Department of Psychology
University of London
KEYWORDS: Dennett, Kinsbourne, cartesian theatre, binding, consciousness, integration
In the visual system, the represented features of individual objects (shape, colour, movement, and so on) are distributed both in space and time within the brain. Representations of inner and outer event sequences arrive through different sense organs at different times, and are likewise distributed. Objects are nevertheless perceived as integrated wholes - and event sequences are experienced to form a coherent "consciousness stream." In their thoughtful article, Dennett & Kinsbourne ask how this is achieved.
According to Descartes (1644), integration requires a single interface between conscious experience and the brain, provided by the pineal gland. It's central location is well suited to transform volitions originating in the soul into motor movements; and it provides a suitable point of convergence where disparate retinal images can be combined into a single, integrated visual field. In current theorizing the status of the pineal is less exalted, but it is still widely assumed that information arriving at the various sense organs is, somehow, integrated (with stored knowledge, needs, intentions, etc.) not only to allow co-ordinated, adaptive response, but also to provide an integrated conscious stream (e.g. Baars 1989; Dixon 1981; Marcel 1983; Navon 1991; van Gulick 1991; Velmans 1991b). A few theorists follow Descartes, in suggesting a central co-ordinating system (for example, Dimond 1980; Penfield 1975); others allow that the integrated neural correlates of consciousness may be widely distributed throughout the brain (e.g., Koch & Crick 1991; John 1976; Pribram 1971, 1986).
According to Dennett & Kinsbourne, these models subscribe to a form of "Cartesian Theatre" in that they assume human information processing relating to any given event to present a final, integrated representation (of that event) to the "footlights" of consciousness. Other than their commitment to cerebral integration, however, there is very little that is Cartesian about these views. No current theory adopts Descartes' neurophysiology, or account of processing. So it might be more accurate to call this the "integrationist" position.
In Dennett & Kinsbourne's "Multiple Drafts" model, there is no integrated, "definitive stream of consciousness" in which information about the world all comes together; there is only a "parallel stream of conflicting and continuously revised contents." They claim that this avoids the "scientific and metaphysical extravagances" of the Cartesian Theatre, which assumes (1), that there is a place in the brain where information from all relevant inputs is presented for a final subjective judgement and (2), that the temporal ordering of experienced events reflects the temporal ordering in which representations of those events arrive in the theatre. By contrast, representations in the "Multiple Drafts" model are (1) distributed both in space and time in the brain, and (2) "a product of the brain's interpretational processes, not a direct reflection of events making up those processes" (Abstract). This characterization of current integrationist views, however, is very misleading.
As noted above, only a few supporters of cerebral integration propose that this occurs in a specific place within the brain, and even fewer would assume that integration requires prior subjective judgement (one exception might be Eccles, 1980). Indeed, with respect to processing there is very little in Dennett & Kinsbourne's description of the Multiple Drafts model, which differentiates it from an integrationist account. Few integrationists would deny that representations are distributed both in space and time within the brain, or take issue with the view that information encoded within neural representational states may enter into subsequent mental functioning, rather than the physical properties of such states (cf Uttal 1978; Velmans 1991b, note 30). Dennett & Kinsbourne accept the integrationist assumption that in order to function with distributed representations, "the brain must be able to 'bind' or 'correlate' and 'compare' various separately discriminated contents" (sect. 1.1, para 6). To enable integration, interpretational processes must combine information about inputs (contained within distributed representations) with stored contextual information in "real time" (cf Blumenthal 1977). Indeed, few integrationists would deny the existence of "multiple drafts." The stream of consciousness might be integrated, but it is in continuous flux and change. It is difficult to envisage how dynamic change could be represented without constant redrafting of representational states. Even in the Cartesian theatre the show must go on.
Consequently, the only difference between the models that "makes a difference" is whether or not an integrated conscious stream, with its corresponding integrated neural correlates exists. This is the core of what Dennett & Kinsbourne deny.
Faced with the obvious - that we normally experience just one world, not multiple worlds, and that we normally do so without conflict about what we perceive, they plead massive cortical self-deception! In the fraction of a second between multiple, conflicting "drafts of experience" and subjective reports, the "Orwellian" brain continuously rewrites history (on the basis of the latest available information) to produce a consistent "party line." Hence, we report having integrated experiences, although there are no prior integrated experiences to report. According to Dennett & Kinsbourne, this scenario is more plausible than the "extravagant" integrationist alternative - a "Stalinesque" doctoring of the conflicting evidence prior to its presentation before the "show trial" of conscious experience.
The integrationist model, however, requires neither doctoring of the evidence, nor adherence to any given party line. On the contrary, it usually assumes the brain to be making the most accurate mental model it can of what is going on, based on the best available information, stored knowledge, etc. at the time of integration. As Dennett & Kinsbourne ably demonstrate, relevant information may arrive at different times, even that relating to temporal sequencing. Provided that the information does not arrive too late, information about temporal sequence (rather than the arrival time of the information itself) is integrated into the way temporal sequence is experienced (thus accounting for Libet's "backward referral" findings). Colour phi, and the cutaneous rabbit, provide persuasive demonstrations of the way cerebral processes integrate what is often partial information, into a coherent, integrated experience.
Contrary to Dennett & Kinsbourne's claim, this integrationist model is theoretically less "extravagant" than their "Multiple Drafts" account. It assumes that within any given time window, information is integrated into a single conscious stream, and that unless there is evidence to the contrary, subjects' reports about what they experience are accurate. Dennett & Kinsbourne require us to believe that subjective reports are not reports of prior experience; for example, subject's claims to have experienced colour phi prior to reporting it are based, they suggest, on rapid forgetting of parallel, conflicting contents. It is what subjects believe at the time of report that defines "what the subject was conscious of" (sect. 2.2, para 36).
This is theoretically self-defeating, for the reason that, at the time of report, subjects believe that they are describing a prior, integrated experience. A theoretical account that denies the accuracy of this belief is in danger of being unfalsifiable. Unless the conditions under which rapid forgetting occurs are precisely stated (cf Velmans 1991, note 6), one can invoke it at will to deny the legitimacy of any theoretically inconvenient reports. In short, Dennett & Kinsbourne can be accused of doctoring the evidence. It is their argument, rather than the integrationist model, which is Stalinesque.
Baars, B.J.(1989) A Cognitive Theory of Consciousness. Cambridge University Press.
Blumenthal, A.L.(1977) The Process of Cognition. Prentice-Hall.
Descartes, R.(1644) Treatise on Man. Trans. by T.S.Hall. Harvard University Press, 1972.
Dimond, S.J.(1980) Neuropsychology: A textbook of systems and psychological functions of the human brain. Butterworths.
Dixon, N.F.(1981) Preconscious Processing. Wiley.
Eccles, J.C.(1980) The Human Psyche. Springer.
John, E.R.(1976) A model of consciousness. In Consciousness and Self-Regulation, ed. G.Schwartz & D. Shapiro. Plenum, pp 6-50.
Koch, C. & Crick, F.(1991) Understanding awareness at the neuronal level. Behavioral and Brain Sciences 14(4):683-685.
Marcel, A.(1983) Conscious and unconscious perception: an approach to the relation between phenomenal experience and perceptual processes. Cognitive Psychology 15:238-300.
Navon, D.(1991) The function of consciousness or of the information processes have on each other? Behavioral and Brain Sciences 14(4): 690-691.
Penfield, W.(1975) The Mystery of the Mind: a critical study of consciousness and the human brain. Princeton University Press.
Pribram, K.H.(1971) Languages of the Brain: experimental paradoxes and principles in neuropsychology. Englewood Cliffs, N.J: Prentice-Hall. 2nd ed. Monterey: Brooks/Cole.
Pribram, K.H.(1986) The cognitive revolution and mind/brain issues. American Psychologist 41:507-520.
Uttal, W.R.(1978) The Psychobiology of Mind. Lawrence Erlbaum.
van Gulick, R.(1991) Consciousness may still have a processing role to play. Behavioral and Brain SCiences 14(4): 699-700.
Velmans, M.(1991b) Consciousness from a first-person perspective. Behavioral and Brain Sciences 14(4): 702-726.