Experience, action and theory of mind Joëlle Proust (CNRS) [email protected] CREA, Ecole Polytechnique, 1 Rue Descartes, 75005 Paris

--------------James Russell assumes that a theory of mind must be grounded "in the firstorder experience of controlling, within limits, one's mental life". The Piagetian flavour in this claim is that knowledge about the self is gained through the exercise of an executive competence. Exercising self-control is the "soil" on which an explicit theory of mental life will eventually flourish. In contrast with the Historical Piagetian theory, Russell further acknowledges that a child's ability to "become a second-order representer" cannot emerge from "simply doing a lot of first-order representation". Russell supplements the Piagetian theory with a representational theory of agency, according to which the structured expectations about the results of one's actions are what allows an infant to grasp object permanence. James Russell has thus brought two important ideas to bear on the theory of mind domain of research. The first is that mentalizing as a representational capacity relies in part on executive capacities (such as resisting prepotent stimuli, maintaining a representation active in working memory in delayed execution, etc). The second is that language -- as an innate capacity for symbolic thought-- is a necessary condition for grasping the relation between propositional attitude and mental content, and for developping a second-order thought. This two-tiered theory offers a welcome functional-developmental alternative to the modular view on mental concepts acquisition. This line of investigation appears to me inspiring and fruitful. I would like to question only a particular aspect of Russell's argumentative strategy. A major theoretical issue that Russell raises by insisting that a theory of mind be grounded in "the first-order experiences of controlling one's mental life" is whether the relevant facts have to be experienced in order to play a causal role in theory of mind acquisition. To use the philosophical jargon : does the "feeling like" associated to R-negation or to structured expectations play a causal role in building up a self ? Several arguments may lead to question the causal relevance of what could be called agency qualia. Although folk-psychology has it that only experienced events and properties can be memorized and recalled, scientific psychology considers that implicit memory may also store regularities and influence behavior (Kelley & Jacoby, 1993). Furthermore, as shown by Nisbett & Wilson (1977), agents are often wrong about why they acted: the personal level may be more appropriate to the demands of social cohesion than to individual psychological explanations of intentional action. Thus one could suggest that what drives mental states understanding is not so much pretheoretical experience of mentality as a practical, largely implicit knowing-how to achieve mental control.

The question of what are the respective roles for consciously accessible states and for informational states and processes (independently from their conscious availability) leads to the question of subpersonal vs.personal explanations of behavior. In his target paper as well as in his (1996) book, Russell accepts the view that agency should be explicated in subpersonal terms, but denies that acting at will can be accounted for in subpersonal terms. He takes the personal level to be an autonomous causal-explanatory level. His main argument is again that the best way of understanding our adult experience of willful agency consists in grounding it on a developing sense of what it feels like to be an agent, while a subpersonal account fails to ground the full range of the endogeneous control of behavior. To this one may object that although the experience of agency is indeed a universal one (at least in normal, healthy people), it is far from obvious that an adequate theory for agency and self-development can be built at that personal level. The global experience of how it feels to be an agent supervenes on sets of specialized subpersonal mechanisms. Objectivity - the sense of externality - could emerge from subpersonal mechanisms for extracting spatial crossmodal invariants from perceptual inputs (Proust, 1997). The sense of directing one's own acts (the sense of who is in control) could depend on various subpersonal mechanisms like efference copy or from a subpersonal comparison of actual versus projected reafferences ( See Jeannerod, in press and Proust, in press). The theoretical choice of defining action as behavior caused by (personal-level) reasons may also be questioned both on psychological and philosophical grounds. Besides the cognitive operations linked to action monitoring, on which Russell concentrates, another class of executive processes may be an important functional precursor of mentalizing capacities, namely the various types of selective focussing that together define the scope of mental actions. A mental action is a sequence of mental operations that are triggered to reach certain desirable informational or motivational states (Proust, 1998). Selective attention, directed recall, as well as the various metacognitive processes by which a subject appreciates the feasibility of a mental task (reasoning, planning, recalling, problem solving) and the validity of an outcome are certainly major executive building blocks for a mentalizing capacity. Investigating those metacognitive capacities might help understand what is at stake in executive abilities in general and illuminate the transfer from tacit to explicit self-knowledge.

Jeannerod, M., (in press 1999), To act or not to act. Perspectives on the representation of actions, Quarterly Journal of Experimental Psychology. Kelley, C.M., & Jacoby, L.L., (1993), The construction of subjective experience : memory attributions, in M. Davies & G.W. Humphreys (dirs.), Consciousness, Psychological and Philosophical Essays, 74-89, Oxford, Blackwell. Nisbett, R.E. & Wilson, T.D., (1977), Telling more than we can know : verbal reports on mental processes, Psychological Review, vol. 84, 3, 231-259. Proust J., (1997), Comment l'esprit vient aux bêtes, Paris, Gallimard. Proust, J., (1998), A Plea for Mental Actions, Rapports et Documents du CREA.

Proust, J., (in press 1999), Awareness of Agency : Three Levels of Analysis in T. Metzinger (ed.), The Neural Correlates of Consciousness, Cambridge, MIT Press. Russell, J. , (1996), Agency, Its role in mental development. Hove, The Psychology Press.