A Brief Overview of Adult Attachment Theory and Research

Research on adult attachment is guided by the assumption that the same motivational system that gives rise to the close emotional bond between parents and their ... Togglenavigation R.ChrisFraley About Research Publications Lab Teaching Resources AdultAttachmentTheoryandResearch ♦ ABriefOverview Summary Researchonadultattachmentisguidedbytheassumptionthatthesamemotivational systemthatgivesrisetothecloseemotionalbondbetweenparentsand theirchildrenisresponsibleforthebondthatdevelopsbetweenadults inemotionallyintimaterelationships.Theobjectiveofthisessayis toprovideabriefoverviewofthehistoryofadultattachmentresearch,thekeytheoretical ideas,andasamplingofsomeoftheresearchfindings.Thisessayhasbeenwrittenforpeoplewhoareinterestedinlearningmoreaboutresearchonadultattachment. Background:Bowlby'sTheoryofAttachment The theoryofattachmentwasoriginallydevelopedbyJohnBowlby(1907-1990), aBritishpsychoanalystwhowasattemptingtounderstandtheintensedistress experiencedbyinfantswhohadbeenseparatedfromtheirparents.Bowlby observedthatseparatedinfantswouldgotoextraordinarylengths(e.g., crying,clinging,franticallysearching)toprevent separationfromtheirparentsortoreestablishproximitytoamissingparent. AtthetimeofBowlby'sinitialwritings,psychoanalyticwritersheldthattheseexpressionsweremanifestations ofimmaturedefensemechanismsthatwereoperatingtorepressemotional pain,butBowlbynotedthatsuchexpressionsarecommontoawidevariety ofmammalianspecies,andspeculatedthatthesebehaviorsmayservean evolutionaryfunction. Drawingonethologicaltheory,Bowlbypostulatedthattheseattachmentbehaviors, suchascryingandsearching,wereadaptiveresponsestoseparationfrom aprimaryattachmentfigure--someonewhoprovidessupport, protection,andcare.Becausehumaninfants,likeothermammalianinfants, cannotfeedorprotectthemselves,theyaredependentuponthecareand protectionof"olderandwiser"adults.Bowlbyarguedthat, overthecourseofevolutionaryhistory,infantswhowereabletomaintain proximitytoanattachmentfigureviaattachmentbehaviorswouldbemorelikelytosurvivetoareproductive age.AccordingtoBowlby,amotivationalsystem,whathecalled theattachmentbehavioralsystem,wasgradually"designed" bynaturalselectiontoregulateproximitytoanattachmentfigure. Theattachmentbehaviorsystemisanimportantconceptinattachmenttheory becauseitprovidestheconceptuallinkagebetweenethologicalmodels ofhumandevelopmentandmoderntheoriesonemotionregulationandpersonality. AccordingtoBowlby,theattachmentsystemessentially"asks" thefollowingfundamentalquestion:Istheattachmentfigurenearby,accessible, andattentive?Ifthechildperceivestheanswertothisquestiontobe "yes,"heorshefeelsloved,secure,andconfident,and,behaviorally, islikelytoexplorehisorherenvironment,playwithothers,andbe sociable.If,however,thechildperceivestheanswertothisquestion tobe"no,"thechildexperiencesanxietyand,behaviorally, islikelytoexhibitattachmentbehaviorsrangingfromsimplevisualsearching onthelowextremetoactivefollowingandvocalsignalingontheother(seeFigure1). Thesebehaviorscontinueuntileitherthechildisabletoreestablishadesirablelevel ofphysicalorpsychologicalproximitytotheattachmentfigure,oruntil thechild"wearsdown,"asmayhappeninthecontextofaprolonged separationorloss.Insuchcases,Bowlbybelievedthatyoungchildrenexperiencedprofounddespairanddepression. IndividualDifferencesinInfantAttachmentPatterns AlthoughBowlbybelievedthatthebasicdynamicsdescribedabovecapturedthenormative dynamicsoftheattachmentbehavioralsystem,herecognizedthatthere areindividualdifferencesinthewaychildrenappraisetheaccessibility oftheattachmentfigureandhowtheyregulatetheirattachmentbehavior inresponsetothreats.However,itwasn'tuntilhiscolleague,Mary Ainsworth(1913–1999),begantosystematicallystudyinfant-parentseparationsthat aformalunderstandingoftheseindividualdifferenceswasarticulated. Ainsworthandherstudentsdevelopedatechniquecalledthestrange situation--alaboratoryparadigmforstudyinginfant-parentattachment. Inthestrangesituation,12-month-oldinfantsandtheirparentsarebrought tothelaboratoryand,systematically,separatedfromandreunitedwithoneanother.Inthe strangesituation,mostchildren(i.e.,about60%)behaveinthewayimplied byBowlby's"normative"theory.Theybecomeupsetwhentheparent leavestheroom,but,whenheorshereturns,theyactivelyseektheparent andareeasilycomfortedbyhimorher.Childrenwhoexhibitthispattern ofbehaviorareoftencalledsecure.Otherchildren(about20% orless)areill-at-easeinitially,and,uponseparation,becomeextremely distressed.Importantly,whenreunitedwiththeirparents,thesechildren haveadifficulttimebeingsoothed,andoftenexhibitconflictingbehaviors thatsuggesttheywanttobecomforted,butthattheyalsowantto"punish" theparentforleaving.Thesechildrenareoftencalledanxious-resistant. ThethirdpatternofattachmentthatAinsworthandhercolleaguesdocumented iscalledavoidant.Avoidantchildren(about20%)don'tappear toodistressedbytheseparation,and,uponreunion,activelyavoidseeking contactwiththeirparent,sometimesturningtheirattentiontoplayobjects onthelaboratoryfloor. Ainsworth'sworkwasimportantforatleastthreereasons.First,sheprovidedone ofthefirstempiricaldemonstrationsofhowattachmentbehaviorispatterned inbothsafeandfrighteningcontexts.Second,sheprovidedthefirst empiricaltaxonomyofindividualdifferencesininfantattachmentpatterns. Accordingtoherresearch,atleastthreetypesofchildrenexist:those whoaresecureintheirrelationshipwiththeirparents,thosewhoare anxious-resistant,andthosewhoareanxious-avoidant.Finally,shedemonstrated thattheseindividualdifferenceswerecorrelatedwithinfant-parentinteractions inthehomeduringthefirstyearoflife.Childrenwhoappearsecure inthestrangesituation,forexample,tendtohaveparentswhoareresponsive totheirneeds.Childrenwhoappearinsecureinthestrangesituation (i.e.,anxious-resistantoravoidant)oftenhaveparentswhoareinsensitive totheirneeds,orinconsistentorrejectinginthecaretheyprovide.Intheyearsthathavefollowed,anumberof researchershavedemonstratedlinksbetweenearlyparentalsensitivityandresponsivenessand attachmentsecurity. AdultRomanticRelationships AlthoughBowlbywasprimarilyfocusedonunderstandingthenatureoftheinfant-caregiver relationship,hebelievedthatattachmentcharacterizedhumanexperience from"thecradletothegrave."Itwasnotuntilthemid-1980's, however,thatresearchersbegantotakeseriouslythepossibilitythat attachmentprocessesmayplayoutinadulthood.HazanandShaver(1987) weretwoofthefirstresearcherstoexploreBowlby'sideasinthecontext ofromanticrelationships.AccordingtoHazanandShaver,theemotional bondthatdevelopsbetweenadultromanticpartnersispartlyafunction ofthesamemotivationalsystem--theattachmentbehavioralsystem--that givesrisetotheemotionalbondbetweeninfantsandtheircaregivers. HazanandShavernotedthattherelationshipbetweeninfantsandcaregiversandtherelationshipbetweenadultromantic partnerssharethefollowingfeatures: bothfeelsafewhentheotherisnearbyandresponsive bothengageinclose,intimate,bodilycontact bothfeelinsecurewhentheotherisinaccessible bothsharediscoverieswithoneanother bothplaywithoneanother'sfacialfeaturesandexhibitamutualfascination andpreoccupationwithoneanother bothengagein"babytalk" Onthebasisoftheseparallels,HazanandShaverarguedthatadultromantic relationships,likeinfant-caregiverrelationships,areattachments,and thatromanticloveisapropertyoftheattachmentbehavioralsystem, aswellasthemotivationalsystemsthatgiverisetocaregivingandsexuality. ThreeImplicationsofAdultAttachmentTheory Theideathatromanticrelationshipsmaybeattachmentrelationshipshashad aprofoundinfluenceonmodernresearchoncloserelationships.There areatleastthreecriticalimplicationsofthisidea.First,ifadult romanticrelationshipsareattachmentrelationships,thenweshouldobserve thesamekindsofindividualdifferencesinadultrelationshipsthatAinsworth observedininfant-caregiverrelationships.Wemayexpectsomeadults, forexample,tobesecureintheirrelationships--tofeelconfident thattheirpartnerswillbethereforthemwhenneeded,andopentodepending onothersandhavingothersdependonthem.Weshouldexpectotheradults, incontrast,tobeinsecureintheirrelationships.Forexample,some insecureadultsmaybeanxious-resistant:theyworrythatothers maynotlovethemcompletely,andbeeasilyfrustratedorangeredwhen theirattachmentneedsgounmet.Othersmaybeavoidant:theymay appearnottocaretoomuchaboutcloserelationships,andmayprefer nottobetoodependentuponotherpeopleortohaveothersbetoodependent uponthem. Second,ifadultromanticrelationshipsareattachmentrelationships,thenthe wayadultrelationships"work"shouldbesimilartotheway infant-caregiverrelationshipswork.Inotherwords,thesamekinds offactorsthatfacilitateexplorationinchildren(i.e.,havingaresponsive caregiver)shouldfacilitateexplorationamongadults(i.e.,havinga responsivepartner).Thekindsofthingsthatmakeanattachmentfigure "desirable"forinfants(i.e.,responsiveness,availability) arethekindsoffactorsadultsshouldfinddesirableinromanticpartners. Inshort,individualdifferencesinattachmentshouldinfluencerelational andpersonalfunctioninginadulthoodinthesamewaytheydoinchildhood. Third,whetheranadultissecureorinsecureinhisorheradultrelationships maybeapartialreflectionofhisorherexperienceswithhisorherprimarycaregivers. Bowlbybelievedthatthementalrepresentations orworkingmodels(i.e.,expectations,beliefs,"rules" or"scripts"forbehavingandthinking)thatachildholdsregarding relationshipsareafunctionofhisorhercaregivingexperiences.For example,asecurechildtendstobelievethatotherswillbetherefor himorherbecausepreviousexperienceshaveledhimorhertothisconclusion. Onceachildhasdevelopedsuchexpectations,heorshewilltendtoseek outrelationalexperiencesthatareconsistentwiththoseexpectations andperceiveothersinawaythatiscoloredbythosebeliefs.According toBowlby,thiskindofprocessshouldpromotecontinuityinattachment patternsoverthelifecourse,althoughitispossiblethataperson's attachmentpatternwillchangeifhisorherrelationalexperiencesare inconsistentwithhisorherexpectations.Inshort,ifweassumethat adultrelationshipsareattachmentrelationships,itispossiblethat childrenwhoaresecureaschildrenwillgrowuptobesecureintheir romanticrelationships.Or,relatedly,thatpeoplewhoaresecureasadultsintheirrelationshipswiththeirparentswillbemorelikelytoforgesecurerelationshipswithnewpartners. InthesectionsbelowIbrieflyaddressthesethreeimplicationsinlight ofearlyandcontemporaryresearchonadultattachment. DoWeObservetheSameKindsofAttachmentPatternsAmongAdultsthatWe ObserveAmongChildren? Theearliestresearchonadultattachmentinvolvedstudyingtheassociation betweenindividualdifferencesinadultattachmentandthewaypeople thinkabouttheirrelationshipsandtheirmemoriesforwhattheirrelationships withtheirparentsarelike.HazanandShaver(1987)developedasimple questionnairetomeasuretheseindividualdifferences.(Theseindividual differencesareoftenreferredtoasattachmentstyles,attachment patterns,attachmentorientations,ordifferencesinthe organizationoftheattachmentsystem.)Inshort,HazanandShaver askedresearchsubjectstoreadthethreeparagraphslistedbelow,and indicatewhichparagraphbestcharacterizedthewaytheythink,feel, andbehaveincloserelationships: A. Iamsomewhatuncomfortablebeingclosetoothers;Ifinditdifficult totrustthemcompletely,difficulttoallowmyselftodependonthem. Iamnervouswhenanyonegetstooclose,andoften,otherswantmeto bemoreintimatethanIfeelcomfortablebeing. B. Ifinditrelativelyeasytogetclosetoothersandamcomfortable dependingonthemandhavingthemdependonme.Idon'tworryabout beingabandonedoraboutsomeonegettingtooclosetome. C. IfindthatothersarereluctanttogetascloseasIwouldlike.I oftenworrythatmypartnerdoesn'treallylovemeorwon'twantto staywithme.Iwanttogetveryclosetomypartner,andthissometimes scarespeopleaway. Basedonthisthree-category measure,HazanandShaverfoundthatthedistributionofcategories wassimilartothatobservedininfancy.Inotherwords,about60%of adultsclassifiedthemselvesassecure(paragraphB),about20%described themselvesasavoidant(paragraphA),andabout20%describedthemselves asanxious-resistant(paragraphC). Althoughthismeasureservedasausefulwaytostudytheassociationbetweenattachment stylesandrelationshipfunctioning,itdidn'tallowafulltestofthe hypothesisthatthesamekindsofindividualdifferencesobservedininfants mightbemanifestamongadults.(Inmanyways,theHazanandShavermeasure assumedthistobetrue.)Subsequentresearchhasexploredthishypothesis inavarietyofways.Forexample,KellyBrennanandhercolleaguescollected anumberofstatements(e.g.,"Ibelievethatotherswillbethere formewhenIneedthem")andstudiedthewaythesestatements"hang together"statistically(Brennan,Clark,&Shaver,1998).Brennan's findingssuggestedthattherearetwofundamentaldimensionswithrespect toadultattachmentpatterns(seeFigure2).Onecriticalvariablehasbeenlabeledattachment-related anxiety.Peoplewhoscorehighonthisvariabletendtoworrywhether theirpartnerisavailable,responsive,attentive,etc.Peoplewhoscore onthelowendofthisvariablearemoresecureintheperceivedresponsiveness oftheirpartners.Theothercriticalvariableiscalledattachment-related avoidance.Peopleonthehighendofthisdimensionprefernotto relyonothersoropenuptoothers.Peopleonthelowendofthisdimension aremorecomfortablebeingintimatewithothersandaremoresecuredepending uponandhavingothersdependuponthem.Aprototypicalsecureadultislow onbothofthesedimensions. Brennan'sfindingsarecriticalbecauserecentanalysesofthestatisticalpatterning ofbehavioramonginfantsinthestrangesituationrevealtwofunctionally similardimensions:onethatcapturesvariabilityintheanxietyandresistance ofthechildandanotherthatcapturesvariabilityinthechild'swillingness tousetheparentasasafehavenforsupport(seeFraley&Spieker, 2003a,2003b).Functionally,thesedimensionsaresimilartothetwo-dimensions uncoveredamongadults,suggestingthatsimilarpatternsofattachment existatdifferentpointsinthelifespan. InlightofBrennan'sfindings,aswellastaxometricresearchpublished byFraleyandWaller(1998),mostresearcherscurrentlyconceptualize andmeasureindividualdifferencesinattachmentdimensionallyrather thancategorically.Thatis,itisassumedthatattachmentstylesarethingsthatvaryindegreeratherthankind. Themostpopularmeasuresofadultattachmentstyle areBrennan,Clark,andShaver's(1998)ECRandFraley,Waller,andBrennan's (2000)ECR-R--arevisedversionoftheECR.[Click heretotakeanon-linequizdesignedtodetermineyourattachmentstyle basedonthesetwodimensions.]Bothoftheseself-reportinstruments providecontinuousscoresonthetwodimensionsofattachment-related anxietyandavoidance.[Click heretolearnmoreaboutself-reportmeasuresofindividualdifferences inadultattachment.] DoAdultRomanticRelationships"Work"intheSameWaythatInfant-Caregiver RelationshipsWork? Thereisnowanincreasingamountofresearchthatsuggeststhatadultromantic relationshipsfunctioninwaysthataresimilartoinfant-caregiverrelationships, withsomenoteworthyexceptions,ofcourse.Naturalisticresearchonadults separatingfromtheirpartnersatanairportdemonstratedthatbehaviors indicativeofattachment-relatedprotestandcaregivingwereevident, andthattheregulationofthesebehaviorswasassociatedwithattachment style(Fraley&Shaver,1998).Forexample,whileseparatingcouples generallyshowedmoreattachmentbehaviorthannonseparatingcouples, highlyavoidantadultsshowedmuchlessattachmentbehaviorthanless avoidantadults.InthesectionsbelowIdiscusssomeoftheparallels thathavebeendiscoveredbetweenthewaythatinfant-caregiverrelationships andadultromanticrelationshipsfunction. Partnerselection Cross-culturalstudiessuggestthatthesecurepatternofattachmentin infancyisuniversallyconsideredthemostdesirablepatternbymothers (seevanIJzendoorn&Sagi,1999).Forobviousreasonsthereisnosimilar studyaskinginfantsiftheywouldpreferasecurity-inducingattachment figure.Adultsseekinglong-termrelationshipsidentifyresponsivecaregiving qualities,suchasattentiveness,warmth,andsensitivity,asmost"attractive" inpotentialdatingpartners(Zeifman&Hazan,1997).Despitetheattractiveness ofsecurequalities,however,notalladultsarepairedwithsecurepartners. Someevidencesuggeststhatpeopleendupinrelationshipswithpartners whoconfirmtheirexistingbeliefsaboutattachmentrelationships(Frazier etal.,1997). Securebaseandsafehavenbehavior Ininfancy,secureinfantstendtobethemostwelladjusted,inthesense thattheyarerelativelyresilient,theygetalongwiththeirpeers,and arewellliked.Similarkindsofpatternshaveemergedinresearchon adultattachment.Overall,secureadultstendtobemoresatisfiedin theirrelationshipsthaninsecureadults.Theirrelationshipsarecharacterized bygreaterlongevity,trust,commitment,andinterdependence(e.g.,Feeney, Noller,&Callan,1994),andtheyaremorelikelytouseromanticpartners asasecurebasefromwhichtoexploretheworld(e.g.,Fraley&Davis, 1997).Alargeproportionofresearchonadultattachmenthasbeendevoted touncoveringthebehavioralandpsychologicalmechanismsthatpromote securityandsecurebasebehaviorinadults.Therehavebeentwomajor discoveriesthusfar.Firstandinaccordancewithattachmenttheory, secureadultsaremorelikelythaninsecureadultstoseeksupportfrom theirpartnerswhendistressed.Furthermore,theyaremorelikelyto providesupporttotheirdistressedpartners(e.g.,Simpsonetal., 1992).Second,theattributionsthatinsecureindividualsmakeconcerning theirpartner'sbehaviorduringandfollowingrelationalconflictsexacerbate, ratherthanalleviate,theirinsecurities(e.g.,Simpsonetal.,1996). AvoidantAttachmentandDefenseMechanisms According toattachmenttheory,childrendifferinthekinds ofstrategiestheyusetoregulateattachment-relatedanxiety.Following aseparationandreunion,forexample,someinsecurechildrenapproach theirparents,butwithambivalenceandresistance,whereasotherswithdraw fromtheirparents,apparentlyminimizingattachment-relatedfeelings andbehavior.Oneofthebigquestionsinthestudyofinfantattachment iswhetherchildrenwhowithdrawfromtheirparents--avoidantchildren--are trulylessdistressedorwhethertheirdefensivebehaviorisacover-up fortheirtruefeelingsofvulnerability.Researchthathasmeasuredthe attentionalcapacityofchildren,heartrate,orstresshormonelevels suggeststhatavoidantchildrenaredistressedbytheseparationdespite thefactthattheycomeacrossinacool,defensivemanner. Recentresearchonadultattachmenthasrevealedsomeinterestingcomplexities concerningtherelationshipsbetweenavoidanceanddefense.Althoughsome avoidantadults,oftencalledfearfully-avoidantadults,arepoorly adjusteddespitetheirdefensivenature,others,oftencalleddismissing-avoidant adults,areabletousedefensivestrategiesinanadaptiveway.Forexample, inanexperimentaltaskinwhichadultswereinstructedtodiscusslosing theirpartner,FraleyandShaver(1997)foundthatdismissingindividuals (i.e.,individualswhoarehighonthedimensionofattachment-related avoidancebutlowonthedimensionofattachment-relatedanxiety)were justasphysiologicallydistressed(asassessedbyskinconductancemeasures) asotherindividuals.Wheninstructedtosuppresstheirthoughtsandfeelings, however,dismissingindividualswereabletodosoeffectively.Thatis, theycoulddeactivatetheirphysiologicalarousaltosomedegreeandminimize theattentiontheypaidtoattachment-relatedthoughts.Fearfully-avoidant individualswerenotassuccessfulinsuppressingtheiremotions. AreAttachmentPatternsStablefromInfancytoAdulthood? Perhapsthemostprovocativeandcontroversialimplicationofadultattachment theoryisthataperson'sattachmentstyleasanadultisshapedbyhis orherinteractionswithparentalattachmentfigures.Althoughtheidea thatearlyattachmentexperiencesmighthaveaninfluenceonattachment styleinromanticrelationshipsisrelativelyuncontroversial,hypotheses aboutthesourceanddegreeofoverlapbetweenthetwokinds ofattachmentorientationshavebeencontroversial. Thereareatleasttwoissuesinvolvedinconsideringthequestionofstability: (a)Howmuchsimilarityistherebetweenthesecuritypeopleexperience withdifferentpeopleintheirlives(e.g.,mothers,fathers,romantic partners)?and(b)Withrespecttoanyoneoftheserelationships,how stableissecurityovertime? Withrespecttothisfirstissue,itappearsthatthereisamodestdegree ofoverlapbetweenhowsecurepeoplefeelwiththeirmothers,forexample, andhowsecuretheyfeelwiththeirromanticpartners.Fraley,forexample, collectedself-reportmeasuresofone'scurrentattachmentstylewith asignificantparentalfigureandacurrentromanticpartnerandfound correlationsrangingbetweenapproximately.20to.50(i.e.,smallto moderate)betweenthetwokindsofattachmentrelationships.[Click heretotakeanon-linequizdesignedtoassessthesimilaritybetween yourattachmentstyleswithdifferentpeopleinyourlife.] Withrespecttothesecondissue,thestabilityofone'sattachmenttoone's parentsappearstobeequaltoacorrelationofabout.25to.39(Fraley, 2002).Thereisonlyonelongitudinalstudyofwhichweareawarethat assessedthelinkbetweensecurityatage1inthestrangesituationand securityofthesamepeople20yearslaterintheiradultromanticrelationships. Thisunpublishedstudyuncoveredacorrelationof.17betweenthesetwo variables(Steele,Waters,Crowell,&Treboux,1998). Theassociationbetweenearlyattachmentexperiencesandadultattachment styleshasalsobeenexaminedinretrospectivestudies.HazanandShaver (1987)foundthatadultswhoweresecureintheirromanticrelationships weremorelikelytorecalltheirchildhoodrelationshipswithparents asbeingaffectionate,caring,andaccepting(seealsoFeeney&Noller, 1990). Basedonthesekindsofstudies,itseemslikelythatattachmentstylesinthe child-parentdomainandattachmentstylesintheromanticrelationship domainareonlymoderatelyrelatedatbest.Whataretheimplications ofsuchfindingsforadultattachmenttheory?Accordingtosomewriters, themostimportantpropositionofthetheoryisthattheattachmentsystem, asystemoriginallyadaptedfortheecologyofinfancy,continuestoinfluence behavior,thought,andfeelinginadulthood(seeFraley&Shaver, 2000).Thispropositionmayholdregardlessofwhetherindividualdifferences inthewaythesystemisorganizedremainstableoveradecadeormore, andstableacrossdifferentkindsofintimaterelationships. Althoughthesocialandcognitivemechanismsinvokedbyattachmenttheoristsimply thatstabilityinattachmentstylemaybetheruleratherthantheexception, thesebasicmechanismscanpredicteitherlong-runcontinuityordiscontinuity, dependingontheprecisewaysinwhichtheyareconceptualized(Fraley, 2002).Fraley(2002)discussedtwomodelsofcontinuityderivedfromattachment theorythatmakedifferentpredictionsaboutlong-termcontinuityeven thoughtheywerederivedfromthesamebasictheoreticalprinciples.Each modelassumesthatindividualdifferencesinattachmentrepresentations areshapedbyvariationinexperienceswithcaregiversinearlychildhood, andthat,inturn,theseearlyrepresentationsshapethequalityofthe individual'ssubsequentattachmentexperiences.However,onemodelassumes thatexistingrepresentationsareupdatedandrevisedinlightofnew experiencessuchthatolderrepresentationsareeventually"overwritten." Mathematicalanalysesrevealedthatthismodelpredictsthatthelong-term stabilityofindividualdifferenceswillapproachzero.Thesecondmodel issimilartothefirst,butmakestheadditionalassumptionthatrepresentational modelsdevelopedinthefirstyearoflifearepreserved(i.e.,theyare notoverwritten)andcontinuetoinfluencerelationalbehaviorthroughout thelifecourse.Analysesofthismodelrevealedthatlong-termstability canapproachanon-zerolimitingvalue.Theimportantpointhereisthat theprinciplesofattachmenttheorycanbeusedtoderivedevelopmental modelsthatmakestrikinglydifferentpredictionsaboutthelong-term stabilityofindividualdifferences.Inlightofthisfinding,theexistence oflong-termstabilityofindividualdifferencesshouldbeconsidered anempiricalquestionratherthananassumptionofthetheory. Outstanding QuestionsandFutureDirectionsforResearchonAdultAttachment Thereareanumberofquestionsthatcurrentandfutureresearchonattachment needstoaddress.Forexample,itisprobablythecasethat,whilesome romanticrelationshipsaregenuineattachmentrelationships,othersare not.Itwillbenecessaryforfutureresearcherstofindwaystobetter determinewhetherarelationshipisactuallyservingattachment-related functions.Second,althoughitisclearwhyattachmentbehaviormayserve animportantevolutionaryfunctionininfancy,itisnotclearwhether attachmentservesanimportantevolutionaryfunctionamongadults.Third, westilldon'thaveastrongunderstandingoftheprecisefactorsthat maychangeaperson'sattachmentstyle.Intheinterestofimprovingpeople's lives,itwillbenecessarytolearnmoreaboutthefactorsthatpromote attachmentsecurityandrelationalwell-being. ©2018R.ChrisFraley Tolearnmoreaboutattachmenttheoryandresearch,pleasecheckoutthebookOmri,Gery,andIwrote.