AISB Workshop on Sensorimotor Theory
The first AISB Members’ workshop was held at Goldsmiths, University of London on the 26th September, co-organised by Mark Bishop and Andrew Martin. Addressing Sensorimotor Theory and described as “A day of discussion on the Sensorimotor account of Perception, Consciousness and Robotics, its development and contemporary state”, and with over 60 delegates attending the event attracted interest across a wide range of disciplines.
Two themes emerged from the presentations, they were: the results of adopting a sensorimotor approach in various ongoing research projects and the comparisons of the philosophical profile of sensorimotor theory with that of (relatively traditional) cognitivist and (relatively dynamic) enactive theories.
J. Kevin O’Regan gave the inaugural keynote presentation, describing the latest developments in sensorimotor theory since his seminal 2001 paper “A sensorimotor account of vision and visual consciousness” co-authored with Alva Noë. O’Regan’s focussed on an account of “raw feel” which provides an account of how the phenomenological profile of conscious experiences differ and the means for their scientific investigation. In extending the theory to the necessary and sufficient criteria for conscious experience, O’Regan includes concepts of cognitive access to one’s activity and a notion of self. As presented by O’Regan, contemporary robotics projects can be described as having the necessary cognitive access resulting in lively discussion on the possibility of computational consciousness. Similarly, the requirement for a notion of self as described in O’Regan’s (fundamentally Dennettian) terms which implied that (for example) babies did not consciously experience pain, provoked further discussion amongst the attendees.
A significant development in O’Regan’s sensorimotor theory is its extension to address the character of conscious experiences so that they may be catergorised and compared. O’Regan described a few terms, namely (partial) insubordinateness, bodiliness, grabbiness. Respectively, they are the levels at which; stimuli may change without a subject’s acting, change as the subject’s body moves, and grab our attention. These terms allow comparison of feelings as diverse as touch (extremely bodily and grabby), thirst (reasonably grabby, but weakly bodily) and proprioception (weakly grabby but highly bodily). Feelings that rate highly on bodiliness and grabbiness are easier to describe as “real”, present senses such as sight, as compared to balance or hunger. Notably, in his 2011 publication O’Regan presented a fourth term in this set that was absent from his presentation.“Richness” describes the level of detail that may present itself to a subject’s experience. When questioned, O’Regan claimed that there are many ongoing sensorimotor couplings present, constituting the metabolism for example, that are arguably very rich, but do not have a correspondingly “real” feel.
The keynote presentation for the afternoon session, given by Daniel D. Hutto praised the move away from cognitivism evident in sensorimotor theory but identified philosophical criticisms that would apply until the theory cut all ties with its classical roots. The implications of the argument being that, in its current form sensorimotor theory will prove valuable in many applications, but in reforming relatively auxiliary aspects of the theory it could prove to have general explanatory power and applicability to consciousness. Hutto presented the theoretical foundations of Radically Enactive/Embodied Cognition as espoused by (amongst others) Varela, Thompson and Rosch. In this way, research into sensorimotor accounts of consciousness may shift emphasis away from a search for a suitable robotic implementation towards an attempt to understand the role sensorimotor interactions play in a larger system comprising the subject’s environment and society.
Outside of the keynote presentations were energetic sessions of presentations describing projects applying a sensorimotor approach. Interesting parallels were drawn between sensorimotor perception and the technique from the field of robotics known as active sensing; by mimicing the morphology and action of whiskered animals Sheffield University’s Active Touch Laboratory demonstrated a novel and successful technique for robotic perception. Chrystopher L. Nehaniv demonstrated the cutting edge of enactive robotics which showed how interactions between humans and socialised robots can result in emergent behaviours in both participants.
There was a general acceptance of the fundamental tenet that perception is grounded in embodied interaction in an environment. Of the criticisms, they only addressed sensorimotor accounts of the constituents of conscious experience, and were always accompanied by alternative formations of the theory that allowed these criticisms to be avoided. Ultimately the workshop showed there was a lot of interest in sensorimotor approaches to various aspects of cognition and a wide range of active research benefitting from the approach.