Block's argument for the overflow of cognitive accessibility by phenomenology relies primarily on the partial-report results of Sperling (Reference Sperling1960) and others. I argue first that those results do not by themselves support that overflow. I then raise concerns about whether phenomenology could be altogether independent of cognitive access, as Block maintains.
Subjects in Sperling's experiment can identify only a few of the presented letters, but they report that they are all letters. Block convincingly argues that this limitation in identifying results from constraints imposed by visual working memory. He also urges that “phenomenology persists beyond the stimulus” (sect. 9, para. 4), since post-stimulus cuing enables subjects to focus on the subset of items they then identify. Because that cuing occurs after the stimulus has ceased, one reasonable hypothesis is that the identifying relies on persisting phenomenology.
Still, subjects identify all the items as alphanumeric independent of any cuing; as Block notes, subjects “can bring them [all] under a general concept like ‘alphanumeric character’” (sect. 9, para. 2). So all the items are at least partially conceptualized independent of cuing, presumably before the stimulus ceases.
So it may well be that, even before the stimulus ceases, the items are all conceptualized in respect of their specific alphanumeric identity. Identification of specific letters might then rely solely on that conceptual information, rather than on persisting phenomenology. The working-memory bottleneck would still be operative; conceptual representations of the specific identities of all the letters would be available prior to cuing, but only about four could get through to figure in subjects' identifications.
On this model, only persisting conceptual representations figure in whatever identifications the bottleneck permits. As with the Global Workspace model of Dehaene and Naccache (Reference Dehaene and Naccache2001), which Block considers, the limit on specific identifications results from a constraint on how much conceptual information can get through.
Perhaps phenomenology does persist, as Block urges. But some conceptual representations plainly persist as well, since subjects can identify all the items as alphanumeric long after phenomenology has clearly ceased. So if phenomenology does persist, it is accompanied by persisting conceptual representations, and those representations may by themselves be the basis of whatever specific identifications subjects can make.
The same holds for the related results of Landman et al. (Reference Landman, Spekreijse and Lamme2003). Block urges that these “[s]ubjects are apparently able to hold the visual experience for up to 1.5 seconds” (sect. 9, para. 7). But, as with the Sperling result, subjects' post-cuing ability to detect which rectangles have changed may be the result, not of persisting phenomenology, but rather, of persisting conceptual representations.
Block takes phenomenology to overflow cognitive access because subjects must have “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). But the working-memory bottleneck may instead operate by restricting the number of fine-grained conceptual representations that can figure in reporting and comparing.
The upshot of the partial-report results, then, would not be “phenomenological overflow” (sect. 9, para. 13), but rather overflow of the working-memory bottleneck by fine-grained conceptual representations. Since the mesh Block argues for is “between the psychological result that phenomenology overflows cognitive accessibility and the neurological result that perceptual representations that do not benefit from attention” (sect. 14, para. 10), that mesh is in jeopardy. So it cannot help resolve Block's puzzle about “whether the cognitive access underlying reportability is a constitutive condition of phenomenal consciousness” (sect. 2, para. 9).
In any case, there is reason to think phenomenology does always involve cognitive access. Block notes that “when one has a phenomenally conscious experience [i.e., an experience with phenomenology], one is in some way aware of having it” (sect. 6, para. 1). This reflects the widespread intuition that phenomenology is a matter of how our mental lives appear to us. There being something it is like to have an experience is just a matter of such mental appearance. Block denies that being thus aware of an experience itself involves some cognitive access to it, but his reasons for that rejection are unconvincing.
One way to explain why phenomenology always involves awareness of the experience is to posit a distinct higher-order state about the experience (Rosenthal Reference Rosenthal2005). Block notes that phenomenology occurs even when we do not notice, attend to, or think about it. But that does not rule out having inattentive, nonconscious thoughts about one's phenomenology.
Block also notes that young infants are often in pain, though they may be incapable of such higher-order states. But it is possible that infants' pains are not conscious pains. A state can have qualitative character even if there is nothing it is like for one to be in that state (Rosenthal Reference Rosenthal2005, Ch. 7); it is question-begging here simply to assume otherwise. And because nonconscious pains have largely the same causal connections to behavior and to a strong desire for pain to cease, they are just about as bad. So infants' pains are compatible with the hypothesis that higher-order cognitive states are required for pains to be conscious – for there to be something it is like for one to have them.
Block also argues that a higher-order theory of consciousness “fits both science and common sense” (sect. 6, para. 3) less well than same-order theories, on which the awareness of an experience is internal to the experience itself. But according to standard same-order theories (Brentano Reference Brentano1874/Reference Brentano1924; Kriegel Reference Kriegel2005), that awareness is every bit as cognitive as on the higher-order-thought hypothesis (Rosenthal Reference Rosenthal2005).
Block's concession that phenomenology always involves being aware of it also raises a difficulty he does not address. Phenomenology may occur, he holds, when there is activation in the fusiform face area of the extinction patient G.K., even though G.K. “not only does not know about, but … cannot know about” (sect. 14, para. 13) that phenomenology. But awareness of things standardly results in one's knowing about them and in being able to express that knowledge in verbal reports. If phenomenology always involves awareness of it, why would such awareness fail here to have those results?
Perhaps Block thinks that G.K. might in this case have only a weak phenomenology, like the nonconscious qualitative character that may occur in infants. But that is not ordinary phenomenology, which involves there being something it is like to have it.
Higher-order theories of consciousness and many same-order theories avoid this difficulty, since they hold that some cognitive awareness is constitutive of all phenomenology. By contrast, the global-workspace theories Block favors do not help explain why ordinary phenomenology always involves some awareness of it.
Block argues for the possibility of phenomenology wholly without cognitive access. But the awareness that always accompanies phenomenology and the compelling intuition that phenomenology is essentially a matter of mental appearance together suggest a close tie between phenomenology and some type of cognitive access.
