Here is a well-informed review of a book we should cite in this thread: Mark Rowlands,
The New Science of Mind.
The review:
An Extensive Enterprise
By
Stephen E. Robbins on August 14, 2013
4 of 4 people found this review helpful
This book [The New Science of the Mind], I must note right away, is truly for the philosophical elite and for the philosophical psychology folks, with the scales tipping heavily to the philosophical side. The work is an extensive and intensive effort to examine notions of embedded mind, embodied mind, enactive mind, extended mind, while arriving at the author's particular view of extended mind. A key component of this framework is "action on external structures [ e.g., the environment] that transforms the information they contain from merely present to available... " Characterizing positions nicely and lucidly, he arrives at a position he terms "amalgamated," combining the embodied and the extended views. In this effort he reviews the visual theories and work of Marrs, O'Regan and Noë, Vygotsky, and Gibson, each discussion clear and well done. Gibson's concept of the "optic array," Rowlands notes, qualifies as the "external structure" of information par excellence, while his emphasis on the organism's action to transform this structure, thus isolating invariants and making this information available also fits perfectly.
The extended mind thesis that Rowlands defends is that at least some mental processes extend into the organism's environment in that they are composed of actions performed on the world around the organism. This therefore extends the notion of cognition beyond the brain, in that the whole body is (or can be) involved. In this general conception, Rowlands reserves a role for the standard notion of representations current in cognitive science, and in this he differs explicitly from the view that Gibson denied such a construct in his view of the brain. In all this, he emphasizes that this is to be viewed as process, rather than cognitive "states." It is obvious that this thesis is not the version of extended mind, e.g., a panpsychic-like framework, that some might be looking for, in fact, from this perspective, the thesis seems very conservative. Yes, it is different from the pure centered-on-the-brain notion that is attributed to the Cartesian view, but it still almost becomes a question of just how could this be really that controversial? Further, it is very unclear just why this view is not already implemented in and furtherable in, robotic/connectionist architecture, in fact, in studying Rowlands' (very interesting) discussion of certain elements of connectionist architecture and the subsumption architectures of situated robotics, it seems that he views this to be the case. In fact, it was not clear to me, given the book, why such robots are not perfectly conscious?
The extended mind thesis that Rowlands might have arrived at would indeed be radical had he not ignored several major questions. The first is the problem of qualia, only obliquely approached in his examination of Nagel, but not come to grips with in terms of Chalmers. This problem is actually more general, for avoiding Chalmers' misleading emphasis on qualia, it is, at least in vision/audition, the problem of the origin of the image of the external world. How does this image arise, or get generated, or in general is it accounted for? If you are still holding, as Rowlands, that the brain employs representations, then you have the "coding" problem that the concept of representations involves. Three dots (...) can stand for an "S" in Morse code, or three tomatoes, or a stove. One must know the domain to which to map the code, and the brain would then, in effect, have to already know what the external world (the domain) looks like to use its "representational" code (for the neural-chemical code cannot look anything like the external world). If to solve this, you are relying surreptitiously on the external structure (or field) , say, the optic array, then at what scale of time are you conceiving this structure/array/field to exist - the scale of a "buzzing" fly in the field, a fly slowly flapping his wings like a heron, a fly as a vibrating ensemble of atoms, or what? If the external field is indeed holographic (as one could also construe the optic array) - a massive interference pattern - then it is un-imageable - at any scale of time. You cannot just assume the environment as a nice "external structure" that the brain uses - as though somehow the image of the external world as we know it already exists! Yet this is what is going on in Rowlands, or in O'Regan and Noë for that matter. Further, the external structure is not static, it is dynamically transforming with buzzing flies, falling leaves, rotating cubes. One does not need action from the body to "make the structure change." But now you open another set of problems, for now you need a theory of the basic "memory" that allows the brain to specify (or us to perceive) these transforming events - as extended events over time. It is a problem hidden currently under the notion of "temporal consciousness" that is equally a problem of qualia, in fact has greater primacy, for all perceived qualia extend over time.
This is where Rowlands goes wrong in holding that Gibson is compatible with representational models. Gibson was well aware of the problem of the origin of the external image of the world, and further, the difficulty with invariants defined only over time, for example form as defined only over flow fields. No static representations can capture this - no symbolic manipulation or connectionist architecture as I think Rowlands fails to see. An invariant defined over time cannot exist as a "bit" traveling along the nerves. This is why Gibson went to the notion that the brain is "resonating" (a time-extended process) to these invariants and in this resonance is "specific to" the environment. This "specific to" has been the problem, for how does this "specific to" actually explain the origin of the image of the external environment and at a certain scale of time? This brings us to the greatest philosopher of the extended mind, apparently unknown, despite Rowlands' attention to Heidegger (with another nice exposition), Husserl, Sartre, and others, and this is Bergson. Bergson, I have argued, presciently (and uniquely) using the essence of holography, fills in the missing piece of Gibson, but Bergson's mode of mind is indeed extended, for the relation and the difference between subject and object is not in terms of space, but of time. (See for example [
Collapsing the Singularity: Bergson, Gibson and the Mythologies of Artificial Intelligence]) In this the brain is seen as an entirely different form of device, with a critical dynamics that is far from achievable by the connectionist or robotic or computer model. The neglect of Bergson in this embedded, embodied, enactive, extended literature - even if only to consider and reject his vision for hopefully solid reasons - is indeed unfortunate.
I could add more analytical objections for I see problems in the conception of cognition, the role of consciousness in cognition, the nature of the problem of memory (one can have no theory of the "storage" of experience when the problem of perception - the very origin of experience - is unresolved). However, the book, as I noted, is an extensive and intensive piece of thought, attacking effectively the Cartesian framework and making a case for a wider view of mind. Rowlands, obviously in tune with the state of various minds in his field, sees this as necessary, noting that his amalgamated (embodied/extended) vision is yet seen in these quarters as "outlandish." As such, the case needs to be made; there is just, imo, far more room to extend."
There are other helpful reviews at this amazon.com link:
http://www.amazon.com/dp/0262014556/?tag=rockoids-20
The beginning pages of Rowland's book at the amazon link summarize clearly the deeper issues concerning mind (and thus consciousness) to which recent neuroscience, studies of cognition, philosophy of mind, and phenomenology have brought us.