AD -A 125 977 EXCEPTIONAL RE RORYIU) CARNEGIE-ELI.Oll WIl PITISIURtoNPA DEPT OF PSYCHOLOOS K A ERICSSON El AL. NOV 82

N ASS IFED 10--N 0048 .. 3SF/4 /10 NE

flfllflffl~fflfr7

ii.

.

i,

LU t

MICROCOPY RESOLUTION TEST CHARTNATIONJAL INgES OF $TMIOAUOS - ASS A

P 4,

i_4

---.--- ~----

I

- :4fr ~-*~ --.. - - -

Excepional Memory

K. Andrs Eftsm n d Will.. GL Chin

D~prtmn~Of Psyd~lhow

Univr*f of Colorado mid

TNt PW t

Ps*MMW In Aowf, 8defW, YOpMiO416,NsmrOsmb l.

Appwd, ftr psAf lo ftbuius wi~ft 148P m WINO Inl or I put is p -M@b for

F.T'S GRA&I(1iT' P? TAB0

but on/

~JA1it Co4l~

77,7W.-7-7-7-z

SECURITY CLASS'IrICATtOH OF TIS PAGE P 1,en Do*a Bflre,.

READ WlSTUCONSREPORT DOCUMENTATION PAGE 3E7REl COMPL9TING FORM1. REPORT HIumsER I' GOVT ACCESSION NO: S. RIIENT1mS CATALOG HNI1191R

4. TITLE (mod Subtitle) 5. TYPE OP REPORT A PERIOD COVERED

EXCEPTIONAL Me"OY Technical ReportG. PERFORINGw ORG. REPORT MUSER9

7. AUTI4Oft~a) L CONTRACT OR RANT NUMBER(A)

K. Anders Ericsson I N0Ol4 -. 02 15William G. Chase 1

S. PERORINGH ORGANIZATION NUR ANC ACORESS Ia. PROGAN LENT P FIT. TASK

Department of Psychology AREA1& WORKMNT 1= 1011

Carnegie-Mellon UniversityPittsburgh, PA 15213

1t. CONTROLLING OFFICE NAME ANDS I@OES 1. REPORT DATE

Personnel and Training Research Piograms ____________

Office of Naval Research 13. WN11SER OF PAGES

Arlington, VA 22217* -. 14. MONITORING AGENCY NAME 1k ADORUW('I Wfffuv hem Casft*flV*Oe0) I8. SECURITY CLASS. (.1 1.O MwV

IS&. gCWPICATION/O111NRADING

19. DISTRISUTION STATEMENT rof $Me [1apIn)

Approved for public release; distribution unlimited.

WT. DisTRoaiviON STATEMET (at a~u wommi in iiii a. it UEwow Amu nhpwe

ISL SUPPLEME1NTARY OEE

To appear In: American Scientist, 70(6), November-December 1982.

NannyComputer simulationhumoaica Skilled Poermzune

*Cognitive Psychology Learning

IL ASSIRACT (C ve.-s wRM ft ieZ70p ";,-*UR IV e .0680

wxrodmry feats of mamesy cam, be matched or ampassed by people with

average memories that hoe bee m awed by Mzining.

Us JAN 147 loUw euev OSOMLwCUMIVFO 11 M ove

~Kr~i!±

K. Anders Ericssn Exceptional MemoryWilliam G. Chase

Extraordinary feats of memory can be matched orsurpassed by people with average memories that havebeen improved by training

-There are scientific records of mem- and attended to, it bs kept available fats Of memory appear to violate theoyfeats that deviate so markedly for a short tim but then is irrevoca- limitation of STM and other charac-

from the normal that they are called bly lost unles it isatnetogin estcsofnrame o f ot oexceptional and are assumed to re- or rehearsed. This temporary srge these memory feat have used num-fleet a memory system structurally system is called short-term memory bars and other kinds of meaninglessdifferent from that of most people. (Sm'). The amount of information material, similar to those used to testSome -recent research involving that nbehlat one limina isT Is S in normal people. Around thememory training of normal people severely limited for normal people. turn of the century Blnet (1894)has led us to question this distinction. To be str" permanently. informa published a study of the exceptional

Wewill fias describe the empirical dion has to be placed in logtrteory of mental calculators and-evidencezreorte in support of the memory (Liii), which consists of an chess masters The calculators were

idea that exceptional memory is dif- essentially unlimited and permanent able to multiply two S-digit numbersferent from normal memory. Then base for stoing Information. Infer, mentally without external memorywe will present our research in sup. nastion in Lii can be rettred only ais and were also abl to commfitport of the assertion that n1orl b precis retrieval cuss, and failure larg matrices of digit I* memorymemory stuctre is sufficient to ex- in retrieval Ia the najormuse of lss ate a brief prsettin Accordngplain ec"eptosa Mu y how if we of. lahform. in Liii. Fmr normal .teverba ffports we thes mentaltake into aePnumt in people it requires conscious effort alculatmW the stoe h -esnpractice Sand prio experience. and considerable time to commnit digits as eihe auditory 'or visual

Let us first describe some laws uneltedl informtion to Liii In a imes thus s o"esin eiePrIm-and generail charcteristics of normal form that makes it available for re- itive copyin rcs n leading tohuman memory and then speeify tiwaL the term photrpl memory."how exceptional memory deviates Fay early in psychology, at- In the most cited ad extensivefrom and contradicts them. The temptsi were made to measure the study of exceptional imemoary. Luuiacontemporary view of the memory capacity of sm Tite mO common (196Q examined the memory of a"onqt in normal adults is that in- -procedure was thememorye newspaper reporter named S. V.

fo-owatlon can be held primtarily in task, in which an mprmne e-Seehvkl(5) for Over 20 years Stwo different storage systems: iase aaberdemtb oefle showed an exceptional ability toshort-term and long-term mmy.Ina orer The itemwar -psntdat eoias ennls nomtoWhen new informatod e rceive a fetyapi rte1Iftem pereecosad) such a nonsense sylla le.mthe-

to sainvlne the a, 'of Informs- matinal formulas, and poe in for-dent =*tavetd to LIK The Inter, etgp languages. ALbu rhkan un-

A"bwio Axntinho in ating concinusioa was tha the fortunately pywv. rof i actualAvvNewof *Wh 0 * xbf~y 1 memoq ry sI limited and Is a&p- docuumnaton ofSsmemory feats,

,C.I . , prudm i Ite the snfee m--- "ype he did publish a detailed descriptiontIN.& " -* ". ofN SAWol aroundf difeetdigit of S. memorization of a matrix con-

,-g* Ior naonants and slightly fewer (5 tming 50 digits. S looked at the ma-ANI#AI Iaw*4 DiW IV ofr' ) coleus viual pausented A-e tnix for minutes and then. after the

Ppdmhp ofwe COs~ ii Usa iwwy aft uq etIca igns. and w ille matrix was taken away, wa able toaattnaesaarw~usewvv (1946) smmedthis seerhby deibe all the informto on the

awpm*.O frm*wdlk11011 . daigtat Ut has th" apct to I a xsiit weseavmlbleto himinCkhO m I%*u P P~dx6IM rean plus ot sanm I Wmbols or a metal picture. Lurka argued thatCd~ INt4IestMMP Am. & chunks. A chunk is a cauleti" of SS exceptional memory wa stuc-

-i A A fi I nepapVC IW I vls such~ as a0 phonme number.. tuly different from neenal ands*Ab MA. ad "h. fima ' 1 ""' lbhatV t l t- -*;as a yuieall the was based on noncegntlve sensory. N o - .ItmeuueyWI - Fw

drl.*d% ~ 1td & I, 1~es 11 1 andw~t there is a A few recntstdies have usedIke 10evale OWLbAthm cunk. modern11 eJ11n4601metod to

* of Indvidual whose ecponlmance, of people With enceptiona

memory. The most notable of theme performance of cheow masters wasstudies is Hunt and Love's (1972) go, closely linked to the presence ofanalysis of VP, an excellent chess meaningful Chen paterns, patternsplayer who displayed the best mem! 00 that have become familiar with yearsory of recent subjects although he is of practice.not up to the same level as Luria's In the rest of this article, we.willS. Iargue that both exceptional and ex-

Several types of evidence have pert memory are consistent with thebeen cited in support of the idea that laws and limitations of normalexceptional memory is qualitatively 0 emory, and that all adults an de-different from normal memory. First, velop these forms of memorythere is a marked difference in per- through extensive practice. In theory,formance, a difference so large that 0 50 100 150 INU IO 30extensive practice creates a largethe exceptional feat is judged to be' P=V"Y' w knowledge base in LThO and ntew in-outside the range of normal subjets. RM L MONt m0081 PeOPO can formation can be stored efficiently inFor example, after having had their. moWe a 4Uopmc of randomS numbers a retrievable form by associating itonly as lon an? dig1ts, and own Peoplememory spant for digits measuredi, With ena~ ,to~ with familiar material in the knowl-most People find it inconceivable .bsIn mmm nU !N gi Yet sF, a edge base.that they Could ever double or triple cal'a ads utiuu m ran loog-dlstaac Our argument Is based on thretheir performance, regardless of the a so of m oucs fdaa Frt.w wl damount of practice. A second differ. imaadN n ro~e aqec;a oI15trate thast normal adults withonce concerns the processes and abn ORes Ali .bmmo)y Van ol0 modest amnounts of practice canstructurst involved in the transfer of mgnehof aml, isho.i. men i. DO, achieve memory performance thatinformation to LYM. Subject S re- am.Ihes*d who wa a WRagdistana equals the recorded performance of

a ~ b viua imgw atain ~as~ - pol with exceptional memories.ported forming his umoy span _s shown is Onr PyOseyeaiOn*h dvlpsimply looking at the digit matrix.BycoeyxaingtedvlpMost people form a verbal string by ment of such a memory skill we willusing rote rehearsal, as they do for show its relation to the limits ofother types of memorized informa. and ability to recail. Investigators normal memory. Second, we willtion such as the name of the months have demonstrated this reation for show tha the cognitive structuresand the national anthem. Further- a Imp number .of skillsing brdeand processes acquired throughmore, motpolerqie xesve ChiiftW),crui l practice cam account for exceptionaleffort and concentraton over pro- ass (FOa nd Schwartz l99Y), n expert minoty. In particular, welonged Pei ods of tim to memorize and i arhcal drawings (Akinin will compare in oe detai the per.

mean~ngs inforrastlom Uf the in- press). Sormance of our trained subjects withformation ksams he~rplal In his original studies Dntthat of peopl wihalgedly excep-copied, It ap pe unlikely that skill (1I94) sent out questionnaires to tional memtory. Finally, -we Willor other cognitive proaesses are in- chess masters, inquiring about their demonstrate that.all normal adultsvolved. The implicit assumption tois mental representation of cheow exhibit skilled and exceptionalthast, event with practice. normal gae.The chews astlers almost in- memory in a domain where they arepeoplelis ammory praues are dio- vably reported having the chess exports.

ply n powrM eoughto generate board stored as a visual image, like ateperforae adhbsd by people mental photogeaph. Howevier, K1not Acquistion of exceptionl

with enmpIoa memories. How- found sanes d1ifferences betwee- me-reyer, matnormal0 people Are Simply regular photographs and theseor

uaaeof this silsw of tech.- nta visual Imm t images did We de-de Athat it would be partic.iqe tha an be used to improve no laduede the eme color and do- ularly interesting to study the effieta

te memory. tailed features of ches plss or the of prate on the digiteopan task,A type of superio mmory re- shadoews cast by the pious and, in which Is generally used to asses the

laOe tO GNCeP001onal memory has general the Images, sme to takw capacity of a personIs sm~ In thisbeenA ebs mn gapu.s for trei en abtAci t or schesmatic chok' task, a sbjec Is nead a sequence ofdoma" of, ne Wor sef a hewbuted actor, randm digit at a rate of about Iis shown IprIfor S scdsabgneat S emoer ,achhas dgtpacn.If the subjct epeat

ch reproduice the location of definitely refte the notion that tesqeccortlthen the nextonly 4 or 9 pieces This is exactly chess ismers we able to ake a v1- sequenc is Increased by I digit;what thes liit for normal ?m svia menal copy of chessboards otherwise, the next sequence is de.woul pinile. However, a chess (Chase and Simown 19h b; de Groot cuease by I digit. The estimatedmuter con, after th se short ox- IW,4) When Investigators briefy spant of a subjec equabs the length ofposure, mou all the pleas on the showed the duemets maplOel adigt sequenc that the subje cancheesboed vitualy perfectly. Im.rni r qo espee rePat coretyhl h e

preedchess piFMs at Inteuse- on a asher the missw coud We #dminred thedgteadielvels of skil shw better-nDd rsaioth oftlmdm 4or5 task to as n erdat aF o

dwi~ha She bapnr,. bet '06cMw N nRbasw than the about I hOursday t Sdsaweek4"ie Ta Isthw- WwleO c~hies 4M with rich for 30 wmnhe'r formae tha 230

a h O~ !1o e (** madam the superioir Iomfaeoytetn.Ateg

7 -A

SF had only avraemr abilide T* hearsed to himself. It was only afterand Mvre intelligence fora college TA I1. W nt onay OfUAOV Sw* he introduced a new leVel of encod-

stdn.hsdigit span steadilyW Am. my ft ing, in which the digit groups wetuproved from?7 to around 80 digits, a aftedbo wnpm Won combined into "super groups," thattruly exceptional memory perfor- SF was able reliably to recall moremance (Fig. 1)Y. Normal subjets have OM M lan 'C1 than four groups. For example, tospans of around 7 digits and only ahilfli <12 remember 25 digits, SF normallyrarely are spans of over l0digitob. MOM1 1 < 1 grouped the digits into three groupsserved. Even individuals with alieg. L..(M < 2 of 4 digits each, three groups of 3edly exceptional memory do not %g EM Lo " pI? digits each, and a 4-digit rehearsalcome dlose to this level of perfor. I. kas 1wyj ASOm. Is group at the end. One indication of

*mance (Table I). The highest digit this grouping structure is that whenspan ever recorded previously is 18 a subject repeats the digits, there is adigits, the span of the German ages 25%, of SF's mnemonic associa- falling intonation in his voice on the

*mathematics professor Rfickle tions. last group of 4 digits and there is a(Mi41ler 1911). As soon as we discovered ST. long pause before he repats the 3-

It is important to note that dur. successful technique of associating digit group. With further practice, SF* ing this entire. study SF was in no digit sequmnces with running times, continued to introduce further levels

way coached or instructed in how to we atte te toconsutrta model of into his hierarchical storage of digitimprove. However, he was highlIy his cognitive processes. We simuist- groups until he reached his highest

* motivated and constantly tried :7i- ed thep peoe involved In recelv- memory span of 82 digits. His orga-ferent mtdsto improve his span. ig amd e^coding a digit sequence by nizatlon of 80 digits is shown inHis skill was thus self-taght. constructing a coputer program Figure 2.

In the JIMs 4 sessions of the ex. that wudtransformn digi sequencees In anothe eries of experiments,periment. SF either rehearsed the into rumnn Hims We used our we denmotrated that SF stre theseentire digit sequence or broke the computer model to generate digit digits in a retrievable form in LTM. assequence into two groups and re- -sequen-ces1 that SF would not am- shown by his ability to recall overhearsed t" second group. He also claw wit running times, When he 90% of the 200-300 digits presentedoccasionally reporte noticing nu- was faced with these uncodable se- .during an entire session. When SF

mralpoetns, ut as 654 andf 424. quences, SF's performance dropped could regulate the speed of presen-This is exactly what weobserve with alof to his beginnng leveL In an- tatlon of digits, after about 100 hours

* other normal subjects. Ouring this other experimenal session, we pro- of practice hie was able to reduce byinitial period. SF*s memory spani sented him with sequences that half the tdme he needed to memorizestayedi withina normal ronge of?7 to toub! alf Ie associated with runnin 50 digits.9 digits. .times. His performance jumped by On the basis of these and other

In the fifth sissae- SIF noticed 2211. from an average of 16 to an av- -experiments, wet Aodue that Si'sthat some digitgroupereminded him envg of 19.5 digits. memory skinl consistd of efficientof running tim for different races. This lest e=eimn also dam- and rapid storage anid retrieval of(SF was a good long-ditanc runner onstrates that SF memor span for knfomation in, LY~i. SF did not ~who compete in races; throughout digits was not unlimited eve when achieve his extraordinary- perfor-

* the eastern United States) As soon as. all the groups of 3 and 4 digits weemance by siply Improving hishe started thinking of some digit meaningful. SF was at that time abe ability to rehearse digits mentally. Ingroups a nnming times. his di~ It t remember only three ot foursuch fact, he relied on rehearsal only to

* did was begin mentally to-oncode 3- at he nd of a sequience, that he re- setdin eact sequence. He gave theand 4-digi Poop as running timefor various tose. For example, hermembered 3492a*3 minutes and49.2 seconds. near wield-rcord time

*for ruigaml."In the earlyphmo rpotehe diecovered

* only asmal nenbers( running-lmi*catogories which mat tha he had

to remember may digit seqences

ally Cora trucied an ebrate sot ofmnemonic aseoclatious based Ini-tially en running times and then

"4.3 powa old a vim OW porso ) --ur-28 -up - -unssspa..& dta .apus ru m si* amd dafts 41044 wag 'w the end of "e~j~ -~hmhlsbispuo dla h. Ile

Me"l WWa ") fe hm qec sheaue f 4disovwolA, I

B~- o M n g ti e r e m " n d ~ b u t u s

other inherently meaningless digits runner, DO, ha been able to improve digit sequences. Rcke reported di-meaningful interpretations as run- his digit span, which is currently at viding 18 digits into three groups ofning times, ages date, and so on. aboutT 5digits (see F g. 1). 6 digits and meaningfully encodingThroughout bis development of the , In this series of studies, we have each of these groups by using hismemy skill there is no evidence demonstrated that normal adults an extensive knowledge of numbers.that SF extended the limits of STm. perform outsand" gfeats ofmemor For8eample, w047ws encoded asThe largest number of digits for without extending any of the s 893 -19 X 47;.047 - 47.which he generated mnemonic as- limitations of normal memory. In From Lura's (1966) analysis of S,socations was four, and the number memory tasks in which normal we know that S mmitted nforma-of encoded digit groups in a single subjecs rely on reheamd in STM, our tion to Lm, as evidenced by his.super group" was never more than trained subjectsspidly coded the ability to recall the information days,four. Even after 20 months of prac- information meningfuly and were weeks, and years later. S reportedtice, he almost never rehearsed more able to stm it in permanent and re- that he genetew meaningful asso-than six digits to himself. We also trievable form in LTh. clations for many typea of nonsensetested his amnory span for another materials, such as foreign poems, tokind of symbo-ie., consonants- LTM t FoaeSe aid memorization. (S denied usingand found that it remained at six such associations for numbers. Weconsonants. Are the cognitive processes of will return to this bcet later.)

SWehave ebewhem rported the m f e dlugandstoagelin Mnemonic associations likeMults of experiments analyzfg r- Lnt toacimto peor- those deacribd by S have beenrors in recall and temporal patterns menm of eqpet and aceptioa known since the time of the ancientof recall that strongly support the memm telWaiey little inferma- Greek orators, who developed tech-description we have given here of tiem is available Of the cognitiv aques to aid them in memorizingSFs skill (Chase and Ericsson 191, processes used by exceptional l1ftofitemsand names.Anexample19826 ricson et al. 1960). We have subjects to commit inlsmation to of thee hniques, which have beenaiso shown that other subjects can memory. However, MAller (1911) refined and extended over theacquire the same skill with practice anlyxe in eta*l how RMdle, who centuries (Lomyne and Lum 1974;and initial coaching. Using SF's had the highest observed digitspan Yates 196), is the generation of atechniques, another long-distance ep op-In theltfesaummemodsi mental image to connect otherwise

unslated vwdL The word pair"tow, barl can be effectively re-

Fig ur t uo 1rg a U2 squ msu o membeud it om & w, an imae ofsoy en ubyA* sdmuiegl 4 I M h a cowli ng d abellmow, Im, gh. SMeboo ai sms* ohowtirly recentlympetbebeea~uaftl~ae 0 4 a thaad@oehmpa NW1 Ieahktbi d a e~e without practice normal

-w 5.dI -m fmbr. Wmad people can ae these techdques tomio. ma l mm e 3 ' 1 4 9proethermemoysinftlye,d I .iowen. fi" W (lower 12). In fa ct people do notIm WLbe" iaeq.i mah k 4 n s infoer-

1 arde" is wN-a IN waMO. a maton to memoy simply by rote

a7~ m have dst~mna~ted that normal- sujecs alostinvariably gnrte

-. _- - associtionm wen theymmelasd notse word (Mont-

__________of__INS 1g~t~W3~tlk9l. For,a" he w dm "m without

a."hMenic technsappear________fet mosfe veeions- dC metlthM peo lple nonMuly ma.

______ __ teremerim lnbm~M, md peole

to ldqes a---b

- ________ y an ot .ibaEIVORY dfferent\i-om the methdm s OW by normal

x W~~~~~pop e ldebwitsr. . . ...__W#_ of ml e sat twm

• l, l ftoe_____ iwiami m Vewi -

_____ _ pas.. prfmm7"om

the exceptional people. We have so- different ways, several of which are of the digit matrix. as his retrievallected memorization of matrices of illustrated In Figure 3. The time time were about the ame regardlessdigits as the task, because S regarded needed by several mental calculators of the order in which the digits werenumnbersas "the simplest type of anid other subje'cts to stuy and then recalled. These dat are the only ob-material" (Luria 1968, p. 60) and be- to retrieve the 25-digit matrix are jective evidence supporting Luria'scause numbers were the only kind of given in Table 2.Table 3 gives similar claim that S had a structurally uniquematerial for which S did not report results for Luria's SO-digit matrix. memory. The rest of S's memorygenerating mnemonic associations. Before we turn to a detailed performance is based on standard

discussion of these results, let us mnemonic techniquesProcess and structure of briefly report how they were onigi- We had our tainedi subject, SF,

excetioal emory naf1y interpreted. Binet (194) argued and a few other undergrauatesexcptinalthat the data shown in Table 2 sup- perform the same tasks (see Tables 2In comparing people with allegedly ported the reports of the mental cal- and 3). Furthermore, we found thatexceptional memory to our trained culators Inaudi, who claimed to en- Mfiller (1917) had collected data onnormal subjects, we are interested in code the digits as auditory symbols, normal subjects and on the mathe-

*showing not only that the perfor. and Diamondi, who claimed to en- matics professor RGc"d for Dintesmance of the two groups is coinpa- code them visuail, As iamondi was 25-digit matrix (see Table 2). We canrable. but also that their cogn~itive much faster than ma-di in retrieving see that the normal subjects tookprocesses and subsequent memory the digits. Luria (1968) argued that much longer than our trained subjectstructures-the ways- in which they the dat shown in Ta"l 3 upheld S's and the exceptional subjects to studystore inforation-are similar. The reports otgetneati at visual image the matrices. However, there is nofirst problem is. to evaluate thememory structure.

Fortunately, there are data that pip 4Aju=w Lud* .w a sedipl w. to ka the muma sibbubIGM& 5have been used to Infer the structure AemuiqasLud..b mmy wrntucwtulyuiqmbesSa"mo aulew digtasof exceptional memy. Suppose thast Weu mmmxis amy a#tM dhen dso wit -qu apaaLa subject is asked to menrie a ma-trix of digits like the oftes shown int01Figure 3 n 4 net (1894 used * * 0 0 *such a'MWM 1rc 0ur tostd tem- a 4 s 2 -

02ry of menalccIam ndL a-usedasiallarpadrtues 4 0 0

idea is quit "ipl Pa, hlkh- - -.

th matrix we caM eamine t6005SImmay srumare bysa" how, the :

Accordift toth theories thatosme o be'ctp.e armea th turn aat the ;;;muy. = aembsicalk Nn 9 a 6ways " aot t cxam tostomd In - a 9 * IF*meamy. It ain be sore, as a lis ofaudlbor syNMol. the Way moANaele tor e the - lhae or the mae a

cm be strd iUult which pVWaem -OMSO slmms UudfMsSurvaes WSW sptk tUctur. Uf the 0*Mari is stored auditoely, then the 00 000 0

informatloncan berapidly utrieved 0 0 0onlyin the same order in which it .0wa committed to memory. Retriev-

ithematrx Inany other way, such

much harder and wouldmbWho kger. .

On the Ofter had* if abSORI muisstored visually, it Should be~dltomttrieve the latomastl~iaaot. a.* *** *

any way with lib the me Seed. * ... ..

The Ilaibl bsvplpI dwst m,tnealsmv es MOSesmao S 1 * 0

To asss how MIN lrMas -_ I* * oil Pwas slosud, the subets we"e Wasavuied to recall the dlltsla mw

- - ..- *.**.y~v- 4

TaWle 2 lm On 0 Cahids ned o study tm re~ov shuB's naomh

(OmmImW NoonA MIan of

ia sOf 46 10 2 12.7 26. 226.6

om 11 7.2 6.7 41.8 24 5hindiAst fe 6.40 uW 7 3 7A .3 2.7 31.2 5mu , -dew - - U 7.0 226 33. 5OWisus so 36 23.9 16.1 64.0 71.6 25upwol stumm go 36 24.6 16.S 56.5 - 25

so 36 26.7 6.5 43.3 73.8 11OlsbsuW 18 $ a7 16.4 92.6 124.0 2

1W6bMUIa 1617

* nu eao 1 I I d . iN ewess fl sIw(6

evidence that the mental calculahos rows? Such a reult would support subjects-which would lead to theand S were aster than SPand Ilek. the idea that the expe t an retriev- low correlations in Table 5. The in-Rather, there is a clear indication to ina information from an uncoded teresting result to notice is thatthe contrary. visual Image, subjects e able to retrieve the entire

Let up now proceed to the times Wetarmthmo matrix row by row as fast as they canrequired to reue the memord tepttenof ren dotimes forany retrieve a single column.digits in different order. A visual two vslbet by calculaftng the cor- Given our knowledge of theinspection of the retrieval times re- relation coefficient. A high positive conitive processes used by trainedveals no dear, systemacdifferences correlation shows a very similar subjects, can we account for the ob-between the ecepional, tained. pattern of retrieval times for the e t of retrieval times? SFand normal sub*t Although It sub*t an4 sugls that thy e commting the matrices totook the normal subects many tim simila memory structuresr the mem. In a manner similar to the

*longer than the expert to memOrine 25iltk mLrlW sA*"y in 'Ag. I.wymwihh eoie iisithe matrices, they we abl- to re- hecoehtaomtwee all I the digitpa task. Uing his mne-trieve parts of the matrc amest as ar very l l adedive; 14 fAlh i hid" -1 1 ' he coded eachfat as the experts. From e m - 4weh ave d the a b- r o(he m aainldttal studies, we know dMtespeed of tweem th a ee etiea time dt group, and he stored each row in aretrieval is dosely relaed sothrat normal aumbM and tht of 0etha mam- " Y structure smIlar to that" stdy time spent b M t*aI re- repial * t b 2. F in.igure2.11. retrIevd the

quited for mem on Ihe w .i e theou'mti o yrw xrce h

for exmple, Newel n " RosnR- epmkf~aomsfr tedw -"ee digt orgts frm eachbloom 191) Thw, it tWbe the We- matrix d n In Pigm 4 a=i row.he men slphasewastios between diffmrnt al subjecls in T" we find thmi to rleen thenetdgit oup;oncetie of a subject tht will give u be low ad w* canevalatl a 1""up = rLtd n theInformation about th memmy (TaAle M5) Rr l m-e b a row was amt m-

* Sruct rathr than th*,a*ohum mat&i apea fA he Aou the'tame mediaeX mla69blreotswrtieval moes. To tak a xee- r rdlem of the retieval Irrde aabua frml OW nwomd subjectsaOpltvmetheexprt Ale*.uw tin. and thi dtpteno a.adwr ive b' Mkle (Mwl

Sa marx by columns fA* as by form dm is Indcated by all 1%91%"- mwellast&em eMoramt

- iTaw T- re n sm m mi im, amsi m toasi alaffio

OP IWb OM 88 V A*S" * 81 SP 8*0 AiweII fWebub

"a ie o mi u u inn it, mI

5- 46 61 'i b * 146 4.4 - 1 115 1' 4

l777 I.M . .1

- - " - N

V? (Hunt and Love 1974)We bam~v~d des&Wpiesuede Tb4 4- Coo**n g~g~W @MM6101,10ImIm reou OWS 40ms t

which uub*x Iwo, and nretlve --.- Ofma NO gWm S tthee wAkdow tow by mow a digit

of Ike nmber of times a now digi Amso 0.W SAW 0.V 0.1W 0"3 CA3Wgrouap or raw is rswed. In twluw %*"AmP "-a) 9EnS 7) 947 94n O946colum at Tablems 2=01 3 we have ~05 .W 03

*m ba babe roftlved. The .deloa imawkal g"od job ofp~

d~ctise~ev ftn TawlU1c *'<GOw asuf fewievfng th

* aul by rows. * whqi nvolves 5digit p f i.*.esr Pi ~u ng om WAG home(o practice (COim 6 Smi.s in mtwry but with practicetheinar~nbycolumms whitch he.an Simat Wb1% b). Mhoug we in lb. ofieft use of Ln. We also

wilves am arpe nd reie is* u 4Am.* A* *wxW~ aculueo beliee w at such superior ammoaygow~aclwwdkuabuttumm bmmI~Ou~.we now Seem an be. ad is, acquired by any nor-

l~f*-bo~gh~ ~ TM I'su befdA Maduh'e~h iddL -Dhcame individuals de-flora. beWenl.'bi p at of *j,,sSk. _so Awwommbkw elp skilt In a variety of domains,dions a" th - obwevdusw;I~hij~ alemdasI AIs d~fcult w ofnda met of skills fhattiMsM ate b.10 (W441 kuai'eumnd 4kMumeu~ee nursa adults hame invaiably

lastng ~ss an e~q hiherIf~w~ . s1~~squ~e evieimf aquied.However. all adults areunusuaL~~W60 mtosy~e h"upe in using their native Ian-

a", smaook u&Aaftera htmidhoumof7740 SIc end prccaninstuctionin a now

-WIeaam,- 4esw rneqo Sm eseI

OWA * a"u. 464r O mawia (Kbft*aW4k I97~ltach and

'010maeo"ae

- ~ wi.. ~--wina

* ~ ~ ~ M X11faeS64 M

p4~~~~~W1 beh ,,Qi~nSIjo I I e,." IJ to, Alpe

14 0ma~u *l It. "Of~-~'-A-- t.e~ vppj lebio Y.tdm4%! mma

40-~tJ{] 411'T"

ma d

~d-L

those in the regular memory-span mechanisms that under' .A excep- the language uage test (r- 0.50, p <test, have shown that normal subjects tional memory. 0.01). The numeric resoning test wasare in fact able to retain for a short- We designed an experiment to only weekly related to these mea-time the exact wording of sentences, evaluate the hypothesis that LTM is sures of accuracy and did not con-markedly surpassing the normal STM responsible for the superior memory tribute any additional information.limit of only 6 or so unrelated words span for sentences. For scrambled Similar results have been reported by(Aaronson and Scarborough 1977; words, little or no LTM would be ex- Daneman and Carpenter (1980).Jarvella 1971). pected. We presented intact and It appears that normal people's

In order to compare memory for scrambled sentences alternately, and memory for prose involves the sameprose directly with the standard es- asked the subjects for immediate mechanisms that underlie experttimates of memory span, Ericsson written recall after each sentence. memory. Peoples memory for proseand Karat (1981) used a procedure The major difference from earlier can exceed their sTu capacity if theyclosely analogous to the digit-span experiments was that we also unex- use their knowledge of semanticstask. They took meaningful sen- pectedly asked the subjects to recall and syntax to store information intences with different numbers of all the presented information after- LTh4. Further, one can interpret thewords from short stories and novels ward, when they were cued by a large Individual differences in proseby Steinbeck. The longer sentences unique word from each sentence. memory as due primarily to differ-were used intact and were also The main result of this experi- ences in language skill. People whoscrambled to form meaningless se- ment was that the subjects' cued re- have spent many years building upquences of the same words. These call of the intact sentences was re- their language skills have acquiredtwo types of stimuli correspond markably high, but their recall of the an extensive verbal knowledge baseclosely to the regular and scrambled scrambled sentences was virtually in LTm that can be used more effec-chessboards studied by Chase and nonexisent. In only 12 of the c.se tively to store the meaning andSimon (1973a, b). Both intact and could the subjects recall anything structure of sentences.scrambled sentences were read in the from the scrambled sentences, and in We noted earlier that feats ofsame monotone and at the same only 4% wre they able to recall more exceptional memory have been ex-speed of one word persecond. At the than a single word. In contrast, they hibited for information that is unfa-end of each sentence, the subjects recalled wods from int't sentences miliar or meaningless to normalwrote down as much of it as they 79% of the time, generally remem- subjects. Normal subjects' memorycould remember. bering more than half the presented for such Information is severely

There was a striking difference words. This result clearly indicates limited and appears to reflect somein the amount remembered between the Involvement of LTM in the supe- fixed structural limits of the cogni-the meaningful sentences and the rior mesmy for sentenmeL tive processing system for brieflyscrambled words. The subjects hd In our experiments we have also premed Information. However, weperfect recall half the time of scram- consistently found systematic indi- have shown that normal subjects can,bled sequences of only about 6 vidual differentes in the ability to through extensive practice., vastlywords, in complete agreement with recall sentences. Using traditional improve their memory for certainthe estimates we cited earlier of methods for calculating memory types of information, even surpass-memory span for unrelated words. span, we found de san for words in ing the performance of individualsAs predicted, the subjects' ability to sentences to range from 11.0 to 20.5 with allegedly exceptional memory.recall the exact words of meaningful words for different subjects. When The patterns of retrieval used bysentences was muchbettesentences we analysIed our data in terms of the these trained normal subects and byof 12 to 14 words wer recalled per- number of perfectly recalled an- people with exceptional memoriesfectly about half the time.Althugh lnces or the percent of recalled forbW matrices of dts are similarthe average percent of perfectly re- words we found reliable individual to the oceses wused by untrainedaied sentmncs decased as de differencs well normal subjects to remember mean-number of words in th sentences We conducted a final expel- ingful sentences.Increased, seveal wel-formed mi- met toexplom de relatnn between In our analysis of large differ-tences with asmenyas 26 word were the subjet' I sg skill and their enc In memory performance. we.alied perfectly bysome subjects, memory span fe ntences. LAn- have found that cerin limits remainTo take oneeample, 2 of 20 subjec*t uae il ws a edby a testof unchanged. We noted that the num-

rtledthe following 25-word m- I 1--pw &M ve"rbale- ber of chunks of infiormsif that ourtence perfedy. "Phe brudd a cloud soning. To be able to refute the iaM- subject SF kept in mind we limitedof hair out of her eyes with the back portance ofgeneral intelgllence, we to fou, regardles of whether theof her glove and left,& amuof alsogve thk subcts a test of nu- chunks correponded to digits re-earth on her cheek in dohing it. Mork reaeoin S . The number of hearsed in smid, t stored as a

In ters of the amouft maled. correctly recalled words was very graup, or ru Otes stored as athese memoy feat by "aoiual well ptedicted by both the language supergroup. mtla we her notsubectseemamo, as impmive smetest(rm 0.72<000)and the found a single ecption to this limitthose e bited by the d. maser verbal reasonit IM (ri 0.74. p < In Out may e de memory per-andit experts. The qumiis D0). Pow the number of Moely ferma ate W st!eI~ expet.

whether~mmor is am skill _ -,1-we'suwrith . e.... -me ses, w -$ - r < ft) 1ibd5 d .awA is.....WO ethel e o w o awo a *W*h OeW 55 -then balsd f'eaened WNa

subjecA., 1 (S@d I I..... ... ...- .' - -"~ *+ + + - ....I " + ++ m

developed through extensive prac. connected speech. 1. Vvmv LeeIug and mediation. 1. Exp. PsprOW. M- 29-33.tice and experience, that allows for V.15,har I&40-16 M.A1leC E 1911. ZurA~y.dSCId..rapid and efficient use of LTM. Fur- Kintsch. W. 1974.71 Th rpesgrat im of MrpiW nhstigikeit und dos VonstellungSverlaufegther, this skill is developed within in Menwar. Hlillsdale, NJ: Erlbaumn. Tedl 1. ZriISCfift fir Pssjwhoili 5.the basic abilities and limits of the lintsch. W. and T. A. Van Dljk. 1978. Toward - 1917. Zur Analyse. der Gedlichtnistfi-normal cognitive system. In every a model of text comprehension and pro- tigkeit und Vorsteilunpiveulaufes: Tell 11.recorded feat of exceptional memory duction. PsoWkoi. Re'. 85:363-94. Zeilxhnrfi fir Pswicd9.e

we have identified the same compo- Loray-c H.. and 1. Lucas. 1974. The Mn'mnrt Neell A n dsP.iS. Rosnloom.n19111

nents: the importance of prior expe- Bok oflatie prctMice.In asSkilluso and the law-

rience and practice, the availabily Luria. A. R. 196. The Mind of Mitteantit. quisiaed. ). L. Anderson. Hillsdale. NJ:ofmenigfl sscitins tg Avon. Erbam

in m andfu efficittienal sofrane Miller. G. A. 1956. The magica number seven, Prytulak. L S. 1971. Natural language media.formand ficiem retrAievgl odeln plus or minus two. Psycau. Rev. 63:81-97. ion. Coapnhse PsycW. 2:1-56.

isadqutetodscib al dutMontague, W.LJ .A. Adams, and H.. Kies. Yates, F. A. 1966. 71e Art ofMemor. London:msaeory. esrbeal dl 1966. Forgetting and natural language Routledge & Iegn Paul.

ReferencesAaronson. D., and H. S. Scarboough. 1977.

Performance theories for sentence codingSome quantitative models. I. Ve4ul endegfand Y~i eraiv mW .277-303.

Akin. 0. In palme. The PvgC~lbJ.A AdiftmUIDen~a London: Pion.

Binet. A. 1894. Ps~haogi des Gread Calcaul.leaps dt Pumps d'&*cs. Paris: Uibralule Ha-

Dower, C. H. 1972 Mental Wmagel anda-ciative learning. In ConH as. in Learsnft andMemory. ad. L W. Gregg Wiley.

Charneas. N. iM7. Component of skill inbridge. Ca.. ISO"pha Sd-O

Chase. W. C.. and IL A. Erimonsi. Int. Skilledmemory. In Cagedavr Skitls end Tlu* Aca psisihsue. ed. 1. R. Anderson. Hillade. NJ: Eri-barna.

- 11112. Skill and working meosy. In

16. ed. G. H. Bower. Acadlemic Press,*Chane. W. CL, andl H. A. Simon. 1973a Per.

ception in chess Ceguftio Psyclita. 4:55-81.

-1973b. The mIndrs eye in chems. In V.snad luefawm Pracesmad. W. C. Chose.Academic Pi.

Daneman. M.. and1 P. A. CarpenWe. 1911. In-dividual differences in soi ng memory'and reading. 1. Ve"u ldam6eg and VerialIehavasv 19:430-66.

doerowu. A. D. 196. Pereeptlon and moeoIveo~ thoughtSomeoldieu andreent

*findings. In P!ra*EeTu Auwa~urd*.Method. aged ?Wear, ad. .nmaufflWiley.

Egan. 0. E.and LI1. khwast W99. Chuningin recall of symbolic drawings. Memauwy

Cauan7:149-56.Ericsson. IC. A.. W. C. Chose, and S. Fats..

19110. Acquisition of a memory skill. Scieue

Orkasf. K. A6. and 1. Knort.smy (eW nin sequences. Pae a4 &t mee-tg Pay.chonomic Sac. Philadelpa 19111.

Hunt. L, and T. Lo. IOM2 How goo canmemoery be? Is Cedhrg Ptrems to temmin"t. ad. * W. Mob"e 8nd K Martint.WaahlI W Wiesief

Hu~AMa. . L 197. An rAW"en mmsPbe .Po 6WtSS.&

Ja0wril. It. 1. 1971. Syntsele penn of

CHU/CMASE October 16, 1930 Page 1

Navy Navy

I Dr. Robert Preaux I Dr. William L. NaloyCode 3-711 Principal Civilian Advisor forNAVTRAEQUIPCEN Education and. TriningOrlando, FL 32813 Naval Training Command, Code OOA

Pensacola, FL 325081 Chief of Naval Education and Training

Liason Office 1 Dr. Kneale MarshallAir Force Human Resource Laboratory Scientific Advisor to DCNO(MPT)Flying Training Division OPOITWILLIAMS AFB, AZ 852241 Washington DC 20370

1 Dr. Richard Elster I CAPT Richard L. Martin, USN.Department of Administrative Sciences Prospective Comanding OfficerNaval Postgraduate School USS Carl Vinson (CYN-70)Monterey, CA 93940 Newport News Shipbuilding and Drydock Cc

Newport News, VA 236071 DR. PAT FEDERICO

NAVY PERSONNEL R&D CENTER 1 Dr William MontagueSAN DIEGO, Ck 92152 Navy Personnel PAD Center

San Diego, CA 921521 Dr. John Ford

Navy Personnel RD Center 1 Library

San Diego, CA 92152 Naval Health Research CenterP. 0. Box 85122

1 Dr. Henry M. Halff San Diego, CA 92138Department of Psychology,C-009University of California at San Diego I Naval Medical R&D CommandLa Jolla, CA 92093 Code 44

National Naval Medical Center1 LT Steven D. Harris, KSC. USN Bethesda, KD 20014

Code 6021Naval Air Development Center I Ted H. I. YellenWarminster, Pennsylvania 18971 Technical Information Office, Code 201

NAVY PERSONNEL R&D CENTER1 Dr..Patrick R. Harrison SAN DIEGO, CA 92152

Psychology Course DirectorLEADERSHIP & LAW DEPT. (7b) 1 Library, Code P201LDIV. OF PROFESSIONAL DEVELOPtMKMET Navy Personnel RlD CenterU.S. HAVAL ACADEMY San Diego, CA 921S2ANNAPOLIS, MD 21402

1 Technical Director1 Dr. Norman J. Kerr Navy Personnel R&D Center

Chief of Naval Technical Training San iego, CA 92152Naval Air Station Memphis (75).llinston, TO 3805 6 ComamdiLn Offloer

"N@al R search, Labertory

Vaahntoa, 9 10390

± 1'-

imp' *±z A'-I*.

CMU/CHASL October 16, 1980 Page 2

Navy Navy

Psychologist I Roger W..Remington, Ph.DOUR Branch Office Code L52Bldg 11i, Section D NAHAL666 Sumer Street Penooola, FL 32508Boston, MA 02210

1 Dr. Bernard Rimland (033)Psychologist Navy Personnel lrD CenterONR Branch Office San Diego, CA 92152536 S. Clark StreetQhicago, IL 60605 1 Dr. Worth Scanland

Chief of Naval Education and TrainingOffice of Naval Research Code N-SCod- 1137 NAS, Pensacola, FL 32508800 N. Quincy SStreotArlington, VA 2227 1 Dr. Robert G. Smith

Office of Chlef of Naval Operations5 Personnel & Training Resarch Programs OP-987H

(Code 45$) Washington, DC 20350Office of Naval ResearchArlington, VA 22217 1 Dr. Alfred F. 3iode

Training Aalysis & Evaluation GroupPsychologist (TARO)ON Braneh Office, Dept. of the Navy1030 tast Green Street Orlando, FL 32813Pasadena, CA 91101

1 Dr. Richard SorensesOffice of' the Chief of Naval Operations Navy Personnel R&D CenterResearch Develolment & Studies Branch San Dego, CA 92152

COP-115)Washington, DC 20350 1 Roger Welssinger-aylon

Department of Administrative SciencesCaptain Donald F. farker, USN Naval Postgraduate SehoolComandlng Officer Monterey, CA 9310Navy Personnel liD CenterSan Diego, CA 92152 1 Dr. Robert Visher

Code 309LT Frank C. Petho, MSC, US3 CPh.D) Navy Personnel PiD CanterCode L51 San Diel., CA 92152Naval Aerospace Medical Research LaboratPensacola, FL 32508

Dr. Gary Poock.Operations Research DepartmentCode s5PIKNaval Postgraduate 3uhoolMonterey, CA 939"0

-•~- .- .'7 - _ -:

CHIU/CHASE October 16, 1980 Page 3

Army Air Force

1 Tecliical Director 1 Air University LibraryU. S. Army Research Institute for the AUL/LSE 76/443

Behavioral and Social Sciences Maxwell AFB, AL 361125001 Eisenhower AvenueAlexandria, VA 22333 1 Dr. Earl A. Alluisi

HO, AIHAL (AFSC)I HO USAREUZ & 7th Army Brooks AF3, TX 78235

ODCSOPSUSAAREUE Director of GED I Dr. Genevieve IbddadAPO New York 09403 Progrm ?lnager

Life Scienees Directorate"1 Dr. Wlchsel Kaplan AF031

U.S. ARMY RESEARCH INSTITUTE Bolling AFB, DC 203325001 EISENHWJER AValUEALEXANDRIA, VA 22333 Dr. Donad G. Hughes

AFIML/OTXI Dr. Milton S. Katz Willims AID, AZ 85224

Training Technical AreaU.S. Army Researih Institute I Dr. Marty Iocknmy5002 Eisenhower Avenue Technical DirectorAlexandria, VA 22333 AFUAL(OT)

Iillims AFB, AZ 58221Director

U.S. Army Human Engineering Labs V Dr. Frank SohufletowskiAttn: DRXHE-DS U.S. Air ForceAberdeen Proving Ground, NO 21005 ATCIPTD

folph AFI, TX 781481 Dr. Harold F. O'Neil, Jr.

Attn: PRI,-O 2 3700 TCHIW/TTGH Stop $2Army Research Institute Sheppard *13, TX 763115001 iaenhower AvenueAlexandria, VA 22333 1 Jack A. 7horp, Mej., US

Lito. Seloese Nreestwte1 Dr. obert Sbmor AFOh

U. S. Army Research Institute for the tolling *fl, DC 20332- ehavieral and Social Sciences

5001 Eisenhowr AvenuAlexandria, VA 22333

1 Dr. Joseph WardU.S. Army Research Institute5001 Eisenhower AvenueAlexandria, VA 22333

I -

CHU/CHA3E October 16. 1980 Page 4

Marines Other DoD

IH. William Greenup 12 Defense Technical Information CenterEducation Advisor CE031) Cameron Station, BldS 5Education Center, * C Alesandria, VA 22314Quantico, VA 22134 Atta: TC

Headquarters, U. S.-Merine Corps I Dr. Dexter FletcherCode NPI-20 ADVANCED RESEARCH PROJECTS AGENCYWashington, DC 20380 1400 WILSON BLVD.

ARLINGTON, VA 222091 Special Assistant for Hrine

Corps Matters 1 Military Assistant for Traiing andCode 1001M Personnel TechnologyOffice of Naval Research Office of the Under Secretary of Defense800 N. Quincy St. for Research & EnglneeringArlington, VA 22217 Room 3D129, The Pentagon

Washington, DC 20301DR. A.L.. SLAFKOSKYSCIEN1TIFIC ADVISOR (CODE I0-1) 1 BEAD, SECTION ON MEDICAL EDUCATIONHQ' U.S. MtAl" CORPS UNIFORMED SERVICES U11EV. (N THEWA HIGTON. C 20360 HEALTH SCIENCIS

6917 ARIUGt MO ROAD

BETIIEDA* MD 2W014

mCU/CRASE October 16, 1980 Page 5

Civil Govt Non Govt

Dr. Susan Chipam 1 Dr. John 3. AndersonLearning and Development Departmnt of PsychologyNational Institute of Education Carnegie Mellon Untiversity1200 19th Street WN Pittsburgh, PA 15213Washington., DC 20208

1 Anderson, Thomas N., FhD.

Dr. Joseph 1. Lipsmn Center fbr the Study of ReadingSEM W-638 174 Children's Research CenterNational Science Foundation 51 Gerty Drivebahington, DC 20550 Champlagn. IL 61820

Dr. rthur M6m ed I Dr. John AnettNational Intitute of Education Department of Psychology1200 19th Street NW University of WarwickWashington, DC 20208 Coventry CV4 7AL

ENGLANDI Dr. Andrew 3. 161nar

Science ducotiom tv.. 1 DI. HICHUL ATWOODand Researoh SCIENCE APPLICATIONS INSTITUTE

Ntional Science Fbondaton 40 DENVER TECH. CENTER WESTWashington. DC 2" 7935 E. PRENTICE AVENUE

ENGLEWOOD, CO 80110I Dr f f. Wallace Sinaiko

Program Director 1 I psychological research unitManpower Research and Advisory Services Dept. of Defense (Army Office)Smithsonian Institution Campbell Park Offices801 North Pitt Street Canberra ACT 2600, AustraliaAlexandria, VA 22314

1 Dr. Alan BoddeleyDr. Frank Withrow Medical Researoh CouncilU. S. Orfle of Education Applied Psychology Unit400 Maryland Ave. SW 15 Chaucer RoadWashington, DC 20202 Cambridge Cl2 2EF

ENGLANDI Dr. Joseph L. Yroung, Director

memory & Cognitive Processes 1 Dr. Patricia aggettNotional Science Foundation Department of PsychologyWashington. DC 20550 University of Denver

University ParkDenver, CO 80208

I Mr Avron BerrDepartment of Computer ScienceStanftrc UniversityStanford, CA 94305

1I17

CMU/CH4ASE October 16, 1930 Page 6

Non Govt Non Govt

Dr. Nicholas A. Bond 1 Dr. William ClanceyDept. of Psychology Department of Computer ScienceSacramento State College Stanford University600 Jay Street Stanford, CA 94305Sacramento, CA .5819

1 Dr. Allan H. CollinsDr. Lyle Boura Bolt Branek & Neman, Inc.Department of Psychology 50 Mouton StreetUniversity of Colorado Cambridge, He 02138Boulder, CO 80309

1 Dr. Lynn A. CooperDr. K(enneth Bovies LRDCInstitute for Information Sciences University of PittsburghUniversity of California at San Diego 3939 O'Hare StreetLa Jolla, CA 92037 Pittsburgh, PA 15213

Dr. John S. Brom 1 Dr. Meredith P. CrawfordXEROX Palo Alto leseareh Center American Psychological Asaolation3333 Coyote Road' 1200 17th Street, N.V.Palo Alto, CA 94304 Washington, DC 20036

Dr. Brume Boshanan 1 Dr. Kenneth B. CrossDepartment of Computer Selence Anacap Slences, Inc.Stanford University P.O. Drawar.QStanford, CA 94305 Santa Barbara, CA 93102

I O. C. VicT= IuMNRSON 1 Dr. Hubert DreyfuiVZCAT INC. Department of PhilosophyUIIVERSIT PLAZA, SUIT 10- University of California1160 30..STATI ST. Berkely, CA 94720ORI, UT 34057

1 LCOL J. C. %genbergerDr. Jhn 8. Carll DIRECTORATE OF PErSMONL APPLIED E1ERCPsychametrie Lab NATIONAL DEFEINCEOUniv. of lb. Carolina 101 COLONEL BY I1WtDvie Hall 013A OTTAWA, CANADA' EIA Q12Chapel Hill, N 1514

I Dr. Ii reig e MCharles Myers. Ubrary Departmemt of Gomputar ScienceLivingstone Huse Stanford UniversityLivingstone Iod Stanford, CA 94305Stratford'London 31S I I Mr. Wallace eurzeigENGLAND Solt Beranek A Newman, Inc.

50 Moulton St.Dr. William i Cobrid*e, *A 02138ftparoo" et PoleNlarCarnegie Bli i'vrnI I

Pittsburgh, l,$Pt.

. . .. .. - + 7

CMU/CHASE October 16i 1980 Page 7

Non Govt Non Govt

I Dr. Edwin A. Fleishman I Dr. Ron HambletonAdvaced Research Resoures Orga, School of EducationSuite 900 University of Massehusetts11330 East West Highway Amherst, MA 01002Washington, DC 20014

1 Dr. Harold HawkinsDR. JOHN D. FOLLEY JR. Department of PsychologyAPPLIED SCIENCES ASOCIATES INC University of OregonVALENCIA, PA 16059 Eugene OR 971103

Dr. John R. Frederiksen 1 Dr. Barbara Hayes-RothBolt" eranek & Nowman The Rand Corporation50 Moulton Street 1700 ain StreetCambridge, 4A 02138 Santa Monica, CA 90406

Dr. AlLnda Friedman 1 Dr. Frederick Hayes-RothDepartment of Psychology 7he Rand CorporationUniversity of Alberta 1700 Main StreetEdmonton, Alberta Santa Honics, CA 9006CAUA TGG 29

1 Dr. Jnes R. HoffmanDr. 3. Edward Gelselman Department of PsychologyDeprtment of Psyohoalosy University of DelawareUniversity of California Newark, D9 19711Los Angeles, CA 9002

I Glenda Oreenwld,. Ed.I DR. ROERtT CLAM Offumen Intelliglene Newsletter*LIOC P.O. Box 1163

UNIVERSITY OF PITSSKM Birmtinn, MI 4i0123939 O'HAA 21111PITI'SSUVIK, PA 15213 1 Dr. Earl Hunt

Dept. of PsychologyI.Dr. Mrvil . oek University of Wshington

217 Stone oll Seattle, WA 98105Cornell olversityIthns, my t#853 I Dr. Steven W. goe

Dept. of" Psychologly

Dr. DoniLel Olher University of Oregonndiatriel a Monagwent logineering Eugene, 08 971103

Tehnion-sreel Institute of TechnologyHaifa 1 Dr. Wlter Klatach•ISRAL Department of Psychology

University of CaleradoDio. JAI4S 0. a D 'Boulder. CO 80302

UNIVIRSITV 9 SPI'TSR I Dr. David hKee3939 O'NARA STrRET Deperum or PsychaftyI numi7Ut , PA 15213 • UI Itr or aria" ,• • Tusm* At W#.

CHU/CHASE October 16, 1930, Page 8

Non Govt Non Govt

Dr. Stephen Kosslyn'1 Dr. Donald A NormanHarvard University Dept. of Psychology C-GO,Department of Psychology Univ. of California, San Diego33 Kirkland Street La Jolla, CA 92093Cambridge, HA 02135.

1 Dr. Jesse OrlanskyHr. Mrlin Kroger Institute for Defense Malyses1117 Via GOleta 400 Army Navy DrivePalos Verdea Estates, CA 90274 Arlington, VA 22202

Dr. Jill Larkin I Dr. Seymour A. PapertDepartment of Psychology Massachusetts Institute of TechnologyCarnegle Mellon University Artificial Intelligence LabPittsburgh, PA 15213 545 Technology Square

Cambridge, MA 02139Dr. Alan LesgoldLearning RID Center 1 Dr. James A. PaulsonUniversity of Pittsburgh Portland Set* UniversityPittsburgh, PA 15260 P.O. low TS1Dr. t~ehel Lv~nePortland, 02 97207

I Dr. Michael LevineDepartment of Educational Psychology 1 MR. LUIGI PEIttILO210 Education WEdg. 2131 W. EDI OOS STREETUniversity of Illinois ARLISGTON. VA 22207Champaign, IL 61801

I DR, PETER L3ON. .Dr. Robert A. Levit DEPT. OF PYCHOLOGYDirector, blvioral Sciences UNVERSITY OP w COLRAThe SO Q.rpwetion BOULDR.I CO 80307915 Jame Brench DrSveMcClean, VA 22101 1 Dr.Fred Naif

Dr. Charles Levis 0/o Physics BapartmatFaculteitl Soel ,Wtemshappen University. of GkUlfwriaRiJksuniveralitt ir~oningon Sarkely, CA 1 0Oude Nateriaanstr"tGro ""e Dr. Andrev . baoNTHERILAN5 Aferioas InstituteS fr Research

1055 hm" AJrferna St. WDr. KArk 1iller Wasbingto.a, DC 007

•Texas fnst.ents, be. 1 Dr. Ernst . t l i emil Station 371, P. SOx 225936 ell LaboratoriesSDallas, T . : •0No Mountain Avewe

I Dr. oil. Nary 11 N t~- :tatics

t S e.._ --u*

CIU/CHASE October 16, 1980 Page 9

lNon Govt Non Govt

I DR. WALTER SCHINEDER I Dr. David StoneDEPT. oF PSYCHOLOGY ED 236UNIVERSITY OF ILLZNO13 SUNY, AlbanyCIWIPAIGN, IL 61820 Albany, NY 12222"

Dr. Alan Sohoenfeld I DR. PATRICK SUPPESDepartment of hthematics INSTITTUE FOR MIATHMATICAL STUDIES INHauilton College THE SOCIAL SCIECESClinton, NY 13323 STANFORD UNIVERSITY

STAWORD, CA 94305Committee on Cognitive Research% Dr'. Lonnle R. Sherrod I Dr. KIumi TatsuokaSocial Sl-nee Research Council Coputer Based Educatlon Research605 Third Avenue Lob6ratoryNew York, NY 10016 252 Engineering Research Laboratory

University of IlinoisRobert S. Siegler Urbana, IL 61801Associate ProfessorCarnegle-Mellon ol~versity 1 r. John. 7ho8sDepartment of Psyohology 1DM Thomas J. Watson Research CenterSchenley Park P9O.n3x 218Pittsburgh, P. 15213 Yorktown lr, NT. NY 10S96

I Dr. Robert Saith 1 DR. PERRY THO2E00iCDepartment of Computer Science THE RAND O.OPORUTOWRutgers University 1700 MAIN STREETNew runbmok, NJ 06903 SANTA MONICA, CA "@06

Dr. Rieherd sow 1 D. outo1a TO""School of MiWstm Univ. of 36. CllftraieIanford Univerity heavlori Wleaoofp Labs, sfbrd, CA 9"% 184 S. flnes .

Redondo each, CA 902??I Dr. tbbert StorCapt. of' Ift meS 1... 0' ; hI

Yale I~iveraity b..Box hA. 1We Station 6271 vM Am

P-wime. CT 0600 Woodland Hill, CA91364

OR A3J9"." Toll 1 Dr. Benton J. MderwlodiDOT DItANio & =IIWm, -two- Dept. of PsAtyelegy.0OlTO STCT; cA mnx ,A tII3S, -J

X4 raste * A1* t"""aoS I,"IV

~ N. 1b 1S. 4,

.... .... .... .... ...... . . . . . . ----

CMU/CHASE October 16, 1980 Pie 10

Non Govt

I Dr. Dsvid J. Weiss3660 Elliott mallUiiversity of Minnesota75 E. River RoadMinneapolis, MN 5555

1 DR. GERSHON WELTmNPERCEPTRONICS INC.6271 VARIEL AVE.WOODLAND HILLS, CA 91367

1 Dr. Keith T. WescourtInformation Sciences Dept.The Rand Corporation1700 Main St.Santa Moniea, CA 90406

1 DR* SUSAN E. WHITELYPSYCHOLOGY DEPARTMENTUNIVERSITY OF KANSASLAWRENCE, KA11SAS 66044

Dr. Christopher WiekensDepartment of PsyohologyUnive/,sityc Ifllinois

& Ii