Psyche 9(03): 'Eliminativism, First-Person Knowledge and Phenomenal Intentionality' by Charles Siewert

1 view
Skip to first unread message

Patrick Wilken

Jan 29, 2003, 7:07:00 PM1/29/03


Charles Siewert
Department of Philosophy
University of Miami
PO Box 248054
Coral Gables, FL 33124

Copyright (c) Charles Siewert 2003

PSYCHE, 9(03), January 2003

KEYWORDS: consciousness; eliminativism; first-person knowledge;
phenomenal color experience; conscious thought.

REPLY TO: Levine, J. (2001) Phenomenal Consciousness and the
First-Person: Comments on *The Significance of Consciousness*,
*Psyche, 7(10).*

ABSTRACT: Levine suggests the following criticisms of my book. First,
the absence of a positive account of first-person knowledge in it
makes it vulnerable to eliminativist refutation. Second, it is a
relative strength of the higher order representation accounts of
consciousness I reject that they offer (as my account does not)
explanations of the subjectivity of conscious states and their special
availability to first-person knowledge. Further, the close connection
I draw between the phenomenal character of experience and
intentionality is unwarranted in the case of both color perception and
conceptual thought. In response to Levine's critique, I argue that
the eliminativist can be rebutted and higher-order representation
theories found wanting, even without offering a positive account of
first-person knowledge. Also, I note that I actually have begun to
offer an account of this based on my conception of phenomenal
consciousness. Finally, it will be seen that Levine's concerns do not
undermine my views on color experience, conscious thought, and
intentionality, once their justification and character are made clear.

I thank Joseph Levine for his thoughtful comments on *The Significance
of Consciousness* (Siewert, 1998). (All pages references will be to this
book.) His criticisms raise important issues, challenging me to develop
my views in directions I agree I need to explore. Levine's main areas of
concern are these. He says that until I provide a positive account of
what first-person knowledge or warrant consists in my account of
consciousness is either vulnerable to eliminativist objections, or (at
best) seriously incomplete. Further, he doubts whether the close
connection between phenomenal character and intentionality that I
(correctly) say holds in the case of visual spatial experience can be
extended, on the basis of first-person reflection, as far as I claim
with respect to color experience, and non-imagistic, conceptual thought.
I will address these issues in turn.


The first concern Levine raises seems the most threatening to my
views. The fact that I leave it open just what the "distinctiveness" of
first-person warrant consists in creates, he suggests, a deep
vulnerability opponents could readily exploit. Where I am silent, they
might offer their own positive account, and use this to justify an
eliminativist position on phenomenal consciousness. More specifically,
they might hold that the specialness of first-person knowledge consists
in our possession of an "internal monitoring" mechanism that is reliable
in its pronouncements -- but reliable only within certain bounds, bounds
that do not include the self-attribution of what I call "phenomenal
consciousness." Then they might say: "Set to the side any attributions
of phenomenally conscious experience, and ask: can we adequately explain
what is left (behavior, information-processing) without re-introducing
any conscious experiences into the picture? The answer is: yes, we can.
Thus the inner monitor's reliability does not extend to whatever
self-attributions of conscious experience it might generate." So
experiences with phenomenal character in my sense are idle, unnecessary
explanatory posits, and my book purports to discuss the significance of
a phenomenon we have no right to say ever occurs.

The critic here assumes that we have first-person warrant for the
application of this class of psychological predicates -- those
expressing phenomenal features -- only if that can be justified by
claiming it is needed to account for explananda conceived of in some way
that thoroughly excludes the members of this class. In my book, I
identify that underlying eliminativist assumption, and question why we
should accept it. This begins in Chapter 2 (pp. 49-61), where I argue
against a certain aspect of Cartesian epistemology that has, ironically,
been adapted to eliminativist ends -- namely, the view that the
third-person application of psychological terms must be justified by
appeal to their indispensability in accounting for expananda that have
been thoroughly "de-mentalized." And I argue that, even if first-person
applications are not warranted without warranted third-person
applications, this does not show that first-person knowledge is
dependent on a third-person knowledge of mind conceived of in this way.
Therefore, we cannot fairly impugn our first-person knowledge of mind by
arguing that it is dependent on a third-person source of justification
that regards all mentalistic talk as a system of explanatory posits that
fails to explain as it should. Then in Chapter 5 (p. 152), I propose
that this sort of response be given, not just, as before, to an
eliminativist who wants to throw out all everyday mentalistic talk on
such a basis, but also to one who would mount such an argument, more
specifically, against all attributions of phenomenal experience. That
is, I challenge the grounds for assuming that the warrant for
attributions of conscious experiences is to be assessed by asking
whether such attribution better explains data conceived of in a way that
excludes any phenomenal experience, either directly or by entailment.

In conversation it has been put to me by Georges Rey, who advocates a
view along the lines Levine describes, that the eliminativist
epistemology is supported by the following analogy. Suppose we met
someone who is firmly convinced that his dreams give him knowledge of a
separate world. He says that in his dreams he visits a peculiar place
that exists even when he does not dream. Though he has no idea where
this place is in relation to places encountered in waking life, nor how
he is transported there, still, he is adamant that he knows it is real.
Of course, we should not believe him. For we should think his warrant
for his beliefs about this "dream place" depends on how well his
dreaming experience is explained by the truth of these beliefs. And that
is to say: not well at all. Now if this is right, we should accept
something similar regarding anyone's assertions that his or her
experience has phenomenal character, understood as I propose. Unless
these outputs of their internal monitors can be justified by saying
their truth better explains data described in such a way that does not
assume any phenomenal features have instances, we should regard whatever
beliefs we hold that they do -- however firmly we hold them -- to be
just as groundless as beliefs about a mythical place one visits in one's

Levine himself does not offer Rey's analogy, but I would like to get to
the bottom of this style of objection, so let's examine it. First, I do
not find initially plausible the idea that my warrant for judging that I
have conscious experience should be assessed in a way similar to that in
which the dreamer's judgments about his dream place should be assessed:
by asking whether accepting their truth yields the best explanation of
something else. Why should I accept that the two cases are analogous?

Perhaps it will seem that the presumption is in favor of the analogy;
unless we can produce some reason to say the situations are not
epistemically analogous, we should believe that they are. However, I
don't think the analogy carries any such presumptive weight. Consider
first: it seems that an eliminativist who draws such an analogy should
recognize some domain to which it does not extend. For ask yourself:
Does it extend, for example, to beliefs about the real places we visit
when awake? Are we to say that just as we should assess the warrant for
beliefs about dream space and the goings-on therein by asking how well
they explain other goings-on, so too we should assess the warrant for
beliefs about waking space, and the goings-on in it: by asking how well
they explain something else entirely? We might want to say this, for we
might consider facts about the phenomenal character of one's experience
to constitute the "neutral" non-question-begging data, the explanation
of which determines the fate of one's beliefs about things in (real)
space. This would be the traditional Cartesian way of viewing one's
epistemic situation. But note that the would-be eliminator of phenomenal
consciousness cannot share it. For then beliefs about space would give
him no standpoint from which he could criticize and reject beliefs about
phenomenal experience.

So eliminativists should recognize some domain where the analogy with
"dream space" beliefs does not hold. And this invites us to oppose the
eliminativist's analogy with another. The class of beliefs about the
phenomenal character of one's experience is, epistemically, not so much
like the class of hypothetical beliefs about dream space, as it is like
the class of beliefs about the space of waking life, at least in this
respect -- neither is in need of justification by viewing it as
comprising a candidate theory to explain evidence that entails the truth
of no one of its members. So at this point we face a choice of
analogies: is our warrant for beliefs about phenomenal experience
similar (in the just mentioned respect) to our warrant for beliefs about
the real space perceived and inhabited in waking life, or, is it (as the
eliminativist says) more like one's warrant for beliefs about the
behavior of things in the purported "space of dreams"? We have no right,
it seems to me, to suppose that the latter analogy enjoys some
presumption of correctness, as against the former.

The eliminativist needs some positive justification for his favored
analogy, if it is to support the epistemological assumptions essential
to his argument. No plausible justification for accepting the analogy
occurs to me. However, I can think of some reasons to reject it. For
consider: just what is supposed to serve as the point of analogy with
the dreamer's experience; what constitutes the data to be explained by
the "postulation" in me of phenomenal experiences? Not yet another level
of experiences -- some experiences of the phenomenally conscious
experience, that would stand to beliefs about phenomenal character much
as the dreamer's dream experiences stand to his beliefs that there are
people and things moving about in dream space. No. Well, what then?
Perhaps the data are my judgments that I have this or that sort of
conscious experience, at one or another time. But this seems
unsatisfactory for another reason. My understanding of the utterances by
which I express such judgments depends on my success in referring to
illustrative instances of the kind of phenomenal character involved. So,
if I don't actually manage to identify for myself real instances of
these ways of seeming (ways in which it looks to me, feels to me, etc.),
then I quite literally do not understand what I mean by the terms I use
for them. But in that case, I do not even make the judgments that were
our most recent candidates for the data to be explained by the
postulation of consciousness. Those "data" do not exist. In that case,
eliminativism winds up doing away with the data to be explained as well
as the "posits" to explain them. Or else the only data to explain now
are bits of unintelligible verbal behavior.

It now emerges that the eliminativist position is not that my reports of
consciousness are false -- rather they are unintelligible or somehow at
bottom incoherent, like beliefs in magic perhaps. Of course the
eliminativist, like the logical positivist of yore, might relish the
prospect of accusing opponents of talking nonsense. But now, as then, we
should recognize that the evidential standards for such an accusation
ought to be high. And we should notice that the earlier argument against
me has, in effect, been abandoned for a new one, claiming something
like: "It is, on examination, unintelligible to suppose that, to someone
in Belinda's situation -- to someone with Connie-like powers to
discriminate left field visual stimuli -- it might look no way at all on
her left." But just what is the argument for that?

The eliminativist may be tempted to retort here that asserting the
reality of phenomenal experience in my sense would leave us with
metaphysical puzzles that admit of no truly satisfactory solution. We
would be condemned to a dismal choice between, on the one hand, some
form of property dualism (which inevitably brings epiphenomenalism), and
on the other, acceptance of some brute materialist necessity (leaving us
with an unclosable explanatory gap). And where some putative property
gives rise to such metaphysical nightmares, we are justified in refusing
to accept that it has instances.

But even if accepting consciousness does leave us with only these
choices (and is our grasp on cause and necessity so firm that we cannot
entertain a doubt about this?), to react by denying the reality of
consciousness would be a cure much worse than the disease. Anyway, one
wonders what entitles us to assume reality should leave us with pleasing
metaphysical options and whether consistently applying the principle,
"metaphysically problematic, therefore not real," will leave us with
much reality at the end of the day.

I want to emphasize that the challenge I pose here is not merely to
provide a reason to believe I have wandered into some kind of
incoherence in my blindsight thought experiment. The challenge I address
to the critic is this. Engage in the process of reflection that starts
with the argument for first-person warrant and with it gathers together
broad classes of examples that I say share the feature of being
conscious (experiences differing in phenomenal character). Proceed
thence to conceiving of various forms of blindsight from the point of
view of one who has them. See then whether you yourself are capable of
deploying (or at least seeming to yourself to deploy) the notion you
want to argue contains some secret incoherence. If you refuse to engage
in this exercise of reflection at some point or another, then I beg you
to consider whether your refusal is justified, or whether it manifests
only an aversion to sources of doubt about your philosophical position.
Of course you may argue for the incoherence of my understanding of
phenomenal consciousness. But I ask that you first give it a fair
hearing, and that means that you yourself at least try to engage in the
reflection that I maintain manifests such understanding. If you do, you
may find your intellectual battle against this idea is as much a
struggle with yourself as it is with me. Must we really strain against
our own minds this way?

There are other objections worth noting to the line of criticism Levine
mentions. We might argue that if the reliable inner mechanism account of
self-knowledge invoked by Levine's eliminativist will not support our
claim to warranted judgments about the phenomenal character of our
experience, well: so much the worse for that reliable inner mechanism
account of self-knowledge. The question this raises is a serious one.
Just when is it justified to use a theory of first-person knowledge to
argue against our warrant for a broad class of first-person judgments?
That is, when can we fairly argue that, since a theory of first-person
knowledge does not show us as having a distinctive warrant for a given
class of judgments (though some are disposed to claim it), the members
of that class are all warrantless? We should treat this form of argument
with suspicion. For we need to ask: where does the warrant for accepting
a theory of first-person knowledge come from anyway? It seems to me that
if a given account of first-person knowledge fails to grant us knowledge
of such and such occurrences that, independently of epistemological
theorizing, we presume ourselves to know to be true in a distinctively
first-person way, then that is, prima facie, a reason to reject the
account. Those arguing that we don't know we have conscious experience
because their theory of first-person knowledge doesn't grant us that
knowledge need to explain why this is not just an indication that they
are offering a defective theory of first-person knowledge. Perhaps the
fault likes not in our claims to knowledge, but in their epistemology.

Finally, there is this objection to the eliminativists. Suppose I play
along with their premises, and ask myself: Just what would be the best
explanation of why I and others are disposed to attribute to ourselves
experiences whose character, on reflection, enables us to engage in the
kind of conceptual exercises embodied in my blindsight
thought-experiment -- phenomenally conscious experience? One
explanation that at least has the merit of being straightforward is
this: we do indeed have these conscious experiences. That's why we think
we do. It is far from clear there is any credible alternative "error
theory" that does not attribute to us confusions of which we are not
guilty or assumptions we do not make.

I conclude that appeal to a positive account of first-person knowledge
does not bolster an eliminativist rejection of phenomenal consciousness.
However, I also do not suppose that the friend of consciousness can
happily stay silent about the distinctive nature of first-person
knowledge. So let me now turn to the next of Levine's concerns that I
mentioned at the start.


I agree with Levine's suggestion that there is an important
connection between an experience's being conscious and its availability
to a distinctively first-person knowledge. And I definitely believe that
a satisfactory philosophical discussion of consciousness needs to
account for this. I also agree that one needs to do justice to a certain
aspect of experience that is difficult to get a hold of, even in a
provisional way -- that which Levine expresses by saying a person's
conscious experiences are experience for that person. My own experiences
are experiences "for me," in some special sense that leads us to want to
speak of their being peculiarly subjective, and of a distinctive
"first-person perspective" on them. However, I find less surface appeal
than does Levine in the way higher order representation theories of
consciousness speak to these concerns, and I find much less alluring
than many do the perennially resurgent image of consciousness as a
theatre, an inner space where the mind's antics "present" themselves.
Nonetheless, I do agree that part of what gives higher order
representation theories their appeal, and what makes the theatre image
so difficult to lay to rest, is that they promise to say something
readily accessible to the intellect or imagination about these
fundamental but elusive aspects of consciousness -- aspects which,
regrettably, I was not able to address in my book.

Still, I steadfastly maintain: neither a higher order representation
theory of consciousness nor the theatre of consciousness metaphor
provides a good way of understanding either of these important aspects.
They illuminate neither the role of consciousness in first-person
knowledge, nor the subjectivity of consciousness. And I make no
concession to the idea that bad theories are better than no theories at

My reasons for rejecting higher order theories are given in Chapters 6
and 7 (and also I elaborate on them in my Psyche responses to Ludwig and
Lycan). I won't recapitulate those arguments here, though I might just
comment on Levine's complaint that I don't answer the higher order
thought theorist's argument that we are justified in positing the
occurrence of nonconscious higher order thoughts shadowing our every
experience, because "[w]hat we think of as conscious experience is
thereby explained."

My response to this (expressed -- admittedly rather tersely -- on p.
201) is that the claim is either question-begging or false. For what is
explained exactly by the positing of non-conscious higher order
thoughts? The fact that the states they target are conscious? I don't
see why one would think they explained this, unless one just assumed
higher order thought was necessary for consciousness. But that is (part
of) what is at issue. Maybe they are supposed to explain the
reportability of conscious states? But to explain this, we need not
postulate the occurrence of nonconscious thinking. We could instead say
that we generally form dispositional belief states about conscious
states, where such a belief is the capacity to think a conscious higher
order thought. To explain the capacity to report, one need posit nothing
more than the capacity to think the thoughts that would be expressed in
a report -- no actual occurrence of thought is needed.

But now: what do I have against the theatre metaphor? First let us ask
why the image of the mind's theatre is attractive. It seems we want some
image or analogy to help us explain why conscious states are available
to the reflection of the person whose states they are, in a way that are
not available to others. And so it is tempting to think of consciousness
as a special place, whose unity and boundaries help explain why things
in it are "observable" only from "inside." But if, like me, one admits
no "inner observation" but certain forms of thought or judgment about
one's own experience, what is wanted is an account of how first-person
judgment is related to experience that helps us understand the extent
and limits of its reach. But it is totally unclear how the proposal that
consciousness occurs in a certain literal or figurative place is
supposed to accomplish that. The theatre is, at best, a misleading
metaphor for what we want explained.

The question then arises: do I think there is a better way to explain
what first-person warrant and the subjective (or first-person)
"perspective" consist in? I think there is. But doing so is a big
project in its own right, one on which I have only begun, and of which I
can here give only a rough idea.

The key challenge to be met, I believe, is to explicate the special
relationship one's conscious experiences have to one another in virtue
of being phenomenally conscious -- what I would call the "phenomenal
unity of experience." Though it is, I believe, the phenomenal unity of
experience that one is tempted to represent as the unity of a place, as
in the theatre metaphor, the unity in question is not a unity of place.
And I would distinguish phenomenal unity from the notion of just being a
conscious experience attributable to a particular person. My approach
involves saying that there is a kind of unity among conscious
experiences that helps account for their special availability to
first-person reflection and for the distinctiveness of the first-person
point of view. But what kind of unity this is, is explicable (like the
notion of phenomenal consciousness itself) only through a use of example
and contrast that appeals to a first-person approach.

It is a challenge to clarify the relevant notion of unity sufficiently.
I have tried to make a start on that job (in "Self-Knowledge and
Phenomenal Unity," Nous 35:4 (2001) 542-568). I will not try to
recapitulate those thoughts here. But supposing I eventually manage this
task of clarification well enough, the next question is: how does that
sort of unity help us to understand the special scope and limits of
first-person judgment? I look for an answer in the following idea.
Underlying our first-person judgments of experience (e.g., 'I feel...',
'It looks to me...') is a capacity to form conscious demonstrative
thoughts both about and phenomenally unified with conscious experience
(e.g., "This is a feeling...', "This is its looking to me as if...').
What gives these thoughts their special demonstrative character, and
makes them distinct from first-person attribution of experience, is the
form of attention on which they are based -- a form that can link
experience only with thoughts that are phenomenally unified with the
experience. Because this way in which one's experience is available to
one's thought is made possible and delimited by the unity of phenomenal
consciousness, and because the special warrant enjoyed by first-person
judgment is explained with reference to that kind of availability,
ultimately first-person warrant and knowledge does get explained by
phenomenal consciousness. This style of account does not depend on, and
in fact repudiates any notion of consciousness as inner observation or
inner stage. On my view, talk of inner observation is a distorted way of
speaking of the kind of attention and judgment made possible by the
unity of consciousness. And talk of the mind's theatre is an ultimately
unhelpful, if imaginatively arresting, metaphor for that unity. On my
account this gives us a way of understanding, not only the
distinctiveness of first-person warrant, but also the subjectivity, the
"for me" quality of consciousness, the specialness of the first-person
point of view.

Let me say a little about this. A conscious experience is "for me"
because its character depends on its belonging to a certain phenomenal
unity of experience, the belonging to which is sufficient for being
mine. But I find it misleading to think of the special access of
first-person judgment to experience that is based on this unity of
consciousness as reflecting a special perspective on the experience.
Since we do not, in any substantive sense, perceive our own experiences,
they are not presented to us from anything like a certain "angle" or
"point of view." What makes the link between judgment and experience
special is not the perspective from which the former speaks of the
latter. Rather, this specialness of first-person reflection is to be
sought in the way judgment and experience together constitute a
single perspective or point of view, on account of their phenomenal
unity. These are only hints at how I can go about addressing the
challenge of accounting for first-person knowledge and the subjectivity
of experience. At least I want to have raised hopes that there are some
alternatives for us to explore, consistent with an embrace of phenomenal
consciousness, and a rejection of higher order theories and theatre


Now I will respond to the worries Levine raises regarding how
thoroughly and richly intentional the phenomenal character of experience
is. I claim that visual phenomenal features are intentional features of
a kind, and Levine finds this acceptable for the phenomenal character of
one's experience of space, partly since it is indeed hard to see how its
seeming as it does for it to look as if something is X-shaped could,
through some kind of reinterpretation, be made accurate or inaccurate,
through the presence or absence of, say, O-shaped things. But he worries
about my claim that the experience of color is inherently intentional.
One might, he points out, say that, e.g., the color green could look to
someone else the way red looks to me (and red could look to him the way
green looks to me), and yet both of us have "accurate" color experience.
It's just that the chromatic "look" that for me "stands for" green, for
this other person "stands for" red. We use, so to speak, a different
phenomenal vocabulary to visually represent the same objective colors.
And the regular causal connections that determine what color our color
experience stands for do indeed have the status of interpreting
conditions, imposing conditions of accuracy on an experience's
phenomenal character that it cannot furnish of itself.

My response? First, I want to concede that I do not pretend my
discussion of the intentionality of color experience (in section 7.7)
settles questions concerning whether the view of color and experience
just sketched or some other view of the matter is correct. My point was
simply this. Even if there is something about the intentionality of
color experience that its phenomenal character leaves undetermined, so
that this has to be added through the environment's "interpreting" the
experience, it is important to recognize that phenomenal character of
normal color experience cannot be broken down into a phenomenal spatial
component that is intentional, and a phenomenal chromatic component that
is not intentional. The reason is this. That different colors appear
differently distributed in space is inseparable from the way they look
-- from the way it seems to have such visual experience of color as one
has. The phenomenal character of the experience makes that experience
one of a color, which looks a certain way, occupying a certain place. It
is not, then, just an experience of a certain place and at the same time
an experience of a certain chromatic look. What this shows, I think, is
that the differences in the phenomenal character between, e.g., my
experience of a red equilateral triangle on a white background, and that
of a green equilateral triangle (with same size, orientation, location)
on a white background, constitute not only differences in the phenomenal
features I have, but in the intentional features I have. I experience
the space differently "chromatically filled" in the two instances, on
account of the phenomenal character of the two experiences. Now this
will, if the theory to which Levine alludes is correct, leave open the
matter of which color these phenomenally different experiences are to be
judged accurate or inaccurate of. On that theory, this would have to be
determined by the environment. But what is not left open is whether the
experiences are of different colors in certain places. That aspect of
the intentionality of color experience is already fixed by phenomenal
character. And, whatever one goes on to say about the nature of color
experience and how its accuracy conditions are determined, it needs to
be compatible with this. For if color A and color B (generally) look
different to x, x's experiences cannot "represent" things that "look A"
as having the same color as things that "look B."


Now the other domain where Levine worries that reliance of
first-person knowledge will not yield the kind of thesis I want relating
phenomenal character and intentionality is the domain of non-iconic,
conceptual thought. He objects that it is unclear to him, when
consulting his own experience, that there is some particular way of
seeming distinctive of, e.g., thoughts about Mondays, distinguishing
them from thoughts about Tuesdays.

I think that what I say about the phenomenal character of thought
probably does not commit me to asserting about Levine's experience what
he finds himself powerless to discover in it. Let me clarify what my
basic claim here is. Often we think, out loud or silently, and our
thinking does not consist (at least, not entirely) in the forming of
mental images (that is, what we are thinking about is nothing we either
visualize or otherwise form an image of). Now there are variations in
the way it seems to one to do such thinking that are not rightly
identified just with differences in the way it seems to us to sense or
image things, for they can be identified only by reference to what one
is non-iconically thinking, the "content" of one's thought, if you like.
To support this claim, I first invite one to recognize, through
first-person reflection, that the way it seems to one to understand and
follow what one (or another) is saying differs from the way it seems to
one to perceive what is said uncomprehendingly, or without following it:
this sort of difference is evidently not just a matter of a difference
in the phenomenal character of sense or imagery experience identifiable
without reference to one's understanding or meaning something by an
utterance. Also I invite you to acknowledge the occurrence of episodes
of sudden unverbalized (though verbalizable) non-imagistic thought. I
take both of these to illustrate the point that there are differences in
the phenomenal character of thinking that are not (or at least not
entirely) differences in the phenomenal character of imagery or sensing.
The next main point is that this non-imagistic, non-sensory difference
is not a single difference. There are distinguishable differences of
this sort in the phenomenal character of thought. The phenomenal
character of thought is continually changing, in a way not identifiable
other than relative to the expression of thought. These differences in
phenomenal character are manifest in the "experience of interpretive
switch" discussed on pp. 278-283. Contary to Levine's suggestion, these
differences are not subject to "wholesale reinterpretation, of the sort
that Quine, say, imagines." For the differences are only identified
relative to some "interpretation" -- some thought. We have no way of
conceiving of these shifts in phenomenal character as uninterpreted
utterances, merely formal items, or bits of symbolic behavior, to which
varying "contents" might be attached in radical interpretation, since
our only way of considering them is relative to such contents.

Notice, the view is not that you know what thought you're thinking by
recognizing its distinctive non-iconic phenomenal character. Certainly
no inference is involved. You identify the phenomenal character by the
thought you are thinking, in a sense. You conceptualize the phenomenal
character by thinking of it as the way it seems to think a certain
thought -- it is in such terms that you know what that way of seeming
is. But in my view there is no knowledge of either, independent of the
other, which serves as the basis of your knowledge of it.

Perhaps it will be clearer whether this conflicts with Levine's
convictions if we apply it to his "Monday/Tuesday" case. Suppose I am
writing a letter, or preparing to leave a voice message, and I need to
say what the day is. The thought occurs to me -- suppose I express it to
myself, silently -- "It's Monday." But then in the next moment I realize
my mistake, and correct myself, thinking, "No, it's Tuesday." So my
claim would be, first, that the way it seems to me to utter these
phrases to myself comprehendingly differs from the way it seems to me to
do so meaninglessly (as I would, for example, if I repeated them to
myself again and again, until they became to me "mere sound"). Second,
the way it seems to me to utter these phrases with understanding
changes: it is not as though there is some unvarying phenomenal
character of understanding that changes not at all with what is

I do not say this is adequately appreciated just by focusing on this
case, when it occurs. For it may not be evident that the phenomenal
change is anything more than might be distinguished by reference to the
different imaged sounds 'Monday' and 'Tuesday.' So the indirect stategy
I employ in the book, applied to this case, would involve appeal to the
occurrence of unverbalized non-iconic thought, and to what I call the
experience of interpretive switch. Here I suppose we could consider an
instance in which, after saying to oneself, "It's Monday,' the sudden
corrective realization that it is Tuesday is not verbalized. This can
happen, I think. And if it happens, there is a positive change in the
phenomenal character of experience that can't be pinned on any changes
in imagery or sense experience. We also could adapt the "interpretive
switch" point to this example. One might, it seems use the sound of
'Tuesday' to mean, not a day of the week, but something like, 'the day
of Tou' -- where 'Tou' is a proper name. (So maybe it is Mr. Tou's turn
today to take out the garbage: it's Tou's day.) Then when someone says,
'It's Monday -- no, it's Tou's day,' we can imagine an experience
interpretive switch, thinking first of the third day of the week, then
of the day on which it is our friend Tou's turn. Assuming you grant all
this, I would argue that such change in phenomenal character is not
confined to these cases where its distinctness from merely sensory or
imagistic differences in phenomenal character are most directly evident.
For that, I believe, can be shown to have unacceptable consequences (pp.

I am not sure what in this Levine would reject or balk at. I suspect
that what he hesitates to accept is not the phenomenology just
indicated, but something else. I suspect what bothers him is the idea
that we can, through first-person reflection, justifiably claim to
discern a unique phenomenal character specific to all and only "Monday"
thoughts. But I think I can honor this skepticism. For I think I can say
what I want to say, that episodes of thought differ in phenomenal
character in ways distinguishable only by reference to what one is
thinking, while allowing also that these episodes cannot be broken up
into phenomenal bits, corresponding to the words with which they are
expressed, such as 'Monday,' each of which have a separately
re-identifiable phenomenal character. The phenomenal character of
thought does not resolve into recognizable phenomenally distinguishable
units corresponding to the words with which the thoughts are expressed,
something on the order of Lockean "simple ideas." But that is consistent
with what I have to say in Chapter 8.

Also I would like to accept that there are considerable disanalogies
between the phenomenology of thought I am now discussing, and the kind
of variation in phenomenal character we are likely to fasten on
initially in sense experience -- such as differences in the way colors
look, tones sound, or bodily sensations feel. These latter differences
admit of certain sorts of stable subjective comparative judgments. We
are able to answer questions such as: 'Does red look more like orange
than blue?', 'Does this tone sound lower than this one that the previous
did?', 'Does this current pain feel more or less intense than the
previous one?' By contrast, it seems absurd to ask whether Monday
thoughts seem more like Tuesday thoughts than Wednesday thoughts, and so
on. But that is fine with me; I don't think that differences in
phenomenal character need to lend themselves to this sort of question --
even if many do. After all, even readily appreciable differences in
phenomenal character of visual experience don't seem amenable to these
kinds of judgments. Think of Gestalt shifts in visual experience. Can we
sensibly ask whether seeing the duck/rabbit as a duck seems more like
seeing the vase/faces as a vase or more like seeing it as faces in

So I want to suggest that resistance to what I have to say about the
phenomenology of thought may come from sources which, properly exposed,
can be seen to provide no real objection to it. I believe that, once we
clearly detach the notion of differences in phenomenal character from
the empiricist tradition's conception of phenomenal simples -- ideas or
impressions -- and if we avoid assuming that the kinds of comparisons to
which some differences in phenomenal character are amenable must be
applicable to all, my claims about the phenomenology of thought will
seem much less objectionable.


Let me now summarize my response to the challenges Levine raises.
First, there is the question of whether I can get away with introducing
and relying on a distinctive first-person warrant for my judgments about
attitudes and experience, while leaving open just what positive account
of the mind's self-knowledge is to be preferred. I think it is actually
desirable to avoid presupposing any such epistemological view at the
start of inquiry, and I deny that this attitude makes my account
vulnerable to eliminativist attack. The eliminativist asks us to assess
our warrant for attributing phenomenal experiences to ourselves
ultimately on the model of our assessment of theoretical postulates or
explanatory posits in scientific theorizing. But why should we adopt
this epistemology? Support may be sought in something like Rey's "dream
place" analogy, but it would be a mistake to suppose such an analogy
enjoys some presumptive favor. We can draw analogies differently, and in
ways counter to the eliminativist's case, with at least as much, if not
more, plausibility. And, the analogy seems to break down when one
considers the importance, in understanding the notion of phenomenal
character, on identification of illustrative instances in the
first-person case, without adverting to other putative data to be
explained by them. So it seems eliminativists are driven to seek another
line of attack. They must argue that, because I determinedly leave open
the possibility of a person such as my blindsighted Belinda, my
conception of phenomenal features suffers from some kind of incoherence,
and that ultimately neither I nor anyone who thinks along my lines knows
what we are talking about. The prospects of a case for this harsher line
are unpromising.

Furthermore, we need to consider whether an epistemological theory's
failure to supply us with knowledge of the phenomenal character of
experience should cast doubt, not on the reality of phenomenal
consciousness, but on the adequacy of that theory of knowledge. And we
should ask whether our tendency to think we have phenomenal experience
is not rather better explained by saying we actually do have it, than by
a rival "error theory."

I do acknowledge the need to build on what I have said in the book about
phenomenal consciousness to provide an account of how it is involved in
the special availability of experience to first-person reflection and
the special warrant enjoyed by first-person judgments. Here I can just
say: I have made a start on this at least. I'll keep working on it.

As to Levine's worries about my treatment of
intentionality-phenomenality connections, I think my claims here will
seem less objectionable once their character and scope are properly
understood. My remarks on the way chromatic experience is infused into
the phenomenal character of spatial vision were not meant to settle the
vexed issues to which Levine alludes, regarding the accuracy conditions
and objectivity of color experience. But I believe they set important
constraints on what will constitute an acceptable answer -- constraints
that may prove crucial. Finally, I would say, if one doubts my remarks
about the phenomenology of thought, one needs to take account of the
actual route of reflection that led me to them and attempt to retrace it
oneself, before declaring one cannot find the way to my destination. I
ask that one take care not to confuse my claims about thought with
certain others, however much they might be suggested by famous
philosophers of the past.


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

Reply all
Reply to author
0 new messages