Sunday, 16 April 2017

Flynn’s Cat – Part 2: An Exploration of Embodiment

© 2017 Simon’s Cat Limited
< Flynn's Cat - Part 1 

So the cat is watching a cursor move about on my laptop screen. Is how the cat solves the problem of figuring out what the cursor is, embedded inside her somewhere sandwiched between perceiving and the movement of her paw, or is it developed and made possible through her interactions with the cursor? 
Even before this problem-solving task, how does the cat go about categorising the world she perceives? Is she constrained or limited by the body she inhabits?  I perceive her as black and white, but is that because in some way, I am physiologically equipped to perceive her as black and white.  The next door neighbour’s dog may perceive her differently. If I twist the can opener around a tin of cat food, does she perceive this, since she has no opposable thumbs to understand the concept of twisting, or does she merely see me move the can? Is there a pre-given world for either of us, or are both of us bringing it forth from our respective histories of structural coupling with artefacts in the world, which will be markedly different?
And if how she categorises differs from how I categorise then presumably any competences which depend on categorisation such as memory will differ between us as well.

Standard cognitive science is criticised for relying on amodal symbols and algorithms that cannot acquire the representational function that standard cognitive science assigns them. Meaning must come from elsewhere, for example, the characteristics of the body and the interactions with the world that the body affords. According to O’Regan & Noe (2001), vision is constituted by exploring, not by the mere activation of neural representations; Vision involves the interaction of the visual apparatus and its environment. If the cat’s legs move in a running motion while she is sleeping on her side, she is not running, even if her legs are making the correct movements, as her feet are not making contact with the ground. They also argue that perception depends on possession of sensorimotor knowledge, and it is the skilful activity of the cat as a whole that allows her to perceive. Her interactions with the cursor obey regularities known as sensorimotor contingencies and her visual experience is the exercise of those sensorimotor contingencies. In a similar manner, if I stroke her, my tactile experience is the knowledge of sensorimotor contingencies associated with tactile simulations. But what is the nature of this knowledge? It is a tacit knowledge, where experience of objects in my environment creates expectations in my nervous system.  Is it a type of know-how or a set of skills I have acquired that allow me to perceive – the set of movements and adjustments that gather information particular to the shape, colour, texture and other visual attributes of the cat. And what does it mean to exercise this sensorimotor knowledge? Is it my potential to perform these actions that allows me to perceive the cat?  Do I actually need to move my eyes over her to reveal sensorimotor knowledge, or does it mean that at some time in the past, I moved my eyes relative to a cat in a way that created knowledge of sensorimotor contingencies, and thus, my perceptual experience now, is the knowledge I acquired from those previous actions. This throws up a problem. If knowledge alone constitutes the perceptual experience, then the mind is again, akin to a brain in a vat, with the sensorimotor contingencies represented in the brain. While this theme does emphasise the importance of the body in perception, this is only weakly different from traditional cognitive science. If on the other hand, my eyes are firmly fixed on my laptop and I do not perform actions now, along with sensorimotor knowledge, will I fail to perceive the cat? This might certainly explain inattentional blindness.

It can certainly be argued that this form of knowledge is know-how rather than symbolical – it requires a brain in a body to acquire the skills, but can’t know-how be represented symbolically? Can’t the brain in a vat encode the know-how?  So, either I do not perceive the cat unless I physically exercise my sensorimotor contingency knowledge, or if I do not exercise this knowledge, my experiences of the cat rest on previously stored knowledge in my brain, which is admittedly functional in nature, but must at the same time, be represented in some way.  

So traditional cognitive science and this particular theme of embodiment directly compete with each other in their interpretation of data. While computational models of vision from traditional cognitive science assume an independent world, separate from the algorithms that derive information from it, embodiment depends on what my body brings to processing the world it encounters. But is there a better way to do cognition? This is the subject of the next post…

Flynn's Cat - Part 3 > 


O'Regan, J. and Noë, A. (2001). A sensorimotor account of vision and visual consciousness. Behavioral and Brain Sciences, 24(05), pp.939-973.

1 comment: