Ancestral priorities and the mnemonic value of survival processing

22 Pages·2010·0.21 MB·English

CognitivePsychology61(2010)1–22
ContentslistsavailableatScienceDirect
Cognitive Psychology
journal homepage: www.elsevier.com/locate/cogpsych
Adaptive memory: Ancestral priorities and the mnemonic
value of survival processing
James S. Nairnea,*, Josefa N.S. Pandeiradab
aPurdueUniversity,UnitedStates
bUniversityofAveiro,Portugal
a r t i c l e i n f o a b s t r a c t
Articlehistory: Evolutionary psychologists often propose that humans carry
Accepted26January2010 around‘‘stone-age”brains,alongwithatoolkitofcognitiveadap-
Availableonline6March2010 tationsdesignedoriginallytosolvehunter–gathererproblems.This
perspective predicts that optimal cognitive performance might
Keywords: sometimesbeinducedbyancestrally-basedproblems,thosepres-
Memory
entinancestralenvironments,ratherthanbyadaptiveproblems
Evolution
faced more commonly in modern environments. This prediction
Survival
was examinedin fourexperiments usingthe survival processing
paradigm, in which retention is tested after participants process
information in terms of its relevance to fitness-based scenarios.
Ineachoftheexperiments,participantsrememberedinformation
betterafter processingitsrelevanceinanancestralenvironment
(the grasslands), compared to a modern urban environment (a
city),despitethefactthatallscenariosdescribedsimilarfitness-
relevantproblems. Thesedatasuggestthatourmemorysystems
maybetunedtoancestralpriorities.
!2010ElsevierInc.Allrightsreserved.
1. Introduction
Thecapacitytorememberevolved,havingbeenshapedandsculptedbytheprocessesofnatural
selection.Specificselectionpressures,relatedtosurvivalandreproduction,conferredselectionadvan-
tagestoorganismscapableofusingthepastintheserviceofthepresent.Noscholarseriouslyques-
tionstheseclaims,althoughthefunctionalpropertiesofmemoryarerarelyconsideredexplicitlyby
modernmemorytheorists(Nairne,2005;Klein,Cosmides,Tooby,&Chance,2002).Instead,themain
* Correspondingauthor.Address:DepartmentofPsychologicalSciences,PurdueUniversity,703ThirdStreet,WestLafayette,
IN47907-2081,UnitedStates.
E-mailaddress:[email protected](J.S.Nairne).
0010-0285/$-seefrontmatter!2010ElsevierInc.Allrightsreserved.
doi:10.1016/j.cogpsych.2010.01.005
2 J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22
empiricalandtheoreticaleffortoverthepastcenturyhasbeendirectedatunderstandingtheproxi-
matemechanismsthatcontrolperformanceonspecifictasks—i.e.,the‘‘how”ratherthanthe‘‘why”
ofremembering(seealso,Bruce,1985;Sherry&Schacter,1987).
Yet,fromanevolutionaryperspective,itisreasonabletoaskwhethervestigesoftheselectionpres-
suresthatledtomemory’sdevelopmentremainapparentinitsoperatingcharacteristics.Ifmemory
systemsevolvedtoenhancefitness(survivalandreproduction),thenareorganismsespeciallygoodat
encodingandretainingmaterialfromfitness-relevantsituations?Thisisultimatelyanempiricalques-
tion, one that need not suffer from the criticisms usually directed at investigations in evolutionary
psychology(e.g.,‘‘just-so”and‘‘Panglossian”reasoning;seeGould&Lewontin,1979).Indeed,there
isnowsubstantialempiricalevidenceshowingthatmemoryisenhancedwhenpeopleprocessinfor-
mationintermsofitsfitnessvalue(Kang,McDermott,&Cohen,2008;Nairne,Thompson,&Pandei-
rada, 2007; Weinstein, Bugg, & Roediger, 2008). For example, Nairne, Pandeirada, and Thompson
(2008)foundthatafewsecondsofsurvivalprocessing,duringwhichparticipantswereaskedtorate
therelevanceofrandomwordstoasurvivalscenario,producedbetterlong-termretentionthanahost
oftraditionallystrongencodingprocedures(e.g.,formingavisualimageorengaginginsemanticpro-
cessing;seealsoNairne&Pandeirada,2008a).
Still, establishing an empirical advantage for fitness-relevant processing tells us little about the
cognitivemechanismsthatproducethoseadvantages,orabouttheirultimateorigins.Onecouldargue
thatorganismsevolveddomain-generalmechanismsforretaininginformation,onesthatarematerial-
andsituation-independent.Infact,memoryresearchersvirtuallyalwaysemploygeneralmechanisms
toexplainsuperiormnemonicperformance,mechanismssuchaselaboration(connectinginformation
tootherthingsinmemory),distinctiveprocessing(creatingunusualrepresentations),orprocessing
thatistransfer-appropriate(encodingrepresentationsthatarelikelytobematchedinatargetedre-
trievalenvironment).Certaintypesofmaterial,orsituations,maynaturallyaffordelaborate,distinc-
tive, or appropriate encodings, but the mechanisms themselves are assumed to operate similarly
acrossdomains(seeSurprenant&Neath,2009).
Evolutionarypsychologiststendtoeschewdomain-generalmechanismsforavarietyofcompelling
reasons.Natureusually‘‘selects”onephysicaldesignoveranotherwheniteffectivelysolvesaspecific
adaptiveproblemrelatedtofitness—thus,wehaveheartsuniquelydesignedtopumpblood,kidneys
tofilterimpurities,andsoforth.Fromasocialandcognitiveperspective,therearetoomanycritical
problemstobesolved,suchasavoidingpredators,locatingnourishment,orselectinganappropriate
mate,torelyonthewhimsofgenerallearningmechanismsthataresimplydesignedtoextractinfor-
mationfromindividualexperiences(e.g.,Cosmides&Tooby,2005).Relevantexperiencesmayoccur
naturally with too low a probability, and problems of combinatorial explosion arise quickly with a
cognitivesystemthatisdesignedtostoreeverything.Selectivestorage,basedoninherentcognitive
‘‘biases”or‘‘tunings,”wouldappeartobenecessaryforspeciestosurviveandreproduceeffectively.
Of course, these are not airtight arguments (see Buller, 2005), but they can serve as the impetus
andfoundationforempiricalinvestigation.
1.1. Ancestralpriorities
Tothe extentthat our brainswere sculptedprimarilyduring thePleistocene, theperiod usually
considered as the environment of evolutionary adaptation (e.g., Symons, 1992), natural selection
wouldhavebeendrivenbyPleistocenepressures—thatis,theuniquesetofadaptiveproblemsfaced
byour hunter–gathererancestors. Evolutionarypsychologists commonlyarguethatwecontinueto
carry around these ‘‘stone-age” brains, along with a toolkit of cognitive adaptations that were de-
signed originally to solve hunter–gatherer problems (Tooby & Cosmides, 1992). One might predict,
therefore,thatoptimalcognitiveperformanceshouldbeinducedbyancestrally-basedproblems,par-
ticularlythosepresentinancestralforagingenvironments,ratherthanbyproblemsfacedinmodern
environments.If,infact,ourcognitivesystemsevolvedtosolveparticularadaptiveproblemsthenthe
fitbetweenproblemandprocessshouldaccount,atleastinpart,forprocessingefficiency.Thisisnota
strongpredictionofevolutionarytheory,becauseadaptationscanbeeffectivelyexaptedtoperform
rolesthatareunrelatedtotheoriginalselectionenvironment(seeAndrews,Gangestad,&Matthews,
2002),butitrepresentsaninterestingempiricaltest.
J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22 3
Thereissomesupportforthenotionofancestralprioritiesincognitiveprocessing.Forexample,it
iswidelyacceptedthatorganisms,includinghumans,haveaninbornattentionalbiasforthreatening
stimuli—particularly stimuli that were relevant in ancestral environments (see Öhman & Mineka,
2001,forareview).Bothadultsandveryyoungchildrendetectevolutionarily-relevantstimuli,such
assnakesandspiders,morequicklythannon-threateningstimuli(LoBue&DeLoache,2008).Even7-
to9-month-oldinfants, whopresumablyhavelimitedexperiencewithsnakes,showan attentional
biastosnakescomparedtootherunfamiliaranimals(DeLoache&LoBue,2009).Inaddition,New,Cos-
mides,andTooby(2007)foundfasterandmoreaccuratedetectionforanimals,bothhumanandnon-
human,thanforinanimateobjectsusingthechange-detectionparadigm,aprocedureinwhichpeople
areaskedtodetectdifferencesbetweentworapidly-alternatingimages.Changedetectionfornonhu-
mananimalswasfasterthanforvehicles,eventhoughthelatterarecommonplaceineverydayexpe-
rienceandcanbesurvival-related.Avarietyofalternativeexplanationsforthedatawerediscounted,
includingtheinterestvalueandlow-levelvisualcharacteristicsofthestimuli,leadingtheauthorsto
concludethatpeoplepossessananimatemonitoringsystemthatis‘‘bettertunedtoancestralthanto
modernpriorities”(Newetal.,2007,p.16603).
Inthelearningdomain,evolutionarily-significantstimulicanbeeasiertoassociatewithaversive
stimuli (e.g., shock or loud noise) than comparable modern controls (Öhman & Mineka, 2003). For
example, some studies have found that snakes and/or spiders are easier to associate with aversive
stimuli,orshowmoreresistancetosubsequentextinction,thanontogeneticfear-relevantstimulisuch
asgunsorelectricaloutlets(seeÖhman&Mineka,2001,forareview).Peoplearealsomorelikelyto
perceiveillusorycorrelationsbetweenpicturesofsnakesandshockthanbetweenpicturesofdamaged
electricalequipmentandshock(Tomarken,Sutton,&Mineka,1995).Inaddition,specificphobiasare
more likely to develop to ancestral stimuli (e.g., predators) than to aversive stimuli exclusively
encounteredinmodernenvironments(e.g.,weapons;seeDeSilva,Rachman,&Seligman,1977).Col-
lectively,thesedatasuggestacertainamountof‘‘preparedness”forassociatingancestrally-relevant
stimuliwithparticularoutcomes.
Demonstratingthatourcognitivesystemsshowancestralprioritiesinprocessingisanimportant
stepinfurtheringevolutionaryaccounts.Asnotedbymany(e.g.,Buller,2005;Richardson,2007),evo-
lutionaryaccountscanbedifficulttodefend,largelybecausecriticaldataareeithermissingorimpos-
sibletoobtain(e.g.,cognitiveprocessescannotbefossilized).Toestablishanevolutionarylocus—that
is,thepresenceofanadaptationsculptedbynaturalselection—requiresbuildingamultiprongedcase
based partly on the systematic ruling out of alternative accounts (see Andrews, Gangestad, & Mat-
thews,2002;Nairne,inpress).Establishingancestralprioritiesisrelevantpartlybecauseitisdifficult
toseehowgenerallearningmechanisms,thoseessentiallytunedbyexperiencesthroughoutdevelop-
ment,couldyieldsuperiorperformanceforsituationsandstimulithatarerootedinthedistantpast
ratherthaninthepresent.Atthesametime,eachofthestudiesconductedtodatesuffersfroman
inherent confound—ancestral versus modern comparisons require one to compare across different
stimuli(e.g.,snakesversusguns)andsuchstimulicandifferinmanyuncontrolledways(Blanchette,
2006;Fox,Griggs,&Mouchlianitis,2007).
1.2. Thecurrentresearch
Thepresentexperimentsweredesignedtoinvestigateancestralprioritiesinthesurvivalprocess-
ingparadigm(Nairneetal.,2007).Asnotedearlier,processinginformationintermsofitsrelevanceto
fitness can improve retention dramatically, at least compared to traditional encoding procedures
(Nairneetal.,2008).Thisparadigmpresentssomemethodologicaladvantagesovertheworkjust-re-
viewed,mainlybecauseallparticipantsareaskedtorememberexactlythesamematerial.Whatmat-
ters to retention is how people process the information—i.e., in terms of survival or some other
relevant control. This means that fitness-based retention differences cannot easily be attributed to
the characteristics of thestimuli, or toassumptionsabout whatthosestimulinaturallyafford (e.g.,
naturalfearversuslearnedfear).
Researchbasedonthesurvivalprocessingparadigmhasfirmlyestablishedthemnemonicpowerof
fitness-relevantprocessing,butthelocusoftheprocessingadvantageremainscontroversial.Thereis
someevidencethatsurvivalprocessingmaybespecial,atleastcomparedtootherformsofencoding,
4 J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22
andnoteasilyinterpretedviastandardtheoreticalframeworks(seeNairne,inpress;Nairne&Pandei-
rada,2008a).However,thefactthatfitness-relevantprocessingmaybepowerfulor‘‘special”doesnot
meanthatitsmnemonicadvantagesaccruefromaspecificevolutionarytuning—moreandvariedevi-
denceisneededtomakesuchacase.Asnotedearlier,establishinganevolutionarylocusforanycog-
nitivephenomenonisdifficultandnosingleempiricalattackislikelytobedefinitive.Demonstrating
thatsurvivalprocessingissensitivetoancestralprioritieswouldexpandtheexistingdatabaseonsur-
vivalprocessingconsiderablyandstronglybolstertheevolutionaryaccount.
ArecentexperimentreportedbyWeinsteinetal.(2008;Experiment2)providesinitialsupportfor
thecontentionthatsurvivalprocessingmaybesensitivetoancestralpriorities.Thoseauthorscom-
pared retention after people processed the relevance of words to a survival situation, but varied
whether the scenario invoked an ancestral or a modern context. In one situation, following Nairne
etal.(2007),peoplewereaskedtoimaginethemselvesstrandedinthegrasslandsofaforeignland
withoutbasicsurvivalmaterials.Overthenextfewmonths,theinstructionsexplained,theywould
needtofindsteadysuppliesoffoodandwaterandprotectthemselvesfrompredators.Theexperimen-
taltaskwastoratetherelevanceofwordstothissurvivalsituation.Inasecondcondition,exactlythe
samescenariowasusedbuttwocriticalwordswerechanged:citywassubstitutedforgrasslandsand
predatorswasreplacedbyattackers.Escapingfrompredatorsinthegrasslands,theauthorsreasoned,
is a closer fit to the problems faced in the environment of evolutionary adaptation; as a result, it
shouldproducebettermemorythanprocessinginamoderncontext,eventhoughthelatterisargu-
ablymorefamiliarandlikelytoleadtogreateramountsofelaboration.Consistentwiththeirhypoth-
esis,betterretentionfortheratedwordswasobtainedinthegroupprocessingtheancestralscenario.
Theexperimentsreportedherehadthreemaingoals:first,giventheirpotentialimportancetothe
evolutionaryhypothesis,wesoughtinitiallytoreplicatethefindingsofWeinsteinetal.(2008).Will
ratingtherelevanceofwordstoanancestralscenarioleadtobetterretentionthanprocessingthose
samewordsinamatchedmoderncontext?Second,tohelpestablishthegeneralityofthephenome-
nonweexploredtheroleofancestralprioritiesintwonewdomains,healingadangerousinfectionand
collectingfoodtoeat.Bothdescribedrecurrentadaptiveproblemsthatwerelikelytobepresentinthe
environmentofevolutionaryadaptation,butthenewdomainsweredesignedtoactivate‘‘gathering”
strategiesratherthanpredatoravoidancestrategiessuchasthoseusedintheWeinsteinetal.exper-
iment.Onceagain,allparticipantswereaskedtoprocesstherelevanceofwordstoasurvivalscenario.
Whatdifferedacrossconditionswastheancestralnatureofthescenario:participantswerepresented
withthesamegeneralsurvivalproblembutineitheranancestral(grasslands)oramoderncontext
(city).Totheextentthatourmemorysystemsare‘‘prepared”toprocessinformationinancestralenvi-
ronments(foragingashunter–gatherersinthegrasslands),weanticipatedbetterretentionafterpro-
cessingwiththeancestralscenario.Finally,inanefforttoprovidegreaterinsightintotheproximate
mechanismsthatmightunderlieancestraladvantages,andtogainadditionalsupportfortheevolu-
tionary account, in the last two experiments we also asked participants to evaluate the ancestral
andmodernscenariosalonganumberofmnemonically-relevantdimensions(e.g.,interest,familiarity,
etc.).
2. Experiment1
Inthesurvivalprocessingparadigm,participantsareaskedtoratetherelevanceofwordstoasur-
vivalscenario.Theratingtaskisthenfollowedbyasurpriseretentiontest,usuallyfreerecall,andper-
formanceiscomparedtoanon-fitness-basedcontrol.FollowingNairneetal.(2007),Weinsteinetal.
(2008)usedaschematiccontrol,movingtoaforeignland,astheircomparison.The‘‘moving”control
iscloselymatchedtothesurvivalscenario,butisnotfitness-relevant(thatis,itdoesnotrequiredeal-
ingwithissuesdirectlyrelatedtosurvivalorreproduction).Betterrecallforthewordsratedwiththe
survivalscenario,comparedtothemovingcondition,definesthebasicsurvivalprocessingadvantage.
FollowingtheWeinsteinetal.(2008)study(Experiment2)wecomparedtwoversionsofthesur-
vival scenario in Experiment 1—one rooted in an ancestral environment (avoiding predators in the
grasslands) and one in a modern context (avoiding attackers in a city). The critical question asked
whetherthesizeofthesurvivalprocessingadvantagewouldvaryacrossthesetwosurvivalscenarios.
J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22 5
Weinsteinetal.foundthatthesurvivalprocessingadvantagewaslargerwhenpeopleratedtherele-
vanceofwordstotheancestral(grasslands)scenario,consistentwithanevolutionaryaccount.
2.1. Method
2.1.1. Participantsandapparatus
EightyPurdueundergraduatesparticipatedforcreditinanintroductorypsychologycourse.Stimuli
werepresentedandcontrolledbypersonalcomputersinsessionslastingapproximately30min.Ses-
sionswereconductedingroupsofuptofourparticipants.
2.1.2. Materialsanddesign
Thirty-twounrelatedconcretenouns,drawnlargelyfromtheupdatedBattigandMontaguenorms
(Van Overschelde, Rawson, & Dunlosky, 2004) or from the extended Paivio norms (Clark & Paivio,
2004), were used as target words in the experiment (six additional concrete nouns were used in a
practice phase). The stimuli are presented in the Appendix A. Everyone rated and recalled exactly
the same 32words, presented in the same order. Withina session participantsrated the relevance
of 16 target words to the survival (S) scenario and 16 to the moving (M) control. Rating condition
wasblockedintrialsofeightwordsintheformSMSMorMSMS;halfoftheparticipantsreceivedeach
version,ensuringthateachwordwasratedequallyoftenunderbothscenarios.Thecontentofthesur-
vivalscenariowasmanipulatedbetween-subjects:onegroupofparticipantswasgiventheancestral
(grasslands) scenario (N=40); the other group received the modern (city) scenario (N=40). Both
groups received the same ‘‘moving” instructions. Immediately following the last block of ratings,
everyonecompletedashortdigit-distractortaskfollowedbyasurprisefreerecalltest.
2.1.3. Procedure
On arrival in the laboratory, participants were assigned randomly to either the ancestral or the
modern group. The instructions for the two survival scenarios are presented below, along with the
instructionsforthemovingcontrol.
2.1.3.1. Ancestral. Inthistask,pleaseimaginethatyouarestrandedinthegrasslandsofaforeignland,
withoutanybasicsurvivalmaterials.Youhaverecentlyreceivedwordthatadangerouspredatorhas
beenseeninthearea.Youwillneedtoavoidand/orescapefromthepredatortoensureyoursurvival.
Wearegoingtoshowyoualistofwords,andwewouldlikeyoutoratehowrelevanteachofthese
wordswouldbeinyourattempttoavoidthepredator.
2.1.3.2. Modern. Inthistask,pleaseimaginethatyouarestrandedinthecityofaforeignland,without
anybasicsurvivalmaterials.Youhaverecentlyreceivedwordthatadangerousattackerhasbeenseen
inyourarea.Youwillneedtoavoidand/orescapefromtheattackertoensureyoursurvival.Weare
goingtoshowyoualistofwords,andwewouldlikeyoutoratehowrelevanteachofthesewords
wouldbeinyourattempttoavoidtheattacker.
2.1.3.3. Moving. Inthistask,pleaseimaginethatyouareplanningtomovetoanewhomeinaforeign
land.Overthenextfewmonths,you’llneedtolocateandpurchaseanewhomeandtransportyour
belongings.Wearegoingtoshowyoualistofwords,andwewouldlikeyoutoratehowrelevanteach
ofthesewordswouldbeforyouinaccomplishingthistask.
Theto-be-ratedwordswerepresentedindividually,centeredonthescreen,for5sapiece.People
wereaskedtoratethewordsona5-pointscale,with1indicatingtotallyirrelevanttothescenario
and5signifyingextremelyrelevant.Theratingscalewasshownjustbeloweachwordandpeoplere-
spondedbyclickingontheirvalueofchoice.Theywerecautionedtorespondquickly,withinthe5-s
presentationwindow,andthelaterretentiontestwasnotmentioned.Ashortpracticesessionpre-
cededeachofthefirsttwoblocks(oneSandoneM).
Afterthelastwordwasrated,instructionsappearedforthedigitrecalltask.Forthistask,sevendig-
its,rangingbetweenzeroandnine,werepresentedsequentiallyfor1sapieceandparticipantswere
requiredtorecallthedigitsinorderbytypingresponsesintoatextbox.Thedigitrecalltasklastedfor
6 J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22
approximately2min.Recallinstructionsthenappeared.Participantswereaskedtowritedownthe
earlier-ratedwords,inanyorder,onaresponsesheet.Thefinalrecallphaselastedfor10minandpar-
ticipantswereaskedtodrawalineontherecallsheet,underthelastrecalledword,aftereachminof
recall.Aclockwasdisplayedonthecomputermonitoranda‘‘beep”soundedeveryminsignalingthe
participantstodrawtheline.Usingthisprocedureallowsonetoconstructcumulativerecallcurves,
buttheyarenotreportedhere.
2.2. Resultsanddiscussion
Thelevelofstatisticalsignificance,unlessotherwisenoted,wassetatp<.05forallcomparisons.
Participantshadlittledifficultyproducingtherelevanceratingswithintheallottedtime,andnosig-
nificantdifferencesincompletionrateswerefoundacrossgroupsorconditions.
The data of main interest are shown in Fig. 1. These data show proportion correct recall for the
rated target words for each of the main comparisons. The left-hand side of the figure shows recall
for thoseparticipantsreceiving the ancestralscenario; the moderndata are shownto the right. An
overall analysis of variance (ANOVA) on these data, with group (ancestral versus modern) as a be-
tween-subjectvariable,andratingcondition(survivalversusmoving)asawithin-subjectvariable,re-
vealedonlyasignificantinteractionbetweengroupandratingcondition,[F(1,78)=6.27,MSe=.016,
g2 :074].Plannedt-testsrevealedasignificantsurvivaladvantageintheancestralgroup(ancestral
p¼
versusmoving;t(39)=2.91,p<.01);thecomparablecomparisoninthemoderngroup(modernversus
moving) failed to approach significance [t(39)<1.0]. Turning to the direct planned comparison be-
tween the ancestral and modern scenarios, participants recalled more words after rating with the
ancestralscenariobutthedifferencewasonlymarginallyreliable[t(78)=1.65,p=.052].Additional
analysesusingitemsastheanalyticunit,ratherthanparticipants,revealedthesamepatternsofsig-
nificance—however,therecalldifferencebetweentheancestralandthemodernscenariosdidreach
conventionallevelsofstatisticalsignificanceintheitemanalysis[t(31)=1.93,p=.031].
These results confirm the main predictions of the evolutionary account. There was a significant
interaction between group and rating condition, indicating a larger survival effect for the ances-
trally-basedgrasslandsscenario.Infact,althoughtherewasarobustfitnesseffectfortheancestral
scenario,therewasnoindicationthatsurvival-basedprocessingproducedanymnemonicadvantages
in the modern scenario. Interestingly, Weinstein et al. (2008) found essentially the same pattern,
althoughtheinteractionofgroupandratingconditionisnotreportedintheirarticle.Theyreportonly
theoverallmeansfortheancestral(.38),modern(.31),andmovingconditions(.28),butthepatternis
clearlysimilartotheonereportedhere.Itissomewhatsurprisingthatthecityscenariofailedtopro-
ducearetentionadvantageoverthemovingcontrol.Survivinginacityandescapingfromattackers
0.60
0.55
call 0.50
e
R
ct 0.45
e
orr 0.40
C
n
ortio 0.35
op 0.30
Pr
0.25
0.20
Ancestral Moving Modern Moving
Fig.1. ProportioncorrectrecallperformanceforeachoftheconditionsinExperiment1.Thedataontheleftarefromthegroup
receivingtheancestral(predator)scenario;datafromthegroupreceivingthemodern(attacker)scenarioareshownonthe
right.Errorbarsrepresentwithin-subject95%confidenceintervals(asperMassonandLoftus(2003)).
J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22 7
Table1
Ratingandresponsetime(ms)averagesforthevariousconditionsinExperiment1.
Ancestral Modern
Predator Moving Attacker Moving
Rating 3.30 2.63 3.07 2.74
Responsetime 2341 2486 2362 2322
areclearlyfitness-relevant,yettherewasnomemoryenhancementforsurvivalprocessinginthiscon-
text.WereturntothisissueinExperiments2and3.
Itisalsoofinteresttoexaminetheaverageratingsandresponsetimesforeachoftheconditions;
thesedataareshowninTable1.Turningfirsttotheratingdata,anoverallANOVArevealedamain
effectofratingcondition[F(1,78)=95.77,MSe=.104,g2 :551];thetargetwordswereratedasmore
p¼
relevanttothesurvivalscenariosthantothemovingscenario.Therewasalsoareliableinteraction
betweengroupandratingcondition[F(1,78)=11.22,MSe=.104,g2 :126],reflectingalargerdiffer-
p¼
enceinratingbetweenthesurvivalscenarioandmovingfortheancestralgroup.Individualpost-hoc
comparisonsdeterminedthatthedifferencesbetweenthesurvivalandmovingconditionswerereli-
ableforboththeancestralandthemoderngroups(botht-testvaluesweregreaterthan4.5)andthere
wasasignificantratingadvantagefortheancestralscenariooverthemodernscenario[t(78)=2.29,
p<.03].Thispatternisalittletroublingbecauseitmirrorsthepatternfoundinrecall,suggestingthat
therecalldifferencesmayhavebeendeterminedbytheratingdifferences.However,thereareseveral
reasonstodiscountthispossibility.First,priorresearchinthesurvivalprocessingparadigmhascon-
sistently shown that rating differences cannot account for observed differences in recall (Nairne &
Pandeirada,2008a;Nairneetal.,2007).Second,supportingthisconclusion,thecurrentdatashowa
highlyreliableratingadvantageforthemodernscenariooverthemovingcontrol,yetthereisnoevi-
denceforacomparabledifferenceinrecall.Finally,therecallresultsofExperiment1essentiallyrep-
licatethoseofWeinsteinetal.(2008),whofoundnosignificantratingdifferencesacrossconditions.
Analysisoftheresponsetimedatashowednomaineffectsofconditionorgroup,butdidshowa
reliablegroup conditioninteraction[F(1,78)=7.27,MSe=46782.17,g2 :085].Post-hoctestsre-
" p¼
vealed that participants in the ancestral group took significantly longer to make relevance ratings
tothemovingscenariothantheydidtothesurvivalscenario[t(39)= 2.97,p<.01];therewereno
#
similarsignificantdifferencesfortheparticipantsinthemoderngroup.Asaresult,differencesinre-
sponsetimescannoteasilybeusedtoexplaintherecalldifferencesthatwereobtained.
The results of Experiment 1 replicate the general patterns found previously by Weinstein et al.
(2008).Therecalldifferencebetweentheancestralandmodernscenarioswassmallerthanreported
by Weinstein et al. (2008), and statistically significant only in the item analysis, but the important
interaction of group by condition was demonstrated unequivocally. The survival processing advan-
tage,definedastheadvantageofsurvivalprocessingoveranon-fitness-basedmovingcontrol,waslar-
ger in the ancestral than in the modern group. This data pattern is broadly consistent with the
evolutionaryaccount,suggestingthatourmemorysystemsaretunedtofitness-relevantprocessing,
butespeciallyincontextsthatmirrorthoselikelytohavebeenpresentintheenvironmentofevolu-
tionaryadaptation.
3. Experiment2
Critics of evolutionary accounts often focus on our lack of knowledge about ancestral environ-
ments,whichsomescholarsbelieveareessentiallyunknowable(e.g.,Buller,2005;Reeve&Sherman,
2007). Moreover, the idea that there is a single ancestral environment,whichuniquely defined the
selectionpressuresthatledtothedevelopmentofhumancognition,isdebatable.Atthesametime,
ourancestorslikelyfacedrecurrentadaptiveproblems,onesthatremainedconstantacrosssituations.
Ourancestorscertainlyneededtofindfood,potentialmates,trackandescapefrompredators,andfind
remedies for injuries and disease (see Tooby & Cosmides, 2005). Cognitive adaptations may have
8 J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22
developedtodealwiththesespecificproblems(seeBarrett,2008)and,moreimportantly,thoseadap-
tationsmaycontinuetobearthegeneralimprintofourspecificforagingpast.
TheancestralscenariousedinExperiment1wasdesignedtotaponesuchrecurrentproblemfaced
byourforagingancestors—survivingandescapingfrompredatorsinagrasslandsenvironment.Exper-
iment 2 explores another ancestral problem,one designed to activate a ‘‘gathering” strategy rather
thanonedealingexclusivelywithpredatoravoidance.Onceagain,twoversionsofthesurvivalsce-
narioweredeveloped,onerootedinagrasslandsenvironment(ancestral)andoneinthecity(mod-
ern). In both cases participants were told they had been hurt and a dangerous infection might be
developing. In the ancestral scenario, the task was to search and find relevant medicinal plants in
anattempttocuretheinfection;inthemodernscenario,thetaskwastofindrelevantantibiotics.Sce-
nariotypewasmanipulatedbetween-subjectsandeachgroupreceivedanon-fitness-basedcontrolas
well.RatherthanusingmovingasourcontrolinExperiment2,however,participantswereaskedto
makepleasantness ratingsabout half of thewords. Pleasantnessratings have been used frequently
as a control in the survival processing paradigm (e.g., Nairne et al., 2008; Weinstein et al., 2008),
andaregenerallyconsideredtobeaneffectiveformofdeepprocessing(e.g.,Packman&Battig,1978).
3.1. Method
3.1.1. Participantsandapparatus
OnehundredandtwentyPurdueundergraduatesparticipatedforcoursecredit.Sixtypeoplewere
assignedtoeachscenariogroup.Stimuliwerepresentedandcontrolledbypersonalcomputers.Ses-
sionswereconductedingroupsofuptofourparticipants.
3.1.2. Materialsanddesign
TargetwordsweredrawnfromthesamepoolusedinExperiment1.Onceagain,everyonerated
andrecalledexactlythesamesetofwordspresentedinthesameorder.However,unlikeinExperi-
ment1,theratingtask(survivalscenarioversuspleasantness)wasnotblocked,butwasdistributed
randomlythroughoutthesession.(Intermixingcanhelptoreducevariabilitycausedbyhavingdiffer-
entconditionsblockedatthebeginningandendofasession.)Presentationorderwasconstrainedas
follows:nomorethantworatingsofagiventypecouldoccurinarowandeachratingtypehadto
occurequallyofteninthefirstandsecondhalfofthelist.Ratingtaskwasalsocounterbalancedacross
participantstoensurethateachtargetwordwasratedforsurvivalandpleasantnessanequalnumber
oftimes.AsinExperiment1,theratingtaskwasfollowedbyashortdistractortaskandthenasurprise
freerecalltest.
3.1.3. Procedure
On arrival in the laboratory, participants were assigned randomly to either the ancestral or the
modern group.The instructionsfor the two survival scenarios are presented below, along with the
instructionsforthepleasantnesscontrol.
3.1.3.1. Ancestral. Forthesurvivalsituation,pleaseimaginethatyouarestrandedinthegrasslandsof
a foreign land, without any basic survival materials. You have recently been hurt and a dangerous
infectionmightbedeveloping.Youwillneedtosearchandfindrelevantmedicinalplantstoensure
your survival. We would like you to rate how relevant the word would be in your attempt to cure
theinfection.Thescaleofrelevancerangesfromonetofive,withone(1)indicatingtotallyirrelevant
andfive(5)signifyingextremelyrelevant.Someofthewordsmayberelevantandothersmaynot—it’s
uptoyoutodecide.
3.1.3.2. Modern. Forthesurvivalsituation,pleaseimaginethatyouarestrandedinthecityofaforeign
land, without any basic survival materials. You have recently been hurt and a dangerous infection
might be developing. You will need to search and find relevant antibiotics to ensure your survival.
We would like you to rate how relevant the word would be in your attempt to cure the infection.
Thescaleofrelevancerangesfromonetofive,withone(1)indicatingtotallyirrelevantandfive(5)
J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22 9
signifyingextremelyrelevant.Someofthewordsmayberelevantandothersmaynot—it’suptoyou
todecide.
3.1.3.3. Pleasantness. Fortheotherdimension,wewouldlikeyoutoratethepleasantnessoftheword.
Thescaleofrelevancerangesfromonetofive,withone(1)indicatingtotallyunpleasantandfive(5)
signifyingextremelypleasant.Someofthewordsmaybepleasantandothersmaynot—it’suptoyou
todecide.
AsinExperiment1,theto-be-ratedwordswerepresentedindividually,centeredonthescreen,for
5sapiece.Theratingscalewasshownjustbeloweachwordandpeoplerespondedbyclickingontheir
valueofchoice.Boththesurvivalscenario(ancestralormodern)andthepleasantnesstaskwerefully
explainedtoeachparticipantatthebeginningofthesession.Inthesessionitself,eachwordwaspre-
sentedalongwithaquestionthatexplainedtheratingtasktobeperformed(Howrelevantisthisword
to the survival situation?; How pleasant is this word?). Participants were cautioned to respond
quickly,withinthe5-spresentationwindow,andthelaterretentiontestwasnotmentioned.Thepro-
cedurefortheshortdistractorperiodandtherecalltestmatchedExperiment1.
3.2. Resultsanddiscussion
AsinExperiment1,participantshadnotroublecompletingtherespectiveratingdecisionswithin
the5-spresentationwindow.ThedataofmaininterestareshowninFig.2,whichshowsproportion
correctrecallperformanceforthewordsratedinthesurvivalandpleasantnesstasks.Theleft-hand
side of the figure shows performance for the group receiving the ancestral scenario; the data for
themoderngroupareshownontheright.
AmixedANOVAwithgroup(ancestralversusmodern)asabetween-subjectvariableandcondition
(survival versus pleasantness) as a within-subject variable revealed a highly reliable main effect of
condition [F(1,118)=45.61, MSe=.010, g2 :279] as well as a reliable interaction between group
p¼
andcondition[F(1,118)=5.31,MSe=.010,g2 :043].Pairedt-testsconfirmedasurvivaladvantage
p¼
(compared to pleasantness) for both the ancestral [t(59)=6.14, p<.001] and the modern groups
[t(59)=3.30, p<.002]. Mostimportantly, rating wordswithrespecttothe ancestralscenarioled to
better recall performance than did rating words about the modern scenario [t(118)=2.19, p=.03].
Additional item analyses revealed the same patterns, including the significant advantage for the
ancestraloverthemodernscenario[t(31)=2.04,p<.05],butthedifferencebetweensurvivalprocess-
ingandpleasantnessdidnotquitereachstatisticalsignificanceinthemoderncondition[t(31)=1.88,
p<.07].
0.60
0.55
all
c 0.50
e
R
ct 0.45
e
Corr 0.40
n
ortio 0.35
p
o 0.30
Pr
0.25
0.20
Ancestral Pleasantness Modern Pleasantness
Fig.2. ProportioncorrectrecallperformanceforeachoftheconditionsinExperiment2.Thedataontheleftarefromthegroup
receivingtheancestral(medicinalplants)scenario;datafromthegroupreceivingthemodern(antibiotics)scenarioareshown
ontheright.Errorbarsrepresentwithin-subject95%confidenceintervals(asperMassonandLoftus(2003)).
10 J.S.Nairne,J.N.S.Pandeirada/CognitivePsychology61(2010)1–22
Table2
Ratingandresponsetime(ms)averagesforthevariousconditionsinExperiment2.
Ancestral Modern
Medicinalplants Pleasantness Antibiotics Pleasantness
Rating 3.50 2.91 3.38 2.92
Responsetime 3104 2864 3101 2849
Average rating and response time data are shown in Table 2 for each group and condition. The
mixedANOVAontheratingdatashowedthatthesurvivalratingsweresignificantlyhigherthanpleas-
antnessratings[F(1,118)=114.67,MSe=.145,g2 :493];neitherthemaineffectofscenarionorthe
p¼
group ratingconditioninteractionwassignificantinthisanalysis.TheresultsofamixedANOVAon
"
responsetimerevealedonlyamaineffectofratingcondition:participantstooksignificantlylongerto
make survival ratings than pleasantness ratings [F(1,118)=81.15, MSe=44685.18, g2 :407]. The
p¼
samepatternsofresultswereobtainedintheitem-basedanalysis.
Generally,theresultsofExperiment2replicatethemainfindingsofExperiment1.Onceagain,the
criticalgroup conditioninteractionwassignificant—therewasalargersurvivalprocessingeffectfor
"
the group receiving the ancestral (grasslands) scenario. Moreover, individual planned comparisons
confirmedthatprocessing withtheancestralscenarioproducedhigher levels ofrecallthan did the
modernscenario—and,importantly,neithertheancestralnorthemodernscenariosproducedsignif-
icant differences on either the rating or the response time measures. Consequently, the ancestral
advantage reported here provides additional supporting evidence for the evolutionary account: fit-
ness-based processing is particularly effective in contexts that presumably mirror those found in
theenvironmentsofevolutionaryadaptation.
Itisinterestingtonotethatsurvivalprocessingproducedenhancedrecallforboththeancestral
andthemodernscenariosinExperiment2,relativetothepleasantnesscontrol(althoughthemod-
ern-pleasantness advantage was only marginally significant in the item analysis). In Experiment 1,
there was a robust survival advantage only for the ancestral scenario. The pattern seen here could
bedue,inpart,toouruseofthe‘‘pleasantness”ratherthan‘‘moving”controlcondition.Ratingitems
forpleasantnessisahighlyeffectiveformofdeepprocessing,butittendstoinduceprimarilyitem-
specificratherthanrelational(oracross-item)processing.Itiswell-knownthatitem-specificprocess-
ingismosteffectivewhento-be-recalleditemsarerelated,suchaswhenalistiscategorized,because
ithelpsonediscriminatebetweenitemsthatdidanddidnotoccur(seeHunt&Einstein,1981;Hunt&
McDaniel,1993).Giventhatourto-be-ratedwordswereunrelated(seeAppendixA),itispossiblethat
both survival scenarios induced greater amounts of relationalprocessing than did the pleasantness
tasktherebyaccounting,atleastinpart,forthesurvivaladvantagesseeninthisexperiment.However,
Nairne and Pandeirada (2008a) have shown that survival processing produces better recall perfor-
mancethanpleasantnessprocessingevenwhenwordlistsarecategorized,soitisunlikelythaten-
hanced relational processing can completely explain the survival advantages seen here, or in other
experimentsusingthesurvivalprocessingparadigm.Moreover,importantly,thecriticalcomparison
betweentheancestralandmodernscenarioswasreliableinboththesubjectandtheitemanalyses
anddoesnotdependontheuseofthepleasantnesscontrol(seealsoExperiment4).
4. Experiment3
TheancestralscenariosofExperiments1and2weredesignedtoevokerecurrentadaptiveprob-
lems faced by our foraging ancestors—escaping from predators and treating a dangerous infection.
Experiment3investigatesyetanotherbasicrequirement—searchingfornourishment.Onceagain,par-
ticipantswererequiredtoratetherelevanceofwordseithertoasurvivalscenario,onerootedinthe
grasslandsandoneinacity,orforpleasantness.Thesurvivaltaskaskedparticipantstoimaginegain-
ingneedednourishmenteitherbysearchingforandbuyingfoodinacity,orbysearchingforandgath-
ering edible plants in the grasslands. Consistent with the evolutionary account, and the results of
Experiments 1 and 2, we anticipated the best final free recall for words rated with respect to the
ancestralscenario.