to decay, so is the visual cache. The inner scribe functions similarly
to that of the subvocal rehearsal, refreshing the image in the visual cache as needed. The
similarity ends there, as the inner scribe is also responsible for the generation of images, as
opposed to the subvocal rehearsal process of conversion of images to phonological form.

xviii
Baddeley, A., Grant, S., Wight, E., & Thomson, N. (1975), Imagery and visual working memory. In P.M.A.
Rabbit & S. Dornic (Eds.), Attention and performance, V, pp. 205-217. London: Academic Press. New York:
Academic Press
xix
Baddeley, A., & Lieberman, K., (1980), Spatial working memory. In R. Nickerson (Ed.), Attention and
performance, VIII, pp. 521-539. Hillsdale, NJ: Lawrence Erlbaum Associates
xx
Logie, R., (1995), Visuo-spatial working memory, Essays in Cognitive Psychology, pp. 126-131, Hove (UK):
Lawrence Erlbaum Associates.

Page 8
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
What roles does the visuo-spatial sketch pad play in working memory? Baddeley’s and Logie’s
suggestions include:

1. Providing temporary storage of visual and spatial information

2. Refreshing images in the sketch pad as needed

3. Generating images

An example of working memory functions

The following is a description of the roles of working memory in a fairly typical situation.
Portions of the situation are presented, followed by an explanation of what happens in working
memory. The actions of working memory are shown in italics.

You are sitting at your desk reading a report.

Phonological store and subvocal rehearsal to facilitate text comprehension. Use of a reading
schema (specifically a report reading schema) allows for some automation in reading.

The phone rings. You pick up the receiver and speak.

Attentional control is moved from visual to aural. Use of a phone answering schema to
answer the phone. Transfer of knowledge (what to say on the phone) from long-term memory
to the central executive, which transmits the information to subvocal rehearsal.

“Hello. This is Brian”, you say into the receiver. “Hi Brian, this is Sally.” In the background,
you hear a familiar sound.

Phonemes are stored in phonological store as they are heard. Subvocal rehearsal occurs to
allow processing of complete sentences. The sound of the ocean is also stored. The central
executive activates traces in long-term memory looking for people named Sally. If the voice
is recognized then this too is used for identification. At the same time, the central executive
activates traces in long-term memory attempting to identify the sound. “Sally” is recalled
from memory. The sound is identified as the ocean. The inner scribe generates an image of
Sally at the beach.

Page 9
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
This is just a brief sample of what working memory does. However, you can appreciate its uses
in everyday life.

Summary

This paper focused on current theories in psychology which suggest the roles of working
memory. The assumption was made that a working memory does exist. Baddeley’s model of
working memory was then presented, followed by a discussion of the function and roles of each
component of the model. These were:

Central executive

 responsible for controlling transmission of material between the phonological loop and
the visuo-spatial sketch pad.

 controls transmission between working memory and other memory stores.

 Involved in mental arithmetic, logical reasoning, and generation of random letters and
numbers.

Phonological loop

 stores phonological information for up to two seconds in the phonological store.

 refreshes the material as needed

 converts non-speech input into phonological form with the subvocal rehearsal process

Visuo-spatial sketch pad

 stores imagery in the visual cache

 refreshes imagery as needed in the visual cache

 generates images.

Page 10
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.
As a side note, it should be pointed out that an additional difference between Baddeley’s model
of working memory and the modified version presented by Logiexxi differ in the path by which
external stimuli reaches working memory. Logie proposes that stimuli reaches working memory
via the knowledge store. Baddeley’s model shows stimuli going directly to the phonological
loop and visuo-spatial sketch pad. Currently, I do not see how this would effect the role of
working memory. It appears to remain functionally the same in both models. However, further
study may show otherwise.

xxi
Logie, R., (1995), Visuo-spatial working memory, Essays in Cognitive Psychology, pp. 126-131, Hove (UK):
Lawrence Erlbaum Associates.

Page 11
© 1996 Douglas Potts. All rights reserved.
No portion of this document may be reproduced without the expressed written consent of the author.