Thursday, January 04, 2007

 

Hypothesis: Consciousness is not an emergent property of sufficiently complex neural (or more generally, computational) circuitry.

Many scientists have made remarks to the effect that the most remarkable thing about the universe is that it is understandable; that it is subject to scientific examination and so perfectly describable by mathematics. It seems to me that there's a kind of epistemologically anthropic principle worth noting, with regard to this kind of statement: Some aspects of the universe happen to be understandable (by us; through a scientific approach), and it is precisely these aspects which constitute the whole of a scientist's perception of the universe. Whatever is not understandable is either simply unseen, or dismissed as not measurable, and thus not an appropriate thing for a scientist to ask questions about. Indeed, the universe might be packed infinitely densely with activity of an infinite variety of different kinds, and so long as none of them are measureable and understandable, we will be utterly blind to them so long as they do not disrupt what regular (measureable & understandable) activity we do see. From this view, the understandability of "the universe" is not a surprise at all; it is in fact inevitable as long as some subset of the universe - even if it is a relatively infinitesimal subset - is understandable.

A pragmatic scientist is perfectly correct to ignore the rest, and focus only on what is measurable; on scientifically testable hypotheses. It's the right way to get things done. However, there is no justification for taking the position that it is the only valid way to look at the world, or understand it. Consider this hypothesis: some real truths are directly knowable, without proof, evidence, or instruction. This doesn't seem so outlandish to most people; for the 90+% of the population with religious convictions it's unquestioned, and it even features prominently in the constitution: "We hold these truths to be self-evident:" Yet it is anathema to the cognitive scientist's view of the world: knowledge, from a materialist, computational perspective, consists of a cognitive state brought into a correct correspondence with the world. On the assumption that the world is fundamentally mechanistic, this kind of correct correspondence can only be reached via processes like deduction, induction, dictation, or (if you're permissive of some degree of nativism) evolution. So who's right? Well, there's no way to tell. It depends on the validity of the scientist's assumption of a mechanistic universe, which is untestable: no finite body of data on mechanistic, rule-abiding behavior actually constitutes evidence against the existence of unobservable behavior of other kinds, as discussed above.

So: Consciousness.

The conscious experience of something like the color yellow is not "the same" as even a theoretically plausible complete description of the neural process of perception, from the retina on as far as a signal could be traced. There is a certain relationship, to be sure: if I had an appropriate instrument, I have no doubt I could measure a characteristic response somewhere in my CNS that would correlate strongly with my perception of yellow. However, I also have no doubt that, with an appropriate instrument, I could measure some characteristic of a rock that would correlate strongly with yellow light being shone on it. So a certain stimulus (yellow light) seems to produce both a characteristic CNS response and a characteristic quale, in me; it also seems to produce a characteristic response in a rock. I have access to my own qualia, but not to those that the rock might hypothetically possess. If it is sensible to ascribe a causal role to my characteristic CNS response in the "generation" of my quale, what does that say about the rock's hypothetical quale? It doesn't say anything definitive, to be sure, because as mentioned, I have no way of obtaining any data about any conscious experience the rock might be having. But this lack of data also means there's no justification for arguing that the difference between the (arguably complex) response produced in me and the (arguably simple) response produced in the rock corresponds to any difference in conscious experience. The data simply aren't there, and can't be obtained.

So what is a rock aware of?

It is possible, of course, to collect data on what types of signals - and what types of responses - correlate with our own consciously experienced qualia, or those communicated by others. But I would argue that what we've found even here doesn't suggest a requisite level of complexity. Brevity of a stimulus, or the presence of distractors, can prevent conscious experience; more dramatically, so can (e.g.) cutting the spinal cord. But very simple stimuli & responses - like pain signals, or white light - can induce conscious experience.

posted by Miles 11:15 AM