On Gilbert Harman’s The Intrinsic Quality

of Experience
Tim Klaassen

I.Introduction
Gilbert Harman’s article The Intrinic Quality of Experience (Harman, 1990)
occupies the 94th place of the 100 most cited articles in the philosophy of
mind in the last couple of decades.1 This is admirable since the number
of article entries on Mindpapers is already 28490.2 In the article, Harman
argues against the thesis that we are directly aware of intrinsic qualities
of experience. This thesis, if true, poses a serious objection to functional
accounts of the mind. For it is said that the phenomenal character of an
experience is fundamentally shaped by such intrinsic qualities, and that no
functional account can be given of these particular phenomenal elements
of experience. By showing that, in fact, we are not aware of such intrinsic
qualities, Harman hopes to show that functionalism can provide an adequate
account of the mind after all.
Although Harman does not use this terminology himself, his article is
usually understood as arguing for what has come to be known as the so-called
“transparency of experience”, or the view that experience is “diaphanous”.3 .
Also, the view that Harman presents in his article has come to be associated
with views of philosophers such as Fred Dretske (Dretske, 1995) and Michael
Tye (Tye, 1995) who maintain that phenomenal character is nothing but a
species of representational content. The transparency of experience is usu-
ally put forth as one of the main motivations for this thesis about phenome-
nal character. The view that phenomenal character is nothing but a species
of phenomenal character has come to be known as “representationalism” or
“representationism”.4
Be that as it may, it is not quite clear to me what exactly it is that
Harman is arguing for, and what it is that he is arguing against. Also, I am
not sure what his arguments are, and whether they are sound. Although
Harman presents his statements as if they are all part of one big argument;
1
“100 most cited works by philosophers in MindPapers according to Google Scholar”,
http://consc.net/mindpapers/sreq/most_cited_phil.html, accessed on 17-01-2014
2
“MindPapers: Contents”, http://consc.net/mindpapers, accessed on 17-01-2014
3
See e.g. (Tye, 1995), (Stoljar, 2004), and (Bourget & Mendelovici, 2013)
4
Ned Block uses this second term, see e.g. (Block, 2003)

1
upon closer consideration the feeling cannot escape that a lot remains un-
clear. This partly has to do with certain terminological issues that need to
be resolved. But besides that, although Harman makes a lot of individual
statements that I am in agreement with, I am not convinced that they are
sufficient to paint the overall picture that Harman wants them to paint.
Hence, this paper has two aims. The first is to identify and resolve certain
terminological difficulties that stand in the way of a proper evaluation of
Harman’s views. My second aim is to show that Harman’s arguments fail
to give an adequate response to the objection that phenomenal character
resists functional explanation.

II. A short exposition of Harman’s arguments

As the title of this section indicates, I would like to begin with a short
exposition of Harman’s arguments. This will enable us to get before us
Harman’s own wording of things. It will also give us some overview of the
issues that we are dealing with. The reason I want to take this approach is
that many of the key terms in the philosophy of mind, such as e.g. “qualia”,
“experience”, “representation” etc. do not have univocal meanings. And the
best way to determine what Harman is likely to mean by certain commonly
used terms is to see how he uses them himself. When one reads the literature
on the philosophy of mind, one really gets the impression that a univocal
terminology is lacking. Or, at the very least, it seems to me that very
often authors take insufficient care to articulate the meanings of the terms
they use. However, this is absolutely required when it comes to meaningful
(philosophical) dialogue. Of course I do not have the pretension that, in this
paper, I am going to solve all these terminological ambiguities. Rather, I
merely hope that, by taking these conceptual\terminological difficulties into
account, we will be led in the end to more clarity.
In the abstract of his article Harman indicates that there are three ob-
jections against psychophysical functionalism but that these objections can
be resolved by noting that they fail to distinguish between the properties of
an experience and properties of the object of an experience. We shall only
be interested here in the first of these objections and the manner in which
Harman tries to deal with it. According to Harman, the objection begins
with the premise that we are directly aware of qualities of experience that
are intrinsic to it. Since functional definitions of mental states are formu-
lated in terms of a mental state’s causal relations to other mental states,
sensory inputs, and behavior; the argument concludes from this that no
functional account can be given of the qualities because, by definition, they
are intrinsic. He cites Thomas Nagel as one of the proponents of this argu-

2
ment.5 However, as Harman points out, this objection fails to distinguish
between the properties of an experience and the properties of the object of
an experience. According to Harman, as long as we keep an eye on this
distinction, we shall have to conclude that we are not in fact aware of any
intrinsic qualities of experience. And because the objection to functionalism
depends on the truth of this premise, the initial objection to functionalism
does not actually pose a threat.
Harman starts his argument with the supposition that experiences have
intentional or representational content. According to Harman this means
that experience represent things as being a certain way. Perhaps this sound
as a rather trivial proposition. However, an important qualification of the
proposition that experiences have representational content is that this in-
volves that an experience has the possibility to misrepresent things. That is,
experiences may convey to us things that are false. This happens in cases of
perceptual illusions and hallucinations. It should be noted that this propo-
sition is a lot less trivial. For one could also suppose that cases of perceptual
illusion or hallucination are not so much cases of having wrong experiences,
but rather having wrong believes about experiences. One way then to ex-
press the idea that experiences have representational or intentional content
is to say that already in and of themselves experiences are truth-conditional.
In addition to expressing his commitment to the thesis that experiences
are intentional, Harman gives other examples of intentionality. He mentions
the story of Ponce de Leon (a 16th century Spanish explorer) who was on a
search in Florida to find the Fountain of Youth. The interesting thing about
this story, is that it shows that we can search for things that do not exist.
After all, for all we know, there is no such thing as the Fountain of Youth.
But as Harman points out, this does not mean that Ponce de Leon was not
actually searching for something. After all, Ponce de Leon was searching for
something. However, what he was searching for simply happens not to exist.
In addition, Harman states that: “We can therefore say that his search had
an intentional object. But the thing that he was looking for, the intentional
object of his search, did not (and does not) exist” (Harman, 1990, p. 34).6
Another example that Harman considers are paintings. As an example
he takes a painting of a unicorn. For all we know, unicorns do not in actual
fact exist. However, this does not mean that the painting is not a depiction
of a unicorn. That is, the fact that no unicorns actually exist does not
prevent the painting from having a certain “intentional content” (Harman,
1990, p. 34). Hence, the fact that a painting has the intentional content that
it does is quite independent from whether or not the object that it depicts
actually exists.
Harman then proceeds to another philosophical insight, an insight which
5
See (Nagel, 1970). Levine makes a similar point (Levine, 1995)
6
We will come back to Harman’s phraseology here later.

3
I take to be rather different from the insight that, apparently, some things
can represent other things that do not exist. He points out that imagining
a unicorn is usually understood as being similar to a painting of a uni-
corn. After all, both seem to have the same non-existent intentional object.
However, and this is Harman’s crucial point, imagining a unicorn does not
involve any awareness of a mental picture of a unicorn. According to Har-
man, this is plain from the fact that imagining a unicorn is not the same
as imagining a picture of a unicorn; they are two acts of imagination with
different intentional objects.
It is in connection with this that Harman points out the importance
of distinguishing between the “properties of a represented object and the
properties of a representation of that object” (Harman, 1990, p. 35). To
make his point, Harman draws attention to the fact that although a unicorn
supposedly has legs and a horn, neither a painting of a unicorn, nor the act
of imagining one, has legs or a horn. Conversely, Harman points out that
although a painting is flat and covered with paint, and imagining a unicorn a
mental activity, a unicorn neither is flat and covered with paint nor a mental
activity. These examples then make clear, on Harman’s account, that there
is a distinction between the properties of a representation and the properties
of a represented object.
Harman thinks that a similar distinction applies to experience, and that
especially the argument from illusion, which is often used to motivate the
sense data theory, fails to honor it. The conclusion that Harman draws
from this is that the argument from illusion is false, and that therefore it
cannot be used in support of the sense data theory. The argument from
illusion starts from the premise that the way things appear to us may not
be the way things really are, such as in cases of illusion or hallucination.
The argument then concludes that, because what appears to us in these
cases are not external physical objects, what appears to us must really be
something “mental or internal” (Harman, 1990, p. 35).
It is the inference from ‘what appears is not external’ to ‘what appears
must be mental or internal’ that Harman thinks is especially fallacious. In
fact, according to Harman it is precisely at this point that the argument from
illusion makes the mistake of not distinguishing between the properties of
an experience and the properties of a represented object of experience.
To demonstrate his point, Harman uses Ponce de Leon’s search for the
Fountain of Youth and the painting of a unicorn as analogous examples. He
claims that an argument about paintings analogous to the argument from
illusion would infer that, because unicorns do not actually exist, the painting
must really be depicting something mental (“for example, the painter’s idea
of a unicorn”) (Harman, 1990, p. 35). Harman claims however that this
inference is fallacious, and that this becomes obvious as soon as we apply
the same trick to the case of Ponce de Leon:

4
 Ponce de Leon was searching for the Fountain of Youth. But
there is no such thing. So he must have been searching for some-
thing mental.” This is just a mistake. From the fact that there
is no Fountain of Youth, it does not follow that Ponce de Leon
was searching for something mental. In particular, he was not
looking for an idea of the Fountain of Youth. He already had the
idea. What he wanted was a real Fountain of Youth, not just
the idea of such a thing.”(Harman, 1990, p. 35-36)

This then is Harman’s criticism of the argument from illusion. Harman

argues that it is wrong to suppose that, because the Fountain of Youth does
not exist, Ponce de Leon was searching for something mental.7 Similarly
then, it is equally wrong to suppose that in cases of illusions or hallucinations
we are aware of something mental as the argument from illusion wants us to
believe. According to Harman, the argument from illusion is simply a bad
argument.
Remember that Harman’s point is to show that we are not in fact aware
of any intrinsic qualities of experiences, so as to counter an important objec-
tion to functionalism. Harman thinks that with the points he has just made,
he has all the resources to do this. How does he think he can accomplish
this?
First of all (as is indicated by the previous passage), Harman equates the
conclusion of the argument from illusion with the thesis that we are aware of
intrinsic qualities of experience. In other words, Harman thinks that ‘being
aware of something mental or internal’ is really the same thing as ‘being
aware of an intrinsic quality of experience’. Next, because Harman believes
he has shown that the argument that has been used in support of this thesis
is false, he concludes that the thesis that we are aware of intrinsic qualities
of experiences must also be false.
To be sure, Harman connects all this with the distinction between the
properties of a representation (or an experience, or an intentional act, etc.)
and the properties of the object that is represented. Also, in addition, he
also argues for this by an appeal to introspection:

When Eloise sees a tree before her, the colors she experiences are
all experienced as features of the tree and its surroundings. None
of them are experienced as intrinsic features of her experience.
... Look at a tree and try to turn your attention to intrinsic
features of your visual experience. I predict you will find that
7
Whether or not it is really obvious that it is wrong to suppose that Ponce de Leon
was searching for something mental, we shall go into in the next section. Perhaps Harman
thinks this is obvious given that, were a contemporary of Ponce de Leon to ask him whether
he was searching for something mental, Ponce de Leon would probably have answered in
the negative (expect of course, if Ponce de Leon considered himself a kind of Berkelean
idealist!).

5
 the only features there to turn your attention to will be features
of the presented tree...(Harman, 1990, p. 39)

Hence, we must say that Harman really gives two arguments against the
view that we are directly aware of intrinsic qualities of experiences. First
Harman argues that the argument form illusion, which is used to support
this thesis, is based on a fallacious inference. Second, Harman claims that
if we introspect our experiences, we shall see that all that we find are quali-
ties objects are represented as having and not any intrinsic qualities of the
experiences themselves. It seems then that Harman appeals to both logical
and phenomenological grounds to make his case.
To conclude, because Harman believes himself to have shown that we
are not in fact aware of any intrinsic qualities of experience, functionalism
is saved from one serious objection to it. After all, if we are really aware of
intrinsic qualities of experience, then functionalism cannot properly account
for it. However, because the antecedent of this objection is nolw shown to
be false, functionalists do not need to worry that they should account for
something that they cannot in fact account for. Ergo, the objection does
not pose a threat to functionalism.

III. Terminological Difficulties and Doctrinal Am-

biguities
When I reflect upon the contents of Harman’s article, several questions enter
my mind. These questions spring on the one hand from certain doubts
about terminological issues that, after reading Harman’s article, I am left
with. On the other hand, they spring from worries about the soundness of
the arguments themselves. Related to this is also that I am just not sure
which view it is that Harman is arguing against. Of course, no doubt these
difficulties are all more or less related to each other.
Let me begin with some of the terminological difficulties that I have. To
be sure, these difficulties do not so much arise out of Harman’s article taken
by itself as they arise within the bigger context of all the different terminolo-
gies that are used by philosophers in dealing with the issues of contemporary
philosophy of mind. As I have already indicated in the previous section, it is
my impression that usually no two philosophers mean the same thing when
they make use of one and the same term. Perhaps resolving such termi-
nological issues is not very exciting task. However, I believe everyone will
agree that the settlement of such issues is absolutely required.
One terminological aspect that confuses me is that Harman’s account is
usually put under the header of “representationalism” and that his argument
is understood to argue for the thesis that experience is “transparent” or

6
“diaphanous”.8 My confusion has to do with the following. One central
theme in Harman’s article is his opposition, to the sense data theory of
experience. And he does this, in part, by an appeal to the transparency
of experience. However, the early 20th century philosopher G.E. Moore is
usually accredited as being the first to have emphasized the diaphanousness
of experience.9 But G.E. Moore was in fact himself a sense data theorist!
Second, in Perception: A Representative Theory by Frank Jackson (Jackson,
1977, Ch. 6) we find a (relatively) contemporary defense of the sense data
theory which also labels itself as “representational”.10 I hope the point that
I want to make is clear; the dichotomies in which Harman speaks (or at least
are now commonly attributed to him) do not seem , on the surface at least,
to be recognized by those he opposes himself to.
Before we can examine Harman’s arguments it is necessary that termi-
nological issues such as these are resolved. Fortunately, I have found some
answers to these terminological difficulties in the literature.
Let me begin by addressing the difficulty of accounting for the sense data
theorist’s and Harman’s shared commitment to the thesis that experience
is diaphanous. Daniel Stoljar (Stoljar, 2004) proposes a solution to this
difficulty.11 First, Stoljar indicates that Harman’s arguments are aimed at
refuting qualia realism. However, Stoljar points out that there are in fact
two kinds of qualia realism12 , and that the argument from diaphanousness
really is only directed at one of these. He explains that the two kinds
of qualia realism differ with respect to their commitment to what he calls
the “relational thesis”. According to the relational thesis, the phenomenal
character of an experience (i.e. “what it is like” to have the experience
(Nagel, 1974)) is wholly determined by the objects that one is related to
in having the experience.13 Now the first kind of qualia realism affirms the
relational thesis, while the other one denies it. The view that is proposed by
the sense data theory is an example of the first kind, while adverbial theories
(Chisholm, 1969) and more contemporary views such as the ones that can
be found in (Shoemaker, 1990) and (Block, 2003) are examples of the second
kind. It is only this second kind of qualia realism against which the argument
of diaphanousness is aimed. The diaphanousness of experience supposedly
shows that there is nothing over and above the objects of experience that
determines phenomenal character.
What then about the terminological issue that springs from the fact that,
8
See e.g. (Shoemaker, 1990), (Dretske, 1995), (Tye, 1995) and (Tye, 2013), (Kind,
2003), and (Stoljar, 2004).
9
See his The Refutation of Idealism (Moore, 1903)
10
It should be noted however that Jackson no longer defends this theory (Jackson, 2000)
11
An article by Tim Crane was also very helpful here (Crane, 2002)
12
Or to put it in the kind of terms that we were already using; Stoljar points out that
there are two ways of construing the notion of an ‘intrinsic quality of experience’. From
now on I’ll use these terminologies interchangeably.
13
This is also called the act-object model of perception. See e.g. (Fish, 2010, p. 16)

7
in the contemporary debate, ‘representationalism’ is usually contrasted with
theories that defend qualia realism (such as the sense data theory) and the
fact that, at the same time, sense data theorists such as Jackson label their
theories as ‘representational’ accounts of experience?14
An answer to this question can be found in an article by Frank Jackson
in which he himself rejects his former sense data theory in favor of “rep-
resentationalism” (Jackson, 2000).15 In the article he states “the reason I
abandoned the sense datum theory was my belated realisation that it fails
to capture the representational nature of perceptual experience”. In the ar-
ticle he makes clear that what he means by the “representational nature of
perceptual experience” is the idea that phenomenal experience is inherently
truth-conditional (a notion that I already elaborated on in the previous sec-
tion). This implies then that when Jackson put his sense-data theory under
the banner of “representationalism” what he had in mind was precisely not
truth-functionality but something else! Cleary, Jackson thinks that in one
important sense his sense data account of perception was representational
while in another very important sense it is not. How should we account for
this ambiguity?
I believe Brad Thompson (Thompson, 2008) gives a useful answer to
this particular question. He points out that we should distinguish between
‘content-based representationalism’and ‘vehicle-based representationalism’.
Content-based representationalism is the view that phenomenal qualities
are properties that are represented. On the other hand, vehicle-based repre-
sentationalism is the view that phenomenal qualities are the vehicles that do
the representing. The distinction may be illustrated as follows. According to
content-based representationalism “dog” represents a particular pattern of
ink, while according to vehicle-based representationalism “dog” represents
the animal that is man’s best friend. While sense data theories are essentially
committed to vehicle-based varieties of representationalism, contemporary
representational accounts of phenomenal experience, such as those of Har-
man, Dretske and Tye, are essentially content-based. It is because of this
content-based variety of representationalism that take themselves to be able
to account for phenomenal character in truth-conditional terms.16
14
See (Price, 1950) and (Robinson, 1994) for two other examples of authors that seem to
hold that the sense data theory is perfectly compatible with ‘representationalism’. Also,
to be sure, in this sentence I refer to what is now in the contemporary debate called ‘strong
representationalism’. See (Chalmers, 2004)for an overview of some of the contemporary
varieties.
15
See also e.g. (Byrne, 2009)
16
Consider the following two linguistic expressions: (a) ‘Amsterdam is the capital city of
the Netherlands’, and (b) ‘Amsterdam vletc stoliceĭ Niderlandov’. Now suppose
both (a) and (b) are two different phenomenal experiences. According to vehicle-based
representationalism both (a) and (b) would represent the same content, viz .that Amster-
dam is the capital city of the Netherlands. According to a content-based account however
both (a) and (b) would represent entirely different things (e.g. a Latin scripted English

8
 To summarize, the terminological difficulties that I have indicated can
be resolved by noting that there are two kinds of qualia realism and two
kinds of ‘representationalism. The two kinds of qualia realism are distin-
guished by their attitude to the thesis that the phenomenal character of
an experience is entirely determined by the kinds of objects to which the
experience is related. The kind of qualia realism that the sense data advo-
cates affirms this thesis, while the other kind of denies it. Furthermore, the
ambiguity of representationalism can be dealt with by making a distinction
between a content-based variety and a vehicle-based variety. Contemporary
representationalists such as Harman advocate the first variety, while sense
data theorists commit themselves to the second.
We are now in a better position to evaluate Harman’s position and the
arguments that he uses. First of all, we should remind ourselves that Harman
argues against the view that we are aware of intrinsic qualities of experience.
Now, as I have pointed out there are two ways to construe the notion of an
intrinsic quality of experience. Hence, the question is; which of these is
Harman arguing against? Unfortunately, Harman is not sufficiently clear
about this.
One the one hand, Harman says that he opposes himself to the sense data
theory. But he does this in part by appealing to the fact that when we try
to attend to our experience of things, all we notice are objects of experience
but never the experience itself. However, on the present account, the sense
data theory is in perfect agreement with this, for it affirms the relational
thesis according to which the phenomenal character of an experience is fully
captured by the objects that are represented in it. Perhaps then it is more
plausible to take Harman as arguing against the other type of qualia realism
according to which there is more to a phenomenal experience than just the
objects the experience is of.17
However, there are perhaps other reasons for holding that Harman really
is arguing against the sense data theory. Perhaps his attack is not so much
directed at the kind of qualia realism that is contrary to the diaphanous-
ness of experience (after all, I have shown that this kind of qualia realism
is not actually part of the sense data theory), but to the fact that the sense
data theory proposes a representational account of experience that is essen-
tially vehicle-based. According to vehicle-based representationalism we are
only indirectly or mediately aware of external objects via other intermediary
expression versus a Cyrillic scripted Russian expression). Note that on a vehicle-based
account it seems especially puzzling why (a) and (b) should appear the way they do. After
all, it is entirely arbitrary that the English use something like (a), and that the Russians
use something like (b), to express that Amsterdam is the capital city of the Netherlands.
Misguided scientists might even argue that there is an explanatory gap between (a) and
(b) qua patters of ink and the fact that Amsterdam is the capital city of the Netherlands!
From a content-based point of view however, the fact that (a) and (b) appear the way
they do is simply because, as a matter of fact, patterns of ink just look like that!
17
Stoljar gives a similar account (Stoljar, 2004).

9
objects. The argument from illusion is used to demonstrate this. Vehicle-
based versions of representationalism want to “reify appearances” (Jackson,
2000, p. 2). Vehicle-based representationalism is essentially committed to
what Howard Robinson calls the “phenomenal principle”. According to the
phenomenal principle, when something appears to a subject with property
P, it follows there actually is something that is P, even in cases of illusion
and hallucination (Robinson, 1994, p. 32). Vehicle-based representational-
ism can allow for this because it holds that phenomenal properties simply
are representational vehicles and not representational contents (which are
truth-functional). According to Harman however, this reification of appear-
ances is misguided, for it appeals to the argument from illusion (which is
fallacious).
Perhaps we should put the whole story as follows. Harman really gives
two arguments.18 The first argument argues that the argument from illusion
is fallacious, and that the vehicle-based representationalism that is built on
it is misguided. Here Harman is really arguing against the sense-data theory.
Harman’s second argument however, by an appeal to the diaphanousness of
experience, argues that all we are aware of in experience are the objects
our experience is of. Here, although he does not seem aware of this him-
self, Harman is not so much arguing against the sense-data theorists as he
is arguing against the qualia realists that deny the relational thesis. Of
course, Harman himself does not explicate this as such. However, he ought
the have, for his claim that his main target is the sense data theory is
conceptually inadequate. In any case, it seems to me that, in light of the
terminological\doctrinal issues that I have addressed, this reconstruction of
Harman’s arguments is the most appropriate and\or adequate.

IV. Examining the arguments

I would now like to proceed to an actual examination of Harman’s argu-
ments. I intuit that they may not be sufficient to save functionalism from
the objection that it cannot account for the phenomenal character of our
mental states. Of course this is not to say that functionalists may not have
other resources to counter this objection. All it says is that to the extent that
a functionalist appeals to the kind of arguments that Harman puts forth,
the functionalist fails to counter the objection that phenomenal character
cannot be fully functionalized. In other words, I am not so much arguing
against functionalism here as I am arguing against Harman’s attempts to
save functionalism from one serious objection to it.
Let me begin by addressing Harman’s criticism of the argument from
illusion. First of all, it should be pointed out that in the context of evaluating
18
I already indicated this somewhat in the previous section

10
objections to functional accounts of the mind, it seems a bit odd, initially
at least, that Harman should want to focus so much of his attention on
the argument from illusion. For the argument from illusion is first and
foremost an argument in favor of indirect realism. And when you read
the literature on functionalism it seems that the epistemological problem
of indirect realism does not really seem to play any significant role. So
why, when it comes to the problem of functionalizing phenomenal character,
would you want to pay attention to the argument from illusion? Moreover,
the argument from illusion seems especially concerned with accounting for
the problematic circumstance thatthings appear to us that turn out not to
exist. It seems to me however that the issue of functionalizing phenomenal
character has more to do with how things phenomenally appear to us; i.e.
with the issue of why mental states are phenomenally characterized the way
they are (e.g. why pains feel the way they do).
Perhaps I misunderstand the role that Harman’s attack on the argument
from illusion plays in his overall attempt to defend functionalism. Perhaps
Harman only uses the argument from illusion to illustrate the kind of faulty
reasoning that the thesis that we are aware of intrinsic qualities of experience
is also guilty of making. In other words, perhaps Harman merely uses the
argument from illusion as an example alongside the other examples of Ponce
de Leon and paintings of unicorns to show that it is wrong to suppose that we
are aware of intrinsic qualities of experience. However, to interpret Harman
in this way, I think, would be mistaken. For according to Harman, the
argument from illusion tries to show that we are only indirectly aware of
external objects by being directly aware of sense data, and he equates being
aware of sense-data with being aware of intrinsic qualities of experience. In
other words, Harman does seem to think that the argument from illusion is
somehow relevant to the issue of functionalizing phenomenal character.
In any case, it is not obvious to me why the argument from illusion should
bear any direct relevance to phenomenological issues of functionalism. But
even if we grant Harman this, I am not sure if his criticism of the argument
from illusion is convincing. More specifically, I doubt whether his criticism
is really sufficient to refute the thesis that we are aware of intrinsic qualities
of experience.
Harman argues that from the fact that the Fountain of Youth does not
exist, it does not follow that Ponce de Leon really was searching for some-
thing mental. True, but it also certainly does not follow that the Fountain
of Youth is not in fact something mental! It seems to me that Harman over-
looks the circumstance that his examples have intensionality. Or at least,
it can be argued that a function such as “Ponce de Leon is searching for x”
ought be treated as intensional. Suppose Ponce de Leon was searching for
the Morningstar, and suppose that he did not know that the Morning Star
is the Evening Star. Then on an intensional reading of “Ponce de Leon was
searching for the Morning Star” it certainly does not follow that Ponce de

11
Leon was searching for the Evening Star, but it also does not follow that
the Morningstar is not in fact the Evening Star.19 As a matter of fact, the
Morningstar is the Evening Star. This then shows that, although Ponce de
Leon was not searching for something mental, this does not preclude the
possibility that the Fountain of Youth is not in fact something mental. In
a similar way can allow for the possibility that illusions and hallucinations
make us aware of something mental.
I am not sure how serious this objection is. But then again, I am also not
that sure why Harman thinks Ponce de Leon is not searching for something
mental. It seems to me that the only reason for this, is that if a contemporary
were to ask Ponce de Leon if he was searching for something mental, he would
have answered in the negative. But we should account for this in intensional
terms. And as I think I have shown, such an intensional account does not
at all show that the Fountain of Youth is not in fact something mental. In
addition to this, I want to point out that Harman’s criticism can also be
applied to some of the things that he himself says. He says: “In fact, there
is no such thing as the Fountain of Youth, but that does not mean Ponce de
Leon wasn’t looking for anything. We can therefore say that his search had
an intentional object. But the thing that he was looking for, the intentional
object of his search, did not (and does not) exist” (Harman, 1990, p. 34).
According to Harman, Ponce de Leon’s search had an intentional object.
However, by Harman’s standards, I would suppose that it would be equally
wrong to say that Ponce de Leon was searching for an intentional object!
Again however, as I have tried to show, from this it does not follow that
the Fountain of Youth is not in actual fact an intentional object.
Although I think there is more to be said about Harman’s evaluation
of the argument from illusion, I want to conclude with a difficulty that I
have with the phenomenological part of Harman’s argument. As I have
pointed out, even if Harman has satisfactorily showed that the argument
from illusion is fallacious, I don’t believe this sufficiently shows that we are
not in fact aware of intrinsic qualities of experience (remember, Harman
thinks that ‘being aware of something mental’ is equivalent to ‘being aware
of an intrinsic quality of experience’). One could therefore ask whether there
are maybe other considerations that favor the view that we are not in fact
aware of such qualities. Now Harman could point out that if we try to find
these intrinsic qualities of experience through introspection, it appears that
such qualities are nowhere to be found. Experience is diaphanous, and the
only thing that we are aware of are the object represented by the experience.
Now I have pointed out that there are in fact two forms of qualia realism;
i.e. two ways of construing the notion of an intrinsic quality of experience.
And I have pointed out that the argument from diaphanousness is only di-
19
Similar points have been put forward by (Smart, 1959), (Brandt & Kim, 1967), and
(Churchland, 1985)

12
rected at one of these. The argument from diaphanousness only attempts to
show that there is nothing more to the phenomenal character of an experi-
ence than the objects that are represented in it. However, this is compatible
with the kind of qualia realism that the sense data theorists commits himself
to. This other form of qualia realism is equally threatening to the prospect
of giving a functional account of the mind. And as far as I can see, none of
Harman’s arguments show that this kind of qualia realism does not in fact
hold true. Let me explain.
The content-based representationalist wants to solve the problem of phe-
nomenal character by holding that phenomenal qualities are the qualities ob-
jects are represented as having. Or to put it differently, the content-based
representationalist wants to solve the problem by “kicking the phenomenal
character downstairs, into the external world” (Shoemaker, 2003, p. 256).
However, the kind of qualia realism that is advocated by the sense data
theory holds that certain phenomenal qualities only exist to the extent that
they are actually being experienced. In other words, although it grants that
there is nothing more to the phenomenal character of experience than the
objects the experience is of, it also holds that at least some of these objects
are entirely mind-dependent. If there are indeed such mind dependent phe-
nomenal objects, then the content-based representationalist has a problem,
because it would seem that phenomenal character cannot be entirely kicked
downstairs into the external world.
I believe the argument from diaphanousness does not show that certain
phenomenal objects are not in fact entirely dependent on the mind. When
I focus my attention on my lips, I feel a certain kind of bodily sensation
whith its own unique phenomenal character. Now when I try to attend to
my experience of the sensation that I feel in my lips, all I end up with is
the sensation itself, which is the object of my experience. This is the point
that the argument from diaphanousness wants to make. However, when I
try to conceive of the peculiar sensation that I feel in my lips apart from my
experiencing of it, I must equally well conclude that I do not know how to
proceed. The same goes when I try to conceive of an unfelt pain; I cannot
conceive of such a thing. Both the sensation that I feel in my lips and
the particular painfulness of a pain each have their own kind of phenomenal
character. However, it seems to me that these phenomenal characters do not
exist, are not instantiated, apart from actually being experienced.20 And to
the extent that we might properly call such phenomenal characters intrinsic
qualities of certain experiences, we must conclude that perhaps we are aware
of intrinsic qualities of experiences after all.21
20
Searle also makes this point about pain (Searle, 1995).
21
We are indeed justified to count this as intrinsic qualities of experience, even though
they really concern the objects of the experience. According to some uses of the word,
‘experience’ can also be identified with the object one is related to in having the experience
(Stoljar, 2004)

13
 Harman’s arguments do not show that certain phenomenal characters are
not, as a matter of fact, only instantiated as a function of experience. This
is neither logically precluded by his criticism of the argument from illusion,
nor is it apparent from the phenomenological observation that experience is
diaphanous. However, I do think that the inconceivability of unfelt pains or
bodily sensations make a good case for the thesis that, as a matter of fact,
certain features of phenomenal character only exist to the extent that they
are actually experienced. Moreover, pains especially have been put forward
as problematic cases for the functionalist.22 Hence, my conclusion is that
the essentially experiential nature of pains and bodily sensations allow us
to say that we are in some sense aware of intrinsic qualities of our experi-
ence. And because Harman’s arguments do not affect this circumstance in
any way, his arguments cannot be used to defend functionalism against the
objection that phenomenal character cannot be accounted for in functional
terms.

V. Final Remarks
In this paper I have paid considerable attention to some of the terminological
and doctrinal difficulties that are involved in the contemporary philosophy
of mind. I maintain that the resolution of such difficulties is of crucial
importance to the understanding of an article such as Harman’s. Without a
doubt, my emphasis on these issues have prevented me from getting to more
detailed examinations of Harman’s arguments. For I think there is still much
left to examine. However, I must conclude that this is only because there is
a lot of philosophical vagueness on Harman’s part. In so far as I understand
Harman’s article, I do think I have been able to expose some of its flaws.
My resolution of certain conceptual issues have especially helped me in this,
for they have showed me that the thesis that Harman is arguing against
(the thesis that we are aware of intrinsic qualities of experience) can be
interpreted in at least two ways, and that Harman’s arguments only poses
a threat to one of them.

References
Block, N. (2003). Mental paint. Reflections and Replies: Essays on the
Philosophy of Tyler Burge, (pp. 165–200).

Bourget, D., & Mendelovici, A. (2013). Tracking representationalism. In

Philosophy of Mind: the Key Thinkers. Bloomsburry.
22
See e.g. (Lewis, 1980) and (Levine, 2003)

14
Brandt, R., & Kim, J. (1967). The logic of the identity theory. The Journal
of Philosophy, 64 (17), 515–537.

Byrne, A. (2009). Experience and content. The Philosophical Quarterly,

59 (236), 429–451.

Chalmers, D. J. (2004). The representational character of experience.

Chisholm, R. M. (1969). On the observability of the self. Philosophy and

Phenomenological Research, 30 (1), 7–21.

Churchland, P. M. (1985). Reduction, qualia, and the direct introspection

of brain states. The Journal of Philosophy, 82 (1), 8–28.

Crane, T. (2002). The origins of qualia. In History of the Mind–Body

problem, (p. 169). London, UK: Routledge.

Dretske, F. (1995). Naturalizing the Mind . Cambridge, Ma: MIT Press.

Fish, W. (2010). Philosophy of Perception: A Contemporary Introduction.

New York, New York: Routledge.

Harman, G. (1990). The intrinsic quality of experience. Philosophical per-

spectives, 4 , 31–52.

Jackson, F. (1977). Perception: A representative theory. CUP Archive.

Jackson, F. (2000). Some reflections on representationalism.

Kind, A. (2003). Whats so transparent about transparency? Philosophical

Studies, 115 (3), 225–244.

Levine, J. (1995). Qualia: intrinsic, relational or what? Conscious experi-

ence, (pp. 277–292).

Levine, J. (2003). Experience and representation. In Q. Smith, & A. Jokić

(Eds.) Consciousness: New Philosophical Perspectives, (pp. 57–76). Ox-
ford, UK: Oxford University Press.

Lewis, D. (1980). Mad pain and martian pain. Readings in the Philosophy
of Psychology, 1 , 216–222.

Moore, G. E. (1903). The refutation of idealism. Mind , 12 (48), 433–453.

Nagel, T. (1970). Armstrong on the mind. The Philosophical Review , 79 (3),

394–403.

Nagel, T. (1974). What is it like to be a bat? The Philosophical Review ,

83 (4), 435–450.

15
Price, H. (1950). Perception. London, UK: Methuen and Co., 2 ed.

Robinson, H. (1994). Perception. New York, New York: Routledge.

Searle, J. (1995). The Construction of Social Reality. London, UK: Penguin.

Shoemaker, S. (1990). Qualities and qualia: What’s in the mind? Philosophy

and Phenomenological Research, 50 , 109–131.

Shoemaker, S. (2003). Content, character and color. Philosophical Issues,

13 (1), 253–278.

Smart, J. J. (1959). Sensations and brain processes. The Philosophical

Review , 68 (2), 141–156.

Stoljar, D. (2004). The argument from diaphanousness. Language, mind

and world. Canadian journal of philosophy, supplementary, 30 .

Thompson, B. (2008). Representationalism and the argument from halluci-

nation. Pacific Philosophical Quarterly, 89 (3), 384–412.

Tye, M. (1995). Ten Problems of Consciousness. Cambridge, Massachusetts:

the MIT Press.

Tye, M. (2013). Transparency, qualia realism and representationalism.

Philosophical Studies, (pp. 1–19).

16