bs_bs_banner

DOI: 10.1111/ejop.12005

Mark Sacks Lecture

Perceptual Experience: Both Relational

and Contentful
John McDowell
University of Pittsburgh

1. What is meant by saying perceptual experience is relational?

Capturing a widely shared intuition, Charles Travis says perception places our
surroundings in view.1 (Obviously he is focusing on visual perception, and I shall
stay with that focus.) If something in one’s surroundings is in view for one, that
is a relation one stands in to it. The intuition is that it is by placing one in such
relations to things that perception enables one to know about them.
Many people—Travis for one—think a relational account of the cognitive role
of perception is inconsistent with explaining the epistemic significance of expe-
riences in terms of content.2 Travis thinks content figures in the truth about
perceptual knowledge only in connection with, as he puts it, what we make of
what perception anyway places in our view, whether we make anything of it or
not. Content does not play a role in how perceptual experience presents us with
things.
I am going to take issue with that. Perception makes knowledge about things
available by placing them in view for us. But it is precisely by virtue of having
content as they do that perceptual experiences put us in such relations to things.

2. What is meant by saying experiences have content?

In the most familiar form of the idea, experiential content is propositional.
Wilfrid Sellars puts this by describing experiences as so to speak making claims.3
On this version of the idea, if an experience enables someone to know there
is something red and rectangular in front of her, in the most straightforward way,
that is because it is an experience of seeing that, in Sellars’s metaphor, so to

1
See “The Silence of the Senses”, Mind vol. 113 (2004), 65.
2
See, e.g., Bill Brewer, “Perception and Content”, in Jakob Lindgaard, ed., John McDowell:
Experience, Norm, and Nature (Malden, MA: Blackwell, 2008); Brewer, Perception and its
Objects (Oxford: OUP, 2011); John Campbell, Reference and Consciousness (Oxford: OUP,
2002).
3
“Empiricism and the Philosophy of Mind”, in Science, Perception and Reality (London:
Routledge, 1963; reprinted Atascadero, CA: Ridgeview, 1991), §16.

European Journal of Philosophy 21:1 ISSN 0966-8373 pp. 144–157 © 2013 Blackwell Publishing Ltd., 9600 Garsington Road,
Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
 Perceptual Experience 145

speak claims that there is something red and rectangular in front of its subject.4
And—here I diverge from Sellars—it makes that knowledge available by making
present to the subject a state of affairs consisting in there being something red and
rectangular in front of her. That there is something red and rectangular in front
of the subject is both the relevant aspect of the experience’s content and what it
enables the subject to know.
On another version of the idea, experiential content is less than propositional;
it is related to the content of claims as noun phrases that apply to things claims
are about are related to sentences uttered in making those claims. This form of
the idea can also be found in Sellars.5
On this version of the idea—again, here I diverge from Sellars—an experience
that enables someone to know there is something red and rectangular in front of
her, again in the most straightforward way, makes present to her not a state of
affairs, but an object: something presented in the experience as red and rectan-
gular and in front of her. On this version we cannot connect the experience’s
content so directly with what it enables its subject to know. But the connection
is still obvious: if an experience presents someone with an object as red and
rectangular and in front of her, it enables her to know there is something red and
rectangular in front of her.
For my purposes, it will make no difference which version we work with.
I have illustrated the idea of experiential content in terms of formulations of
content that can only be partial. Even if one has nothing else in view but a red
rectangular surface, there will be more specificity in the content of one’s
experience than is captured by “red” and “rectangular” and “in front of me”.
That will not matter for my purposes.

3. On a relational conception, experience enables us to know things about the

environment by placing us in cognitively significant relations to environmental
realities. And in spelling out the idea that experiences have content, I presup-
posed something on those lines. I sketched an account in which the content of
an experience of seeing enables it to make present to its subject a certain
environmental reality, a state of affairs or an object, thereby putting her in a
position to have associated knowledge about her environment. If a relational
conception is inconsistent with explaining the cognitive significance of experi-
ences in terms of content, what I said must be incoherent.
So where is the incoherence? Why is it thought that perceptual experience
cannot be both relational and contentful, indeed relational precisely by virtue of
being contentful in the way it is?

4
In the most straightforward way: we can imagine scenarios in which an experience
might make that knowledge available to one indirectly, by inference from knowledge
available to one by virtue of having something in view.
5
See, e.g., “Some Reflections on Perceptual Consciousness”, in Kant’s Transcendental
Metaphysics, edited by Jeffrey F. Sicha (Atascadero, CA: Ridgeview, 2002).

© 2013 Blackwell Publishing Ltd.

146 John McDowell

I shall consider two responses to that question.

4. One goes like this: if we credit experiences with content, we imply that what
a subject encounters in an experience, even one we would like to conceive as
enabling her to know something about her environment, is not the things in her
surroundings that, according to a relational conception, her experience relates
her to, but the content of her experience. We have to conceive the content as
intervening between the subject and her environment. So we lose the direct
access to the environment that figures in a relational conception.
I shall give this response short shrift.
Nothing in the idea that experiences have content implies denying that if an
experience is one of seeing, what its subject encounters in her experience is an
environmental reality. There is no implication that an experience’s content would
usurp the role in which a relational conception casts things in the subject’s
surroundings. We have, so far, no reason not to suppose experiences of seeing
place our surroundings in view precisely by virtue of having the content they do.

5. A more interesting response can be set out in the form of an argument, on the
following lines.6
The epistemic significance of an experience must be available to the subject
in enjoying the experience. It must reside in some aspect of the experience’s
subjective character. So on the content conception, an experience’s having
content as it does must be constitutive of its subjective character, in so far as its
subjective character is relevant to its putting its subject in touch with her
surroundings, or at least seeming to. That is the argument’s first premise.7
The second premise is that experiences can appear to make knowledge about
the environment available when they do not.
It can seem to follow that if an experience enables someone to know
something about her environment, it cannot be by putting her in relation to a
suitable environmental reality. The second premise can seem to entail that in the
relevant aspect of its subjective character, an experience that makes knowledge
available is just like one that merely appears to do that, and so certainly does not
put its subject in a cognitively relevant relation to a suitable environmental
reality. It would follow that if the epistemic significance of experiences is
explicable in terms of their having content, it is indifferent to the possible
non-existence of any relevant environmental reality. But a relation cannot be
indifferent to the possible non-existence of one of its relata.

6
In formulating the argument I have been greatly helped by the first chapter of William
Fish, Perception, Hallucination, and Illusion (Oxford: OUP, 2009).
7
The point of the “in so far as . . .” proviso is that there may be more to the subjective
character of an experience than its epistemic significance. For instance, in a visual
experience of a shortsighted person not wearing corrective lenses, things are out of focus.
That belongs to the experience’s subjective character, but not to its epistemic significance.

© 2013 Blackwell Publishing Ltd.

 Perceptual Experience 147

6. We should not find this argument convincing.

It is true that if what we say of an experience is only that its content is such
as to make it appear to put its subject in touch with a certain environmental
reality—otherwise put: such as to determine it to represent such an environ-
mental reality—we leave open a possibility that there is no such environmental
reality.
But of some experiences we can say something stronger than that, without
going beyond describing their subjective character. If an experience is a seeing,
we can say the representing it does is its revealing or disclosing a certain
environmental reality: that is, its bringing that environmental reality into view
for the subject. Experiential representing, the representing by experiences that
consists in their having content in the way they do, comes in two kinds: bringing
environmental realities into view and merely seeming to do that. If an experi-
ence’s representing is of the first kind, what is relevantly constitutive of its
subjective character is that it brings a certain environmental reality into view for
its subject.
On the content conception, the epistemic significance of an experience consists
in its having content in the way it does. An experience that is a seeing can be like
an experience that merely appears to put its subject in touch with a correspond-
ing environmental reality in respect of what content it has. But a seeing is unlike
a mere appearing in how it has its content. Seeings have their content in a way
that is characteristic of seeings; they make environmental realities present to
their subject. Thereby, on either version of the idea of experiential content, they
put the subject in a position that leaves open no possibility of things not being
as they would be believed to be in suitably related beliefs.
So the supposed inconsistency with the idea that perceptual experience puts
its subject in relation to environmental realities disappears. By virtue of having
content in the way they do, experiences of seeing bring environmental realities
into view for their subjects, and thereby provide them with conclusive warrant
for believing suitably related things about their environment.

7. The argument I have been considering purports to establish this conditional:

if the epistemic significance of an experience consists in its having content in the
way it does, then it cannot be by bringing environmental realities into view that
experiences enable their subjects to know things about the environment.
That conditional can be exploited in either of two directions.
The argument I have been considering is an argument by modus tollens, to the
conclusion that the epistemic significance of an experience does not consist in its
having content in the way it does.
But the conditional can also figure in an argument by modus ponens. Assume
that the epistemic significance of experiences does consist in their having content
in the way they do. Now the conditional yields this conclusion: if an experience,
by virtue of its content, enables its subject to know something about her
environment, it cannot be by placing her in relation to a suitable environmental
reality. If the conditional is right, the content conception entails that the epistemic

© 2013 Blackwell Publishing Ltd.

148 John McDowell

significance of an experience, even one we want to conceive as enabling its

subject to know something about her environment, is indeed indifferent to the
possible non-existence of any relevant environmental reality. When someone
knows something about her environment on the basis of an experience, that
things are as she believes them to be is not guaranteed by the experience’s
epistemic significance. Experiences provide warrants for beliefs, but even when
we want to count the beliefs as knowledgeable, the warrants provided for them
by the relevant experiences are less than conclusive.
Many people have contrived to find this tolerable. They have persuaded
themselves that a belief on the basis of less than conclusive warrant can
intelligibly count as knowledgeable.
But I do not think anyone has ever given a satisfactory answer to this
question: if one acknowledges that one’s warrant for believing something leaves
open a possibility that things are not as one believes them to be, how does that
differ from acknowledging that for all one knows things are not as one believes
them to be, so that one does not know them to be that way? I doubt that anyone
would go on supposing that a belief based on less than conclusive warrant can
be knowledgeable, refusing to be embarrassed by that question, were it not for
being unable to envisage any alternative—apart, that is, from explicitly conced-
ing that we ought to be sceptical about the possibility of experience-based
knowledge.8
But I have described an alternative, which is consistent with the content
conception of the epistemic significance of an experience. The representing an
experience does can be its making an environmental reality present to its subject,
even though experiences can seem to make knowledge of the environment
available when they do not. We do not need to accept a forced choice between
the modus tollens version of the argument and its modus ponens version.

8. More needs to be said about this. It can be hard to see how we can reject that
conditional.
The argument I considered is right to focus on the subjective character of
experiences. Our topic is, or at least should be, how experience figures in the
warrant for knowledge of a sort that is distinctive of rational subjects: knowledge
of a sort we can conceive, following Sellars, as a standing in the space of
reasons.9 Knowledge of this sort is an act of its subject’s rationality. Someone
who has knowledge of this sort must be in a position to know the warrant by
virtue of which her state counts as knowledge. No doubt one does not typically
make explicit, even to oneself, how one’s knowledgeable beliefs are warranted,
or even what it is that one knowledgeably believes. But if an experience warrants
one in knowledge of the sort that is an act of rationality, the warrant-constituting

8
See Sebastian Rödl’s remark about putting forward scepticism as a theory of knowledge,
in Self-Consciousness (Cambridge, MA: Harvard University Press, 2007), 149.
9
See “Empiricism and the Philosophy of Mind”, §36.

© 2013 Blackwell Publishing Ltd.

 Perceptual Experience 149

status of the experience must be part of the content of an at least implicit

self-consciousness that belongs to one’s cognitive state in knowing what one
does. In the terms I have been using, the experience’s epistemic significance must
be part of its subjective character.
Now consider again the argument’s second premise. An experience can seem
to put knowledge about the environment at one’s disposal, in whatever way
experiences do, when it does not do that. And it can seem urgent to ask: if that
is so, how can the subjective character of an experience include its being such as
to warrant one in a belief about the environment in a way that leaves open no
possibility that things are not as one believes them to be? How can one’s
self-consciousness in enjoying the experience enable one to know that it is not
one of the sort one is liable to mistake for an experience that makes knowledge
available: one that merely seems to make knowledge available, and so certainly
does not leave open no possibility that things are not as one believes them to be?
Suppose we find this line of thought convincing. What options does that leave
us, in framing an epistemology for perceptual knowledge?
We would be immune to the puzzlement those questions express if we
accepted a relational conception of the sort that refuses to conceive the epistemic
significance of experience in terms of content. On this view, experiences con-
tribute to one’s having knowledge about one’s environment, but only by placing
one’s surroundings in view. If one knows something about one’s environment
through perception, that is because one has made something of things an
experience places in view for one. One has brought them under concepts,
exercising cognitive capacities that are extra to having the things in view. If
someone has a knowledge-constituting warrant for a belief about how things are
in her environment, the warrant depends on the epistemic credentials of the
capacities she has exercised in applying concepts to things her experience
anyway places in view for her. So the puzzlement does not arise. There is no
application for the conception that causes it; there is no such thing as an
experience that itself provides its subject with a conclusive warrant, or indeed
any warrant, for believing something about her environment.
That corresponds to the conclusion of the modus tollens version of the
argument. And it can seem that the only alternative is the conclusion of the
modus ponens version: a conception according to which experiences themselves
do warrant beliefs, but only less than conclusively, even in the case of experi-
ences we want to conceive as making knowledge about the environment
available.
If an experience could by itself guarantee that things in the environment are
a certain way, the subject would have to be in a position to know, in her
self-consciousness in enjoying the experience, that it had that epistemic signifi-
cance. As I said, that just reflects the fact that the knowledge we are concerned
with is an act of its subject’s rationality. The difficulty is to see how that
requirement could be met, consistently with the fact that an experience can seem
to make knowledge available, in whatever way experiences do that, when it does
not.

© 2013 Blackwell Publishing Ltd.

150 John McDowell

We might be tempted by a thought on these lines: if one takes an experience

to make knowledge available when it does not, that reflects some flaw in one’s
cognitive conduct—haste, inattention, or whatever. That way, we could suppose
an experience can conclusively warrant a belief about the environment, consist-
ently with the requirement that one’s self-consciousness in enjoying it must put
one in a position to know it has that epistemic significance. It is just that to avail
oneself of that possibility of knowing the epistemic significance of an experience,
one needs to ensure that one’s cognitive conduct, in exercising one’s capacity to
know the epistemic significance of one’s experiences, is flawless. The idea would
be that if one exercised that capacity without such flaws in cognitive conduct as
haste or inattention, one would not be at risk of taking an experience to enable
one to know something about the environment when it does not.
But the trouble with this idea is that it is not credible. Taking an experience
to make knowledge available when it does not can be blameless. This is not a
way to avoid the conclusion that if experiences themselves do warrant beliefs,
they do so only less than conclusively, even in the case of experiences we want
to conceive as making knowledge available.
So we seem to be stuck with that forced choice. There seems to be no room
for the alternative I described.

9. I acknowledged that if experiences can warrant knowledge, the warrant-

constituting significance of an experience must belong to its subjective character;
it must be something the subject can know just from her self-consciousness in
enjoying the experience. As I said, that is a consequence of the fact that the
knowledge whose warrant we are trying to understand is an act of the knower’s
rationality. If someone has a bit of knowledge of the kind that is an act of her
rationality, she must be in a position to know, just in being in the state that
constitutes her knowing, how her rationality is operative in it.
What makes it seem that we are restricted to those two options is an
interpretation of that requirement on these lines: knowledge of the warrant-
constituting significance of an experience would have to be available to one in
a way that is epistemically prior to the availability to one of the knowledge that
the experience enables one to have about the environment.
That is what makes it seem that experiences could have an epistemic signifi-
cance consisting in making environmental realities present to their subjects only
if knowledge that an experience has such an epistemic significance were avail-
able to its subject in an act of a cognitive capacity that would be infallible, if she
avoided haste or inattention or the like in her exercises of it. The second premise
of our argument is, in effect, that there is no such cognitive capacity. So it comes
to seem that the epistemic significance of an experience, since it must be
introspectively knowable, can never be such as to exclude all possibility that
things are not a certain way in the environment.
But that is not how the requirement should be understood.
If a state of affairs consisting in there being something red and rectangular in
front of one is present to one in an experience, one is thereby in a position to

© 2013 Blackwell Publishing Ltd.

 Perceptual Experience 151

know that that is how it is with one, and thus to recognize that one is
conclusively warranted in believing there is something red and rectangular in
front of one. That is to formulate the point in a way that fits the idea that
experiential content is propositional, so that it is states of affairs that perception
makes present to one. But a corresponding formulation fits the idea that
experiential content is less than propositional: if one’s experience makes an
object present to one as red and rectangular and in front of one, one is thereby
in a position to know that that is how it is with one, and thus to recognize that
one is conclusively warranted in believing there is something red and rectan-
gular in front of one. As I said, the difference between the two versions does not
matter for my purposes; on either version, to have a suitable environmental
reality present to one is to have a conclusive warrant for a certain belief about
one’s environment.
Either way, the essential point is this: it is not by exercising a self-standing
capacity to know how it is with one that one knows one’s experience has an
epistemic significance of that sort. One’s knowledge that there is something red
and rectangular in front of one includes knowledge of its own credentials as
knowledge. And it is the knowledge it is because it is a non-defective act of a
capacity to know such things through perception. That capacity is a capacity to
be in positions in which one knowingly has such environmental realities present
to one.
The premise that has been giving us trouble is that one can take an experience
to make knowledge available when it does not. That does not show that when
an experience makes knowledge available, it is not by bringing a suitable
environmental reality into view for one. The point is just that the capacity to be
in such positions is fallible.
We take in stride the fact that a capacity to know through perception such
things as that there is something red and rectangular in front of one is fallible.
We must, if we are going to suppose we ever have such knowledge. What
makes it look as if the capacity cannot work in the way I have described—
yielding knowledge whose title to count as such depends on experience
making environmental realities present to us and thereby giving us conclusive
warrant for beliefs about the environment—is the thought that knowing the
epistemic significance of an experience that warrants a bit of knowledge
we have through that fallible capacity would have to be an act of another
capacity: a separate capacity to know such things, which would have to be in
principle infallible. And the argument’s second premise is that there is no
such infallible capacity. But the argument fails, because the capacity by
which one knows the epistemic significance of one’s experiences is not another
capacity, but just the capacity whose fallibility we anyway have to take in
stride.
To repeat: if one knows through perception that there is something red and
rectangular in front of one, one’s knowledge includes knowledge, at least
implicit, of its own credentials as knowledge. And it is the knowledge it is
because it is a non-defective act of a capacity to be in positions in which one

© 2013 Blackwell Publishing Ltd.

152 John McDowell

knowingly has present to one environmental realities such that in knowingly

having them present to one, one knowingly has conclusive warrant for rel-
evantly related beliefs. The knowing alluded to in those occurrences of “know-
ingly” is an act of a capacity that has the fallibility we are anyway committed
to taking in stride. One can take an experience to make knowledge available
when it does not, but that just reflects the fact that the capacity is fallible. When
the capacity is non-defectively in act, one knowingly has present to one an
environmental reality such that in having it present to one, one has a guarantee
that things are as one believes them to be.

10. There are various ways in which acts of a capacity of this kind can be
defective.
There is defectiveness of the most obvious kind when an experience makes
things seem to be a certain way when they are not. Perhaps the lighting is such
that something looks red when it is not, or an illusion of perspective makes a
surface look rectangular when it is not.
But acts of a capacity of this kind can be defective even if things in the
environment are just the way an experience makes them seem to be. Philoso-
phers often talk as what is important is whether an experience is veridical.
But for an experience to be veridical would be for things in the environment
to be the way it makes them seem to be, and that does not suffice for an act
of a capacity of the kind we are considering to be non-defective. For a non-
defective act of a capacity to know by looking that something in front of one
is, say, red, it is not enough that one sees a red thing in a good light for
getting to know the colours of things, even if one’s colour vision is in perfect
working order.
Suppose someone is in the hands of scientists experimenting with illumina-
tion that is undetectably unsuitable for knowing colours by looking: in a few
trials things look as if they have the colours they have, and in most trials they
do not. When this subject is confronted by a thing that is, say, red, and it looks
red to her through the effect of ordinary light on her visual equipment, which
is functioning properly, she is not thereby in a position to know that the thing
is red. Though her experience is veridical and normally caused, it is a defective
act of her capacity to know colours by looking. It is an act of that capacity, in that
the thing looks red to her. But it is not an act in which she is in a position to
know the thing’s colour by looking.10
The same holds if she has been told she is in an experiment of that sort. (It
does not matter whether what she has been told is true.) Even if on some
occasion the light is suitable for knowing colours by looking, for all she knows
this is one of the occasions she has been told there will be, on which colour

10
This example is structurally analogous to a “barn façade” case.

© 2013 Blackwell Publishing Ltd.

 Perceptual Experience 153

appearances are misleading. Here again her normally caused veridical experi-
ence is a defective act of the capacity to know things by looking.11
In either version of this case, the subject does not know the thing’s colour,
because she is not entitled to take its colour to be visually present to her in her
experience. If she does not know she is in the experimental situation, she may
wrongly take herself to be so entitled. If she does know, or just thinks she knows,
she is in the experimental situation, she will not take herself to be in a position
to know that her experience presents her with the actual colour of the thing she
sees.
It is important to see that the second version of the case is not an entering
wedge for a renewed threat to the conception I have been recommending.
This subject knows she is not entitled to take the colour of something she sees
to be visually present to her. That is so even on an occasion on which the light
is suitable for knowing colours by looking, so that in one sense the thing’s colour
is visually present to her. On any occasion, for all she knows the light is
unsuitable for knowing colours by looking, as she has been told it will be on
some occasions.
Now apart from situations like the experiment, illumination may be
undetectably unsuitable for knowing colours by looking; that is just one way acts
of the capacity to know colours by looking may be defective, one mode of
actualization of the fallibility everyone has to accept that the capacity has. So
ought we to say that on any occasion on which someone takes herself to have
something’s colour visually present to her, she is not in a position to know that
her experience has that epistemic significance, because for all she knows the
illumination is undetectably unsuitable for knowing colours by looking? Maybe
she could check the illumination. But surely one cannot establish, for all
conceivable scenarios in which acts of the capacity would be defective, that those
scenarios are not actual, as a precondition for taking a current act of the capacity
to yield knowledge. For one thing, how could one know one had thought of all
the possible ways an act of the capacity could be defective?
This line of thought is wrong in a way that is parallel to what goes wrong in
the argument I considered. There the idea was this: one could know that one’s
experience conclusively warranted a belief about one’s environment only in an
exercise of an infallible capacity to know such things. Here the idea is this: one
could know such a thing only by entitling oneself, as a precondition for taking
one’s experience to yield knowledge, to exclude any scenario in which one’s
experience would seem to make knowledge available when it does not. In both
cases knowing the epistemic significance of one’s experiences is conceived as an

11
This example is structurally analogous to one with which Crispin Wright tries to make
trouble for the conception I am recommending. See “Comment on John McDowell’s ‘The
Disjunctive Conception of Experience as Material for a Transcendental Argument’ ”, in
Adrian Haddock and Fiona Macpherson, eds., Disjunctivism: Perception, Action, and
Knowledge (Oxford: OUP, 2008).

© 2013 Blackwell Publishing Ltd.

154 John McDowell

act of a capacity separate from the capacity whose fallibility everyone has to
acknowledge, the capacity to know the relevant things about one’s environment.
The effect is that fallibility about the epistemic significance of one’s experiences
is made out to pose problems not addressed by acknowledging the fallibility
everyone has to acknowledge.
As before, the apparent problems dissolve when we recognize that knowing
one’s experience provides conclusive warrant for a belief about the environment
is a non-defective act of the same fallible capacity that is non-defectively in act
in knowing the relevant things about the environment. In a non-defective act of
the capacity to know colours by looking, one knows that something’s colour is
visually present to one, and so knows, at least implicitly, that no spoiling
scenario is actual. In claiming to have knowledge that is an act of a fallible
capacity, one has to admit that one risks having one’s claim turn out to be false.
If there is reason to suspect a spoiling scenario, one ought to be especially
cautious. But acknowledging that one runs that risk in one’s claim to know is not
acknowledging that for all one knows some scenario is actual in which one
would not know what one is claiming to know.
The line of thought I have been considering makes it seem that the ultimate
ground for a claim to know something about the environment would have to be
a ground one independently has for excluding spoiling scenarios. But the line of
thought is wrong. The ultimate ground for a claim to know something about the
environment is that one’s experience makes a relevant environmental reality
present to one. In claiming that warrant, one must be ready to acknowledge that
the claim issues from a capacity that is fallible, so there is a risk of its turning out
to be wrong. But if it is not wrong, one’s knowledge about the environment is
warranted in just the way in which, knowingly running that risk, one claims it is.

11. Defective acts of a capacity to know by looking that there is something red
and rectangular in front of one—that is, experiences that at best only seem to put
one in a position to know that there is something red and rectangular in front
of one—are just like non-defective acts of that capacity in respect of their content.
The defective and non-defective acts alike have their content determined by
their being, among other things, acts of a capacity to have instances of redness
visually present to one and acts of a capacity to have instances of rectangularity
visually present to one.12

12
Those capacities must be in act with a certain togetherness, because what is present to
one in a non-defective act is not just an instance of redness and an instance of
rectangularity, but something that is both red and rectangular.
When we take note of the fact that a capacity to know, by looking, such things as that
there is something red and rectangular in front of one can be articulated into sub-
capacities, we can register that an act may be defective in respect of one sub-capacity but
non-defective in respect of another. Something that looks red and rectangular to one may
be red but not rectangular, or rectangular but not red.

© 2013 Blackwell Publishing Ltd.

 Perceptual Experience 155

To describe an experience as, in part, an act of a capacity to have instances of

redness present to one is to group it with other experiences, actual or possible,
in a way that loses some specificity in what is, or seems to be, relevantly present
to one in any one of them. That does not matter for my purposes; it is enough
that the description is correct as far as it goes.
I said that what differentiates defective and non-defective acts of a capacity to
know by looking that there is something red and rectangular in front of one is
not their content. What differentiates them is that in non-defective acts an
experience has its content in the manner characteristic of a seeing: it makes a
suitable environmental reality present to one, and thereby enables one to know
there is something red and rectangular in front of one. Whereas in defective acts
the experience at best seems to do that (it does not even seem to do that in
the case of the subject who knows, or thinks she knows, that she is in an
experiment).
It may seem surprising that I say it is not content that differentiates defective
from non-defective acts. It is often assumed that the content of an experience of
seeing would have to be special, in a way that would distinguish it from the
content of an experience of merely seeming to see. I have urged that if we
explain the epistemic significance of experiences in terms of their having content
in the way they do, we can capture the idea that perceptual experience relates
subjects to things in their surroundings. And it is often assumed that if an
experience of seeing relates its subject to things in her surroundings, that would
have to be completely provided for by its content.
Sellars suggests that on the non-propositional conception of experiential
content, the content of an experience of seeing can be given by a demonstrative
expression, on the lines of “That red rectangular thing”.13 On a certain construal
(which cannot, I think, be the one Sellars intends), content expressible like that
would be itself relational; it would owe its being available to be expressed, or
thought, to a relation in which the subject of the experience in question stands
to something she can refer to like that. I think that way of construing perceptual
demonstratives is right, and just for that reason I think we should reject Sellars’s
suggestion. A demonstrative expression of that sort expresses, not the content of
an experience, but a thinking the subject is enabled to engage in because of how
her experience, by virtue not just of its content but of its being a seeing with that
content, relates her to something she is thereby enabled to refer to like that.
Visual experiences as such, whether seeings or not, have content that enables
those that are seeings to place their subjects in such relations to objects. But that
is provided for, not by a de re character that belongs to the contents of seeings,

13
See, e.g., Science and Metaphysics (London: Routledge, 1967; reprinted Atascadero, CA:
Ridgeview, 1992), 3. Talking about an interpretation of Kant on intuitions, Sellars says:
“On this model, which I take to be, on the whole, the correct interpretation, intuitions
would be representations of thises and would be conceptual in that peculiar way in which
to represent something as a this is conceptual.”

© 2013 Blackwell Publishing Ltd.

156 John McDowell

but by a de se character that belongs to the contents of visual experiences as

such.14 When I formulate an aspect of the content of an experience with the
words “(There is) something red and rectangular in front of me”,15 “in front of
me” is schematic for a much more specific placement, in relation to me, of an
apparent red rectangular thing. That is enough to secure that if the experience
is a seeing, it enables me to refer, with a visually grounded demonstrative, to the
red rectangular thing I see.16

12. I have argued that if we take the epistemic significance of experiences to

consist in their having content in the way they do, we can accommodate the
intuition that motivates a relational conception of perceptual experiences. The
intuition does not require us to restrict content to cognitive acts that are extra to
having things in view. It can be by having content in the way they do that
experiences of seeing place things in view for us.
The conception I have recommended seems clearly better than one of the two
other conceptions I have considered, a conception that requires us to pretend a
belief can count as knowledgeable even though one’s warrant for it leaves open
a possibility that it is not true. But I have so far argued only that my conception
is available. I have not argued that it is better than the other of the two
competitors, a conception according to which content figures in the epistemology
of perception only in connection with our bringing under concepts things
perception anyway places in view for us.
According to this competitor, any knowledge one has by virtue of an expe-
rience is an act of capacities that are extra to having things in view, which is what
the experience provides.
That seems the right line to take about some knowledge one has by virtue of
an experience. In having something red and rectangular in view, I might know
that it is a book (what I see is its front cover); perhaps that it is a copy of
Wittgenstein’s Tractatus; perhaps even that it is my copy of Wittgenstein’s
Tractatus. These bits of knowledge would be acts of cognitive capacities that
would not need to be in operation for me to have the thing in view at all.17

14
See David K. Lewis, “Attitudes De Dicto and De Se”, Philosophical Review 88 (1979).
15
“There is” is to be included or not depending on whether we are working with the
propositional or the non-propositional conception of experiential content. It makes no
difference to the point I am making.
16
In this section I correct something I said in “Intentionality as a Relation” (Essay 3 of
Having the World in View: Essays on Kant, Hegel, and Sellars [Cambridge, Mass.: Harvard
University Press, 2008]). I thought I had to hold that a merely seeming intuition merely
seems to have content of the kind an intuition would have, which I understood as Sellars
does. But I was wrong to think a merely seeming intuition and a genuine intuition would
need to differ in content.
17
The concession need not be taken to imply anything about the phenomenology of
experiencing.

© 2013 Blackwell Publishing Ltd.

 Perceptual Experience 157

But I can make that concession and still insist that having the thing in view
at all is already an act of cognitive capacities that involve content. Experience
places things like books in view for us, so that we can bring them under concepts
like the concept of a book, by making their visible features present to us. In my
example, what is in fact the front cover of the book is present to me as red and
rectangular. If I come to know that it is red and rectangular, it is not by exploiting
cognitive capacities that are extra to what is required for having it in view at all.
It is by making explicit—with a loss of specificity—content, relating to the thing’s
visible features, that belongs to the experience in which I have the thing in view.
How might we adjudicate this difference?
Our topic is, as I said, knowledge of a kind that is an act of its possessor’s
rationality. That formulation is explicitly restricted to a kind of knowledge; it
leaves unchallenged the common-sense view that it is not only rational animals
that know things through perception.
Now on the competing conception, in just having things in view we are not
exploiting cognitive capacities of the sort that are distinctive of rational animals.
Distinctively rational capacities are operative only in our making something of
what experience anyway brings into view for us. Our ability to make something
of what we see is special to us as rational animals, but in isolating what
experience does for us we have to abstract from that ability. So our capacity to
have things in view, considered by itself, is indistinguishable from the capacity
to have things in view possessed by animals that are not capable of that
distinctive kind of knowledge. The special character of knowledge that is an act
of its possessor’s rationality does not extend into all the capacities that are
operative in acquiring it.
On this picture, rationality is operative in bringing things under concepts, but
not in being given things to bring under concepts. To a follower of Sellars, that
looks like a form of the Myth of the Given. Evidently that does not worry
proponents of the competitor conception, and obviously I cannot say much
about this now. Let me end with this remark. Proponents of the competitor
conception may be confident that there cannot be anything in the accusation of
falling into the Myth of the Given, because they think seeing things their way is
compulsory if we are to accommodate the intuition that perceptual experience
puts us in relation to things in our surroundings. But in this lecture I have
undermined that supposed ground for confidence in brushing the accusation
aside.

© 2013 Blackwell Publishing Ltd.