1131316

Wiley and Society for Research in Child Development are collaborating with JSTOR to digitize, preserve and extend access toChild Development.

http://www.jstor.org

Theory of Mind Is Contagious: You Catch It from Your Sibs Author(s): Josef Perner, Ted Ruffman and Susan R. Leekam Source: Child Development, Vol. 65, No. 4 (Aug., 1994), pp. 1228-1238Published by: on behalf of the Wiley Society for Research in Child DevelopmentStable URL: http://www.jstor.org/stable/1131316Accessed: 10-03-2015 02:43 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

This content downloaded from 181.118.153.129 on Tue, 10 Mar 2015 02:43:47 UTCAll use subject to JSTOR Terms and Conditions

Theory of Mind Is Contagious: You Catch It from Your Sibs

Josef Perner and Ted Ruffman University of Sussex at Brighton

Susan R. Leekam University of Kent at Canterbury

PERNER, JOSEF; RUFFMAN, TED; and LEEKAM, SUSAN R. Theory of Mind Is Contagious: You Catch It from Your Sibs. CHILD DEVELOPMENT, 1994, 65, 1228-1238. This study investigated the relation between family size and "theory of mind." Results from an experiment with 3- and 4-year-olds showed that children from larger families were better able than children from smaller families to predict a story character's mistaken (false-belief based) action. Results from a second experiment on children with exactly 1 sibling failed to show any superiority of older over younger siblings in promoting earlier understanding of belief. The data are interpreted as suggesting that sibling interaction provides a rich "data base" for building a theory of mind, and this is dis- cussed in relation to other studies showing that particular kinds of interaction between sibling and child, or caregiver and child, have a beneficial effect on the understanding of false belief.

In recent years much research has been devoted to tracing the developmental course of children's understanding of the mind. This research has tended to focus on the nor- mative question of what children understand about mental concepts at different ages, and in so doing sheds some light on the structure of the conceptual system that the child acquires. One very influential view about the type of conceptual structure that is acquired is that children acquire a theory of mind. One should point out that the term "theory of mind" is used by some to emphasize that the child has a coherent body of knowledge that goes beyond the immediately perceptible and which provides pre- dictive power (Premack & Woodruff, 1978; Sellars, 1956) without any implication of how such a theory might be acquired. And some go even so far as to suggest that it is not really acquired at all but that it is innate (Fodor, 1987) or that its essential substrate emerges through maturation (Leslie, 1987).

Most developmental theorists who sub- scribe to the theory view, however, find it attractive to view intellectual development as a process of theory acquisition exactly because it promises an explanation of how gen- uinely new conceptual systems can be ac-

quired. They look at children as acquiring new ways of thinking in an essentially similar (albeit more guided) way to that in which scientists develop their theories (Carey, 1985, 1988; Gopnik, 1988; Karmiloff-Smith, 1988; Keil, 1989; Kitcher, 1988; Perner, 1991; Wellman, 1990). This research then gives important insights on the type of conceptual structure the child is acquiring, but as yet there is little research on the origin of and driving force behind this acquisition process and the individual differences to which this process gives rise.

In science, one thing is fairly clear; new theories and insights do not simply mature but need hard work and devotion. Only scientists who immerse themselves in creating and contemplating relevant data are likely to make significant contributions to their field. Obsession, in the sense of a perhaps inextri- cable web of high motivation and great expertise, pays, or as Einstein once remarked, "Genius is 10% inspiration and 90% perspi- ration." So, if children's understanding of the mind develops along similar principles, one would expect that children, who live in an environment with greater incentive and opportunity to benefit from an advanced theory of mind, will develop such a theory more

Requests for reprints should be directed to the first author at the following address: Experi- mental Psychology, University of Sussex, Brighton, BN1 9QG, England.

[Child Development, 1994, 65, 1228-1238. ? 1994 by the Society for Research in Child Development, Inc. All rights reserved. 0009-3920/94/6504-0003$01.00]


Perner, Ruffman, and Leekam 1229

quickly than children with less of such an opportunity. What might such an environment be?

Dunn's (e.g., 1984, chap. 3, 1988) work has shown that it is mostly in their interaction with siblings that children around the age of 2 and 3 years have their most intensive social experience. It is in those situa- tions that they are most creative in their play (Dunn & Dale, 1984). Sibling conflicts also tend to elicit the most maternal talk about feeling states (Dunn, Bretherton, & Munn, 1987) and moral issues which elicit much explicit reflection on motives, intentions, and knowledge. Consequently, we might expect that children who have siblings to work and play with should have a noticeable advantage over only children who are stuck with their parents as their only source of social entertainment.

We tested this prediction by investigat- ing whether the number of siblings does have a beneficial effect on the age at which children show an understanding of false belief (Wimmer & Perner, 1983; Perner, Leekam, & Wimmer, 1987), since such understanding marks an important step in children's acquisition of a theory of mind (Astington, Harris, & Olson, 1988; Perner, 1991; Wellman, 1990). Understanding of false belief takes such a central position in the development of "children's theory of mind" because it requires understanding of several features that mark the mind as a representational system (Flavell, 1988; Forgu- son & Gopnik, 1988; Perner, 1991; Wellman, 1990). For instance it requires understanding of the conditions under which a belief is formed, for example, if a person does not see (or get other information) that an object is moved to a new location then that person's mental representation of the object's location remains unchanged. As a consequence the person has a misrepresentation of where the object is. To understand this children have to be able to represent not only the. world as it is not (object in the old location) but also to appreciate that the mistaken person takes this (mis)representation as a representation of where the object really is and, therefore, will look in the wrong place even though he wants to go where it really is.

One can intuitively see that it would be difficult to explain how a false belief is formed and why it leads its holder to look in the wrong place without going into a "representational theory" of how the mind works. In the case of false belief, information (or

misinformation) about the world creates something (a representation) in the mind which governs the person's actions in the real world. And it does so regardless of whether or not it is an accurate reflection of the world, because it functions as a representation of the world. Since it is hard to see how false belief can be understood without developing such a, at least tacit, representational theory of mind, the false belief test seems to be a good indicator of a substantial intellectual step in children's understanding of the mind.

In contrast, a superficially similar question that has to do, for example, with knowing where an object is when it has been moved, can be answered on the basis of a simple rule, namely, that informational access (e.g., seeing) leads to knowing and that without such informational access a person does not know where something is. Because this relatively simple rule is sufficient, correct answers to this knowledge question do not reflect as deep an understanding of the mind as those to the belief question about where the mistaken person will look for the object (Perner, 1991, chaps. 7 and 8). Indeed it has been found that 3-4-year-old children's ability to answer the knowledge question fluctuates much more with changes in experimental procedure (Hogrefe, Wimmer, & Perner, 1986; Pillow, 1989; Pratt & Bry- ant, 1990) than is typical for belief questions.

In our first experiment we use both types of question, but for the reasons just outlined we will rely in the data analysis more on children's responses to the belief question than answers to the knowledge question. We expect children with siblings to do better on the belief question than only children because interaction with siblings provides a more intensive data base and motivation for developing a representational theory of mind.

Experiment 1 Method

Subjects.--Eighty children from one nursery school in Brighton, England, aged 3 years, 1 month to 4 years, 9 months (34 girls and 46 boys) participated in this experiment. They came from a predominantly white, working- and middle-class social back- ground. All children spoke English as their first language. We were unable to obtain data on family size for two children, whose test data were therefore omitted from further analysis. In two cases, two children from the


1230 Child Development same family had been tested. To avoid statistical problems of nonindependence one of the children in each case was dropped from the study at random. Of the remaining 76 children, 22 were only children, 42 came from a family of two children (18 had an older sibling, 23 a younger sibling, and one had a twin), 11 came from a family of three (four were second born, six were third born, and one had an older sibling and a twin), and one was the last of a family of four. The number of siblings was not known to the experimenter when testing each child.

Procedure and material. -Each child was told a false-belief story enacted with dolls in two model rooms. For half the children it was an Unexpected Change story as introduced by Wimmer and Perner (1983) about Max who helps his mother unpack the shopping bag. In the story, Max puts the chocolate into a particular cupboard and goes to play. In his absence his mother needs some of the chocolate and then puts it into a different cupboard. In the meantime Max gets hungry on the playground and wants some chocolate. Children are asked two test questions, one about Max's knowledge about the chocolate's location (Knowl- edge Question), "Does Max know where the chocolate is or does he not know where it is?" and then one about Max's belief-based action (Belief Question), "Where will Max look for the chocolate?"

Each subject's answer to the belief question was immediately followed by a se- ries of Memory Questions assessing his or her understanding of the basic story facts:

1. Is it in there? (If no, Where is it?) 2. Remember at the beginning, where did

Max put the chocolate? 3. Remember when mother moved the choco-

late? Where did she put it? 4. Did Max see her put it there?

For the other half of the children, the story involved Misinformation. On the playground Max is joined by his brother Sam, who asks him where the chocolate is. Max tells him what he mistakenly thinks, namely, that it is where he put it. Children are then asked the same test questions, but about Sam rather than Max, and the Memory Ques- tion 2 is changed to: "Remember when Max talked to Sam on the playground? Where did Max say that the chocolate was?"

Further detail about materials and the verbatim version of the stories can be found in Perner and Wimmer (1988).

Results and Discussion The dependent variable of main interest

is the proportion of children giving correct answers to the belief question. The independent variables were age, sex, family size, memory questions (mean number of correct answers to the four questions), and origin of belief (unexpected change vs. misinformation stories). Table 1 shows the means, standard deviations, and intercorrelations (df = 74) for all independent variables (except the experimentally manipulated variable). The effect of the independent variables was analyzed by logistic regression (BMDP LR, Dixon et al., 1981) specifying the model "age *sex* family-size *memory-questions *belief-origin," where age, family size, and memory-question were specified as interval variables. Terms were introduced according to the following considerations. Using the hierarchical rule, interaction effects were considered only after all component effects were already in the model. The constant term of the regression equation was in the model at the beginning. Since the theoretically interesting factors could be partly confounded with age, which has a known effect on belief understanding, we first introduced age into the model and then checked which of the other effects if added to the model would reduce the error variance by a significant amount. The effect of age was significant, X2(1, N = 76) = 11.14, p < .001. After age had been introduced into the model the only other significant effect was family size, X2(N = 76) = 3.99, p < .05.

The two significant main effects reflect an increase in the proportion of children who gave correct answers to the belief question with age and with family size. Figure 1 shows these increases (in rough approxima- tion for the age effect by grouping children into two age groups). Also included are plots of false belief performance predicted by the regression equations for age and size.

The magnitude of the improvement with family size is quite impressive, consid- ering that its magnitude almost matches the improvement with age from young 3- to older 4-year-olds. In other words, the benefit children get from interacting with two siblings rather than none is worth about as much as 1 year of experience, which at this young age is considerable, namely, one- fourth of the young child's entire life.

To investigate further the role of siblings, we asked whether it makes any difference whether the siblings are older or



TABLE 1

MEANS, STANDARD DEVIATIONS, AND INTERCORRELATIONS BETWEEN PREDICTOR VARIABLES IN EXPERIMENT 1

CORRELATION WITH VARIABLES

VARIABLE RANGE MEAN SD Age Sex Family Size

Age .................................... 3.1-4.7 3.9 .47 ... - .07 .06 Sex ..................................... 1-2 1.42 .50 - .07 ... -.06 Family size ....................... 1-4 1.88 .69 .06 -.06 Control question .............. 0-4 3.60 .73 .54 -.06 .08

NOTE.-N = 76.

younger than the subject because Light (1979, p. 76) reported that children with an older sibling do better on role-taking measures than children with a younger sibling. Indeed, there is an indication that older siblings are more of a help. Firstborns of two gave fewer correct responses (56.5%, n = 23) than second borns (66.7%, n = 18), and second borns of three were less well off (50.0%, n = 4) than third borns (85.7%, n = 6, including one fourth born of four). There were no firstborn of three at the age investigated. To test this potential effect of birth order, we contrasted performance by children with only older siblings (i.e., second borns of two, third borns of three, and the one fourth born of four; 70.8% correct, n =

25) with that by children with only younger siblings (i.e., firstborns of two, 56.5% correct, n = 23). This difference was, however, not statistically significant, )2(1, N = 47) = 1.04, p > .30.

Taking children's performance on the false belief question as an indication of their understanding of false belief is premised on the assumption that they understand the story details. Their degree of understanding is assessed by answers to the battery of four memory questions. Children's performance on these questions was remarkably good with a mean of 3.60 correct answers, and only seven 3-year-olds and one 4-year-old gave more than one wrong answer. Most of

100

80

o 60

cM Data From 0- Regression

o 40

a 0- Raw Data 20

3 4 0 1 2-3 n=38 n=38 n=22 n=42 n =12

Age Group Number of Siblings

FIG. 1.-Percent children giving correct answer to the belief question in Experiment 1, according to age and family size. Three-year-olds: mean = 3-6; 4-year-olds: mean = 4-4; regression equation for age: Y = 1.78x - 6.66; regression equation for size: Y = .74x - 1.11.


1232 Child Development the errors (84%) were committed in about equal portions on questions 2 and 4.

To the degree that there were problems with the memory questions this raises the possibility that the observed correlation between family size and answers to the false belief question is mediated by failure to understand story facts. To protect our interpretation of results against this possibility we repeated the original logistic regression for family size on answers to the belief question, but with both age, X2(1, N = 76) = 9.55, p < .01, and control questions, X2(1, N = 76) = .27, p > .60, as independent variables forced into the model. Family size still ac- counted for a significant amount of variance, X2(1, n = 76) = 4.14, p < .05, in false belief understanding.

The same logistic regression used for the belief question was carried out on responses to the knowledge question. Only age was a significant factor, X2(1, N = 76) = 12.49, p < .001, which can be approximately illustrated by the mean correct responses in the three age groups. Correct responses rose from 62.5% among young 3-year-olds, to 72.4% among old 3-year-olds, to 100% among 4-year-olds.

In sum, the important result of this experiment is the effect of family size on children's understanding of false belief. In the next experiment we investigate further the reasons for this effect in terms of sibling's age.

Experiment 2 Several possibilities come to mind con-

cerning why siblings may have a beneficial effect on children's development of a theory of mind, which all relate to the age difference between child and sibling. One possibility is that siblings teach, or pass on in some other way, relevant knowledge. In this case we would expect that older siblings should have a much more pronounced effect than younger siblings. As mentioned before, Light's (1979) findings support this expecta- tion, and there was some indication of a possible difference in the results of our first experiment, although not a statistically significant one.

A second possibility is that siblings provide an opportunity for a child to sharpen his or her understanding by teaching the siblings. In fact, this possibility has been used by Zajonc and Markus (1975) to explain why the IQ of last borns has been found to be

consistently lower than that of their elder siblings: because they did not have anybody younger to teach. The prediction from this hypothesis is that younger siblings should be more beneficial than older siblings. In Experiment 1, however, there was little support for this.

A third possibility is that siblings are beneficial because they provide more opportunity for interaction in relevant activities. From this it follows that older as well as younger siblings provide comparable bene- fits as long as they are not too distant in age. For neither a newborn baby nor a much older sibling (who would be more like an- other parent) provide much opportunity to engage in intensive joint play activities.

To test these different possibilities we looked only at children with one sibling. Children were told stories in which a doll was led to a false belief about the contents of a box. The doll saw one object inside the box. Then, in the doll's absence, the object was removed from the box and replaced with a different object. The doll then returned, and either children were asked what she would think was inside the box (before the doll had a chance to look inside), or the doll looked inside and children were asked what she had thought was inside prior to looking.

We examined how different age relations between sibling and child affect their ability to answer this question. We analyzed for effects of sibling's age, of distance in age (absolute difference between sibling's and child's age), and of whether the sibling is older or younger than the child. Method

Subjects.-All 43 children from three nurseries and the reception class of a pri- mary school who had exactly one sibling were included in this study. The nurseries and school were in a largely white, working- and lower middle-class neighborhood of Brighton. All children spoke English as their first language. The ages of their siblings ranged from 7 months to 19 years with a mean age of 41/2 years (SD = 31/2 years). The 19-year-old sibling was an extreme outlier, 4 SDs away from the mean. This sibling was also very distant in age from the next oldest sibling of 11 years (1.8 SDs away from the mean). Since such outliers can exert undue influence on the statistical analysis we dropped the child with the oldest sibling from our sample. The ages of the remaining 23 girls and 19 boys ranged from 3-2 to 5-9 (mean 4-10); 15 of them had older and 27



younger siblings. At the time of testing, the experimenter was not aware of the age of subjects' siblings.

Materials.-A doll named "Katy" (33 cm tall) was used to represent the story character. There were six pairs of materials (one pair per story): (a) spoon, bandaid; (b) pen- cil, sock; (c) keys, paper; (d) rock, matches; (e) cup, envelope; and (f) twig, Q-tip. One of the objects from each pair was placed in a box (11 x 11 x 11 cm).

Procedure.-To begin with, the experimenter introduced children to the doll named Katy and asked them to pretend she was a real person "just like you and me." The experimenter explained that Katy vis- ited her grandmother every week and that each time she liked to bring her something. Each week Katy could take either of two items (see the "Materials" section above). The experimenter instructed the child to place one of the items inside the box. Katy then looked inside the box, and the experimenter assured children that she saw what was inside. Katy then left to go to the "playground" (was put in a bag out of the child's view), and the experimenter explained that "Katy can't see us when she is in the playground." The child then removed the first object from the box and replaced it with the other object. To ensure that children understood Katy had not seen this transfer the experimenter then asked, "Did Katy see us put the (second object) inside the box?" Katy then returned. Each child was told six stories of this kind. In three stories the experimenter asked children the Belief question before Katy had a chance to look inside: "What does Katy think is in the box?" In the other three stories the sequence of events was the same except that Katy discovered the new object in the box before the experimenter asked the Belief question: "What did Katy think was in the box when she was in the playground?" In all six stories children were asked a memory question about the ac- tual contents of the box after they were asked the Belief question: "What's in the box?"

Stories were presented in four random orders. The pairs of materials were assigned to each story on a random basis. Results

Children's answers to the memory questions were perfect. The pattern of performance on the test question over the six stories was strongly bimodal: 11 children gave all wrong answers, 3 children gave one cor-

rect answer, no children gave 2 or 3 correct answers, 1 child gave 4 and 2 children gave 5, and 25 gave all correct answers. It seemed, therefore, appropriate to classify children into two groups-those who understood false belief (4 or more correct answers) and those who failed to understand (1 or 0 correct)-and use this classification as the dependent variable rather than the number of correct responses. The independent variables were: age, sex, sibling's age, distance in age (absolute value of sibling's age minus subject's age), and the interaction between the latter two factors. Table 2 shows the means, standard deviations, and intercorrelations for these predictor variables. Their effect on the dependent variable was analyzed by logistic regression (BMDP LR: Dixon et al., 1981).

Factors were introduced according to the same rationale as in Experiment 1. The first factor introduced was age which was significant, X2(1, N = 42) = 15.22, p < .001. The size of the effect is roughly reflected by the improvement from 43% of children passing false belief in the youngest third (3-2 to 4-6), to 67% in the middle third (4-7 to 5-3), to finally 92% among the oldest third (5-4 to 5-9). No other effect was significant after age had been introduced, nor, in sup- plementary analyses, were there any significant effects before age had been partialed out. In particular, none of the theoretically interesting child-sibling relations was significant: sibling's age, X2(1, N = 42) = 0.26, p > .60, difference in age, X2(1, N = 42) = 0.05, p > .80, or the interaction between sibling's age and difference in age, x2(1, N = 42) = 0.05, p > .80. We also checked whether there was a difference between children with an older and those with a younger sibling, but there was no difference apparent: 10 of the 15 (two-thirds) children with older siblings and 18 of the 27 (two-thirds) children with younger siblings passed false belief.

These results confirm the negative finding in the previous experiment that the age of a child's sibling does not seem to matter. The present results, furthermore, suggest that closeness in age between sibling and child does not matter much either (at least not within a range of about 7 years).

General Discussion The main finding of these experiments

is that the number of siblings in a family is related positively to young children's rea-


1234 Child Development TABLE 2

MEANS, STANDARD DEVIATIONS, AND INTERCORRELATIONS BETWEEN PREDICTOR VARIABLES IN EXPERIMENT 2

CORRELATION WITH VARIABLES

Sibling's VARIABLE RANGE MEAN SD Age Sex Age

Age .................................. 3.2-5.7 4.8 .72 ... .05 .14 Sex ..................................... 1-2 1.55 .50 .05 ... .14 Sibling's age .................... .58-11 4.11 2.89 .14 .14 Distance in age ................. .58-6.8 2.72 1.10 .08 .18 .28

NOTE.-N = 42. Ages are being expressed in years, correct to the first decimal place.

soning in theory of mind tasks. These results are particularly interesting in view of recent findings by Dunn, Brown, Slomkowski, Tesla, and Youngblade (1991) who reported that measures of the interaction between second-born children, their mothers, and older siblings at 33 months correlated with these children's ability to use false belief in explaining behavior at 40 months.

The study by Dunn et al. and our study complement each other in an important way because the pattern of correlations in the study by Dunn et al. suggests which kind of interaction between child, sibling, and mother may be responsible for our observed effect. In particular the set of correlations reported in the lower half of their table 3 (Dunn et al., 1991, p. 1361) showed a relevant, distinct pattern of specific correlations between later belief understanding and types of earlier interactions. From the large set of possible correlations only two were clearly and significantly different from zero: (1) the amount of control mother exerted on the sibling and (2) the child's attempts to interact cooperatively with the sibling.

It is tempting to interpret these correlations causally, that is, one is inclined to as- sume that the earlier patterns of interaction help the child to acquire an understanding of belief more quickly. However, as Dunn et al. warn us, one needs to be cautious about causal interpretation of correlational data. In fact, individual differences in intellectual and/or social maturity could account for the correlations. For instance, generally more precocious children may interact more cooperatively with their siblings, on the one hand, but, on the other hand, when something goes wrong they may be more able to make it appear that their sibling was at fault, which might lead mother to exert control on

the sibling. The important point is that the general precocity of these children could also be directly responsible for their earlier onset of belief reasoning, rather than their more frequent, earlier social experiences.

The present results complement the study by Dunn et al. because they help firm up the causal interpretation of relations between early interaction and later belief reasoning. The causal interpretation is strength- ened because the pattern of results we obtained for birth order and false belief understanding are quite different from how birth order affects IQ (Belmont & Marolla, 1973; Zajonc & Markus, 1975). One of the clearest effects pertaining to IQ is that last borns have lower IQs than earlier borns. In contrast, with false belief understanding there was no such difference. In Experiment 2 second borns of two were as good (two- thirds passed false belief) as firstborns (two- thirds passed), and in Experiment 1 there was actually a trend in the opposite direc- tion. Second borns of two were slightly better (66.7%) than firstborns (56.5%), and from families of three the third borns were better (83.3%) than the second borns (50.0%). This makes it difficult to argue that the sibling effects we have documented in this study are confounded with cognitive maturity. Also, in response to an earlier version of this paper, Jenkins and Astington (personal com- munication) looked at their existing data from 68 3-5-year-olds for an effect of family size on belief understanding. Their prelimi- nary analysis showed that family size was correlated with false belief understanding even after a general measure of language competence had been partialed out.

To the degree that our data do support the intuitively plausible causal interpretation of the data by Dunn et al., they pose a



serious problem for nativist proposals and various developmental explanations relying on internal maturation. For instance, Fodor (1987) made the strong nativist claim that a theory of mind is innate and that the observed developmental improvements on various tasks are due to increased information- processing abilities (Fodor, 1992). This proposal is difficult to maintain in view of our and Astington and Jenkin's data since family size seems to affect belief reasoning specifically and not information-processing abilities in general. One should, however, mention that the data are compatible with Fodor's (1981) earlier, less radical claims about the innateness of concepts where he allowed experience to play a necessary role in "triggering concepts." Within this frame- work the number of siblings could be seen as providing the necessary experiences for triggering concepts like false belief.

The data also speak against so called neo-Piagetian theories that explain developmental progression as a maturational increase in central processing capacity (Pas- cual-Leone, 1970). Our data are, however, not so clearly incompatible with Case's (1985) revision of this theory, which he used to explain developmental advances on theory of mind tasks (Case, 1989). Although Case shares with Pascual-Leone the assumption of limited processing capacity, he admits of two ways in which this limitation can progressively be overcome. One way is through an increase in processing speed due to neurological maturation; the other way, however, is through increased efficiency due to practice. So Case could argue that the effect of family size is due to greater social practice in sibling interaction which increases the efficiency of social reasoning strategies.

For similar reasons our data fail to rule out (though they may contradict in spirit) Leslie's (1987; Leslie & Roth, 1993) proposal that the necessary computational mechanism for a theory of mind matures around the age of 11/2 years, since he does not commit himself as to what factors may effect the subsequent improvements in the workings of this basic mechanism.

The data also pose a certain challenge for the sociolinguistic criticisms originally leveled against Piaget's conservation experiments by Rose and Blank (1974) and Don- aldson (1978; McCarrigle & Donaldson, 1974-1975), later extended to egocentrism studies (Hughes & Donaldson, 1979) and re-

cently applied to theory of mind experiments (Lewis & Osborne, 1990; Siegal & Beattie, 1991). The criticism was that Pia- getian tests do not accurately assess what they purport to assess, namely, the development of certain concepts. Rather, they assess young children's growing sensitivity to communicative intentions in their interaction with an adult experimenter asking strange questions.

To assimilate our findings the sociolinguistic criticism theorists need to explain why increased interaction with siblings helps children to better understand adults' communicative intentions. If this challenge can be met the interesting prediction follows that family size should affect performance on various Piagetian tasks (e.g., conservation tasks), to which the sociolinguistic criticism has been applied, in a similar way to that in which it affects performance on the false belief task.

On the more positive side, the finding that siblings help develop a theory of mind is compatible with the sociocognitive tradi- tion of Vygotsky (1965), Doise (1985), and Perret-Clermont (1980), who emphasize intellectual progress as a function of social interaction among peers and view intellectual growth as a process of internalizing the knowledge already incorporated in the social interaction. However, we did not find a consistent superiority of children with older over younger siblings. Although it is possible that this result was due to a lack of statistical power, our result may pose a problem for the usual assumption in sibling interaction studies of this kind that interaction with an intellectually slightly advanced partner is optimal. The data gathered by Jenkins and Astington underline this potential problem. They obtained a slightly negative partial correlation between birth order and false belief understanding after the effects of family size (with which birth order tends to be positively correlated) had been partialed out. However, this apparent discrepancy between peer interaction studies and our family-size effects (and lack of birth-order effects) can be resolved by pointing out that the effect of family size may be mediated by parents' intervention in sibling interaction (Dunn et al., 1991) and it may matter little whether the sibling is older or younger, while in peer interaction studies children gain from their interaction directly but only from a more experienced partner.

Our results may also throw light on why


1236 Child Development traditional attempts to train theory of mind concepts like the appearance-reality distinction (Flavell, Green, & Flavell, 1986, study 3) have met with failure. One interesting reason why family size may have a more per- vasive and reliable effect than the teaching methods employed by Flavell et al. and the peer interaction engendered in laboratory studies, may be that family size provides a natural learning environment with a density of learning experience unattainable by a laboratory study. For this reason we suggest that gathering data of this kind, in particular, using the fine-grained approach adopted by Dunn et al. (1991) might be a good way to provide details about the important factors involved in the acquisition of a "theory of mind."

Here are but some speculations about which early activities between siblings may especially stimulate false belief understanding. One candidate is deception because it involves the manipulation of belief. How- ever, our finding that there was no advantage of having an older over having a younger sibling poses a problem. The problem arises since only children who already understand false belief tend to deceive (LaFreniere, 1988; Peskin, 1992; Russell, Mauthner, Sharpe, & Tidswell, 1991; Sodian, 1991; Stouthamer-Loeber, 1986; Wimmer & Per- ner, 1983). Claims that deception occurs before understanding of false belief (Chandler, Fritz, & Hala, 1989; Hala, Chandler, & Fritz, 1991) have met with methodological criticism and replication difficulties (Doherty, 1992; Ruffman, Olson, Ash, & Keenan, 1993; Sodian, Taylor, Harris, & Perner, 1991). Hence a child who does not understand false belief and does not engage in deceitful actions is unlikely to be confronted with de- ceptive actions by an even younger, less able sibling. Only older siblings who are already able to deceive will provide opportu- nities to learn from deception.

Moreover, the finding by Dunn et al. (1991) that false belief understanding corre- lates specifically with the child's attempts to interact cooperatively with the sibling, but not with competitive interactions, also makes it unlikely that sibling deception is a crucial factor. Fortunately, there are also cooperative activities that very young children engage in and that provide a plausible preparation for belief understanding.

One such cooperative type of activity is pretend play. Theorists of different persua- sions have seen pretence as an important

prerequisite for understanding false belief (Harris, 1991, in support of the simulation or role-taking view; Leslie, 1987, and Perner, 1991, in support of the theory formation view). Belief and pretence are conceptually very closely related because both lead to act- ing-as-if (e.g., both Max pretending and Max thinking the chocolate is in the [actually empty] cupboard lead him to look in that cupboard, i.e., act as if the chocolate were in there; Perner, Baker, & Hutton, in press). Furthermore, as Dunn and Dale (1984) suggest, children do engage more frequently in creative social role taking with siblings than with anybody else. And since a benefit can be gained from joint pretence with a younger as well as with an older sibling, pretend play is perhaps our best candidate for a cooperative activity which furthers the eventual understanding of false belief.

Finally we need to point out why it is so important to study the socially facilitating factors for developing a theory of mind and, in particular, understanding of false belief. Its importance is clearly illustrated by Chan- dler, Fritz, and Hala's (1991) report that individual differences in understanding false belief covary with parents' and teachers' ratings of social and interpersonal maturity. And one can see why such a relation should exist. As mentioned above, understanding of false belief is of central importance for deception, which is an important skill to acquire if the child does not want to become an easy victim for her more malevolent peers. It is also essential for important moral distinctions such as that between mistakes and lies, a grasp of which is achieved in the wake of understanding false belief (Wim- mer, Gruber, & Perner, 1984). And understanding of higher-order mental states (e.g., of second-order beliefs, i.e., of what a person thinks someone else thinks) is involved in the distinguishing of various speech acts (Perner, 1988), such as distinguishing lies from jokes (Leekam, 1991) and deception from irony (Winner & Leekam, 1991). It may also be part of a deeper understanding of social commitment (Mant & Perner, 1988). Because of the impact that the understanding of false belief has on such a variety of socially relevant distinctions, the study of social factors helping the child to acquire such an understanding is crucial.

References Astington, J. W., Harris, P. L., & Olson, D. R.

(Eds.). (1988). Developing theories of mind. New York: Cambridge University Press.



Belmont, L., & Marolla, F. A. (1973). Birth order, family size, and intelligence. Science, 182, 1096-1101.

Carey, S. (1985). Conceptual change in childhood. Cambridge, MA: MIT Press.

Carey, S. (1988). Conceptual differences between children and adults. Mind and Language, 3, 167-181.

Case, R. (1985). Intellectual development: Birth to adulthood. Orlando, FL: Academic Press.

Case, R. (1989). A neo-Piagetian analysis of the child's understanding of other people, and the internal conditions which motivate their behavior. Paper presented at the biennial meeting of the Society for Research in Child Development, Kansas City, MO, April 1989.

Chandler, M. J., Fritz, A. S., & Hala, S. M. (1989). Small scale deceit: Deception as a marker of two-, three-, and four-year-olds' early theories of mind. Child Development, 60, 1263-1277.

Chandler, M. J., Fritz, A. S., & Hala, S. M. (1991). Children's theories of mental life and social practices. Paper presented at the biennial meeting of the Society for Research in Child Development, Seattle, WA, April 1991.

Dixon, W. J., Brown, M. B., Engelman, L., Frane, J. W., Hill, M. A., Jennrich, R. I., & Toporek, J. D. (1981). BMDP Statistical Software. Berkeley: University of California Press.

Doherty, M. J. (1992). Getting the controls right: Is early deception really flexible? Paper presented at the British Psychological Society Developmental Psychology Section's annual conference, University of Edinburgh, Sep- tember 1992.

Doise, W. (1985). On the social development of the intellect. In V. L. Shulman, L. C. R. Re- staino-Baumann, & L. Butler (Eds.), The fu- ture of Piagetian theory: The neo-Piagetians (pp. 99-121). New York: Plenum.

Donaldson, M. (1978). Children's minds. London: Fontana.

Dunn, J. (1984). Sisters and brothers. London: Fontana.

Dunn, J. (1988). The beginnings of social understanding. Oxford: Blackwell.

Dunn, J., Bretherton, I., & Munn, P. (1987). Con- versations about feeling states between mothers and their young children. Developmental Psychology, 23, 1-8.

Dunn, J., Brown, J., Slomkowski, C., Tesla, C., & Youngblade, L. (1991). Young children's understanding of other people's feelings and beliefs: Individual differences and their ante- cedents. Child Development, 62, 1352-1366.

Dunn, J., & Dale, N. (1984). I a daddy: 2-year-olds' collaboration in joint pretend with sibling and with mother. In I. Bretherton (Ed.), Symbolic play (pp. 131-158). New York: Academic Press.

Flavell, J. H. (1988). The development of children's knowledge about the mind: From cognitive connections to mental representations. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 244-267). New York: Cambridge Univer- sity Press.

Flavell, J. H., Green, F. L., & Flavell, E. R. (1986). Development of knowledge about the appearance-reality distinction. Monographs of the Society for Research in Child Development, 51(1, Serial No. 212).

Fodor, J. A. (1981). The current state of the innateness controversy. In J. A. Fodor, Repre- sentations. Cambridge, MA: MIT Press.

Fodor, J. A. (1987). Psychosemantics: The problem of meaning in the philosophy of mind. Cambridge, MA: MIT Press.

Fodor, J. A. (1992). A theory of the child's theory of mind. Unpublished manuscript, Rutgers Center for Cognitive Science.

Forguson, L., & Gopnik, A. (1988). The ontogeny of common sense. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 226-243). New York: Cam- bridge University Press.

Gopnik, A. (1988). Conceptual and semantic development as theory change. Mind and Lan- guage, 3, 197-217.

Hala, S., Chandler, M., & Fritz, A. S. (1991). Fledgling theories of mind: Deception as a marker of three-year-olds' understanding of false belief. Child Development, 62, 83-97.

Harris, P. L. (1991). The work of the imagination. In A. Whiten (Ed.), Natural theories of mind: The evolution, development and simulation of everyday mindreading (chap. 19). Oxford: Blackwell.

Hogrefe, G. J., Wimmer, H., & Perner, J. (1986). Ignorance versus false belief: A developmental lag in attribution of epistemic states. Child Development, 57, 567-582.

Hughes, M., & Donaldson, M. (1979). The use of hiding games for studying the coordination of viewpoints. Educational Review, 31, 133- 140.

Karmiloff-Smith, A. (1988). The child is a theoreti- cian, not an inductivist. Mind and Language, 3, 182-195.

Keil, F. C. (1989). Concepts, kinds, and cognitive development. Cambridge, MA: MIT Press.

Kitcher, P. (1988). The child as parent of the scien- tist. Mind and Language, 3, 217-228.

LaFreniere, P. J. (1988). The ontogeny of tactical deception in humans. In R. W. Byrne & A. Whiten (Eds.), Machiavellian intelligence: Social expertise and the evolution of intellect in monkeys, apes, and humans (pp. 238-252). Oxford: Clarendon Press.

Leekam, S. (1991). Jokes and lies: Children's un-


1238 Child Development

derstanding of intentional falsehood. In A. Whiten (Ed.), Natural theories of mind (pp. 159-174). Oxford: Blackwell.

Leslie, A. M. (1987). Pretense and representation: The origins of "Theory of Mind." Psychologi- cal Review, 94, 412-426.

Leslie, A. M., & Roth, D. (1993). What autism teaches us about metarepresentation. In S. Baron-Cohen, H. Tager-Flusberg, & D. Co- hen (Eds.), Understanding other minds: Per- spectives from autism (pp. 83-111). Oxford: Oxford University Press.

Lewis, C., & Osborne, A. (1990). Three-year olds' problems with false beliefs: Conceptual deficit or linguistic artifact? Child Develop- ment, 61, 1514-1519.

Light, P. (1979). The development of social sensitivity. Cambridge: Cambridge University Press.

Mant, C. M., & Perner, J. (1988). The child's understanding of commitment. Developmental Psychology, 24, 343-351.

McGarrigle, J., & Donaldson, M. (1974-1975). Con- servation accidents. Cognition, 3, 341-350.

Pascual-Leone, J. (1970). A mathematical model for the transition rule in Piaget's developmental stages. Acta Psychologica, 32, 301-345.

Perner, J. (1988). Higher-order beliefs and intentions in children's understanding of social interaction. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind (pp. 271-294). New York: Cambridge University Press.

Perner, J. (1991). Understanding the representational mind. Cambridge, MA: MIT Press.

Perner, J., Baker, S., & Hutton, D. (in press). Pre- lief: The conceptual origins of belief and pretence. In C. Lewis & P. Mitchell (Eds.), Ori- gins of an understanding of mind. Hove, East Sussex: Erlbaum.

Perner, J., Leekam, S. R., & Wimmer, H. (1987). Three-year olds' difficulty with false belief: The case for a conceptual deficit. BritishJour- nal of Developmental Psychology, 5, 125- 137.

Perner, J., & Wimmer, H. (1988). Misinformation and unexpected change: Testing the development of epistemic-state attribution. Psycho- logical Research, 50, 191-197.

Perret-Clermont, A. N. (1980). Social interaction and cognitive development in children. Lon- don: Academic Press.

Peskin, J. (1992). Ruse and representation: On children's ability to conceal information. De- velopmental Psychology, 28, 84-89.

Pillow, B. H. (1989). Early understanding of perception as a source of knowledge. Journal of Experimental Child Psychology, 47, 116-129.

Pratt, C., & Bryant, P. (1990). Young children understand that looking leads to knowing (so

long as they are looking into a single barrel). Child Development, 61, 973-982.

Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behav- ioral and Brain Sciences, 1, 516-526.

Rose, S. A., & Blank, M. (1974). The potency of context in children's cognition: An illustra- tion through conservation. Child Develop- ment, 45, 499-502.

Ruffman, T., Olson, D. R., Ash, T., & Keenan, T. (1993). The ABCs of deception: Do young children understand deception in the same way as adults? Developmental Psychology, 29, 74-87.

Russell, J., Mauthner, N., Sharpe, S., & Tidswell, T. (1991). The "windows task" as a measure of strategic deception in preschoolers and au- tistic subjects. British Journal of Develop- mental Psychology, 9, 331-349.

Sellars, W. (1956). Empiricism and the philosophy of mind. In K. Gunderson (Ed.), Minnesota studies in the philosophy of science (Vol. 1, pp. 253-329). Minneapolis: University of Minnesota Press.

Siegal, M., & Beattie, K. (1991). Where to look first for children's knowledge of false belief. Cognition, 38, 1-12.

Sodian, B. (1991). The development of deception in young children. British Journal of Devel- opmental Psychology, 9, 173-188.

Sodian, B., Taylor, C., Harris, P. L., & Perner, J. (1991). Early deception and the child's theory of mind: False trails and genuine markers. Child Development, 62, 468-483.

Stouthamer-Loeber, M. (1986). Adults' perception of verbal misrepresentation of reality in four- year-olds. Unpublished manuscript, Univer- sity of Pittsburgh, Western Psychiatric Insti- tute and Clinic.

Vygotsky, L. S. (1965). Thought and language. Cambridge, MA: MIT Press.

Wellman, H. M. (1990). The child's theory of mind. Cambridge, MA: MIT Press.

Wimmer, H., Gruber, S., & Perner, J. (1984). Young children's conception of lying: Lexical realism-moral subjectivism. Journal of Ex- perimental Child Psychology, 37, 1-30.

Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception. Cognition, 13, 103-128.

Winner, E., & Leekam, S. (1991). Distinguishing irony from deception: Understanding the speaker's second-order intention. British Journal of Developmental Psychology, 9, 257-270.

Zajonc, R. B., & Markus, G. B. (1975). Birth order and intellectual development. Psychological Review, 82, 74-88.


Article Contentsp. [1228]p. 1229p. 1230p. 1231p. 1232p. 1233p. 1234p. 1235p. 1236p. 1237p. 1238

Issue Table of ContentsChild Development, Vol. 65, No. 4, Aug., 1994Front MatterReview PaperThe Insecure/Ambivalent Pattern of Attachment: Theory and Research [pp. 971 - 991]

Sleeping out of Home in a Kibbutz Communal Arrangement: It Makes a Difference for Infant-Mother Attachment [pp. 992 - 1004]The Quechua Manta Pouch: A Caretaking Practice for Buffering the Peruvian Infant against the Multiple Stressors of High Altitude [pp. 1005 - 1013]Attachment Patterns at Age Six in South Germany: Predictability from Infancy and Implications for Preschool Behavior [pp. 1014 - 1027]Responses of Free-Ranging Rhesus Monkeys to a Natural Form of Social Separation. I. Parallels with Mother-Infant Separation in Captivity [pp. 1028 - 1041]Androgen and the Development of Human Sex-Typical Behavior: Rough-and-Tumble Play and Sex of Preferred Playmates in Children with Congenital Adrenal Hyperplasia (CAH) [pp. 1042 - 1053]Children's Moral Reasoning about Family and Peer Violence: The Role of Provocation and Retribution [pp. 1054 - 1067]Group Social Context and Children's Aggressive Behavior [pp. 1068 - 1079]Factors Related to the Achievement and Adjustment of Young African American Children [pp. 1080 - 1094]Developmental Changes in Achievement Evaluation: Motivational Implications of Self-Other Differences [pp. 1095 - 1110]Beyond Parental Control and Authoritarian Parenting Style: Understanding Chinese Parenting Through the Cultural Notion of Training [pp. 1111 - 1119]Associations between Parental Psychological and Behavioral Control and Youth Internalized and Externalized Behaviors [pp. 1120 - 1136]Family Acceptance and Family Control as Predictors of Adjustment in Young Adolescents: Linear, Curvilinear, or Interactive Effects? [pp. 1137 - 1146]Adolescents' and Parents' Conceptions of Parental Authority and Personal Autonomy [pp. 1147 - 1162]Two-Year-Olds Readily Learn Multiple Labels for the Same Basic-Level Category [pp. 1163 - 1177]Children's Use of Context in Interpreting "Big" and "Little" [pp. 1178 - 1192]Children's Perception of Proportion in Graphs [pp. 1193 - 1213]Discrepant Utterances Resulting from a False Belief: Children's Evaluations [pp. 1214 - 1227]Theory of Mind Is Contagious: You Catch It from Your Sibs [pp. 1228 - 1238]Back Matter [pp. 1239 - 1241]

1131316

Documents