Transcription of THE CONTROL PROCESSES OF SHORT-TERM MEMORY …
1 , ,CALIFORNIA'rHECON' , , ,WilliamJames, ,thissentenceisinyourcurrentawareness,bu tthewinnerofthe1968 WorldSeries,whileprobablyinmemory, , 'swhenanumberofpsychologists,particu~arl yDonaldBroadbentinEngland,DonaldHebbinCa nada,andGeorgeMillerintheUnitedStates,1r eintroducedit( ,InformationandMemory,ScientificAmerican ,1956,195(2),42-47). ,theshort-termmemorysystem,whichwewillca llshort-termstore(orSTS), ;theycanbecalledintoplayatthesUbject'sdi scretion, , ,anovertorcovertrepetitionofinformation, isemployedinnumeroussituations:whenremem beringaphonenumberuntilitcanbewrittendow n,whenre-member; ,andwhencopyingapassagefromabook, ,studentssometimeslearnthetwelvecranialn erveswiththeuseofthemnemonic"S:nS:ldS:ly mpusI-,!,iny-,!,op!2!innJ2ndQerman~iewed !
2 :!omeBops," ~ (orspeeches) ,inc~udingdecisionrules,organizationalsc hemes,retrievalstrategiesandproblemsolvi ngtechniques, ,thecontrolproceSsesareselectedatthesubj ect'sdiscretion;theymayvary, ~memorysystemisbestdescribedintermsofthe flowofinformationintoando~tofSTSandthesu bject' "image"or"trace"(thus"image"doesnotneces sarilyimplyavisualrepresentation).Ifthep air"TKM-4"ispresentedformemory,theimages toredmightincludethesizeofthecardonwhich thepairisprinted,thetypeofprint,thesound ofthevarioussymbols,thesemanticcodes, (STS).Theinformationremainstemporarilyin STS, (LTS).WhileinformationremainsinSTS, ,ifapictureofatriangleispresented, ,sincetheverbalname"triangle"isassociate dwiththisvisualinformationinLTS, IVISUALIIAUDITORY1 .. ,ICONTROLPROCESSESIIIIREHEARSAL,IICODING ,IIDECISIONS, 3bmodalities, , , , ~(orLTS); , ;infact,weassumethatduringtheperiodanima geresidesinSTS, , ,awordmaybepresentedvisu-ally,butimmedia telyafterinputthearticulatory-verbal"nam e" , ,forexample, , ,inourthinkingwetendtoequateSTSwith"cons ciousness"; , , ,andbecaUsecontrolprocessesarecenteredin andactthroughSTS,thisstoreiscQnsideredto bea"workingmemory":astoreinwhichdecision saremade,problemsaresolved, , ,placesitinSTS,andthenscansSTSfQrthedesi redimage(whichmaynotbepresentinthecurren tSUbset).
3 "probeinformation" ; , "searchcset."Eeforethesearch-setislostfr omSTS, ,thesearchiseitherstopped, T SANDITSTRANSFERTOSTSDECISIONTOCONTINUEOR TERMINATESEARCHRESPONSECHOICEANDITSOUTPU T5bthe"searqh"set."Thesubjeqtseleqtsfrom thesearqh-setsomeimage, :hasthedesiredinformationbeenfound?Ifso, , , ,dependinguponthesubject' , (inwhichcasethesameprobeinformationmaybe usedoneachsearchcycle)orhemaygeneratesuc cessivestatesinaregulargeographicorder(i nwhichcasetheprobeinformationissystemati callychangedfromoneqyqletothenext).Itqan beshownthatstrategiesinwhiqhtheprobein-f ormationissystematiqallyqhangedwillmoreo ftenresultinsucqessfulretrieval,butwillt akelongertodosothanalternative"random" ,informationstorageandretrievalisbestdes qribedintermsoftheflowofinformationthrou ghSTS,andintermsofthesubjeqt' ; ; "freerecall.
4 "Becausetheexperimentstobeconsideredhere arebasedononeoranothervariationofthispar adigm, (usuallycommonEnglishwords) ,EndelTulvingatYaleUniversity, ; (a)Thebasicserialpositioncurve:theriseon therightiscalledtherecencyeffectbecauset hesewordsarethemostrecentlypresented;the riseontheleftiscalledtheprimacyeffectbec ausethesearethefirstwordspresented.(b) ,andiseliminatedwhenarithmeticcausesthew ordstobelostfromSTS.(c) (d) ,andarebasedonexperimentsreportedbyJames Deese,BennetMurdock, ~.._--'--'-~ ':r---i"'~"'~"'~"'1-0-a-------;----=r;;i K~~.40 WORDLISTJ~--20 WORDLISTWITHINTERVENINGARITHMETICWITHOUT INTERVENINGARITHMETIC-_ < ~--, > ,..------------,..,-JIX!<lIX! ,.. ,,..,.5___--2 , , ,forexample, ,therecencyeffectiseliminatedwhenthisexp erimentisperformed;furthermore,theearlie rportionsoftheserialpositioncurveareunaf fected(seeFigure3b).
5 ,therecencyportionoftheserialpositioncur veshouldberelativelyunaffected, ,a wordinalongerlistislesslikelytoberecalle d, ,increasesinpresentationratedecreasethel ikelihoodofrecallingwordspriortotherecen cyregion,butleavetherecencyeffectlargely unaffected(seeFigure3d). , , , ~-- ,_~~-~~~ ~_,_____,~ <.> > ~..Jm '---'-~~~--""""'----~~-'-~-~-~~--'----'5 1015 SERIALPRESENTATIONPOSITION8bsummary,theJ ;' , ;' ,includingtheitemmostJ;'ecentlypJ;' ;'ffistorewhenthelastlistitemhasbeenpres entedwereassumedtobeoutputatonce(givingr isetotherecencyeffect). ~eritems,theywereassumedtoJ;' (thusgivingrisetotheprimacyeffect). , ;' ;'trehearsalwastape-recorded, "rehearsalset"isusedtorefertothoseitemso vertlyrehearsedbetweensuccessivepresenta tionsofwOJ;' ;fromthefourthwordonthenumberofdifferent wordsintherehearsalsetremainedfairlycons tant( )untiltheendofthelist(seeFigure5a).
6 ;intherecencyregion,however, , ,therecallprobabilityfora , , , ' (a)Exampleofasubject' 'sovertrehearsa19ir) ~sgivenmoretotalrehearsalsthanlaterwords .(b)Probabilityofrecallcomparedwithtotal numberofrehearsals, , ,LTSstoragedependsonthenumberofrehearsal sgivenanitem, 'IrII'-::-'tIIJIIIIIIiII15 .. ,Hoof,Reactionf; ' ,Reaction,Hoof, ,Hoof,Research,Reaction>-~ ;HardshipHardship,Hoof,Hardship,HooflIJ ,candy,Hardship,HoofI~.5 ,Kindness,Candy, ,Alcohol,Misery,CellarSubjectsweretraine dtoengageinoneoftwotypesofrehearsal(seeF igure6a).Inthefirst(One-itemrehearsalset )themostrecentlypresentedi tell! (Three-itemrehearsalset)thesubjectrehear sedthethreemostrecentlypresenteditems)on ceeachbeforepresentationofthenextitem;th efirstrehearsalsetcontainedthreerehearsa lsofthefirstword,thesecondrehearsalsetco ntainedtworehearsalsofthesecondwordandon erehearsalofthefirstword, , , , ,evenwhendroppedfromrehearsal, , ,adelayedrecalltestforallwordswasgivenat theendoftheexperimentalsession(similarto theproce-duretheresultsofwhicharegivenin Figure4).
7 Thedatafromthis11 FigureCaptionFigure6.(a) , ,thelastthreeitemspresentedarerehearsed; (b) , ,whichreflectLTSretrievalonly, ~ I ARITHMETICTASKII]TEST~~[TI~I???@ wordmeaning"tolookat"youmightsay"pear."B utifaskedtonamealivingthingwhichsoundsli keawordmeaning"tolookat,"theprobabilityo fre-sponding"pear" ,supposeyouattendapartywithfiveotherpeop le,onenamed"JohnSmith."Ifaskedlatertonam ethepeopleattheparty,youmightbelikelytor emember"JohnSmith."Iftherehadbeen20peopl eattheparty, ,besidesafailureofthememorysearch, "JohnSmith" ,itcouldbearguedthatthenamesofthemanyoth erpeoplemetatthelargerpartycoulderodeord estroythememorytracefor"JohnSmith."Areth eseobjectionsreasonable? , ,ratherthanuponthenumberofwordsinterveni ngbetweenpresentationandrecall, :19inmanyfreerecallexperimentswhichvaryl istlengththesubjectsdo\ ~lessthefirstseveralwordsarerecalledatdi fferentlevels(seeFigure3c).
8 Todisposeoftheinterferenceexplanation(Wh ich,implicatesthenumberofwordsbetweenpre sentationandrecall) ( ,Forgetting,ScientificAmerican,1964,210( 3),91-99,andJohnCeraso,TheInterferenceTh eoryofForgetting,ScientificAmerican,1967 ,217(4),117-124).Inthesetheories,forgett inghasoftenbeenseenasamatteroferosionoft hememorytrace--usuallycausedbyitemsprese ntedfollowingtheitemtoberemembered, , , , (asinthetypical20freerecallprocedure) , , :5-5,5-20,20-5, ; ;consequently, :wordsinlistsoflength5arerecalledmuchbet terthanwordsinlistsoflength20,buttheleng thoftheinterveninglisthaslittle, (probablytemporalinnature) , 'sproba-bilityofrecallisdependentuponthe lengthoflistinwhichitisimbedded, ~ ,INTERVENINGLISTOF5 WORDS~INTERVENINGLISTOF20 WORDSOF20 WORDOF5 WORDS'\V-INTERVENINGLIST\\\\\"-\'; (.)
9 > ~ehimtoselectprobeinformationpertainingt othedesired~ ,thense~ , ,thedecisiontoacceptagivenimageastheones ought, ,buthopefullyithasclarifiedoUrapproach, ~sandmathematicalmodelsinthememoryareaha vebecomeincreasinglysophisticated, , , ,&Shiffrin, (Eds.),Thepsychologyoflearningandmotivat ion:Advancesinresearchandtheory, :AcademicPress,1968,89-195 Atkinson, ,&Wickens, (Ed.), , ,Inc.,1971, , ,1971, , :Traceerosionorretrievalfailure?Science, 1970,168, , (Ed.), :AcademicPress,1970.
