Symposium
Chairs: Claudia Uller and Shaun Nichols
Discussant: Alan M. Leslie
One of the most fascinating topics in human cognition is our remarkablecapacity to understand other minds. The origins of this capacity have beenstudied by two different research traditions. One tradition has exploredtheory of mind capacities in nonhuman primates in an effort to chart thephylogenetic origins of theory of mind. The other, more recent tradition,has used looking time measures to explore the ontogenetic origins of theoryof mind in human infants. Experimental data indicate that chimpanzees failsimple theory of mind tasks that are easily passed by human children. Onthe other hand, recent evidence on human infants indicates that a primitivecapacity for theory of mind emerges in infancy.This symposium will include contributions from researchers in bothtraditions in an attempt to address theoretical questions about whethertheory of mind depends on capacities that are distinctively human or whetheryoung children and nonhuman primates have common shared mechanisms. Onemain objective of this symposium is to begin to generate fruitfulconversations between the two traditions. The symposium will start with apresentation by Daniel Povinelli entitled 93Theory of mind is not aninevitable byproduct of social evolution94. This presentation will befollowed by Tomasello and Call's 93Chimpanzees know what con-specifics do anddo not see94. Uller and Nichols will give a presentation entitled93Precursors of theory of mind in chimpanzees: A comparison with humaninfants94, followed by Gergely and Csibra's 93The teleological origins ofnaEFve theory of mind in infancy94. Alan Leslie will give the finalpresentation as the discussant of the session. We are confident that theproposed symposium will advance our knowledge of the extraordinary humancapacity to understand other minds by investigating primitive theory of mindcapacities in infants and nonhuman primates.
Details of individual items:
paper
It has been proposed that theory of mind evolved to cope with complex socialproblems that evolved in the context of group-living. On this view, the mosteffective solution to reasoning about the behaviors of others was to makeinferences about their mental states. I dispute this view of the evolutionof theory of mind. Empirically, I show that humanity's nearest livingrelative, the highly social chimpanzee, does not appear to reason aboutmental states. Theoretically, I offer a series of reasons to suppose thattheory of mind probably did not evolve to solve on-line social problems. Ipropose an alternative model which posits that the majority of the socialbehaviors that humans and other primates share in common emerged long beforethe human lineage evolved the psychological means of interpreting thosebehaviors in mentalistic terms. Although humans and chimpanzees are bothexperts at reasoning about the behavior of others, and both species may besaid to possess mental states, humans alone may have evolved a cognitivespecialization for
paper
Three studies of chimpanzee social problem solving are reported. In eachstudy subordinant and a dominant individual were put into competition overtwo pieces of food. In all studies dominants obtained virtually all of thefoods to which they had good visual and physical access. Subordinates weresuccessful quite often, however, in three different situations in which theyhad better visual access to the food than the dominant, for example, whenthe food was positioned so that only the subordinate (and not the dominant)could see it. In some cases (e.g., Studies 1 and 2) the subordinate mayhave been monitoring the behavior of the dominant directly and so simplyavoiding the food that the dominant was moving toward (which just happenedto be the one it could see). In other cases, however (e.g., Study 3), thispossibility was ruled out by giving subordinates a small head-start andforcing them to make their choice (to go to the food that both competitorscould see, or to go to the food that only they could see) before thedominant was released into the area. Together with other recent studies,the current studies suggest that chimpanzees know what con-specifics can andcannot see, and, further, that they use this knowledge to devise effectivesocial-cognitive strategies in naturally occurring social problem solvingsituations.
paper
Over the last few years, several studies show that chimpanzees fail a rangeof relatively simple theory of mind tasks on the attribution of perceptualstates (e.g., Povinelli & Eddy, 1996). In light of these and other data,many researchers on nonhuman primates have suggested that nonhuman primatesin the wild may not have any understanding of the mind at all. However,others have suggested that the failures of chimpanzees in these experimentscan be attributed to excessive task demands. We maintain that recent workon human infants suggests an alternative way to explore these issues.Researchers in the developmental tradition have used the looking timemeasure to explore the understanding of the mind in human infants. Recentresults using this measure have been interpreted as evidence that humaninfants attribute goals (Gergely et al. 1995; Sodian & Thoermer, submitted;Woodward 1998). In an attempt to bridge these two research communities, wepresented chimpanzees with stimuli modeled on the infant studies (e.g.,Gergely et al. 1995). Four chimpanzees were shown a computer-generatedstimuli task in which a block moves above a 93barrier94 and makes contact witha ball. After familiarization trials, the chimpanzees saw test trials inwhich there is no barrier between the block and the ball. The trajectory ofthe block is either the same as in familiarization trials (parabolic) or itis a straight line towards the ball. If the chimpanzees interpret the blockas trying to achieve a goal in the most direct way, then they should looklonger at the old trajectory event than at the new trajectory event. If, onthe other hand, the chimpanzees simply prefer perceptual novelty, then theyshould look longer at the novel trajectory event than at the old trajectoryevent. In our study, the chimpanzees looked longer at the old trajectoryevent than the new trajectory event. These results show that thechimpanzees responded the way infants do. Thus, if the evidence withinfants reveals that infants attribute goals, then our evidence suggeststhat chimpanzees also attribute goals. Furthermore, if this interpretationis right, these results indicate that the capacity for goal attribution doesnot depend on language or on enculturation because these chimpanzees wereneither language trained nor enculturated.
paper
In this talk we shall summarize the results of a series of visualhabituation studies (Gergely et al., 1995; Csibra et al., in press) with 9-and 12-month-old infants which indicate a precocious ability to interpretbehaviour in terms of attributed goals and to evaluate the rationality ofgoal approach when viewing the spatial behaviour of abstractcomputer-animated figures. We shall consider and contrast severaltheoretical proposals (e.g., Premack & Premack, 1997; Kelemen, 1999; Gergely& Csibra, 1997; Csibra & Gergely, 1998) that have been advanced concerningthe nature of the interpretational system underlying this early competenceand its developmental relation to the young child12s theory of mind. Inparticular, we shall consider arguments for and against two alternativepositions. According to the first, the demonstrated ability of one-year-oldsto interpret behaviour in terms of goals and to infer the novel meansactions of agents in new situations reflects the early presence of an innatetheory of mind which generates causal mentalistic action explanations interms of intentional mind states. In contrast, we shall argue for thealternative position according to which one-year-olds apply a reality-based93teleological interpretational system94 which does not yet involverepresenting intentional mind states. This nonmentalistic interpretationalstance is, however, theory-like in so far as it exhibits inferentialgenerativity and systematicity and it applies an abstract principle ofreasoning, the principle of rational action. We propose that the causalmentalistic action explanations of later theory of mind are driven by thesame inferential principle of rationality and that theory of mind developsout of the (ontogenetically and possibly philogenetically) earlier andindependent teleological interpretational system by extending itsontological domain to include representations of fictional (mental) statesas well as representations of reality. Finally, we shall consider thepotentially crucial role of comparative studies in evaluating thealternative positions discussed. We shall argue that demonstrations ofteleological reasoning in apes (Uller & Nichols, forthcoming) and/or inchildren with autism in conjunction with a concurrent lack of understandingintentional mind states (e.g., Povinelly & Eddy, 1996; Baron-Cohen et al.,1985) may provide empirical support for the claim that the nonmentalistic93teleological stance94 is a separate phylogenetic adaptation to interpretgoal-directed behaviour that is independent (but at the same time may formthe developmental core) of naive theory of mind.