Block says he wants to sidestep the “explanatory gap,” but I'm afraid that it is too wide to walk around – you have to find a way to get across it.
Let us suppose we had a completely worked-out explanation of how information flows between the various chunks of brain discussed in the target article. Let's stipulate, with the author, that the explanations are roughly at the level of neural networks as described by Koch (Reference Koch2004). They will therefore be computational, in the sense that they would leave little room for distinctions between the actual brain and any device that passed equivalent coded messages between nodes standing for assemblies of neurons.
It seems likely that we will possess such explanations in the foreseeable future. At that point we will know how “access” in the author's sense will happen; that is, let us suppose, how information is distributed to a global workplace in Baars' (Reference Baars1997) framework. Will we also know how “phenomenology” happens? (I mean “phenomenology” in the sense Block uses it in sentences such as, “Sperling's clever idea was to test … whether the phenomenology persists after the stimulus was turned off by…”; target article, sect. 9, para. 2.) Well, no, we won't. The computational explanation of how information flows in order to enable subjects to report a row of Sperling's (Reference Sperling1960) array after hearing a tone will of course not refer to anything like phenomenology, but only to neural structures playing the role of buffers and the like. This is what Chalmers (Reference Chalmers1996) calls the Paradox of Phenomenal Judgment, which is just another angle on the Explanatory Gap.
Block would like it to be the case that phenomena play a role in our psychological abilities. To extract a few exemplary phrases from section 9:
[S]ubjects are able to deploy working memory so as to access only half the rectangles despite the fact that … subjects' reported phenomenology is of seeing all or almost all of the rectangles. … [Suppose] the subjects are continuing to maintain a visual representation of the whole array – as subjects say they are doing … (sect. 9, para. 6)
and
The subject has persisting experiences as of more specific shapes than can be brought under the concepts required to report or compare those specific shapes with others. (sect. 9, para. 10)
Notice how terms like “working memory,” “phenomenology,” “visual representation,” “experiences,” and “concepts” are used as though they all belong at roughly the same level in the explanatory hierarchy; but, in fact, in our envisioned computational explanation the “phenomenology” and “experiences” will drop out, as they seemingly always do.
Our intuition that experiences are somehow inspected in making phenomenal judgments is,Footnote 1 as I argue at greater length in McDermott (Reference McDermott2001), exactly backwards. The inspections are computational events that must precede the experiences, causally and temporally. If an experience is reported, or accessed in some weaker sense, then we can explain it by reference to how the brain thinks about itself.Footnote 2 But if an experience happens without being accessible, then it becomes a phantom.
Let me draw an analogy here. If we continue to suppose that cognitive science finds a satisfying computational explanation of visual processing and access consciousness, the unaccessed experiences will be in a position analogous to unobserved electrons in a two-slit experiment demonstrating interference of their wave functions. Some physicists, such as David Bohm (Bohm & Hiley Reference Bohm and Hiley1993), found it hard to believe that an electron could shed its particlehood for all except brief flashes of time, and devised theories in which electrons always remained particles, whose positions ingeniously served as indices to wave functions. Few in the mainstream physics community found this theory attractive, for some technical reasons, but also for an easily grasped one: The “positions” in Bohm's theory, being in principle completely unobservable, were obviously vestigial ornamentation on a theory that was otherwise an inelegant variant of ordinary (nonrelatavistic) quantum mechanics.
It seems clear that if we get a cognitive-scientific (i.e., a neuroscientific or a cognitive-psychological) explanation of access to experience, we are going to go through an intuition change at least as violent as that brought about by quantum physics. I think it is inevitable that our understanding of consciousness will change as radically as our understanding of, say, life, has since the seventeenth century. Will this be a new “definition” of consciousness? Block supposes (sect. 3) that we can always rephrase questions about phenomenology using the popular but at best vacuous formulation asking “what it is like” to have a particular experience. However, one has to ask to whom it might be “like something” for the fusiform face area of brain-lesion patient G.K. to light up when it is not like anything to G.K. himself?
I am actually in sympathy with the author's stance on methodology. I think the Occamish arguments he uses to justify his conclusions can be valid. I just don't think Block's methodological razor is a match for the scientific buzz saw that is cutting through this area of science.
