Block's original distinction between phenomenal consciousness and access consciousness has provoked a lot of critical comment in the last decade or so, my own included. I see two distinct sources of uneasiness with the distinction between phenomenal and access consciousness, and while these two sources have not usually been distinguished, some of Block's remarks in the present target article seem to acknowledge the need to do so. On the one hand, coming principally from a functionalist perspective, many philosophers and psychologists believe that there has to be a constitutive connection between what we are conscious of and what we can report. The idea is that consciousness is somehow reducible to this kind of access, and only if we can so reduce consciousness will it be amenable to scientific investigation. It is this view that is Block's target here.
On the other hand, some, like myself, have been uneasy with Block's distinction because the idea of phenomenal consciousness totally divorced from any access by the subject does not really seem like any kind of consciousness at all. As Block notes here, we have complained that the very phrase that serves to canonically express the notion of the phenomenal – “what it's like for x to …” – explicitly refers to the phenomenal state in question being “for” the subject. The way I would put it now is: Phenomenal states/properties are not merely instantiated in the subject, but are experienced by the subject. Experience is more than mere instantiation, and part of what that “more” involves is some kind of access.
So one possibility, suggested by Block's discussion here, is that for a state to be phenomenally conscious it must be accessible to the subject in some sense, but not necessarily to the mechanisms responsible for report. Block supports the view by appeal to the Sperling and the Landman et al. experiments, together with the associated neurological data, which, he argues, show that phenomenal consciousness overflows the “working space” that is report-accessible. Is this appeal persuasive?
One thing to notice right away is that Block's own argument actually relies on building some notion of access into the notion of what is phenomenally conscious, which reinforces the argument mentioned above. After all, in the Sperling experiment, the evidence that is supposed to show that we are phenomenally conscious of more than we can report is that there is information concerning the identity of the letters in the array that we are aware of, though we lose it when reporting mechanisms are engaged. The evidence that it was there in phenomenal consciousness is primarily that we say we saw all the letters. My point is not about the saying, but about the fact that what we report is that we did have a kind of access to this information; and it is because the information is available phenomenally that, when prompted appropriately, we can report a portion of it.
But now one might object that we can't really tell from the experimental data that the identities of the letters were phenomenally conscious. One might claim that, instead, one was phenomenally aware of more generic information, and that the specific information manifested in the partial report was stored unconsciously. Block explicitly addresses the objection that perhaps the information persists retinally only, but one might claim, in response to his demonstration that it goes higher up, that it is still a matter of unconscious, or sub-personal storage. How can one ever rule that out, after all?
Of course, Block is not claiming to have a knock-down refutation of his opponents' interpretation of the data, just a more plausible version. I find myself largely sympathetic to his position, and so wonder about the source of the resistance to it. Why could it not be pretty much as he says? It seems to me that fueling this resistance is a nagging sense that access just has to be constitutive of conscious experience, that we just cannot understand what it would be to be conscious without it. But given the concession already made that some access is involved in phenomenal consciousness, why not go along with Block's interpretation of the Sperling and other data? Why insist on report-access in particular?
I cannot speak for Block's opponents, but I suspect that reasoning along the following lines underlies a good part of the resistance to his position- First, there is the relatively conceptual point that consciousness is constitutively related to subjective access – my aforementioned point about experience being more than mere instantiation. Once the notion of the subject – the “for whom” it is like what it's like – enters the picture, we then ask who or what this subject is. Well, the natural thought is that the subject is the person, and the person is the entity that both plays the highest executive role in deliberation and planning action, and reports to others (and to itself). Hence, the idea of a subjective access that is divorced from reporting mechanisms seems hard to swallow, as there doesn't seem to be a suitable candidate for the subject whose access is involved in the bit of phenomenal consciousness at issue.
If I am right about the source of the resistance, then I think there are two lines of reply. First, it might be that subjective experience does not entail the existence of a subject, at least not anything like what we normally take a subject to be. While what is consciously attended to might involve access by the sort of high-level executive we associate with the personal level, and we naturally think of as the subject, the person, it seems plausible that there are centers of experience more loosely connected in the mind and that do not involve an entity we would call a full-fledged subject. I take Block's argument here to show that this is plausible, given the data so far. The argument can be contested, of course, but I wonder what, besides a commitment to the conceptual binding of experience to a substantial subject, motivates the search for these alternative – and to my ear – strained, interpretations of the data.
Finally, one might retain the idea of a substantial subject as a necessary component of experience, and even retain the idea that the subject is that which reports. As Block noted in discussing Dehaene and Naccache's division of the global workspace into I2 and I3, the only dispute here is whether what's phenomenally conscious has to be in I3. Block isn't denying that it must at least be in I2. But if I2 is understood as that pool of information from which I3 draws, then what counts as the subject can be identified with the union of I2 and I3. The point is that whatever bottleneck exists due to the processing that gets an item from I2 into I3 shouldn't be taken to restrict what we count as the full-fledged subject of experience. Of course it might have been that way, and in the end it is an empirical question, as Block insists. But why think, as so many who insist on the constitutive connection between consciousness and reportability seem to, that it just has to be that way?
