0066-4308/96/0201-0431$08.00 431

Annu.Rev. Psychol. 1996.47:431–59Copyright© 1996by AnnualReviewsInc. All rightsreserved

ORIGINSAND EARLYDEVELOPMENT OF PERCEPTION,ACTION, AND REPRESENTATION

BennettI. Bertenthal

Department of Psychology, University of Virginia, Charlottesvil le, Virginia22903

KEY WORDS: perceptual development, motor development, conceptual development, cognitivedevelopment, sensorimotor development, perception-action, representation

ABSTRACT

Research relevant to the origins and early development of two functionallydissociable perceptual systems is summarized. Onesystem is concerned withtheperceptual control andguidanceof actions,theotherwith theperceptionandrecognition of objects and events. Perceptually controlled actionsfunction inreal timeand aremodularly organized. Infantsperceivewherethey areandwhatthey aredoing. By contrast, research on object recognition suggests that evenyoung infants representsomeof the defining featuresand physical constraintsthat specify the identity andcontinuity of objects. Diff erent factors contributeto developmental changes within the two systems; it is difficult to generalizefrom oneresponsesystem to another; and neitherperception,action,nor repre-sentation quali fies asontogenetically privileged. All threeprocesses developfrom birth asa function of intrinsic processingconstraintsandexperience.

CONTENTSINTRODUCTION..................................................................................................................... 432PERCEPTUAL CONTROL OF ACTIONS............................................................................. 434

Newborn Actions AreSpatially Coordinated...................................................................... 434Reciprocity BetweenActionand Perception ....................................................................... 436Perception Is Prospective.................................................................................................... 438Perception Is Multimodal .................................................................................................... 440Perception Is ContextSpecific............................................................................................. 443

OBJECT PERCEPTION AND RECOGNITION..................................................................... 445Newborn Recognitionof Objects......................................................................................... 446

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

Implicit Knowledgeof Objects............................................................................................ 447Explicit Knowledgeof Objects............................................................................................ 450ObjectRecognition Is Domain Specific............................................................................... 451

CONCLUSIONS....................................................................................................................... 452

INTRODUCTION

Recentfindings on the perceptual,motor, and conceptualcompetenciesofyounginfantschallenge long-heldbeliefs aboutearly development. Accordingto mostclassicaldevelopmental theories(Baldwin 1906,Bruner1973,Piaget1952),newbornsareendowedwith only averysimplerepertoireof sensorimo-tor behaviorsthat aregraduallyintegratedand internalized.The capacityforrepresentationandconceptualizationis presumedto emergefrom thisdevelop-mental process.It is not clear, however,that this position is still tenable.Indeed, thecapacity for representationmay be availableat birth or soonthereafter(Carey& Gelman1991;Eimas1994;Leslie1988;Karmiloff-Smith1992;Mandler1988,1992;Mounoud1993;Spelke1994).

This new view requiresa reconceptualizationof the developmental rela-tionsamongperception,action,andrepresentation.Most modelsof perceptualprocessingnow suggestthat different sensoryinputsconvergeinto a unifiedrepresentationthatprecedesboth thoughtandaction(Marr 1982,Ungerleider& Mishkin 1982).This monolithic view of perceptionsuggeststhatassessingwhat is perceivedis independentof whethertheresponsemeasureis basedonan action or a perceptualjudgment. From a developmental perspective,thisview impliesthatevidencefor representationof objectsshouldbemanifestedby thoughtandactionat the sameage.It is now apparent,however,that thisview is obsolete.Theparadigmatic casefor thisassertion istheconflicting andcontradictoryevidenceon object permanence. Most infants do not reachfor anobject hidden by an occluderuntil they are 8–9 months old (Butterworth1982).This failure to recoverthehiddenobjectis interpretedasevidencethatinfants do not think about objectsthat are not perceptuallypresent(Piaget1954). Nevertheless, infants asyoung as 3monthsold showevidencethattheyrepresentthecontinuity andsolidity of objects—atleastwhenthetestof objectpermanencedemandsnothingmorethana visual fixation by the infant (Bail-largeon1987,Diamond1991,Spelkeet al 1992).In orderto eliminate theseconfusionsandcontradictionsin theliteraturewemustadoptanewframeworkfor understandingthe early perceptual,motor, andcognitivedevelopmentofinfants.

An intriguing possibility is suggestedby Goodale& Milner (1992),whoproposethat the visual systemis divided into two functionally dissociablepathways.One is concernedprimarily with the perceptualcontrol andguid-anceof actions,the other primarily with the perceptionand recognitionofobjectsandevents.This dichotomyresemblestheoneadvancedpreviouslyby

432 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

Ungerleider& Mishkin (1982)thatthevisualbrainconsistsof two systems,a“what” anda “where” system. It is neverthelessdifferent,becausetheempha-sis is not on the input side of visual processingbut the output side or theresponseselicitedby thevisual information. “What” vs “how,” not “what” vs“where,” bestcaptures thisfunctionaldissociation.

This functionaldissociation betweencontrolandrecognitionrepresentsthedeparturepoint for thecurrentreview.Althoughvirtually all theevidenceforthis dissociation is basedon neurophysiological findings with monkeysandneuropsychological findings with humanadults,it is plausiblethat this disso-ciation is presentearly in development,given that the proposeddivision oflabor mapsonto different neuralpathwaysthat areall developingwithin thefirst year(Johnson1990).Onemayspeculatethat this functionaldissociationis notlimi ted tovisualprocessing,butextendsto other modalities as well.

A brief review of theprocessingdifferencesbetweenthesetwo perceptualsystemswill clarify furtherwhy behaviorsmediatedby theperceptualcontrolsystemare functionally dissociablefrom behaviorsmediatedby the objectrecognitionsystem.

1. Object recognitionincludesprocessesthat makecontactwith informa-tion perceivedat someprior time and storedin somerepresentationalform.Successfulrecognitiondependson bothhow thevisualsceneis parsedandontherepresentationalformatof thestoredinformation. By contrast,thepercep-tion and control of actions isdirected toward presentinformation and, ifanything,includesa prospectiveview toward information in order to offsetdelays producedby neuraltransmission and theinertiaof bodysegments.

2. The seconddifferenceis relatedto the first andinvolvesthe coordinatesystemfor perceivingobjects.Perceptionof objectsinvolvesanallocentricorworld-basedcoordinatesystemsuchthat displacements areseenrelativeto astableor constantworld. By contrast,acting on an object requiresthat theobject be referencedegocentrically—i.e. relative to the effector systemin-volvedin theaction(Paillard 1991).

3. A third difference concerns thecodingand preservationof modality-spe-cific information. Objectsare typically specifiedby multimodal sourcesofinput, but the informationis storedin a modality-specificformat.This provi-sion is necessaryto explain how observers recognize specificfeatures ofobjects,suchasits color or pitch, aswell asthecovariationbetweenfeaturesspecifiedboth within and betweenmodalities. By contrast,perceptualinfor-mationis representedby theactionsystemin anamodalformatcomprisedofbody-scaledinformation. This format transformsall sensoryinputs into theappropriatemusclesynergiesnecessaryfor producingcoordinatedactionsinresponseto localconditions. Itis thusnotessentialto representthemodality ofthesensoryinput,becausethefunctionof this information is thesameregard-less ofits source.

PERCEPTION, ACTION, ANDREPRESENTATION 433

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

4. The last processingdifferenceconcernsthe role of awarenessin theperceptionof information. Objectsthat arenot consciouslyprocessedareneitherrecognizednor storedfor future recall. Recognitionrequiresthat ob-serversdirect their attentiontowardselectedobjectsandknow whentheyareperceivingthe relevant information. Conversely,information necessaryfordetectingself-motionandcontrolling otheractionsoperateswithout any nec-essaryawarenessby the observer.For example,patientswith brain lesionswho areunableto recognizeobjectsareneverthelessable to reachfor theseobjectsandto anticipatetheir sizeandshapecorrectlywhile reaching(Goo-dale et al1991,Weiskrantzetal 1974).

In the remainderof this chapter, Isummarize recentresearchrelevant tothedevelopmentof thesetwo perceptualsystems, noting that different factorscontributeto developmental changeswithin thetwo systems, thatit is difficultto generalizefrom oneresponse system to another,andthatneitherperception,action, nor representationqualifies as ontogenetically privileged. All threeprocessesdevelopfrom birth asa function of intrinsic processingconstraintsand experience.

PERCEPTUAL CONTROLOFACTIONS

All spatially coordinatedbehaviors,such as visual tracking, reaching,andsitting, require that perceptual informationand action are coupled. Inthewordsof JamesGibson(1979),“We mustperceivein orderto move,but wemustalsomovein orderto perceive” (p.223).Perceptualinformationrelevantto theregulationof movementsincludesspatio-temporal patterns ofoptic flowatamovingeye,hapticpatternsof joint, muscle,andskindeformations,andsoon. All of this information changesin ways that are lawfully relatedto thepropertiesof theenvironmentandtheactionitself. For example,reachingforan object is guided by perceptualinformation that changesas the reachisexecuted.Theseperceptualchangesmodulatethe effectorsto insurethat thereach is successful(Jeannerod1994). Similarly, perceptualinformation isnecessaryto maintainposturalequilibrium, but againtheinformation changesasthepostureis adjusted inresponseto thatinformation(Howard1986).Fromthis perspectiveit is arbitraryandmisleadingto conceptualizeperceptionandaction as independentprocesses.It is more parsimonious to view thesetwoprocessesas oppositepolesof a functional unit or action system, along thelinessuggestedby Reed(1982,1989).

NewbornActions Are Spatially Coordinated

When are perceptionand action first coupled?Until recently,the answertothis question wasdominatedby Piaget’s (1952)view of sensorimotor develop-ment.He assertedthatperceptionsandactionsareinitially independentproc-

434 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

essesthat arecoordinatedgraduallywith experience.The implicationof thisproposalis that the early behaviorof the neonateis essentiallyrandomandinsensitive to contextualinformation.Recentresearchsuggeststhat somere-thinking of thisextremepositionis necessary.

During thepastdecade,researchershaveobservedthatnewborninfantsarecapableof performingmanyactionsthatareregulatedby perceptualinforma-tion. For example,newborn infantsorientto sound (Clifton et al1981, Muir &Field1979,Zelazoetal 1984),scandifferently in differentstimulusconditions(Haith 1980), visually track moving targets(Bloch & Carchon1992, Kre-menitzer etal 1979), increasethefrequencyof hand-to-mouth contactsfollow-ing oral delivery of a sucrosesolution (Blasset al 1989,Rochatet al 1988),andshowhandextensionstowarda visible moving object(Trevarthen1984,von Hofsten1982). Of course,thesebehaviorsare fragile and inconsistent,which explains why theywereoverlooked for quite sometime. Subtlechangesin postureor stimulus parametersareoften sufficient to disruptthesecoordi-natedbehaviors.For example,Roucouxet al (1983)haveshownthatneonatessometimes experiencedifficulty in tracking objectsvisually becauseof theinstability of their trunks,which do not yetmove independently of theirheads.

It thus appearsthat newbornsenter the world preparedto perceptuallyregulate actionsthatare essentialto the survivaland adaptation of theneonate.An intriguing suggestionis that behaviorspracticedin the womb show anadvantageat birth (von Hofsten1993).For example,proprioceptive guidanceof the hand to the mouth is readily observedin neonates(Butterworth &Hopkins 1988, Rochat et al 1988). Furthermore,Butterworth & Hopkins(1988)reportthat themouthis morelikely to remainopenduringarmmove-mentswhenthehandgoesdirectly to themouthratherthanfirst touching otherportionsof theface.Soonafterbirth, this responseis observedmorefrequentlyprior to feeding than following feeding (Lew & Butterworth 1995). Takentogether,thesefindings suggestsignificantspecificity in the coordinationofhand andmouthat birth.

Theevidencefor thecouplingof perceptionandactionsat birth shouldnotbe misconstruedassuggesting that thesesystemsarefully developedor thatnewcouplings will not emerge.Contemporary theoristsemphasizethatdevel-opment involvesaconfluence offactorsthat include neural andbiomechanicalchangesaswell asenvironmentalandtaskfactors(Newell 1986,Savelsbergh& vanderKamp1993,Thelen1995).Practiceandexperiencewith a specificaction systemcontributeto its development.Someof the bestexamplesin-volve behaviorstraditionally viewedasmotor skills, suchaspostureandgait(Sveistrup &Woollacott1995,Thelen &Ulrich 1991).

One reasonthat practiceand experienceare not sufficient to capturetheprocessof developmentalchangeis that the infant is also changingin bodyproportionsandneuralconnectivity. For example,Banks(1988) reportsthat

PERCEPTION, ACTION, ANDREPRESENTATION 435

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

theopticalcomponentsof theeyearestill growingatbirth, thephotoreceptorswill matureandmigrateduring thefirst few months,andthedendriticarbori-zationof the centralvisual pathwayswill continueto developfor sometime.Thesechangesinform us that the resolution and projectivestructureof thevisual image will improve with development. Likewise, the perceptionofspatiallayoutandof therelativedepthsanddistancesof objectswill improvewith development(Yonas& Owsley1987).

It is perhapsevenmoreimportant to point out thatoculomotor functioningwill showsignificantimprovementduringearlydevelopment.Saccadiclocal-ization of stationaryand moving targetsinvolves a direct mappingbetweenretinal location and neuromuscular stimulation of the relevanteye muscles.Initially, this localization processis impreciseandinvolvesmultiple saccadesbefore thetargetis foveated(Aslin & Salapatek1975).No doubtsomeexperi-enceis necessaryto learnthe preciserelationbetweenthe neuralpulsedura-tion andsaccademagnitude necessaryfor rotatingtheeyeto thecorrectposi-tion. It is still somewhatsurprisingthat the calibrationprocessrequiresoverfour monthsto complete,especiallywhenestimatessuggestthat infantsmakebetween3 milli on and6 milli on eyemovementsby 3.5 monthsof age(Haithet al 1988).Oneespeciallyintriguing hypothesisaboutthis lengthyprocessisthat themappingof retinal locusontoanoculomotorcommandis constrainedby thechangingdistribution of photoreceptorsin theretina(Aslin 1988).It isthusnecessaryfor theinfant to adaptcontinually to thischangingsensorimotorrelationduringearly development.

This lastexampleillustratesespeciallywell thatactionsarespatiallycoor-dinatedfrom birth but becomebetter tunedor coordinatedas a function ofneuraldevelopmentandexperience.Although the processby which percep-tuomotor behaviorsdevelopis rarely investigated,researchershavebeguntorecognizethat motor skills arenot only productsof this developmentalproc-ess, butare intimately involvedin theprocess itself.

ReciprocityBetween Action andPerception

Perceptualcontrolof behaviordependson thedetectionof therelevantpercep-tual informationaswell asthefunctionality of theactionsavailableto infants.As simple actionssuch as visual tracking or sucking are practicedand re-peated,they becomebettercoordinatedandcontrolled,andperceptualinfor-mationis detectedwith increasingspecificity.An excellentexampleof thesechangesis revealedby researchon the minimum audibleanglenecessaryfordetectionof thedirectionof a soundsource. In this task infants areexpectedtoturn their headsto theright or left of midline if theyarecapableof localizingthe sound(Ashmeadet al 1987,Morrongiello 1988).Ashmeadet al (1991)summarizethe data from a numberof studies to show that the minimumdetectabledifferencedecreasesrapidly between8 and 24 weeksof ageand

436 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

then continuesto decreasemore gradually through 80 weeksof age. It isnoteworthythat themostrapid improvementoccursduringandjust followingthe timethatinfantsaredevelopingindependent control of their heads(Bayley1969).

The precedingexampleis an excellentillustrationof the reciprocity thatexistsbetweenactionandperceptionin development.As newactionsbecomeavailable,new opportunities for exploring the fit betweenthe self and theenvironmentemerge(Adolphetal 1993a).Anotherexampleof thisproposalisassociatedwith the developmentof crawling or self-producedlocomotion.Perceptualguidanceis necesaryto assuremovementwithout collisionson asafe and sturdysurface of support(Gibson& Schmuckler 1989).

Bertenthal& Campos (1990) report that perceptualsensitivity to objectsandsurfaceschangessignificantly following someexperiencewith crawling.For example,Camposet al (1992) report a seriesof studiesshowing thatprecrawlinginfantsshowno evidenceof fear (asindexedby heartrateaccel-eration) when lowered onto the deepside of a visual cliff (simulating anapparentdrop-off in height), whereas crawlinginfants show a significantdegreeof fear. Fear is also shownby precrawlinginfants if they are givensufficient experiencewith self-locomotion in baby-walkers.Apparently,suchexperiencewith perceptualguidanceof self-locomotion changesinfants’ per-ceptualappreciationof anapparentcliff. Precrawlinginfantsdo not showfearof heightsnot becausethey cannotperceive depth (Yonas& Owsley1987)butbecausetheydonotyetneedto coordinatetheperceptionof surfaceswith theirdirectionof heading.

Similar findings arereportedby Gibsonet al (1987),who testedcrawlingand newly walking infants ontheir mode of locomotion on two surfacesvarying in rigidity (plywoodvs waterbed).Infantscapableof upright locomo-tion differentially exploredthe two surfacesand choseto walk only on therigid surface.Crawling infants did not show different behavioron the twosurfaces.In more recent research,Adolph et al (1993b) report that newlywalking infants,but not crawling infants,differentiatebetweeninclined anddeclinedsurfacesby choosinga morestableposture,suchassitting or crawl-ing backwards, when traversingthedown-slopingsurface.

Anothercompelling exampleof thereciprocitybetweenperceptionandthedevelopmentof newactionsis offeredby Bushnell& Boudreau(1993).Adultsdetect many different propertiesof objects,such as size,texture, weight,hardness,and temperature,from haptic explorations (Lederman& Klatzky1987).Someof theseproperties,suchassizeandtemperature,demandmini-mal control of thehandandfingers,whereasotherproperties,suchasweightandshape,requiremuchgreatercontrol. Bushnell& Boudreaureviewedtheagesat which infants first discriminatedifferent object propertiesand con-cludedthat thesequencecorrespondsto developmentalchangesin thecontrol

PERCEPTION, ACTION, ANDREPRESENTATION 437

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

of the handandfingers.For example,infantsdetectsizewithin the first fewmonths,but texture,temperature,andhardnessarenot detecteduntil around6monthsof age,and weight and shapeare not detecteduntil even later. Al -thoughtheevidencefor thisclaim is still incomplete, alternativeinterpetations(e.g.that theseobservationsresult from differential exposureto different ob-ject properties)are unlikely,giventheecologyof the infant’s environment.

New perceptionsand new actionsare relatedthrougha dynamicprocessinvolving the selectionof new behaviorsin responseto new sourcesof vari-ability in theorganismandin theenvironment(Bertenthalet al 1994,Manoel& Connolly 1995;Thelen1989,1995).Considerthecoordination of thelimbsduringforwardproneprogression.Most infantscrawl with their abdomensonthegroundbeforecrawlingon hands-and-knees.Oncetheydevelopsufficientstrengthto supportthemselves on hands-and-kneesthey briefly show manydifferentpatternsof interlimb coordinationbeforeconvergingon a patternofmoving diagonally oppositelimbs simultaneously(Freedland& Bertenthal1994). The selectionof this specific patternis a function of perceivingtheoptimal coordinativestructureto insurebalancewhile minimizing theexpen-ditureof energy.

Another examplerelatedto the developmentof crawling experiencein-volves the spatial coding of a hidden object. Numerousstudiesreport thatcrawling infantsshowimproved localizationof objectsfollowing a displace-ment ofthe infantor theobject (Bremner& Bryant 1985,Horobin& Acredolo1986). Precrawling infants tend to code the location of an object with abody-centeredframeof reference,presumablybecausethis codingis initiallynecessaryfor successfulorientation tothe object. With theemergenceofcrawling,infantsshowa transitional periodduringwhich theirresponsesvary.Eventually,they learnto updatetheir initial spatialcodingin responseto theperceived displacement(Bai & Bertenthal1992).

Thelenandcolleaguesoffer additional examplesof howthedevelopmentofnew actions,suchas infant stepping,emergefollowing periodsof increasedvariability (Thelen& Ulrich 1991,Ulrich et al 1991).Overall, this researchsuggeststhatperceptuomotordevelopmentis anemergentprocessin which agoal-directedorganismseeks stableoutcomesto specifictasks.

Perception Is Prospective

Our actions,like thoseof all animals,are coupledto the spatial layout anddemandperceptual guidanceand control(Lee 1993).In locomotion, for exam-ple, we must make contactwith some surfaceswhile avoiding others. Ingeneral,it is necessaryto controlactionsprospectivelyandnot retrospectively(i.e. following feedbackfrom the action) in order to insuresmoothandsafemovement(von Hofsten1993).The inertia of the limbs andthe time lagsofneuralconductiondemandsomeanticipationof future actions(Haith 1994,

438 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

von Hofsten 1993). Information neededfor the specificationof upcomingeventsis availablein theoptic andacousticarraysandis usedfor controllingfutureactions.As adults,we readilyappreciatethetemporalcomponentin thecontrolof actions.For example,we know thatit is necessaryto bein therightplaceat theright time to catcha ball, meeta person,or give a lecture.Recentfindings in the literaturerevealsomeremarkableexamplesof future-orientedbehavior byinfants.

Oneof the earliest examples of prospective behavior isseen in thedevelop-mentof smooth visual pursuit of moving targets.In order to track an objectsmoothly it is necessaryto anticipateits futureposition: Theprogramming ofeye movements takestime. Shea& Aslin (1990) presentedinfants with 2°white squaresthat moved at a rangeof fixed velocities between3 and 12degreespersecond.Theyreportedthatthepursuitsystemis clearlyfunctionalby 7 weeksof ageandsuggested thatslowerspeedsandlargertargetscouldbedetectedat youngerages.This suggestion is consistent with the findings ofotherinvestigators(Bloch & Carchon1992,Kremenitzeret al 1979,Roucouxet al 1983), who report brief segmentsof smooth pursuit in newbornsandone-month-oldinfants.The successof pursuit trackingat suchyoungagesisespeciallyimpressivewhenit is recognizedthat eyemovementsmayalsohaveto compensatefor unrelatedheadmovements.In studieswherethe headwasunrestrainedduring testing, young infants trackeda moving target with acombination of headand eye movements(Daniel & Lee 1990, Regalet al1983,Roucouxet al1983).

Anothereyemovement paradigmthat showsearly evidenceof future-ori-entedbehavioris thevisualexpectationparadigmpioneeredby Haith (1993):Infantsobservesmall picturesthat alternatebetweenthe left andright of thecenterof thescreen.After a few repetitions,2- and3-month-oldinfantsbeginto show anticipatoryfixations to the location of the appearanceof the nextpicture,evenwhenthe timing andlocationof thealternationpatternsbecomemore complex (Canfield & Haith 1991, Wentworth & Haith 1992). It thusappearsthat evenvery young infants can learn quickly how to control thelocationof their fixations in orderto explorethechangingpatternof informa-tion availablein theirvisualworld.

Oneof the most remarkableexamplesof prospectivebehaviorby infantsinvolves their reachingfor moving objects.Von Hofsten(1983) studiedin-fants’ reachingfor stationaryandmoving objectsandreportedthattheybeganto contactobjectsin both conditionsat the sameage.By 18 weeksof age,infantscouldcatchanobjectmoving at30cm/s,andby 8 monthsinfantscouldcatchobjects moving at125cm/s. Inthis study,areachdid notcorrespondto asimplereflex response.Theobjectswerecontactedat variouslocationsalongtheir trajectory,and the aiming and timing errors were quite small. Theseobservationssuggestthattheinfantswere sucessfullycontrollingtheir reaches

PERCEPTION, ACTION, ANDREPRESENTATION 439

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

by extrapolating fromthe trajectoriesof the objectsand modulating theirmotor responses.This modulation is a complexprocessinvolving thepercep-tion of both passiveand active forcesthat vary from one reachto the next(Zernicke& Schneider1993).In a relatedstudy,Robinet al (1995)observed5- and 7.5-month-old infants reachingfor a horizontally moving object. In-fantsusuallyreachwith their ipsilateralhandfor a stationaryobject(Perris&Clifton 1988).In this study,infantsshiftedto reachingwith their contralateralhand when the object was moving, which increasedthe time available tointerceptthetarget.Kinematicmeasures,suchasvelocity anddurationof armmovements,convergedwith thosereportedby von Hofsten to suggestthatinfantsaimedtheir reachesin anticipation of thefuturepositionof themovingobject.

Prospectivebehaviorsarealsoevidencedwhen infants learnto posturallycompensatefor a lossof balance(Bertenthalet al 1995,von Hofsten1993),lean toward objectsthat are out of reach(McKenzie et al 1993, Rochat&Goubet1995,Yonas& Hartman1993),anticipatethesize,shape,andorienta-tion of objectsthat they are attempting to grasp(Lockmanet al 1984, vonHofsten & Ronnqvist 1988), and guide their locomotion aroundobstacles(Gibson& Schmuckler1989,Schmuckler& Gibson1989).One interpretationof all of thesefindings is that infantsbecomesuccessfulacrosstasksastheydevelopthecapacityto representfutureevents.Theproblemwith this cogni-tive interpretationis that it ignoresdevelopmentaldifferencesattributable tothecoordination of differentmotorskills. A moreparsimonious interpretationis that the prospectivebehaviordisplayedby infants is not contingenton acentralrepresentationbut ratheremergespiecemealfrom the specificexperi-encesthat infantsencounterthrough their actions. It isthus thedynamicinterplaybetweenactionsandoutcomesthat fostersthe developmentof pro-spectivecontrol.As infantsexperiencenewtasksthatdemandgreatercontrol,the precise timing of their actionswill improve.

Perception Is Multimodal

In mostsituations,theperceptualinformation availablefor controlling actionsis multimodal. Consider, for example,the control of postureduring inde-pendentstance.Postureis specifiedby proprioceptive, vestibular, andvisualflow information (Lishman& Lee 1973). Itis agoal-directed behavior, even ifit is not consciouslycontrolled (Howard 1986). The individual’s goal is topositiontheheadandbodyrelativeto gravityandthesurfaceof support.Whena perturbationof this position is sensed,a posturalcompensation is initiated.One reasonthat this perceptuomotorresponseis so successfulis that it isspecified bymultiple andredundant sensoryinputs. Thisredundancyincreasesthe likelihood of detectionby evenyoung infants who show rapid develop-mentof thisperceptual-motorresponse.

440 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

Much of theresearchon thedevelopmentof posturalcontrol tests infants ina “moving room.” In this paradigm,the infant sits or standson a stationaryfloor while thewalls andceilingmoveforward and backward. Thismovementproducesvisual informationcongruentwith the headmoving in the oppositedirection.If theopticalflow is perceivedasspecifyingself-motion(asopposedto objectmotion),thentheinfantwill showaposturalcompensationthatvarieswith age and experience (Bertenthal& Rose1995).

Lee & Aronson(1974)werethe first to showthat independentlystandinginfants compensateposturally in a directionally appropriatemannerin re-sponseto suchvisualflow information. Otherssubsequently demonstratedthatopticalflow information restrictedto theperipheralportions of thevisualfieldwassufficientto induceposturalcompensations(Bertenthal& Bai 1989,Stof-fregenet al 1987).Additional evidencefor the couplingbetweenvision andposturewas reportedby Butterworth& Hicks (1977) and Bertenthal& Bai(1989),who showedthat infantswho could sit independently alsorespondedwith posturalcompensationsof their trunk whentestedin themovingroom.Itappearsthat an evenearlier form of this coupling is presentat birth. Jouen(1990) reportedthat newborninfants show posturalcompensations of theirheadwhenstimulatedby an optical flow patternof blinking lights locatedinthe peripheryof the visualfield.

In themoving room paradigmposturalcompensations areinducedby vis-ual information, but it would be misleading to suggestthat the responseiscontrolledexclusivelyby visual inputs.Posturalswayis specifiedvisually byoptical flow information, but it is alsospecifiedby moreproximalstimulationfrom muscles,joints,andtheinnerear.Studiesinvolving mechanicalperturba-tions of a platform reveal that somatosensoryand vestibular systemsalsoinduceposturalresponses(Hirschfeld& Forrsberg1994,Woollacott & Sveis-trup 1994). Developmentalchangesin posturalcontrol involve learning toregulatethe amountof force to compensatefor the perceiveddisplacement.Bertenthaletal (1995)studiedinfantsduringtheperiodwhentheyarelearningto sit without support.They reportthat the compensatoryforcesnecessarytomaintainposturalequilibrium becomemorepreciselyscaledto the perceiveddisplacementof the trunk between5 and9 months of age.Performanceim-provesnot only becausecompensatorypostural responsesare more finelymodulated,but also becausethe perceiveddisplacementsare detectedmorerapidlyand precisely.

How do infantslearnthatdifferentsourcesof sensoryinput areequivalentand convergeon the samemotor response?If sensoryinformation is firstrepresentedin a modality-specific format, thenit would appearnecessaryforinfantsto learnthe relationsbetweendifferent sensoryinputs, suchasvisionandtouch.Indeed,this form of learningto coordinatedifferent modalities is

PERCEPTION, ACTION, ANDREPRESENTATION 441

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

thebedrockof Piaget’s (1954)theoryof sensorimotordevelopment.Yet cur-rent evidence suggeststhatthis form of associative learning isunnecessary.

An alternativeproposalis that all sensoryinputs relatedto self-producedactionsare representedin a common amodalformatthatmapsdirectlyontoanappropriatepatternof muscleactivations(Lee 1993, Warren1990). Suchacommonformatis thoughtto insure that all sourcesof sensory informationaretransformedinto the samebody-scaledinformationnecessaryfor modulatingthe motor response synergies involvedin the performanceof coordinatedmovements(Bertenthal& Rose1995,Savelsbergh& van der Kamp 1993).For example,when an individual detectsthat supportis perturbed,it is notimportant to determinewhich sensoryinput channelspecified this loss ofbalance.The goal is simply to restoreequilibrium, and this involvesscalingthe compensatoryforcesto theperceived displacement.

Anotherexampleof theequivalenceof differentsensoryinputsfor control-ling actionsinvolvesthedevelopmentof reaching.Historically, theprevailingview hasbeenthatreachingis initially visually guided(Bushnell1985,Piaget1952, White et al 1964), but more recentstudies show that infants reachreadily and accuratelyin the dark for soundingas well as luminousobjects(Clifton et al 1991,Clifton et al 1994,Stacket al 1989).In onestudy(Cliftonet al 1993),infantsbetween6 and25 weeksof ageweretestedlongitudinallyto determinewhetherthey requiredsight of their handswhen beginningtoreachfor, contact,andgraspobjects.Eachsessionincludedtrials of objectspresentedin the light andtrials of glowing andsoundingobjectspresentedincompletedarkness.Theresultsrevealedlittl e variationasa functionof experi-mentalcondition. Overall, infantsfirst contactedtheobjectin bothconditionsat comparableages(light 12.3weeks;dark11.9weeks),andtheyfirst graspedthe object in the light at 16.0 weeksand in the dark at 14.7 weeks.Infantscouldnot seetheir handsor armsin thedark;their earlysuccessin contactingthe glowing and soundingobjectsindicatesthat proprioceptive informationwas sufficient to guide reaching.It thus appearsthat no single sourceofsensoryinformation(e.g.visual,proprioceptive,or vestibular)is privilegedininitially guidingactions.

Someof themostdramaticevidencefor theamodalrepresentationof sen-sory inputs is revealedby studiesof neonatalimitation. Meltzoff & Moore(1983,1989,1994) and many others(seeAnisfeld 1991 for a review) haveshownconvincingly thatnewborninfantsimitate specificfacial gestures(e.g.mouthopening)producedby anadultmodel.Suchgesturescannotbevisuallyguided becausethey involve movementsthat the infant cannotsee—i.e.move-ment of the infant’s own face. The correspondencebetweenthe perceivedfacial gestureand action by the newbornsuggeststhat visual informationconcerningthe adult’s face is perceivedamodallyin a format that mapsdi-rectly onto the appropriatemuscle activation patterns(Meltzoff & Moore

442 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

1994).Currently,thisclaimremainsfairly controversialbecauseit necessitatesthedetectionof a correspondencebetweenthevisualperceptionof theactionsof a modelandthe proprioceptiveperceptionof one’s own actions.No com-pleteexplanationfor thismatchinghasyetbeenpresented,althoughthespatio-temporalcodingof themodel’s gesturesmay providemorespecificinforma-tion thanassumedby mostinvestigators (Bertenthal& Pinto 1993).

In sum, actionsare most often guided by multimodal information. Thisredundancymay help to explainwhy the perceptionandcontrol of adaptivebehaviors,such as reaching,sitting, and walking, developrapidly oncethenecessarymusclesynergiesareavailable.Moreover,theavailability of multi-ple sourcesof informationfor modulating actionsin responseto local condi-tions increases theconsistency,stability, and flexibility of anyadaptivebehav-ior.

Perception Is ContextSpecific

Recenttheoreticalandempiricaladvancesin the studyof motor control andcoordinationhighlight thatactionsarea productof multiple factorsincludingphysical,physiological, and energeticcomponents(Freedland& Bertenthal1994,Goldfield 1993,Manoel& Connelly 1995,Thelen1995,Turvey1990).A principal implication of this view is thatthesameactionswill not necessar-ily be observedin different contexts.For example,Grenier(1981) reportedthat reachingmovements by newbornsaremuchbettercoordinatedwhenthehead is stabilized than when itis unsupported. This finding isespeciallyimportantbecauseit emphasizesthat contextinvolvesnot only externalfac-tors, but also the ways in which different body segmentsareconfiguredandinteract.Zernicke &Schneider(1993)show explicitly that the forcesresponsi-ble for moving limb segmentsarea function both of activeforces(producedby musclecontractions)andof passiveforces(correspondingto gravity andtheinertial forcesfrom theothermovingbodyparts).Thus,it is apparentthatthereexistmultiple constraintsthatdeterminewhetheror not anactionwill beperformed ina specific context.

This contextualspecificityis illustratedby thefinding thatnewborninfantsperformalternatingstep-likemovementswhenhelduprightwith their feetonasupportsurface.Within a few months, thesemovements disappear,presum-ably becausethey are inhibited by the developmentof higher-levelcorticalstructures(Zelazo 1984). Curiously, however,similar movements arestillobservedin babieslying on their stomachsor backs(Thelen& Fisher1982).Thesefindingsmay beexplained by asimplebiomechanical calculationshow-ing that moreenergyis neededto lift a leg to full flexion while upright thanwhile supine.Althoughgravity is a constantforce in theenvironment,it onlybecomesa constraintafter thenewbornperiodwheninfantsbeginexperienc-ing rapidweightgainsthatdecreasetheratio of muscleto subcutaneousfat in

PERCEPTION, ACTION, ANDREPRESENTATION 443

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

thelegs.Experimentalmanipulationsthatchangedtheweightof theleg or theresistanceof the leg to flexion (e.g.submerginginfants in torso-deepwater)showedthat thepresenceor absenceof steppingwassystematically relatedtothebiomechanicalconstraintsof thesituation(Thelenet al 1984,Thelenet al1987).This simulation of developmentalchangehighlightstheimportantcon-tributionof contextualvariables.

Thedevelopmentof reachingin differentposturesis anotherexampleof thecontextspecificityof motor control.Coordinatedreachingis only possible inthecontextof a stablebodyposture(von Hofsten1993,Paillard1991).Wheninfantsincapable ofsitting withoutsupport(22–26weeks ofage)areplacedina fully supportedposture(e.g. supineor reclined), they tend to reachforobjectswith bothhands(Rochat& Senders1991,Rochat1992).By contrast,infantscapable of sitting withoutsupport(28–38 weeksof age)reachwith onehand,regardlessof theirposture.Theyoungerinfantsalsoreachwith onehandwhenplacedin a seatedposition, becausetheymustcompensatefor a lossofbalanceby recruitingthe otherhandto helpstabilize themselvesin this posi-tion. Notethatin this caseinfantsshift to a differentresponsebecause the taskis different,not becausetheyhaveundergonea changein neuralor muscularcontrol.Theselectionof a morestableresponseinducedby behaviorbecom-ing morevariable in anewsituation or taskappearsto representoneof thefewgeneralprocessesin thedevelopmentof newactions(Freedland& Bertenthal1994,Manoel& Connolly 1995,Thelen1995).

Additionalevidencesuggeststhatperception-actioncouplingsarerelativelyspecificandthusdo not generalizeto similar actions.For example,Rochat&Senders(1991) report a progressionin hand-mouthcoordinationfrom bi-manualactionorganizedin mirror imagesymmetry towardan asymmetricalinvolvementof the hands.This sameprogressionis repeatedwhen infantsbegin to visually explore objects that aregrasped, even though bimanualreachingis less flexible or functional. Likewise, visual control of a sittingposturedoesnot generalizeto visualcontrolof a standingposture(Bertenthal& Bai 1989,Woollacott& Sveistrup1994).It appearsthat infantsmustlearnto modulateor control eachnewmotor responsedenovo,evenif thepercep-tual information (e.g.opticalflow specifyingself-motion) is readilyprocessed.

Thereis an importantmoral to this section.It is misleadingandcontradic-tory to ascribea specific perceptualskill to an infant basedon one task.Performanceis basedon multiple factors;seemingly insignificant variables,suchasposturalstability or orientation,produceprofoundeffects.Moreover,thepresenceor absenceof the skill in questionwill also dependon thespecificmotorresponsenecessaryfor performingthetask.For this reason,researchersarewell advisednot to discusstheseperceptualskills independentof the taskor response assessed.A betterapproach isto determinewhether theperceptualinformationis transformedinto theappropriatemotorresponsesynergies,and

444 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

to specifyall thefactorsthatcontributeto this process.This strategyshiftsthefocus from assessingonsetof perceptualskills to understandinghow theseskills develop.

OBJECTPERCEPTION AND RECOGNITION

The recognitionsystemis distinguishedfrom the perception-actionsystemprimarily by thefact thatrecognitionis definedwith referenceto thepast.If Irecognizea studentor a friend, for example,someimmediately perceivedinformationmustmatchsomepreviouslystoredinformation. Although theo-rists (Biederman1987,Marr 1982,Rock 1984)offer little consensuson therepresentationalformat of this information, it is reasonablycertain that thestoredinformation is not an exactcopy of the perceivedscene.Logical andfunctionalimperativesdictatethat storingall the availableinformationin thesceneis neithernecessarynor usefulfor specifictasks.Theordinaryenviron-mentconsists of a hierarchicalnestingof informationat multiple scalesrang-ing from largeobjects, suchasmountainsandtrees,to verysmallobjects,suchas leavesand cells; it is rich in structureand consistsof places,surfaces,layouts, people,animals,etc. From any point of observation,a plenum ofstructuredinformation is available,including the textureandcomposition ofindividual surfaces,thearrangementof thosesurfacesin thespatiallayout,andthe binding of someinto distinct objects.What is perceivedand recognizeddependson theintentionsandgoalsof theobserver.Thetaskof recognizingabook is thesameregardlessof its orientation, position,or locationrelativetotheobserver.It is thusunlikely thatinformation aboutthespatialpropertiesofthe book will be storedfor purposesof book-recognition,becausethesewillchangewith theposition of theobserver.Unlike theperception-actionsystem,which is viewer-centered,the recognitionsystemencodesand storesthosepropertiesthatareinvariantacrossmultiple perspectivetransformationsof theobject.

In this section,we reviewevidencethatrecognitionbeginsat birth andthatinfantsareendowedwith perceptualdecodingprinciplesthatexploit someofthe mostimportant regularitiesin the physical world. (This discussion isrestrictedto theprocessingof visual information becausesuchinformationismostrelevantto theproposedfunctionaldissociationthatguidesthis review.)Theseprinciplesare initially availablein implicit form andguideperceptionby constrainingor privileging certaininterpretations of the visual scene.Asinfants perceivethe sameinformation repeatedly,the storedrepresentationsderivedfrom theseexperiencesbecomeincreasinglyrich andabstract(Eimas1994).Perceptualandconceptualknowledgeblendtogetherin this framework,becausestoredrepresentationsareaccessiblefor bothrecognitionandreason-ing (Spelke 1994).

PERCEPTION, ACTION, ANDREPRESENTATION 445

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

NewbornRecognitionof Objects

Thetraditionalview is thatinfantsareendowedwith very simplecapacitiestolook at objects—capacitiesthat enablethemto perceivethe world piecemealvia fleeting images(e.g.Piaget1954).This view acknowledgesthat recogni-tion of somekind occursearly in life, but it is of a kind linked to previouslyproducedactionsandnot to storedinformationaboutthe world. Sucha pro-posalis generallyconsistentwith thedevelopmentof thosebehaviorsassoci-atedwith theaforementionedperception-actionsystem, but it doesnot gener-alizeat all to thedevelopmentof theobjectrecognitionsystem.Nevertheless,recent studies on object recognition reveal evidenceof very early repre-sentation,beginning with recognitionmemoryin neonates(Slateret al 1984,1990a,b).

Thehabituation paradigmis foundational to thestudyof perceptualrecog-nition by infants(Bornstein 1985)andthusdeservesa brief overview.In thisparadigminfantsarepresentedwith a specificstimulus for a numberof trialsuntil their attention declines.One reasonthat infants’ attentivenessdeclinesover trials is presumablythat as they developa storedrepresentationof thestimulus it becomesless interesting. The encodingand storageof stimulusinformationaretestedby presentinga novelstimulusfollowing somecriteriondecreasein responding. If theinfant’s decline in responsiveness occurredbecausethe first stimulusbecamefamiliar, thena novelstimulusshouldrein-itiate responsiveness. Conversely, the novel stimulus shouldnot produceanincreasein responsivenessif the previousdeclinewas simply a function offatigue.Note that the sensitivity of this paradigmwith very young infants isattributableto requiringonly very simple responses,such as visualfixation.

The finding that neonateshabituateto visual displaysis provocativebe-causeit confirms that they begin to storeperceptualinformation from theirfirst encounterswith theworld. It is notnecessarilythecase,however,thatthisinformationis storedbeyondtheperiodof thestudy.Recently,Rovee-Collier(1995) introducedthe conceptof a “time window” to explainwhenandhownew information would be integratedinto memory. In essence,this theorypredictsthatrepeatedencounterswith thesame information over shortperiodsincrease the likelihood oflong-termretentionof thatinformation.

Studies ofneonates’ face recognitioncapabilitiessuggestthat the timewindow is functional from birth. Bushnellet al (1989)showedthat neonateslook longerat their mothers’ facesthanat strangers’ faces,evenwhenolfac-tory informationwascontrolledby maskingtheodorof themother.Walton etal (1992)reportedthat this preferencewasalsoobservedwhenthefaceswerevideorecorded.More recently,Pascaliset al (1995)replicatedthis finding butshowedthatthispreferencewasextinguishedwhenwomenworeheadscarves.Theauthors suggestthatneonatesstorearepresentationof theirmothers’ faces

446 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

in which thehairlineandoutercontourplay a prominent role. This finding isconsistentwith the evidencethat very young infants are biasedtoward theperception of low spatial frequencies, i.e. large-scale pattern information(Banks& Dannemiller1987).As the spatialresolutionof the visual systemimproves, infantsrespondmoreto theinternalfeaturesof theface(deSchonen& Mathivet 1989,Morton & Johnson1991).

Thesefindings convergeto show that infants begin to store frequentlyrepeatedperceptualinformation from birth. Presumably,this information isfunctionally significant and engagesthe infants’ attention much more thanotherinformationavailable tothem.Although themotheris typically specifiedby multimodalinformation,it is intriguing to find thatneonatesshowrecogni-tion of modality-specific information. This finding is consistent with otherevidence thatperceptual information isstoredin a modality-specific formatbythe recognitionsystem(Bertenthal& Rose1995).Lessclearis the organiza-tion of the perceptualinformation storedby the neonate.We addresssthisissuein thenext section.

Implicit Knowledge ofObjects

Beforevisual information is storedby infants,theymustbind and/orsegmentit into units likely to be perceivedwhen the sameinformation is presentedagain.Thecomplexity of mostvisualscenesmakesthis a formidable taskforyounginfants.Theinfant’s visualworld mayincludea wide varietyof objectsrangingfrom blanketsandstuffedanimalsto people,machines,andfurniture.Most objectsarenot completelyvisible: Portionsareoccluded,boundariesarenot alwaysdelineated,andtheprojectivestructureof this information changescontinuously as infants and objectsmove.The first task for the infant is toascertainwhich surfacesand features inthe optic array compriseobjectsdistinctfrom otherobjects.

Object segmentation is readily accomplishedby adult observers,who or-ganize the visualscenebasedon perceptualgrouping principles, physicalknowledgeof objects,and pastexperience(Needham& Baillargeon1995).Suchgroupingprinciplesapplyacrossmanydifferentcontextsandsituations.Recentresearchsuggeststhat evenyoung infants apply someof thesesameprocesseswhen viewing the visual world. For example,Spelke& van deWalle (1993) reportthat 3-month-old infantsperceivetwo objectsasdistinct iftheyareseparatedin depthor moveindependently. By contrast,younginfantsdo not perceiveboundariesbetweenobjectsthat arestationary andadjacent,even if the objectsdiffer in color, texture, and form. It is suggestedthatinfants’ perceptionof the spatial layout follows two specific principles orprocessingconstraints.In essence,theseprinciples assertthat surfacesareperceivedasconnectedif, andonly if, they movetogether(principle of con-tact) or lieon a single object (principleof cohesion).

PERCEPTION, ACTION, ANDREPRESENTATION 447

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

Additional studiesrevealthat younginfantsperceivetheunity of partiallyoccludedobjectswhen the visible partsareseento move togetherin a rigidfashion (henceforthreferredto as“rigid motion”) (Baillargeon 1987,Craton&Yonas1990,Kellmanet al 1986,Johnson& Nanez1995,Slateret al 1990b).In an early study by Kellman & Spelke(1983), 4-month-old infants wererepeatedlypresentedwith a vertically orientedrod that movedhorizontallybackandforth behinda block thatoccluded thecenter of the rod.Once infantsbecamehabituatedto this event,theywerepresentedon alternatingtrials withtwo noveldisplays—a completerod andabroken rodconsisting of two collin-earsegmentsseparatedby a gap in the middle. Infantsshoweda significantincreasein visual attentionto the broken rod but not to the completerod,suggestingthattheyperceivedthepreviouslypresentedpartially occludedrodas a unitary object. Later studiesrevealedthat any rigid translationof thepartially occludedrod enabled4-month-old infants to perceivethe visibleportionsas unified (Kellman et al 1986); on the other hand,staticgroupingprinciplessuchasgoodcontinuation, collinearity,andsimilarity of textureorcolor were notsufficientto suggestunity (Kellman& Spelke 1983).

As infants grow older they begin to exploit additionalregularitiesof thephysicalworld in theprocessof segmentingobjects.Needham& Baillargeon(1995)reviewanumber ofrecentfindingsshowing thatinfantsby 8 monthsofageemploy featuralproperties,suchas color and shape,and physicalcon-straints,suchas the impenetrability of surfaces,to help them interpret ambigu-ousarrangementsof objectsin thevisualscene.Their findingsalsosuggestanimportantcaveatto interpretingage-relatedchangesin object segmentation.Wheninfantsweregivenprior experience withthe objectsthatwere displayedduring the experiment, thecorrect perceptualresponse wasshown at 4.5monthsinsteadof at 8 months of age.Apparently,objectsegmentationproc-essesinteractwith pastexperiencewheninfantsperceptuallyorganizea visualscene.

Theinterplay betweenperceptualgroupingprocessesandpastexperienceisalsoillustratedby researchon infants’ perceptionof biologicalmotions(Ber-tenthal 1993). Thesemotions are depictedby points of light moving as ifattachedto the major joints and headof a personwalking. Adult observers,who do not recognizestatic displays of such point-lights in any consistentway, recognizethemoving point-lightsasdepictinga humanform in less than0.5 s (Johansson1973,Bertenthal& Pinto 1994).Three-andfive-month-oldinfants discriminate thesesamemoving point-light displaysfrom ones inwhich the temporalpatterningof the lights are perturbed(Bertenthalet al1987,Proffitt & Bertenthal1990). It is conjecturedthat multiple processingconstraints,including storedknowledgeof the humanform, contributeto theinterpretationof these point-light displays(Bertenthal 1993,Bertenthal &Pinto 1994).This conjectureis supportedby findings showingthat 5-month-

448 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

old infants do not discriminatepoint-light displaysdepictingunfamiliar ob-jects, suchas a four-leggedspider,from a perturbedversion(Bertenthal&Pinto1993).

At a general level, thepoint-light display is similar to thepartiallyoccludedrod becausebothdepictunitaryobjectsthataremoreor lessoccluded.Never-theless,an importantdifferencebetweenthe displays is that the visible seg-mentsof the rod move rigidly, whereasthe point-lights do not. The contactprinciple proposedby Spelke& van de Walle (1993;seeabove)exploitstherigid motionsof spatiallyseparatedsurfacesto organizethe visual scene.Inthe caseof biological motions, this principle will not suffice. Note that theobjectmotionsin thesetwo displayscorrespondto very different categoriesofknowledge(i.e. physicalvs biological), suggestingthat different processingprinciplesmaybeassociatedwith differentcoredomains(Leslie1988,Carey& Spelke1994).

Oneof the most importantreasonsthat the precedingdecodingprinciplesappearso generalizableis that they mirror thephysicalconstraintsgoverningan object’s behavior(Spelkeet al 1992).A similar conclusionis reachedbyShepard (1994) based onhis researchon adult perception ofapparent motions.In the absenceof any physicallypresentedmotion, the perceivedpathof anapparentmotionis underspecifiedandmustthereforereflectcertainorganizingprinciples in the visual system.Shepardcontendsthatthevisual systemselectsa particularmotion from amongtheinfinite setof possiblemotionsaccordingto theconstraintsof kinematicgeometry,which governthe relativemotionsofrigid objects,or of local parts of nonrigid objects,during brief moments.Accordingto Shepard,thesephysicalconstraintsrepresentsomeof the mostpervasivepropertiesabouttheworld thathaveenduredthroughout evolution;thus,naturalselectionshouldhavefavoredgenesthat internalizedthesecon-straintsas processingprinciples.

Computationalresearchinvestigatinginfants’ perceptionof structurefrommotionofferssomeadditionalsupportfor theinternalization of thosephysicalconstraintsthat contribute to the perception ofobjects.When perspectiveinformation is excludedfrom the display,a transforming2-dimensional pro-jectionof a3-dimensionalstimulus is underdetermined. Itis thusnecessaryforobserversto implementadditional processingconstraints,suchas a rigidityassumption, in order to extracta uniquestructurefrom the projection(Marr1982).Arterberry& Yonas(1988)testedinfants’ perceptionof structurefrommotionby testing themwith a 2-dimensionalprojectionof a spherefilled witha randomdistribution of elements.Whenviewedstatically, the2-dimensionalimageappearsambiguous, but assoonastheelementsbeginto rotate,a rigidform is perceived—aslong asthe imageis interpretedasthe projectionof a3-dimensional rigid object.Arterberry& Yonasshowedthat infantsasyoungas4 monthsof ageperceive3-dimensional formsfrom 2-dimensional projec-

PERCEPTION, ACTION, ANDREPRESENTATION 449

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

tionsthatchangeovertime.Similar findingsarepresentedby Kellman(1984)andKellman & Short(1987).Collectively, thesefindings suggestthat younginfantsaresensitiveto someof thesameprocessingconstraintsusedby adultsfor perceiving3-dimensionalobjects.

Explicit Knowledge ofObjects

Theoreticalopinion is divided over whetherthe processingconstraints dis-cussedaboveareinducedfrom earlyperceptualexperiencesor correspondtoinnate cognitive representationsof the physical world (Baillargeon 1993,1995; Karmiloff-Smith1992; Leslie 1988; Mandler1988,1992; Spelke1994).In eithercase,anabstract representationof some kindguidestheperceptionofobjectsby 3 monthsof ageor younger.Moreover,recentfindingssuggestthatthese representations not onlyguide perception but also provide an earlyfoundationfor reasoningabouttheworld. In otherwords,theseabstractrepre-sentationsabout the motionsof objectsare accessibleto infants as explicitknowledge(Baillargeon1993,1995;Leslie1988;Spelke etal 1992).

Theprincipalevidencefor thisknowledgederivesfrom occlusionstudiesinwhich inferencesare requiredbecausethe entire event is not visible. Forexample, Spelke et al (1992) tested 2.5–4.5-month-old infants’ reasoningabout the continuity and solidity of objectsby habituating them to variousevents.In oneexperiment, aninitially visibleball wasdroppedbehindascreenand thenthe screen was raisedto reveal theball on thefloor of the stage. Afterinfantsbecamehabituatedto this event,a brightly coloredsurfacewasaddedabovethe floor, the screenwas loweredto coverboth surfaces,andthentheball wasdroppedagain.On alternatingtrials thescreenwasraisedto revealaball on thenewsurfaceor onthefloor of thestage.Adult observersreasonthatthe former eventis (covertmanipulation aside)impossible becauseit wouldrequirethe ball to passthroughor jump over the uppersurface.Suchmove-mentswouldviolateboththecontinuity principleof objects(i.e.objectsfollowonecontinuouspathover spaceand time) and the solidity principle (i.e. thepartsof distinct objectsmay nevercoincidein placeand time). Apparently,infants appreciatethese principles,becausethey paidconsiderably morevisualattentionto theimpossible eventthanto thepossibleevent.Additional experi-mentsby Spelkeet al (1992)andby Baillargeon(1993)convergeto showthatyounginfantsreasonaboutthe continuity andsolidity of objectsinvolved insimplephysicalevents.

Certainotherphysicalconcepts,suchasgravity andinertia,arenot under-stoodby infants until closeto the end of their first year (Spelkeet al 1992,Spelkeetal 1994).This finding highlightsanimportantdifferencebetweentheperception-action system and the object-recognition system. Research onyoung infants’ reachingfor moving targets(reviewedabove)suggeststhatthey are implicitly sensitive to gravity and inertia by 5 to 6 months of age;

450 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

otherwise,it is difficult to imaginehow theycouldsuccessfullyanticipatethefuture position of moving targets.This implicit knowledgeis not generalizable,however,becauseit is encapsulatedin the actionandis thusnot repre-sentedin a formataccessibleby thecognitive system.By contrast,knowledgeaboutgravity andinertia that is representedby the object-recognitionsystemwill generalize across situations onceit is storedas a representation.

The conclusionthat young infants reasonaboutphysical eventsis contestedby sometheorists(e.g.Fischer& Biddell 1991,Oakes& Cohen1995,Siegler1993). In evaluting this research,it is important to distinguish betweentheevidencefor infants’ reasoningand the conclusionthat theseresultsreflectinnatetheoriesof the physicalworld. The empiricalevidencethat reasoningoccursis muchmoredefensiblethantheevidencefor aninnatecoreof knowl-edge,especiallybecausemoststudiesof infants’ physicalreasoningfocusoninfants3 monthsold or older.Resolution of this issuerequiresanexplanationof how coreknowledge about the physical world (e.g.about thecontinuity andsolidity of objects)developsfrom perceptualinformation (Spelke1994).Thisis a dauntingrequirement,but preliminary ideasare beginning to appearin theliterature.

Baillargeon(1995) suggeststhat the development of physicalknowledgereflectsa highly constrainedlearningmechanismthat processesthe availableperceptual information. Regrettably, the details ofthis learning mechanism arevague,althoughsomeof the predictionsarenot. Baillargeonclaims that in-fantsfirst learnto encodethecausalpropertiesof a physicalevent,suchasacollision, at a very simplelevel andthenbeginto encodebothqualitative andquantitative variablesabouttheevent.For example,a series of experimentsbyKotovsky & Baillargeon (reviewedin Baillargeon1995) revealsthat 2.5-month-old infants expecta stationaryobject to be displacedwhen hit by amovingobject,but it is not until 5.5–6.5monthsof agethat infantsrecognizethat a stationaryobject should be displacedfurther by a larger than by asmallermoving object.It is conjecturedthat infantswill beginto encodeothervariables,suchasmass,speed,anddistance,following additional experiencewith relevant events(Baillargeon1995).

ObjectRecognition Is Domain Specific

An importantimplication ofthe precedingtheory isthatinfants’ knowledge ofphysicaleventsdevelopspiecemealandnot all at once.Someobjectproper-ties,suchassize,aremoreeasilydetectedthanothers,suchasmass,andthusconceptualknowledgeof the physicalworld developsgraduallyas itis tutoredby experience.A relateddevelopmental trend is that infants appearto usespatio-temporal informationto specifytheidentity of objectsprior to individu-atingtheseobjectson thebasisof featuralinformation (Cohen& Oakes1993,Xu & Carey1995).Eventhoughinfantsaresensitiveby 6 to 7 monthsof age

PERCEPTION, ACTION, ANDREPRESENTATION 451

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

to launching events andother events involving spatio-temporal continuity(Baillargeon1993,Leslie& Keeble1987,Oakes1995),recentevidencesug-geststhat they do not detectthe individuating propertiesof objectsin thesedynamiceventsuntil 10 to 12 monthsof age(Xu & Carey1995). Evidence forthis latterconclusionis somewhatsuspect,however,becauseinfantscanper-ceptuallycategorizecertainbasickindsof objectsby 3 monthsof age(Quinn& Eimas1986).Perceptualcategorizationrequiresthatinfantsbothdistinguishbetweenand generalizeaboutobjectsbasedon specific features;thus, thesecategorizationfindings suggestthat young infants do representidentifyingfeatures of objects.

The apparentdiscrepancybetweenthe findings cited abovedmay be ex-plainedby notingthatevidencefor perceptualcategorizationof objectsprior to10 monthsof ageis restrictedprimarily to animateobjects,suchaspeopleandanimals(Eimas& Quinn1994, Eimas et al1994, Quinn& Eimas1986, Quinnet al 1993). By contrast,studiesinvestigating infants’ knowledgeof objectmotionsand causalinteractionsinvolve inanimateobjects,suchas carsandblocks (e.g. Baillargeon1995, Leslie 1988, Oakes& Cohen1995). Recentneuropsychological evidencesuggeststhat adultsrepresentand perceptuallyprocessliving andnonliving thingsdifferently (Farah 1992).In general,livingthingsarerepresentedholistically whereasnonliving thingsare representedbytherelationsamongtheirparts.Theoriginsof thesedifferencesmaybepresentearly in development,a possibility that would help to explain the finding ofdifferent representationsfor animateandinanimateobjectsby infants.

Although sucha reconciliationbetweenfindings is somewhatspeculative,it is certainly consistentwith recentproposalsthat conceptualknowledgeispresentearly and organizedby core principles of domain-specific knowl-edge—e.g. people, objects, number(Cary& Spelke1994, Wellman& Gelman1992). One explanationfor early differencesthat emergebetweencore do-mainsfollows from theproposalby Eimas(1994)thatcategoricalknowledgeemergesfrom a progressiveabstractionof the storedperceptualinformation.Presumably,the perceptualpropertiesthat are initially encodedand storeddiffer asafunctionof thetaskbeingperformedalongwith theseprocesses,andthus initial knowledgeaboutpeopleandobjectswill differ becauseour earlyinteractionswith theseentitiesare so different.

CONCLUSIONS

The researchdiscussedin this review is noteworthyfor many reasons,butthreearehighlighted.First, recentfindings on newbornsrevealthat they areconsiderablymorecompetentthanoncebelieved.Neonatesareendowedwithmuch more than reflexesor fixed actionpatterns.Their actionsaregoal di-rectedand spatially coordinated,eventhoughthey lack consistencyand are

452 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

oftenobscuredby other factors,such as posture.Also, neonatesshowrecogni-tion memoryfrom birth, andby 3 months of ageinfantsreasonabouteventsthat areperceptuallyoccluded.During the first yearof life, theseactionsandrepresentationsbecomemoreaccessibleandmoregeneralizable,but theclearmessagefrom therecentresearchis thattherepresentationsavailableto infantsarenotdirectlya functionof the coordinationand internalization of actions.

Second,this review was organizedby the proposalthat perceptionandactionand object recognitionand representationare functionally dissociableprocessesthat follow different developmentaltrajectories.Although it is pre-matureto offer specific predictionsbasedon this proposal,it is possible tooffer a few generalizationsregarding thedevelopmentof bothsystems.

The findings on thedevelopmentof perceptionandactionconcurwith theearliercharacterizationof theperceptualcontrolsystemasfunctioningin realtime with litt le explicit referenceto pastexperience.Perceptuallycontrolledactionsaremodularlyorganizedandnot representedin a form that is recalledor usedto guidetheproductionof otherresponses.It is sufficient that infantsperceivewheretheyareandwhattheyaredoing.All that mattersis thepresentfit betweenthe infant and the environment.Pastexperiencedoesof coursecontribute to the developmentof thesesystemsin the form of perceptuallearning.With exploration of their own actionsandthe environment,infantsshowincreasingsensitivity to perceptualchangesandfiner controlof actionsthat are guidedby this information. In essence,then, learningis implicit orprocedural;it is elicitedby context,not by recallof explicit informationabouthow tocoordinatesensorimotor behaviors.

By contrast,researchon the developmentof object recognitionconfirmsthat evenyoung infants representdefining featuresand physicalconstraintsthat specify the unity andboundednessof objects.It is significant that thesedefiningproperties,suchasstructurefrom rigid motions,areinvariantacrossperspectivetransformations.This invarianceincreasesthe likelihood that ob-jects are recognizedin new orientations and contexts.The representationsnecessaryfor object recognitionare organizedas domain-specificconceptsthat accumulatenewperceptualinformation over time. Perceptualexperienceis the principal engineby which representationsbecomemore abstractanddifferentiated.Theserepresentationsnot only guideperceptualprocessing,butalsoareavailablein explicit form to guideinfants’ reasoningaboutthe mostfundamentaland frequentlyencounteredpropertiesof their world.

The third and last conclusionemerging from this review concernsthegeneralizabilityof specificresearch findings.It is always advisableto exercisecautionwhen generalizingbetweendifferent contextsand measures(Thelen1995), and the current review offers new reasonsfor suchcaution.As dis-cussedrepeatedlyin this review, competenciesmanifestedby one responsesystem,suchas eye movements,are not necessarilymanifestedby another

PERCEPTION, ACTION, ANDREPRESENTATION 453

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

system,such as reaching,at the sameage.An additional complicationbe-comesevident when inferring conceptualcompetencies,e.g. object perma-nence,from variousresponsemeasures,becausethe developmentof concep-tual knowledgeandsystemsaremediatedby dissociable processesthatfollowindependentdevelopmentaltrajectories.Thus, the developmentof a specificresponsemeasurecouldeasilylagbehindor otherwisemisrepresentthedevel-opmentof a specificconcept.

The proposedfunctional dissociation betweenperceptuomotorandobjectrecognitionprocessesis not meantto imply thatthesetwo processingsystemsarecompletely independent.Clearly,thoughtandaction interact atsome level,anddevelopmentalchangesin onesystemsurelyaffecttheother.An importantresearchtask forthefutureis to investigatewhenandhownewrepresentationsand actionsare coordinatedin theprocess of development.

Any Annual Reviewchapter, aswell asany arti clecited in an Annual Reviewchapter,may bepurchased fromthe Annual ReviewsPreprints and Reprints service.

1-800-347-8007, 415-259-5017, email: arpr@class.org

LiteratureCited

Adolph KE, Eppler MA, Gibson EJ. 1993a.Development of perception of affordances.Adv.Infancy Res.8:51–98

Adolph KE, Eppler MA, Gibson EJ. 1993b.Crawling versuswalking infants’ percep-tion of aff ordances for locomotion oversloping surfaces.Child Dev.64:1158–74

AnisfeldM. 1991. Review: neonatal imitation.Dev. Rev.11:60–97

AshmeadDH, Clifton RK, Perris EE. 1987.Precision of auditory localization in humaninfants.Dev.Psychol. 23:641–47

AshmeadDH, Davis DL, Whalen T, OdomRD. 1991. Sound localization andsensitiv-ity to interaural time differencesin humaninfants.Child Dev.62:1211–26

Asli n RN. 1988. Anatomical constraints onoculomotor development: implications forinfant perception. In The Minnesota Sym-posiaonChild Psychology: Perceptual De-velopment in Infancy, ed. A Yonas, pp.67–104. Hil lsdale,NJ:Erlbaum

Aslin RN, Salapatek P. 1975. Saccadiclocal-ization of visualtargetsby the very younghuman infant. Percept. Psychophys.17:293–302

ArterberryME, YonasA. 1988. Infants’ sensi-tivity to kinetic information for three-di-mensional object shape.Percept. Psycho-phys.44:1–6

Bai DL, BertenthalBI. 1992. Locomotor statusand the development of spatial searchskills. Child Dev.63:215–26

Baill argeon R. 1987.Object permanence in

3.5- and 4.5-month-old infants. Dev. Psy-chol. 23:655–64

BaillargeonR. 1993. The object conceptrevis-ited: new directions in the investigation ofinfants’ physical knowledge. In CarnegieSymposium on Cognition: Visual Percep-tion and Cognition in Infancy, ed. CEGranrud, pp. 265–315. Hi l lsdale, NJ:Erlbaum

BaillargeonR. 1995. A model of physical rea-soning in infancy. Adv. Infancy Res. Inpress

Baldwin JM. 1906. Mental Development in theChild and the Race.NewYork: Macmillan

Banks MS. 1988. Visual recalibration and thedevelopment of contrast and optical flowperception. In TheMinnesota Symposia onChild Psychology: Perceptual Develop-mentin Infancy, ed.A Yonas,pp. 145–96.Hil lsdale,NJ:Erlbaum

BanksMS, Dannemiller JL. 1987. Infantvisualpsychophysics.In Handbook of Infant Per-ception. Vol. 1: From Sensation to Percep-tion, ed.PSalapatek, L Cohen,pp. 115–84.Orlando: Academic

BayleyN. 1969. Bayley Scalesof Infant Devel-opment. New York: Psychological Corp.

Ber tenthal BI . 1993. Percept i on ofbiomechanicalmotionsby infants: intrinsicimage and knowledge-basedconstraints. InCarnegie Symposium on Cognition: VisualPerception and Cognition in Infancy,ed. CGranrud, pp. 175–214. Hi l lsdale. NJ:Erlbaum

454 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

Bertenthal BI, Bai DL. 1989. Infants’ sensitiv-ity to optical flow for controll ing posture.Dev. Psychol. 25:936–45

Bertenthal BI, Campos JJ. 1990. A systemsapproachto the organizing effectsof self-produced locomotion during infancy. Adv.Infancy Res.6:51–98

Bertenthal BI, Campos JJ,Kermoian R. 1994.An epigenetic perspective on the develop-ment of self-produced locomotion and itsconsequences.Curr. Dir. Psychol. Sci. 5:140–45

Bertenthal BI, Pinto J. 1993. Dynamical con-straints in the perception andproduction ofhuman movements. See Thelen & Smith1993, pp. 209–39

BertenthalBI, Pinto J.1994. Globalprocessingof biological motions. Psychol. Sci. 5:221–25

Bertenthal BI, Proffit t DR, Kramer SJ. 1987.The perception of biomechanicalmotions.J. Exp. Psychol.: Hum. Percept.Perform.23:171–78

Bertenthal BI, Rose JL. 1995. Two modes ofperceiving the self. In The Self in Infancy:Theory and Research, ed. P Rochat. Inpress

Bertenthal BI, RoseJL, Bai DL. 1995. Percep-tion-action coupling in the developmentofvisualcontrol of posture. J. Exp. Psychol.:Hum.Percept.Perform. In press

Biederman I. 1987. Recogniti on-by-compo-nents: a theory of human image under-standing. Psychol. Rev.94:115–47

Blass EM, Fillion TJ, Rochat P, HoffmeyerLB, MetzgerMA. 1989. Hand-mouth coor-dination in neonates. Dev. Psychol. 25:963–75

Bloch H, Carchon I. 1992. On the onset ofeye-head coordination in infants. Behav.Brain Res.49:85–90

Bornstein MH. 1985. Habituation of attentionas a measureof visualinformationprocess-ing in humaninfants: summary, systemati-zation, and synthesis. In Measurement ofAudition and Vision in the First Year ofPostnatal Life: A Methodological Over-view, ed. G Gottlieb, NA Krasnegor, pp.253–300. Norwood, NJ:Ablex

Bremner JG, Bryant PE. 1985. Active move-mentanddevelopment of spatial abili tiesininfancy. In Children’s Searching: the De-velopmentof Search Skill and Spatial Rep-resentation, ed. HM Wellman,pp. 53–72.Hil lsdale,NJ:Erlbaum

BrunerJ. 1973. Beyond the InformationGiven:Studiesin thePsychologyof Knowing. NewYork: Norton

Bushnell E. 1985. The decli ne of visuallyguided reaching during infancy. Infant Be-hav. Dev.8:139–55

Bushnell EW, BoudreauJP. 1993. Motor de-velopment andthe mind: the potential roleof motor abili ties as a determinant of as-

pects of perceptual development. ChildDev.64:1005–21

Bushnell IWR, Sai F,Mull in JT.1989. Neona-tal recognition of the mother’s face.Br. J.Dev.Psychol. 7:3–15

Butterworth G. 1982. Object permanenceandidentity in Piaget’s theory of infant cogni-tion. In Infancy and Epistemology: AnEvaluation of Piaget’s Theory, ed. G But-terworth, pp. 137–69. New York: St. Mar-ten’s

Butterworth G, Hicks L. 1977. Visual proprio-ception andpostural stabili ty in infancy: adevelopmental study. Perception 5:255–63

Butterworth GE, Hopkins B. 1988. Hand-mouth coordination in the newborn baby.Br. J. Dev.Psychol. 6:303–14

Campos JJ,Bertenthal BI, Kermoian R. 1992.Early experience and emotional develop-ment: the emergence of fear of heights.Psychol. Sci. 3:61–64

Canfield RL, Haith MM. 1991. Young infants’visual expectations for symmetric andasymmetric stimulus sequences.Dev.Psy-chol. 27:198–208

CareyS,GelmanR, eds.1991. TheEpigenesisof Mind: Essayson Biology and Cognition.Hil lsdale,NJ:Erlbaum

Carey S, Spelke ES.1994. Domain specificknowledge and conceptual change. Do-main-specificity in Cognition and Culture,ed.L Hirschfeld& SGelman, pp. 169–200.NewYork: CambridgeUniv. Press

Clif ton RK, Morrongiello BA, Kulig JW,Dowd JM. 1981. Developmental changesin auditory localization in infancy. In De-velopment of Perception, Vol. 1, Psychobi-ological Perspectives, ed. R Aslin, J Al-berts, M Petersen,pp. 141–60. New York:Academic

Clif ton RK, Muir DW, Ashmead DH, Clark-son MG. 1993. Is visually guidedreachingin earl y infancy a myth? Child Dev.64:1099–110

Clif ton R, Rochat P, Litovsky R, Perris E.1991. Objectrepresentation guidesinfants’reaching in the dark. J. Exp. Psychol.:Hum.Percept. Perform. 17:323–29

Clif ton R, Rochat P, Robin D, Berthier N.1994. Multimodal perception in the controlof infant reaching. J. Exp.Psychol.: Hum.Percept. Perform. 20:876–86

CohenLB, OakesLM. 1993. How infants per-ceive a simple causalevent. Dev.Psychol.29:421–33

Craton L, YonasA. 1990. The role of motionin infant perception of occlusion. In TheDevelopment ofAttention: ResearchandTheory,ed.J Enns,pp. 21–46. Amsterdam:Elsevier

Daniel BM, Lee D. 1990. Development oflooking with headandeyes.J. Exp. ChildPsychol. 50:200–16

de Schonen S, Mathivet E. 1989. First come

PERCEPTION, ACTION, ANDREPRESENTATION 455

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

first served. A scenarioabout developmentof hemisphericspecialization in facerecog-nit ion during infancy. Eur. Bull . Cogn.Psychol. 9:3–44

DiamondA. 1991. Neurophysiological insightsinto the meaning of object conceptdevel-opment. See Carey & Gelman1991, pp.67–110

Eimas PD. 1994. Categorization in early in-fancy and the continuity of development.Cognition50:83–93

Eimas PD,Quinn PC. 1994. Studies on theformation of perceptually based basic-levelcategories in young infants. Child Dev.65:903–17

EimasPD, Quinn PC,CowanP. 1994. Devel-opmentof exclusivity in perceptually basedcategories of young infants. J. Exp. ChildPsychol. 58:418–31

FarahMJ. 1992. Is an object an object an ob-ject?Cognitive andneuropsychological in-vestigationsof domainspecifici ty in visualobject recognition. Curr. Dir. Psychol. Sci.1:164–69

FischerKW, Bidell T. 1991. Constraining na-tivist inferencesabout cognitive capacities.See Carey& Gelman1991, pp. 199–236

Freedland RL, Bertenthal BI. 1994. Develop-mental changes in interlimb coordination:transition to hands-and-knees crawling.Psychol. Sci. 5:26–32

GibsonJJ.1979. The Ecological Approach toVisual Perception. Boston: Houghton-Mif -fli n

GibsonEJ,RiccioG, SchmucklerMA, Stoffre-gen TA, Rosenberg D, Taormina J. 1987.Detection of the traversability of surfacesby crawling and walking infants. J. Exp.Psychol.: Hum. Percept. Perform. 13:533–44

Gibson EJ, Schmuckler MA. 1989. Goingsomewhere:an ecological andexperimen-tal approach to development of mobil ity.Ecol. Psychol. 1:3–25

Goldfield EC. 1993. Dynamic systemsin de-velopment: action systems.SeeThelen &Smith 1993, pp. 51–70

Goodale MA, Milner AD. 1992. Separate vis-ual pathways for perception and action.TrendsNeurosci.15:20–25

Goodale MA, Milner AD, Jakobson LS, CareyDP. 1991. A neurological dissociation be-tween perceiving objects and graspingthem.Nature349:154–56

Grenier A. 1981. ‘Motoricité libérée’ par fixa-tionmanuelle dela nuque aucours despre-mières semaines de la vie [‘ Liberatedmovements’ by manual fixation of theneckduring the first weeksof li fe]. Arch. Fr.Pediatr. 38:557–61

Haith MM. 1980. Rules that Babies Look By:TheOrganizationof Newborn Visual Activ-ity. Hil lsdale,NJ:Erlbaum

Haith MM. 1993. Future-orientedprocessesin

infancy: the caseof visualexpectations. InCarnegie Symposium on Cognition: VisualPerception and Cognition in Infancy, ed.CE Granrud, pp.235–64. Hillsdale, NJ:Erlbaum

Haith MM. 1994. Visual expectations as thefirststeptowardthedevelopment of future-oriented processes.In TheDevelopment ofFuture-OrientedProcesses,ed.MH Haith,JBBenson, RJRoberts,BF Pennington, pp.11–38. Chicago: Univ. Chicago Press.Inpress

Haith MM, HazanC, GoodmanGS.1988. Ex-pectation andanticipation of dynamic vis-ual events by 3.5-month-old babies. ChildDev.59: 467–79

Hirschfeld H, ForssbergH. 1994. Epigeneticdevelopment of postural responsesfor sit-ting during infancy. Exp. Brain Res.97:528–40

Horobin K, Acredolo L. 1986. The role of at-tentiveness, mobil ity history andseparationof hiding sites onStage IV behavior. J.Exp.Child Psychol. 41:114–27

Howard IP. 1986. The perception of posture,self-motion, and the visual vertical. InHandbook of Perception and Human Per-formance. Vol. 1: Sensory ProcessesandPerception, ed. KR Boff, L Kaufman, JPThomas,18:1–62. New York: Wiley

Jeannerod M. 1994. Object oriented action. InInsights into the Reachto Grasp Move-ment, ed. KMB Bennett, U Castiello, pp.3–15. Amsterdam:Elsevier

JohanssonG. 1973. Visual perception of bio-logical motion and amodel for its analysis.Percept. Psychophys.14:201–11

Johnson MH. 1990. Cortical maturation andthe development of visualattention in earlyinfancy. J. Cogn. Neurosci.2:81–95

Johnson SP, NanezJE. 1995. Young infants’perception of object unity in two-dimen-sional displays. Infant Behav. Dev. 18:133–43

JouenF. 1990. Early visual-vestibular interac-tions and postural development. In Sen-sory-Motor Organizationand Developmentin Infancy and Early Childhood, ed. HBloch, BI Bertenthal, pp. 199–216. Dor-drecht: Kluwer

Karmiloff-Smith A. 1992. Beyond Modularity.Cambridge,MA: Bradford/MIT Press

Kellman PJ.1984. Perception of three-dimen-sional form by human infants. Percept.Psychophys.36:353–58

Kellman PJ, Short K. 1987. Development ofthree-dimensional form perception. J. Exp.Psychol.: Hum. Percept. Perform. 13:545–57

Kellman PJ, Spelke ES. 1983. Perception ofpartly occluded objects in infancy. Cogn.Psychol. 15:483–524

KellmanPJ,SpelkeES,Short KR. 1986. Infantperception of object unity from translatory

456 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

motion in depth and vertical translation.Child Dev.57:72–86

Kremenitzer JP, Vaughan HG, Kurtzberg D.Dowling K. 1979. Smooth-pursuit eyemovements in the newborn infant. ChildDev.50:442–48

LedermanSJ,Klatzky RL. 1987. Handmove-ments: a window into haptic object recog-nition. Cogn. Psychol. 19:342–68

LeeDN. 1993. Body-environment coupling. InEcological and Interpersonal Knowledgeof the Self, ed.U Neisser,pp. 68–88. Cam-bridge:CambridgeUniv. Press

LeeDN, AronsonE. 1974. Visual propriocep-tive control of standing in human infants.Percept.Psychophys.15:529–32

Leslie AM. 1988. The necessity of il lusion:perception andthought in infancy. ThoughtWithout Language, ed. L Weiskrantz, pp.185–210. Oxford: Clarendon

Leslie AM, KeebleS. 1987. Do six-month-oldinfants perceive causality? Cognition 25:265–88

Lew A, Butterworth G. 1995. The effectsofhunger on hand-mouth coordination innewborn infants. Dev. Psychol . 31:456–63

LishmanJR,LeeDN. 1973. The autonomy ofvisualkinaesthesis.Perception 2:287–94

Lockman JJ, Ashmead DH, Bushnell EW.1984. The development of anticipatoryhand orientation during infancy. J. Exp.Child Psychol. 37:176–86

MandlerJM. 1988. How to build ababy: onthedevelopment of an accessible repre-sentationalsystem.Cogn. Dev.3:113–36

Mandler JM. 1992. How to build a baby: II.Conceptual primitives. Psychol. Rev. 99:587–604

Manoel EdeJ, Connolly KJ. 1995. Variabil ityandthe development of skilledactions.Int.J. Psychophysiol. In press

Marr D. 1982. Vision. San Francisco, CA:Freeman

McKenzie BE, SkouterisH, Day RH, HartmanB, YonasA. 1993. Effective action by in-fants to contactobjectsby reachingor lean-ing. Child Dev.64:415–29

Meltzoff AN, Moore MK. 1983. Newborn in-fants imitateadult facial gestures. ChildDev.54:702–9

Meltzoff AN, Moore MK. 1989. Imitation innewborn infants: exploring the range ofgestures imitated and the underlyingmechanisms.Dev. Psychol. 25:954–62

Meltzoff AN, Moore MK . 1994. Imitation,memory, andthe representation of persons.Infant Behav.Dev.17:83–100

Morrongiello BA. 1988. Infants’ localizationof sounds inthe horizontal plane:estimatesof minimum audible angle. Dev. Psychol.24:8–13

Morton J, Johnson M. 1991. CONSPEC and

CONLERN: a two processtheoryof infantface recognition. Psychol. Rev.98:164–81

Mounoud P. 1993. The emergence of newskills: dialectic relations between knowl-edge systems.SeeSavelsbergh 1993, pp.13–46

Muir D, Field J. 1979. Newborn infants orientto sounds.Child Dev.50:431–36

NeedhamA, BaillargeonR. 1995. Object seg-regation in infancy. Adv. Infancy Res. Inpress

Newell KM. 1986. Constraints on the develop-ment of coordination. In Motor Develop-mentin Children: Aspects of Coordinationand Control, ed.MG Wade, HTA Whiting,pp. 341–60. Boston: MartinusNijhoff

OakesLM. 1994. Development of infants’ useof continuity cues in their perception ofcausality. Dev.Psychol. 30:869–79

OakesLM, CohenLB. 1990. Infant perceptionof a causalevent. Cogn. Dev.5:193–207

OakesLM, CohenLB. 1995. Infants’ percep-tionof causality. Adv. Infancy Res.In press

Paillard J. 1991. Motor and representationalframing of space.In Brain and Space,ed.J. Paillard, pp. 163–82. New York: OxfordUniv. Press

PascalisO, de SchonenS, Morton J, DeruelleC, Fabre-Grenet M. 1995. Mother’s facerecognition by neonates: a replication andan extension. Infant Behav. Dev. 18:79–85

PerrisE, Clifton R.1988. Reaching in the darktoward sound asa measureof auditory lo-calization in infants. Infant Behav. Dev.11:473–92

Piaget J. 1952. The Origins of IntelligenceinChildren.New York: Int. Univ. Press

Piaget J. 1954. The Construction of Reality inthe Child. New York: BasicBooks

Proffit t DR, Bertenthal BI. 1990. Convergingoperations revisited:assessing what infantsperceive using discrimination measures.Percept. Psychophys.41:1–12

Quinn PC,EimasPD. 1986. On categorizationin early infancy. Merr ill-Palmer Q. 32:331–63

Quinn PC, EimasPD, Rosenkrantz SL. 1993.Evidence for representations of perceptu-ally similar natural categoriesby 3-month-old and 4-month-old infants. Perception22:463–75

ReedES.1982. An outlineof a theory of actionsystems.J. Motor Behav.14:98–134

ReedES. 1989. Changing theoriesof posturaldevelopment. In TheDevelopmentof Pos-ture and Gait acrossthe Lifespan, ed. MWoollacott, A Shumway-Cook, pp. 3–24.Columbia:Univ. S.CarolinaPress

RegalDM, AshmeadDH, SalapatekP. 1983.The coordination of eye and headmove-ments during early infancy: a selective re-view.Behav. Brain Res.10:133–39

Robin D, Berthier N, Clifton R. 1995. Infants’

PERCEPTION, ACTION, ANDREPRESENTATION 457

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

predictive reaching for moving objects inthe dark. Dev.Psychol. In press

Rochat P. 1992. Self-sitting and reaching in5–8 month old infants: the impact of pos-ture and its development onearly eye-handcoordination. J. Motor Behav. 24:210–20

Rochat P, Blass EM, Hoffmeyer LB. 1988.Oropharyngeal control of hand-mouth co-ordination in newborn infants. Dev. Psy-chol. 24:459–63

Rochat P, Goubet N. 1995. Development ofsitting and reaching in 5- to 6-month-oldinfants. Infant Behav.Dev.18:53–68

Rochat P, Senders SJ. 1991. Active touch ininfancy: action systemsin development. InNewborn Attention: Biological Constraintsand the Influenceof Experience,ed. MJWeiss,PR Zelazo,pp. 412–42. Norwood,NJ:Ablex

Rock I. 1984. The World of Perception. NewYork: W. H. Freeman

Roucoux A, Culee C, Roucoux M. 1983. De-velopment of f ixation and pursuit eyemovements in humaninfants.Behav. BrainRes.10:133–39

Rovee-Collier C. 1995. Timewindows in cog-nitive development. Dev. Psychol. 51:147–69

Savelsbergh GJP,ed. 1993. TheDevelopmentof Coordination in Infancy. Amsterdam:Elsevier

Savelsbergh GJP,van der Kamp J. 1993. Thecoordination of infant’s reaching, grasping,catchingandposture:anaturalphysicalap-proach. See Savel sberg 1993, pp.289–317

Schmuckler MA, GibsonEJ.1989. The effectof imposed optical flow onguidedlocomo-tion in young walkers.Br. J. Dev.Psychol.7:193–206

Shea SL, Aslin RN. 1990. Oculomotor re-sponsesto step-ramptargetsby young hu-maninfants.VisionRes.30:1077–92

ShepardRN. 1994. Perceptual-cognitive uni-versals as reflections of the world. Psy-chon. Bull . Rev.1:2–28

Siegler RS. 1993. Commentary: cheers andlamentations. In Carnegie Symposium onCognition: Visual Perception and Cogni-tion in Infancy, ed. CE Granrud, pp.333–44. Hil lsdale,NJ:Erlbaum

SlaterA, MattockA, Brown E. 1990. Sizecon-stancyat birth: newborn infants’ responsesto retinal andreal size.J. Exp.Child Psy-chol. 49:314–22

Slater A, Morison V, Rose D. 1984. Habitu-ation in the newborn. Infant Behav. Dev.7:183–200

Slater A, Morison V, Somers M, Mattock A,Brown E, Taylor D. 1990b. Newborn andolder infants’ perceptionof partly occludedobjects.Infant Behav. Dev.13:33–49

SpelkeE. 1994. Initial knowledge: six sugges-tions.Cognition 50:431–45

Spelke ES,BreinlingerK, Macomber J,Jacob-son K. 1992. Origins of knowledge. Psy-chol. Rev.99:605–32

Spelke ES, Katz G, Purcell SE, Ehrlich SM,Breinlinger K. 1994. Early knowledge ofobject motion: continuity andinertia. Cog-nition51:131–76

Spelke ES,Van de Walle G. 1993. Perceivingandreasoning about objects: insights frominfants. In Spatial Representation, ed. NEilan,W Brewer,R McCarthy, pp. 132–61.Oxford: Blackwell

StackD, Muir D, Sherriff F, RomanJ. 1989.Development ofinfant reaching in the darkto luminous objects and“invisible sounds.”Perception 18:69–82

Stoffregen TA, Schmuckler MA, Gibson EJ.1987. Useof central andperipheralopticalflow in stanceand locomotion in youngwalkers.Perception16:113–19

Sveistrup H, Woollacott MH. 1995. Canprac-tice modify the developing automatic pos-tural response.Child Dev.In press

Thelen E. 1989. Self-organization in develop-mental processes:cansystemsapproacheswork? In The Minnesota Symposia onChild Psychology: Systems in Develop-ment, ed.M Gunnar, E Thelen,pp. 77–117.Hil lsdale,NJ:Erlbaum

Thelen E. 1995. Motor development: a newsynthesis.Am.Psychol. 50:79–95

Thelen E, Fisher DM. 1982. Newborn step-ping: an explanation for a ‘disappearing’reflex. Dev.Psychol. 18:760–75

ThelenE, FisherDM, Ridley-JohnsonR. 1984.The relationship betweenphysical growthand a newborn reflex. Infant Behav. Dev.7:479–93

Thelen E, Skala KD, Kelso JAS. 1987. Thedynamic nature of early coordination:evidence from bilateral leg movementsin young infants. Dev. Psychol. 23:179–86

ThelenE, Smith L, eds.1993. Dynamical Sys-temsin Development. Vol. 2: Applications.Cambridge,MA: Bradford Books

Thelen E, Ulrich BD. 1991. Hidden skills: adynamic systemsanalysisof treadmill step-ping during the first year. Monogr. Soc.Res.Child Dev.(Ser.No. 223) 56:1

TrevarthenC. 1984. How control of movementdevelops. In Human Motor Actions: Bern-stein Reassessed,ed. HTA Whiting, pp.223–61. Amsterdam:Elsevier

Turvey MT. 1990. Coordination. Am.Psychol.4:938–53

Ulrich BD, JensenJL, Thelen E. 1991. Stabil -ity andvariation in the development of in-fant stepping: implications for control. InAdaptabil ity of Human Gait, ed.AE Patla,pp. 145–64. Amsterdam:Elsevier

Ungerleider LG, Mishkin M. 1982. Two cor-tical visual systems.In Analysis of VisualBehavior, ed. DJIngle, MA Goodale,RJW

458 BERTENTHAL

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

Mansfield, pp. 549–86. Cambridge, MA:MIT Press

von HofstenC. 1982. Eye-handcoordination innewborns.Dev.Psychol. 18:450–61

von Hofsten C. 1983. Catching skills in in-fancy. J. Exp. Psychol.: Hum. Percept.Perform. 9:75–85

von Hofsten C. 1993. Prospective control: abasicaspectof action development. Hum.Dev.36:253–70

von HofstenC, RonnqvistL. 1988. Preparationfor grasping an object: a developmentalstudy. J. Exp. Psychol.: Hum. Percept.Perform. 14:610–21

Walton GE, Bower NJA, Bower TGR. 1992.Recognition of familiar facesby newborns.Infant Behav.Dev.15:265–69

Warren W. 1990. The perception-action cou-pling. In Sensory-Motor Organization andDevelopment in Infancy and Early Child-hood, ed. H Bloch, BI Bertenthal, pp.23–38. Dordrecht: Kluwer

Wellman HM, GelmanSA. 1992. Cognitivedevelopment: foundational theoriesof coredomains.Annu. Rev.Psychol. 43:337–75

Weiskrantz L, Warrington E, Sander MD,Marshall J. 1974. Visual capacity in thehemianopic field following restricted oc-cipital ablation. Brain 97:709–29

Wentworth N, Haith MM. 1992. Event-specificexpectations of 2- and3-month-old infants.Dev. Psychol. 28:842–50

White B, Castle P, HeldR. 1964. Observationson the development of visually -directedreaching. Child Dev.35:349–64

Woollacott MH, Sveistrup H. 1994. The devel-opmentof sensorimotor integration under-lying posture control in infants during thetransition to independent stance.In Inter-limbCoordination: Neural, Dynamical andCognitive Constraints, ed. SP Swinnen, JMassion, H Heuer,pp. 371–89. San Diego:Academic

Xu F, Carey S. 1995. Infants’ metaphysics:thecaseof numerical identity. Cogn. Psychol.In press

YonasA, HartmanB. 1993. Perceiving the af-fordance of contact in 4- and5-month-oldinfants.Child Dev.64:298–308

Yonas A, Owsley C. 1987. Development ofvisualspaceperception. In Handbook of In-fant Perception. Vol. 1: From Sensation toPerception, ed. P Salapatek,L Cohen,pp.80–122. Orlando: Academic

ZelazoPR,Brody LR, ChaikanH. 1984. Neo-natal habituation anddishabituationof headturning to rattle sounds. Infant Behav.Dev.7:311–21

ZelazoPR 1984. “Learning to walk”: recogni-tion of higher-order influences?Adv. In-fancyRes. 3:251–60

Zerni cke RF, Schneider K . 1993.Biomechanics and developmental neuro-motor control. Child Dev.64:982–1004

PERCEPTION, ACTION, ANDREPRESENTATION 459

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.

Annual Review of Psychology Volume 47, 1996

CONTENTSASPECTS OF THE SEARCH FOR NEURAL MECHANISMS OF MEMORY, Mark R. Rosenzweig 1 THEORETICAL FOUNDATIONS OF COGNITIVE-BEHAVIOR THERAPY FOR ANXIETY AND DEPRESSION, Chris R. Brewin 33THE DESIGN AND ANALYSIS OF SOCIAL-INTERACTION RESEARCH, David A. Kenny 59PERSONALITY: Individual Differences and Clinical Assessment, James N. Butcher, Steven V. Rouse 87HEALTH PSYCHOLOGY: Psychological Factors and Physical Disease from the Perspective of Human Psychoneuroimmunology, Sheldon Cohen, Tracy B. Herbert 113VERBAL LEARNING AND MEMORY: Does the Modal Model Still Work?, Alice F. Healy, Danielle S. McNamara 143LONG-TERM POTENTIATION AND LEARNING, Joe L. Martinez Jr., Brian E. Derrick 173CROSS-CULTURAL SOCIAL AND ORGANIZATIONAL PSYCHOLOGY, Michael Harris Bond, Peter B. Smith 205STEREOTYPES, James L. Hilton, William von Hippel 237

EXPERT AND EXCEPTIONAL PERFORMANCE: Evidence of Maximal Adaptation to Task Constraints, K. A. Ericsson, A. C. Lehmann 273TEAMS IN ORGANIZATIONS: Recent Research on Performance and Effectiveness, Richard A. Guzzo, Marcus W. Dickson 307PSYCHOLOGY IN CANADA, J. G. Adair, A. Paivio, P. Ritchie 341METHODOLOGICAL ISSUES IN PSYCHOPATHOLOGY RESEARCH, K. J. Sher, T. J. Trull 371THE SOCIAL STRUCTURE OF SCHOOLING, Sanford M. Dornbusch, Kristan L. Glasgow, I-Chun Lin 401ORIGINS AND EARLY DEVELOPMENT OF PERCEPTION, ACTION, AND REPRESENTATION, Bennett I. Bertenthal 431AUDITORY PSYCHOPHYSICS AND PERCEPTION, Ira J. Hirsh, Charles S. Watson 461ENVIRONMENTAL PSYCHOLOGY 1989–1994, Eric Sundstrom, Paul A. Bell, Paul L. Busby, Cheryl Asmus 485COGNITIVE SKILL ACQUISITION, Kurt VanLehn 513ATTACHMENT AND SEPARATION IN YOUNG CHILDREN, Tiffany Field 541COVARIANCE STRUCTURE ANALYSIS: Statistical Practice, Theory, and Directions, Peter M. Bentler, Paul Dudgeon 563THE MOTIVATIONAL IMPACT OF TEMPORAL FOCUS: Thinking About the Future and the Past, Rachel Karniol, Michael Ross 593

Ann

u. R

ev. P

sych

ol. 1

996.

47:4

31-4

59. D

ownl

oade

d fr

om a

rjou

rnal

s.an

nual

revi

ews.

org

by C

OL

LE

GE

DE

FR

AN

CE

on

12/2

6/06

. For

per

sona

l use

onl

y.