PRIVATE Running Head: THEORY OF MIND AND DIVERGENT THINKING Theory of Mind and the Origin of Divergent Thinking Thomas Suddendorf and Claire M. Fletcher-Flinn Department of Psychology University of Auckland Auckland, New Zealand ph: 649-3737599 e-mail: t.suddendorf@auckland.ac.nz Abstract The development of a `theory of mind' may not only be important for understanding the minds of others but also for using one's own mind. To investigate this supposition, forty children between the ages of three and four were given false- belief and creativity tasks. The numbers of appropriate and of original responses in the creativity test were found to correlate positively with performance on false-belief tasks. This association was robust, as it continued to be strong and significant even when age and verbal intelligence were partialled out. The results support the hypothesis that the metarepresentational skills involved in theory of mind also affect the way children can access and scan their own mental repertoire beyond the areas of currently activated content (i.e. divergent thinking). With the advent of theory of mind a basic cognitive shift takes place in human development, and possibly took place in cognitive evolution. Theory of Mind and the Origin of Divergent Thinking Introduction Theory of mind (ToM), or the appreciation of the representational nature of mind and its relation to behavior, has become the subject of much research in developmental and comparative psychology (e.g. Cheney & Seyfarth, 1990; Gopnik, 1993; Perner, 1991; Premack & Woodruff, 1978). Although evidence has accumulated that great apes (e.g. Byrne & Whiten, 1992; Premack & Woodruff, 1978) and very young children (e.g. Gopnik & Slaughter, 1991; Wellman, 1991) have some understanding of their own and others' desires and intentions, neither have been shown to understand informational states, such as those of knowledge and belief (e.g. Premack & Dasser, 1991; Wimmer & Perner, 1983). This latter ability appears to be reserved for humans older than about three years, and only then can a representational ToM be acertained. Tasks involving the attribution of false beliefs have become the standard tool for assessing the acquisition of ToM. The attribution of a false belief implies an understanding of the fact that mental states are attitudes to representations of the world rather than attitudes to the "real world" itself. Since the representational contents of true beliefs correspond by definition to actual states of the "real world", only predictions of somebody's behavior when it is based on a false belief unequivocally show an appreciation of the representational nature of mental states. ToM is evident, for example, when the child understands that people may search for a desired object at the place where they think it is rather than at the place where the child knows it is. The individual thus has to acknowledge that a representation can be false while the metarepresentation remains true (e.g., it is true that the other person holds a belief that is false) (e.g. Baron-Cohen, Leslie & Frith, 1985; Dennett, 1987; Flavell, 1993; Gopnik, 1993; Gopnik & Astington, 1988; Wimmer & Perner, 1983). False-belief tasks may also involve an executive component because the individual is required to disengage from the current representational content in order to assume somebody else's (or one's own former) different beliefs (e.g. Russel & Jarrold, in review; Suddendorf & Corballis, in press). When this level of understanding has been achieved (i.e. ToM has been acquired), the child can impute mental states even when these contradict his or her own current mental states, and the diversity of the heterogenous mental world becomes accessible. This mental feat has obvious significance for the development and evolution of social intelligence and most research has focused on these social effects (e.g. Astington & Jenkins, 1995; Baron-Cohen, 1995; Baron-Cohen, Leslie & Frith, 1985; Byrne & Whiten, 1992; Dunn, Brown, Slomkowski, Tesla & Youngblade, 1991; Wimmer & Perner, 1983). At the same time, however, it has been claimed that ToM is important also for self-understanding, self-monitoring and self-regulation (e.g. Humphrey, 1986; Perner, 1991; Povinelli, in press; Suddendorf, 1994). Only when the representational nature of mind is understood may one truly reflect and introspect, that is, form beliefs about beliefs (e.g., I must be right with my view that...), attitudes about knowledge (e.g., I don't want to know), second-order motives (e.g., I don't want my desire to play to interfere with my work), and so forth. It might thus be expected that skills that apparently depend upon mental access to one's own mind (e.g. knowing that and what one knows) improve dramatically with the acquisition of ToM (i.e. when false-belief tasks are passed). Divergent thinking might be such a skill. Several researchers have identified metarepresentation as an important factor in creativity. Determining whether possible solutions fulfill the criteria of the problem, for example, might be a function of metacognition (e.g. Ebert, 1994; Feldhusen, 1995). Divergent thinking by its very definition appears to require the individual to search his/her own knowledge base beyond the currently activated domain of mental content. This may entail the same basic process of executive control or disengagement from current perceptions and knowledge as is required for assuming a belief that is evidently false. On a higher plane (at a later age) disengaging from a current paradigm and "investing" in disregarded areas is, of course, the key to creative new insights (e.g., Sternberg and Lubart, 1991). Further, active scanning of one's knowledge base in search of appropriate answers appears to imply the ability to metarepresent (i.e., to know what one knows). Knowing what others know (as in ToM tasks) and knowing what oneself knows might be very closely related skills. Indeed, some might suggest that the former is an extension of the latter (cf., Harris, 1991). Thus, if it is true that divergent thinking and attribution of false beliefs draw on the same mental developments, one should expect to find strong correlations between tasks that measure these skills. In other words, children who pass false-belief tasks would be expected to do much better on divergent thinking tasks than children who do not, because they should be able to scan knowledge from diverse, otherwise unrelated, domains in the process of generating new, divergent answers that fit the problem criteria (cf. Wallach, 1970). Karmiloff-Smith (1990) described the development of flexible access to originally domain-specific knowledge in a study of children's creative drawings. She, too, argued that changes in children's imaginative power come about because they develop explicit representations of knowledge they already possess implicitly. In other words, her finding supports the idea that the ability to metarepresent (and to disengage from current representations) is critical for generating divergent solutions. The present study was designed to address whether or not children who have a ToM are in fact better at searching their own minds for creative answers. We administered false-belief and creativity tasks to three- and four-year-old children. Further, we included a verbal intelligence test to control for the potentially confounding influence of mental age on correlations between ToM and creativity. Method Subjects Forty children, 23 girls and 17 boys, all of whom took part in a broader study on the relation between computers, gender, and social thinking, were tested. The majority of the children were of middle SES and the mean age was 50 months (range= 36 to 58 months). The children were recruited from three playcentres (on Waiheke Island, New Zealand), with the consent of parents, staff and the children themselves. Individual testing took place in a quiet room in each of these centres. Tasks Creativity Task. The creativity task consisted of two subtasks of Wallach and Kogan's (1965) creativity test. We followed Ward's (1968) adaptation of the task for preschoolers. In the instances subtask, the children were asked to declare all the things that they could think of that were round, or had wheels, or were red. In the uses subtask, they were asked to declare all the ways in which they might use, or play with, a newspaper, a cup, a table knife, and a coat hanger. In a permissive testing situation each suggestion was followed by liberal praise and the child was encouraged to think of more answers by phrases such as: "What else?, Can you think of something else?...". Both tests were presented as games and the time spent on each question continued until the child said that s/he had no further ideas. To standardize the procedure the experimenter offered encouragement at least three times before suggesting that they move to the next item. However, when the child was uncomfortable the questioning was stopped and either the next item was introduced or testing was completed on another occasion. Performance was scored on two different measures: fluency and uniqueness. First, however, all unique responses, that is, ideas that were put forward only by one child, were given to two independent raters and those that both deemed inappropriate (e.g. Q: what is red? A: An orange) were eliminated and not scored. Identical replies from two or more children were automatically considered appropriate. The remaining responses were then scored for each individual a) by adding up the absolute number of ideas (fluency-score), b) by adding up the total number of unique ideas (uniqueness-score). For further details of the procedure see Ward (1968). False-Belief Task. The false-belief task was administered in three trials, the first two following Prior, Dahlstrom and Squires' (1990) description (from a paradigm originally developed by Baron-Cohen, Leslie & Frith, 1985) and the third implementing a suggestion by Russel and Jarrold (in review). Two dolls, Sally and Anne, were introduced to the child and, when the names were learnt, the following story was told and enacted: "Sally has a marble and she puts it into her basket and closes the lid. She says good-bye and goes out to play. Now `naughty Anne' takes the marble out of the basket and places it into the box and closes both lids." The child is then asked a memory and a reality question: "Where did Sally put the marble?" and "Where is the marble now?". (In the exceptional case of a child failing on the memory or reality question, the child was reminded of the true situation and the procedure was repeated from the start.) Then the false-belief question is asked: "When Sally comes back where will she look first for her marble?" The second trial is identical to the first except that the hiding place is the tester's pocket rather than the box. After these two trials the procedure (using the box as the place were Anne hid the marble) was followed once more but the belief question was altered (following Russel & Jarrold's suggestion) to: "Will Sally look at the right or at the wrong place? [right and wrong were counterbalanced]", followed by: "Show me the right place, and show me the wrong place [asking for the indicated, either right or wrong place, first]". Children received zero points for no correct answer to the first two belief questions, one point for one correct answer and two points for answering both questions correctly. The third trial, the variation suggested by Russel and Jarrold, was scored separately with a zero for an incorrect and a one for a correct reply. A reply was only considered correct when the answer "at the wrong place" was followed by prompted correct pointing to the (empty) basket. Verbal Intelligence Test. The British Picture Vocabulary Scale (BPVS; Dunn, Dunn, Whetton & Pintillie, 1982) was introduced to the children as a picture book game and standard format was followed. The raw scores were converted into IQ and mental age. Results The mean IQ of the sample was 97 (sd: 12), ranging from 74 to 120 (Note that British norms were applied to New Zealanders, which possibly acounts for the slight deviation at the lower end of scores). When converted into mental age the mean translated into 48 months (sd: 13) and ranged from 25 to 75 months. False-Belief Tasks Twenty-seven children answered both false-belief questions correctly and 10 failed on both, while 3 gave one correct and one incorrect answer. The Russel and Jarrold version revealed a similar distribution with 26 passing and 14 children failing the test. However there was a discrepancy between the two measurements. Four children who passed both questions of the classic version failed when asked whether Sally would look at the right or the wrong place. And three children who failed at least on one of the first two trials answered this third question correctly. Of the three children (7.5%) who performed ambivalently (i.e. one correct and one incorrect answer) on the two classic trials, one passed the Russel and Jarrold version and was therefore classified as having a ToM (receiving 2 points in total), while the other two failed this task and were classified as not yet having a ToM ( <2 points). The single ToM measure resulted therefore in 28 children with and 12 children without a ToM. Creativity Tasks Ten unique responses (items mentioned only by one child in the study) were judged inappropriate by two independent raters and thus were excluded from further analysis. This constituted 4.8% of all unique items. The remaining items were then scored as fluency (total number of responses) and uniqueness (number of unique responses) values for the instances and uses conditions. The top of Table 1 presents the means and standard deviations of these measures. _____________________________ Insert Table 1 about here _____________________________ Relationship Between False-Belief and Creativity Tasks A t-test to examine the effect of having or not having a ToM on the performance on the creativity tasks was performed. A significant effect was found on both, the total fluency scores [t(38) = 4.85, p < .001] and the uniqueness scores [t(38) = 3.44, p < .005]. Biserial correlations between ToM and total fluency and uniqueness scores were .62 and .48 respectively, accounting for 38 and 23% of the variance of performance (see Table 1). Mental age (chronological age and IQ combined) correlated significantly with the creativity measures (fluency: r = .47, p < .01; uniqueness: r = .40, p < .05), although Wallach and Kogan (1965) specifically designed the test to measure creativity independently of general intelligence. Thus, the association between ToM and creativity might have been caused by the mediating variables of age and intelligence. In order to determine whether this was the case, we calculated partial correlations controlling for these variables. Neither partialling out chronological age (fluency: r = .58, p < .001; uniqueness: r = .45, p < .005) nor BPVS scores (fluency: r = .52, p < .001; uniqueness: r = .40, p < .01) reduced the association between ToM and creativity below significance. In fact, the correlations remained significant and relatively strong (fluency: r = .51, p < .005; uniqueness: r = .40, p < .01) even when age and scores on the BPVS were partialled out. Thus, results show a robust association between ToM and creativity. Discussion This is the first demonstration of a relationship between the attribution of false beliefs and the amount and uniqueness of creative responses in preschoolers. The fluency scores can be regarded as a quantitative measure of the success of mind search. The uniqueness scores, while not independent from fluency scores, provide an estimate for divergence of the search. That no other child produced the item suggests that it is semantically removed from close association in most peers (perhaps even within the child's own semantic net). Producing many unique items therefore suggests either that the individual has a very distinct semantic network from those of the other children or, perhaps more plausibly, that the child retrieved the items from a wider search (i.e. divergent thinking). Because the creativity tasks did not involve any obvious kind of mental attribution, this finding points to another factor underlying both measures. The prime candidates, since the relationship holds even when intelligence and age are partialled out, is improved metarepresentational capacity and the ability to disengage from the immediate present. Understanding false beliefs in others requires the individual to dissociate from the immediate situation and to form a representation of the other's representation. Similarly, one may argue that the creativity task requires the children to dissociate from the immediate situation and to represent one's own knowledge, scanning it for items with a particular feature. This theoretical argument is consistent with the informal observation that during the testing procedure younger children tended to look for answers in their immediate environment (e.g. eyes=round, shirt=red), while older children gazed at the ceiling, apparently looking "inside" for appropriate responses. The data support the hypothesis that a general, rather than a specifically social, representational improvement takes place between age 3 and 4. However, we cannot rule out that other factors that we did not measure, and thus did not partial out, might have been responsible for the association we observed. Controlling for differences in age and verbal intelligence may cover the most likely interfering variables, but others are possible. Future research has to address these possibilities and preferably tackle the developmental changes in a longitudinal study. Despite the preliminary nature of the finding, if supported by other research, the association between ToM and creative thought might have far-reaching consequences on the way we view not only the ontogeny but also the phylogeny of the human mind. Creative thought, like language, requires informational access to varied domains of knowledge in the generative process of combining and recombining items into virtually infinite numbers of novel sequences. Metaknowledge may be essential for this process to properly unfold. Corballis (1991) claimed that generativity is unique to the human species. Considering the fact that to date there is no convincing evidence for metaknowledge in animals (Cheney & Seyfarth, 1990; Heyes, 1993; Premack & Dasser, 1991), the present data support this claim. Cheney and Seyfarth (1990) examined the evidence for ToM in nonhuman primates and noted that monkeys possess what they called a laser beam intelligence. That is, monkeys act apparently intelligently in one domain, while being unable to transfer that knowledge for application in another. Humans are very skilled at such transfer. With metarepresentation the domain specifity may have been overcome. Evolutionarily, the development of ToM would then have been an important factor not only in the social domain and in understanding the self, but also in the utilisation of mental capacities through metarepresentation, as in the case of divergent thinking. In this way social intelligence, which according to the Machiavellian intelligence hypothesis (Byrne & Whiten, 1992; Humphrey, 1976; Jolly, 1967) gave rise to human intellectual evolution, might have paved the way for new ways of using the mind. Flexible transfer of knowledge between different domains is one of the hallmarks of humans' relentless creativity and invention. Summary Passing false-belief tasks has been found to correlate positively with the amount and originality of answers produced on the creativity task. The association was found to be robust as it remained strong and significant even when age and verbal intelligence were partialled out. This result was taken as preliminary support for the hypothesis that both performances are dependent upon metarepresentational skills and an ability to disengage from current mental content. The emerging theory of mind may therefore not only be important for social understanding but also for understanding and utilizing one's own mind. This suggests that by the age of four children have developed the cognitive foundation which allows them to actively engage in divergent thinking. References Astington, J.W., & Jenkins, J.M. (1995). Theory of mind development and social understanding. Cognition and Emotion, 9, 151-65. Baron-Cohen, S. (1995). Mindblindness: an essay on autism and theory of mind. Cambridge, MA: MIT Press. Baron-Cohen, S., Leslie, A., & Frith, U. (1985). Does the autistic child have a `theory of mind'? Cognition, 21, 37-46. Byrne, R.W., & Whiten, A. (1992). Cognitive evolution in primates. Man, 27, 609-627. Cheney, D.L., & Seyfarth, R.M. (1990). How monkeys see the world. Chicago: University of Chicago Press. Corballis, M.C. (1991). The lopsided ape. New York: Oxford University Press. Dennett, D. (1987). The intentional stance. Cambridge, MA: MIT Press. 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 antecedents. Child Development, 62, 1352-66. Dunn, L.M., Dunn, L., Whetton, C., & Pintillie, D. (1982). British picture vocabulary scale. London: Nfer-Nelson. Flavell, J.H. (1993). The development of children's understanding of false belief and the appearance-reality distinction. International Journal of Psychology, 28, 595-604. Fletcher-Flinn, C.M., & Suddendorf, T. (in press). Do computers affect the mind? Journal of Educational Computing Research. Gopnik, A. (1993). How we know our own minds: The illusion of first-person knowledge of intentionality. Behavioral and Brain Sciences, 16, 1-14. Gopnik, A., & Astington J.W. (1988). Children's understanding of representational change and its relation to the understanding of false belief and the appearance-reality distinction. Child Development, 59, 26-37. Gopnik, A., & Slaughter, V. (1991). Young children's understanding of changes in their mental states. Child Development, 62, 98-110. Harris, P.L. (1991). The work of imagination. In A. Whiten (Ed.), Natural theories of mind: Evolution, development, and simulation of everyday mindreading (pp. 283- 304). Oxford: Basil Blackwell. Heyes, C.M. (1993). Anecdotes, training, trapping, and triangulation: Do animals attribute mental states? Animal Behavior, 46, 177-188. Humphrey, N.K. (1976). The social function of intellect. In P.P.G. Bateson & R.A. Hinde (Eds.), Growing points in ethology (pp. 303-317). Cambridge: Cambridge University Press. Humphrey, N.K. (1986). The inner eye. London: Faber & Faber. Karmiloff-Smith, A. (1990). Constraints on representational change: evidence from children's drawings. Cognition, 34, 57-83. Perner, J. (1991). Understanding the representational mind. Cambridge, MA: MIT Press. Povinelli, D.J. (in press). The undublicated self. In P. Rochat (Ed.), The self in early infancy. Amsterdam: North Holland-Elsevier. Premack, D., & Dasser, V. (1991). Perceptual origins and conceptual evidence for theory of mind in apes and children. In A. Whiten (Ed.), Natural theories of mind: evolution, development and simulation of everyday mindreading (pp. 253- 266). Oxford: Basil Blackwell. Premack, D. & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral & Brain Sciences, 4, 515- 526. Prior, M., Dahlstrom, B., & Squires, T.L. (1990). Autistic children's knowledge of thinking and feeling states in other people. Journal of Child Psychology and Psychiatry, 51, 587-601. Russel, J., & Jarrold, C. (in review). Executive factors in the false belief task, and prospects of an executive account of theory of mind development. Sternberg, R.J., & Lubart, T.I. (1991). An investment theory of creativity and its development. Human Development, 34, 1-31. Suddendorf, T, & Corballis, M.C. (in press). Mental time travel and the evolution of the human mind. Genetic, Social, and General Psychology Monographs. Suddendorf, T. (1994). Discovery of the fourth dimension: Mental time travel and human evolution. M.Soc.Sci. Thesis, University of Waikato, Hamilton, New Zealand (available on request from the author). Wallach, M.A. (1970). Creativity. In A. Mussen (Ed.), Carmichel's manual of child psychology (3rd ed.) (pp. 1211- 72). New York: Wiley. Wallach, M.A., & Kogan, N. (1965). A new look at the creativity-intelligence distinction. Journal of Personality, 33, 348-69. Ward, W.C. (1968). Creativity in young children. Child Development, 39, 737-754. Wellman, H.M. (1991). From desires to beliefs: Acquisition of a theory of mind. In A. Whiten (Ed.), Natural theories of mind: Evolution, development, and simulation of everyday mindreading (pp. 19-38). Oxford: Blackwell. 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. Author Note Thomas Suddendorf, Department of Psychology; Claire M. Fletcher-Flinn, Department of Psychology. The performance of this sample on the false-belief tasks has been reported elsewhere in the context of a study on preschoolers computer use (Fletcher-Flinn and Suddendorf, in press). This research was supported by a Telecom, New Zealand research grant to the second author. We thank the teachers, parents and children of the Waiheke Kindergarten, Chris' Creche, and Waiheke Playcenter for their participation, hospitality and support. We also thank M.C. Corballis for valuable comments on an earlier version of this paper. Correspondence concerning this article should be addressed to Thomas Suddendorf, Department of Psychology, University of Auckland, Private Bag 92019, Auckland, New Zealand. Electronic mail may be send to t.suddendorf@auckland.ac.nz Table 1 Response Means, Standard Deviations, and Correlations for Creativity Measures --------------------------------------------------------------- Measures -------------------------------------------------------- Instances Uses Total -------------------------------------------------------- Uniqueness Fluency Uniqueness Fluency Uniqueness Fluency --------------------------------------------------------------- Mean 2.53 7.80 2.45 7.08 5.00 14.88 SD 2.08 4.56 2.10 2.81 3.71 6.65 --------------------------------------------------------------- Correlations: ToM .49** .57** .35* .55** .48** .62** Mental age .31 .40* .37* .48** .40* .47** --------------------------------------------------------------- * p < .05 ** p < .01 ToM and divergent thinking