(This is another post in a series, starting with this one, on the so-called "explanatory gap" in any current, or perhaps any possible, theory of consciousness and the reality of conscious experience.)
"Why is the performance of these functions [that are 'in the vicinity of experience'] accompanied by experience?" Chalmers asks, in the paper that re-introduced the idea of an "explanatory gap" in all attempts to construct an explanation of consciousness. A little later he puts the same question a bit differently: "Why doesn't all this information-processing go on 'in the dark', free of any inner feel?" It was, presumably, his inability to find an answer to such questions that lay behind his use of the "zombie" thought-experiment to argue against a materialist, and in favor of a dualist, approach to comprehending consciousness as a phenomenon. My argument here, however, is that he gave up too quickly.
In asking why there is experience, as in an "inner feel", Chalmers appears to be asking for a functional reason for experience, not necessarily for a mechanism (which may or may not be the case, but let's assume so). With that understanding, then, we can compare a conscious response to a stimulus with another type of response which really does go on "in the dark", as Chalmers puts it -- and then ask what additional functionality does consciousness or feeling supply? That other type of response is the reflex: if you accidentally touch a hot surface, for example, your hand moves away before you're able to feel anything -- a very simple instance of information-processing "in the dark", as it were. Yet, slightly later, you still do feel the pain as well -- why? Because the feeling is an essential component of a more sophisticated kind of response-generator or behavior control. It is the bearer of necessary information, about the location, type, and severity of the injury, for example, but it also provides that information in a significantly passive form -- i.e., simply as feeling, not as a direct connection to a response -- that is the key to the functional adaptability of consciousness as a control mechanism. Of course, any sort of pain is a form of feeling that, unlike sight, say, or hearing, has a built-in motivational component of varying strength, but the point of the feeling is precisely that, while it may motivate a response, it doesn't direct one, and so the motivation may be over-ridden under dire enough circumstances -- conscious experience, in other words, is a vital component of a behavior control mechanism of astonishing flexibility, without which we would be "in the dark" in a more than just literal sense.
Which is the problem with Chalmers' hypothetical zombies -- without the "light" of experience, such entities lack the form of information that provides consciousness with the free play needed for its flexibility. Now Chalmers, of course, starts out by viewing experience as something inherently different from a mechanism of any sort, and so will always return to his insistence that, given any mechanism, even one of an allegedly "conscious" kind, one can always view it's processes as "dark", or without feeling, which therefore always makes the feeling (for him) appear as an addition to the mechanical, causal processes, or as an "epiphenomenon". In fact, Chalmers must insist not just that you can view any possible mechanical process as dark, but that you must so view it, since feeling and mechanism are fundamentally distinct.
And this, despite everything I've said to this point, might be considered to be half right -- it accurately reports one's intuition from one of two possible perspectives on a phenomenon. Consider, for example, a researcher studying the difference between reflex and conscious response in another organism, and who doesn't, obviously, have direct observational access to the experience itself, but must rely on proxies of one sort or another (e.g. verbal report, other behavioral signs, neural activity, etc.) -- from her perspective, even if she could trace every single neural signal involved in the two different processes, and even though one might be more complex than the other, both would be as apparently "dark", since no trace of "feeling" would ever be observed. As soon as the same researcher studies her own reactions to the same stimulus, on the other hand, it's immediately evident that, while the reflex is as dark as before, the conscious response is inextricably connected with feeling -- indeed, "feeling" is the very meaning of such a response. The difference between the two cases is solely one of situation or perspective -- in the first case she was external to the phenomenon; in the second, "she" was a part of the phenomenon. What Chalmers does, to generate the intuition of an "explanatory gap", is to superimpose the two perspectives, in effect, which produces a rather odd and puzzling sort of double vision, certainly, but which has nothing in itself to do with an explanatory deficiency.
UPDATE: Here's the conclusion to this series [see "The 'explanatory gap' series: a summary and a Q and A" above]..