Machine Consciousness: An enactive analysis of the state of the art
Toward a Science of Consciousness
23-26 July 2007, Budapest, Hungary
In the last few decades the field of cognitive science has undergone a significant shift from pure cognitivism towards a more embodied-embedded understanding of mind and cognition (Wheeler 2005). This change has been strongly influenced by practical dissatisfaction with the problems faced by cognitivist GOFAI in real-world contexts (Dreyfus & Dreyfus 1988), as well as the development of a new style of AI and robotics in which intelligent behavior is seen as emerging from the dynamics of a brain-bodyenvironment systemic whole (e.g. Beer 1995). More recently, a further shift is gaining momentum in cognitive science, namely towards an understanding of cognition as “enaction” (Varela, Thompson & Rosch 1991). This approach builds on the insights of the embodied-embedded perspective, but is additionally driven by a central concern with subjectivity in its dual aspects of biological autonomy and the phenomenology of lived body experience (Thompson 2005). It is with respect to this more recent shift that the role of AI in cognitive science needs to be reevaluated, in particular because most work in machine consciousness continues to ignore the enactive account of subjectivity, as evidenced by the sample of work that appeared in a recent JCS volume (Holland 2003).
The aim of this talk is therefore to propose ways in which the field can move beyond its current preoccupation with engineering the emergence of sensorimotor loops, and thereby maintain a mutually informative dialogue
with the enactive paradigm. In this regard it is helpful to consider that the enactive viewpoint distinguishes three interrelated “dimensions of embodiment”, each relating to the subjectivity of a conscious agent in a distinctive way: 1) organismic regulation, 2) sensorimotor coupling, and 3) intersubjective interaction (Thompson & Varela 2001). Most of the recent work on embodied-embedded AI and robotics has mainly been concerned with 2), but some significant advances have already been
made with 3), namely in terms of social interaction (e.g. Di Paolo 2000). In the near future 3) is probably the most promising area for more investigation, and indeed, if we accept the claim that the conscious observer arises in the relational domain of language (Maturana 2006), then this kind of research could even provide us with another potential path toward the development of machine consciousness. Nevertheless, the principles of biological autonomy underlying the subjectivity which enactivism associates with 1) are still practically unexplored by AI research and remain a significant challenge to the current methodologies (Di Paolo 2003).
In order to clarify this situation the talk will outline a continuum of increasing organizational requirements which, according to the enactive viewpoint, culminates in the constitution of a subjective self. It will be argued that a more concerted attempt at
expanding the field of AI and robotics along this direction will not only make sure that its findings will continue to be relevant to the development of an enactive cognitive science, but also that, following Ziemke (2007), it has the potential to provide current research in machine consciousness with a much needed operational foundation of subjectivity.
Beer, R.D. (1995), “A dynamical systems perspective on agent-environment interaction”, Artificial Intelligence, 72(1-2), pp. 173-215
Di Paolo, E. A. (2000), `Behavioral coordination, structural congruence and entrainment in a simulation of acoustically coupled agents’, Adaptive Behavior, 8(1), pp. 25-46
Di Paolo, E.A. (2003), “Organismically inspired robotics: homeostatic adaptation and teleology beyond the closed sensorimotor loop”, in: K. Murase & T. Asakura (eds.), Dynamical Systems Approach to Embodiment and Sociality, Adelaide, Australia: Advanced Knowledge International, pp. 19-42
Dreyfus, H.L. & Dreyfus, S.E. (1988), “Making a mind versus modelling the brain: artificial intelligence back at a branch-point”, Daedalus, 117(1), p. 15-44
Holland, O. (2003), “Editorial Introduction”, Special Double Issue: Machine Consciousness, Journal of Consciousness Studies, 10(4-5), pp. 1-6
Maturana, H.R. (2006), “Self-consciousness: How? When? Where?”, Constructivist Foundations, 1(3), pp. 91-102
Thompson, E. (2005), “Sensorimotor subjectivity and the enactive approach to experience”, Phenomenology and the Cognitive Sciences, 4(4), pp. 407-427
Thompson, E. & Varela, F.J. (2001), “Radical embodiment: neural dynamics and consciousness”, Trends in Cognitive Sciences, 5(10), pp. 418-425
Varela, F.J, Thompson, E. & Rosch, E. (1991), The Embodied Mind: Cognitive Science and Human Experience, Cambridge, MA: The MIT Press
Wheeler, M. (2005), Reconstructing the Cognitive World: The Next Step, Cambridge, MA: The MIT Press
Ziemke, T. (2007), “What’s life got to do with it?”, in: A. Chella & R. Manzotti (eds.), Artificial Consciousness, Exeter, UK: Imprint Academic, pp. 48-66