W.S. Sossin, J.-C. Lacaille, V.F. Castellucci & S. Belleville (Eds.

)
Progress in Brain Research, Vol. 169
ISSN 0079-6123
Copyright r 2008 Elsevier B.V. All rights reserved

CHAPTER 20

What are the differences between long-term,

short-term, and working memory?

Nelson Cowan�

Department of Psychological Sciences, University of Missouri, 18 McAlester Hall, Columbia, MO 65211, USA

Abstract: In the recent literature there has been considerable confusion about the three types of memory:
long-term, short-term, and working memory. This chapter strives to reduce that confusion and makes up-
to-date assessments of these types of memory. Long- and short-term memory could differ in two
fundamental ways, with only short-term memory demonstrating (1) temporal decay and (2) chunk capacity
limits. Both properties of short-term memory are still controversial but the current literature is rather
encouraging regarding the existence of both decay and capacity limits. Working memory has been
conceived and defined in three different, slightly discrepant ways: as short-term memory applied to
cognitive tasks, as a multi-component system that holds and manipulates information in short-term
memory, and as the use of attention to manage short-term memory. Regardless of the definition, there are
some measures of memory in the short term that seem routine and do not correlate well with cognitive
aptitudes and other measures (those usually identified with the term ‘‘working memory’’) that seem more
attention demanding and do correlate well with these aptitudes. The evidence is evaluated and placed
within a theoretical framework depicted in Fig. 1.

Keywords: attention; capacity of working memory; control of attention; decay of short-term memory; focus
of attention; long-term memory; short-term memory; working memory

Historical roots of a basic scientific question and yet it might be evident that his ability to
capture the names of new students, or to recall
How many phases of a memory are there? In a which student made what comment in an ongoing
naı̈ve view of memory, it could be made all of one conversation, has diminished over the years.
cloth. Some people have a good ability to capture The scientific study of memory is usually traced
facts and events in memory, whereas others have back to Hermann Ebbinghaus (1885/1913 transla-
less such ability. Yet, long before there were true tion), who examined his own acquisition and
psychological laboratories, a more careful obser- forgetting of new information in the form of series
vation must have shown that there are separable of nonsense syllables tested at various periods upto
aspects of memory. An elderly teacher might be 31 days. Among many important observations,
seen relating old lessons as vividly as he ever did, Ebbinghaus noticed that he often had a ‘‘first
fleeting grasp y of the series in moments of
special concentration’’ (p. 33) but that this
�Corresponding author. Tel.: +1 573-882-4232; immediate memory did not ensure that the series
Fax: +1 573-882-7710; E-mail: CowanN@missouri.edu had been memorized in a way that would allow its

DOI: 10.1016/S0079-6123(07)00020-9 323

324

recall later on. Stable memorization sometimes memory but rather of interference during the
required further repetitions of the series. Soon retention interval; one could find situations in
afterward, James (1890) proposed a distinction which memory improved, rather than diminish,
between primary memory, the small amount of over time. From this perspective, one might view
information held as the trailing edge of the what appeared to be forgetting from primary
conscious present, and secondary memory, the vast memory as the profound effect of interference
body of knowledge stored over a lifetime. The from other items on memory for any one item,
primary memory of James is like the first fleeting with interference effects continuing forever but not
grasp of Ebbinghaus. totally destroying a given memory. This perspective
The Industrial Revolution made some new has been maintained and developed over the years
demands on what James (1890) called primary by a steady line of researchers believing in the unity
memory. In the 1850s, telegraph operators had to of memory, including, among others, Melton (1963),
remember and interpret rapid series of dots and Bjork and Whitten (1974), Wickelgren (1974),
dashes conveyed acoustically. In 1876, the tele- Crowder (1982, 1993), Glenberg and Swanson
phone was invented. Three years later, operators in (1986), Brown et al. (2000), Nairne (2002), Neath
Lowell, Massachusetts started using telephone and Surprenant (2003), and Lewandowsky et al.
numbers for more than 200 subscribers so that (2004).
substitute operators could be more easily trained if
the town’s four regular operators succumbed to a
raging measles epidemic. This use of telephone Description of three kinds of memory
numbers, complemented by a word prefix, of
course spread. (The author’s telephone number In this chapter I will assess the strength of evidence
in 1957 was WHitehall 2-6742; the number is still for three types of memory: long-term memory,
assigned, albeit as a seven-digit number.) Even short-term memory, and working memory. Long-
before the book by Ebbinghaus, Nipher (1878) term memory is a vast store of knowledge and a
reported on the serial position curve obtained record of prior events, and it exists according to all
among the digits in logarithms that he tried to theoretical views; it would be difficult to deny that
recall. The nonsense syllables that Ebbinghaus each normal person has at his or her command a
had invented as a tool can be seen to have rich, although not flawless or complete, set of long-
acquired more ecological validity in an industrial term memories.
age with expanding information demands, perhaps Short-term memory is related to the primary
highlighting the practical importance of primary memory of James (1890) and is a term that
memory in daily life. Primary memory seems Broadbent (1958) and Atkinson and Shiffrin
taxed as one is asked to keep in mind aspects of (1968) used in slightly different ways. Like
an unfamiliar situation, such as names, places, Atkinson and Shiffrin, I take it to reflect faculties
things, and ideas that one has not encountered of the human mind that can hold a limited amount
before. of information in a very accessible state tempora-
Yet, the subjective experience of a difference rily. One difference between the term ‘‘short-term
between primary and secondary memory does not memory’’ and the term ‘‘primary memory’’ is that
automatically guarantee that these types of memo- the latter might be considered to be more
ry separately contribute to the science of remem- restricted. It is possible that not every temporarily
bering. Researchers from a different perspective accessible idea is, or even was, in conscious
have long hoped that they could write a single awareness. For example, by this conception, if you
equation, or a single set of principles at least, that are speaking to a person with a foreign accent and
would capture all of memory, from the very inadvertently alter your speech to match the
immediate to the very long-term. McGeoch foreign speaker’s accent, you are influenced by
(1932) illustrated that forgetting over time was what was until that point an unconscious (and
not simply a matter of an inevitable decay of therefore uncontrollable) aspect of your short-term
 325

memory. One might relate short-term memory to a by organizing and grouping information in work-
pattern of neural firing that represents a particular ing memory into a smaller number of units (Miller,
idea and one might consider the idea to be in short- 1956; Ericsson and Kintsch, 1995). For example,
term memory only when the firing pattern, or cell the letter series IRSCIAFBI can be remembered
assembly, is active (Hebb, 1949). The individual much more easily as a series of acronyms for three
might or might not be aware of the idea during federal agencies of the United States of America:
that period of activation. the Internal Revenue Service (IRS), the Central
Working memory is not completely distinct from Intelligence Agency (CIA), and the Federal Bureau
short-term memory. It is a term that was used by of Investigation (FBI). However, that factor was
Miller et al. (1960) to refer to memory as it is used not emphasized in the well-known model of
to plan and carry out behavior. One relies on Baddeley (1986).
working memory to retain the partial results while What is clear from my definition is that working
solving an arithmetic problem without paper, to memory includes short-term memory and other
combine the premises in a lengthy rhetorical processing mechanisms that help to make use of
argument, or to bake a cake without making the short-term memory. This definition is different
unfortunate mistake of adding the same ingredient from the one used by some other researchers (e.g.,
twice. (Your working memory would have been Engle, 2002), who would like to reserve the term
more heavily taxed while reading the previous working memory to refer only to the attention-
sentence if I had saved the phrase ‘‘one relies on related aspects of short-term memory. This,
working memory’’ until the end of the sentence, however, is not so much a debate about substance,
which I did in within my first draft of that but rather a slightly confusing discrepancy in the
sentence; working memory thus affects good usage of terms.
writing.) The term ‘‘working memory’’ became One reason to pursue the term working memory
much more dominant in the field after Baddeley is that measures of working memory have been
and Hitch (1974) demonstrated that a single found to correlate with intellectual aptitudes (and
module could not account for all kinds of especially fluid intelligence) better than measures
temporary memory. Their thinking led to an of short-term memory and, in fact, possibly better
influential model (Baddeley, 1986) in which than measures of any other particular psychologi-
verbal-phonological and visual-spatial representa- cal process (e.g., Daneman and Carpenter, 1980;
tions were held separately, and were managed and Kyllonen and Christal, 1990; Daneman and
manipulated with the help of attention-related Merikle, 1996; Engle et al., 1999; Conway et al.,
processes, termed the central executive. In the 1974 2005). It has been thought that this reflects the use
paper, this central executive possibly had its own of measures that incorporate not only storage but
memory that crossed domains of representation. also processing, the notion being that both storage
By 1986, this general memory had been eliminated and processing have to be engaged concurrently to
from the model, but it was added back again by assess working memory capacity in a way that is
Baddeley (2000) in the form of an episodic buffer. related to cognitive aptitude. More recently, Engle
That seemed necessary to explain short-term et al. (1999) introduced the notion that aptitudes
memory of features that did not match the other and working memory both depend on the ability
stores (particularly semantic information in mem- to control attention, or to apply the control of
ory) and to explain cross-domain associations in attention to the management of both primary and
working memory, such as the retention of links secondary memory (Unsworth and Engle, 2007).
between names and faces. Because of the work of However, more research is needed on exactly what
Baddeley et al. (1975), working memory is we learn from the high correlation between work-
generally viewed as the combination of multiple ing memory and intellectual aptitudes, and this
components working together. Some even include issue will be discussed further after the more basic
in that bundle the heavy contribution of long-term issue of the short-term versus the long-term
memory, which reduces the working memory load memory distinction is addressed.
326

Meanwhile, it may be helpful to summarize a which these stores may differ: in duration, and in
theoretical framework (Cowan, 1988, 1995, 1999, capacity. A duration difference means that items in
2001, 2005) based on past research. This frame- short-term storage decay from this sort of storage
work, illustrated in Fig. 1, helps to account for the as a function of time. A capacity difference means
relation between long-term, short-term, and work- that there is a limit in how many items short-term
ing memory mechanisms and explains what I see as storage can hold. If there is only a limit in
the relation between them. In this framework, capacity, a number of items smaller than the
short-term memory is derived from a temporarily capacity limit could remain in short-term storage
activated subset of information in long-term mem- until they are replaced by other items. Both types
ory. This activated subset may decay as a function of limit are controversial. Therefore, in order to
of time unless it is refreshed, although the evidence assess the usefulness of the short-term storage
for decay is still tentative at best. A subset of the concept, duration and capacity limits will be
activated information is the focus of attention, assessed in turn.
which appears to be limited in chunk capacity (how
many separate items can be included at once). New
associations between activated elements can form Duration limits
the focus of attention. Now the evidence related to
this modeling framework will be discussed. The concept of short-term memory limited by
decay over time was present even at the beginning
of cognitive psychology, for example in the work
of Broadbent (1958). If decay were the only
The short-term memory/long-term memory principle affecting performance in an immediate
distinction memory experiment, it would perhaps be easy to
detect this decay. However, even in Broadbent’s
If there is a difference between short- and long- work contaminating variables were recognized. To
term memory stores, there are two possible ways in assess decay one must take into account, or

Fig. 1. A depiction of the theoretical modeling framework. Modified from Cowan (1988) and refined in further work by Cowan (1995,
1999, 2005).
 327

overcome, contaminating effects of rehearsal, suggested that the process of mentally attending
long-term retrieval, and temporal distinctiveness, to words or searching through the list, an
which will be discussed one at a time in conjunc- attention-demanding process, could serve to reac-
tion with evidence for and against decay. tivate items to be recalled in a manner similar to
covert verbal rehearsal. The key difference is that it
would not be expected that articulatory suppres-
Overcoming contamination from rehearsal sion would prevent that type of rehearsal. Instead,
to prevent that type of rehearsal an attention-
According to various researchers there is a process demanding task would have to be used.
whereby one imagines how the words on the list Barrouillet et al. (2004, 2007) have results that
are pronounced without saying them aloud, a do seem to suggest that there is another, more
process called covert verbal rehearsal. With attention-demanding type of rehearsal. They have
practice, this process comes to occur with a interposed materials between items to be recalled
minimum of attention. Guttentag (1984) used a that require choices; they can be numbers to read
secondary task to show that rehearsal of a list to aloud or multi-choice reaction times. It is found
be recalled was effortful in young children, but not that these interfere with retention to an extent
in adults. If, in a particular experimental proce- commensurate to the proportion of the inter-item
dure, no loss of short-term memory is observed, interval used up attending to the distracting items.
one can attribute that response pattern to rehear- As the rate of the distracting items goes up, fewer
sal. Therefore, steps have been taken to eliminate of the to-be-recalled items are recalled. The notion
rehearsal through a process termed articulatory is that when the distracting task does not require
suppression, in which a simple utterance such as attention, the freed-up attention allows an atten-
the word ‘‘the’’ is repeatedly pronounced by the tion-based rehearsal of the items to be recalled.
participant during part or all of the short-term When the interposed task is more automatic and
memory task (e.g., Baddeley et al., 1975). There is does not require as much attention (e.g., an
still the possible objection that whatever utterance articulatory suppression task) there is much less
is used to suppress rehearsal unfortunately causes effect of the rate of these interposed items.
interference, which could be the true reason for Based on this logic, one could imagine a version
memory loss over time instead of decay. of Lewandowsky’s task in which not articulatory
That problem of interference would appear suppression but attention-demanding verbal sti-
moot in light of the findings of Lewandowsky et muli are placed between items in the response, and
al. (2004). They presented lists of letters to be in which the duration of this filled time between
recalled and varied how long the participant was items in the response varies from trial to trial. The
supposed to take to recall each item in the list. In verbal, attention-demanding stimuli should pre-
some conditions, they added articulatory suppres- vent both attention-based rehearsal and articula-
sion to prevent rehearsal. Despite that suppres- tion-based rehearsal. If there is decay, then
sion, they observed no difference in performance performance should decline across serial positions
with the time between items in the response more severely when longer filled intervals are
varying between 400 and 1600 ms (or between placed between items in the response. Unfortu-
conditions in which the word ‘‘super’’ was nately, though, such results might be accounted for
pronounced one, two, or three times between alternatively as the result of interference from the
consecutive items in the response). They found no distracting stimuli, without the need to invoke
evidence of memory decay. decay.
A limitation of this finding, though, is that What seems to be needed, then, is a procedure to
covert verbal rehearsal may not be the only type of prevent both articulation-based and attention-
rehearsal that participants can use. Perhaps there based rehearsal without introducing interference.
are types that are not prevented by articulatory Cowan and Aubuchon (in press) tried out one type
suppression. In particular, Cowan (1992) of procedure that may accomplish this. They
328

presented lists of seven printed digits in which the out to be only slightly different from the partici-
time between items varied within a list. In addition pant’s telephone number, the participant might be
to some randomly timed filler lists, there were four able to memorize the new number quickly and
critical trial types, in which the six inter-digit blank repeat it from long-term memory. The dual-store
intervals were all short (0.5 s following each item) theories of memory allow this. Although Broad-
or all long (2 s following each item), or comprised bent (1958) and Atkinson and Shiffrin (1968) drew
three short and then three long intervals, or three their models of information processing as a series
long and then three short intervals. Moreover, of boxes representing different memory stores,
there were two post-list response cues. According with long-term memory following short-term
to one cue, the participant was to recall the list memory, these boxes do not imply that memory
with the items in the presented order, but at any is exclusively in one box or another; they are better
rate they wished. According to the other response interpreted as the relative times of the first entry of
cue, the list was to be recalled using the same information from a stimulus into one store and
timing in which it was presented. The expectation then the next. The question remains, then as to
was that the need to remember the timing in the how one can determine if a response comes from
latter response condition would prevent rehearsal short-term memory.
of either type. As a consequence, performance Waugh and Norman (1965) developed a mathe-
should be impaired on trials in which the first three matical model to accomplish this. The model
response intervals are long because, on these trials, operated with the assumption that long-term
there is more time for forgetting of most of the list memory occurs for the entire list, including a
items. Just as predicted, there was a significant plateau in the middle of the list. In contrast, by the
interaction between the response cue and the time of recall, short-term memory is said to remain
length of the first half of the response intervals. only at the end of the list. This model assumes
When participants were free to recall items at their that, for any particular serial position within a list,
own pace, performance was no better with a short the likelihood of successful short-term storage (S)
first half (M = .71) than with a long first half and long-term storage (L) are independent, so
(M = .74). The slight benefit of a long first half in that the likelihood of recalling the item is
that situation could occur because it allowed the S + L � SL.
list to be rehearsed early on in the response. In A slightly different assumption is that short- and
contrast, when the timing of recall had to match long-term stores are not independent but are used
the timing of the list presentation, performance in a complementary fashion. The availability of
was better with a short first half (M = .70) than short-term memory of an item may allow resources
with a long first half (M = .67). This, then, needed for long-term memorization to be shifted
suggests there could be decay in short-term to elsewhere in the list. The data seem more
memory. consistent with that assumption. In several studies,
lists to be recalled have been presented to patients
with Korsakoff’s amnesia and normal control
Overcoming contamination from long-term retrieval participants (Baddeley and Warrington, 1970;
Carlesimo et al., 1995). These studies show that,
If there is more than one type of memory storage in immediate recall, performance in amnesic
then there still is the problem of which store individuals is preserved at the last few serial
provided the information underlying a response. positions of the list. It is as if the performance
There is no guarantee that, just because a in those serial positions is based mostly or entirely
procedure is considered a test of short-term on short-term storage, and that there is no
storage, the long-term store will not be used. For decrease in that kind of storage in the amnesic
example, in a simple digit span task, a series of patients. In delayed recall, the amnesic patients
digits is presented and is to be repeated immedi- show a deficit at all serial positions, as one would
ately afterward from memory. If that series turned expect if short-term memory for the end of the list
 329

is lost as a function of a filled delay period Yet, there may be a true decay effect at shorter
(Glanzer and Cunitz, 1966). test intervals. Baddeley and Scott (1971) set up a
trailer in a shopping mall so that they could test a
large number of participants for one trial each, so
Overcoming contamination from temporal as to avoid proactive interference. They found an
distinctiveness effect of the test delay within the first 5 s but not at
longer delays. Still, it seems that the concept of
Last, it has been argued that the loss of memory decay is not yet on very firm ground and warrants
over time is not necessarily the result of decay. further study. It may be that decay actually reflects
Instead, it can be caused by temporal distinctive- not a gradual degradation of the quality of the
ness in retrieval. This kind of theory assumes that short-term memory trace, but a sudden collapse at
the temporal context of an item serves as a a point that varies from trial to trial. With a
retrieval cue for that item, even in free recall. An control for temporal distinctiveness, Cowan et al.
item separated in time from all other items is (1997a) found what could be a sudden collapse in
relatively distinctive and easy to recall, whereas an the representation of memory for a tone with
item that is relatively close to other items is more delays of between 5 and 10 s.
difficult to recall because it shares their temporal
cues to retrieval. Shortly after a list is presented the
most recent items are the most distinct temporally Chunk capacity limits
(much like the distinctness of a telephone pole you
are practically touching compared to poles extend- The concept of capacity limits was raised several
ing further down the road). Across a retention times in the history of cognitive psychology. Miller
interval, the relative distinctiveness of the most (1956) famously discussed the ‘‘magical number
recent items decreases (much like standing far seven plus or minus two’’ as a constant in short-
away from even the last pole in a series). term processing, including list recall, absolute
Although there are data that can be interpreted judgment, and numerical estimation experiments.
according to distinctiveness, there also are what However, his autobiographical essay (Miller, 1989)
look like dissociations between the effects of indicates that he was never very serious about the
distinctiveness and a genuine short-term memory number seven; it was a rhetorical device that he
effect. One can see this, for example, in the classic used to tie together the otherwise unrelated strands
procedure of Peterson and Peterson (1959) in of his research for a talk. Although it is true that
which letter trigrams are to be recalled immedi- memory span is approximately seven items in
ately or only after a distracting task, counting adults, there is no guarantee that each item is a
backward from a starting number by three, for a separate entity. Perhaps the most important point
period lasting up to 18 s. Peterson and Peterson of Miller’s (1956) article was that multiple items
found severe memory loss for the letter trigram as can be combined into a larger, meaningful unit.
the filled delay was increased. However, subse- Later studies suggested that the limit in capacity is
quently, sceptics argued that the memory loss more typically only three or four units (Broadbent,
occurred because the temporal distinctiveness of 1975; Cowan, 2001). That conclusion was based on
the current letter trigram diminished as the filled an attempt to take into account strategies that
delay increased. In particular, this delay effect was often increase the efficiency of use of a limited
said to occur because of the increase across test capacity, or that allow the maintenance of addi-
delays in the proactive interference from previous tional information separate from that limited
trials. On the first few trials, the delay does not capacity. To understand these methods of discuss-
matter (Keppel and Underwood, 1962) and no ing capacity limits I will again mention three types
detrimental effect of delay is observed if delays of of contamination. These come from chunking and
5, 10, 15, and 20 s are tested in separate trial blocks the use of long-term memory, from rehearsal, and
(Turvey et al., 1970; Greene, 1996). from non-capacity-limited types of storage.
330

Overcoming contamination from chunking and the recalled an average of four to six chunks. Cowan
use of long-term memory et al. (2004) tried to refine that method by testing
serial recall of eight-word lists, which were
A participant’s response in an immediate-memory composed of four pairs of words that previously
task depends on how the information to be had been associated with various levels of learning
recalled is grouped to form multi-item chunks (0, 1, 2, or 4 prior word–word pairings). Each
(Miller, 1956). Because it is not usually clear what word used in the list was presented an equal
chunks have been used in recall, it is not clear how number of times (four, except in a non-studied
many chunks can be retained and whether the control condition) but what varied was how many
number is truly fixed. Broadbent (1975) proposed of those presentations were as singletons and how
some situations in which multi-item chunk forma- many were as a consistent pairing. The number of
tion was not a factor, and suggested on the basis of paired prior exposures was held constant across
results from such procedures that the true capacity the four pairs in a list. A mathematical model was
limit is three items (each serving as a single-item used to estimate the proportion of recalled pairs
chunk). For example, although memory span is that could be attributed to the learned association
often about seven items, errors are made with (i.e., to a two-word chunk) as opposed to separate
seven-item lists and the error-free limit is typically recall of the two words in a pair. This model
three items. When people must recall items from a suggested that the capacity limit was about 3.5
category in long-term memory, such as the states chunks in every learning condition, but that the
of the United States, they do so in spurts of about ratio of two-word chunks to one-word chunks
three items on average. It is as if the bucket of increased as a function of the number of prior
short-term memory is filled from the well of long- exposures to the pairs in the list.
term memory and must be emptied before it is
refilled. Cowan (2001) noted other such situations
in which multi-item chunks cannot be formed. For Overcoming contamination from rehearsal
example, in running memory span, a long list of
items is presented with an unpredictable endpoint, The issue of rehearsal is not entirely separate from
making grouping impossible. When the list ends, the issue of chunk formation. In the traditional
the participant is to recall a certain number of concept of rehearsal (e.g., Baddeley, 1986), one
items from the end of the list. Typically, people can imagines that the items are covertly articulated in
recall three or four items from the end of the list, the presented order at an even pace. However,
although the exact number depends on task another possibility is that rehearsal involves the
demands (Bunting et al., 2006). Individuals differ use of articulatory processes in order to put the
in capacity, which ranges from about two to six items into groups. In fact, Cowan et al. (2006a)
items in adults (and fewer in children), and the asked participants in a digit span experiment how
individual capacity limit is a strong correlate of they carried out the task and by far the most
cognitive aptitude. common answer among adults was that they
Another way to take into account the role of grouped the items; participants rarely mentioned
multi-item chunk formation is to set up the task in saying the items to themselves. Yet, it is clear that
a manner that allows chunks to be observed. suppressing rehearsal affects performance.
Tulving and Patkau (1962) studied free recall of Presumably, the situations in which items
word lists with various levels of structure, ranging cannot be rehearsed are for the most part the
from random words to well-formed English same as the situations in which items cannot be
sentences, with several different levels of coherence grouped. For example, Cowan et al. (2005) relied
in between. A chunk was defined as a series of on a running memory span procedure in which the
words reproduced by the participant in the same items were presented at the rapid rate of 4 per
order in which the words had been presented. It second. At that rate, it is impossible to rehearse the
was estimated that, in all conditions, participants items as they are presented. Instead, the task is
 331

probably accomplished by retaining a passive store Overcoming contamination from non-capacity-

(sensory or phonological memory) and then limited types of storage
transferring the last few items from that store into
a more attention-related store at the time of recall. It is difficult to demonstrate a true capacity limit
In fact, with a fast presentation rate in running that is related to attention if, as I believe, there are
span, instructions to rehearse the items is detri- other types of short-term memory mechanisms
mental, not helpful, to performance (Hockey, that complicate the results. A general capacity
1973). Another example is memory for lists that should include chunks of information of all sorts:
were ignored at the time of their presentation for example, information derived from both
(Cowan et al., 1999). In these cases, the capacity acoustic and visual stimuli, and from both verbal
limit is close to the three or four items suggested by and nonverbal stimuli. If this is the case, there
Broadbent (1975) and Cowan (2001). should be cross-interference between one type of
It is still quite possible that there is a speech- memory load and another. However, the literature
based short-term storage mechanism that is by and often has shown that there is much more
large independent of the chunk-based mechanism. interference between similar types of memoranda,
In terms of the popular model of Baddeley (2000), such as two visual arrays of objects or two
the former is the phonological loop and the latter, acoustically presented word lists, than there is
the episodic buffer. In terms of Cowan (1988, between two dissimilar types, such as one visual
1995, 1999, 2005), the former is part of activated array and one verbal list. Cocchini et al. (2002)
memory, which may have a time limit due to suggested that there is little or no interference
decay, and the latter is the focus of attention, between dissimilar lists. If so, that would appear to
which is assumed to have a chunk capacity limit. provide an argument against the presence of a
Chen and Cowan (2005) showed that the time general, cross-domain, short-term memory store.
limit and chunk capacity limit in short-term Morey and Cowan (2004, 2005) questioned this
memory are separate. They repeated the proce- conclusion. They presented a visual array of colored
dure of Cowan et al. (2004) in which pairs of spots to be compared to a second array that
words sometimes were presented in a training matched the first or differed from it in one spot’s
session preceding the list recall test. They com- color. Before the first array or just after it,
bined lists composed of pairs as in that study. participants sometimes heard a list of digits that
Now, however, both free and serial recall tasks was then to be recited between the two arrays. In a
were used, and the length of list varied. For long low-load condition, the list was their own seven-
lists and free recall, the chunk capacity limit digit telephone number whereas, in a high-load
governed the recall. For example, lists of six well- condition, it was a random seven-digit number.
learned pairs were recalled as well as lists of six Only the latter condition interfered with array-
unpaired singletons (i.e., were recalled at similar comparison performance, and then only if the list
proportions of words correct). For shorter lists was to be recited aloud between the arrays. This
and serial recall strictly scored, the time limit suggests that retrieving seven random digits in a way
instead governed the recall. For example, lists of that also engages rehearsal processes relies upon
four well-learned pairs were not recalled nearly as some type of short-term memory mechanism that
well as lists of four unpaired singletons, but only also is needed for the visual arrays. That shared
as well as lists of eight unpaired singletons. For mechanism may be the focus of attention, with its
intermediate conditions it appeared as if chunk capacity limit. Apparently, though, if the list was
capacity limits and time limits operate together to maintained silently rather than being recited aloud,
govern recall. Perhaps the capacity-limited this silent maintenance occurred without much use
mechanism holds items and the rehearsal mecha- of the common, attention-based storage mechanism,
nism preserves some serial order memory for so visual array performance was not much affected.
those held items. The exact way in which these The types of short-term memory whose con-
limits work together is not yet clear. tribution to recall may obscure the capacity limit
332

can include any types of activated memory that fall conditions, due to strain on the cross-modality
outside of the focus of attention. In the modeling store. That is how the results turned out. Moreover,
framework depicted in Fig. 1, this can include if the cross-modality, capacity-limited store were
sensory memory features as well as semantic the only type of storage used, then the sum of
features. Sperling (1960) famously illustrated the visual and auditory capacities in the dual-modality
difference between unlimited sensory memory and condition should be no greater than the larger of
capacity-limited categorical memory. If an array of the two unimodal capacities (which happened to
characters was followed by a partial report cue be the visual capacity). The reason is that the
shortly after the array, most of the characters in limited-capacity store would hold the same num-
the cued row could be recalled. If the cue was ber of units no matter whether they were all
delayed about 1 s, most of the sensory information from one modality or were from two modalities
had decayed and performance was limited to about combined. That prediction was confirmed, but
four characters, regardless of the size of the array. only if there was a post-perceptual mask in
Based on this study, the four-character limit could both modalities at once following the array to be
be seen as either a limit in the capacity of short- remembered. The post-perceptual mask included a
term memory or a limit in the rate with which multicolored spot at each visual object location
information could be transferred from sensory and a sound composed of all possible digits
memory into a categorical form before it decayed. overlaid, from each loudspeaker. It was presented
However, Darwin et al. (1972) carried out an long enough after the arrays to be recalled that
analogous auditory experiment and found a limit their perception would have been complete (e.g.,
of about four items even though the observed 1 s afterward; cf. Vogel et al., 2006). Presumably,
decay period for sensory memory was about 4 s. the mask was capable of overwriting various types
Given the striking differences between Sperling of sensory-specific features in activated memory,
and Darwin et al. in the time period available for leaving behind only the more generic, categorical
the transfer of information to a categorical form, information present in the focus of attention,
the common four-item limit is best viewed as a which presumably is protected from masking
capacity limit rather than a rate limit. interference by the attention process. The limit of
Saults and Cowan (2007) tested this conceptual the focus of attention was again shown to be
framework in a series of experiments in which between three and four items, for either unimodal
arrays were presented in two modalities at once or, visual or bimodal stimuli.
in another procedure, one after the other. A visual Even without using masking stimuli, it may be
array of colored spots was supplemented by an possible to find a phase of the short-term memory
array of spoken digits occurring in four separate process that is general across domains. Cowan and
loudspeakers, each one consistently assigned to a Morey (2007) presented two stimulus sets to be
different voice to ease perception. On some trials, recalled (or, in control conditions, only one set).
participants knew that they were responsible for The two stimulus sets could include two spoken
both modalities at once whereas, in other trials, lists of digits, two spatial arrays of colored spots,
participants knew that they were responsible for or one of each, in either order. Following this
only the visual or only the acoustic stimuli. They presentation, a cue indicated that the participant
received a probe array that was the same as the would be responsible for only the first array, only
previous array (or the same as one modality in the second array, or both arrays. Three seconds
that previous array) or differed from the previous followed before a probe. The effect of memory
array in the identity of one stimulus. The task was to load could be compared in two ways. Performance
determine if there was a change. The use of cross- on those trials in which two sets of stimuli were
modality, capacity-limited storage predicts a parti- presented and both were cued for retention could
cular pattern of results. It predicts that performance be compared either to trials in which only one set
on either modality should be diminished in the was presented, or it could be compared to trials in
dual-modality condition compared to the unimodal which both sets were presented but the cue later
 333

indicated that only one set had to be retained. The with the finding, mentioned earlier, that memory for
part of working memory preceding the cue showed the last few list items is spared in Korsakoff’s
modality-specific dual-task effects: encoding a amnesia (Baddeley and Warrington, 1970; Carlesimo
stimulus set of one type was hurt more by also et al., 1995). In these studies, the part of the recency
encoding another set if both sets were in the same effect based on short-term memory could reflect a
modality. However, the retention of information short amount of time between presentation and
following the cue showed dual-task effects that recall of the last few items, or it could reflect the
were not modality-specific. When two sets had absence of interference between presentation and
been presented, retaining both of them was recall of the last few items. Thus, we can say that
detrimental compared to retaining only one set short-term memory exists, but often without great
(as specified by the post-stimulus retention cue to clarity as to whether the limit is a time limit or a
retain one versus both sets), and this dual-task chunk capacity limit.
effect was similar in magnitude no matter whether
the sets were in the same or different modalities.
After the initial encoding, working memory The short-term memory/working memory
storage across several seconds thus may occur distinction
abstractly, in the focus of attention.
The distinction between short-term memory and
Other evidence for a separate short-term storage working memory is clouded in a bit of confusion
but that is largely the result of different investiga-
Last, there is other evidence that does not directly tors using different definitions. Miller et al. (1960)
support either temporal decay or a capacity limit used the term ‘‘working memory’’ to refer to
specifically, but implies that one or the other of temporary memory from a functional standpoint,
these limits exist. Bjork and Whitten (1974) and so from their point of view there is no clear
Tzeng (1973) made temporal distinctiveness argu- distinction between short-term and working mem-
ments on the basis of what is called continual ory. Baddeley and Hitch (1974) were fairly
distractor list recall, in which a recency effect consistent with this definition but overlaid some
persists even when the list is followed by a descriptions on the terms that distinguished them.
distracter-filled delay before recall. The filled delay They thought of short-term memory as the unitary
should have destroyed short-term memory but the holding place as described by, for example,
recency effect occurs anyway, provided that the Atkinson and Shiffrin (1968). When they realized
items in the list also are separated by distracter- that the evidence actually was consistent with a
filled delays to increase their distinctiveness from multi-component system that could not be reduced
one another. In favor of short-term storage, to a unitary short-term store, they used the term
though, other studies have shown dissociations working memory to describe that entire system.
between what is found in ordinary immediate Cowan (1988) maintained a multi-component
recall versus continual distractor recall (e.g., word view, like Baddeley and Hitch, but without a
length effects reversed in continual distractor commitment to precisely their components;
recall: Cowan et al., 1997b; proactive interference instead, the basic subdivisions of working memory
at the most recent list positions in continual were said to be the short-term storage components
distractor recall only: Craik & Birtwistle, 1971; (activated memory along with the focus of atten-
Davelaar et al., 2005). tion within it, shown in Fig. 1) and central
There is also additional neuroimaging evidence executive processes that manipulate stored infor-
for short-term storage. Talmi et al. (2005) found mation. By Cowan’s account, Baddeley’s (1986)
that recognition of earlier portions of a list, but phonological loop and visuospatial sketchpad
not the last few items, activated areas within the would be viewed as just two of many aspects
hippocampal system that is generally associated of activated memory, which are susceptible to
with long-term memory retrieval. This is consistent interference to a degree that depends upon the
334

similarity between features of the activated and made an assertion about working memory similar
interfering information sources. Baddeley’s (2000) to that of Engle and colleagues, but a bit more
episodic buffer is possibly the same as the complex. They proposed, on the basis of some
information saved in Cowan’s focus of attention, developmental and correlational evidence, that
or at least is a closely similar concept. multiple functions of attention are relevant to
There has been some shift in the definition or individual differences in aptitudes. The control of
description of working memory along with a shift attention is relevant, but there is an independent
in the explanation of why the newer working contribution from the number of items that can be
memory tasks correlate with intelligence and held in attention, or its scope. According to this
aptitude measures so much more highly than do view, what may be necessary for a working
simple, traditional, short-term memory tasks such memory procedure to correlate well with cognitive
as serial recall. Daneman and Carpenter (1980) aptitudes is that the task must prevent covert
had assumed that what is critical is to use working verbal rehearsal so that the participant must rely
memory tasks that include both storage and on more attention-demanding processing and/or
processing components, so as to engage all of the storage to carry out the task. Cowan et al. (2005)
parts of working memory as described, for found that the task can be much simpler than the
example, by Baddeley and Hitch (1974). Instead, storage-and-processing procedures. For example,
Engle et al. (1999) and Kane et al. (2001) proposed in a version of the running memory span test,
that what is critical is whether the working digits are presented very quickly and the series
memory task is challenging in terms of the control stops at an unpredictable point, after which the
of attention. For example, Kane et al. found that participant is to recall as many items as possible
working memory span storage-and-processing from the end of the list. Rehearsal is impossible
tasks correlates well with the ability to inhibit the and, when the list ends, information presumably
natural tendency to look toward a suddenly must be retrieved from activated sensory or
appearing stimulus and instead to look the other phonological features into the focus of attention.
way, the antisaccade task. Similarly, Conway et al. This type of task correlated with aptitudes, as did
(2001) found that individuals scoring high on several other measures of the scope of attention
storage-and-processing tests of working memory (Cowan et al., 2005, 2006b). In children too young
notice their names in a channel to be ignored in to use covert verbal rehearsal (unlike older
dichotic listening much less often than low-span children and adults), even a simple digit span task
individuals; the high-span individuals apparently served as an excellent correlate with aptitudes.
are better able to make their primary task Other research verifies this idea that a working
performance less vulnerable to distraction, but memory test will correlate well with cognitive
this comes at the expense of being a bit oblivious aptitudes to the extent that it requires that
to irrelevant aspects of their surroundings. In attention be used for storage and/or processing.
response to such research, Engle and colleagues Gavens and Barrouillet (2004) carried out a
sometimes used the term working memory to refer developmental study in which they controlled the
only to the processes related to controlling difficulty and duration of a processing task that
attention. By doing so, their definition of working came between items to be recalled. There still was a
memory seems at odds with previous definitions developmental difference in span, which they
but that new definition allows the simple statement attributed to the development of a basic capacity,
that working memory correlates highly with which could reflect a developmental increase in the
aptitudes, whereas short-term memory (redefined scope of attention (cf. Cowan et al., 2005). Lépine
to include only the non-attention-related aspects of et al. (2005) showed that what was important for a
memory storage) does not correlate so highly with storage-and-processing type of span task to
aptitudes. correlate well with aptitudes is for the processing
Cowan et al. (2006b), while adhering to the component of the task (in this case, reading letters
more traditional definition of working memory, aloud) to occur quickly enough to prevent various
 335

types of rehearsal to sneak in between (see also be interesting to know whether the same type of
Conlin et al., 2005). result could be obtained for high versus low span
Several papers have pitted storage and proces- normal individuals, or whether that comparison
sing (perhaps the scope versus control of atten- instead would show a control-of-attention differ-
tion?) against one another to see which is more ence between these groups as Vogel et al. (2005)
important in accounting for individual differences. must predict. Friedman et al. (2006) found that not
Vogel et al. (2005) used a visual array task all central executive functions correlated with
modified for use with a component of event- aptitudes; updating working memory did, but
related potentials that indicates storage in visual inhibition and shifting of attention did not. On
working memory, termed contralateral delay the other hand, recall that Cowan et al. (2006b) did
activity (CDA). This activity was found to depend find was that a control-of-attention task was
not only on the number of relevant objects in the related to aptitudes.
display (e.g., red bars at varying angles to be In sum, the question of whether short-term
remembered), but sometimes also on the number memory and working memory are different may be
of irrelevant objects to be ignored (e.g., blue bars). a matter of semantics. There are clearly differences
For high-span individuals, the CDA for two between simple serial recall tasks that do not
relevant objects was found to be similar whether correlate very well with aptitude tests in adults,
or not there also were two irrelevant objects in the and other tasks requiring memory and processing,
display. However, for low-span individuals, the or memory without the possibility of rehearsal,
CDA for two relevant objects combined with two that correlate much better with aptitudes. Whether
irrelevant objects was similar to the CDA for to use the term working memory for the latter set
displays with four relevant objects alone, as if the of tasks, or whether to reserve that term for the
irrelevant objects could not be excluded from entire system of short-term memory preservation
working memory. One limitation of the study is and manipulation, is a matter of taste. The more
that the separation of participants into high versus important, substantive question may be why some
low span was based on the CDA also, and the task tasks correlate with aptitude much better than
used to measure the CDA inevitably required others.
selective attention (to one half of the display) on
every trial, whether or not it included objects of an
irrelevant color. Conclusion
Gold et al. (2006) investigated similar issues in a
behavioral design, and testing the difference The distinction between long-term and short-term
between schizophrenic patients and normal con- memory depends on whether it can be demon-
trol participants. Each trial started with a cue to strated that there are properties specific to short-
attend to one part of the display at the expense of term memory; the main candidates include
another (e.g., bars of one relevant color but not temporal decay and a chunk capacity limit. The
another, irrelevant color). The probe display was a question of decay is still pretty much open to
set that was cued for relevance on most trials (in debate, whereas there is growing support for a
some experiments, 75%) whereas, occasionally, chunk capacity limit. These limits were discussed
the probe display was a set that was not cued. This in a framework shown in Fig. 1.
allowed a separate measure of the control of The distinction between short-term memory and
attention (the advantage for cued items over working memory is one that depends on the
uncued items) and the storage capacity of working definition that one accepts. Nevertheless, the
memory (the mean number of items recalled from substantive question is why some tests of memory
each array, adding across cued and uncued sets). over the short term serve as some of the best
Unlike the initial expectations, the clear result was correlates of cognitive aptitudes, whereas others
that the difference between groups was in the do not. The answer seems to point to the
capacity, not in the control of attention. It would importance of an attentional system used both for
336

processing and for storage. The efficiency of this Bunting, M.F., Cowan, N. and Saults, J.S. (2006) How does
system and its use in working memory seem to running memory span work? Q. J. Exp. Psychol., 59: 1691–
1700.
differ substantially across individuals (e.g., Conway
Carlesimo, G.A., Sabbadini, M., Fadda, L. and Caltagirone, C.
et al., 2002; Kane et al., 2004; Cowan et al., 2005, (1995) Different components in word-list forgetting of pure
2006b), as well as improving with development in amnesics, degenerative demented and healthy subjects.
childhood (Cowan et al., 2005, 2006b) and declining Cortex, 31: 735–745.
in old age (Naveh-Benjamin et al., 2007; Stoltzfus et Chen, Z. and Cowan, N. (2005) Chunk limits and length limits
in immediate recall: a reconciliation. J. Exp. Psychol. Learn.
al., 1996; Cowan et al., 2006c).
Mem. Cogn., 31: 1235–1249.
Cocchini, G., Logie, R.H., Della Sala, S., MacPherson, S.E.
Acknowledgment and Baddeley, A.D. (2002) Concurrent performance of two
memory tasks: evidence for domain-specific working memory
This work was completed with the assistance of systems. Mem. Cogn., 30: 1086–1095.
Conlin, J.A., Gathercole, S.E. and Adams, J.W. (2005)
NIH Grant R01 HD-21338.
Children’s working memory: investigating performance limi-
tations in complex span tasks. J. Exp. Child Psychol., 90:
303–317.
References Conway, A.R.A., Cowan, N. and Bunting, M.F. (2001) The
cocktail party phenomenon revisited: the importance of
Atkinson, R.C. and Shiffrin, R.M. (1968) Human memory: a working memory capacity. Psychon. Bull. Rev., 8: 331–335.
proposed system and its control processes. In: Spence K.W. Conway, A.R.A., Cowan, N., Bunting, M.F., Therriault, D.J.
and Spence J.T. (Eds.), The Psychology of Learning and and Minkoff, S.R.B. (2002) A latent variable analysis of
Motivation: Advances in Research and Theory, Vol. 2. working memory capacity, short-term memory capacity,
Academic Press, New York, pp. 89–195. processing speed, and general fluid intelligence. Intelligence,
Baddeley, A. (2000) The episodic buffer: a new component of 30: 163–183.
working memory? Trends Cogn. Sci., 4: 417–423. Conway, A.R.A., Kane, M.J., Bunting, M.F., Hambrick, D.Z.,
Baddeley, A.D. (1986) Working Memory. Oxford Psychology Wilhelm, O. and Engle, R.W. (2005) Working memory span
Series No. 11. Clarendon Press, Oxford. tasks: a methodological review and user’s guide. Psychon.
Baddeley, A.D. and Hitch, G. (1974) Working memory. In: Bull. Rev., 12: 769–786.
Bower G.H. (Ed.), The Psychology of Learning and Cowan, N. (1988) Evolving conceptions of memory storage,
Motivation, Vol. 8. Academic Press, New York, pp. 47–89. selective attention, and their mutual constraints within the
Baddeley, A.D. and Scott, D. (1971) Short-term forgetting in human information processing system. Psychol. Bull., 104:
the absence of proactive inhibition. Q. J. Exp. Psychol., 23: 163–191.
275–283. Cowan, N. (1992) Verbal memory span and the timing of
Baddeley, A.D., Thomson, N. and Buchanan, M. (1975) Word spoken recall. J. Mem. Lang., 31: 668–684.
length and the structure of short-term memory. J. Verbal Cowan, N. (1995) Attention and memory: an integrated
Learn. Verbal Behav., 14: 575–589. framework. Oxford Psychology Series, No. 26. Oxford
Baddeley, A.D. and Warrington, E.K. (1970) Amnesia and the University Press, New York.
distinction between long- and short-term memory. J. Verbal Cowan, N. (1999) An embedded-processes model of working
Learn. Verbal Behav., 9: 176–189. memory. In: Miyake A. and Shah P. (Eds.), Models of
Barrouillet, P., Bernardin, S. and Camos, V. (2004) Time Working Memory: Mechanisms of Active Maintenance and
constraints and resource sharing in adults’ working memory Executive Control. Cambridge University Press, Cambridge,
spans. J. Exp. Psychol.: Gen., 133: 83–100. UK, pp. 62–101.
Barrouillet, P., Bernardin, S., Portrat, S., Vergauwe, E. and Cowan, N. (2001) The magical number 4 in short-term memory:
Camos, V. (2007) Time and cognitive load in working a reconsideration of mental storage capacity. Behav. Brain
memory. J. Exp. Psychol. Learn. Mem. Cogn., 33: 570–585. Sci., 24: 87–185.
Bjork, R.A. and Whitten, W.B. (1974) Recency-sensitive Cowan, N. (2005) Working memory capacity. Psychology
retrieval processes in long-term free recall. Cogn. Psychol., Press, Hove, East Sussex, UK.
6: 173–189. Cowan, N. and Aubuchon, A. M. (in press) Short-term memory
Broadbent, D.E. (1958) Perception and Communication. loss over time without retroactive stimulus interference.
Pergamon Press, New York. Psychon. Bull. Rev.
Broadbent, D.E. (1975) The magic number seven after fifteen Cowan, N., Chen, Z. and Rouder, J.N. (2004) Constant
years. In: Kennedy A. and Wilkes A. (Eds.), Studies in Long- capacity in an immediate serial-recall task: a logical sequel
Term Memory. Wiley, Oxford, England, pp. 3–18. to Miller (1956). Psychol. Sci., 15: 634–640.
Brown, G.D.A., Preece, T. and Hulme, C. (2000) Oscillator- Cowan, N., Elliott, E.M., Saults, J.S., Morey, C.C., Mattox, S.,
based memory for serial order. Psychol. Rev., 107: 127–181. Hismjatullina, A. and Conway, A.R.A. (2005) On the
 337

capacity of attention: its estimation and its role in Engle, R.W., Tuholski, S.W., Laughlin, J.E. and Conway,
working memory and cognitive aptitudes. Cogn. Psychol., A.R.A. (1999) Working memory, short-term memory, and
51: 42–100. general fluid intelligence: a latent-variable approach. J. Exp.
Cowan, N., Elliott, E.M., Saults, J.S., Nugent, L.D., Bomb, P. Psychol. Gen., 128: 309–331.
and Hismjatullina, A. (2006a) Rethinking speed theories of Ericsson, K.A. and Kintsch, W. (1995) Long-term working
cognitive development: increasing the rate of recall without memory. Psychol. Rev., 102: 211–245.
affecting accuracy. Psychol. Sci., 17: 67–73. Friedman, N.P., Miyake, A., Corley, R.P., Young, S.E.,
Cowan, N., Fristoe, N.M., Elliott, E.M., Brunner, R.P. and DeFries, J.C. and Hewitt, J.K. (2006) Not all execu-
Saults, J.S. (2006b) Scope of attention, control of attention, tive functions are related to intelligence. Psychol. Sci., 17:
and intelligence in children and adults. Mem. Cogn., 34: 172–179.
1754–1768. Gavens, N. and Barrouillet, P. (2004) Delays of retention,
Cowan, N. and Morey, C.C. (2007) How can dual-task working processing efficiency, and attentional resources in
memory retention limits be investigated? Psychol. Sci., 18: working memory span development. J. Mem. Lang., 51:
686–688. 644–657.
Cowan, N., Naveh-Benjamin, M., Kilb, A. and Saults, J.S. Glanzer, M. and Cunitz, A.R. (1966) Two storage mechanisms
(2006c) Life-span development of visual working memory: in free recall. J. Verbal Learn. Verbal Behav., 5: 351–360.
when is feature binding difficult? Dev. Psychol., 42: Glenberg, A.M. and Swanson, N.C. (1986) A temporal distin-
1089–1102. ctiveness theory of recency and modality effects. J. Exp.
Cowan, N., Nugent, L.D., Elliott, E.M., Ponomarev, I. and Psychol. Learn. Mem. Cogn., 12: 3–15.
Saults, J.S. (1999) The role of attention in the development of Gold, J.M., Fuller, R.L., Robinson, B.M., McMahon, R.P.,
short-term memory: age differences in the verbal span of Braun, E.L. and Luck, S.J. (2006) Intact attentional control
apprehension. Child Dev., 70: 1082–1097. of working memory encoding in schizophrenia. J. Abnorm.
Cowan, N., Saults, J.S. and Nugent, L.D. (1997a) The role of Psychol., 115: 658–673.
absolute and relative amounts of time in forgetting within Greene, R.L. (1996) The influence of experimental design: the
immediate memory: the case of tone pitch comparisons. example of the Brown–Peterson paradigm. Can. J. Exp.
Psychon. Bull. Rev., 4: 393–397. Psychol., 50: 240–242.
Cowan, N., Wood, N.L., Nugent, L.D. and Treisman, M. Guttentag, R.E. (1984) The mental effort requirement of
(1997b) There are two word length effects in verbal short- cumulative rehearsal: a developmental study. J. Exp. Child
term memory: opposed effects of duration and complexity. Psychol., 37: 92–106.
Psychol. Sci., 8: 290–295. Hebb, D.O. (1949) Organization of Behavior. Wiley, New York.
Craik, F.I.M. and Birtwistle, J. (1971) Proactive inhibition in Hockey, R. (1973) Rate of presentation in running memory and
free recall. J. Exp. Psychol., 91: 120–123. direct manipulation of input-processing strategies. Q. J. Exp.
Crowder, R.G. (1982) The demise of short-term memory. Acta Psychol. A, 25: 104–111.
Psychol., 50: 291–323. James, W. (1890) The Principles of Psychology. Henry Holt,
Crowder, R.G. (1993) Short-term memory: where do we stand? New York.
Mem. Cogn., 21: 142–145. Kane, M.J., Bleckley, M.K., Conway, A.R.A. and Engle, R.W.
Daneman, M. and Carpenter, P.A. (1980) Individual differences (2001) A controlled-attention view of working-memory
in working memory and reading. J Verbal Learn. Verbal capacity. J. Exp. Psychol. Gen., 130: 169–183.
Behav., 19: 450–466. Kane, M.J., Hambrick, D.Z., Tuholski, S.W., Wilhelm, O.,
Daneman, M. and Merikle, P.M. (1996) Working memory and Payne, T.W. and Engle, R.E. (2004) The generality of
language comprehension: a meta-analysis. Psychon. Bull. working-memory capacity: a latent-variable approach to
Rev., 3: 422–433. verbal and visuo-spatial memory span and reasoning. J. Exp.
Darwin, C.J., Turvey, M.T. and Crowder, R.G. (1972) An Psychol. Gen., 133: 189–217.
auditory analogue of the Sperling partial report procedure: Keppel, G. and Underwood, B.J. (1962) Proactive inhibition in
evidence for brief auditory storage. Cogn. Psychol., 3: short-term retention of single items. J. Verbal Learn. Verbal
255–267. Behav., 1: 153–161.
Davelaar, E.J., Goshen-Gottstein, Y., Ashkenazi, A., Haar- Kyllonen, P.C. and Christal, R.E. (1990) Reasoning ability is
man, H.J. and Usher, M. (2005) The demise of short-term (little more than) working-memory capacity? Intelligence, 14:
memory revisited: empirical and computational investiga- 389–433
tions of recency effects. Psychol. Rev., 112: 3–42. Lépine, R., Barrouillet, P. and Camos, V. (2005) What makes
Ebbinghaus, H. (1885/1913) Memory: a contribution to working memory spans so predictive of high level cognition?
experimental psychology. Translated by H.A. Ruger and Psychon. Bull. Rev., 12: 165–170.
C.E. Bussenius. Teachers College, Columbia University, Lewandowsky, S., Duncan, M. and Brown, G.D.A. (2004)
New York. (Originally in German, Ueber das gedächtnis: Time does not cause forgetting in short-term serial recall.
Untersuchen zur experimentellen psychologie). Psychon. Bull. Rev., 11: 771–790.
Engle, R.W. (2002) Working memory capacity as executive McGeoch, J.A. (1932) Forgetting and the law of disuse.
attention. Curr. Dir. Psychol. Sci., 11: 19–23. Psychol. Rev., 39: 352–370.
338

Melton, A.W. (1963) Implications of short-term memory for a Sperling, G. (1960) The information available in brief visual
general theory of memory. J. Verbal Learn. Verbal Behav., 2: presentations. Psychol. Monogr., 74. (Whole No. 498.)
1–21. Stoltzfus, E.R., Hasher, L. and Zacks, R.T. (1996) Working
Miller, G.A. (1956) The magical number seven, plus or minus memory and retrieval: an inhibition-resource approach. In:
two: some limits on our capacity for processing information. Richardson J.T.E., Engle R.W., Hasher L., Logie R.H.,
Psychol. Rev., 63: 81–97. Stoltzfus E.R. and Zacks R.T. (Eds.), Working Memory and
Miller, G.A. (1989) George A. Miller. In: Lindzey G. (Ed.), A Human Cognition. Oxford University Press, New York,
History of Psychology in Autobiography, Vol. VIII. Stanford pp. 66–88.
University Press, Stanford, CA, pp. 391–418. Talmi, D., Grady, C.L., Goshen-Gottstein, Y. and Moscovitch,
Miller, G.A., Galanter, E. and Pribram, K.H. (1960) Plans and M. (2005) Neuroimaging the serial position curve: a test of
the structure of behavior. Holt, Rinehart and Winston, Inc., single-store versus dual-store models. Psychol. Sci., 16:
New York. 716–723.
Morey, C.C. and Cowan, N. (2004) When visual and verbal Tulving, E. and Patkau, J.E. (1962) Concurrent effects of
memories compete: evidence of cross-domain limits in work- contextual constraint and word frequency on immediate recall
ing memory. Psychon. Bull. Rev., 11: 296–301. and learning of verbal material. Can. J. Psychol., 16: 83–95.
Morey, C.C. and Cowan, N. (2005) When do visual and verbal Turvey, M.T., Brick, P. and Osborn, J. (1970) Proactive
memories conflict? The importance of working-memory load interference in short-term memory as a function of prior-
and retrieval. J. Exp. Psychol. Learn. Mem. Cogn., 31: 703– item retention interval. Q. J. Exp. Psychol., 22: 142–147.
713. Tzeng, O.J.L. (1973) Positive recency effect in a delayed free
Nairne, J.S. (2002) Remembering over the short-term: the case recall. J. Verbal Learn. Verbal Behav., 12: 436–439.
against the standard model. Annu. Rev. Psychol., 53: 53–81. Unsworth, N. and Engle, R.W. (2007) The nature of individual
Naveh-Benjamin, M., Cowan, N., Kilb, A. and Chen, Z. (2007) differences in working memory capacity: active maintenance
Age-related differences in immediate serial recall: dissociating in primary memory and controlled search from secondary
chunk formation and capacity. Mem. Cognit., 35: 724–737. memory. Psychol. Rev., 114: 104–132.
Neath, I. and Surprenant, A. (2003) Human memory (2nd ed.). Vogel, E.K., McCollough, A.W. and Machizawa, M.G. (2005)
Wadsworth, Belmont, CA. Neural measures reveal individual differences in controlling
Nipher, F.E. (1878) On the distribution of errors in numbers access to working memory. Nature, 438: 500–503.
written from memory. Trans. Acad. Sci. St. Louis, 3: ccx– Vogel, E.K., Woodman, G.F. and Luck, S.J. (2006) The time
ccxi. course of consolidation in visual working memory. J. Exp.
Peterson, L.R. and Peterson, M.J. (1959) Short-term retention Psychol. Hum. Percept. Perform., 32: 1436–1451.
of individual verbal items. J. Exp. Psychol., 58: 193–198. Waugh, N.C. and Norman, D.A. (1965) Primary memory.
Saults, J.S. and Cowan, N. (2007) A central capacity limit to the Psychol. Rev., 72: 89–104.
simultaneous storage of visual and auditory arrays in Wickelgren, W.A. (1974) Single-trace fragility theory of
working memory. J. Exp. Psychol., 136: 663–684. memory dynamics. Mem. Cogn., 2: 775–780.