Psyche 9(06): 'First-Person Reflection and Hidden Physical Features: A Reply to Witmer' by Charles Siewert

Skip to first unread message

Patrick Wilken

Feb 14, 2003, 3:02:45 PM2/14/03


Charles Siewert
Department of Philosophy
University of Miami
P.O. Box 248054
Coral Gables, FL 331244

Copyright (c) Charles Siewert 2003

PSYCHE, 9(06), February 2003

KEYWORDS: First-Person Approach; Blindsight; Conceivability;
Possibility; First-Person Concepts.

REPLY TO: Witmer, G. (2001). Experience, Appearance and Hidden
Features: Comments on *The Significance of Consciousness*, PSYCHE,

ABSTRACT: My response to Witmer comes in three sections: In the first
I address concerns about my book's blindsight thought-experiment,
remarking specifically on the role imagination plays in it, and my
grounds for thinking (in the face of Witmer's doubts) that a
first-person approach is valuable here. In Section Two I consider the
relation of the thought-experiment to theses regarding possibility and
necessity, and Witmer's discussion of ways of arguing for the
impossibility of "Belinda-style" blindsight, despite its apparent
conceivability. Finally, in Section Three, I consider Witmer's
suggestion that we build on my discussion of blindsight to support the
thesis that consciousness is a hidden physical feature.


I am grateful to Gene Witmer for his detailed and scrupulous remarks on
*The Significance of Consciousness* (Siewert, 1998). (All pages
references will be to this book.) I must apologize for not responding in
commensurate detail to every interesting turn in his intricate
discussion. However, that must not be taken to show any lack of
appreciation for his clear and fair-minded comments.

In my book, I ask the reader to conduct a thought-experiment in which
one takes up the point of view of a hypothetical subject of experience,
Belinda -- a spontaneous, amblyopic, reflective blindsighter. In asking
you to conceive of the situation from her point of view, I ask you to
conceive not just that there is such a person as Belinda -- I ask you to
conceive of being such a person. In this sense my thought-experiment
involves a first-person approach. And there is a serious question --
pressed by Witmer -- about the rationale for taking the first-person
approach I propose. The reason I give in the book (as Witmer notes) is
to avoid certain dangers I say come with taking up a "third-person point
of view" in considering the case. But whatever dangers this may involve,
one might wonder (as does Witmer) -- why not avoid them simply by
refraining from imaginatively inserting oneself into the situation at
all, either as the subject, or her observer? Why not conceive of the
scenario of Belinda's blindsight from *no* point of view within it? This
may well seem preferable -- for the use of imagination I appear to
invite suggests special problems of its own, and in any case arguably
cannot get us what we want, since, concerning many crucial features of
the situation, there is nothing to do by way imagining being someone in
that situation, beyond what would be done simply by conceiving of
someone in that situation. So why not simply consider the matter thus:
describe the blindsight case, and examine this description as best one
can for hidden incoherences? If one finds none, then one can declare
that one has successfully conceived of the case as described, and count
this as evidence that one has indeed described a possible case.

To answer these concerns, I need first to clarify a couple of points
about my procedure. Note that I do recognize that there is a sense in
which (imagistic, iconic) imagining (e.g., visualizing, auralizing) is
distinct from conceiving or thinking (pp.98, 263-4). And I state that it
is specifically *conceiving* that I require in my thought experiment.
Now it is plausible to say that when you conceive of being a person in
this situation (as distinct from just conceiving of there being some
person in a situation of that description), often (maybe inevitably) a
certain exercise of visualization will be involved: one will visualize
in a manner that resembles the visual experience a person in the
situation would have, from her spatial perspective. And that is to be
contrasted with a visualization resembling the experience that would be
had by someone observing that person in that situation. So: if
conceiving of being Belinda involves such an exercise of "centered"
visualization, then part of what one will do here is visualize in a way
that corresponds to having visual experience in one's right field, while
lacking it of things on one's left (which ordinarily one would be able
to see).

However, it is crucial here to avoid certain misunderstandings. First, I
do not require any special talent on the part of the reader to produce
particularly vivid or complete visualizations. I make no explicit
requirement that visual imagery be a part of the exercise, and I do not
think it is necessary for my purposes to take a stand on whether use of
imagery is essential to it. Also I do not assume that every relevant
aspect of the situation described will be something that one can
visualize (or otherwise image) as opposed to merely conceive of. And
finally I do not ask that one conceive of the situation in such a
thorough way as to furnish a determinate answer to just about any
question one might decide to ask about Belinda or her experience. (I
mention this last point in part because I have encountered some
(apparently rather unfriendly) critics (not Witmer!) who dismiss my
thought-experiment as unfeasible, because they think I must be asking
them to imagine "what it is like" to be Belinda (which way of phrasing
the request, incidentally, I do not employ). And they take imagining
what it is like to be someone to require a very rich and far-reaching
imaginative identification with her, which they regard as beyond their
(and maybe anyone's) powers.)

One might accept my clarifications here, but still press the question:
why do I not just save myself the trouble of fending off the
misunderstandings they are meant to avert, and not bother with the
first-person approach at all? I have two reasons. One of these I mention
in the book (and Witmer discusses); the other is not in the book, though
I will try to explain it here. First let me say a word or two to clarify
the rationale I do offer in the book. The danger I see stemming from a
"third-person approach" to the blindsight thought experiment does not
essentially involve any centered use of mental imagery: it's not that I
fear something bad will happen because we visualize Belinda sitting in a
chair with the screen of flashing X's and O's before her. Rather, I
worry that in trying to get clear about what is meant by 'conscious
experience,' one will be, if only covertly, guided by this question:
what evidence would an observer have for attributing a given type of
experience to someone, or for denying that she had it? My worry is that
one will be tempted to give an account of what one means by 'conscious'
in terms of what one would count as evidence for third person
attributions of conscious experience. One may reflect, "When do I think
someone else's states are conscious? When she reports that she has them,
or uses them to make rational choices or inferences. So it seems that I
*mean* by 'conscious state' then, is a state that one can report or use
in self-avowed reasoning." If one starts from some such thought, it can
seem attractive to conceive of consciousness in a behaviorist fashion,
or in terms of some manifest functional role. Thinking along these lines
would then, I believe, lead us away from a proper understanding of
phenomenal consciousness.

Why do I believe that some are tempted to think along these lines? It is
because there are philosophical views that, by my lights, neglect
phenomenal consciousness, in favor of some manifest role, and I find it
plausible to suppose that the appeal of these views is explicable
because there are reasons why some are drawn to confusing *the
occurrence or absence of experience* with what *warrants third-person
attributions and denials of its occurrence*. What are those reasons?
First, notice, in ordinary contexts, we are sometimes inclined to
explain or clarify the use of a term by reference to conditions under
which we would have warrant for applying it. We are thus generally at
some risk of confusing what it is for a term to apply truly (or not),
with what warrants our thinking it does or does not apply. Now, when one
adds to this general disposition to collapse fact and warrant, certain
methodological commitments regarding the study of mind, of the sort that
inspired behaviorism, and a worry that dualism and skepticism are to be
avoided at all costs, it won't be too surprising if one conflates the
phenomenal, experiential facts (e.g., its looking or feeling some way to
someone) with conditions of third-person warrant (e.g., subjects'
ability to discriminate sensory stimuli, or report or express their own
mental states). This confluence of motives has, I think, helped give
behaviorist and functionalist theories of mind their appeal.

But now, one may wonder, if a third-person perspective on the mind can
mislead in this way, why would adopting a first-person approach make us
any better off? As long as there is the possibility of getting it wrong
about one's own experience, won't there also be a fact-warrant gap in
the first-person case that one risks collapsing? In response: I allow
that first-person claims that one has or doesn't have a given experience
are not infallible. But I do think one is not in danger of confusing
what warrants first-person judgments about conscious experience with
conscious experience itself, because *that* identification would not be
entirely mistaken. For I think that the distinctive first-person warrant
or "evidence" that one has a given experience lies, in part, in the
experience itself, and the distinctive first-person warrant that one
doesn't have it is, in part, due to the fact that one doesn't have it.

Of course, you may not share these views of mine, and you may doubt that
third-person perspectives lead to mistakes in the way I have speculated.
My belief that they do helps explain why I proceed as I do, but it is
not essential for me to defend that belief, in order for it to be
legitimate for me adopt a first-person approach in considering Belinda.

However, this still leaves a basic concern of Witmer's unaddressed.
Suppose there is, as I say, this danger of adopting the third person
point of view, which doesn't attach to a first-person approach. Why not
avoid it just by leaving "point of view" out of it altogether? I admit
that the rationale I offer in my book for the first-person approach is
not adequate to answer this question. I now wish to supplement those

We may deprive ourselves of the sorts of philosophical lessons I think
can be derived from Belinda, in a manner that does not (at least
directly) reflect some "third-person" bias. Suppose we approached the
matter as Witmer suggests: we simply consider the description and
examine it for conceptual incoherence. There is a way of doing this, I
believe, which will turn up no conceptual incoherence, but which will,
nonetheless, not involve our successfully conceiving of there being such
a blindsighter as Belinda. This would make use of the idea (mentioned by
Witmer) that we have a purely "recognitional" concept of experience, and
that when one considers the situation in which a subject has ability A,
but not experience E, one understands 'E' by making use of this
recognitional concept. On this basis, one concludes that Belinda's
blindsight is "conceivable," but only in this sense: the possession of
the recognitional concept of experience employed just does not include
the ability either to affirm or deny, with legitimacy, the presence or
absence of entailments among relevant statements, or the possibility of
situations described. (Someone might add: such a concept does not enable
one to "rule Belinda out, a priori," simply because its possession
doesn't enable one to do much of *anything* a priori.)

So, on this view, all we can say here is that one has no right to say
Belinda is conceptually impossible, based on use of this cognitively
*impoverished* recognitional, first-person concept of experience. We may
acknowledge that Belinda's blindsight is not inconceivable, using the
first-person concept of experience, but still think that she is
inconceivable or conceptually impossible, when we utilize any *other*
concept of experience, one which *is* pertinent to the task of trying to
decide whether some state of affairs is conceptually possible.
Alternatively, we might argue that Belinda's not being inconceivable,
using first-person concepts of visual experience, gives us no reason to
resist the idea that she is, in some more-than-nomological,
"metaphysical" sense, quite impossible. Thus merely proceeding in the
way Witmer suggests with the blindsight cases could undermine their use
in trying to assess theories of consciousness.

Now I do not in fact believe that the evident lack of conceptual
incoherence in Belinda's description is due just to some such poverty in
the concepts (the "phenomenal" or "subjective" concepts) we employ, when
engaging in first-person reflection on experience. And we would be
mistaken, in my view, in this way to drain first-person reflection of
philosophical significance and condemn to uselessness a specifically
first-person concept of consciousness, in an assessment of theoretical
accounts of what it is. Of course, I do not ask everyone just to assume
that what I say would be a mistake is in fact a mistake. But, lest we be
led into some (at least potentially mistaken) deflationary assessment of
the significance of Belinda's conceivability, on the basis of some
theory of first-person or phenomenal concepts that yields that result,
we should first try to be sure we are considering Belinda's description
in a manner that really involves the first-person or phenomenal concepts
that are actually available to us.

Therefore, I propose we explicitly begin, not with a theory about our
concepts and their limitations, but with an effort to engage in the very
thought experiment that I urge on my readers, which involves a
"first-person approach" in the sense I explain there. I am not saying
that just by doing this we can show the deflationary view of
first-person concepts of experience would be a mistake. But I do offer
this as an alternative to proceeding in a way that begs the question in
favor of the deflationary view. Rather than just say that Belinda is
conceivable (or at least not inconceivable) on the basis of a view of
our concepts that delivers that result while insuring its theoretical
irrelevance, I ask that instead you affirm her conceivability just by
employing the relevant concepts yourself in conceiving of the blindsight

At this point, someone may say, "OK, I can cooperate with you to this
extent -- I can take up Belinda's point of view in imagination. That is:
I can visualize things as I believe they would look to Belinda. I can
visualize in a way that corresponds to having right field visual
experience of X's and O's, and so on, while visualizing in a way that
correspond to having no visual experience of left field stimuli -- that
is to say, while not visualizing anything to the left at all." But here
I would need to emphasize: my injunction is to *conceive* and not merely
*imagine* the situation from the subject's point of view it (that is,
not merely conjure relevant mental imagery). To conceive of Belinda's
having right field experience, while lacking left field experience, it
is not enough just to visualize right field stimuli, while not
visualizing left field stimuli. To conceive of an absence of visual
experience is not the same as intentionally abstaining from the
corresponding visualization. So, when you conceive of being a
spontaneous reflective blindsighter, you should conceive in a way that
does not consist entirely in constructing (or refraining from
constructing) certain mental images.

However, someone might here insist: "Then I simply cannot do as you ask.
For in trying to 'conceive' of being someone with Belinda's blindsight,
I employ no concept of conscious visual experience that puts me in a
position to say one way or the other whether it follows from having
Belinda's discriminatory powers toward left-field stimuli that one has
conscious visual experience of them -- that they look some way to one."

Now, if *you* say this, I would ask first: what positive reason do you
have to believe your concept of experience limits you in this way?
Second: are you really willing to accept the evident consequences of
this position? For instance, do you really want to say you have no basis
on which to object to, say, an extreme behaviorist analysis of the
concept of experience? Suppose someone says that, on such an analysis
the Venus flytrap's response in seizing its prey, and the sunflower's
phototropic response, entail that they have, respectively, tactile and
visual experience. It appears that you could not then object that there
is a sense of 'experience' in which you think of yourself as having
experience, but can conceive of these plants' lacking any experience
when they make their responses. For on the deflationary conception of
experiential (or first-person or phenomenal) concepts, such conceiving
does not bear on the question of entailment.

Let me summarize the point of this section. Witmer raises the concern
that my first-person approach to the thought-experiment is unnecessary
to avoid the pitfalls I wish to avoid through it, and leads to problems
of its own that attend the use of imagining, as distinct from
conceiving, in the conduct of such conceptual exercises. My response is
that I do not deny -- I even insist upon -- a distinction between
conceiving of a situation, and forming mental imagery. Though I by no
means forbid you to do the second, I definitely ask you to do the first,
in conceiving of being someone with Belinda's blindsight. Also I ask you
to conceive of the matter from the subject's point of view not only (as
mentioned in the book) in order that you will not fail to distinguish
sufficiently the conditions that warrant third-person affirmations and
denials of experience from the fact of its occurrence or absence. I
also, by this strategy, wish to keep readers from employing an
understanding of what is involved in conceiving of Belinda-style
blindsight that would render it unsuitable for mounting the kinds of
theoretical challenge I think emerges from it. To consider Belinda's
situation as Witmer suggests, from no point of view within it, leaves us
open to interpret the conceivability of Belinda in a manner that would
preempt its philosophical significance. However, if we see our task here
as that of conceiving of being someone with her blindsight, we may
succeed in using a first-person concept of experience that is not purely
recognitional and logically vacuous. Of course, someone may deny that we
can succeed in doing this, and hold that our first-person concept of
experience is indeed cognitively impoverished. But then one needs to
explain why that view is maintained, and how one can accept its evident
consequences, when there is, evidently, a more reasonable alternative.


Suppose we do not neutralize my thought experiment with some thin
conception of our first-person concept of experience that would make it
irrelevant. Still there are other strategies one might try for arguing,
in the face of the apparent conceivability of Belinda's blindsight, that
she is, after all, in some more-than-nomological sense, quite
impossible. One might say the problem is not that the concepts we employ
are inherently ill-suited to modal reflection, but that we are confused
about, and misdescribe, just which possibility we focus on, in deploying
them. Witmer discusses some suggestions of this nature that I now wish
to remark on.

First, I want to make a couple of comments on the dialectical situation
at this juncture. My opponent here faces the task of justifying the
claim that what it seems we can conceive of, is after all impossible
(either conceptually or metaphysically), and of explaining why the
necessity in question is one to which our grasp of the relevant concepts
leaves us oblivious. It's not that I assert a possibility my opponent
denies. Rather, I say I can conceive of a situation my opponent would
say is impossible. And I would deny that it is impossible -- which isn't
quite the same as asserting that it is possible. (For such denial is
consistent with withholding commitment about modal facts.) So the burden
of proof rests more heavily on my opponents in this situation. Moreover,
it is not enough for them just to propose a re-description of what I am
doing when I profess to be conceiving of Belinda, and point out that
they can use this description to state a possibility, even while they
hold Belinda's blindsight to be impossible. They also must *justify* the
claim that this correctly describes what I am *really* doing when I
(benightedly) claim to be conceiving of blindsight in which the subject
exercises abilities A without experience E. This is an important point
to which I will return.

Now Witmer would allow that certain ways of trying to discharge the
burden are not successful, which appeal to an analogy with the
superficial appearance of natural kinds. And he notes what my response
would be to a critic's accusation that I wrongly assume that if x lacks
the thought that x has left field visual experience, then x must not
have left field visual experience. The criticism would be that I merely
conceive of the possibility of this missing thought, and confuse this
with the possibility of the missing experience. However, Witmer thinks
there is a further suggestion that poses a bit more of a challenge to me
(which, nonetheless, he thinks ultimately surmountable). The proposal is
that I mistakenly assume that if x has the thought that x lacks left
field visual experience, then x must not have left field visual
experience -- and this leads me to conflate the (genuine) possibility of
Belinda thinking she lacks an experience, with the (spurious)
possibility of her lacking it.

Witmer is right that I would respond to the first suggestion (that I
hold that lack of a thought entails lack of a visual experience) partly
by denying that I hold this view. For it would be contrary to things I
explicitly claim (e.g., on p.208), to hold that higher-order thought is
essential to experience. And, assuming that there is some presumption in
favor of my first-person claims about what I do and don't believe, the
burden is on my opponent to provide evidence that I somehow covertly
believe the very proposition I sincerely deny believing. However, I want
to say much the same thing about the further deflationary move Witmer
suggests. I would also steadfastly deny that your thinking you lack an
experience entails that you lack it. For I would, and do, recognize the
possibility of "hysterical blindness," and I would acknowledge the
possibility of philosophically motivated denials of experience, in my
sense, coming from people whom I believe are, nonetheless, every bit as
visually conscious as I am (see p. 179). And here as before: if someone
would wish to maintain that I covertly believe something, even while I
sincerely deny believing it, and recognize its inconsistency with other
things I profess sincerely to believe, then a heavy burden lies on this
opponent to show that I am so deeply confused about my own beliefs.

But there is more to say here. Suppose, counterfactually, I *did*
believe that thinking you lacked experience guaranteed you did lack it.
Would my opponent then have a case that I am mistaken in thinking I can
conceive of Belinda, as I described her, and all I am really conceiving
of is a consciously sighted subject who falsely denies her experience?
Not at all. For the suggestion would be that I attempt to conceive of
Belinda's blindsight by conceiving of someone thinking she lacks an
experience, and then assuming that if she thinks she lacks it, she must
lack it, and inferring from this that -- well, by golly, she must lack
the experience then. But that just is not the procedure I follow. The
procedure I adopt is the one I explicitly describe in framing my
first-person thought experiment, which relies on no such inference. And
again, what is the evidence that, despite my protestations, this is what
I am really doing? Similarly, I could hold (as in fact do not), that
higher order thought is essential to experience, without thereby
relinquishing the right to say I am conceiving of Belinda. I would say
that while I do believe that Belinda's lack of a thought that she has an
experience guarantees that she lacks it, I do not rely on an inference
from this thesis in conceiving of Belinda. I do not conceive of her
lacking the thought, and then infer that she lacks the experience. What
would be the evidence that would establish I do?

So, I maintain that there is no evidence that I am covertly making the
inferences the critic would impute to me here-and there still wouldn't
be, even if I were to become convinced of the assumptions on which the
inferences depend. What's more, there seems to be some evidence
*against* the critic's accusation. For I ask myself: if I were to
suspend judgment altogether on the truth or falsity of the premises on
which the relevant inferences depend, would I then be disinclined to
assert that I can conceive of Belinda-style blindsight? The answer is:

The conclusion I wish to draw in this section is this. Witmer considers
attempts to argue that Belinda is impossible by maintaining that when we
think we are conceiving of her spontaneous amblyopic reflective
blindsight, really all we're conceiving of is something else, which we
misdescribe in this way. The idea is that our claim to conceive of
Belinda, as described, stems from one of two false assumptions, either:
(a) that lacking a thought about your experience guarantees you lack the
experience; or (b) that thinking you lack an experience guarantees you
lack it. Witmer does, I believe, accurately point out genuine weaknesses
in such arguments, and I welcome the support this provides my view. But
I would bolster my position further. Not only does my honest and
consistent disavowal of (a) undermine the idea that my accepting it
gives me some misconception about what I conceive of, when I say I
conceive of Belinda. My similar disavowal of (b) deprives my critic of
warrant for asserting that (b) dupes me into some conflation or faulty
inference. And I would go a bit further still by pointing out that, even
if I did falsely believe either (a) or (b), this would not show that I
was incorrect in thinking I had conceived of Belinda's lack of
experience (and not merely either: her lack of a thought, or her thought
of a lack). For, to make that case, one would need to show also that I
do not recognize a distinction between: conceiving of (1) Belinda's lack
of visual experience; and conceiving of (2) her lack of a thought or her
thought of a lack. Or else one would need to show that I think I can
conceive of (1) by inference from (2). But there is apparently only
evidence against the hypothesis that I am blind to such distinctions or
rely on such inferences, and there is none in its favor.


What about Witmer's discussion of the prospects for a hidden feature
theory of consciousness? It was instructive for me to see how one might
try to turn my discussion of blindsight to help frame a physicalist
theory of consciousness of the sort Witmer favors. The aspect of my
discussion that seems most crucial here is this. In the argument I
maintain that the concept of consciousness is not just the concept of a
certain manifest functional role, on the basis of a thought experiment
in which the manifest role of a certain kind of conscious visual
experience is mostly filled by something other than a phenomenally
conscious experience. I say "mostly" filled, because I do not maintain,
as would those who argue from "absent qualia" that one can conceive of
the exact same manifest functional role that is filled by the
experience, being filled without it.

I do not claim complete functional equivalence of the phenomenal and
non-phenomenal states is conceivable, because there are certain
attitudes -- for instance, certain judgments about one's own experience,
and desires to have (or not to have) experience of a certain phenomenal
character -- that it seems to me one could not have, if one never had
the relevant sort of phenomenally conscious experience. If the capacity
to give one such judgments and desires is considered part of the
functional role of experience, than it is an aspect of the role that
couldn't be matched by some non-phenomenal substitute. This does not
rescue functionalism, however, because those aspects of what phenomenal
experience gives us that could not (I think) conceivably be gotten
otherwise do not furnish anything with which we can reasonably identify
the difference between having a certain conscious visual experience and
lacking it, in a case where we make the two otherwise as functionally
similar as I am willing to claim we can conceive them to be.

Witmer does not wish to challenge this. Instead he wants us to focus on
this ineradicable residuum of functional difference (as I'll call it),
and see how it might be put to work in his physicalism. I think the
suggestion, ultimately, comes to something like this. We agree that the
difference between having a conscious experience and lacking it does not
consist in some manifest functional difference. But then the suggestion
is: it does consist in what is *responsible* for that functional
difference. And that is something that can be clear to us a priori, if
we do find it inconceivable that a blindsighter could have *all* of the
manifest abilities Connie would have. This isn't to say that our concept
of phenomenal character is simply the concept of whatever accounts for
these abilities. But our concept does set (in Witmer's phrase) "a priori
constraints," according to which phenomenal character is responsible for
abilities that could not otherwise be had. Now if (as seems plausible)
what is responsible for the abilities in question is some hidden
physical feature of the brain, we have reason to *identify* the
phenomenal character of experience with such a hidden feature. Though of
course the discovery of just what hidden feature that is will require
empirical investigation, and the knowledge acquired upon that discovery
will be a posteriori.

Witmer's discussion here is complex and subtle. And I think he goes some
way towards his goal of making intelligible his physicalist thesis that
the first-person concept of consciousness sets the stage for an a
posteriori identification of phenomenal features with hidden physical
features. Still, I have doubts about how much we can use conceptual
investigations that take off from the blindsight thought experiments to
*justify* Witmer's physicalist identifications. Maybe that is just
because of what underlies my sense that there is some conceptually
ineradicable functional difference between visual consciousness and its
lack. So let me try to spell this out a bit, in hopes that this will
make the situation clearer.

My view (and here I am only asserting it, not defending it) is that
there are first-person judgments we make about our experience that
involve the use of first-person (or phenomenal) concepts of experience
-- concepts brought to the fore by conducting the process of reflection
that starts with paradigmatic instances of phenomenal consciousness in
one's own case, and leads to various hypothetical scenarios in which
conscious experience is contrasted with its absence -- as in my
blindsight stories. And I also believe we can make the first-person
judgments employing these first-person concepts of experience only if we
are capable of a certain kind of demonstratively expressible thought
about our own experience and its character: (e.g.) "This feels this way"
(where the first 'this' refers to one's experience, and 'this way' picks
out some type in principle recognizable by me on its recurrence (say, a
specific sort of "burning, nervous" feeling in my arms). For it is on
the basis of such thoughts that we are able to consider paradigms of
conscious experience in ways that exhibit our grasp of the first-person
concept of experience.

Now granting me for the moment that something like this is true, it
seems we may have reason to deny the possibility of an anti-physicalist
blindsighter of the following sort. This blindsighter would not only
have Belinda's capacities, she would also have the hidden feature that
in Connie is at least nomologically sufficient to generate the visual
experience Belinda lacks (so she is to this extent like Witmer's
character "Melinda"). But she would have more than Melinda. This
blindsighter's visual states also enable her to have all the same
manifest abilities (including those for first-person judgment and
demonstrative thought) that Connie's conscious visual experience gives
her. The possibility of a blindsighter of this sort -- allow me to call
her 'Impossilinda' -- would be ruled out, since (ex hypothesi), nothing
*but* experience of the phenomenal character Connie has could, on
reflection, conceivably give one the capacity for the experiential
judgments that those with conscious experience in fact have with regards
to it.

But now, from this am I entitled to take the further step, to hidden
feature physicalism? To do so, it seems that I need to reason in
something like the following manner. The capacity for the reflective
judgments that we balked at attributing to Impossilinda is (necessarily)
due to our having phenomenal experience of the sort they would be about.
But we make such judgments due to hidden physical features of our
brains. Therefore the phenomenal character of experience is one and the
same as some hidden physical feature of the brain.

My problem here is that it is not clear that the manner in which the
phenomenal character of experience is responsible for the capacity for
experiential judgment is the same as the manner in which my neurobiology
is responsible for this. Perhaps the *phenomenal character* enables me
to judge insofar as it puts in place a conceptually necessary condition,
whose presence, together with other factors, can thus explain why I was
able to judge in a certain way on a certain occasion. Perhaps *hidden
physical features*, on the other hand, enable me to judge, just by being
nomologically sufficient for the occurrence of a judgment of that type
(together with experience of the relevant phenomenal character). But
then we cannot (without equivocation) use the notion that both the
phenomenal character and the hidden feature enable me to judge as a
basis for justifying the physicalist identification. For what it is for
each to be responsible for my ability to judge differs in the two cases.

Another way to bring out my difficulty here: it seems to me that
acceptance of something like my reason for ruling out Impossilinda is
compatible with rejecting physicalism on the grounds of some "knowledge
argument" or the possibility of Chalmers' zombie world. (My zombie twin
would not share my capacities for forming and employing phenomenal,
first-person concepts and judgments, though of course he would speak as
if he had!)

It may be that Witmer does not mean to suggest that reflection on our
concept of experience will do any more by way of justifying the
identification of manifest phenomenal features with hidden physical ones
than I have allowed here. And perhaps he would allow that further work
is needed to show that the sense in which the phenomenal character of
experience is (a priori) "responsible for" experiential judgment or
thought is univocal with that in which hidden physical features are (a
posteriori) "responsible for" experiential judgment or thought. But then
at least perhaps I have located a challenge for him to meet, in offering
a fuller justification of his version of physicalism.

In any case, it does seem that physicalists of Witmer's stripe should
want to be friendly to my blindsight argument in order to defeat
manifest functionalist views to which they are opposed. So at least my
view has that much aid to offer him. And he has helped me to see that I
need more fully to work out, defend, and consider the implications of my
views regarding the nature of specifically first-person or phenomenal
concepts of experience.


Siewert, C. (1998). *The Significance of Consciousness.* Princeton:
Princeton University Press.

Reply all
Reply to author
0 new messages