�

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 twofundamental ways, with only short-term memory demonstrating (1) temporal decay and (2) chunk capacitylimits. Both properties of short-term memory are still controversial but the current literature is ratherencouraging regarding the existence of both decay and capacity limits. Working memory has beenconceived and defined in three different, slightly discrepant ways: as short-term memory applied tocognitive tasks, as a multi-component system that holds and manipulates information in short-termmemory, and as the use of attention to manage short-term memory. Regardless of the definition, there aresome measures of memory in the short term that seem routine and do not correlate well with cognitiveaptitudes and other measures (those usually identified with the term ‘‘working memory’’) that seem moreattention demanding and do correlate well with these aptitudes. The evidence is evaluated and placedwithin a theoretical framework depicted in Fig. 1.

Keywords: attention; capacity of working memory; control of attention; decay of short-term memory; focusof 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 tocapture 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 ongoingnaı̈ve 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 tracedfacts 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 andpsychological laboratories, a more careful obser- forgetting of new information in the form of seriesvation must have shown that there are separable of nonsense syllables tested at various periods uptoaspects 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 ofspecial concentration’’ (p. 33) but that this

� immediate memory did not ensure that the seriesCorresponding author. Tel.: +1 573-882-4232;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

dx.doi.org/10.1016/S0079-6123(07)00020-9.3d

324

recall later on. Stable memorization sometimesrequired further repetitions of the series. Soonafterward, James (1890) proposed a distinctionbetween primary memory, the small amount ofinformation held as the trailing edge of theconscious present, and secondary memory, the vastbody of knowledge stored over a lifetime. Theprimary memory of James is like the first fleetinggrasp of Ebbinghaus.

The Industrial Revolution made some newdemands on what James (1890) called primarymemory. In the 1850s, telegraph operators had toremember and interpret rapid series of dots anddashes conveyed acoustically. In 1876, the tele-phone was invented. Three years later, operators inLowell, Massachusetts started using telephonenumbers for more than 200 subscribers so thatsubstitute operators could be more easily trained ifthe town’s four regular operators succumbed to araging measles epidemic. This use of telephonenumbers, complemented by a word prefix, ofcourse spread. (The author’s telephone numberin 1957 was WHitehall 2-6742; the number is stillassigned, albeit as a seven-digit number.) Evenbefore the book by Ebbinghaus, Nipher (1878)reported on the serial position curve obtainedamong the digits in logarithms that he tried torecall. The nonsense syllables that Ebbinghaushad invented as a tool can be seen to haveacquired more ecological validity in an industrialage with expanding information demands, perhapshighlighting the practical importance of primarymemory in daily life. Primary memory seemstaxed as one is asked to keep in mind aspects ofan unfamiliar situation, such as names, places,things, and ideas that one has not encounteredbefore.

Yet, the subjective experience of a differencebetween primary and secondary memory does notautomatically guarantee that these types of memo-ry separately contribute to the science of remem-bering. Researchers from a different perspectivehave long hoped that they could write a singleequation, or a single set of principles at least, thatwould capture all of memory, from the veryimmediate to the very long-term. McGeoch(1932) illustrated that forgetting over time wasnot simply a matter of an inevitable decay of

memory but rather of interference during theretention interval; one could find situations inwhich memory improved, rather than diminish,over time. From this perspective, one might viewwhat appeared to be forgetting from primarymemory as the profound effect of interferencefrom other items on memory for any one item,with interference effects continuing forever but nottotally destroying a given memory. This perspectivehas been maintained and developed over the yearsby a steady line of researchers believing in the unityof memory, including, among others, Melton (1963),Bjork and Whitten (1974), Wickelgren (1974),Crowder (1982, 1993), Glenberg and Swanson(1986), Brown et al. (2000), Nairne (2002), Neathand Surprenant (2003), and Lewandowsky et al.(2004).

Description of three kinds of memory

In this chapter I will assess the strength of evidencefor three types of memory: long-term memory,short-term memory, and working memory. Long-term memory is a vast store of knowledge and arecord of prior events, and it exists according to alltheoretical views; it would be difficult to deny thateach normal person has at his or her command arich, although not flawless or complete, set of long-term memories.

Short-term memory is related to the primarymemory of James (1890) and is a term thatBroadbent (1958) and Atkinson and Shiffrin(1968) used in slightly different ways. LikeAtkinson and Shiffrin, I take it to reflect facultiesof the human mind that can hold a limited amountof information in a very accessible state tempora-rily. One difference between the term ‘‘short-termmemory’’ and the term ‘‘primary memory’’ is thatthe latter might be considered to be morerestricted. It is possible that not every temporarilyaccessible idea is, or even was, in consciousawareness. For example, by this conception, if youare speaking to a person with a foreign accent andinadvertently alter your speech to match theforeign speaker’s accent, you are influenced bywhat was until that point an unconscious (andtherefore uncontrollable) aspect of your short-term

325

memory. One might relate short-term memory to apattern of neural firing that represents a particularidea and one might consider the idea to be in short-term memory only when the firing pattern, or cellassembly, is active (Hebb, 1949). The individualmight or might not be aware of the idea duringthat period of activation.

Working memory is not completely distinct fromshort-term memory. It is a term that was used byMiller et al. (1960) to refer to memory as it is usedto plan and carry out behavior. One relies onworking memory to retain the partial results whilesolving an arithmetic problem without paper, tocombine the premises in a lengthy rhetoricalargument, or to bake a cake without making theunfortunate mistake of adding the same ingredienttwice. (Your working memory would have beenmore heavily taxed while reading the previoussentence if I had saved the phrase ‘‘one relies onworking memory’’ until the end of the sentence,which I did in within my first draft of thatsentence; working memory thus affects goodwriting.) The term ‘‘working memory’’ becamemuch more dominant in the field after Baddeleyand Hitch (1974) demonstrated that a singlemodule could not account for all kinds oftemporary memory. Their thinking led to aninfluential model (Baddeley, 1986) in whichverbal-phonological and visual-spatial representa-tions were held separately, and were managed andmanipulated with the help of attention-relatedprocesses, termed the central executive. In the 1974paper, this central executive possibly had its ownmemory that crossed domains of representation.By 1986, this general memory had been eliminatedfrom the model, but it was added back again byBaddeley (2000) in the form of an episodic buffer.That seemed necessary to explain short-termmemory of features that did not match the otherstores (particularly semantic information in mem-ory) and to explain cross-domain associations inworking memory, such as the retention of linksbetween names and faces. Because of the work ofBaddeley et al. (1975), working memory isgenerally viewed as the combination of multiplecomponents working together. Some even includein that bundle the heavy contribution of long-termmemory, which reduces the working memory load

by organizing and grouping information in work-ing memory into a smaller number of units (Miller,1956; Ericsson and Kintsch, 1995). For example,the letter series IRSCIAFBI can be rememberedmuch more easily as a series of acronyms for threefederal agencies of the United States of America:the Internal Revenue Service (IRS), the CentralIntelligence Agency (CIA), and the Federal Bureauof Investigation (FBI). However, that factor wasnot emphasized in the well-known model ofBaddeley (1986).

What is clear from my definition is that workingmemory includes short-term memory and otherprocessing mechanisms that help to make use ofshort-term memory. This definition is differentfrom the one used by some other researchers (e.g.,Engle, 2002), who would like to reserve the termworking memory to refer only to the attention-related aspects of short-term memory. This,however, is not so much a debate about substance,but rather a slightly confusing discrepancy in theusage of terms.

One reason to pursue the term working memoryis that measures of working memory have beenfound to correlate with intellectual aptitudes (andespecially fluid intelligence) better than measuresof short-term memory and, in fact, possibly betterthan measures of any other particular psychologi-cal process (e.g., Daneman and Carpenter, 1980;Kyllonen and Christal, 1990; Daneman andMerikle, 1996; Engle et al., 1999; Conway et al.,2005). It has been thought that this reflects the useof measures that incorporate not only storage butalso processing, the notion being that both storageand processing have to be engaged concurrently toassess working memory capacity in a way that isrelated to cognitive aptitude. More recently, Engleet al. (1999) introduced the notion that aptitudesand working memory both depend on the abilityto control attention, or to apply the control ofattention to the management of both primary andsecondary memory (Unsworth and Engle, 2007).However, more research is needed on exactly whatwe learn from the high correlation between work-ing memory and intellectual aptitudes, and thisissue will be discussed further after the more basicissue of the short-term versus the long-termmemory distinction is addressed.

326

Meanwhile, it may be helpful to summarize atheoretical framework (Cowan, 1988, 1995, 1999,2001, 2005) based on past research. This frame-work, illustrated in Fig. 1, helps to account for therelation between long-term, short-term, and work-ing memory mechanisms and explains what I see asthe relation between them. In this framework,short-term memory is derived from a temporarilyactivated subset of information in long-term mem-ory. This activated subset may decay as a functionof time unless it is refreshed, although the evidencefor decay is still tentative at best. A subset of theactivated information is the focus of attention,which appears to be limited in chunk capacity (howmany separate items can be included at once). Newassociations between activated elements can formthe focus of attention. Now the evidence related tothis modeling framework will be discussed.

The short-term memory/long-term memorydistinction

If there is a difference between short- and long-term memory stores, there are two possible ways in

which these stores may differ: in duration, and incapacity. A duration difference means that items inshort-term storage decay from this sort of storageas a function of time. A capacity difference meansthat there is a limit in how many items short-termstorage can hold. If there is only a limit incapacity, a number of items smaller than thecapacity limit could remain in short-term storageuntil they are replaced by other items. Both typesof limit are controversial. Therefore, in order toassess the usefulness of the short-term storageconcept, duration and capacity limits will beassessed in turn.

Duration limits

The concept of short-term memory limited bydecay over time was present even at the beginningof cognitive psychology, for example in the workof Broadbent (1958). If decay were the onlyprinciple affecting performance in an immediatememory experiment, it would perhaps be easy todetect this decay. However, even in Broadbent’swork contaminating variables were recognized. Toassess 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,long-term retrieval, and temporal distinctiveness,which will be discussed one at a time in conjunc-tion with evidence for and against decay.

Overcoming contamination from rehearsal

According to various researchers there is a processwhereby one imagines how the words on the listare pronounced without saying them aloud, aprocess called covert verbal rehearsal. Withpractice, this process comes to occur with aminimum of attention. Guttentag (1984) used asecondary task to show that rehearsal of a list tobe recalled was effortful in young children, but notin adults. If, in a particular experimental proce-dure, no loss of short-term memory is observed,one can attribute that response pattern to rehear-sal. Therefore, steps have been taken to eliminaterehearsal through a process termed articulatorysuppression, in which a simple utterance such asthe word ‘‘the’’ is repeatedly pronounced by theparticipant during part or all of the short-termmemory task (e.g., Baddeley et al., 1975). There isstill the possible objection that whatever utteranceis used to suppress rehearsal unfortunately causesinterference, which could be the true reason formemory loss over time instead of decay.

That problem of interference would appearmoot in light of the findings of Lewandowsky etal. (2004). They presented lists of letters to berecalled and varied how long the participant wassupposed to take to recall each item in the list. Insome conditions, they added articulatory suppres-sion to prevent rehearsal. Despite that suppres-sion, they observed no difference in performancewith the time between items in the responsevarying between 400 and 1600ms (or betweenconditions in which the word ‘‘super’’ waspronounced one, two, or three times betweenconsecutive items in the response). They found noevidence of memory decay.

A limitation of this finding, though, is thatcovert verbal rehearsal may not be the only type ofrehearsal that participants can use. Perhaps thereare types that are not prevented by articulatorysuppression. In particular, Cowan (1992)

suggested that the process of mentally attendingto words or searching through the list, anattention-demanding process, could serve to reac-tivate items to be recalled in a manner similar tocovert verbal rehearsal. The key difference is that itwould not be expected that articulatory suppres-sion would prevent that type of rehearsal. Instead,to prevent that type of rehearsal an attention-demanding task would have to be used.

Barrouillet et al. (2004, 2007) have results thatdo seem to suggest that there is another, moreattention-demanding type of rehearsal. They haveinterposed materials between items to be recalledthat require choices; they can be numbers to readaloud or multi-choice reaction times. It is foundthat these interfere with retention to an extentcommensurate to the proportion of the inter-iteminterval used up attending to the distracting items.As the rate of the distracting items goes up, fewerof the to-be-recalled items are recalled. The notionis that when the distracting task does not requireattention, the freed-up attention allows an atten-tion-based rehearsal of the items to be recalled.When the interposed task is more automatic anddoes not require as much attention (e.g., anarticulatory suppression task) there is much lesseffect of the rate of these interposed items.

Based on this logic, one could imagine a versionof Lewandowsky’s task in which not articulatorysuppression but attention-demanding verbal sti-muli are placed between items in the response, andin which the duration of this filled time betweenitems in the response varies from trial to trial. Theverbal, attention-demanding stimuli should pre-vent both attention-based rehearsal and articula-tion-based rehearsal. If there is decay, thenperformance should decline across serial positionsmore severely when longer filled intervals areplaced between items in the response. Unfortu-nately, though, such results might be accounted foralternatively as the result of interference from thedistracting stimuli, without the need to invokedecay.

What seems to be needed, then, is a procedure toprevent both articulation-based and attention-based rehearsal without introducing interference.Cowan and Aubuchon (in press) tried out one typeof procedure that may accomplish this. They

328

presented lists of seven printed digits in which thetime between items varied within a list. In additionto some randomly timed filler lists, there were fourcritical trial types, in which the six inter-digit blankintervals were all short (0.5 s following each item)or all long (2 s following each item), or comprisedthree short and then three long intervals, or threelong and then three short intervals. Moreover,there were two post-list response cues. Accordingto one cue, the participant was to recall the listwith the items in the presented order, but at anyrate they wished. According to the other responsecue, the list was to be recalled using the sametiming in which it was presented. The expectationwas that the need to remember the timing in thelatter response condition would prevent rehearsalof either type. As a consequence, performanceshould be impaired on trials in which the first threeresponse intervals are long because, on these trials,there is more time for forgetting of most of the listitems. Just as predicted, there was a significantinteraction between the response cue and thelength of the first half of the response intervals.When participants were free to recall items at theirown pace, performance was no better with a shortfirst half (M=.71) than with a long first half(M=.74). The slight benefit of a long first half inthat situation could occur because it allowed thelist to be rehearsed early on in the response. Incontrast, when the timing of recall had to matchthe timing of the list presentation, performancewas better with a short first half (M = .70) thanwith a long first half (M = .67). This, then,suggests there could be decay in short-termmemory.

Overcoming contamination from long-term retrieval

If there is more than one type of memory storagethen there still is the problem of which storeprovided the information underlying a response.There is no guarantee that, just because aprocedure is considered a test of short-termstorage, the long-term store will not be used. Forexample, in a simple digit span task, a series ofdigits is presented and is to be repeated immedi-ately afterward from memory. If that series turned

out to be only slightly different from the partici-pant’s telephone number, the participant might beable to memorize the new number quickly andrepeat it from long-term memory. The dual-storetheories of memory allow this. Although Broad-bent (1958) and Atkinson and Shiffrin (1968) drewtheir models of information processing as a seriesof boxes representing different memory stores,with long-term memory following short-termmemory, these boxes do not imply that memoryis exclusively in one box or another; they are betterinterpreted as the relative times of the first entry ofinformation from a stimulus into one store andthen the next. The question remains, then as tohow one can determine if a response comes fromshort-term memory.

Waugh and Norman (1965) developed a mathe-matical model to accomplish this. The modeloperated with the assumption that long-termmemory occurs for the entire list, including aplateau in the middle of the list. In contrast, by thetime of recall, short-term memory is said to remainonly at the end of the list. This model assumesthat, for any particular serial position within a list,the likelihood of successful short-term storage (S)and long-term storage (L) are independent, sothat the likelihood of recalling the item isS+L�SL.

A slightly different assumption is that short- andlong-term stores are not independent but are usedin a complementary fashion. The availability ofshort-term memory of an item may allow resourcesneeded for long-term memorization to be shiftedto elsewhere in the list. The data seem moreconsistent with that assumption. In several studies,lists to be recalled have been presented to patientswith Korsakoff’s amnesia and normal controlparticipants (Baddeley and Warrington, 1970;Carlesimo et al., 1995). These studies show that,in immediate recall, performance in amnesicindividuals is preserved at the last few serialpositions of the list. It is as if the performancein those serial positions is based mostly or entirelyon short-term storage, and that there is nodecrease in that kind of storage in the amnesicpatients. In delayed recall, the amnesic patientsshow a deficit at all serial positions, as one wouldexpect if short-term memory for the end of the list

329

is lost as a function of a filled delay period(Glanzer and Cunitz, 1966).

Overcoming contamination from temporal

distinctiveness

Last, it has been argued that the loss of memoryover time is not necessarily the result of decay.Instead, it can be caused by temporal distinctive-ness in retrieval. This kind of theory assumes thatthe temporal context of an item serves as aretrieval cue for that item, even in free recall. Anitem separated in time from all other items isrelatively distinctive and easy to recall, whereas anitem that is relatively close to other items is moredifficult to recall because it shares their temporalcues to retrieval. Shortly after a list is presented themost recent items are the most distinct temporally(much like the distinctness of a telephone pole youare practically touching compared to poles extend-ing further down the road). Across a retentioninterval, the relative distinctiveness of the mostrecent items decreases (much like standing faraway from even the last pole in a series).

Although there are data that can be interpretedaccording to distinctiveness, there also are whatlook like dissociations between the effects ofdistinctiveness and a genuine short-term memoryeffect. One can see this, for example, in the classicprocedure of Peterson and Peterson (1959) inwhich letter trigrams are to be recalled immedi-ately or only after a distracting task, countingbackward from a starting number by three, for aperiod lasting up to 18 s. Peterson and Petersonfound severe memory loss for the letter trigram asthe filled delay was increased. However, subse-quently, sceptics argued that the memory lossoccurred because the temporal distinctiveness ofthe current letter trigram diminished as the filleddelay increased. In particular, this delay effect wassaid to occur because of the increase across testdelays in the proactive interference from previoustrials. On the first few trials, the delay does notmatter (Keppel and Underwood, 1962) and nodetrimental effect of delay is observed if delays of5, 10, 15, and 20 s are tested in separate trial blocks(Turvey et al., 1970; Greene, 1996).

Yet, there may be a true decay effect at shortertest intervals. Baddeley and Scott (1971) set up atrailer in a shopping mall so that they could test alarge number of participants for one trial each, soas to avoid proactive interference. They found aneffect of the test delay within the first 5 s but not atlonger delays. Still, it seems that the concept ofdecay is not yet on very firm ground and warrantsfurther study. It may be that decay actually reflectsnot a gradual degradation of the quality of theshort-term memory trace, but a sudden collapse ata point that varies from trial to trial. With acontrol for temporal distinctiveness, Cowan et al.(1997a) found what could be a sudden collapse inthe representation of memory for a tone withdelays of between 5 and 10 s.

Chunk capacity limits

The concept of capacity limits was raised severaltimes in the history of cognitive psychology. Miller(1956) famously discussed the ‘‘magical numberseven plus or minus two’’ as a constant in short-term processing, including list recall, absolutejudgment, and numerical estimation experiments.However, his autobiographical essay (Miller, 1989)indicates that he was never very serious about thenumber seven; it was a rhetorical device that heused to tie together the otherwise unrelated strandsof his research for a talk. Although it is true thatmemory span is approximately seven items inadults, there is no guarantee that each item is aseparate entity. Perhaps the most important pointof Miller’s (1956) article was that multiple itemscan be combined into a larger, meaningful unit.Later studies suggested that the limit in capacity ismore typically only three or four units (Broadbent,1975; Cowan, 2001). That conclusion was based onan attempt to take into account strategies thatoften increase the efficiency of use of a limitedcapacity, or that allow the maintenance of addi-tional information separate from that limitedcapacity. To understand these methods of discuss-ing capacity limits I will again mention three typesof contamination. These come from chunking andthe use of long-term memory, from rehearsal, andfrom non-capacity-limited types of storage.

330

Overcoming contamination from chunking and the

use of long-term memory

A participant’s response in an immediate-memorytask depends on how the information to berecalled is grouped to form multi-item chunks(Miller, 1956). Because it is not usually clear whatchunks have been used in recall, it is not clear howmany chunks can be retained and whether thenumber is truly fixed. Broadbent (1975) proposedsome situations in which multi-item chunk forma-tion was not a factor, and suggested on the basis ofresults from such procedures that the true capacitylimit is three items (each serving as a single-itemchunk). For example, although memory span isoften about seven items, errors are made withseven-item lists and the error-free limit is typicallythree items. When people must recall items from acategory in long-term memory, such as the statesof the United States, they do so in spurts of aboutthree items on average. It is as if the bucket ofshort-term memory is filled from the well of long-term memory and must be emptied before it isrefilled. Cowan (2001) noted other such situationsin which multi-item chunks cannot be formed. Forexample, in running memory span, a long list ofitems is presented with an unpredictable endpoint,making grouping impossible. When the list ends,the participant is to recall a certain number ofitems from the end of the list. Typically, people canrecall three or four items from the end of the list,although the exact number depends on taskdemands (Bunting et al., 2006). Individuals differin capacity, which ranges from about two to sixitems in adults (and fewer in children), and theindividual capacity limit is a strong correlate ofcognitive aptitude.

Another way to take into account the role ofmulti-item chunk formation is to set up the task ina manner that allows chunks to be observed.Tulving and Patkau (1962) studied free recall ofword lists with various levels of structure, rangingfrom random words to well-formed Englishsentences, with several different levels of coherencein between. A chunk was defined as a series ofwords reproduced by the participant in the sameorder in which the words had been presented. Itwas estimated that, in all conditions, participants

recalled an average of four to six chunks. Cowanet al. (2004) tried to refine that method by testingserial recall of eight-word lists, which werecomposed of four pairs of words that previouslyhad been associated with various levels of learning(0, 1, 2, or 4 prior word–word pairings). Eachword used in the list was presented an equalnumber of times (four, except in a non-studiedcontrol condition) but what varied was how manyof those presentations were as singletons and howmany were as a consistent pairing. The number ofpaired prior exposures was held constant acrossthe four pairs in a list. A mathematical model wasused to estimate the proportion of recalled pairsthat could be attributed to the learned association(i.e., to a two-word chunk) as opposed to separaterecall of the two words in a pair. This modelsuggested that the capacity limit was about 3.5chunks in every learning condition, but that theratio of two-word chunks to one-word chunksincreased as a function of the number of priorexposures to the pairs in the list.

Overcoming contamination from rehearsal

The issue of rehearsal is not entirely separate fromthe issue of chunk formation. In the traditionalconcept of rehearsal (e.g., Baddeley, 1986), oneimagines that the items are covertly articulated inthe presented order at an even pace. However,another possibility is that rehearsal involves theuse of articulatory processes in order to put theitems into groups. In fact, Cowan et al. (2006a)asked participants in a digit span experiment howthey carried out the task and by far the mostcommon answer among adults was that theygrouped the items; participants rarely mentionedsaying the items to themselves. Yet, it is clear thatsuppressing rehearsal affects performance.

Presumably, the situations in which itemscannot be rehearsed are for the most part thesame as the situations in which items cannot begrouped. For example, Cowan et al. (2005) reliedon a running memory span procedure in which theitems were presented at the rapid rate of 4 persecond. At that rate, it is impossible to rehearse theitems as they are presented. Instead, the task is

331

probably accomplished by retaining a passive store(sensory or phonological memory) and thentransferring the last few items from that store intoa more attention-related store at the time of recall.In fact, with a fast presentation rate in runningspan, instructions to rehearse the items is detri-mental, not helpful, to performance (Hockey,1973). Another example is memory for lists thatwere ignored at the time of their presentation(Cowan et al., 1999). In these cases, the capacitylimit is close to the three or four items suggested byBroadbent (1975) and Cowan (2001).

It is still quite possible that there is a speech-based short-term storage mechanism that is by andlarge independent of the chunk-based mechanism.In terms of the popular model of Baddeley (2000),the former is the phonological loop and the latter,the episodic buffer. In terms of Cowan (1988,1995, 1999, 2005), the former is part of activatedmemory, which may have a time limit due todecay, and the latter is the focus of attention,which is assumed to have a chunk capacity limit.

Chen and Cowan (2005) showed that the timelimit and chunk capacity limit in short-termmemory are separate. They repeated the proce-dure of Cowan et al. (2004) in which pairs ofwords sometimes were presented in a trainingsession preceding the list recall test. They com-bined lists composed of pairs as in that study.Now, however, both free and serial recall taskswere used, and the length of list varied. For longlists and free recall, the chunk capacity limitgoverned the recall. For example, lists of six well-learned pairs were recalled as well as lists of sixunpaired singletons (i.e., were recalled at similarproportions of words correct). For shorter listsand serial recall strictly scored, the time limitinstead governed the recall. For example, lists offour well-learned pairs were not recalled nearly aswell as lists of four unpaired singletons, but onlyas well as lists of eight unpaired singletons. Forintermediate conditions it appeared as if chunkcapacity limits and time limits operate together togovern recall. Perhaps the capacity-limitedmechanism holds items and the rehearsal mecha-nism preserves some serial order memory forthose held items. The exact way in which theselimits work together is not yet clear.

Overcoming contamination from non-capacity-

limited types of storage

It is difficult to demonstrate a true capacity limitthat is related to attention if, as I believe, there areother types of short-term memory mechanismsthat complicate the results. A general capacityshould include chunks of information of all sorts:for example, information derived from bothacoustic and visual stimuli, and from both verbaland nonverbal stimuli. If this is the case, thereshould be cross-interference between one type ofmemory load and another. However, the literatureoften has shown that there is much moreinterference between similar types of memoranda,such as two visual arrays of objects or twoacoustically presented word lists, than there isbetween two dissimilar types, such as one visualarray and one verbal list. Cocchini et al. (2002)suggested that there is little or no interferencebetween dissimilar lists. If so, that would appear toprovide an argument against the presence of ageneral, cross-domain, short-term memory store.

Morey and Cowan (2004, 2005) questioned thisconclusion. They presented a visual array of coloredspots to be compared to a second array thatmatched the first or differed from it in one spot’scolor. Before the first array or just after it,participants sometimes heard a list of digits thatwas then to be recited between the two arrays. In alow-load condition, the list was their own seven-digit telephone number whereas, in a high-loadcondition, it was a random seven-digit number.Only the latter condition interfered with array-comparison performance, and then only if the listwas to be recited aloud between the arrays. Thissuggests that retrieving seven random digits in a waythat also engages rehearsal processes relies uponsome type of short-term memory mechanism thatalso is needed for the visual arrays. That sharedmechanism may be the focus of attention, with itscapacity limit. Apparently, though, if the list wasmaintained silently rather than being recited aloud,this silent maintenance occurred without much useof the common, attention-based storage mechanism,so visual array performance was not much affected.

The types of short-term memory whose con-tribution to recall may obscure the capacity limit

332

can include any types of activated memory that falloutside of the focus of attention. In the modelingframework depicted in Fig. 1, this can includesensory memory features as well as semanticfeatures. Sperling (1960) famously illustrated thedifference between unlimited sensory memory andcapacity-limited categorical memory. If an array ofcharacters was followed by a partial report cueshortly after the array, most of the characters inthe cued row could be recalled. If the cue wasdelayed about 1 s, most of the sensory informationhad decayed and performance was limited to aboutfour characters, regardless of the size of the array.Based on this study, the four-character limit couldbe seen as either a limit in the capacity of short-term memory or a limit in the rate with whichinformation could be transferred from sensorymemory into a categorical form before it decayed.However, Darwin et al. (1972) carried out ananalogous auditory experiment and found a limitof about four items even though the observeddecay period for sensory memory was about 4 s.Given the striking differences between Sperlingand Darwin et al. in the time period available forthe transfer of information to a categorical form,the common four-item limit is best viewed as acapacity limit rather than a rate limit.

Saults and Cowan (2007) tested this conceptualframework in a series of experiments in whicharrays were presented in two modalities at once or,in another procedure, one after the other. A visualarray of colored spots was supplemented by anarray of spoken digits occurring in four separateloudspeakers, each one consistently assigned to adifferent voice to ease perception. On some trials,participants knew that they were responsible forboth modalities at once whereas, in other trials,participants knew that they were responsible foronly the visual or only the acoustic stimuli. Theyreceived a probe array that was the same as theprevious array (or the same as one modality inthat previous array) or differed from the previousarray in the identity of one stimulus. The task was todetermine if there was a change. The use of cross-modality, capacity-limited storage predicts a parti-cular pattern of results. It predicts that performanceon either modality should be diminished in thedual-modality condition compared to the unimodal

conditions, due to strain on the cross-modalitystore. That is how the results turned out. Moreover,if the cross-modality, capacity-limited store werethe only type of storage used, then the sum ofvisual and auditory capacities in the dual-modalitycondition should be no greater than the larger ofthe two unimodal capacities (which happened tobe the visual capacity). The reason is that thelimited-capacity store would hold the same num-ber of units no matter whether they were allfrom one modality or were from two modalitiescombined. That prediction was confirmed, butonly if there was a post-perceptual mask inboth modalities at once following the array to beremembered. The post-perceptual mask included amulticolored spot at each visual object locationand a sound composed of all possible digitsoverlaid, from each loudspeaker. It was presentedlong enough after the arrays to be recalled thattheir perception would have been complete (e.g.,1 s afterward; cf. Vogel et al., 2006). Presumably,the mask was capable of overwriting various typesof sensory-specific features in activated memory,leaving behind only the more generic, categoricalinformation present in the focus of attention,which presumably is protected from maskinginterference by the attention process. The limit ofthe focus of attention was again shown to bebetween three and four items, for either unimodalvisual or bimodal stimuli.

Even without using masking stimuli, it may bepossible to find a phase of the short-term memoryprocess that is general across domains. Cowan andMorey (2007) presented two stimulus sets to berecalled (or, in control conditions, only one set).The two stimulus sets could include two spokenlists of digits, two spatial arrays of colored spots,or one of each, in either order. Following thispresentation, a cue indicated that the participantwould be responsible for only the first array, onlythe second array, or both arrays. Three secondsfollowed before a probe. The effect of memoryload could be compared in two ways. Performanceon those trials in which two sets of stimuli werepresented and both were cued for retention couldbe compared either to trials in which only one setwas presented, or it could be compared to trials inwhich both sets were presented but the cue later

333

indicated that only one set had to be retained. Thepart of working memory preceding the cue showedmodality-specific dual-task effects: encoding astimulus set of one type was hurt more by alsoencoding another set if both sets were in the samemodality. However, the retention of informationfollowing the cue showed dual-task effects thatwere not modality-specific. When two sets hadbeen presented, retaining both of them wasdetrimental compared to retaining only one set(as specified by the post-stimulus retention cue toretain one versus both sets), and this dual-taskeffect was similar in magnitude no matter whetherthe sets were in the same or different modalities.After the initial encoding, working memorystorage across several seconds thus may occurabstractly, in the focus of attention.

Other evidence for a separate short-term storage

Last, there is other evidence that does not directlysupport either temporal decay or a capacity limitspecifically, but implies that one or the other ofthese limits exist. Bjork and Whitten (1974) andTzeng (1973) made temporal distinctiveness argu-ments on the basis of what is called continualdistractor list recall, in which a recency effectpersists even when the list is followed by adistracter-filled delay before recall. The filled delayshould have destroyed short-term memory but therecency effect occurs anyway, provided that theitems in the list also are separated by distracter-filled delays to increase their distinctiveness fromone another. In favor of short-term storage,though, other studies have shown dissociationsbetween what is found in ordinary immediaterecall versus continual distractor recall (e.g., wordlength effects reversed in continual distractorrecall: Cowan et al., 1997b; proactive interferenceat the most recent list positions in continualdistractor recall only: Craik & Birtwistle, 1971;Davelaar et al., 2005).

There is also additional neuroimaging evidencefor short-term storage. Talmi et al. (2005) foundthat recognition of earlier portions of a list, butnot the last few items, activated areas within thehippocampal system that is generally associatedwith long-term memory retrieval. This is consistent

with the finding, mentioned earlier, that memory forthe last few list items is spared in Korsakoff’samnesia (Baddeley andWarrington, 1970; Carlesimoet al., 1995). In these studies, the part of the recencyeffect based on short-term memory could reflect ashort amount of time between presentation andrecall of the last few items, or it could reflect theabsence of interference between presentation andrecall of the last few items. Thus, we can say thatshort-term memory exists, but often without greatclarity as to whether the limit is a time limit or achunk capacity limit.

The short-term memory/working memorydistinction

The distinction between short-term memory andworking memory is clouded in a bit of confusionbut that is largely the result of different investiga-tors using different definitions. Miller et al. (1960)used the term ‘‘working memory’’ to refer totemporary memory from a functional standpoint,so from their point of view there is no cleardistinction between short-term and working mem-ory. Baddeley and Hitch (1974) were fairlyconsistent with this definition but overlaid somedescriptions on the terms that distinguished them.They thought of short-term memory as the unitaryholding place as described by, for example,Atkinson and Shiffrin (1968). When they realizedthat the evidence actually was consistent with amulti-component system that could not be reducedto a unitary short-term store, they used the termworking memory to describe that entire system.Cowan (1988) maintained a multi-componentview, like Baddeley and Hitch, but without acommitment to precisely their components;instead, the basic subdivisions of working memorywere said to be the short-term storage components(activated memory along with the focus of atten-tion within it, shown in Fig. 1) and centralexecutive processes that manipulate stored infor-mation. By Cowan’s account, Baddeley’s (1986)phonological loop and visuospatial sketchpadwould be viewed as just two of many aspectsof activated memory, which are susceptible tointerference to a degree that depends upon the

334

similarity between features of the activated andinterfering information sources. Baddeley’s (2000)episodic buffer is possibly the same as theinformation saved in Cowan’s focus of attention,or at least is a closely similar concept.

There has been some shift in the definition ordescription of working memory along with a shiftin the explanation of why the newer workingmemory tasks correlate with intelligence andaptitude measures so much more highly than dosimple, traditional, short-term memory tasks suchas serial recall. Daneman and Carpenter (1980)had assumed that what is critical is to use workingmemory tasks that include both storage andprocessing components, so as to engage all of theparts of working memory as described, forexample, by Baddeley and Hitch (1974). Instead,Engle et al. (1999) and Kane et al. (2001) proposedthat what is critical is whether the workingmemory task is challenging in terms of the controlof attention. For example, Kane et al. found thatworking memory span storage-and-processingtasks correlates well with the ability to inhibit thenatural tendency to look toward a suddenlyappearing stimulus and instead to look the otherway, the antisaccade task. Similarly, Conway et al.(2001) found that individuals scoring high onstorage-and-processing tests of working memorynotice their names in a channel to be ignored indichotic listening much less often than low-spanindividuals; the high-span individuals apparentlyare better able to make their primary taskperformance less vulnerable to distraction, butthis comes at the expense of being a bit obliviousto irrelevant aspects of their surroundings. Inresponse to such research, Engle and colleaguessometimes used the term working memory to referonly to the processes related to controllingattention. By doing so, their definition of workingmemory seems at odds with previous definitionsbut that new definition allows the simple statementthat working memory correlates highly withaptitudes, whereas short-term memory (redefinedto include only the non-attention-related aspects ofmemory storage) does not correlate so highly withaptitudes.

Cowan et al. (2006b), while adhering to themore traditional definition of working memory,

made an assertion about working memory similarto that of Engle and colleagues, but a bit morecomplex. They proposed, on the basis of somedevelopmental and correlational evidence, thatmultiple functions of attention are relevant toindividual differences in aptitudes. The control ofattention is relevant, but there is an independentcontribution from the number of items that can beheld in attention, or its scope. According to thisview, what may be necessary for a workingmemory procedure to correlate well with cognitiveaptitudes is that the task must prevent covertverbal rehearsal so that the participant must relyon more attention-demanding processing and/orstorage to carry out the task. Cowan et al. (2005)found that the task can be much simpler than thestorage-and-processing procedures. For example,in a version of the running memory span test,digits are presented very quickly and the seriesstops at an unpredictable point, after which theparticipant is to recall as many items as possiblefrom the end of the list. Rehearsal is impossibleand, when the list ends, information presumablymust be retrieved from activated sensory orphonological features into the focus of attention.This type of task correlated with aptitudes, as didseveral other measures of the scope of attention(Cowan et al., 2005, 2006b). In children too youngto use covert verbal rehearsal (unlike olderchildren and adults), even a simple digit span taskserved as an excellent correlate with aptitudes.

Other research verifies this idea that a workingmemory test will correlate well with cognitiveaptitudes to the extent that it requires thatattention be used for storage and/or processing.Gavens and Barrouillet (2004) carried out adevelopmental study in which they controlled thedifficulty and duration of a processing task thatcame between items to be recalled. There still was adevelopmental difference in span, which theyattributed to the development of a basic capacity,which could reflect a developmental increase in thescope of attention (cf. Cowan et al., 2005). Lépineet al. (2005) showed that what was important for astorage-and-processing type of span task tocorrelate well with aptitudes is for the processingcomponent of the task (in this case, reading lettersaloud) to occur quickly enough to prevent various

335

types of rehearsal to sneak in between (see alsoConlin et al., 2005).

Several papers have pitted storage and proces-sing (perhaps the scope versus control of atten-tion?) against one another to see which is moreimportant in accounting for individual differences.Vogel et al. (2005) used a visual array taskmodified for use with a component of event-related potentials that indicates storage in visualworking memory, termed contralateral delayactivity (CDA). This activity was found to dependnot only on the number of relevant objects in thedisplay (e.g., red bars at varying angles to beremembered), but sometimes also on the numberof irrelevant objects to be ignored (e.g., blue bars).For high-span individuals, the CDA for tworelevant objects was found to be similar whetheror not there also were two irrelevant objects in thedisplay. However, for low-span individuals, theCDA for two relevant objects combined with twoirrelevant objects was similar to the CDA fordisplays with four relevant objects alone, as if theirrelevant objects could not be excluded fromworking memory. One limitation of the study isthat the separation of participants into high versuslow span was based on the CDA also, and the taskused to measure the CDA inevitably requiredselective attention (to one half of the display) onevery trial, whether or not it included objects of anirrelevant color.

Gold et al. (2006) investigated similar issues in abehavioral design, and testing the differencebetween schizophrenic patients and normal con-trol participants. Each trial started with a cue toattend to one part of the display at the expense ofanother (e.g., bars of one relevant color but notanother, irrelevant color). The probe display was aset that was cued for relevance on most trials (insome experiments, 75%) whereas, occasionally,the probe display was a set that was not cued. Thisallowed a separate measure of the control ofattention (the advantage for cued items overuncued items) and the storage capacity of workingmemory (the mean number of items recalled fromeach array, adding across cued and uncued sets).Unlike the initial expectations, the clear result wasthat the difference between groups was in thecapacity, not in the control of attention. It would

be interesting to know whether the same type ofresult could be obtained for high versus low spannormal individuals, or whether that comparisoninstead would show a control-of-attention differ-ence between these groups as Vogel et al. (2005)must predict. Friedman et al. (2006) found that notall central executive functions correlated withaptitudes; updating working memory did, butinhibition and shifting of attention did not. Onthe other hand, recall that Cowan et al. (2006b) didfind was that a control-of-attention task wasrelated to aptitudes.

In sum, the question of whether short-termmemory and working memory are different may bea matter of semantics. There are clearly differencesbetween simple serial recall tasks that do notcorrelate very well with aptitude tests in adults,and other tasks requiring memory and processing,or memory without the possibility of rehearsal,that correlate much better with aptitudes. Whetherto use the term working memory for the latter setof tasks, or whether to reserve that term for theentire system of short-term memory preservationand manipulation, is a matter of taste. The moreimportant, substantive question may be why sometasks correlate with aptitude much better thanothers.

Conclusion

The distinction between long-term and short-termmemory depends on whether it can be demon-strated that there are properties specific to short-term memory; the main candidates includetemporal decay and a chunk capacity limit. Thequestion of decay is still pretty much open todebate, whereas there is growing support for achunk capacity limit. These limits were discussedin a framework shown in Fig. 1.

The distinction between short-term memory andworking memory is one that depends on thedefinition that one accepts. Nevertheless, thesubstantive question is why some tests of memoryover the short term serve as some of the bestcorrelates of cognitive aptitudes, whereas othersdo not. The answer seems to point to theimportance of an attentional system used both for

336

processing and for storage. The efficiency of thissystem and its use in working memory seem todiffer substantially across individuals (e.g., Conwayet al., 2002; Kane et al., 2004; Cowan et al., 2005,2006b), as well as improving with development inchildhood (Cowan et al., 2005, 2006b) and decliningin old age (Naveh-Benjamin et al., 2007; Stoltzfus etal., 1996; Cowan et al., 2006c).

Acknowledgment

This work was completed with the assistance ofNIH Grant R01 HD-21338.

References

Atkinson, R.C. and Shiffrin, R.M. (1968) Human memory: a

proposed system and its control processes. In: Spence K.W.

and Spence J.T. (Eds.), The Psychology of Learning and

Motivation: Advances in Research and Theory, Vol. 2.

Academic Press, New York, pp. 89–195.

Baddeley, A. (2000) The episodic buffer: a new component of

working memory? Trends Cogn. Sci., 4: 417–423.

Baddeley, A.D. (1986) Working Memory. Oxford Psychology

Series No. 11. Clarendon Press, Oxford.

Baddeley, A.D. and Hitch, G. (1974) Working memory. In:

Bower G.H. (Ed.), The Psychology of Learning and

Motivation, Vol. 8. Academic Press, New York, pp. 47–89.

Baddeley, A.D. and Scott, D. (1971) Short-term forgetting in

the absence of proactive inhibition. Q. J. Exp. Psychol., 23:

275–283.

Baddeley, A.D., Thomson, N. and Buchanan, M. (1975) Word

length and the structure of short-term memory. J. Verbal

Learn. Verbal Behav., 14: 575–589.

Baddeley, A.D. and Warrington, E.K. (1970) Amnesia and the

distinction between long- and short-term memory. J. Verbal

Learn. Verbal Behav., 9: 176–189.

Barrouillet, P., Bernardin, S. and Camos, V. (2004) Time

constraints and resource sharing in adults’ working memory

spans. J. Exp. Psychol.: Gen., 133: 83–100.

Barrouillet, P., Bernardin, S., Portrat, S., Vergauwe, E. and

Camos, V. (2007) Time and cognitive load in working

memory. J. Exp. Psychol. Learn. Mem. Cogn., 33: 570–585.

Bjork, R.A. and Whitten, W.B. (1974) Recency-sensitive

retrieval processes in long-term free recall. Cogn. Psychol.,

6: 173–189.

Broadbent, D.E. (1958) Perception and Communication.

Pergamon Press, New York.

Broadbent, D.E. (1975) The magic number seven after fifteen

years. In: Kennedy A. and Wilkes A. (Eds.), Studies in Long-

Term Memory. Wiley, Oxford, England, pp. 3–18.

Brown, G.D.A., Preece, T. and Hulme, C. (2000) Oscillator-

based memory for serial order. Psychol. Rev., 107: 127–181.

Bunting, M.F., Cowan, N. and Saults, J.S. (2006) How does

running memory span work? Q. J. Exp. Psychol., 59: 1691–

1700.

Carlesimo, G.A., Sabbadini, M., Fadda, L. and Caltagirone, C.

(1995) Different components in word-list forgetting of pure

amnesics, degenerative demented and healthy subjects.

Cortex, 31: 735–745.

Chen, Z. and Cowan, N. (2005) Chunk limits and length limits

in immediate recall: a reconciliation. J. Exp. Psychol. Learn.

Mem. Cogn., 31: 1235–1249.

Cocchini, G., Logie, R.H., Della Sala, S., MacPherson, S.E.

and Baddeley, A.D. (2002) Concurrent performance of two

memory tasks: evidence for domain-specific working memory

systems. Mem. Cogn., 30: 1086–1095.

Conlin, J.A., Gathercole, S.E. and Adams, J.W. (2005)

Children’s working memory: investigating performance limi-

tations in complex span tasks. J. Exp. Child Psychol., 90:

303–317.

Conway, A.R.A., Cowan, N. and Bunting, M.F. (2001) The

cocktail party phenomenon revisited: the importance of

working memory capacity. Psychon. Bull. Rev., 8: 331–335.

Conway, A.R.A., Cowan, N., Bunting, M.F., Therriault, D.J.

and Minkoff, S.R.B. (2002) A latent variable analysis of

working memory capacity, short-term memory capacity,

processing speed, and general fluid intelligence. Intelligence,

30: 163–183.

Conway, A.R.A., Kane, M.J., Bunting, M.F., Hambrick, D.Z.,

Wilhelm, O. and Engle, R.W. (2005) Working memory span

tasks: a methodological review and user’s guide. Psychon.

Bull. Rev., 12: 769–786.

Cowan, N. (1988) Evolving conceptions of memory storage,

selective attention, and their mutual constraints within the

human information processing system. Psychol. Bull., 104:

163–191.

Cowan, N. (1992) Verbal memory span and the timing of

spoken recall. J. Mem. Lang., 31: 668–684.

Cowan, N. (1995) Attention and memory: an integrated

framework. Oxford Psychology Series, No. 26. Oxford

University Press, New York.

Cowan, N. (1999) An embedded-processes model of working

memory. In: Miyake A. and Shah P. (Eds.), Models of

Working Memory: Mechanisms of Active Maintenance and

Executive Control. Cambridge University Press, Cambridge,

UK, pp. 62–101.

Cowan, N. (2001) The magical number 4 in short-term memory:

a reconsideration of mental storage capacity. Behav. Brain

Sci., 24: 87–185.

Cowan, N. (2005) Working memory capacity. Psychology

Press, Hove, East Sussex, UK.

Cowan, N. and Aubuchon, A. M. (in press) Short-term memory

loss over time without retroactive stimulus interference.

Psychon. Bull. Rev.

Cowan, N., Chen, Z. and Rouder, J.N. (2004) Constant

capacity in an immediate serial-recall task: a logical sequel

to Miller (1956). Psychol. Sci., 15: 634–640.

Cowan, N., Elliott, E.M., Saults, J.S., Morey, C.C., Mattox, S.,

Hismjatullina, A. and Conway, A.R.A. (2005) On the

337

capacity of attention: its estimation and its role in

working memory and cognitive aptitudes. Cogn. Psychol.,

51: 42–100.

Cowan, N., Elliott, E.M., Saults, J.S., Nugent, L.D., Bomb, P.

and Hismjatullina, A. (2006a) Rethinking speed theories of

cognitive development: increasing the rate of recall without

affecting accuracy. Psychol. Sci., 17: 67–73.

Cowan, N., Fristoe, N.M., Elliott, E.M., Brunner, R.P. and

Saults, J.S. (2006b) Scope of attention, control of attention,

and intelligence in children and adults. Mem. Cogn., 34:

1754–1768.

Cowan, N. and Morey, C.C. (2007) How can dual-task working

memory retention limits be investigated? Psychol. Sci., 18:

686–688.

Cowan, N., Naveh-Benjamin, M., Kilb, A. and Saults, J.S.

(2006c) Life-span development of visual working memory:

when is feature binding difficult? Dev. Psychol., 42:

1089–1102.

Cowan, N., Nugent, L.D., Elliott, E.M., Ponomarev, I. and

Saults, J.S. (1999) The role of attention in the development of

short-term memory: age differences in the verbal span of

apprehension. Child Dev., 70: 1082–1097.

Cowan, N., Saults, J.S. and Nugent, L.D. (1997a) The role of

absolute and relative amounts of time in forgetting within

immediate memory: the case of tone pitch comparisons.

Psychon. Bull. Rev., 4: 393–397.

Cowan, N., Wood, N.L., Nugent, L.D. and Treisman, M.

(1997b) There are two word length effects in verbal short-

term memory: opposed effects of duration and complexity.

Psychol. Sci., 8: 290–295.

Craik, F.I.M. and Birtwistle, J. (1971) Proactive inhibition in

free recall. J. Exp. Psychol., 91: 120–123.

Crowder, R.G. (1982) The demise of short-term memory. Acta

Psychol., 50: 291–323.

Crowder, R.G. (1993) Short-term memory: where do we stand?

Mem. Cogn., 21: 142–145.

Daneman, M. and Carpenter, P.A. (1980) Individual differences

in working memory and reading. J Verbal Learn. Verbal

Behav., 19: 450–466.

Daneman, M. and Merikle, P.M. (1996) Working memory and

language comprehension: a meta-analysis. Psychon. Bull.

Rev., 3: 422–433.

Darwin, C.J., Turvey, M.T. and Crowder, R.G. (1972) An

auditory analogue of the Sperling partial report procedure:

evidence for brief auditory storage. Cogn. Psychol., 3:

255–267.

Davelaar, E.J., Goshen-Gottstein, Y., Ashkenazi, A., Haar-

man, H.J. and Usher, M. (2005) The demise of short-term

memory revisited: empirical and computational investiga-

tions of recency effects. Psychol. Rev., 112: 3–42.

Ebbinghaus, H. (1885/1913) Memory: a contribution to

experimental psychology. Translated by H.A. Ruger and

C.E. Bussenius. Teachers College, Columbia University,

New York. (Originally in German, Ueber das gedächtnis:

Untersuchen zur experimentellen psychologie).

Engle, R.W. (2002) Working memory capacity as executive

attention. Curr. Dir. Psychol. Sci., 11: 19–23.

Engle, R.W., Tuholski, S.W., Laughlin, J.E. and Conway,

A.R.A. (1999) Working memory, short-term memory, and

general fluid intelligence: a latent-variable approach. J. Exp.

Psychol. Gen., 128: 309–331.

Ericsson, K.A. and Kintsch, W. (1995) Long-term working

memory. Psychol. Rev., 102: 211–245.

Friedman, N.P., Miyake, A., Corley, R.P., Young, S.E.,

DeFries, J.C. and Hewitt, J.K. (2006) Not all execu-

tive functions are related to intelligence. Psychol. Sci., 17:

172–179.

Gavens, N. and Barrouillet, P. (2004) Delays of retention,

processing efficiency, and attentional resources in

working memory span development. J. Mem. Lang., 51:

644–657.

Glanzer, M. and Cunitz, A.R. (1966) Two storage mechanisms

in free recall. J. Verbal Learn. Verbal Behav., 5: 351–360.

Glenberg, A.M. and Swanson, N.C. (1986) A temporal distin-

ctiveness theory of recency and modality effects. J. Exp.

Psychol. Learn. Mem. Cogn., 12: 3–15.

Gold, J.M., Fuller, R.L., Robinson, B.M., McMahon, R.P.,

Braun, E.L. and Luck, S.J. (2006) Intact attentional control

of working memory encoding in schizophrenia. J. Abnorm.

Psychol., 115: 658–673.

Greene, R.L. (1996) The influence of experimental design: the

example of the Brown–Peterson paradigm. Can. J. Exp.

Psychol., 50: 240–242.

Guttentag, R.E. (1984) The mental effort requirement of

cumulative rehearsal: a developmental study. J. Exp. Child

Psychol., 37: 92–106.

Hebb, D.O. (1949) Organization of Behavior. Wiley, New York.

Hockey, R. (1973) Rate of presentation in running memory and

direct manipulation of input-processing strategies. Q. J. Exp.

Psychol. A, 25: 104–111.

James, W. (1890) The Principles of Psychology. Henry Holt,

New York.

Kane, M.J., Bleckley, M.K., Conway, A.R.A. and Engle, R.W.

(2001) A controlled-attention view of working-memory

capacity. J. Exp. Psychol. Gen., 130: 169–183.

Kane, M.J., Hambrick, D.Z., Tuholski, S.W., Wilhelm, O.,

Payne, T.W. and Engle, R.E. (2004) The generality of

working-memory capacity: a latent-variable approach to

verbal and visuo-spatial memory span and reasoning. J. Exp.

Psychol. Gen., 133: 189–217.

Keppel, G. and Underwood, B.J. (1962) Proactive inhibition in

short-term retention of single items. J. Verbal Learn. Verbal

Behav., 1: 153–161.

Kyllonen, P.C. and Christal, R.E. (1990) Reasoning ability is

(little more than) working-memory capacity? Intelligence, 14:

389–433

Lépine, R., Barrouillet, P. and Camos, V. (2005) What makes

working memory spans so predictive of high level cognition?

Psychon. Bull. Rev., 12: 165–170.

Lewandowsky, S., Duncan, M. and Brown, G.D.A. (2004)

Time does not cause forgetting in short-term serial recall.

Psychon. Bull. Rev., 11: 771–790.

McGeoch, J.A. (1932) Forgetting and the law of disuse.

Psychol. Rev., 39: 352–370.

338

Melton, A.W. (1963) Implications of short-term memory for a

general theory of memory. J. Verbal Learn. Verbal Behav., 2:

1–21.

Miller, G.A. (1956) The magical number seven, plus or minus

two: some limits on our capacity for processing information.

Psychol. Rev., 63: 81–97.

Miller, G.A. (1989) George A. Miller. In: Lindzey G. (Ed.), A

History of Psychology in Autobiography, Vol. VIII. Stanford

University Press, Stanford, CA, pp. 391–418.

Miller, G.A., Galanter, E. and Pribram, K.H. (1960) Plans and

the structure of behavior. Holt, Rinehart and Winston, Inc.,

New York.

Morey, C.C. and Cowan, N. (2004) When visual and verbal

memories compete: evidence of cross-domain limits in work-

ing memory. Psychon. Bull. Rev., 11: 296–301.

Morey, C.C. and Cowan, N. (2005) When do visual and verbal

memories conflict? The importance of working-memory load

and retrieval. J. Exp. Psychol. Learn. Mem. Cogn., 31: 703–

713.

Nairne, J.S. (2002) Remembering over the short-term: the case

against the standard model. Annu. Rev. Psychol., 53: 53–81.

Naveh-Benjamin, M., Cowan, N., Kilb, A. and Chen, Z. (2007)

Age-related differences in immediate serial recall: dissociating

chunk formation and capacity. Mem. Cognit., 35: 724–737.

Neath, I. and Surprenant, A. (2003) Human memory (2nd ed.).

Wadsworth, Belmont, CA.

Nipher, F.E. (1878) On the distribution of errors in numbers

written from memory. Trans. Acad. Sci. St. Louis, 3: ccx–

ccxi.

Peterson, L.R. and Peterson, M.J. (1959) Short-term retention

of individual verbal items. J. Exp. Psychol., 58: 193–198.

Saults, J.S. and Cowan, N. (2007) A central capacity limit to the

simultaneous storage of visual and auditory arrays in

working memory. J. Exp. Psychol., 136: 663–684.

Sperling, G. (1960) The information available in brief visual

presentations. Psychol. Monogr., 74. (Whole No. 498.)

Stoltzfus, E.R., Hasher, L. and Zacks, R.T. (1996) Working

memory and retrieval: an inhibition-resource approach. In:

Richardson J.T.E., Engle R.W., Hasher L., Logie R.H.,

Stoltzfus E.R. and Zacks R.T. (Eds.), Working Memory and

Human Cognition. Oxford University Press, New York,

pp. 66–88.

Talmi, D., Grady, C.L., Goshen-Gottstein, Y. and Moscovitch,

M. (2005) Neuroimaging the serial position curve: a test of

single-store versus dual-store models. Psychol. Sci., 16:

716–723.

Tulving, E. and Patkau, J.E. (1962) Concurrent effects of

contextual constraint and word frequency on immediate recall

and learning of verbal material. Can. J. Psychol., 16: 83–95.

Turvey, M.T., Brick, P. and Osborn, J. (1970) Proactive

interference in short-term memory as a function of prior-

item retention interval. Q. J. Exp. Psychol., 22: 142–147.

Tzeng, O.J.L. (1973) Positive recency effect in a delayed free

recall. J. Verbal Learn. Verbal Behav., 12: 436–439.

Unsworth, N. and Engle, R.W. (2007) The nature of individual

differences in working memory capacity: active maintenance

in primary memory and controlled search from secondary

memory. Psychol. Rev., 114: 104–132.

Vogel, E.K., McCollough, A.W. and Machizawa, M.G. (2005)

Neural measures reveal individual differences in controlling

access to working memory. Nature, 438: 500–503.

Vogel, E.K., Woodman, G.F. and Luck, S.J. (2006) The time

course of consolidation in visual working memory. J. Exp.

Psychol. Hum. Percept. Perform., 32: 1436–1451.

Waugh, N.C. and Norman, D.A. (1965) Primary memory.

Psychol. Rev., 72: 89–104.

Wickelgren, W.A. (1974) Single-trace fragility theory of

memory dynamics. Mem. Cogn., 2: 775–780.

What are the differences between long-term, short-term, and working memory?Historical roots of a basic scientific questionDescription of three kinds of memoryThe short-term memory/long-term memory distinctionDuration limitsOvercoming contamination from rehearsalOvercoming contamination from long-term retrievalOvercoming contamination from temporal distinctiveness

Chunk capacity limitsOvercoming contamination from chunking and the use of long-term memoryOvercoming contamination from rehearsalOvercoming contamination from non-capacity-limited types of storage

Other evidence for a separate short-term storage

The short-term memory/working memory distinctionConclusionAcknowledgmentReferences