Becoming a Social Partner With Peers. Cooperation and Social Understanding in One- And Two-Year-Olds...

Becoming a Social Partner With Peers: Cooperation and Social Understanding

in One- and Two-Year-Olds

Celia A. Brownell, Geetha B. Ramani, and Stephanie ZerwasUniversity of Pittsburgh

One- and two-year-old peer dyads were presented with a simple cooperative task. Age differences were foundin amount of coordinated activity, monitoring the peers activity and location in relation to the goal, and at-tempting to achieve the goal when the peer was (or was not) available as a partner. One-year-olds coordinatedactions appeared more coincidental than cooperative whereas older children appeared to be more activelycooperating toward a shared goal. Differences in coordinated activity with peers were associated with differ-ences in attention sharing with an adult and with language about self and other. The ability to cooperate withpeers, becoming a true social partner, develops over the 2nd and 3rd years of life in concert with growing socialunderstanding.

Development of the ability to cooperate with peers isa signal achievement. Theorists have long main-tained that critical social and cognitive develop-ments are born in the context of cooperative playwith other children (Dunn, 1988; Garvey, 1990;Hartup, 1983; Piaget, 1932). Indeed, cooperation mayconstitute a social imperative (Hartup, 1996) and itforms the foundation for human culture (Rogoff,1990; Tomasello, 1999a). The beginnings of peer co-operation represent the childs entry into peer cul-ture. However, the origins and early development ofchildrens ability to cooperate with peers are littlestudied and remain poorly understood. Moreover,despite the likely involvement of developing socialunderstanding in the growth of early peer interac-tion (Brownell, 1986; Brownell & Hazen, 1999; Dunn,1988; Eckerman & Peterman, 2001) and the lastdecades burgeoning research on very young chil-drens understanding of others desires, intentions,and goals, there have been few efforts to relate tod-dlers nascent social skills to their emerging social

understanding. The purpose of this study was two-fold: to examine developments in childrens earlycooperative abilities with peers, and to examine as-sociations between cooperation and social under-standing in a period when both are rapidlydeveloping.

Early Developments in Cooperation With Peers

In the 2nd year of life, children begin to move out ofthe exclusivity of adult child relationships into thelarger, novel world of peers. Before 18 months, peerinteractions are rare, primitive, and minimally co-ordinated, even though children exhibit social in-terest in one another as early as 6 months of age (seeBrownell & Brown, 1992; Eckerman & Peterman,2001, for reviews). In a longitudinal study, Ecker-man, Davis, & Didow, (1989) found a dramatic in-crease between 20 and 24 months of age inspontaneous, novel cooperative activity betweenunfamiliar toddler peers. This was almost entirelyaccounted for by mutual imitation of nonverbal ac-tions such as jumping or running. Others have alsoreported marked increases in imitative and/or re-ciprocal peer play at about 24 months (Eckerman &Whitehead, 1999; Howes, 1988; Maudry & Nekula,1939; Ross, 1982), and one study using a structuredproblem-solving task found that 24- and 30-month-old children could reliably cooperate with each otherbut younger children could not (Brownell & Car-riger, 1990). Thus, peer cooperation appears toemerge at the end of the 2nd year of life. During the3rd year, childrens cooperative play becomes more

r 2006 by the Society for Research in Child Development, Inc.All rights reserved. 0009-3920/2006/7704-0001

This research was supported in part by a grant from the Na-tional Institute of Child Health & Human Development (R03HD043971) to the first author. We thank the children and parentswho volunteered their time to participate in this research. Specialappreciation is also extended to Anita Adalja, Chris Byers, Am-anda Kimmel, and Amy Mariaskin for assistance with data col-lection and coding, to Sara Nichols, Margarita Svetlova, andanonymous reviewers for helpful comments on an earlier draft,and to Frank Valentich for building the cooperation and trainingtasks. Portions of this research were presented at meetings of theSociety for Research in Child Development and the InternationalSociety for Infant Studies.Correspondence concerning this article should be addressed to

Celia A. Brownell, Department of Psychology, University ofPittsburgh, Pittsburgh, PA 15260. Electronic mail may be sent [email protected].

Child Development, July/August 2006, Volume 77, Number 4, Pages 803 821

responsive to their peers actions and desires, andthey actively influence one anothers behavior andgoals. For example, between 27 and 40 months chil-dren are more likely to take their peers desires orneeds into account, such as offering or sharing toysthat the peer actually wants (Smiley, 2001); they alsoproduce more other-focused speech during peer playand cooperation, such as asking or telling the peer todo something (Brownell & Carriger, 1991; Eckerman& Didow, 1996; Smiley, 2001). Therefore, childrenscooperative activities with peers appear to beginwith imitative games and simple routines at the endof the 2nd year, and become more coordinated andfully cooperative during the 3rd year when childrenconsider, accommodate to, and influence one an-others actions and goals. However, little is knownabout the earliest forms of peer cooperation, whenthey emerge, or under what conditions very youngchildren can share activities and goals to cooperatewith one another.

Interestingly, this general progression in earlypeer cooperation is consistent with the sequence ofcomplexity in coordinated behavior identified in aformal analysis of cooperative hunting among non-human primates (Boesch & Boesch, 1989). Althoughnot developmental, this analysis presents a detailedarticulation of four increasingly complex levels ofcoordinated behavior that may provide an analog forage-related changes in childrens early cooperation.At the lowest level, hunters perform similar actionstoward the prey at the same time but without anyother spatio-temporal relations among them, i.e.,they act independently and do not consider the be-havior of their partners. At the second level, they stillbehave similarly but now coordinate their behaviortemporally, for example adjusting their relativerunning speeds to remain near each other. At thethird level they adjust both their position and theirspeed relative to each other, and at the highest levelthey adopt complementary roles and actions, i.e.,different hunters behave differently such as onechasing the prey and one blocking the escape. In thisframework, actors at the lowest level do not takeothers or their behavior into account although theiractions are sometimes coordinated, while at thehighest level actors behavior appears genuinelycollaborative, with multiple accommodations intime, space, and behavior relative to their partnersand the goal. Thus, cooperation varies from simple tocomplex as a function of how the actors organizetheir behavior relative to one another and to thecommon goal.

This conceptual framework can serve as a usefulheuristic for characterizing developmental progres-

sions in young childrens cooperation. Primitiveforms of coordinated activity among 1-year-olds maybe based on similar but independent actions per-formed at the same time toward a common goal suchas acting together on the same toy (Eckerman et al.,1989; Mueller & Brenner, 1977), which then progressto spatially and temporally coordinated cooperativebehavior in 2-year-olds based on a shared goal(Brownell & Carriger, 1990; Howes, 1988). Complexcoordinations of roles and behavioral timing withmultiple shared goals emerge still later among 3-year-olds and indicate explicit collaboration (Ashley& Tomasello, 1998; Smiley, 2001). One purpose of thecurrent study was to examine age differences in theearliest forms of childrens cooperation with peers.To this end, a cooperative task was designed thatpermitted children to behave imitatively in pursuinga single goal to correspond to the putatively mostprimitive level of coordination based on similar ac-tions produced at the same time. We also designed asequential version of the task that required temporalcoordination of behavior to achieve the goal, tocorrespond to the second, more difficult level. Weexpected that 1-year-olds actions would be coordi-nated largely by chance if they act at the same timebut independently and cannot yet share the goal andorganize their behavior jointly around it, whereas2-year-olds would be able to share the goal and co-ordinate their behavior with one another more sys-tematically to achieve it.

Social Understanding and Peer Cooperation

The preceding conceptualization of the developmentof cooperation is anchored in assumptions aboutyoung childrens social understanding, especiallywhat they understand about one anothers inten-tions and goals. At the simplest level identifiedabove, in which actors perform similar actions in-dependently while achieving a single goal, primitiveforms of coordination do not necessarily rest onawareness of a partners intentions and goals, andcoordinated activity can be achieved without ac-tively considering the partners behavior. However,the ability to cooperate by coordinating behaviortemporally and spatially with a partner presumablydoes require such awareness (Brownell & Carriger,1991; Moore, in press; Povinelli & ONeill, 2000;Smiley, 2001; Tomasello, 1999a). Developmentalchanges in early peer cooperation may thus be atleast partly a function of growth in childrens rep-resentation of their own and others intentions andgoals, and of relations between their own behaviorand their partners behavior. Links between social

804 Brownell, Ramani, and Zerwas

understanding and social behavior have rarely beeninvestigated in this age group, however. Thus, thesecond purpose of this research is to examine howyoung childrens emerging ability to cooperate witha peer partner relates to their early social under-standing.

Early Developments in Social Understanding

Recent empirical work points to rapid and pro-found changes in the first 2 years of life in childrenssocial understanding. Particularly relevant for thecurrent study, considerable evidence suggests that1-year-olds attend to and perhaps understand othersperceptions, desires, intentions, and goals (Brooks& Meltzoff, 2002; Caron, Kiel, Dayton, & Butler,2002; Carpenter, Nagell, & Tomasello, 1998; Dunphy-Lelii & Wellman, 2004; Gergely, Nadasdy, Csibra,& Biro, 1995; Kuhlmeier, Wynn, & Bloom, 2003; Moll& Tomasello, 2004; Repacholi, 1998; Tomasello &Haberl, 2003; Wellman, Phillips, & Spelke, 2002;Woodward, 1999, 2003), and are able to take theseinto account in reasoning about others actions(Behne, Carpenter, Call, & Tomasello, 2005; Carpenter,Call, & Tomasello, 2005; Gergely, Bekkering, &Kiraly, 2002; Liszkowski, Carpenter, Henning,Striano, & Tomasello, 2004), even when others in-tentions are not fully realized in their actions(Bellagamba & Tomasello, 1999; Carpenter, Ahktar, &Tomasello, 1998; Meltzoff, 1995) and when othersdesires conflict with the childs own (Repacholi &Gopnik, 1997). Thus, by 915 months of age infantsappear able to detect others intentions and goals intheir behavior and to use this awareness to governtheir own behavior.

This nascent social understanding continues todevelop over the 2nd year and beyond. For example,1-year-olds are limited in the range of actions theycan detect as goal directed (Poulin-duBois & Forbes,2002; Sodian & Thoermer, 2004), whereas 2-year-oldscan infer agents intentions from a variety of moreand less subtle behavioral cues (Hollich, Hirsh-Pas-ek, & Golinkoff, 2000; Poulin-duBois & Forbes, 2002).Moreover, 1-year-olds can read adults goals fromtheir actions (Meltzoff, Gopnik, & Repacholi, 1999),i.e., they appear to understand intentions-in-action, but they do not yet understand prior inten-tions, the mental states that precede and motivateactions (Meltzoff, 1995). The latter understandingdevelops over the 2nd and 3rd years of life as chil-dren come to understand others as knowledgeableabout the world and not only as actors (Baldwin &Moses, 1996; Carpenter, Call, & Tomasello, 2002;Moore, in press; Smiley, 2001). Growth in social un-

derstanding over this age period is reflected in manyaspects of early development, including expandedability to use and respond to others communicativeintentions (Baldwin, 1993, 1995; Schwe & Markman,1997; Smiley, 2001; Tomasello, 2001) as well asgrowth of internal state words to refer to othersemotions, desires, perceptions, and physiologicalstates (Bartsch & Wellman, 1995; Bretherton & Bee-ghly, 1982; Shatz, Wellman, & Silber, 1983), increas-ing use of action words to describe or comment onones own or others goal-directed behavior (Bartsch& Wellman, 1995; Smiley, 2001), and using personalpronouns to denote the distinct perspectives of selfand other on the world (Bates, 1990; Lewis & Ram-say, 2004; Ricard, Girouard, & Decarie, 1999).

Links With Early Cooperation

Insofar as the ability to cooperate depends onbeing able to represent and to share goals and in-tentions with a partner, growth in early social un-derstanding should support developing cooperativeabilities. Infants can cooperate with adults in simplegames and routines by 1218 months of age (Hay,1979; Ross & Lollis, 1987; Warnecken, Chen, Liebal,& Tomasello, 2005). But when they must take a moreactive role in structuring the interaction, their abilityto cooperate with adults improves dramatically after24 months of age (Rutter & Durkin, 1987). Similarly,as reviewed above, children are unable to cooperatewith peers until the close of the 2nd year of life orlater. This suggests that 1-year-olds ability to attendto adults goal-directed behavior and to detectadults intentions may not yet support the ability tointegrate their own intentions with a peers inten-tions to achieve goals jointly. As childrens socialunderstanding develops over the 2nd and 3rd yearsof life, perhaps through cooperative games and in-teractions with adults, and children are better able toinfer and to integrate their own and others inten-tions (Barresi & Moore, 1996), they should be able tocooperate with one another more systematically aswell, recognizing and generating joint intentions andgoals. Thus, they should become progressively moreable to take their peers goal-related activity intoaccount in concert with their own, and to adjust theirown behavior accordingly by monitoring, timing,and sequencing their behavior together with the peerto attain a shared goal.

The Current Study

To examine age-related changes in early coopera-tion with peers and associations with early social

Toddler Peer Cooperation 805

understanding, we created a task in which 18- to 30-month-old children had to coordinate their activitywith one another to achieve a single goal. The goalwas provided by the task itself so that the childrendid not have to invent or discover it on their own,and the means to achieve the goal were demon-strated by an adult. The task also permitted childrento capitalize on the imitative skills that they use intheir earliest instances of joint play with peers (As-endorpf & Baudonniere, 1993; Camaioni, Ba-umgartner, & Perucchini, 1991; Eckerman et al.,1989). It was thus well within the capabilities ofyoung toddlers and much simpler than previouslyused cooperation tasks (Ashley & Tomasello, 1998;Brownell & Carriger, 1990), making it possible toobserve very early forms of peer cooperation.

We constructed two versions of the task; in bothversions each child had to perform the same singlebehavior (pull a handle) to activate an interesting toy.In the first version each child pulled a separatehandle and the two handles had to be pulled at ap-proximately the same time to activate the toy. In thesecond version children again performed the samebehavior with two separate handles, but sequential-ly, pulling the handles one after the other. Both ver-sions of the task were first demonstrated to thechildren by an adult; in the second version childrenwere also individually trained to produce the se-quential handle-pulling actions on a smaller versionof the task before being presented with the larger,cooperative version.

On the first version of the task, in which the goalcould be achieved imitatively and required nominaltemporal coordination, coordinated activity couldarise adventitiously, with minimal social under-standing, as a result of independently performingsimilar behaviors to achieve the same goal. Or itcould be the result of the children using theiremerging intention understanding to share the goalwith their peer partner and cooperating to achieve it.On the basis of the preceding conceptualization weexpected age differences on this version of the task,with 1-year-olds coordinations remaining primitiveand infrequent, as a result of failing to share the goalwith their partner, whereas older children would beable to cooperate more readily and with greater co-ordination of their goal-related activity. The secondversion of the task required more complex temporalcoordination in the form of turn taking. We expectedthis version to be more difficult for all children, butthat older children would again succeed more oftenby explicitly cooperating to achieve a shared goal.

To index early social understanding, we assessedchildrens ability to share attention with an adult and

their language about self and other. These representdifferent but related components of understandingothers perceptions, intentions, desires, and goals.We assessed childrens ability to share attention withan adult because gaze and attention provide infor-mation about what someone could be wanting orintending (Dunphy-Lelii & Wellman, 2004; Wellmanet al., 2002), especially when combined with the di-rection of their body movements (Mundy, in press).Establishing joint attention by following anothersgaze or point may also require that the child repre-sent the others communicative intention (Baldwin,1993; Brooks & Meltzoff, 2002; Dunphy-Lelii &Wellman, 2004; Tomasello, 1999b), and some scholarshave suggested that joint attention is necessary forcooperation (Brinck & Gardenfors, 2003).

Second, we assessed language about self andother, particularly personal pronouns, action words,and internal state words. Language development hasbeen shown to relate to the development of mentalstate understanding across a variety of paradigms,and one central mechanism appears to be the childsexposure through discourse to varying perspectivesbetween self and others (Harris, deRosnay, & Pons,2005; Symons, 2004). For example, toddlers under-standing of others behavior and internal states isassociated with their mothers talk about the chil-drens own goals in response to the childrens re-quests (Smiley & Greene, 1995), and familyconversation at age 2 about feelings and mentalstates predicts a childs emotion understanding andemotion talk in the preschool years (Dunn, Brown,Slomkowski, Tesla, & Youngblade, 1991; Jenkins,Turrell, Kogushi, Lollis, & Ross, 2003). Syntax ac-quisition may also provide children with the meansfor representing reality as embedded in mental states(He thinks its really a pencil) (Astington & Jenk-ins, 1999; deVilliers & deVilliers, 2000; Lohmann &Tomasello, 2003; but see Cheung et al., 2004), andeven earlier in childhood, the means for encodingand coordinating intentions (He wants me to getit) (Bartsch & Wellman, 1995; Smiley, 2001).

In very young children, before syntax, develop-mentally earlier aspects of language also relate toemerging social understanding (Bloom & Tinker,2001; Farrar & Maag, 2002; Hollich et al., 2000).Relevant to the current study, personal pronouncomprehension and use reflects an understanding ofthe complementary and reversible equivalence rela-tion between self and other (Bates, 1990; Kagan, 1981;Lewis & Ramsay, 2004; Ricard et al. 1999). This issimilar to the self other representations involved injoint goal-directed behavior with a partner (As-endorpf & Baudonniere, 1993; Brownell & Carriger,


1990, 1991). Similarly, early verb usage may build onchildrens understanding of intentional action inthemselves and others (e.g., Kagan, 1981; Poulin-Dubois & Forbes, 2002; Smiley, 2001; Smiley & Hut-tenlocher, 1995), which is critical to being able toshare intentions and goals with others. Finally, in-ternal state terms are presumed to reflect childrensearly understanding of others psychological states,including others goals and intentions (Bretherton &Beeghly, 1982; Shatz et al., 1983). Insofar as earlylanguage can act as a catalyst in developing an initialunderstanding of others internal states (Baldwin &Moses, 1996; Shatz, in press), children with moreadvanced language about self and others may be alsobetter able to use their linguistic skills to learn aboutand to infer others intentions, desires, and goals.Thus, because young childrens language both con-tributes to and indexes their growing social under-standing, we expected these language-basedindicators of early social understanding to be asso-ciated with developments in childrens ability tocooperate with a peer partner to achieve a joint goal.

The first hypothesis examined was that the de-velopment of cooperation with peers progressesfrom a relatively primitive form in 1-year-olds, basedon chance coordinations of independent but similaractions, with little evidence of understanding orsharing the peer partners intentions or goals, tomore genuine cooperation in 2-year-olds, whenchildren can coordinate their behavior by sharing thegoal and monitoring and accommodating their be-havior to one another and their joint goal. We thusexpected age differences in the frequency of coordi-nated behavior with a peer, in monitoring the peersactivity when children themselves were attemptingto achieve the goal, in pulling the handle when thepeer was nearby and available as a partner, and ininterference with the peers goal-directed activity.Second, we hypothesized that developments in co-operation would occur in concert with growth inchildrens social understanding, and thus that chil-dren who were better at coordinating attention withan adult and more skillful in using language refer-ring to self and other would be more successful incooperating with a peer.

Method

Participants

Participants were 88 children (48 girls) dividedinto three age groups: 19 months (SD5 0.91; N5 34;18 girls); 23 months (SD5 0.95; N5 28; 16 girls); and27 months (SD5 1.1; N5 26; 14 girls). Children were

scheduled together with a same-age (within 1month), same-sex unfamiliar peer. Families werefrom a medium-sized urban area and varied fromworking class to upper middle class by parent re-port; 94% were Caucasian, 5% African-American,and 1% Hispanic. Most children had some play-group or child-care experience; data from the chil-dren who were in child care did not differ from thosewho were not. Most children had no siblings; thenumber of siblings did not differ by age and therewere no sibling effects on any dependent measure.Several recruited dyads could not be used for thefollowing reasons: one child in the dyad refused toengage the cooperation task (N5 9); one child in thedyad became distressed (N5 2); experimenter error(N5 2).

General Procedure

The session began with a 1015-min period ofwarm-up freeplay, which included a standard set oftoys that children could play with individually ortogether. The cooperation task was presented in oneroom and videotaped through a one-way mirror. TheEarly Social Communication Scales (ESCS, describedbelow) were administered individually in a secondroom, with a video camera placed unobtrusively in acorner approximately 2.5m away from the child.Parents remained with their children at all times.

Cooperation Task

The cooperation task was modeled after tasksused by Ashley and Tomasello (1998), Brownell andCarriger (1990), and Chalmeau, Visalberghi, andGallo (1997). It consisted of a colorful wooden box(1m .4m .3m) with two small plastic handlesprotruding horizontally from one side of the box (seeFigure 1), mounted on a small table approximately.4m high. The handles were just below shoulderheight for most children and moved freely back andforth, in and out of the box. An animated musical toy,mounted on top of the box behind a Plexiglas barrier,was surreptitiously activated by remote controlwhen both handles were pulled. From the childrensperspective, their handle pulling was the apparentcausal mechanism. To make the task cooperative, thetwo handles were mounted too far apart for onechild to reach them both, requiring each child to pullone of the handles.

A female experimenter demonstrated the taskthree times in succession, pulling both handles, witha simple directive, Watch! If you pull both handles,the doggie will sing! Parents remained uninvolved.


The experimenter encouraged the children withnondirective comments. Once the children weresuccessful, the task was repeated as often as thechildren wished. If one or both children were notengaged, or if they were unsuccessful after the 1stminute, the experimenter initiated a series ofprompts such as Can you help each other? or Pullyour handles to make the doggie sing. These moredirective prompts were infrequent, occurring infewer than a third of the dyads, generally only once.Two versions of the task were presented to all chil-dren, as described below.

Simultaneous actions version. In this version of thetask, the two handles were the same color. The ex-perimenter demonstrated the task by pulling thehandles simultaneously. Each child had to pull onehandle within 3 s of the other childs pull of the otherhandle; no other coordination was necessary to ac-tivate the toy. For example, one child might repeat-edly push and pull one handle back and forth whilethe other child pulled and held the second handleextended. Such minimal coordination would besufficient to activate the toy. This version of the taskwas always presented first because both our con-ceptual analysis and pilot testing established thatthis version was easier; by presenting it first, wehoped to maximize the youngest childrens successon the second, more difficult version of the task.

Sequential actions version. In this version the twohandles were different colors. The experimenterdemonstrated that each handle had to be pulled inturn, one after the other, to activate the toy; eitherhandle could be pulled first, and one handle wasfully extended before initiating the second handlepull. Children had to pull the second handle within3 s of the first; if the children pulled the handles

simultaneously, nothing happened. Thus, this versionrequired greater temporal coordination than the firstversion and the children had rudimentary roles toplay, one child starting and the other child following.Before they were presented with the second version

of the task, children were individually trained to pullthe two handles sequentially on an identical smallerversion (0.3m 0.6m 0.35m) complete with ani-mated toy, placed on the floor; the children couldreach both handles on this smaller version. The ex-perimenter used a verbal cue while training the childthat would be used later as a reminder in the coop-eration version: Pull both handles, One! Two! Allchildren were successfully trained to a criterion ofthree successive correct handle pulls or four totalcorrect handle pulls.The second version of the cooperation task was

administered much like the first. In this case, whenthe experimenter demonstrated the task she used theverbal cue to remind the children of their experiencewith the individual training task and of the sequentialactions. There was no evidence of fatigue or loss ofinterest on the second version of the task as a result ofits being presented last.

Cooperation Coding

The videotapes from each task were event codedusing the Noldust Observer 3.0 observation soft-ware. Specific task-related behaviors adapted fromBrownell and Carriger (1990) and Chalmeau et al.(1997) were defined (see Table 1 for definitions).These were grouped into four categories: coordinat-ed behaviors; communicative behaviors; uncoordi-nated behaviors; time on task (a single durationscore). The set of coordinated behaviors reflects thechildrens appropriate timing of their handle pullingrelative to one another, their appropriate location atthe task and the handles relative to the peers loca-tion, and their monitoring of the peers activity andlocation at the task. Communicative behaviors indexchildrens gestures and verbalizations to one anotherabout the task. Uncoordinated behaviors reflect thevarious ways in which children could interfere withone anothers activity at the task as well as their at-tempts to achieve the goal on their own without thepeer.

Interobserver reliability was established betweentwo raters who independently coded tapes from 31children (35%), approximately equally distributedover age and the two versions of the task. Percentagreement and interrater correlations are shown inTable 1. Because individual behaviors in the codingsystem were not mutually exclusive (e.g., monitoring

Figure 1. The cooperation task.


peers and pulling the handle could be coded at thesame time), ks were not calculated. Directed coor-dinated pulls and declarative communications wereinfrequent, and were excluded from analyses be-cause of their low reliability.

Social Understanding

Two measures of social understanding were ob-tained. One was based on a procedure individuallyadministered by a female experimenter to indexchildrens social understanding in social and com-municative exchanges with adults. The second wasbased on questionnaires about childrens languagethat parents completed during the visit.

Joint attention. Each child was individually ad-ministered the ESCS (Seibert, Hogan, & Mundy,1982), a standardized, structured procedure to indexsocial understanding in children between 8 and 30months of age using nonverbal communicationtasks. Childrens behavior was coded from video-tapes using the criteria established by Mundy, Hog-

an, and Doehring (1996). Joint attention indexes thedegree to which children understand and respondappropriately to others communicative and socialintentions and is the measure used in the currentstudy. Joint attention is further categorized into ini-tiating joint attention (IJA) and responding to joint at-tention (RJA). High-level joint attention (child pointsor shows to share experiences with the experimenter;child turns head and eyes to follow experimentersvisual regard and pointing gesture to something outof his or her immediate vision) is presumed to reflectcommunicative intent or awareness of anotherscommunicative intent, and is distinguished fromlower level behaviors such as simple eye contact.Because understanding others goals and intentionsis at the heart of cooperation, we focused on thesehigh-level dimensions. The frequencies of IJA andRJA were standardized for analysis. Interobserverreliability was established between two raters whoindependently coded data from 24 children (27%)approximately equally distributed across age. Per-cent agreement ranged from 79.5% to 90%. Correla-

Table 1

Behaviors Coded From Cooperation Task and Interrater Reliabilities

Behavior Definition % Agree

Interrater

correlation

Coordinated behaviors

Spontaneous coordinated pull Child pulls handle within 3 s of peers handle pull 96 0.95

Directed coordinated pull Child pulls handle within 3 s after highest level verbal directive from adult 50 0.03Cued coordinated pull Child pulls handle within 3 s after One . . . Two! cue from adult

(Sequential task only)

90 0.87

Monitor partners activity Child focuses gaze on peers face or hands within 2 s preceding or

following his/her own handle pull

97 0.81

Peer proximal pull Child pulls his or her own handle when peer is within arms length of other

handle

97 0.98

Communicative behaviors

Declarative communications Child gestures or verbalizes to peer about task in a nondirective manner

(e.g., points to toy or says Doggie!)

28 0.94

Imperative communications Child gestures to peer about task in a clearly directive manner (e.g., points

to peers handle while looking at peer) or verbally directs peers

behavior on task or requests peer to perform an action (e.g., Pull it,

Help)

75 1.00

Uncoordinated behaviors

Individual pull Child pulls own handle and peers handle is not pulled within 3 s

preceding or following

98 0.97

Control Child interferes with peers manipulation of his or her handle (e.g.,

displaces peer at handle)

100 0.85

Join Child moves next to peer at peers handle 92 0.86

Leave task Child moves more than arms length away and disengages from task for 5 s

or longer

81 0.90

Time on task

Duration (s) Child is facing task and visually or manually engaged, with interruptions

of less than 5 s

na 0.94


tions between the two raters for the frequency ofhigh-level IJA and high-level RJA over the 24 chil-dren were r5 .89 and .95, po.001, respectively.

Language about self and others. Parents completedthe toddler form of the MacArthur CommunicativeDevelopment Inventory (CDI; Fenson et al., 1994), awidely used instrument to assess language compre-hension and production in 12- to 36-month-oldchildren. It yields frequency measures for specificcategories of words and gestures as well as foroverall vocabulary. Because we were interested inwords referring to self and other and to goals andintentions, we used the Pronouns subscale and theAction Words subscale. These scales indicatewhether the child understands (scored as 1) or un-derstands and says (scored as 2) eight personalpronouns (e.g., her, you, me) and 55 action words(e.g., bite, feed, run). The scores were summed overall words checked (M5 81.8; range5 12126 for thetwo scales combined).Parents also completed the State Words Checklist

(SWCL; see appendix), adapted for this study fromBretherton and Beeghly (1982) and Shatz et al. (1983),to index childrens use of words referring to emotion,perception, desire, intention, pretense, thought, andphysiological state. The parent indicated for each of70 words how often the child had used the word inthe past 6 months (05 never; 15 once or twice; 25 35 times; 35 often). Ratings were summed over all

words for a total internal state words usage score(M5 44.1; range 5 0155).Two composite scores were created. Childrens

scores on the two CDI scales and the SWCL werecorrelated; therefore, the three scores were stand-ardized and summed to create a total self otherlanguage score (Cronbachs a5 .90). Because olderchildrens generally larger vocabularies could ac-count for age differences on these subscales, we alsoweighted childrens individual scores by creatingproportions for each one with the total CDI as thedenominator. We then standardized and summedthese proportion scores to yield a weighted summaryscore of language about self and others.

Results

Means and standard deviations for the dependentmeasures are presented in Table 2. Preliminary t testsconfirmed that there were no sex differences; thus allanalyses were conducted on data collapsed over sex.T tests were also conducted to test for age differencesin the amount of time that children spent on task andfor differences between the two task versions fortime on task. There were no significant differences.

We first report results for age and task-versiondifferences in childrens ability to coordinate theirbehavior with the peer, using the measures of coor-dinated handle pulls and other coordinated behavior

Table 2

Means and Standard Deviations for Childrens Performance as a Function of Age

19 months 23 months 27 months

M SD M SD M SD

Cooperation taska

Coordinated behavior

Spontaneous coordinated pulls 3.03 2.9 5.66 3.7 7.85 3.1

Cued coordinated pulls 0.50 0.55 0.95 0.66 1.67 0.79

Peer proximal pullsb 3.46 2.6 5.89 3.9 6.44 5.3

Monitor peer 5.38 2.5 7.23 2.1 8.80 3.3

Imperative communication 0.21 0.20 0.35 0.47 0.48 0.83

Uncoordinated behavior

Individual pulls 8.04 2.2 9.38 5.6 10.10 7.9

Other uncoordinated behavior 1.94 1.4 0.91 0.90 1.50 1.9

Joint attentionc

Initiate 2.9 3.5 5.4 6.5 4.1 3.6

Respond 2.5 0.83 2.9 0.64 2.9 1.0

Self/other Languaged 1.2 1.9 0.15 1.7 0.96 1.5aFrequencies averaged over the two children in a dyad and over the two versions of the task.bDoes not include coordinated pulls.cFrequencies of high-level joint attention.dZ scores.


(see Tables 1 and 2 for definitions and descriptivestatistics). We supplement these analyses with per-son-level non-parametric analyses to clarify andstrengthen the findings. Second, we report results forage and task-version differences in uncoordinatedbehavior, i.e., individual handle pulls and other un-coordinated behavior (see Tables 1 and 2). Third, wereport results from a series of analyses addressingwhether coordinated behavior was coincidental orcooperative, also considered as a function of age.Fourth, we report results for age differences in themeasures of social understanding (joint attention;language about self and others), and finally wepresent analyses of associations between measures ofpeer cooperation and measures of social under-standing.

Age and Task Differences in Coordinated Behavior

To test for age and task-version differences in co-ordinated behavior, we conducted a 2 3 repeatedmeasures multivariate analysis of variance (MA-NOVA) with task version (simultaneous; sequential)as the within-subjects factor and age group as thebetween-subjects factor (19 months; 23 months; 27months) on the set of coordinated behavior measures(spontaneous coordinated pulls; monitor peer; peerproximal pulls). Significant multivariate effectsemerged for age, F(6, 78)5 2.80, po.05, and for taskversion, F(3, 39)5 14.45, po.001, on the measures ofcoordinated behavior (see Figure 2). The interactionbetween age and task version was not significant(similar results were obtained using age as a con-tinuous variable in regression analyses). The find-ings from the post hoc univariate analyses followingthe main effects for age and task version are reportedbelow.

Age differences: Coordinated pulls. Across tasks,older children more often coordinated their handlepulling with their peer partner than did youngerchildren (see Table 2), F(2, 41)5 7.37, p5 .002. Spe-cifically, 19-month-old dyads produced the fewestcoordinated pulls (M5 3.0), followed by 23-month-olds (M5 5.7) and 27-month-olds (M5 7.6), all ofwhom differed significantly from one another. Like-wise, the proportion of toddler dyads total handlepulls that were coordinated with the peer doubledbetween 19 and 27 months, from 27% at 19 monthsto 54% at 27 months, a significant difference, F(2, 41)5 4.8, po.01. An additional one-way ANOVA withage group as the factor was conducted on cued co-ordinated pulls, which were unique to the sequentialversion. Significant effects emerged for age on cuedcoordinated pulls, F(2, 41)5 11.46, po.001. Theyoungest toddlers produced significantly fewer cuedcoordinated actions on the sequential version of thetask than did older children (Ms5 0.50, 0.95, and1.67 for 19-month-olds, 23-month-olds, and 27-month-olds, respectively), and 23-month-olds pro-duced significantly fewer than did 27-month-olds.Thus, 1-year-old peers coordinated their behaviorrelatively infrequently overall, whereas 2-year-oldsdid so regularly.Not only were 1-year-olds efforts less often coor-

dinated, but more of them failed to produce anyspontaneous coordinated pulls at all. On the simul-taneous-actions version of the task, 24% of the 1-year-olds were unable to coordinate their behavior atall and another 12% could do so only once, unable torepeat their initial success. This compares to 14% ofthe 23-month-olds who never coordinated their be-havior with the peer, and 14% who could do so onlyonce. All of the 27-month-old dyads coordinatedtheir actions at least once and, with only one ex-ception, were able to cooperate multiple times. Theseage differences, while striking, were not significant.On the sequential version, for which they had re-ceived individual training, 29% of 1-year-old dyadswere unable to produce any coordinated handlepulling and another 35% were unable to coordinatetheir behavior more than once. This compares to 7%of the 23-month-old dyads with no coordination and14% with just one instance. All of the 27-month-oldswere able to cooperate with the peer multiple times.These differences were significant, likelihood ratio(df5 4)5 23.76, po.001.Finally, few 1-year-olds were able to achieve

multiple successes on both versions of the coopera-tion task whereas nearly all of the older 2-year-oldscould do so, suggesting that the older children weremore skilled at achieving a goal jointly. Only 30% of

0

1

2

3

4

5

6

CoordPulls

MonitorPeer

Peer ProxPulls

CoordPulls

MonitorPeer

Peer ProxPulls

19 m23 m27 m

SIMULTANEOUS-ACTIONS SEQUENTIAL-ACTIONS Figure 2. Age differences in frequency of cooperative behaviors foreach task version.


the 19-month-old dyads were successful more thanonce on both versions of the task, compared to 64%of the 23-month-old dyads and 92% of the 27-month-old dyads, w2 (2)5 12.28, po.01.

Age differences: Other coordinated behavior. We alsoexamined whether there were age differences inchildrens task-relevant monitoring of the peer andin adjusting their goal-directed actions in relation tothe peers location relative to themselves and thegoal (see Table 2). Univariate tests showed that the 27month-olds more often monitored their peer partnerwhile they were pulling their own handle (M5 8.80)than did 19-month-old children (M5 5.38),F(2, 41)5 6.18, po.01. The 27-month-old childrenwere also nearly twice as likely to pull their ownhandle when the peer was in proximity to the otherhandle (M5 6.44) than were 19-month-olds(M5 3.44), although the univariate test was onlymarginally significant, F(2, 41)5 2.49, po.10.A separate 2 (task version) 3 (age group) re-

peated measures univariate ANOVAwas conductedon imperative communication because it was notcorrelated with other measures. There were no sig-nificant effects for either age or task version on im-perative communications.

Task differences: Coordinated pulls. The sequentialversion of the task proved more difficult for allchildren than the simultaneous-actions version,F(1, 41)5 34.12, po.001, with fewer spontaneouscoordinated pulls per dyad in the sequential(M5 3.6) than in the simultaneous version (M5 7.2).

Task differences: Other coordinated behavior. Therewere no significant multivariate or univariate effectsas a function of task version on monitoring the peersactivity at the task or in pulling the handle when thepeer was near the other handle (see Figure 2).

Age and Task Differences in Uncoordinated Behavior

A second 2 3 repeated measures MANOVAwith task version (simultaneous; sequential) as thewithin-subjects factor and age group as the between-subjects factor (19 months; 23 months; 27 months)was conducted on the measures of uncoordinatedbehavior (individual pulls; sum of control, join, andleave task). There were no significant multivariateeffects for either age or task version on uncoordi-nated behavior. Children at all three ages oftenpulled the handles by themselves and these indi-vidual attempts to achieve the goal did not declinesignificantly with age (see Table 2). Moreover, thefrequency of childrens individual pulls was notsignificantly associated with their coordinated pulls,r(44)5 .016, ns; thus pulling the handle by oneself

more often was not related to greater success on thetask. Children also engaged in a variety of other in-dividual behaviors that either did not consider theirpartners role and activity in achieving the goal (i.e.,join) or interfered with rather than coordinated withthe partners behavior (i.e., control; leave task); thesealso did not decline significantly with age. Not sur-prisingly, a dyads uncoordinated behavior wasnegatively associated with the frequency of theircoordinated handle pulls, r(44)5 .33, po.05;therefore, engaging in noncooperative behavior wasrelated to lower chances of success. Nevertheless,children at all three ages engaged in similar rates ofthese unproductive behaviors. Hence, even the 27-month-old children continued to display relativelyhigh frequencies of uncoordinated behavior at thesame time as they were also able to coordinate be-havior with a peer to achieve a shared goal.

To What Extent Was Coordinated Behavior Coincidental?

Because the task used in this study was simplifiedto permit primitive, low-level forms of cooperativeactivity, it was possible for childrens behavior to becoordinated without actively taking into accounttheir partners goal-directed actions. That is, childrencould simply attempt to achieve the goal individu-ally when both happened to be positioned at thetask. In this case, the goal is not shared; neither childis adjusting or accommodating behavior to the otherin the service of a common goal, rather they simplyhappen to act in parallel as they each independentlypull their own handle to achieve their own personalgoal. It is important, therefore, to determine not onlywhether older childrens behavior was more oftencoordinated with their peers behavior, but alsowhether they were actually more cooperative, that is,whether they appear to have understood their part-ners goal and to have shared it. There are severalindirect means of addressing this question that con-verge into a coherent pattern. These strategies areadapted from work with nonhuman primates inwhich similar issues were addressed (e.g., Chalmeauet al., 1997; Povinelli & ONeill, 2000; Visalberghi,Quarantotti, & Tranchida, 2000).

First, we asked whether older dyads were morelikely to pull their handles cooperatively than indi-vidually. The ratio of dyads coordinated pulls toindividual pulls increased significantly with age,albeit marginally, F(2, 41)5 2.99, p5 .056, with theyoungest children differing significantly from theoldest group. The 19-month-old dyads more oftenpulled their handles individually than cooperatively;indeed, they were more than twice as likely to pull


the handle individually than cooperatively (M ratio ofcoordinated pulls:individual pulls5 0.40). The 27-month-olds, on the other hand, were almost twice aslikely to pull their handles cooperatively than indi-vidually (M ratio5 1.6). The 23-month-old dyads fellin the middle, pulling about equally often individu-ally and cooperatively (M ratio5 1.1), and did notdiffer from either the youngest or the oldest groups.Thus, the predominant pattern in the 1-year-olddyads was to try to achieve the goal independently,whereas the predominant pattern among the 2-year-olds was to attempt to achieve the goal together. Thissuggests that coordination in the youngest dyads mayhave been serendipitous, whereas the oldest dyadshad figured out how to achieve the goal cooperatively.

Next, we asked whether among all of a childshandle pulls, more of them occurred when the peerwas near the other handle and available as a partner.The ratio of peer-proximal pulls to all handle pullsincreased significantly with age, F(2, 41)5 8.57,po.001. Post hoc tests showed that the youngestdyads differed from both older groups, and the twoolder groups did not differ significantly from oneanother. Among the 27-month-olds, 83% of theirhandle pulls occurred when the peer was availableas a partner and 78% of 23-month-olds handle pullsoccurred when the peer was nearby, in contrast to55% of handle pulls among the 19-month-olds. Thus,among 1-year-olds only half of their handle pullingoccurred when their peer was close enough to theother handle to serve as a partner, whereas amongthe older children most of their handle pulling oc-curred when the partner was near the handle. One-year-olds generally were not taking their partnersposition into account when they pulled their ownhandle, but 2-year-olds were considering the part-ners location in relation to themselves and the goal.

Finally, we asked whether childrens individualhandle pulls were more probable when a partnerwas available than when the child was all alone atthe task. That is, when children pulled the handle bythemselves, was it mostly when they were all aloneat the task or was it mostly when the peer was nearthe task? To address this question, we compared thefrequency of each childs solitary handle pulls, whentheir peer partner was off-task and more than armslength away from the other handle, with the fre-quency of his or her peer-proximal individual handlepulls (see Figure 3). Among 19-month-olds, therewas no significant difference between the number ofindividual handle pulls when children were alone atthe task (M5 4.6) and when the peer was present(M5 3.4). However, in the two older age groupschildren pulled their own handles significantly more

often when the peer was nearby (M5 5.9, 6.5 for 23-month-olds and 27-month-olds, respectively) thanwhen the children were alone at the task (M5 3.5,3.6, respectively), paired samples t(27)5 2.04, po.05,and t(25)5 2.56, po.05, respectively. Thus, whenolder children were pulling their handles on theirown, without a partners coordinating pull, theywere more likely to do so when the peer was near theother handle than when no partner was available,capitalizing on the possibility that the peer wouldbegin to pull his or her handle and the two couldachieve the goal together. One-year-olds, in contrast,were just as likely to pull their handle when theywere all alone at the task as when a potential partnerwas nearby, indicating that their peers position andactivity did not affect their decision to pull their ownhandle in trying to achieve the goal.

Age Differences in Joint Attention and Self OtherLanguage

To identify age differences in the two indices ofsocial understanding, separate one-way ANOVAswere conducted with age group as the factor on thescores of individual children.

Joint attention. There were no age differences inhigh-level IJA (Ms5 2.9, 5.4, 4.1 at 19, 23, and 27months, respectively). For high-level RJA, the effectfor age was marginally significant, F(2, 85)5 2.41,po.10, and post hoc tests were not significant(Ms5 2.5, 2.9, 2.9 at 19, 23, and 27 months, respec-tively). Thus, communicating with the experimenterabout toys and objects by pointing to them and

0

1

2

3

4

5

6

7

19 mos 23 mos 27 mos

Pull When AlonePull When PeerNearby

Figure 3. Age differences in frequency of solitary and peer-proxi-mal individual handle pulls.


showing them did not change with age, whereasfollowing the experimenters gaze and pointing totargets outside the childs own visual field trendedupward between 19 and 23 months slightly, but notsignificantly.

Language about self and others. Age differencesemerged in the total self other language score,F(2, 74)5 19.89, po.001. The 27-month-olds hadhigher scores than did 23-month-olds, who hadhigher scores than 19-month-olds (M z-scores5 2.0,0.23, 1.9, respectively). The summary scoreweighted by total vocabulary size also exhibited agedifferences, F(2, 74)5 10.32, po.001, with 27-month-olds producing a higher proportion of languageabout self and others than did 23-month-olds, whoproduced a higher proportion than did 19-month-olds (M z-scores5 0.96, 0.15, 1.2, respectively).Thus, childrens production and comprehension ofpersonal pronouns, actions, and internal state wordsincreased between 19 and 27 months in both absolutenumber and as a proportion of their total vocabulary.

Associations Between Coordinating Behavior With Peersand Social Understanding

To determine whether childrens social under-standing predicted their individual behavior in at-tempting to cooperate with a peer, Pearsoncorrelations were conducted using individual scores(the same results were obtained when parallel anal-yses were conducted for each dyad member sepa-rately). Children who were better at following andsharing the adults attention (RJA) also coordinatedtheir handle pulling with the peer more often,r(88)5 .30, po.005, monitored their peer more oftenwhile they pulled their own handle, r(88)5 .25,po.01, and more often pulled their own handlewhen the peer was within arms length of his or herhandle, r(88)5 .18, po.05. Conversely, the rate ofuncoordinated behavior was negatively associatedwith RJA, r(88)5 .20, po.05. Similar associationsemerged with childrens language about self andothers (using the weighted summary score), whichwas positively correlated with childrens coordinat-ed handle pulls, r(88)5 .34, po.001, monitoring thepeer while pulling their own handle, r(88)5 .38,po.001, and peer-proximal handle pulls, r(88)5 .36,po.001 (parallel results were obtained when chil-drens self other language was categorized via me-dian split and then used as the independent variablein ANOVAs). Language about self and others wasnot correlated with uncoordinated behavior.

Generally these associations remained significantwith age partialled out, although reduced in mag-

nitude (rs 5 .18 .26), confirming that the relationsbetween social skill and social understanding werenot a simple artifact of general age trends. The twoexceptions were the association between RJA andpeer-proximal handle pulls which was no longersignificant with age controlled, and between lan-guage about self and others and coordinated handlepulls which became marginally significant,r(88)5 .15, po.10.

To determine the relative contributions of age,joint attention, and language about self and others tochildrens coordinated behavior, they were enteredtogether as predictors into separate multiple regres-sion analyses using coordinated pulls, monitor peerwhile pulling, and peer-proximal pulls as dependentmeasures. Three hierarchical linear regression mod-els were run, one for each dependent measure. Ineach analysis age was entered first, followed by RJA,followed by the weighted measure of language aboutself and others. The language measure was enteredlast because language develops later than, and de-pends on, joint attention (Baldwin, 1995; Tomasello,1999a). The results are shown in Table 3. For coor-dinated pulls, the model predicted nearly a third ofthe variance, and RJA continued to predict coordi-nated pulls with age controlled. However, languageabout self and others did not predict coordinatedpulls when both age and RJA were controlled. Formonitoring the peer while pulling ones own handle,the model explained about 17% of the variance, butonly language about self and others was predictiveafter controlling age and RJA. Age was associatedwith childrens tendency to pull their own handlewhen the peer was nearby, but neither RJA nor lan-guage about self and others was predictive in thisjoint model.

Table 3

Hierarchical Regression Predicting Toddlers Coordinated Behaviors on

the Cooperation Task From Their Age, Response to Joint Attention, and

Language about Self and Other

Coordinated

pulls

Monitor

peer

Peer-proximal

pulls

R2 0.32 0.17 0.27

F of R2 change 11.12 5.14 8.84

Agea 0.44 0.18 0.42

RJAa 0.24 0.09 0.04

Langa 0.06 0.27 0.15

Note. RJA5 response to joint attention, high; Lang5 languageabout self and othersastandardized bs.po.05. po.01. po.001.


Discussion

In this study we examined very early developmentsin childrens ability to cooperate with peers. Wefound age-related changes in toddlers ability to co-ordinate their activities with one another on a simpletask to achieve a common goal, and associations withtheir growing social understanding. Coordinatedactivity between 19-month-old peers was sporadicand primitive at best. Although they sometimesproduced coordinated behavior, their efforts ap-peared to be more coincidental than cooperative andmany dyads could not coordinate their behavioreven once. Thus, 1-year-olds were relatively un-skilled at taking a peers behavior into account evenwhen this was relevant for their own attempts toachieve a desired outcome. In contrast, 2-year-oldsappeared to be more actively cooperating toward ashared goal. By 27 months of age children wereconsiderably more skilled at coordinating their be-havior with a peer, both monitoring and accommo-dating to the partners activity and location as theyattempted to achieve the goal together. Childrensability to share attention with an adult and to un-derstand and use language about self and otherswere associated with their skill in peer cooperation.These findings suggest that children begin to becometrue social partners with one another late in the 2ndyear or at the beginning of the 3rd year of life, inconcert with developing social understanding,thereby initiating their entry into peer culture.

Development of Peer Cooperation in Toddlers

We capitalized on toddlers propensity to imitateothers to create a simple task with a single, moti-vating goal so that we could examine the earliestforms of cooperative activity among young peers. Inthe only other published study of goal-directed co-operation with peers in this age group, Brownell andCarriger (1990) found that 18-month-olds could notsystematically coordinate their behavior on tasksthat required complementary behavior. Similarly, onthe simpler tasks used in the current study, 19-month-old children performed quite poorly overalland they did not otherwise appear more capable orcooperative than did the 18-month-olds of Brownelland Carriger (1990). Although adults demonstratedand reminded the children of the common goal, re-moving the need for children to discover or generatea goal together, this was not sufficient to assist1-year-olds in sharing the goal and cooperating toachieve it. Even on the easier simultaneous-actionsversion of the task, a quarter of the 1-year-old peer

dyads were unable to generate any coordinated be-havior at all, and many others were able to generateonly one instance. In contrast, all of the 27-month-olds were able to cooperate readily with their peerpartners and did so often, repeating their initialsuccess several times, suggesting true cooperation.

Children at all three ages found it more difficult tocoordinate their behavior on the sequential versionof the task that required them to take turns. How-ever, nearly all of the older children were able to doso multiple times, whereas two-thirds of theyoungest dyads were never able to coordinate theiractivity or could do so only once. This is despite thefact that the children had been individually trainedon the turn-taking version: they knew that it tooktwo handles to activate the toy and they knew howto pull the handles sequentially to make it go. No-tably, the only difference between the individual,trained version of the sequential task and the coop-erative version was that a partner was needed in thelatter. In spite of their training, 1-year-olds wereunable to generalize their understanding from theindividual version to the cooperative version of thetask, even though they had already interacted withthe same partner on the easier task. Together thesefindings suggest that despite their social interest inone another, 1-year-olds are not yet aware of theirpeers as potential partners and cannot yet share in-tentions and goals with peers to create cooperativeactivity. These abilities emerge at the end of the 2ndyear and continue to develop into the 3rd year.

This conclusion is reinforced by the findingssuggesting that coordinated activity among 1-year-old dyads, when it occurred, was inadvertent,whereas 2-year-olds were more actively and explic-itly cooperating. The oldest dyads predominantlypulled their respective handles cooperatively, but 1-year-olds pulled their handles on their own morethan twice as often as they acted in concert. Moreo-ver, the youngest childrens efforts failed to take thepeers position or activity at the task into account;they tried just as often to achieve the goal when itwas impossible because they were all alone at thetask as they did when the peer was near the otherhandle and available as a partner. Older children, incontrast, were more likely to pull their own handlewhen the peer was nearby and available than whenthey were alone at the task. Thus, 1-year-olds coor-dinated their behavior much less frequently than did2-year-olds, they monitored one anothers goal-directed activity less, and they were more likely totry to achieve the goal entirely on their own. Thissuggests that 1-year-olds do not understand the roleof their peers as potential partners even at the most


elemental level. Their coordinated activity appearedto be serendipitous, a chance outcome of acting in-dependently to achieve their own goal irrespective ofthe whereabouts or activity of the peer and potentialpartner. In comparison, 2-year-olds were more oftencooperating, accommodating their behavior to thepeers location or activity, and thus acting togetherwith the peer as a partner in achieving the goaljointly.

Although the older children were better able tocoordinate their behavior with a peer, they werehardly perfect. Nearly half of the oldest dyads ef-forts were not coordinated with each other, andmany of their handle pulls occurred when the peerwas not at the other handle. Moreover, children at allages engaged in similar rates of uncoordinated be-havior, including leaving their own side of the task tojoin the peer next to the other handle and thusmaking it impossible to achieve the goal, or inter-fering with the peers activity by taking over thepeers handle, or leaving the task altogether as thepeer was pulling his or her own handle. Thus, it re-mains unclear how much even the 2-year-olds, as agroup, actually shared the goal. It is possible thatthey were responding more basically to the behavi-oral topology of the task, recognizing that their ownefforts were more likely to be successful when theother child was in a particular location or engaging ina particular action, but not understanding that theother child held the same intention or goal. They mayhave understood the peers goal-directed behaviorswithout understanding goals as internal states.

Together, these results extend prior research infinding that coordinated activity with peers emergesbetween 18 and 30 months of age even on highlysimplified cooperation tasks in which the joint goaland the means to achieve it are provided so thatchildren do not have to generate either by them-selves, the goal can be achieved imitatively withidentical behavioral roles, and requirements forspatial and temporal coordination of behavior areminimized. The findings are consistent with the hy-pothesis that early developments in peer cooperationprogress from a primitive form of simple, unsys-tematic, serendipitous coordination of activityamong 1-year-olds, with limited recognition of thepeer as a social partner, to more collaborative andtemporally coordinated cooperative behavior among2-year-olds that takes into account the partners ac-tions and location in relation to their common goal.Becoming a social partner with peers may begrounded in early forms of coordinated activity thatchildren can manage without fully understandingtheir peers as partners. Participating in such inter-

actions may help children progress from being ableto coordinate goal-directed actions with one anotherto being able to share goals and intentions coopera-tively.

Associations With Social Understanding

Although behavior can be coordinated betweenactors fortuitously, without understanding the part-ner as an agent of his or her own behavior or rep-resenting the partners goals or intentions, whichsome nonhuman primates can do (Chalmeau et al.,1997; Povinelli & ONeill, 2000; Visalberghi et al.,2000), uniquely human forms of cooperation dependon understanding, representing, and sharing goalsand intentions (Tomasello, Carpenter, Call, Behne, &Moll, 2005). We thus expected that very young chil-drens developing understanding of others desires,intentions, and goals would be reflected in theirperformance on the cooperation tasks. Consistentwith this hypothesis we found that children withmore advanced social understanding were betterable to cooperate with a peer partner.

Children can put their incipient social under-standing to work in cooperative activities withadults by late infancy (Hay, 1979; Ross & Lollis, 1987;Warnecken et al., 2005). As infants come to produceand share intentions with adults in the frame ofestablished activity (Nelson, in press), they candetect and interpret adults intentional actions anduse their own gestures and words to affect thoseactions. This also permits them to engage in coop-erative play with adults well before they can do sowith agemates. In the current study, 1-year-olds werenot able to cooperate with one another even thoughthe goal was demonstrated by adults and the adultsreminded them how to achieve it, much as in adult child cooperative play.

Cooperative activities with peers may represent aboundary condition in very young childrens abilityto reason about their own and others intentions andbehavior and to put their understanding to use ininteraction. To cooperate with one another, childrenmust integrate their own intentions and their peersintentions with respect to a common goal, whichmeans that they must be able to recognize or infertheir peers intentions and goals from their behavior.Toddlers nascent understanding of goals and in-tentions is put to a stringent test in exchanges withone another, in part because intentions are likely notas evident in peers behavior as they are in adultsbehavior when adults interact with young children.Adults communications are clearer than thoseof toddlers, and adults social behavior is more


predictable and perhaps more rational, even instrange experimental settings, than that of toddlerswith limited and rudimentary social skills (Ross &Kay, 1980). When others intentions and goals are notdirectly and unambiguously conveyed in their im-mediate behavior, very young children may find itespecially difficult to detect them, much less to join,accommodate to, or influence them. Peers may be abit of a mystery, their intentions inscrutable, theirbehavior often uninterpretable.

Moreover, in childrens early cooperative playwith adults, the adult partner structures and sup-ports the interaction, establishing and defining thejoint goal or discerning the childs goal and inten-tions and accommodating behavior accordingly(Bakeman & Adamson, 1984, 1986; Bruner, 1982;Rocissano, Slade, & Lynch, 1987). By uniquely timingand adjusting their behavior to the immaturities inchildrens social skill and social understanding(Bornstein et al., 1992; Callanan & Sabbagh, 2004;Flom & Pick, 2003; Kochanska & Aksan, 2004; Lillard& Witherington, 2004; ONeill, Bard, Linnell, &Fluck, 2005; Pan, Imbens-Bailey, Winner, & Snow,1996), adults help children integrate their own in-tentions with the adults intentions to achieve goalsjointly. Even when adults violate toddlers socialexpectations in experimental manipulations (Ag-netta & Rochat, 2004; Ross & Lollis, 1987; Warneckenet al., 2005), the violations are scripted within pre-dictable, contingent sequences of adult-definedgames and communicative routines (Ross & Kay,1980). However, children receive no such support orscaffolding from one another. Thus, putting theiremerging social understanding to work in coopera-tive interactions with peers may be especially chal-lenging for toddlers because it is difficult toapprehend or infer one anothers intentions andgoals from unclear or unpredictable behavior, andbecause the children must figure out by themselveshow to integrate their intentions with a peer partnerand organize their behavior with respect to a jointgoal when they have no frames of established ac-tivity or active support from their partner in doingso.

Although the simplifications in the tasks used inthe current study were meant to compensate forsome of these demands unique to interactions be-tween young children, the fact that 1-year-olds werestill unable to coordinate their behavior with oneanother suggests that their social understanding maybe too immature to support peer cooperation. Thismay be especially true when the cooperation task orsetting is relatively unfamiliar. Under such circum-stances lower level strategies might carry over from

one task or trial to another, especially if they havebeen successful, because childrens fragile socialunderstanding is not readily accessible or is easilydisrupted when the task is difficult or the settingnonroutine and unfamiliar. As social understandingdevelops over the 2nd and 3rd years of life, andtoddlers are increasingly able to detect, generate, andshare intentions with others outside of establishedsocial and communicative routines and familiar ac-tivity, young peers can invent and share novel goalsand intentions with one another and can monitorand join one anothers behavior related to thosegoals, permitting true cooperation. Perhaps for veryyoung children, social understanding emerges andconsolidates in play and communicative routinesduring the dynamic give-and-take of social interac-tion with adults and then later generalizes to the lessfamiliar, less routine, and less predictable world ofpeers.

Consistent with this conjecture, we found thattoddlers who demonstrated greater social under-standing in social interaction with an adult were alsomore skilled in coordinating their behavior with apeer partner. Children who were more skilled insharing an adults perspective and could better talkabout their own and others actions and internalstates and refer to themselves and others using per-sonal pronouns were also better at cooperating withtheir peers. This research design does not tell us, ofcourse, whether advances in social understandingwith adults precede advances in social skill withpeers, whether growth in adult child and childchild social understanding and social skill occur intandem, or whether participation in progressivelymore challenging social interactions with adultsand/or peers drives changes in social understanding(Carpendale & Lewis, 2004; Zerwas, Balaraman, &Brownell, 2004). Longitudinal studies will be neces-sary to distinguish possible patterns of influence.

Responding to joint attention bids by an adult andlanguage about self and other are distinct measuresof social understanding, and both were associatedwith measures of peer cooperation in the currentstudy. When examined together, taking into accounttheir shared variance, they uniquely predicted dif-ferent aspects of childrens cooperative efforts. Skillin joint attention was uniquely associated with morecoordinated behavior with peers, whereas self otherlanguage was uniquely related to monitoring thepeer while engaging in goal-directed activity oneself.

We can only speculate about why these particularassociations emerged. With respect to joint attention,following into and sharing anothers attention tothings in the world is a form of coordinated activity


much like cooperation. Indeed, true collaboration,whether in communication, game playing, or prob-lem solving, depends on shared attention to the sameexternal object or goal. Joint attention is thus a basicprerequisite for cooperative activity (Brinck & Gar-denfors, 2003; Tomasello et al., 2005). With respect tolanguage about self and other, our measure wasderived in part from childrens use of words refer-ring to others actions and behavior. Greater interestin and talk about others actions may reflect greaterawareness of the goal directedness of others be-havior. More generally, however, using wordsgrounded in self other representations and talkingabout others feelings, actions, and internal statesinvolves relatively sophisticated interpersonal un-derstanding (Bates, 1990; Lewis & Ramsay, 2004),including both differentiation of self and other andunderstanding of self other equivalence. Insofar aslanguage about self and other and cooperation bothrequire the integration of first-person and third-personal intentional perspectives (Moore, in press),monitoring a partners actions in relation to onesown intentions and to the same shared goal mayreflect this understanding. At the same time, partic-ipating in communicative routines and cooperativeplay may promote and fine-tune developing socialunderstanding. Questions about relations betweensocial understanding and social skills are the focus ofincreasing research interest in preschoolers; suchquestions may be especially fruitfully addressedduring the toddler years when both are undergoingrapid transition, and in the context of cooperativeactivity with peers where children are independentlyputting their emerging social understanding to use.

References

Agnetta, B., & Rochat, P. (2004). Imitative games by 9-, 14-,

and 18-month-old infants. Infancy, 6, 1 36.Asendorpf, J. B., & Baudonniere, P. M. (1993). Self-aware-

ness and other-awareness: Mirror self-recognition and

synchronic imitation among unfamiliar peers. Develop-mental Psychology, 29, 88 95.

Ashley, J., & Tomasello, M. (1998). Cooperative problem-

solving and teaching in preschoolers. Social Development,7, 143 163.

Astington, J., & Jenkins, J. (1999). A longitudinal study of

the relation between language and theory-of-mind de-

velopment. Developmental Psychology, 35, 1211 1320.Bakeman, R., & Adamson, L. (1984). Coordinating atten-

tion people and objects in mother infant and peer in-

fant interaction. Child Development, 55, 1278 1289.Bakeman, R., & Adamson, L. (1986). Infants convention-

alized acts: Gestures and words with mothers and peers.

Infant Behavior and Development, 9, 215 230.

Baldwin, D. A. (1993). Early referential understanding:Infants ability to recognize referential acts for what theyare. Developmental Psychology, 29, 832 843.

Baldwin, D. A. (1995). Understanding the link betweenjoint attention and language. In C. Moore & P. Dunham(Eds.), Joint attention: Its origins and role in development(pp. 131 158). New Jersey: Erlbaum.

Baldwin, D., & Moses, L. (1996). The ontogeny of socialinformation. Child Development, 67, 1919 1939.

Barresi, J., & Moore, C. (1996). Intentional relations andsocial understanding. Behavioral and Brain Sciences, 19,107 154.

Bartsch, K., & Wellman, H. M. (1995). Children talk about themind. London: Oxford University Press.

Bates, E. (1990). Language about me and you: Pronominalreference and the emerging concept of self. In D. Cic-chetti & M. Beeghly (Eds.), The self in transition: Infancy tochildhood (pp. 165 182). Chicago: University of ChicagoPress.

Behne, T., Carpenter, M., Call, J., & Tomasello, M. (2005).Unwilling versus unable: Infants understanding of in-tentional action. Developmental Psychology, 41, 328 337.

Bellagamba, F., & Tomasello, M. (1999). Re-enacting in-tended acts: Comparing 12- and 18-month-olds. InfantBehavior and Development, 22, 277 282.

Bloom, L., & Tinker, E. (2001). The intentionality modeland language acquisition. . Monographs of the Societyfor Research in Child Development, 66(4; Serial No.267).

Boesch, C., & Boesch, H. (1989). Hunting behavior of wildchimpanzees in the Tai National Park. American Journalof Physical Anthropology, 78, 547 573.

Bornstein, M., Tamis-LeMonda, C., Tal, J., Ludemann, P.,Toda, S., Rahn, C., et al. (1992). Maternal responsivenessto infants in 3 societies: The United States, France andJapan. Child Development, 63, 808 821.

Bretherton, I., & Beeghly, M. (1982). Talking about internalstates: The acquisition of an explicit theory of mind.Developmental Psychology, 18, 906 921.

Brinck, I., & Gardenfors, P. (2003). Co-operation and com-munication in apes and humans. Mind & Language, 18,484 501.

Brooks, R., & Meltzoff, A. N. (2002). The importance ofeyes: How infants interpret adult looking behavior. ChildDevelopment, 38, 958 966.

Brownell, C. (1986). Convergent developments: Cognitive-developmental correlates of growth in the infant/tod-dler peer skills. Child Development, 57, 275 286.

Brownell, C., & Brown, E. (1992). Peers and play in infantsand toddlers. In V. B. Van Hasselt (Eds.), Handbook ofsocial development: A lifespan perspective (pp. 183 200).New York: Plenum Press.

Brownell, C., & Carriger, M. (1990). Changes in cooperationand self other differentiation during the second year.Child Development, 61, 1164 1174.

Brownell, C., & Carriger, M. (1991). Collaborations amongtoddler peers: Individual contributions to social con-texts. In L. B. Resnick (Eds.), Perspectives on socially shared


cognition (pp. 365 383). Washington, DC: AmericanPsychological Association.

Brownell, C., & Hazen, N. (1999). Early peer interaction: Aresearch agenda. Early Education and Development, 10,403 413.

Bruner, J. (1982). The organization of action and the natureof adult-infant transactions. In E. Tronick (Ed.), Socialinterchange in infancy: Affect, cognition, and communication(pp. 23 36). Baltimore: University Park Press.

Callanan, M., & Sabbagh, M. (2004). Multiple labels forobjects in conversations with young children: Parentslanguage and childrens developing expectations aboutword meanings. Developmental Psychology, 40, 746 763.

Camaioni, L., Baumgartner, E., & Perucchini, P. (1991).Content and structure in toddlers social competencewith peers from 12 to 36 months of age. Early ChildDevelopment and Care, 67, 17 27.

Caron, A., Kiel, E., Dayton, M., & Butler, S. (2002). Com-prehension of the referential nature of looking andpointing between 12 and 15 months. Journal of Cognitionand Development, 3, 445 464.

Carpendale, J., & Lewis, C. (2004). Constructing an un-derstanding of mind: The development of childrenssocial understanding within social interaction. Behavioraland Brain Sciences, 27, 79 96.

Carpenter, M., Akhtar, N., & Tomasello, M. (1998). Four-teen- through 18-month-old infants differentially imitateintentional and accidental actions. Infant Behavior andDevelopment, 21, 315 330.

Carpenter, M., Call, J., & Tomasello, M. (2002). Under-standing prior intentions enables 2-year-olds to imi-tatively learn a complex task. Child Development, 73,1431 1441.

Carpenter, M., Call, J., & Tomasello, M. (2005). Twelve- and18-month-olds copy actions in terms of goals. Develop-mental Science, 8, 13 20.

Carpenter, M., Nagell, K., & Tomasello, M. (1998). Socialcognition, joint attention, and communicative compe-tence from 9 to 15 months of age. Monographs of theSociety for Research in Child Development, 63(4, SerialNo. 255).

Chalmeau, R., Visalberghi, E., & Gallo, A. (1997). Capuchinmonkeys, Cebus apella, fail to understand a cooperativetask. Animal Behavior, 54, 1215 1225.

Cheung, H., Hsuan-Chih, C., Creed, N., Ng, L., Wang, S., &Mo, L. (2004). Relative roles of general complementationlanguage in theory-of-mind development: Evidencefrom Cantonese and English. Child Development, 75,1155 1170.

deVilliers, J., & deVilliers, P. (2000). Linguistic determinismand the understanding of false beliefs. In P. Mitchell &K. Riggs (Eds.), Childrens reasoning and the mind (pp.191 228). Hove, UK: Psychology Press.

Dunn, J. (1988). The beginnings of social understanding.Cambridge, MA: Harvard University Press.

Dunn, J., Brown, J., Slomkowski, C., Tesla, C., & Young-blade, L. (1991). Young childrens understanding ofother peoples feelings and beliefs: Individual differ-

ences and their antecedents. Child Development, 62,1352 1366.

Dunphy-Lelii, S., & Wellman, H. (2004). Infants under-standing of occlusion of others line of sight: Implica-tions for an emerging theory of mind. European Journal ofDevelopmental Psychology, 1, 49 66.

Eckerman, C. O., Davis, C. C., & Didow, S. M. (1989).Toddlers emerging ways of achieving\social coordina-tions with a peer. Child Development, 60, 440 453.

Eckerman, C. O., & Didow, S. M. (1996). Nonverbal imitationand toddlers mastery of verbal means of achieving co-ordinated action. Developmental Psychology, 32, 141 152.

Eckerman, C. O., & Peterman, K. (2001). Peers and infantsocial/communicative development. In A. Fogel & G.Bremner (Eds.), Blackwell handbook of infant development(pp. 326 350). Malden, MA: Blackwell.

Eckerman, C., & Whitehead, H. (1999). How toddler peersgenerate coordinated action: A cross-cultural explora-tion. Early Education & Development, 10, 241 266.

Farrar, J., & Maag, L. (2002). Early language developmentand the emergence of a theory of mind. First Language,22, 197 213.

Fenson, L., Dale, P. S., Reznick, J. S., Bates, E., Thal, . D.,Pethick, S. J., et al. (1994). Variability in early commu-nicative development. Monographs of the Society forResearch in Child Development, 59(5, Serial no. 242).

Flom, R., & Pick, A. (2003). Verbal envouragement andjoint attention in 18-month-old infants. Infant Behaviorand Development, 26, 121 314.

Garvey, C. (1990). Play. Cambridge, MA: Harvard Univer-sity Press.

Gergely, G., Bekkering, H., & Kiraly, I. (2002). Rationalimitation in preverbal infants. Nature, 415, 755.

Gergely, G., Nadasdy, Z., Csibra, G., & Biro, S. (1995).Taking the intentional stance at 12 months of age. Cog-nition, 56, 165 193.

Harris, P., de Rosnay, M., & Pons, F. (2005). Language andchildrens understanding of mental states. Current Di-rections in Psychological Science, 14, 69 73.

Hartup, W. (1983). Peer relations. In P. Mussen (Series Ed.)E. Hetherington (Vol. Ed.), Handbook of child psychology,Vol. 4: Socialization, personality, and social development (pp.103 196). New York: Wiley.

Hartup, W. (1996). Cooperation, close relationships & cog-nitive development. In W. Bukowski, A. Newcomb, &W.Hartup (Eds.), The company they keep: Friendships in child-hood & adolescence. Cambridge, UK: Cambridge Univer-sity Press.

Hay, D. F. (1979). Cooperative interactions and sharingbetween very young children and their parents. Devel-opmental Psychology, 15, 647 653.

Hollich, G., Hirsh-Pasek, K., & Golinkoff, R. (2000). Breakingthe language barrier: An emergentist coalition model ofthe origins of word learning. Monographs of the Societyfor Research in Child Development, 65(3, Serial No. 262).

Howes, C. (1988). Peer interaction of young children.Monographs of the Society for Research in Child De-velopment, 53(1, Serial No. 217).


Jenkins, J., Turrell, S., Kogushi, Y., Lollis, S., & Ross, H.(2003). A longitudinal investigation of the dynamicsof mental state talk in families. Child Development, 74,905 920.

Kagan, J. (1981). The second year: The emergence of self-awareness. Cambridge, MA: Harvard University Press.

Kochanska, G., & Aksan, N. (2004). Development of mu-tual responsiveness between parents and their youngchildren. Child Development, 75, 1657 1676.

Kuhlmeier, V., Wynn, K., & Bloom, P. (2003). Attribution ofdispositional states by 12-month-olds. Psychological Sci-ence, 14, 402 408.

Lewis, M., & Ramsay, D. (2004). Development of self-rec-ognition, personal pronoun use, and pretend play dur-ing the 2nd year. Child Development, 75, 1821 1831.

Lillard, A., & Witherington, D. (2004). Mothers behaviormodifications during pretense and their possible signalvalue for toddlers. Developmental Psychology, 40, 95 113.

Liszkowski, U., Carpenter, M., Henning, A., Striano, T., &Tomasello, M. (2004). Twelve-month-olds point to shareattention and interest. Developmental Science, 7, 297 307.

Lohmann, H., & Tomasello, M. (2003). The role of languagein the development of false-belief understanding: Atraining study. Child Development, 74, 1130 1144.

Maudry, M., & Nekula, M. (1939). Social relations betweenchildren of the same age during the first two years oflife. Journal of Genetic Psychology, 54, 193 215.

Meltzoff, A. (1995). Understanding the intentions of others:Re-enactment of intended acts by 18-month-old chil-dren. Development Psychology, 31, 1 16.

Meltzoff, A., Gopnik, A., & Repacholi, B. (1999). Toddlersunderstanding of intentions, desires and emotions: Ex-plorations of the dark ages. In P. Zelazo, J. Astington, &D. Olson (Eds.), Developing theories of intention: socialunderstanding and self-control (pp. 17 41). Mahwah, NJ:Erlbaum.

Moll, H., & Tomasello, M. (2004). Twelve- and 18-month-old infants follow gaze to spaces behind barriers. De-velopmental Science, 7, 1 9.

Moore, C. (in press) Understanding self and others in thesecond year. In C. Brownell & C. Kopp (Eds.), Transitionsin early socioemotional development: The toddler years. NewYork: Guilford Press.

Mueller, E., & Brenner, J. (1977). The origins of social skillsand interaction among playgroup toddlers. Child De-velopment, 48, 854 861.

Mundy, P. (2006) Motivation, self-regulation and the ne-urodevelopment of intention sharing. Behavioral andBrain Sciences, 28, 709.

Mundy, P., Hogan, A., & Doehring, P. (1996). A preliminarymanual for the Early Social Communication Scales (ESCS).University of Miami Psychology Department, CoralGables, FL, (http://www.psy.miami.edu/faculty/pmundy/ESCS.pdf).

Nelson, K. (in press). Becoming a language user: Entering asymbolic world. In C. Brownell & C. Kopp (Eds.),Transitions in early socioemotional development: The toddleryears. New York: Guilford Press.

ONeill, M., Bard, K., Linnell, M., & Fluck, M. (2005). Ma-ternal gestures with 20-month-old infants in two con-texts. Developmental Science, 8, 352 359.

Pan, B., Imbens-Bailey, A., Winner, K., & Snow, C. (1996).Communicative intents expressed by parents in inter-action with their young children. Merrill-Palmer Quar-terly, 42, 248 266.

Piaget, J. (1932). The moral judgment of the child. London:Routledge & Keegan Paul.

Poulin-Dubois, D., & Forbes, J. (2002). Toddlers attentionto intentions-in-action in learning novel action words.Developmental Psychology, 38, 104 114.

Povinelli, D., & ONeill, D. (2000). Do chimpanzees usetheir gestures to instruct each other? In S. Baron-Cohen,H. Tager-Flusberg, & D. Cohen (Eds.), Understandingother minds: perspectives from developmental cognitive ne-uroscience (2nd ed., pp. 459 487). Oxford, UK: OxfordUniversity Press.

Repacholi, B. M. (1998). Infants use of attentional cues toidentify the referent of another persons emotional ex-pression. Developmental Psychology, 34, 1017 1025.

Repacholi, B. M., & Gopnik, A. (1997). Early reasoningabout desires: Evidence from 14- and 18-month-olds.Developmental Psychology, 33, 2 21.

Ricard, M., Girouard, P. C., & Gouin Decarie, T. (1999).Personal pronouns and perspective taking in toddlers.Journal of Child Language, 26, 681 697.

Rocissano, L., Slade, A., & Lynch, V. (1987). Dyadic sync-hrony and toddler compliance. Developmental Psychology,23, 698 704.

Rogoff, B. (1990). Apprenticeship in thinking: Cognitive de-velopment in social context. New York: Oxford UniversityPress.

Ross, H. S. (1982). Toddler peer relations: Differentiation ofgames and conflicts. Canadian Journal of Behavioural Sci-ence, 14, 364 379.

Ross, H. S., & Kay, D. (1980). The origins of social games. InK. Rubin (Ed.), Childrens play (pp. 17 31). San

Date post:	04-Nov-2015
Category:	Documents
Upload:	eduardo-aguirre-davila
View:	6 times
Download:	0 times

Becoming a Social Partner With Peers. Cooperation and Social Understanding in One- And Two-Year-Olds...

Documents