Inductive Skepticism and the Methodological Argument

I want to give a short response to the two excellent and extremely helpful commentaries.

First, Jennifer Corns pointed out some subtleties about Mole’s arguments that I missed in the original paper. Most importantly, that Mole is not intending to make an argument of the inductive skeptic’s form, as primarily evinced by his use of common sense reasoning at the start of his paper. I found Mole’s reasoning on this point unconvincing because of both method and content, but may have been incorrect to leave these thoughts out of the paper, given my strong claims. Mole supports his claim that the commonsense view of attention is that attention is not necessary for consciousness but consciousness is necessary for attention by asking 19 women to describe the case of the baby’s cry waking its mother: is she “conscious of the cry” and/or “paying attention to the cry”, Mole asks. Unfortunately, as “paying attention” requires the subject to direct their attention, this form of the question begs the question in favor of the subjects saying that one must be conscious to attend. A neutral phrasing would have been, “Is the mother attending to the baby’s cry?” I think, in this case, that the responses would have been evenly split. My reasoning for this is that I find the language of attention to be neutral to whether we are always attending or not.

Some further notes about Jennifer’s commentary:

1. Although I do not want to blame this reading on Jennifer, as it may well have been my writing that misled her, I do not think that all objections to ANC must take the form of logical inductive skepticism. In fact, I try to push this type of objection aside so that I can look at what I consider the more serious objections.

2. I do not want to say that Ned Block holds Mole’s view. Block has not written about attention, but access, and the former is not limited to the latter. That is, one can attend to something without having access to it, so I think that there is no reason for him to claim that phenomenal conscious states cannot involve attention (contra Mole).

Second, John is quite right to point out that I am not clear enough about my view of attention. My view is that attention does contribute to working memory, but that it does not give working memory its distinct character. That is, attention is necessary for access, but only because it is also necessary for conscious perception (and thus access to conscious attention). The character of having epistemic access to one’s conscious perceptions does not come from attention, but from something else. Perhaps whatever is special about access is also special about one type of attention, namely, active attention. However, attention comes in both active and passive types.

Finally, I very much like John’s suggestion about testing some of these issues. I think that the Sperling test was answered in this way already by Kouider when he showed that subjects do not notice when a letter is degraded, but perhaps a test that does not rely on the subject noticing anything would be better. I will have to think more about these possibilities.

In close, I am very grateful to the time of my commentators, as they have helped me to see a better direction for the type of response I make to the Methodological Objection.

I’d like to make a few follow-up comments in response to Carolyn’s post. Many thanks to her for a rich discussion both here in this public forum and in the emails we’ve been exchanging privately. Some of what I say here will draw from that correspondence. I recognize and warn that this means my comments are somewhat lengthy.

First of all, my apologies for what seems to have been a misinterpretation of Carolyn’s view on two main points. First, on whether she thinks all rejections of ANC can be construed as skeptical and second, on exactly what claim she is targeting with ‘ANC’.

The pages of Carolyn’s argument that I referred to in my comments and took to support an interpretation of her as making the strong claim that all rejections of ANC are skeptical still seems to me to be the best interpretation of her comments there. She says:”Objections to the claim that attention is necessary for conscious perception (ANC) use inductive skepticism to reject the projection of the property of requiring attention to certain tokens of conscious perception,” (3) and as there was no qualifier there, I assumed she meant all objections (perhaps there’s a scope amiguity here, but I didn’t think so as I take all necessity claims to involve projection to all tokens of a type). In her next few sentences of that paragraph she breaks down the objections into two kinds: both forms of indcutive skepticism and then concludes: “Thus I divide the objections against ANC into two types….” (4) I took that as definitive. I have, however, been persuaded in conversation with Carloyn and by perusing her other work, that this interpretation is misplaced. This does leave me to wonder, however, just how pervasive she thinks skeptical rejection of ANC actually is. As I’ve argued, I don’t think Mole is guilty of it and, as Carolyn has agreed, neither is Block.

On the second interpretive point, as I say in my comments, I had taken ANC to be a claim about all conscious states, but in conversation it’s become clear that this is not what Carloyn had intended. Rather, ANC, as she was defending it, is a claim only about conscious perceptual states. I think that this sort of conservatism is perfectly well placed and have no objections to the weakening of the claim. Moreover, I don’t think anything substantive in what’s been at issue between Carloyn and I turns on it. As a methodological point, however, I think excessive carefulness with the terms one employs in these discussions is particularly important when there are still live issues about what sort of states can be conscious, what causes and constitutes consciousness, and whether some states always or never occur consciously. Until we all start working with a common vocabulary (perhaps a pipe-dream), one cannot be too clear or careful.

This last point leads me to Carolyn’s earlier points and recent question about Block’s view and it’s relation to ANC and Mole.

In her most recent comment, Carolyn asks for reasons for thinking some states could be phenomenally conscious without being access conscious. I myself had intended to stay neutral on the question (largely because I’m rather unpersuaded that we need to posit two kinds of consciousness at all and arguing that would take us far afield), but I think that if one is going to engage with Block (either as such or as it relates to Mole), then one has to wrestle with the reasons that Block gives. My point about Block was an interpretive and methodological one.

I hadn’t intended to imply that I thought Carolyn had directly against Block as she did against Mole. What I *do* think, however, is that if Block’s mesh argument goes through, it undermines Carolyn’s argument against Mole instead of, as she seems to take it, supporting that argument.

This leads me to Carolyn and my disagreements about whether Block actually *is* committed to a rejection of ANC. I continue to think that he is, as I continue to understand him to think that attention is necessary for access consciousness, but *not* for phenomenal consciousness. In her first comment, Carolyn says that access is not limited to attention, and if by this she means that attention is not necessary for access consciousness then, while that may be true, I do not think it is Block’s view. I have conceded in discussion with Carolyn that I don’t think Block is very clear in his the mesh article about this, though I think section 7 and his discussion of Prinz’s AIR theory points in that direction. He has, however, confirmed with me in conversation that he does think attention is necessary for access consciousness, but not for phenomenal consciousness. As above, if his mesh argument that there are instances of the latter that aren’t instances of the former goes through, then this amounts to a rejection of ANC. That was the reason I thought that Carolyn needed to deal directly with Block’s mesh argument.

Moving to Mole, I want to agree with Carolyn that the study Mole provides is insufficient to establish his claim that commons sense denies ANC. Within the context of Mole’s paper, however, I took the study as an example, not as a proof. It would be interesting, I think, to get more experimental evidence for what the folk *do* think. The dangers of imputing unsubstantiated views to the folk seem dangerous indeed.

Of course, it’s only important to get the folk right if we think that what they think is important. For my part, I think getting the folk right gives us a starting place for further investigation. Once we’ve used the folk to ascertain the reference of their terms, however, I think all bets are off. The decrees of the folk can, and should, be revised in light of sufficient theoretical or experimental pressure. I take this to be Mole’s methodology also, though in the article targeted by Carolyn, the examination is of the experimental evidence only. The interesting questions, then, are whether he *has* the folk right (I don’t think the piece of research he cites is sufficient one way or the other) and, if so, whether there’s sufficient pressure to revise them.

Lastly, I’d like to stress, as I didn’t have the space to do in my original commentary, that I think Carolyn’s approach to these issues is original and creative. It’s easy to be complacent about our methods, and I appreciate the opportunity to think and converse more deeply about them with her.

Hi Carolyn,

I enjoyed reading your paper and reading Jennifer and John’s comments. The discussion inspired me to read Mole’s 2008 paper and some other stuff on the relationship between attention and consciousness. I wanted to raise some questions about the best interpretation of Mole’s argument, in particular whether it should be construed as a form of inductive skepticism, whether logical or pragmatic. The claim that attention is necessary for consciousness is presented as a claim that goes against the commonsense notion that “one does not pay attention to all the things that one is conscious of.” In responding to the Mack and Rock 1998 evidence about inattentional blindness (among other points), Mole presents the quote you discuss in the paper: “attention is not necessary for consciousness, but it is necessary if oneʼs experience is to provide one with knowledge of the sort probed by the experimenter’s questions” (95). Here, Mole seems to be distinguishing between consciousness of one’s perceptions and knowledge necessary for reporting one’s perceptions. Could the debate between you and Mole (or the logical inductive skeptic) about whether it counts as a form of inductive skepticism hinge on whether we take ‘reportable’ to mean ‘observable’ and ‘not reportable’ to mean ‘unobservable’? Mole’s professed aim is a kind of deflationary argument that would bring us back to the commonsense notion. Mole’s quote at p. 95, though, if you stress the italicized term ‘consciousness,’ might lead to the interpretation that this is a form of open question argument (, and that may be no different than logical inductive skepticism…). But, it seems like Mole motivates this open question argument by considering how consciousness of perceptions and reportability of perceptions might come apart. Taking inspiration from McDowell and Brewer, he suggests that reportability of perceptual experiences should be construed in terms of a two-tiered account of experience, according to which reportability is dependent upon entertaining a conscious perception with conceptual content (top tier) which depends upon a form of non-conceptual form of conscious perception (bottom tier). (I admit I’m being pretty liberal in interpreting him this way…, but there seems to be a contrast between conceptually articulated experience and perceptual experience that differs from the conceptual articulated kind…). The main quote at issue in the article seems to be, “So long as we continue to operationalize attention and consciousness in the usual ways we will remain unable to adjudicate experimentally between the hypothesis that (1) unattended items are consciously experienced but are not epistemically useful until they have been ‘brought into the space of reasons’ by the (attention-involving) deployment of demonstrative concepts and (2) the hypothesis that unattended items are, ipso facto, not consciously experienced ones” (97). The interpretation of Mole’s argument as a form of logical inductive skepticism might go through if he took the terms ‘conscious perception’ (F) and ‘attention’ (G) to be fixed. Because, then he would be saying something like “Simply because some observed Fs are Gs it doesn’t follow that all Fs are Gs,” where ‘conscious perception’ is understood to depend upon being reported, for instance. But, the quote above suggests that he is not keeping those terms fixed, but instead, expressing that it’s an open question whether we should change the way we operationalize ‘conscious perception’ and ‘attention.’ In particular the focus seems to be on ‘conscious perception’ while keeping ‘attention’ relatively fixed, because he seems to want to separate ‘conscious perception’ from ‘report of perception.’ (There is a similar issue in the outline of Simons and Chablis’s argument about inattentional blindness argument, you (Mole?) have (has) “3. From 1 and 2, lack of attention is sufficient for a failure to report C(g). 4. Failure to report C(g) indicates an absence of C(g).” But, it seems to me it should be “3. From 1 and 2, lack of attention is sufficient for a failure to report (g). 4. Failure to report (g) indicates an absence of C(g).” I wonder also if 3 should be rephrased as “lack of attention is the best explanation of the failure to report (g)”.) So, I wonder if one way to paraphrase your critique is that it is not sufficient for Mole to argue that there are non-conceptual conscious perceptions from the fact that McDowell and Brewer can conceive of a two-tiered conception of experience, but instead Mole has to provide arguments or evidence that there are non-conceptual conscious perceptions that do not involve attention. (I’m not sure if that gets at the logical/pragmatic distinction you want… or whether it can be used as a limiting reply to the shadowy opponent you discussed above.) In that case, your critique of Mole can be raised independently of whether it depends upon Block’s distinction between a-consciousness and p-consciousness, but instead you would be critiquing the distinction between non-conceptual and conceptual contents of conscious perception. Then, the issue seems to turn on whether you think that these non-conceptual conscious perceptions that do not involve attention are observable or unobservable.