object permanence in young infants: further...

Object Permanence in Young Infants:Further Evidence

Ren^e Baillargeon and Julie DeVosUniversity of Illinois

BAILLARGEON, RENfiE, AND DEVOS, JULIE. Object Permanence in Young Infants: Further Evi-dence. CHILD DEVELOPMENT, 1991, 62, 1227-1246. Recent evidence suggests that 4.5- and even3.5-month-old infants realize that objects continue to exist when hidden. The goal of the presentexperiments was to obtain converging evidence of object permanence in young iniants. Experi-ments were conducted using paradigms previously used to d«nonstrate object permanence in5.5-month-old infants and 6.5-month-old infants. In one experiment, 3.5-month-old infantswatched a short or a tall carrot slide along a track. The track's center was hidden by a screenwith a large window in its upper half. The short carrot was shorter than the window's lower edgeand so did not appear in the window when passing behind the screen; the tall carrot was tallerthan the window's lower edge and hence should have appeared in the window but did not. Theinfants looked reliably longer at the tall than at the short carrot event, suggesting that they(a) represented the existence, height, and trajectory of each carrot behind the screen and(b) expected the tall carrot to appear in the screen window and were surprised that it did not.Gontrol trials supported this interpretation. In another experiment, 4.0-month-old infants saw atoy car roll along a track that was partly hidden by a screen. A large toy mouse was placed be-hind the screen, either on top or in back of the track. The female infants looked reliably longerwhen the mouse stood on top as opposed to in back of the track, suggesting that they (a) repre-sented the existence and trajectory of the car behind the screen, (b) represented the existenceand location of the mouse behind the screen, and (c) were surprised to see the car reappearfrom behind the screen when the mouse stood in its path. A second experiment supported thisinterpretation. The results of these experiments provide further evidence that infants aged 3.5months and older are able to represent and to reason about hidden objects.

When adults see an object occlude an- til about 9 months of age, Piaget maintained,other object, they typically make three as- infants do not understand that objects con-sumptions. The first is that the occluded ob- tinue to exist when occluded. They believeject continues to exist behind the occluding that objects cease to exist when they ceaseobject. The second is that the occluded ob- to be visible and begin to exist anew whenject retains the physical and spatial proper- they become visible again. At about 9ties it possessed prior to occlusion. Finally, months of age, infants begin to view objectsthe third is that the occluded object is still as permanent entities that continue to existsubject to physical laws: its displacements when masked by other objects. However,and interactions with other objects do not this permanence is still limited. Infants dobecome capricious or arbitrary but remain not yet conceive of occluded objects as occu-regular and predictable. When do infants pying objective locations in space. Rather,begin to share these assumptions? Piaget they tend to confer on occluded objects "a(1954) was the first to address this question, sort of absolute position" (p. 46), the firstDetailed analyses of infants' performance in place in which they were found. It is notmanual search tasks led him to conclude that until about 12 months of age, Piaget held,infants' beliefs about occluded objects de- that infants begin to attend systematically tovelop slowly over the course of infancy. Un- visible displacements and assume that oc-

This research was supported by a grant from the National Institute of Child Health andHuman Development to Rende Baillargeon (PHS 5-PO1-HD-05951-17). We thank Wendy Greenfor her help with the design and construction of the test stimuli; Stephanie Lyons-OIsen, Eliza-beth Mazenko, and Amy Needham for their help with the data analyses; and the undergraduatestudents working in the Infant Gognition Laboratory at the University of Illinois for their helpwith the data collection. Finally, we thank the parents who kindly agreed to have their infantsparticipate in the studies. Gorrespondence conceming this article should be addressed to: Ren^eBaillargeon, Department of Psychology, University of Illinois, 603 E. Daniel, Champaign, IL61820.

IChild Development, 1991,62,1227-1246. © 1991 by the Society for Research in Child Development, Inc.All rights reserved. 0009-3920/91/6206-0016$01.00]

1228 Child Development

eluded objects reside in whatever locationsthey occupied immediately prior to occlu-sion. The final advance in the developmentof infants' beliefs about occluded objects fol-lows the emergence of symbolic representa-tion, at about 18 months of age. Because oftheir new representational capacity, infantsbecome able to imagine invisible displace-ments and hence to infer occluded objects'locations. According to Piaget, objects' ap-pearances and disappearances are then nolonger mysterious but follow known, pre-dictable patterns. Occluded objects are un-derstood to be subservient to the same spa-tial and kinematic laws as visible objects.

Over the past 2 decades, many investi-gators have tested Piaget's description of thedevelopment of infants' beliefs about oc-cluded objects (see Baillargeon, in press-a,in press-b; Bremner, 1985; Gratch, 1975,1976; Harris, 1987, 1989; Schuberth, 1983;Sophian, 1984; and Wellman, Cross, &Bartsch, 1987, for reviews). In these tests,investigators have used Piaget's manualsearch tasks as well as novel visual tasks.The introduction of visual tasks stemmedfrom a concern that young infants might per-form poorly on search tasks, not becausetheir concept of object permanence is lack-ing or incomplete, but because their abilityto plan means-end search sequences is lim-ited. The results obtained with visual taskshave substantiated this concern: they indi-cate that Piaget seriously underestimatedyoung infants' understanding of occlusionevents (e.g., Baillargeon, 1986, 1987a,1987b, 1991, in press-a, in press-b; Bail-largeon, DeVos, & Graber, 1989; Baillargeon& Graber, 1987, 1988; Baillargeon, Graber,DeVos, & Black, 1990; Baillargeon, Spelke,& Wasserman, 1985; Spelke, 1988).

One experiment, for example, revealedthat infants as young as 6.5 months of ageare able to reason about the existence, loca-tion, and trajectory of occluded objects (Bail-largeon, 1986). The infants sat in front of asmall screen; to the left of this screen was along, inclined ramp. The infants were habit-uated to the following event: the screen wasraised (to show the infants that there wasnothing behind it) and lowered, and a toycar rolled down the ramp, passed behind thescreen, and exited the apparatus to the right.Following habituation, the infants saw twotest events. These were identical to the ha-bituation event except that a box was placedbehind the screen. This box was revealedwhen the screen was raised. In one event(possible event), the box stood in back of the

car's tracks; in the other event (impossibleevent), the box stood on top of the car'stracks, blocking its path. The results indi-cated that the infants looked reliably longerat the impossible than at the possible event,suggesting that they were surprised to seethe car reappear from behind the screenwhen the box stood in its path. A secondcondition in which the box was hidden infront (possible event) or on top (impossibleevent) of the car's tracks yielded the sameresults. Together, the results of these twoconditions indicated that the infants (a) real-ized that the box and the car continued toexist behind the screen, (b) believed that thebox retained its location behind the screen,(c) assumed that the car pursued its trajec-tory behind the screen, (d) understood thatthe car could not roll through the space occu-pied by the box, and hence (e) were sur-prised in the impossible event to see the carreappear from behind the screen.

Another experiment provided evidencethat infants as young as 5.5 months of ageare able to reason about the existence,height, and trajectory of occluded objects(Baillargeon & Graber, 1987). The infantswere habituated to a toy rabbit sliding backand forth along a horizontal track whose cen-ter was occluded by a screen. On alternatetrials, the infants saw a short or a tall rabbitslide along the track. Following habituation,the midsection of the screen's upper halfwas removed, creating a large window. Theinfants saw two test events. In one (possibleevent), the short rabbit moved back and forthalong the track; this rabbit was shorter thanthe window's lower edge and so did not ap-pear in the window when passing behindthe screen. In the other event (impossibleevent), the tall rabbit moved back and forthalong the track; this rabbit was taller thanthe window's lower edge and hence shouldhave appeared in the window but did not.The results indicated that the infants lookedequally at the short and the tall rabbit habit-uation events but looked reliably longer atthe impossible than at the possible testevent, suggesting that they (a) realized thateach rabbit continued to exist behind thescreen, (fo) believed that each rabbit retainedits height behind the screen, (c) assumedthat each rabbit pursued its trajectory be-hind the screen, and hence (d) expected thetall rabbit to appear in the screen windowand were surprised that it did not. This inter-pretation was supported by the results of asecond condition that was identical to theexperimental condition with one important

Baillargeon and DeVos 1229

exception: prior to the habituation trials, theinfants received two pretest trials in whichthey saw two short or two tall rabbits stand-ing motionless on either side of the win-dowless habituation screen. Half of the in-fants saw the two short rabbits in the firsttrial and the two tall rabbits in the secondtrial; the other infants saw the rabbits in theopposite order. Unlike the infants in the ex-perimental condition, the infants in this pre-tests condition looked equally at the impos-sible and the possible events. These resultssuggested that the infants were able to usethe information presented in the pretest tri-als to make sense of the impossible event.Specifically, the infants understood that thetall rabbit did not appear in the screen win-dow because it did not in fact travel the dis-tance behind the screen: instead, one rabbittraveled along the left half of the track andanother rabbit along the right half.

Yet another experiment provided evi-dence that 4.5- and even 3.5-month-old in-fants are able to reason about the existenceof an occluded object (Baillargeon, 1987a).The infants were habituated to a screen thatrotated back and forth through a 180° arc,in the manner of a drawbridge. Followinghabituation, a box was placed behind thescreen, and the infants saw two test events.In one (possible event), the screen rotateduntil it reached the occluded box; in theother (impossible event), the screen rotatedthrough a full 180° arc, as though the boxwere no longer behind it. The results indi-cated that the 4.5-month-oId infants, and the3.S-month-oId infants who were fast habitua-tprs, looked reliably longer at the impossiblethan at the possible event, suggesting thatthey (a) believed that the box continued toexist behind the screen, (b) understood thatthe screen could not rotate through the spaceoccupied by the box, and hence (c) expectedthe screen to stop in the impossible eventand were surprised that it did not. Supportfor this interpretation came from a controlcondition that was identical to the experi-mental condition except that no box wasplaced behind the screen. Unlike the infantsin the experimental condition, the infants inthis control condition tended to look equallyat the shorter (112° arc) and the longer (180°arc) screen rotations. Together, these resultsindicated that infants as young as 3.5 monthsof age are aware that objects continue to ex-ist when occluded.

The goal of the present research wastwofold. The first goal was to obtain converg-ing evidence for Baillargeon's (1987a) con-

clusion that 3.5-month-old infants are able torepresent and to reason about the existenceof occluded objects. The second goal was todetermine whether young infants can repre-sent and reason about not only the existencebut also some of the properties—such as theheight, location, and trajectory—of occludedobjects. Subjects were 3 to 4 months of age.Infants were tested with either the slidingrabbit task Baillargeon and Graber (1987)used successfully with 5.5-month-old infants(Experiments 1 and 2), or with the rollingcar task Baillargeon (1986) used successfullywith 6.5-month-old infants (Experiments 3,3A, and 4). We reasoned that evidence thatyoung infants performed successfully inthese two tasks would have important impli-cations for models of the development of ob-ject permanence in infancy.

Experiment 1Subjects in Experiment 1 were 3.5-

month-old infants. The method of this exper-iment was similar to that used by Baillju'-geon and Graber (1987) with 5.5-month-oldinfants. The only departure from the de-scription given above was that carrots wereused instead of rabbits (see Fig. 1). In pilotwork, some infants were found to be scaredof the rabbits so they were replaced withless threatening-looking carrots.

METHOD

SubjectsSubjects were 32 healthy, full-term in-

fants ranging in age from 3 months, 5 daysto 3 months, 27 days (M = 3 months, 15days). An additional 19 infants were ex-cluded from the experiment because theyfailed to complete at least two pairs of testtrials (see below), 13 due to fussiness, 3 dueto drowsiness, and 3 due to procedural error.The infants' names in this experiment andin the following experiments were obtainedfrom birth announcements in the local news-paper. Parents were contacted by letters andfollow-up phone calls. They were offered re-imbursement for their travel expenses butwere not compensated for their partici-pation.

Half of the infants were assigned to theexperimental condition (M = 3 months, 18days) and half to the pretests condition(Af = 3 months, 12 days).Apparatus

The apparatus consisted of a largewooden box 180 cm high, 136 cm wide, and66 cm deep. The infant faced an opening 47


Habituation Events

Short Carrot Event Tall Carrot Event

Test Events

Possible Event Impossible Event

FIG. 1.—Schematic representation of the habituation and test events shown to the infants in theexperimental condition in Experiment 1.

cm high and 108 cm wide in the front wallof the apparatus. The back and side wallsof the apparatus were covered with colorfulcontact paper; the floor was painted black.

In the floor of the apparatus, parallel tothe back wall and centered between the sidewalls, was a narrow track 126 cm long. Twocarriers moved back and forth along thistrack, one along the left half and the otheralong the right half. Each carrier consistedof a styrofoam strip 9.5 cm high, 2 cm wide,and 1.5 cm thick. Inserted into the strip wasa metal rod 9 cm high and 0.5 cm in diame-ter. The lower portion of this rod wasattached nndemeath the floor of the appara-tus to a cubical metal base 2 cm a side thatslid along a metal guide rod 136 cm long and0.75 cm in diameter. The base of each carrierwas connected by a thin cable to a pulleyand balance weight system on the side of theapparatus (left side for the left carrier andright side for the right carrier). Lowering thebalance weight of a carrier down the side ofthe apparatus caused the carrier to slide fromthe center of the apparatus toward the sidewall; conversely, raising the balance weightof a carrier caused it to slide back toward thecenter of the apparatus. To help the experi-menters raise and lower the balance weightsof the carriers at an even pace, a column ofequally spaced marks was placed on eachside of the apparatus; in addition, the experi-menters listened through headphones to ametronome beating once per second.

Identical tall or short flat carrots wereplaced on the left and right carriers. Thesecarrots were made of thick orange cardboardand were decorated with small, black-inkedfeatures and green cardboard bow ties; theyalso had leaves at their upper, larger endsmade of green cardboard and decorated withgreen pom-poms. The tall carrots were 27cm high, 0.6 cm thick, and 6.25 cm wide attheir widest point; the short carrots were 15cm high, 0.6 cm thick, and 6.25 cm wide attheir widest point. Because the carriersstood 0.5 cm above the floor of the apparatus,the tall and the short carrots' total heightswere 27.5 and 15.5 cm, respectively. Stripsof Velcro were glued to the back of the car-rots to attach them to the carriers.

Centered between the side walls, at adistance of 5.5 cm from the track and 29.5 cmfrom the back wall, was a three-sided metalframe consisting of two vertical bars, each30.5 cm high and 2.5 cm wide, standing 26cm apart, and connected at their base by ametal bar 1.5 cm high and 33.5 cm long. Acardboard screen could be attached to thevertical bars by strips of Velcro. Two screenswere used in the experiment: a yellowscreen 32 cm high and 42 cm wide and ablue screen also 32 cm high and 42 cm widebut with a window 16 cm high and 21 cmwide in the center of its upper half.

The infant was tested in a brightly litroom. Four clip-on lights (each with a


40-watt lightbulb) were attached to the backand side walls of the apparatus to provideadditional light. These lights were arrangedso as to prevent tell-tale shadows. Twoframes, each 180 cm high and 60 cm wideand covered with blue cloth, stood at anangle on either side of the apparatus. Theseframes served to isolate the infant from theexperimental room. At the end of each trial,a curtain consisting of a muslin-coveredframe 52 cm high and 108 cm wide was low-ered in front of the opening in the front wallof the apparatus.

EventsTwo experimenters worked in concert

to produce the events; the flrst operated theleft carrier and the second operated the rightcarrier.

Tall carrot habituation event.—In thetall carrot habituation event, the windowlessyellow screen occluded the center of thetrack and the tall carrots stood on the left andright carriers.

At the start of the trial, the carrot placedon the left carrier stood visible at the leftend of the track; the carrot placed on theright carrier stood just inside the right edgeof the screen, hidden from the infant. Aftera 1-sec pause, the flrst experimenter slid theleft carrot at the speed of about 21 cm/secuntil it had slid 42 cm and stood just insidethe left edge of the screen, hidden from theinfant. After a 2-sec pause, the second exper-imenter slid the right carrot at the samespeed of about 21 cm/sec until it had slid 42cm and stood at the right end of the track.After a 1-sec pause, the entire process wasrepeated in reverse. The second experi-menter returned the right carrot to its start-ing position behind the screen's right edge;the flrst experimenter waited 2 sec and thenslid the left carrot from behind the screen'sleft edge back to its starting position at theleft end of the track. Each event cycle thuslasted about 14 sec. Cycles were repeateduntil the computer signaled that the trial hadended (see below). When this occurred, thesecond experimenter lowered the curtain infront of the apparatus.

Short carrot habituation event.—Theshort carrot habituation event was identicalto the tall carrot habituation event exceptthat the short carrots were substituted for thetall carrots on the carriers.

Impossible and possible test events.—The impossible and the possible testevents were identical to the tall and theshort carrot habituation events, respectively.

except that the windowless yellow screenwas replaced by the blue screen with thelarge window. We hoped that the change inthe screen's color would draw the infants'attention to the screen, thus making themmore likely to notice the presence of thewindow.

ProcedurePrior to the experiment, each infant was

allowed to manipulate a tall and a short car-rot for a few minutes while his or her parentfllled out consent forms. During the experi-ment, the infant sat on the parent's lap infront of the apparatus, facing the screen. Theinfant's head was approximately 66 cm fromthe screen and 96 cm from the back wall ofthe apparatus. The parent was asked not tointeract with the infant during the experi-ment and to close his or her eyes during thetest trials.

The infant's looking behavior was moni-tored by two observers who viewed the in-fant through peepholes in the cloth-coveredframes on eidier side of the apparatus. Theobservers could not see the events from theirviewpoints and they did not know the orderin which the events were presented. Eachobserver held a button box linked to a MI-CRO/PDP-11 computer and depressed thebutton when the infant attended to theevents. Each trial was divided into l(K)-msecintervals, and the computer determined ineach interval whether the two observersagreed on the direction of the infant's gaze.Interobserver agreement was calculated foreach trial on the basis of the number of in-tervals in which the computer registeredagreement, out of the total number of inter-vals in the trial. Agreement in this experi-ment and in the following experiments aver-aged 91% or more per trial per infant. Thelooking times recorded by the primary ob-server were used to determine when a trialhad ended (see below).

The infants in the experimental condi-tion participated in a two-phase procedureconsisting of a habituation phase and a testphase. During the habituation phase, the in-fants saw the tall and the short carrot habitu-ation events described above on alternatetrials. These trials served two purposes: theyserved to acquaint the infants with the car-rots and their trajectories, and they made itpossible to assess whether the infants foundthe tall carrot intrinsically more interestingthan the short carrot. Each trial ended whenthe infant (a) looked away from the event for2 consecutive sec after having looked at itfor at least 6 cumulative sec or (b) looked


at the event for 60 cumulative sec withoutlooking away for 2 consecutive sec. Habitua-tion trials continued until the infant (a) satis-fled a criterion of habituation of a 50% orgreater decrease in looking time on threeconsecutive trials, relative to the infant'slooking time on the flrst three trials, or (fo)completed nine habituation trials. There-fore, the minimum number of habituationtrials an infant could receive was six, and themaximum number was nine. During the testphase, the infants saw the impossible andthe possible test events described above onalternate trials until they had completedthree pairs of test trials.' The criteria usedto determine the end of each test trial werethe same as for the habituation trials. The6-sec minimum value was chosen to ensurethat the infants had sufficient information todistinguish between the impossible and thepossible test events. Half of the infants sawthe habituation and test events with the tallcarrot flrst; the other infants saw the habitua-tion and test events with the short carrotflrst.

The infants in the pretests conditionparticipated in a three-phase procedurecomprising a pretest phase, a habituationphase, and a test phase. The habituation andtest phases were identical to those in the ex-perimental condition. During the pretestphase, the infants received a trial in whichthey saw two tall carrots standing mo-tionless on either side of the windowless ha-bituation screen, and a trial in which theysaw two short carrots standing motionlesson either side of the same screen. The car-rots were positioned about halfway betweenthe edges of the screen and the ends of thetrack. Analysis of the infants' looking timesduring these trials revealed no preferencefor the tall (M = 22.4) over the short (M =23.8) carrots, F(l,15) = 0.46. Half of the in-fants saw the pretest, habituation, and testevents with the tall carrot(s) flrst, and halfwith the short carrot(s) flrst.

Of the 32 infants in the experiment, 18completed nine habituation trials withoutsatisfying the criterion of habituation; theother infants took an average of 6.57 trialsto reach the criterion. Four infants failed tocomplete the full complement of three pairsof test trials. These infants completed onlytwo pairs, three because of fussiness and one

because of drowsiness. All subjects (in thisexperiment as well as in the subsequent ex-periments) were included in the data analy-ses, whether or not they had completed allthree pairs of test trials. Preliminary analy-ses revealed no significant effect of order orsex on the infants' looking times at the im-possible and the possible events during thethree pairs of test trials, all F's < 1.61,p > .05. The data were therefore collapsedin subsequent analyses.

RESULTS

Figure 2 shows the mean looking timesof the infants in the experimental and thepretests conditions during the last threepairs of habituation trials and the three pairsof test trials. The infants' looking times dur-ing these trials were analyzed by means of a2 x 2 x 3 x 2 mixed-model analysis of vari-ance with condition (experimental or pre-tests condition) as the between-subjects fac-tor and with block (habituation or test trials),pair (flrst, second, or third pair of trials), andevent (tall carrot/impossible or short carrot/possible event) as the within-subjects fac-tors. Recause the design was unbalanced,the SAS GLM procedure was used to calcu-late the analysis of variance (SAS Institute,1985). There was a signiflcant main effectof event, F(l,176) = 3.97, p < .05, and asigniflcant condition x block x event inter-action, F(l,176) = 3.88, p = .05. Plannedcomparisons revealed that the infants inthe experimental condition looked aboutequally at the tall (M = 32.9) and the short(M = 32.4) carrot habituation events,F(l,176) = 0.03, but looked reliably longerat the impossible (M = 34.1) than at the pos-sible (M = 26.0) test event, F(l,176) = 5.69,p < .02. In contrast, no reliable differencewas found between the looking times of theinfants in the pretests condition at the tall(M = 34.5) and the short (M = 29.2) carrothabituation events, F(l,176) = 2.53, p > .05,or at the impossible (M = 29.9) and the pos-sible (M = 30.7) test events, F(l,176) =0.05.

The initial analysis of variance also re-vealed a significant main effect of pair,F(2,176) = 19.49, p < .0001, indicating thatthe infants looked reliably less as the habitu-ation and test phases of the experiment pro-gressed.

' Baillargeon and Graber (1987) gave their subjects four pairs of test trials but were unableto use the data from the fourth test pair because many of their subjects were fussy on that pair.We therefore decided to use only three test pairs in our experiment.

BaiUargeon and DeVos 1233

Experimental Condition

-3 -2 -1Habituation Trials Test Trials

FIG. 2.—Mean looking times of the infants in the experimental and the pretests conditions inExperiment 1 at the habituation and test events.

DISCUSSION

The infants in the experimental condi-tion tended to look equally at the tall and theshort carrot habituation events, but lookedreliably longer at the impossible than at thepossible test event. These results indicatethat the infants (a) realized that each carrotcontinued to exist after it slid behind thescreen, (fo) assumed that each carrot retainedits height behind the screen, (c) believedthat each carrot pursued its trajectory behindthe screen, and therefore (d) expected thetall carrot to be visible in the screen windowand were surprised that it was not. Theseresults confirm Raillargeon's (1987a) conclu-sion that infants as young as 3.5 months ofage are aware that objects continue to existwhen occluded. In addition, the results ex-tend Baillargeon's (1987a) conclusion in thatthey indicate that 3.5-month-old infants canrepresent and reason about not only the exis-tence but also the height and trajectory ofoccluded objects.

In contrast to the infants in the experi-mental condition, the infants in the pretestscondition looked about equally at the impos-sible and the possible test events. At leasttwo interpretations can be offered for thisfinding. One interpretation is that the infantsmade use of the information conveyed in thepretest trials to generate an explanation forthe impossible event. Specifically, the in-fants understood that the tall carrot did notappear in the screen window in the impossi-ble event because the tall carrot did not infact travel the distance behind the screen.Instead, two separate carrots traveled alongthe track: one carrot traveled from the leftend of the track to the left edge of the screenand stopped just inside this edge; a second,identical carrot then emerged from behindthe right edge of the screen and traveled tothe right end of the track. This interpreta-tion, if valid, would provide strong addi-tional support for the conclusion that 3.5-month-old infants are able to reason about


the existence as well as the trajectory of oc-cluded objects.

The other, less impressive interpreta-tion for the results of the pretests conditionis that the pretest trials biased the infants toattend to the two sides of the screen simulta-neously, causing them to scan the habitua-tion and the test events inappropriately. Ac-cording to this explanation, the infants in thepretests condition looked equally at the im-possible and the possible events, not be-cause they realized that two carrots wereused to produce each event, but becausethey were too confused to distinguish be-tween the events. There are, however, sev-eral reasons to doubt this alternative inter-pretation. First, if the infants had beenconfused when watching the habituationevents, one might have expected this confu-sion to have had some effect on their lookingbehavior. Yet statistical analyses revealed nodifferences (a) between the number of habit-uation trials completed by the infants in theexperimental (M = 7.8) and the pretests (M= 8.1) conditions, F(l,28) = 0.60; (fo) be-tween the total looking time during the ha-bituation trials of the infants in the experi-mental (M = 247.3) and the pretests (M =240.4) conditions, F(l,28) = 0.03; (c) be-tween the mean looking time during the ha-bituation trials of the infants in the experi-mental (M = 33.5) and the pretests (M =31.1) conditions, F(l,28) = 0.30; (cf) be-tween the mean looking times during thefirst six habituation trials of the infants in theexperimental (M = 34.6) and the pretests(M = 31.5) conditions, F(l,28) = 0.38; andfinally (e) between the mean looking timeduring the last six habituation trials of theinfants in the experimental (M = 32.6) andthe pretests (M = 31.9) conditions, F(l,28)= 0.04. The marked similarity between thehabituation pattems of the infants in the ex-perimental and the pretests conditions is in-consistent with the notion that one grouphad a straightforward interpretation of theevents, while the other group was hope-lessly confused by them. Second, even if theinfants in the pretests condition had experi-enced some initial confusion left undetectedby the aforementioned analyses, it is diffi-cult to believe that the infants would still

have been confused by the events at the endof the habituation phase, after witnessingthe carrots slide along the track over 30times on the average (if the infants hadwatched the habituation events for a total of240.4 consecutive sec across trials, theywould have seen the carrots move from oneend of the track to the other 34.3 times, sinceeach half cycle lasted 7 sec). It seems highlyunlikely that the infants would have failedafter 30 observations to appreciate the sim-ple translation pattems they were shown, es-pecially in light of Haith's (in press) findingsthat infants this age can readily detect farmore complex event regularities. A final rea-son for doubting the hypothesis that the in-fants in the pretests condition scanned thehabituation and test events inappropriatelyis that it is inconsistent with observers' de-scriptions of the infants' performance. Ob-servers reported that most infants in the ex-perimental and the pretests conditionsrapidly settled into following each carrotfrom left to right and right to left across theapparatus. Indeed, parents often mistakenlyassumed that the true goal of the experimentwas to establish how quickly their infantsengaged in this following pattern.

In light of these arguments, it seemslikely that the infants in the pretests condi-tion looked equally at the impossible andthe possible test events, not because theynever fiilly perceived these events and socould not distinguish between them, but be-cause they understood that two carrots wereused to produce each event. In the absenceof more direct supportive evidence,^ how-ever, this conclusion must remain tentative.

Experiment 2Would results similar to those of Experi-

ment 1 be obtained with infants less than3,5 months of age? To answer this question,3.0-month-old infants were tested in Experi-ment 2 using the same method as in Experi-ment 1.

METHOD


fants ranging in age from 2 months, 14 days

^ Such evidence could take several forms, such as (a) videotape data indicating that thelooking behavior of infants in the experimental and the pretest conditions did not differ reliably,or (b) looking time data indicating that infants still showed a reliable preference for the impossi-ble event when given two pretest trials with a tall and a short carrot standing motionless oneither side of the screen (in this case, infants would no longer be given a "clue" as to how theimpossible event was produced but would presumably still be biased to attend to the two sidesof the screen simultaneously).


to 3 months, 4 days (M = 2 months, 26 days).An additional 21 infants were eliminatedfrom the experiment because they failedto complete at least two test pairs, 15 dueto fussiness, 4 due to drowsiness, 1 due toequipment failure, and 1 due to the primaryobserver's inability to follow the direction ofthe infant's gaze. Sixteen infants were as-signed to the experimental condition (M =2 months, 27 days) and 13 to the pretestscondition (M = 2 months, 24 days).

Apparatus, Events, and ProcedureThe apparatus, events, and procedure

used in Experiment 2 were identical to thosein Experiment 1. Analysis of the lookingtimes of the infants in the pretests conditionduring the two pretest trials revealed nopreference for the tall (M = 22.1) over theshort (M = 23.2) carrots, F(l,12) = 0.48. Ofthe 29 infants in the experimental and thepretests conditions, 13 failed to satisfy the

habituation criterion within nine trials; theother infants took an average of 7.44 trials toreach the criterion. Four infants completedonly two pairs of test trials, one because offussiness, one because of equipment failure,and two because the primary observer couldnot follow the direction of their gaze. Pre-liminary analyses revealed no significant ef-fect of order or sex on the infants' lookingtimes at the impossible and the possibleevents during the three pairs of test trials, allF's < 2.04, p > .05. The data were thereforecollapsed in subsequent analyses.

RESULTS

Figure 3 shows the mean looking timesof the infants in the experimental and thepretests conditions during the last threepairs of habituation trials and the three pairsof test trials. The infants' looking times wereanalyzed as in Experiment 1. The main ef-

Experimental Condition

a>EF

'̂oo_Jc(C0

50

45

40

35

30

25

20

15

10

50

45

40

35

30

25

20

15

10

Tall CarrotShort Carrot

Control Condition

Both Conditions

Habituation Trials Test TrialsFIG. 3.—Mean looking times of the infants in the experimental and the pretests conditions in

Experiment 2 at the habituation and test events.


fects of condition, F(l,27) = 0.01, and event,F(l,208) = 3.18, p > .05, were not signifi-cant, nor were any of the interactions involv-ing these factors, all F's < 2.42, p > .05.There were thus no reliable differences be-tween the looking times of the infants in theexperimental and the pretests conditions atthe tall and the short carrot habituationevents or at the impossible and the possibletest events.

DISCUSSION

Unlike the 3.5-month-oId infants in theexperimental condition in Experiment 1, the3.0-month-old infants in the experimentalcondition in Experiment 2 did not show areliable preference for the impossible overthe possible test event As is often the casewith negative results, this finding is open toseveral different interpretations. To illus-trate, the infants' failure to show a reliablepreference for the impossible event couldbe taken to suggest that the infants lacked anotion of object permanence and so did notrealize that each carrot continued to exist,retained its height, and pursued its trajectorybehind the screen. Altematively, it could beproposed that the infants possessed a notionof object permanence but were preventedfrom detecting the violation embedded inthe impossible event by procedural limita-tions. One such limitation might have beenthat the infants could perceive the surprisingaspect of the impossible event only if theylooked at the screen window during the 1sec (per half cycle) the tall carrot was to haveappeared there. Further research is neces-sary to determine which, if either, of theseaccounts best explains the poor performanceof the infants in the experimental conditionin Experiment 2.

Like the 3.5-month-oId infants in thepretests condition in Experiment 1, the 3.0-month-old infants in the pretests conditionin Experiment 2 showed no reliable prefer-ence for the impossible over the possibleevent. Because the 3.0-month-old infants inthe experimental condition in Experiment 2also failed to look reliably longer at the im-possible than at the possible event, how-ever, no clear conclusions can be drawnabout the results of the pretests condition.

Experiment 3

The results of Experiment 1 indicatedthat a paradigm used to demonstrate 5.5-month-old infants' ability to reason about theexistence and properties of occluded objects(Baillargeon & Graber, 1987) could also be

used successfully with 3.5-month-old in-fants. Civen this finding, we were encour-aged to ask whether another paradigm, firstused to reveal 6.5-month-old infants' under-standing of occlusion events (Baillargeon,1986), would also yield positive results withyounger infants.

Subjects in Experiment 3 were 4.0-month-old infants. The method of this exper-iment was similar to that used by Baillar-geon (1986). The only departure from thedescription given in the introduction wasthat a large toy mouse, rather than a box, wasplaced on top or in back of the car's path (seeFig. 4).

METHOD


fants ranging in age from 3 months, 23 daysto 4 months, 13 days (M = 3 months, 29days). Four additional infants were excludedfrom the experiment because they failed tocomplete at least two pairs of test trials, twodue to fussiness and two due to proceduralproblems.

ApparatusThe apparatus consisted of a large un-

painted wooden box 89 cm high, 152 cmwide, and 60 cm deep. The infant faced anopening 45 cm high and 150 cm wide in thefront wall of the apparatus. The back wall ofthe apparatus was covered with blue cloth.

A wooden ramp 61 cm long and 13 cmwide was centered against the left wall ofthe apparatus below an opening 15 cm highand 15 cm wide. The ramp was 15 cm highat its highest point and sloped downward ata 16° angle. There was a rail 1 cm high and0.5 cm wide on either side of the ramp. Twowooden tracks, each 91 cm long, 1 cm high,and 1 cm thick, lay 6 cm apart between thelower end of the ramp and an opening 13 cmhigh and 20 cm wide in the right wall of theapparatus. A toy car 8 cm high, 17 cm long,and 8 cm wide could roll onto the rampthrough the opening in the left wall. The carrolled down the ramp, between the rails, andthen rolled across the apparatus, along thetracks, until it disappeared through theopening in the right wall. The car waspainted white and was decorated with redstrips and green pom-poms. Black felt cov-ered the ramp and tracks to minimize thecar's noise. White muslin curtains hid theopenings in the side walls of the apparatus;these curtains opened as the car wentthrough and closed behind it.


HABITUATION EVENT

1 '—

TEST EVENTSPossible Event

Impossible Event

1-:

— 1

•' >'•'• ' ~ " ' ^ J . H

["• ' • •

1—«U

1

FIG. 4.—Schematic representation of the habituation and test events shown to the infants in Exper-iment 3.

A yellow plastic screen 28 cm high and26 cm wide stood 23 cm in front of the tracksat a distance of 61 cm from the left wall and65 cm from the right wall. A wooden handle91 cm high, 1.5 cm wide, and 1 cm thick wasaffixed to the back of the screen and pro-truded through the ceiling of the apparatus.The top portion of the handle fit into a verti-cal slit mounted inside the front wall of theapparatus. By raising and lowering the han-dle within this slit (from above), an experi-menter could raise and lower the screen.

A brightly colored plastic toy mouse, ap-proximately 17 cm high, 13 cm wide, and10 cm thick, could be introduced into theapparatus through a hidden opening in theback wall. This toy represented a smilingMickey Mouse, in a sitting position, playingwith alphabet blocks.

The infant was tested in a brightly litroom. Four lights (each with a 40-watt lightbulb) were attached to the front and sidewalls of the apparatus to provide additionallight. These lights were arranged so as toeliminate tell-tale shadows. Two woodenframes, each 183 cm high and 70 cm wideand covered with blue cloth, stood at anangle on either side of the apparatus. Theseframes isolated the infant from the experi-mental room. At the end of each trial, a cur-tain consisting of a muslin-covered frame 63cm high and 152 cm wide was lowered infront of the opening in the front wall of theapparatus.

EventsHabituation event.—Two experiment-

ers worked in concert to produce the habitu-ation event The first operated the screenand the second operated the car. To start,the first experimenter lifted the screen 27cm, taking about 1 sec to complete tbis ac-tion; she paused for about 2 sec, and thenlowered the screen to its initial position,again taking about 1 sec to perform this ac-tion. After a 2.5-sec pause, the second exper-imenter pushed the car through the curtainat the top of the inclined ramp. The car thenrolled down the ramp and across the appara-tus, passing behind the screen, and finallyexiting the apparatus to the right The cartook about 2.5 sec to roll in and out of theapparatus. About 2 sec after the car emergedfrom the apparatus, the first experimenteragain lifted the screen, beginning a newevent cycle. Each cycle thus lasted approxi-mately 11 sec. Cycles were repeated withoutstop until the computer signaled the endingof the trial (see below). When this occurred,the second experimenter lowered the cur-tain in front of the apparatus.

Impossible test event.—The impossibletest event was identical to the habituationevent with two exceptions. First, the mousewas placed on top of the car's tracks, cen-tered behind the screen; the mouse was re-vealed when the screen was raised at thestart of each event cycle. Second, after thescreen was lowered, a third experimenter


surreptitiously reached through the hiddenopening in the apparatus's back wall and re-moved the mouse from the car's path. Afterthe car rolled past the screen, this same ex-perimenter quickly replaced the mouse ontop of the tracks so that when the screen wasnext raised, the mouse stood intact in thesame location as before. As in the habitua-tion event, each cycle lasted approximately11 sec, and the mouse was totally occludedfor the last 7 of these 11 sec. The delay be-tween the occlusion of the mouse and thereappearance of the car from behind thescreen was about 4.5 sec.

Possible test event.—The possible testevent was identical to the impossible testevent except that the mouse was placed 10cm behind the car's tracks. As in the impossi-ble event, the third experimenter reachedinto the apparatus after the screen was low-ered and grasped the mouse. This insuredthat any faint sounds associated with themouse's surreptitious movement during theimpossible event were also present duringthe possible event.

ProcedurePrior to the experiment, the infant was

allowed to manipulate the car and the mousefor a few minutes while his or her parentfilled out consent forms. During the experi-ment, the infant sat on the parent's lap infront of the apparatus, facing the screen. Theinfant's head was approximately 60 cm fromthe screen and 117 cm from the back wall ofthe apparatus.

Each infant participated in a three-phase procedure consisting of a familiariza-tion phase, a habituation phase, and a testphase. During the familiarization phase, theinfant received two trials designed to ac-quaint him or her with the mouse's two pos-sible locations. The screen remained liftedthroughout these trials. In one trial, themouse stood on top of the tracks; in the othertrial, it stood behind the tracks. Each trialended when the infant either (a) lookedaway from the display for 2 consecutive secafter having looked at it for at least 4 cumula-tive sec or (b) looked at the display for 60cumulative sec without looking away for 2consecutive sec.

Following the familiarization trials,each infant was habituated to the habitua-tion event described above. The main pur-pose of this habituation phase was to ac-quaint the infant with the movements of thescreen and car. Each habituation trial endedwhen the infant (a) looked away from the

event for 2 consecutive sec after havinglooked at it for at least 9 cumulative sec or(b) looked at the event for 60 cumulative secwithout looking away for 2 consecutive sec.Habituation trials continued until the infanteither (a) met a habituation criterion of a50% or greater decrease in looking time onthree consecutive trials, relative to the in-fant's looking time on the first three trials, or(b) completed nine habituation trials. Of the32 infants in the experiment, 14 completednine trials without satisfying the habituationcriterion; the remaining infants took an aver-age of 7.22 trials to meet the criterion.

During the test phase, the infants sawthe impossible and the possible test eventsdescribed above on alternate trials until theyhad completed four pairs of test trials. At thebeginning of each test trial, the first experi-menter waited to lower the screen until thecomputer signaled that the infant had lookedat the mouse for 3 cumulative sec. This en-sured that the infant had noted the presenceand the location of the mouse behind thescreen. Each test trial ended when the infant(a) looked away from the event for 2 consec-utive sec after having looked at it for at least6 cumulative sec (beginning at the end ofthe pretrial, when the first experimenterlowered the screen in front of the toy) or(b) looked at the event for 60 cumulative secwithout looking away for 2 consecutive sec.Like the 9-sec value in the habituation trial,the 6-sec value was chosen to ensure thatthe infant had the opportunity to see the carreappear from behind the screen. Half of theinfants saw the mouse on top of the tracksfirst during the familiarization and test trials;the other infants saw the mouse behind thetracks first.

Eleven of the 32 infants in the experi-ment completed fewer than four pairs of testtrials. Four infants completed only threepairs, three because of fussiness and one be-cause of drowsiness; the other infants com-pleted only two pairs, six because of fussi-ness and one because of procedural error.Preliminary analyses revealed no significanteffect of order on the infants' looking timesat the impossible and the possible eventsduring the four pairs of test trials, all F's <0.98. The data were therefore collapsed insubsequent analyses.

RESULTS

The infants' looking times at the testevents were analyzed by means of a 4 x 2mixed-model analysis of" variance with pair


(first, second, third, or fourth pair of test tri-als) and event (impossible or possible event)as the within-subject factors. The main effectofevent was not significant, F(l,181) = 0.80,indicating that the infants did not look reli-ably longer at the impossible (M = 27.4)than at the possible (M = 25.8) event. Noother result was significant.

Sex differences.—Examination of thedata suggested that the pattem revealed bythe preceding analysis—statistically equallooking times at the impossible and the pos-sible events—represented the average oftwo distinct looking patterns. Specifically, itappeared that the female infants (n = 16,M = 4 months, 0 day) in the experimenttended to look longer at the impossible thanat the possible event, whereas the male in-fants (n = 16, M = 3 months, 29 days)tended to look equally at the two events (seeFig. 5).

An additional analysis was therefore car-ried out comparing the looking times of themale and female infants at the test events.This analysis was a 2 x 4 x 2 mixed-modelanalysis of variance with sex as thebetween-subjects factor, and with pair and

event as the within-subjects factors, as in thepreceding analysis. The only significant ef-fect was the interaction between sex andevent, F(l,174) = 5.18, p < .05. Follow-upanalyses confirmed that the female infantslooked reliably longer at the impossible(M = 31.1) than at the possible event (M =23.7), F(l,174) = 4.15, p < .05, whereas themale infants looked about equally at the twoevents, F(l,174) = 1.07, p > .05 (impossible:M = 24.1, possible: M = 27.6).

Upon the obtention of these results, fur-ther analyses were undertaken comparingthe responses of the male and female infantsto the familiarization and habituation events.The results of these analyses indicated thatthe male and female infants in the experi-ment did not differ significantly in (a) theirlooking times at the two familiarizationevents, F(l,30) = 0.58; (b) the number ofhabituation trials they received, F(l,30) =1.30, p > .05; (c) their total looking timesduring the habituation trials, F(l,30) = 0.17;(d) their mean looking times during the ha-bituation trials, F(l,30) = 0.08; and (e) theirlooking times during their last six habitua-tion trials, F(l,30) = 0.53. The male and fe-male infants in Experiment 3 thus differed

Female Infants

-4 -3 -2 -1

Habituation TrialsFIG. 5.—Mean looking times of the male and female infants in Experiment 3 at the habituation

and test events.

2 3

Test Trials


in their reactions to the test but not the fa-miliarization and habituation events, as as-sessed by these measures.

The female infants' reliable preferencefor the impossible over the possible testevent suggested that they were surprised tosee the car reappear from behind the screenwhen the mouse stood on top of the tracks.However, another possible interpretation forthis finding was that the female infantsfound the mouse especially attractive and solooked longer when it stood closer to them.Analysis of the female infants' responses tothe two familiarization events provided evi-dence against this interpretation: the femaleinfants' looking times did not differ reliablywhen the mouse stood on top (M = 27.6) orin back (M = 31.3) of the tracks, F(l,15) =0.46. To provide further evidence againstthis alternative interpretation, and at thesame time confirm female infants' surpriseat the impossible test event, an additionalgroup of 4.0-month-old female infants wasrun in Experiment 3A. This experiment'smethod was identical to that of Experiment3 except that the mouse was placed in frontrather than in back of the tracks in the possi-ble test event (see Baillargeon, 1986). Wereasoned that if the female infants in Experi-ment 3 looked longer at the impossibleevent because the mouse stood closer tothem, then the infants in Experiment 3Ashould look longer at the possible (front)than at the impossible (top) event.

The performance of the male infants inExperiment 3 will be discussed after thepresentation of the results of Experiments3A and 4.

Experiment 3A

METHOD

SubjectsSubjects were 16 healthy, full-term, fe-

male infants ranging in age from 3 months,23 days to 4 months, 14 days (M = 4 months,2 days). Three additional infants were elimi-nated from the experiment because of fuss-iness.


used in Experiment 3A were identical tothose in Experiment 3 except that the mousewas positioned 10 cm in front instead of inback of the tracks in the familiarization andtest events.

Analysis of the infants' looking times atthe two familiarization events indicated that

they looked about equally when the mousewas positioned on top (Af = 29.4) or in front(M = 30.5) of the tracks, F(l,15) = 0.03.Four infants completed nine habituation tri-als without satisfying the habituation crite-rion; the other infants took an average of 7.17trials to reach the criterion. Three infantscompleted fewer than four pairs of test trials.Two infants completed only three pairs, onebecause of fussiness and one because theprimary observer could not follow the direc-tion of the infant's gaze; the remaining infantcompleted only two pairs, because of fussi-ness. Preliminary analyses revealed no sig-nificant effect of order on the infants' lookingtimes at the impossible and the possibleevents during the four pairs of test trials, allF's < 1.55, p > .05. The data were thereforecollapsed in subsequent analyses.

RESULTS

Figure 6 shows the infants' mean look-ing times at the impossible and the possibleevents. For purposes of comparison, themean looking times of the female infants inExperiment 3 are also presented. It can beseen that both groups of infants tended tolook longer at the impossible event.

The looking times of the infants in Ex-periment 3A were compared to those of thefemale infants in Experiment 3 by means ofa 2 X 4 • X 2 mixed-model analysis of vari-ance with experiment (3 or 3A) as thebetween-subjects factor and with pair (first,second, third, or fourth test pair) and event(impossible or possible event) as thewithin-subjects factors. The main effect ofevent was significant, F(l,104) = 9.67, p <.003, indicating that the infants looked reli-ably longer overall at the impossible (M =34.6) than at the possible (M = 28.4) event.The effect of condition was not significant,F(l,30) = 2.78, p > .05, nor was any of theinteractions involving this factor, all F's <2.64, p > .05, indicating that there were noreliable differences between the lookingtimes of the female infants in Experiments3 and 3A at the impossible and the possibleevents (Experiment 3: impossible, M =31.1, possible, M = 23.7; Experiment 3A:impossible, M = 37.7, possible, M = 32.4).No other result was significant.

DISCUSSION

Like the female infants in Experiment3, the female infants in Experiment 3Alooked reliably longer at the impossible than


Experiments: Top/Back

Expenment 3A: Top/Front

~~*'~ Impossible Event- - • - Possible Event

-4 -3 -1

Habituation Trials Test TrialsFIG. 6.—Mean looking times of the female infants in Experiments 3 and 3A at the habituation and

test events.

at the possible event. Together, these resultssuggest that the female infants in these twoexperiments (a) realized that the mouse andthe car continued to exist behind the screen,(b) believed that the mouse retained its loca-tion behind the screen, (c) assumed that thecar pursued its trajectory behind the screen,(d) understood that the car could not rollthrough the space occupied by the mouse,and hence (e) were surprised to see the carreappear from behind the screen when themouse stood on top of the tracks. The resultsof Experiments 3 and 3A are thus consistentwith those of Experiment 1 in suggestingthat young infants can represent and reasonabout not only the existence but also someof the properties—such as the location andtrajectory—of occluded objects.

Experiment 4Would results similar to those obtained

with the female infants in Experiment 3 alsobe obtained with younger female infants? Tofind out, 3.5-month-old female infants weretested using the same procedure as in Exper-iment 3.

METHOD


fants ranging in age from 3 months, 6 daysto 3 months, 22 days (Af = 3 months, 16days). One additional infant was excludedfrom the experiment because she failed tocomplete at least two pairs of test trials dueto fussiness.̂

^ The reader may be puzzled by the fact that the attrition rate due to fussiness in Experi-ments 3, 3A, and 4 was so much smaller than that in Experiments 1 and 2, despite the generalsimilarity of the subjects' ages. There were 32 infants in Experiment 3, and two additional infantswere eliminated because of fussiness. The corresponding figures for the other experiments were:3A, 16 and 3; 4, 20 and 1; 1, 32 and 13; and 2, 29 and 15. We are not entirely clear as to thereasons for this difference. T"he infants generally loved the car experiment, but seemed more



used in Experiment 4 were identical to thosein Experiment 3. During the two familiariza-tion trials, the infants looked about equallywhen the mouse stood on top (Af = 31.2)and in front (M = 28.5) of the tracks, F(l,19)= 0.28. Ten of the 20 infants in the experi-ment completed nine habituation trials with-out reaching the habituation criterion; theremaining infants took an average of 7.30 tri-als to reach the criterion. Eight infants failedto complete the full complement of fourpairs of test trials. Four infants completedonly three pairs, three because of fussinessand one because of drowsiness; the other in-fants completed only two pairs, one becauseof fussiness, one because of drowsiness, andtwo because of procedural error. Preliminaryanalyses revealed no significant effect of or-der on the infants' looking times at the im-possible and the possible events during thefour pairs of test trials, all F's < 0.78. Thedata were therefore collapsed in subsequentanalyses.

RESULTS

Figure 7 presents the infants' meanlooking times at the impossible and the pos-sible test events. The data were analyzed bymeans of a 4 x 2 mixed-model analysis ofvariance with pair (first, second, third, orfourth test pair) and event (impossible orpossible event) as within-subject factors.The main effect of event was not significant,F(l,64) = 2.41, p > .05, indicating that theinfants did not look reliably longer at theimpossible (Af = 30.9) than at the possible(Af = 27.4) event. No other result was sig-nificant.

DISCUSSION

Like the 4.0-month-old male infants inExperiment 3, the 3.5-month-old female in-fants in Experiment 4 failed to show a reli-able preference for the impossible over thepossible event, suggesting that they werenot surprised to see the car roll past thescreen when the mouse stood in its path.

How can we explain the discrepancy be-tween these negative findings and the posi-tive findings obtained with the 4.0-month-old female infants in Experiments 3 and 3A?Given the results of Experiment 1 and thoseof Baillargeon (1987a), it is plausible that the4.0-nionth-old male infants in Experiment 3and the 3.5-month-old female infants in Ex-periment 4 (a) believed that the mouse con-tinued to exist behind the screen, (b) as-sumed that the car continued to exist andpursued its trajectory behind the screen, and(c) understood that the car could not rollthrough the space occupied by the mouse.Nevertheless, the infants may have lackedsome other conceptual ability necessary fordetecting the surprising character of the im-possible event, such as the ability to keeptrack of the mouse's location on each trial,or the ability to reason about the interactionof two simultaneously occluded objects,such as the mouse and the car.

Another explanation for the infants' fail-ure to notice the violation embedded in theimpossible event is that this failure stemmednot from a conceptual but from a perceptuallimitation. Clearly, if the infants lacked thevisual skills necessary to determine that themouse stood on top or in back of the car'stracks, they could not have distinguishedbetween the impossible and the possibleevents. Recent research on the developmentof stereopsis may be relevant here (e.g..Birch, Gwiazda, & Held, 1982; Gwiazda,Bauer, & Held, 1989a, 1989b; Held, Birch,& Gwiazda, 1980; Held, in press). Investiga-tors have found that following the onset ofcoarse stereopsis (30 or more min of dispar-ity), sensitivity to disparity rapidly increasesto 1 min of arc (Birch et al., 1982; Held etal., 1980). In these developments, male in-fants lag behind female infants by severalweeks. For example, Gwiazda et al. (1989a)found the mean age of onset of stereopsis tobe 9.1 weeks for female infants, as comparedwith 12.1 weeks for meile infants. These andrelated findings suggest the following specu-lation. It may be that the 4.0-month-oId fe-male infants in Experiments 3 and 3A hadachieved sufficient stereoacuity to assess ac-

ambivalent about the carrot experiment. One reason for this difference may have been that tliecar events were more complex and thus more engaging. Another reason may have been that theinfants were troubled by the conflict between the animate and inanimate features of our carrots(e.g., they possessed Eacial features and moved in some respects like animate organisms, stoppingand reversing direction without obvious extemal forces being applied; yet they were clearlytoylike objects). This second reason is consistent with observers' comments that the infants whodid not complete Experiments 1 and 2 often appeared scared of the carrots. Researchers inter-ested in replicating these experiments might be better served by using less animate-like stimuli,such as simple geometric shapes.

i-

o

COd)

5045403530252015105

BaiUargeon and DeVos 1243

Impossible EventPossible Event

-4 -3 -2 -1

Habituation Trials Test TrialsFIG. 7.—Mean looking times of the female infants in Experiment 4 at the habituation and test

events.

curately the location of the mouse relative tothe car's tracks, but that the 3.5-month-oldfemale infants in Experiment 4 and the 4.0-month-old male infants in Experiment 3 hadnot. As already mentioned, if these infantswere unable to determine whether themouse stood in or out of the path of the car,they could not have distinguished betweenthe possible and the impossible events sincethese events were identical in all other re-spects. This explanation could perhaps betested by altering our current experimentalprocedure so as to provide infants with bet-ter or richer depth infonnation. Such alter-ations might include (a) increasing the dis-tance between the mouse and the car'stracks in the possible event, to make this dis-tance more salient and hence more easilydetectable; (b) rocking the infant gently backand forth in front of the display when themouse is in view, to enhance kinetic depthinformation; and/or (c) moving the mousegently back and forth in its location when inview, for the same reason.

ConclusionWhen adults see an object occlude an-

other object, they typically make three as-sumptions. The first is that the occluded ob-ject continues to exist behind the occludingobject. The results of the present experi-ments indicate that young infants also sharethis assumption. Thus, the 3.5-month-old in-fants in Experiment 1 believed that each car-rot continued to exist after it slid behind thescreen. Similarly, the 4.0-month-old femaleinfants in Experiments 3 and 3A believedthat (a) the mouse continued to exist after it

was occluded by the screen and (b) the carcontinued to exist after it rolled behind thescreen. These results are consistent with andprovide converging evidence for Baillar-geon's (1987a) claim that infants as young as3.5 months of age represent the continuedexistence of occluded objects.

The second assumption adults generallyhold about occluded objects is that they re-tain the physical and spatial properties theypossessed prior to occlusion. The present re-sults suggest that young infants also sharethis second assumption. The 3.5-month-oldinfants in Experiment 1 realized that eachcarrot retained its height behind the screen,and the 4.0-month-old female infants in Ex-periments 3 and 3A believed that the mouseretained its location behind the screen.

Finally, the third assumption adultshold about occluded objects is that they re-main subject to physical laws: their displace-ments and interactions with other objects donot become capricious or arbitrary but fol-low the same regular pattems as visible ob-jects. The results of the present experimentscannot tell us whether young infants haveany conception of physical laws; indeed, itseems highly unlikely that young infantscould appreciate the necessary character ofsuch laws. However, the present results dosuggest that young infants expect occludedobjects to behave in the same predictablemanner as visible objects. Thus, the 3.5-month-old infants in the experimental condi-tion in Experiment 1 believed that each car-rot pursued a spatially continuous trajectorybehind the screen, just as it did on eitherside of the screen. Similarly, the 4.0-month-

1244 ChUd Development

old female infants in Experiments 3 and 3Aassumed that the car rolled along a continu-ous path behind the screen, just as it didbefore and after its passage behind thescreen. Furthermore, the 4.0-month-old fe-male infants in Experiments 3 and 3A ap-preciated that the occluded car could not rollthrough the space occupied by the occludedmouse. This last result is consistent withBaillargeon's (1987a) observation that in-fants aged 3.5 months and older understandthat a rigid screen cannot rotate through thespace occupied by a box placed behind thescreen. Together, these results suggest thatyoung infants have general expectationsabout objects' displacements and interac-tions and believe, very sensibly, that theseexpectations apply to visible as well as tooccluded objects.

Piaget (1954) described the develop-ment of object permanence in terms of theslow amalgamation of two initially separateworlds. To start, Piaget maintained, infantsdistinguish between the visible world, filledwith solid entities whose behaviors can beknown and understood, and the occludedworld, a void which objects, like occult spir-its, enter and leave without any discernibledesign (pp. 11-13). By the end of the secondyear, however, infants regard occluded ob-jects as substantial entities that obey thesame laws as visible objects.

The present results, together with thoseof Baillargeon (1987a), provide little supportfor Piaget's characterization of the develop-ment of the object concept in infancy. Onthe contrary, the present results suggest that,from very early on, infants conceive of oc-cluded objects in the same general manneras adults, as inhabiting the same world andas conforming to the same patterns as visibleobjects.

To say that infants conceive of occludedobjects as adults do—because they share thesame three basic assumptions about theseobjects—is not to say that infants always rea-son about occluded objects as adults do.Young infants' ability to represent and toreason about occluded objects is clearlymore limited than that of adults and mustdevelop through infancy and childhood. Toillustrate, consider the habituation and testevents shown to the infants in the experi-mental condition in Experiment 1. Adultsmight have noticed from the start that thenoise that accompanied the movement ofeach carrot to the left or the right of thescreen stopped abruptly when the carrot slidbehind the screen. On the basis of this cue.

adults might have concluded that two sepa-rate carrots were used in producing the ha-bituation and test events, thereby account-ing for the tall carrot's failure to appear inthe screen window in the impossible testevent. The results of the experimental con-dition in Experiment 1 suggest that the 3.5-month-old infants in this condition, like the5.5-month-oId infants tested by Baillargeonand Graber (1987), did not attend to or failedto comprehend the implications of these au-ditory cues.

There already have been attempts atcharting some of the ways in which infants'ability to reason about occluded objects de-velops over time (e.g., Baillargeon, 1991, inpress-a, in press—b; Baillargeon & DeVos,1991). One hypothesis to emerge from thisresearch is that infants succeed in solvingocclusion problems requiring qualitativereasoning strategies before they succeed insolving occlusion problems requiring quan-titative strategies. Beasoning strategies arereferred to as quantitative if they require in-fants to reason about specific quantities, andas qualitative if they do not.

This distinction gives rise to interestingspeculations about the way in which the in-fants in the present experiments solved theproblems they were given. To solve the roll-ing car problem used in Experiments 3, 3A,and 4, the infants did not need to representand to reason about specific quantities: allthey had to do was to note the location ofthe mouse relative to the path of the car.However, in order to solve the sliding carrotproblem used in Experiments 1 and 2, theinfants could use one of two strategies. Onequantitative strategy involved mentally com-paring the height of each carrot, after it slidbehind the screen, to that of the window'slower edge to see whether the former wasgreater than the latter. This strategy is re-ferred to as quantitative because it requiredthe infants to represent the specific heightof each carrot. The other, qualitative strategyinvolved visually comparing each carrot asit approached the screen to determinewhether the carrot was taller than the win-dow's lower edge. This strategy is said to bequalitative because it did not necessitate therepresentation of specific quantities, onlythe encoding of relative quantities. All thatthe infants needed to do was to note whetherthe carrot was taller than the window: thetwo objects' specific heights were irrelevant.

The present data are insufficient to de-termine whether the 3.5-month-old infantsin Experiment 1 used a quantitative or a

qualitative strategy to reason that the tall car-rot should appear in the screen window.However, data collected by Baillargeon(1991), using the rotating screen paradigm,suggest that the latter possibility is morelikely. In a series of experiments, Baillar-geon found that it was not until 6.5 monthsof age that infants could predict quantita-tively at what point a rotating screen shouldcontact the box placed behind it and stop(i.e., by using their representation of the oc-cluded box's height and location). At 4.5months of age, infants could only predict thescreen's stopping point qualitatively, by us-ing as reference point a second, identicalbox placed to the side of the occluded box,out of the path of the screen. It seems likelythat, like the 4.5-month-old infants in Bail-largeon's experiments, the 3.5-month-old in-fants in Experiment 1 were using a qualita-tive, visual alignment strategy to predictwhether each carrot should be visible in thescreen window.

Beyond the theoretical issues addressedin the previous discussion, the results of thepresent experiments also have importantmethodological implications. One such im-plication is that the present results reinforcethe previously noted (e.g., Baillargeon, inpress-a; Baillargeon et al., 1985, 1990) dis-crepancy between investigations of infants'physical world that have relied on visualtasks as opposed to manual search tasks. Theresults of the present experiments suggestthat by 3.5 to 4.0 months of age, infants areable to represent the existence and locationof occluded objects. Yet it is not until severalmonths later that infants (o) begin to searchfor objects they have observed being hiddenand (b) correctly search for objects hiddenin one of two locations (e.g.. Diamond, 1985;Piaget, 1954; Wellman et al., 1987). Thereader is referred to Baillargeon (in press-a;Baillargeon et al., 1989, 1990) for an accountof the late emergence and slow developmentof infants' manual search in terms of the lim-itations of infants' problem-solving skills.

Another methodological implication ofthe present results and of those found byBaillargeon (1987a) concems researchers'selection of object permanence tests for usewith young infants. These results make clearthat whether a test yields positive findingsvery much depends on which test is used;all tests are not created equal. The rotat-ing screen paradigm used by Baillargeon(1987a) provided evidence of object perma-nence in 4.5-month-old infants and in fastbut not in slow habituating 3.5-month-old


infants. The sliding carrot paradigm usedin Experiment 1 yielded positive evidencewith 3.5-month-old infants. Finally, the roll-ing car paradigm used in Experiments 3, 3A,and 4 generated negative results with 3.5-month-old infants, albeit positive resultswith 4.0-month-old female infants. Detailedtask analyses are needed to determine whycertain tasks prove more difficult for infantsthan others, and why these difficulties resultat times in differences between fast and slowhabituators and at times in differences be-tween male and female infants.

In conclusion, the results of the presentexperiments point to remarkable knowledgeand abilities on the part of young infants.Young infants are aware that objects (a) can-not exist at two successive points in timewithout having existed during the intervalbetween them, (fo) cannot appear at two sep-arate points in space without having traveledthe distance between them, and (c) cannotmove through the space occupied by otherobjects. Furthermore, infants are able to usethis knowledge to make (qualitative and per-haps quantitative) predictions about objects'displacements and interactions with otherobjects. Though unexpected from the pointof view of traditional developmental theory,these findings are nevertheless consistentwith recent investigations of other facets ofyoung infants' physical world (e.g., Baillar-geon, 1991; Baillargeon & Hanko-Summers,1990; Needham & Baillargeon, 1991). To-gether, these results underscore the richnessand sophistication of young infants' physicalworld and raise important questions aboutthe origins and development of this world.

ReferencesBaillargeon, R. (1986). Representing the existence

and the location of hidden objects: Objectpermanence in 6- and 8-month-old infants.Cognition, 23, 21-41.

Baillargeon, R. (1987a). Object permanence in 3.5-and 4.5-month-oId infants. DevelopmentalPsychology, 23, 655-664.

Baillargeon, R. (1987b). Young infants' reasoningabout the physical and spatial properties ofa hidden object. Cognitive Development 2179-200.

Baillargeon, R. (1991). Reasoning about the heightand location of a hidden object in 4.5- and6.5-month-old infants. Cognition, 38, 13-42.

Baillargeon, R. (in press-a). The object conceptrevisited: New directions in the investigationof infants' physical knowledge. In C. E. Gran-rud (Ed.), Visual perception and cognition in


infancy: Carnegie-Mellon symposia on cogni-tion (Vol. 23). Hillsdale, NJ: Erlbaum.

Baillargeon, R. (in press-b). The development ofinfants' physical reasoning. In H. W. Reese(Ed.), Advances in child development and be-havior (Vol. 23). New York: Academic Press.

Baillargeon, R., & DeVos, J. (1991). Qualitativeand quantitative inferences about hidden ob-jects in young infants. Manuscript submittedfor publication.

Baillargeon, R., DeVos, J., & Graber, M. (1989).Location memory in 8-month-old infants in anon-search AB task: Further evidence. Cogni-tive Development, 4, 345-367.

Baillargeon, R., & Graber, M. (1987). Where is therabbit? 5.5-month-old infants' representationof the height of a hidden object. CognitiveDevelopment, 2, 375-392.

Baillargeon, R., & Graber, M. (1988). Evidence oflocation memory in 8-month-old infants in anon-search AB task. Developmental Psychol-ogy, 24, 502-511.

Baillargeon, R., Graber, M., DeVos, J., & Black, J.(1990). Why do young infants fail to search forhidden objects? Cognition, 36, 225-284.^

Baillargeon, R., & Hanko-Summers, S. (1990). Isthe top object adequately supported by thebottom object? Young infants' understandingof support relations. Cognitive Development,5, 29-53.

Baillargeon, R. (1991). The development of in-fants' intuitions about support. Manuscriptsubmitted for publication.

Baillargeon, R., Spelke, E. S., & Wasserman, S.(1985). Object permanence in 5-month-old in-fants. Cognition, 20, 191-208.

Birch, E. E., Gwiazda, J., & Held, R. (1982). Ste-reoacuity development for crossed and un-crossed disparities in human infants. VisionResearch, 22, 1019-1024.

Bremner, J. G. (1985). Object tracking and searchin infancy: A review of data and a theoreti-cal evaluation. Developmental Review, 5,371-396.

Diamond, A. (1985). Development of the abilityto use recall to guide action, as indicated byinfants' performance on AB. Child Develop-ment, 56, 868-883.

Gratch, G. (1975). Recent studies based on Pia-get's view of object concept development. InL. B. Cohen & P. Salapatek (Eds.), Infant per-ception: From sensation to cognition (Vol. 2,pp. 51-99). New York: Academic Press.

Gratch, G. (1976). A review of Piagetian infancyresearch: Object concept development. InW. F. Overton & J. M. Gallagher (Eds.),Knowledge and development: Advances in re-search and theory. New York: Plenum.

Gwiazda, J., Bauer, J., & Held, R. (1989a). Binocu-

lar function in human infants: Gorrelation ofstereoptic and fusion-rivalry discriminations.Journal of Pediatric Ophthalmology andStrabismus, 26, 128-132.

Gwiazda, J., Bauer, J., & Held, R. (1989b). Fromvisual acuity to hyperacuity: A 10-year up-date. Canadian Journal of Psychology, 43,109-120.

Haith, M. M. (in press). The early development ofvisual expectations. In C. E. Granrud (Ed.),Visual perception and cognition in infancy:Carnegie-Mellon symposia on cognition (Vol.23). Hillsdale, NJ: Erlbaum.

Harris, P. L. (1987). The development of search.In P. Salapatek & L. B. Gohen (Eds.), Hand-book of infant perception (Vol. 2, pp. 155-207). New York: Academic Press.

Harris, P. L. (1989). Object permanence in in-fancy. In A. Slater & J. G. Bremner (Eds.),Infant development (pp. 103-121). Hillsdale,NJ: Erlbaum.

Held, R., Birch, E. E., & Gwiazda, J. (1980). Ste-reoacuity of human infants. Proceedings ofthe National Academy of Sciences, USA, 77,5572-5574.

Held, R. (in press). Development of cortically me-diated visual processes in human infants. InG. von Euler, H. Forssberg, & H. Lagercrantz(Eds.), Neurobiology of early infant behav-iour. London: Macmillan.

Needham, A., & Baillargeon, R. (1991). Intuitionsabout support in 4.5-month-old infants.Manuscript submitted for publication.

Piaget, J. (1954). The construction of reality in thechild. New York: Basic.

SAS InsHtute. (1985). SAS user's guide: Statistics(1985 ed.). Gary, NG: SAS Institute.

Schuberth, R. E. (1983). The infant's search forobjects: Alternatives to Piaget's theory of con-cept development. In L. P. Lipsitt & G. K.Rovee-Gollier (Eds.), Advances in infancy re-search (Vol. 2, pp. 137-182). Norwood, NJ:Ablex.

Sophian, G. (1984). Developing search skills ininfancy and early childhood. In G. Sophian(Ed.), Origins of cognitive skills (pp. 27-56).Hillsdale, NJ: Erlbaum.

Spelke, E. S. (1988). Where perceiving ends andthinking begins: The apprehension of objectsin infancy. In A. Yonas (Ed.), Perceptual de-velopment in infancy: Minnesota Symposiaon Child Psychology (Vol. 20, pp. 197-234).Hillsdale, NJ: Erlbaum.

Wellman, H. M., Gross, D., & Bartsch, K. (1987).Infant search and object permanence: Ameta-analysis of the A-not-B error. Mono-graphs of the Society for Research in ChildDevelopment, 51(3, Serial No. 214).

object permanence in young infants: further...

Documents