Psyche 7(10): Phenomenal Consciousness and the First-Person by Joseph Levine

	Charles Siewert's book, The Significance of Consciousness, can be purchased from Amazon.Com

Phenomenal Consciousness and the First-Person

Joseph Levine
Department of Philosophy
350 University Hall
The Ohio State University
Columbus, OH 43210-1365
U.S.A.

levine.120@osu.edu

PSYCHE, 7(10), April 2001
http://psyche.cs.monash.edu.au/v7/psyche-7-10-levine.html

KEYWORDS: consciousness, eliminativism, first-person warrant, higher-order theory, intentionality, phenomenal character, qualia

COMMENTARY ON: C. Siewert (1998) The Significance of Consciousness. Princeton: Princeton University Press. x + 374pp. ISBN: 0-691-02724-2. Price: $42.50 hardcover.

ABSTRACT: Siewert's book revolves around three theses: that there is a distinctive style of epistemic warrant associated with the first-person point of view, that if we pay close attention to the deliverances of this first-person point of view, we will see that phenomenal consciousness is both real and yet neglected by many current theories that purport to explain consciousness, and that phenomenal consciousness is inherently intentional; one cannot divorce what phenomenal character presents to us from what it's like to have it. Among several points made on the relations among these three theses, it is argued that Siewert's argument for the distinctive status of first-person warrant does not provide him with the support necessary to employ that thesis in his defense of the significance of phenomenal consciousness.

Siewert's book aims to establish, as the title announces, the "significance" of consciousness. He does this by arguing in great detail for three principal theses: First, that there is a distinctive style of epistemic warrant associated with the first-person point of view. Second, that if we pay close attention to the deliverances of this first-person point of view, we will see that phenomenal consciousness - defined with respect to a particular set of thought experiments - is both real and yet neglected by many current theories that purport to explain consciousness. Finally, again by reference to what can be known with this distinctive first-person warrant, he argues that phenomenal consciousness is inherently intentional; one cannot divorce what phenomenal character presents to us from what it's like to have it. The third point, of great interest in its own right, is then used to show that neglect of phenomenal consciousness involves a much more serious lack in a theory than one might have otherwise thought.

The argument for the claim that there is a distinctive first-person type of warrant for beliefs about our own mental states is based on two considerations: that we are normally warranted in our beliefs about our own mental states without having made the sorts of observations of our behavior that others would have to make to be so warranted, and that we are often warranted in our self-directed beliefs even when no third-person accessible evidence at all is available. Given especially the second consideration, it seems to follow that first-person warrant is of a distinct type from third- person warrant.

Depending of course on what one reads into the phrase "distinct type of warrant", I don't think many would doubt that we can come to have warranted belief, indeed knowledge, of our own mental states without recourse to the sorts of observations others would have to make to support identical mental attributions to us. We don't have to see how we behave, or listen to what we say. (Yes, there are those instances where we learn about our own mental states from observing our own behavior, or listening to what we've said. But this doesn't impugn the basic point.) But just how much weight is this notion of a distinct type of warrant supposed to bear, and is it up to it?

In particular, consider the following way of understanding what is distinctive about first-person warrant: it is distinctive in that it involves a mode of observation that is available only to the subject. The idea is that the mind is set up with internal monitors that directly access the presence of certain sorts of mental states (roughly, the conscious ones), and therefore, due to the reliability of such monitors, when one finds oneself thinking that one believes such-and-such, or perceives such-and-such, one is warranted in thinking it. Though it's pretty clear that Siewert doesn't agree with this way of understanding first-person warrant, the question I want to pursue now is whether the use he makes of first-person warrant would be compromised by understanding it in this way.

Siewert asks us to consider a case of "spontaneous blindsight" as a way of specifying what he means by "phenomenal consciousness". Blindsight involves the ability of certain brain-damaged patients who report blindness in a certain region of their visual field nevertheless to detect various stimuli presented in that region when requested to guess what's there. By a "spontaneous blindsight" case Siewert imagines a case where the patient spontaneously makes these judgments about what's in that region of their visual field, without the instigating requests to guess. Though this doesn't in fact happen, it certainly could happen, argues Siewert. Now, consider what's missing from their visually induced state, that feature which distinguishes it from the states they occupy when presented with stimuli in the other areas of their visual field. There is a way it seems to the subject to see an "X", say, in their good side, that is missing from how it seems to them when blindsightedly detecting the presence of the "X" in the bad side (even when done spontaneously). This missing feature is the phenomenal character of the experience. Having phenomenal character of this sort is what constitutes having phenomenal consciousness.

So where does appeal to first-person warrant come in? Well, someone might object to belief in phenomenal character on the grounds that it is an unnecessary theoretical posit. We can adequately explain all we need to about a subject's behavior, including their verbal behavior, without recourse to phenomenal character. This is of course a standard argumentative strategy of qualia-eliminativists. One line of reply is to argue that phenomenal character doesn't enter the picture merely as a theoretical posit, but rather as a datum itself. We experience phenomenal character first hand, and it is part of the job of a theory of mind to explain it. Since its role as a datum is primarily manifest from within the first-person point of view, legitimating first-person warrant becomes a crucial part of the defense of belief in phenomenal character.

Suppose we adopt the internal monitoring account of first-person warrant described above. The idea then is that our internal monitors detect phenomenal character, and this is how phenomenal features emerge as data for theorizing about consciousness. But now the alleged neglector of phenomenal consciousness, in particular the eliminativist, can challenge Siewert as follows. Our epistemic access to phenomenal features, as with epistemic access to any other data, is a matter of observation (of a sort). In particular, it is a matter of some cognitive state arising as a relatively direct effect of some internal state. So long as one can explain how such cognitive states might misconstrue the nature of their objects, and thus explain the "appearances" away, and so long as we have good theoretical reasons to think that phenomenal features don't really exist (apart from their functional/behavioral correlates, that is), there is nothing special about first-person warrant that should stand in the way of accepting the claim that phenomenal consciousness is just an illusion of sorts.

Now I don't buy this internal monitoring story, and clearly neither does Siewert. But I wish I had a really good argument against it, and I don't see one in Siewert's discussion. He does argue against the internal monitoring account of consciousness, showing persuasively, I think, that to be phenomenally conscious is not to have a second-order state intentionally directed on a first-order mental state. But for all that discussion shows, it's still possible that when we do explicitly form beliefs about our first-order mental states, including those we would classify as phenomenally conscious ones, some sort of internal monitoring architecture is at work. Arguing against the internal monitoring account of consciousness is not the same as arguing against the internal monitoring account of first-person warrant.

In addition, if we look back to the original argument supporting the claim that there is a distinctive type of first-person warrant, it is perfectly compatible with the internal monitoring account. Yes, I don't have to observe my behavior to tell what mental state I'm in, and I can tell what mental state I'm in even when no third-person accessible evidence is available, but this is because I am built so that my mental states are normally monitored. I have a cognitive capacity with access to my mental states that no one else does. However, the workings of this capacity are in principle assimilable to those of my cognitive capacities that detect external features of the world.

I agree with Siewert that something special is involved in first-person warrant, or access, and I also agree that only by taking this seriously can we keep our eyes on the prize and cease neglecting conscious experience while supposedly explaining it. However, what I think we need, and perhaps Siewert would agree, is a better understanding of what this first-person access consists in. In other words, we need to see not just that first-person warrant is distinctive, but how it is distinctive, if we are to fend off the consciousness neglectors.

It's clear that Siewert's strategy is to avoid any theorizing, whether about first- person warrant or consciousness itself. This way he keeps his commitments as weak and subject to general agreement as possible, and then he can show how powerful the consequences are nevertheless. But the notion of first-person warrant that emerges from merely noticing that we can tell what's going on with ourselves mentally when others can't doesn't, as far as I can see, support very powerful consequences. What's more, I think it causes Siewert to miss something crucial about phenomenal consciousness itself.

My concern emerges in his discussion of what is wrong with higher-order theories of consciousness. He is concerned to combat the idea that consciousness is a matter of having some other mental state intentionally directed on the target conscious mental state. I agree with him in his rejection of this theory. I don't believe that phenomenal consciousness is analyzable as a relation between two mental states, the one about the other. But I have two worries about his argument for rejecting this theory.

First, I find his dismissal of the idea that we might have non-conscious judgments that accompany our conscious mental episodes (thereby making them conscious) too quick. His argument seems to come down to this: he sees no reason to believe in such unconscious judgments. Well, the higher-order theorist might claim that the virtues of the theory itself give one reason to posit such states. What we think of as conscious experience is thereby explained, goes the story. Now I don't think it does explain what we want explained, but I think Siewert needs to say more about why it doesn't.

However, my main worry is one that connects directly with my remarks about first-person warrant above. What endows the higher-order theory with whatever plausibility it has, it seems to me, is that conscious states are those the subject is aware of. True, as Siewert correctly notes, some theorists have been too quick to interpret this phrase, "aware of" (or "conscious of") in the way that leads to adoption of a higher-order account. But I don't think all the intuitive support for the view derives from this. After all, is it an accident that the mental states with respect to which we have first-person warrant are the conscious ones? Isn't it the case that when a state is conscious, even if it isn't a state that I explicitly take mental note of, or introspect, but just experience, that the state, its phenomenal character, is present to me, "for me", in a way that non-conscious states are not? And isn't this subjective character steeped in epistemological significance and of the essence of what it is to be phenomenally conscious?

Of course this is precisely the question: what is essential to being phenomenally conscious? Siewert (wisely, no doubt) never tackles this question head on. He characterizes the target of his investigation - phenomenal consciousness - by reference to what is missing from the blindsight case, but this of course doesn't provide a very rich characterization. The one feature that he does deem essential, though presumably not exhaustive, is intentionality, which brings us to the third principal argument of the book. According to Siewert, phenomenal features are intentional, and of necessity. I want to spend a little time exploring this quite intriguing idea, and then bring it back to my concern about subjectivity and first-person warrant.

Siewert argues, with respect to perceptual experiences, imagistic experiences, and pure conscious thoughts (i.e. thoughts unaccompanied by visual or aural images, what he calls "non-iconic" thoughts), that these episodes of phenomenal consciousness possess intentional features, and not merely contingently. So, for instance, my visual state of its seeming to me as if there is an "X" in my visual field at a certain location, is necessarily such that it is accurate if there is an "X" and inaccurate if not. Similarly, its seeming to me the way it does when I think that today is Monday is necessarily a thought about Monday, and therefore is inaccurate (or false) if today isn't Monday. It wouldn't seem as it does if I were in fact thinking about Tuesday.

The contrasting view is one that sees phenomenal character as itself non- intentional, and therefore phenomenal states require some further condition to interpret them. On this view, as opposed to Siewert's, it's possible for two subjects to occupy phenomenally indistinguishable but intentionally distinct states. So, for instance, though it might seem to both me and my phenomenal twin as it does for me when it looks as if there's an "X" at 2 o'clock, my experience is accurate just in case there's an "X" at 2 o'clock, whereas my phenomenal twin's experience is accurate just in case there's an "O" at 2 o'clock. Siewert finds this possibility extremely implausible, and therefore concludes that the accuracy-conditions for the experience are essential to its phenomenal character.

It seems to me that Siewert's argument is quite strong with respect to visual experiences of spatial properties. It really is hard to see how a phenomenally conscious visual experience of an "X" could be about anything other than that shape; certainly not about an "O". It's much less clear to me that he has such a strong argument when we consider other features, whether it be tastes and sounds, colors, or non-iconic thoughts. It's been a staple of anti-functionalist and anti-representationalist arguments that red surfaces could have looked to us the way green ones in fact do, and the same seems to be true of tastes like bitter/sweet, or salty/non-salty, and the sounds of various pitches. I wasn't convinced by Siewert's discussion of color that this wasn't so. Similarly, is there really some particular way it seems to think about Monday that is distinguishable from the way it seems to think about Tuesday (leaving aside the aural sensation of saying "Monday" to oneself, which is another issue)? I'm not sure my intuitions are up to deciding that question.

In fact, leaving aside the case of phenomenal representations of spatial properties (about which I really don't know what to say), I think there's some reason to be quite suspicious of the claim that a conscious thought's intentional features - its accuracy conditions - are determined by its phenomenal character alone. The problem is this. We are asked to consider two non-iconic thoughts, say that today is Monday and that today is Tuesday, and then note that there seems to be a difference between them; or, better, that there is a difference in the way they seem. But it's one thing to notice a difference in the way they seem, and another to attach interpretations to the way they seem. For example, how do I know that the one is about Monday and the other about Tuesday, and not the other way around? All I have to go on is my ability to detect, in a first-person way, what I'm thinking, but that is just more thought. There doesn't seem to be a way for me to forestall the possibility of wholesale reinterpretation, of the sort that Quine, say, imagines. I'm not saying there is no determinacy of interpretation; rather, that whatever determinacy there is can't be pinned on the phenomenal character of the thoughts in question.

Still, it is worth pursuing the question: just what is it for there to be a way it seems to have a non-iconic thought? At times I wonder whether there really is any non-iconic phenomenal character, and it isn't really all a matter of the aural and visual imagery associated with a thought. But at other times Siewert's claim seems right on the money. For instance, I just got up to get myself another cup of coffee. This was a conscious intention. I knew why I got up, though I never said to myself anything like "I think I'll go get a cup of coffee". If you asked me what I was thinking when I got up, I'd tell you. What's more, it's clear to me that it was a desire for a cup of coffee that I had in mind, not for something to eat, or to make a telephone call (even if I ended up doing those things as well).

On the other hand, along the lines suggested above, we can characterize the situation this way. When I have a conscious thought, I'm aware of the fact that I'm having it, and there is something like a phenomenal character involved in having it. I can tell when two thoughts differ, or when a suggested expression for my non-iconic thought is accurate or not, but this need not be a matter of detecting a difference in the way they seem. Rather, I have a way of knowing what I'm thinking, and this is to think some more, perhaps non-iconically. Direct revelation of the thought's intentional character through its phenomenal character is not involved.

One final word about the connection between phenomenal and intentional character before returning to the question of first-person warrant and awareness. It's interesting to chart a bit of the dialectic among the functionalist/representationalists, the strict separators of phenomenal and intentional character, and those, like Siewert, who are anti-reductionist about phenomenal character, but see a necessary link between phenomenal and intentional character. The first group claim that phenomenal character is nothing but intentional character. The reddishness of a visual sensation is just a representation of redness. This is clearly a reductionist view (sometimes with eliminativist overtones as well).

The standard response to the reduction of phenomenal character to intentional content has been to show how one could divorce the two, as in the inverted spectrum hypothesis. If one could have two phenomenal twins with distinct intentional contents, then of course one can't reduce phenomenal character to intentional content. The representationalist reductionists have therefore expended a lot of energy trying to undermine these thought experiments that allegedly show that the two notions can be divorced.

Along comes Siewert who says both sides are wrong. You can't divorce phenomenal character and intentional content, but that doesn't mean you can reduce the former to the latter either. Well, if the anti-reductionist strategy of divorce is wrong- headed, what is supposed to block the reduction now? Presumably, the spontaneous blindsight thought experiment. That is, if we could reduce phenomenal character to intentional content, then, since the relevant intentional content is present in the blindsight case, there shouldn't be a difference between blindsightedly detecting a stimulus and seeing it. This difference is phenomenal character.

But if the two states - blindsight detection and seeing - both share an intentional content and differ only in one having and the other lacking phenomenal character, doesn't that entail that the phenomenal character is detachable from the intentional content? No, it doesn't. One could characterize the situation this way. A phenomenally conscious state can share intentional content with a non-conscious state. However, when an intentional state is phenomenally conscious, that state's phenomenal character is distinctive. There is a particular way it seems to consciously entertain a particular intentional content. Thus Siewert's version of anti-reductionism constitutes a viable alternative to the divorce view which has been dominant among anti-reductionists. For the reasons given above I'm not totally convinced it's right, but I value having this alternative articulated.

Let me now return to the question of first-person warrant, with which I started. I claimed above that Siewert's characterization of phenomenal consciousness left out something essential, namely the way that what it is to be phenomenally conscious is intimately connected with awareness, being "for the subject". Though I find the higher- order theory inadequate, I do think part of what drives it is recognition of how essential awareness, as an epistemic relation, is to being conscious. I agree with Siewert that something is missing if we characterize what distinguishes phenomenally conscious sight from blindsight purely in terms of access - especially if that is reduced to non- phenomenal, purely intentional terms. However, I also think that we lose something if we don't recognize that an essential component of what distinguishes phenomenal sight from blindsight is that with the former the state I'm in is epistemically available to me in a special way - a way peculiar to conscious states. It is specifically this distinctive, conscious way of being epistemically available that gives first-person warrant its distinctive character.

Of course this feature of "availability" (for want of a better name) motivates not only the higher-order theory but also the traditional Cartesian theater model of consciousness. Siewert is right that commitment to the reality of phenomenal consciousness does not commit one to the idea of a Cartesian theater, but it is not as if the theater model has nothing to recommend it. The idea that what's conscious is there, in the "theater", to be taken notice of in a way that what is not conscious is not, seems quite close to what distinguishes conscious from non-conscious states. The main problem - leaving aside metaphysical issues about the spatiotemporal location of the theater - has to do with the alleged observer, the "audience" in the theater, and its relation to what's conscious. There doesn't seem to be an observer, and even if there were, the relation of observation - if understood on the model of normal perception of external objects and features - doesn't seem to accurately capture the phenomenon of conscious availability. What model does capture it? Beats me.

Earlier I argued that Siewert's support for the claim that there is a distinctive type of first-person warrant wasn't sufficiently strong to fend off certain "deflationist" interpretations (such as the internal monitoring account) that might undermine the use to which he wants to put the notion. In addition, it seems to me that he doesn't adequately incorporate what is distinctive about the first-person point of view into the very notion of phenomenal consciousness. On the whole, however, I found Siewert's investigation into phenomenal consciousness sensitive, intriguing, and a useful antidote to much of what he so aptly calls "consciousness neglect".