Why Perception Matters: the Case of the
Global Clairvoyant
Many philosophers have supposed that perception plays an indispensable role in making it possible for us to have thoughts and beliefs about the empirical world. If we didn’t have any perceptual experiences, on this view, we wouldn’t merely lack knowledge about what the world is like; it would be impossible for us to have any thoughts or beliefs about the world at all.
Philosophers have typically believed that perception is indispensable for one of two reasons. According to one traditional empiricist view, perception is indispensable because our most basic empirical concepts must be derived from perceptual experience. Some—more complex—concepts can be understood in terms of their logical relations to other concepts. But to avoid a regress, on this view, we must eventually bottom out in primitive concepts that can only be understood ostensively—by pointing at or “attending to” the things in our experience to which the concept applies. We can find a second reason for thinking that perception is indispensable in John McDowell’s book, Mind and World. According to McDowell, our beliefs cannot have any empirical content unless they can be justified by experience. Unless experiences can serve as reasons for holding beliefs, our thoughts would not be answerable to the facts they are purportedly about, and would therefore fail to have any content at all. Our thinking, in that case, would turn into what McDowell calls a “frictionless spinning in a void.”
I don’t think that either of these reasons for thinking that
perception is indispensable work in the end, though I will not argue for that
here. Instead, I will sketch what I
think is a more promising way of understanding why we cannot have any beliefs
about the world without perception. The view I will consider does not focus
on how our primitive concepts are acquired or how our beliefs are justified,
but rather on the links between perception and other important capacities that
are themselves essential for having thoughts about the world.
I think it is helpful to introduce this discussion by raising a question about content externalism. Content externalists believe that causality plays an essential role in determining the contents of our thoughts and beliefs. They think that in the most basic cases, our beliefs are necessarily about the things that normally cause them. I will not question the truth of this thesis here. Instead, I would like to see if there are any restrictions on how our beliefs are caused by the things they are about. As it happens, events in the world cause us to adopt our most basic beliefs through sense perception. It is by means of my senses that the presence of a table in front of me causes me to believe that a table is in front of me; I believe that the table is there because I see it there before me. I am interested in the significance of the fact that it is typically through perceptual channels that we are caused to form beliefs by the things they are about. Is this merely a contingent fact about the way human beings are caused to form beliefs? Might things have been otherwise?
How might we evaluate the significance of
the fact that we are typically caused to form many of our beliefs through
perception? I propose that we proceed by just trying to imagine a person who is
caused to form all of her beliefs in a deviant, non-perceptual way. We will see that such a person is more
difficult to comprehend than one might think.
One occasionally hears fantastic stories
about “clairvoyants” who acquire accurate information through mysterious
channels. We hear of people who
accurately predict a natural disaster or divine where a murder victim is
buried. They presumably don’t witness these
events themselves. Nevertheless there
is some kind of mysterious, underlying causal connection between their beliefs
and the events their beliefs are about.[1] A clairvoyant might therefore be the sort of
person we are trying to imagine.
Admittedly, clairvoyants are often said to have “visions” of the events
in question. The clairvoyant might
“see,” in some sense of the word, an image of the place where a murder victim
is buried. Since having visions might
count as a kind of perception (I will discuss the issue of what counts as
perception in the next section), one might wonder if it is appropriate to use
cases of clairvoyance to illustrate the question of whether it is possible for
someone to acquire her beliefs through non-perceptual channels. I would like to avoid these worries by
limiting the discussion to cases in which such visions are absent. After all, we can easily imagine—and perhaps
we have even heard of—people who just find themselves with a strong (and
unfortunately accurate) conviction that something terrible has happened; the
causal relation between the conviction and the event in question is not
mediated by visions of any kind. These
people are not plagued by images of the misfortune; they just find themselves
believing that the tragedy has occurred.
So
far, we have been discussing cases of what we might call “local
clairvoyance.” The people we generally
hear about are supposedly clairvoyant only with respect to some of their beliefs; the rest of their beliefs are formed in the
usual way. While a clairvoyant may come
to believe that a disaster has occurred through some mysterious causal
mechanism, she generally adopts beliefs about her current, immediate
environment through ordinary perceptual channels. Because of their “local” nature, these cases of clairvoyance seem
far-fetched, but not conceptually
impossible. But what would we say
about a person who forms all of her
beliefs about the world by means of clairvoyance? Suppose this person does not perceive anything at all, but nevertheless
comes to have largely accurate beliefs about some part of the world. There is a causal mechanism through which
she comes to have these beliefs, though this mechanism has nothing to do with
perception. This is a case of what I’ll
call “global clairvoyance.” The
question, then, is whether this situation is ultimately intelligible. If we suppose that the global clairvoyant
perceives nothing at all, can we still hold on to the idea that she has
beliefs?
Before we can make any
progress on this question, it will help to think more about what it is to
perceive something. Otherwise, it is
not clear what we are trying to imagine when we suppose that someone perceives
nothing at all. That will be the topic
of the following section. I will argue
there that perception is characteristically perspectival and that any perceiver
who is capable of having beliefs will therefore have a robust body of
“egocentric” beliefs about the location and orientation of things with respect
to her body. Then in section three, I
will try to show that we cannot suppose that a subject lacks all egocentric
beliefs without making it impossible to view her as having any beliefs at
all. Finally in section four, we’ll see
that it is only through perception that a subject could have the resources to
have any egocentric beliefs. The upshot
is that the scenario we are trying to imagine is incoherent: by supposing that
a person lacks the faculty of perception we make it impossible to see her as a
thinking subject with beliefs about the world.
Perception is a causal
process. When I perceive that it is
raining, there is a causal chain from the meteorological event to my coming to
believe that it is raining.[2] But we generally think that perception
involves something more than just being caused to adopt beliefs. What more, then, does perception involve? We
often speak of perception as if it essentially involves the use of our eyes,
ears, nose, skin, etc. Perception,
according to this way of speaking, is distinguished from other ways of being
caused to form beliefs by the fact that, in perception, the causation takes
place through a person’s sense organs.
This is not to say that this is the only
thing that distinguishes perception; it is just to say that, on this view,
perceiving something necessarily requires the use of one’s sense organs. This
is thought not only to be an empirical truth about human beings, but an
analytic truth based on the concept of perception. If we inspect this idea more closely, we will find that it is actually
the conjunction of two theses. The
first is that one cannot see without
eyes, hear without ears, and so
on. In other words, our sense
modalities (seeing, hearing, etc.) necessarily involve the kinds of sense
organs that human beings in fact possess.
The second is that perception
necessarily involves either seeing, hearing, smelling, feeling, etc. That is, perception necessarily encompasses
all (and only) the sense modalities that
we actually have. It is impossible, on
this view, for there to be a form of perception that is not seeing, hearing,
smelling, feeling, or tasting.[3]
Although we do often speak
as if it is part of what it means to perceive something that the perceived
object is affecting the kinds of sense organs we actually have, I don’t think
it is much of a stretch of the concept of perception to imagine perceiving
things in bizarre ways. We can do this
either by imagining someone who “perceives” by means of a peculiar sense
modality or by imagining someone who has an ordinary sense modality that is not
realized in an ordinary sense organ.
Sydney Shoemaker, in his book Self-knowledge and Self-Identity,
takes the latter route when he considers a case in which we would be tempted to
say that a man is able to “see” things behind his back. He says:
One can
imagine, first of all, hearing someone claim that he sees things from the back
of his head…Of course, someone’s claiming this would not by itself be
sufficient to show that what he claims is true. But it is also imaginable that someone might regularly make true
statements about things situated, or events happening, behind his back, and
that we might find, on investigation, that he does not know the truth of these
statements in any of the ways in which a normal person might know them (he is
not using a mirror, is not receiving signals from a confederate, and so
on). If this happened, and if the
person expressed his statements as perceptual statements, we would certainly be
inclined to say that he is able to see things behind his back.[4]
Shoemaker admits that we normally use the
word “see” as if it were “self-contradictory to say that someone literally sees
things without using his eyes.”[5] Nevertheless, he says that the interesting
question is “whether there could be a form of perception which, whether or not
the current use of the word ‘see’ would permit us to call it seeing, is exactly
like seeing except that it does not involve the use of the eyes.”[6] And Shoemaker thinks there could be such a
form of perception.
Shoemaker
is on to something here. If we insist
that, strictly speaking, perception requires the use of the kinds of sense
organs human beings actually have, then the issue of whether it is possible to
have beliefs without perception is not very interesting. The question I want to address is not whether it’s possible to have
beliefs without having the kinds of sense organs (or sense modalities) that we
happen to possess. The answer to that question is obviously that it is
possible. There is no apparent reason
why the man that Shoemaker imagines would fail to have beliefs. Having eyes or ears is not a necessary
condition for having beliefs about the world.
We can get closer to the question I am interested in by relaxing the
notion of perception so that it does not necessarily involve the possession of
sense organs or modalities that we happen to have. We can then look in more
detail at what circumstances would make us say that someone “perceives”
something, in the looser sense of the word, even though she lacks the usual sense
organs or modalities. What is it to be
“exactly like perceiving” except that it does not involve the use of any sense
organs? In other words, what is the
difference between someone who perceives in that broader sense and a person who
is globally clairvoyant? If we can
identify this difference, we can go on to ask whether it is relevant to one’s
capacity to have beliefs.
The
passage I quoted above from Shoemaker does not say quite enough about why we
are tempted to say that the man “perceives” things from the back of his
head. The mere fact that he is able to
make true statements about objects and events behind his back does not suffice
to distinguish him from a clairvoyant.
Clairvoyants, after all, can also make true statements about various
objects and events in the world. Nor
does it help much when Shoemaker adds that the man makes “perceptual
statements” about the things behind his back.
We don’t merely want to know whether a person says that he sees something; we want to know what a person must be
like in order for what he says to be true. When the man in Shoemaker’s example says, “I
see a car,” he says something true. If
the clairvoyant says the same thing, she says something false. What facts about the two cases would account
for this?
Fortunately, Shoemaker
goes on to say more about the distinction between clairvoyance and sight:
So in the case of the man
who is able to make true statements about events occurring behind his back, we
must ask what would have to be true of the case, in addition to his making
these statements and their being true, in order for it to be correct to say
that he sees those events, as opposed
to knowing of them “by clairvoyance,” i.e., simply being able to make true
statements about them. What is
required, I think, is that he should speak and behave in such a way as to
enable us to pick out some point on his body as the point of view from which he sees.[7]
Shoemaker is saying that a
significant difference between clairvoyance and sight is that we see things from
a particular perspective or point of view.
I think there is something right and important about this. There is something essentially perspectival
about the way objects are represented in visual perception. Compare, for example, what is involved in seeing
that there is a lectern on the table in Howison Library to what is involved in
believing the same thing to be the case.
When I see that there’s a lectern on the table, I see the lectern and
the table from a particular perspective. The perspective from which I see these
things is determined by where they are in relation to my body, or—more
specifically—the “point of view” on my body from which I see them (my
eyes). I see the lectern from a
particular angle or orientation, I see one side and not the other, etc. I not only see the lectern to be on the
table, but I also see the lectern and the table to be in a particular location
and orientation with respect to my point of view. The things I see, in other words, are represented as standing in
a certain spatial relation to me. That
is what it means for visual perception to be perspectival.
My belief that there’s a
lectern on the table in Howison Library, on the other hand, merely represents
the lectern and the table to be in particular locations with respect to each
other. There is nothing in the content
of that belief about where these things are located or oriented with respect to
me. I could, of course, have an additional belief about the positions of
the lectern and table in relation to me.
I may believe, for example, that the table is three feet away on my
left. That belief would represent where the table is in relation to me, however it is
not essential that all of my beliefs are like that, and many of my beliefs are
not. While some beliefs represent
things as standing in a certain spatial relation to the believer, this is not
characteristic of the nature of belief in the way that it is essential to the
way things are represented to us when we see them. It is possible for me to have a belief about something without
believing that thing to be in a certain orientation or location with respect to
me, but it seems bizarre to say that I see something but do not see it to stand
in some spatial relation to the point from which I am viewing it.[8]
Let me try to put this
point another way. Our experiences are
sometimes thought to have “conditions of satisfaction”[9]:
these are conditions that must be fulfilled in order for our experiences to
count as veridical. Visual experiences
are perspectival in the sense that their conditions of satisfaction depend in
part on where the relevant object is actually located with respect to the
subject. In order for the experience I
am having of the lectern on top of the table to be veridical, for example, the
lectern must not only be on top of the table, but must also, along with the
table, stand in a specific spatial relation to the point on my body from which
I see it. Imagine, for example, that
there is a lectern on the table in Howison Library, but there’s also an elaborate
system of mirrors that makes the lectern appear as if it’s in front of me when
in fact it is behind my back.[10] In that case, my perceptual experience would
not be veridical. The conditions of
satisfaction of my belief that a
lectern is on the table, on the other hand, do not depend on where these things
are with respect to me—the belief is true as long as there’s a lectern on top
of the table in question. Or to give
another example, suppose that a number of different perceivers are looking at
the same lectern from different perspectives.
They all believe the same
thing: that there’s a lectern on the table.
We can even say that they all see
that the lectern is on the table. But they nevertheless have different
perceptual experiences with different conditions of satisfaction, depending on
their various points of view.
While Shoemaker is content
to link the idea of a “point of view” with just visual perception, I believe it to be a characteristic feature of
perception in general that the perceived object is represented as standing in a
particular spatial relation to the perceiver’s body. When a blind person hears the voice of a friend, or bumps into
the table in front of him, he hears or feels the relevant things to be in a
certain position (and, at least in the case of the table, orientation) with
respect to his body. And if we come
across a hitherto unknown sense modality—say, if we meet aliens with a very
different physiology than our own—I think our willingness to call that modality
a kind of perception will depend in part on whether the perceived objects are
represented to the creature from a perspective.[11]
Some cases that appear to
be counter-examples to this claim can in fact be seen, on closer inspection, to
have some kind of perspectival component,
however minimal. For example, it may be
possible to hear a sound that does not appear to come from any specific
direction.[12] But even
then, the sound might still be represented as coming from somewhere near the perceiver’s body. My claim that perception is perspectival
need not imply that things are always represented in perception as being in a
very specific location with respect to the perceiver. Other apparent counter-examples have to do with sense modalities,
such as taste or smell, that do not typically provide information about the
layout of the perceiver’s immediate environment. It may seem odd to say that when I bite into a lemon, I taste
the sourness from a particular point of view or that I taste it to be in some
spatial relation to my body. And it
sounds even stranger to say that I smell a rose from some perspective or point
of view. But again, I think that there
is still a kind of perspective at work in such cases. Although I do not taste the sourness of the lemon to be three
feet away on my right, the taste is at least represented as being in my mouth, not in my ear or in my big
toe. And this is not merely to say that
my mouth is the causal conduit through which I taste the sourness of the lemon;
rather, it is part of the content of
the experience that I taste the sourness to be inside of my mouth. This point may be more difficult to make in
the case of smell, since it is not obvious that smells are represented as being
in one’s nose. But even then, smells
still seem to be represented as being in the vicinity of one’s nose.[13]
Someone could insist that
there might be some cases of perception that fail to have even a minimal amount
of perspective. It might be possible to
hear a sound that does not appear to come from anywhere at all: the perceiver
cannot even tell whether the sound is coming from outside of her or if it is
just a ringing in her ears. Even so, I
don’t believe that such cases pose a threat to my claim that perceptual states
are characteristically perspectival. If
one insists that there are perceptual experiences that have no perspectival
element whatsoever, it becomes difficult to maintain that these are cases of perception in the first place.[14]
Perception is essentially representational: a perceptual state represents the world to be a certain way. I have been proposing that part of what
distinguishes perceptual states from other representational states—such as
beliefs—has to do with the fact that they represent the world from a particular
perspective. I think it will be quite
difficult to come up with an example of something we are inclined to call a
perceptual state that is representational, but not perspectival. To the extent that the examples above
appear not to be perspectival, they also fail to be representational. If I hear a sound that comes from no
particular direction, for example, I fail to see how the sound I hear could
have any representational content: it does not represent the world to be a
certain way at all.[15]
An
important consequence of the fact that perception is perspectival is that we
expect perceivers to have a robust body of what I will call “egocentric
beliefs.” Egocentric beliefs are
beliefs whose truth-conditions depend on where the relevant objects are located
or orientated in relation to the believer.[16] They are expressed by statements containing
words such as here, there, left, right, up, down,
which identify things in terms of their position with respect to the
speaker. Because of the perspectival
way in which things are represented in perception, the perceiver will not
merely have beliefs about how things are in the world; she will also have
beliefs about how things are situated in
relation to her. This is not surprising in light of the
obvious fact that we often come to believe that something is the case because
we perceive it to be so. If I see a
lectern on top of a table, I will generally believe that there’s a lectern on
the table. In the same way, if I see
the lectern and table to be on my right,
then I will generally believe that they are on my right.
What I have said so far
may not apply to infants and some non-human animals. It seems reasonable to say that animals and infants have
perceptual states in which things are presented from a particular perspective
or point of view. Yet they will lack
egocentric beliefs if they are not capable of having any beliefs in the first place.
Whether they have beliefs or not depends, of course, on how robust we
take the conditions for having beliefs to be.
If having beliefs requires sophisticated conceptual capacities,[17]
then it is likely that infants and many animals lack beliefs. To be safe, I will restrict my claim that
perceivers have egocentric beliefs to subjects who are capable of having
beliefs. We are trying, after all, to imagine a person who has beliefs but no
perceptual states, not a person who has perceptual states but no beliefs. Nevertheless, there are interesting and
difficult issues here about what exactly animals and infants do get through perception, if we don’t
wish to say that they get egocentric beliefs. Consider a dog that jumps up to snatch a
juicy steak off the counter top. Even
if we do not want to say that the dog has a perceptually based egocentric belief about the location of the steak,
it is clear that he has something
that prompts him to jump up on the counter as opposed to moving in some other
direction. Perhaps we could call it
egocentric “information.” What it is to
have such information is difficult to characterize, though we can start by
saying that it plays a role in the causation of behavior. The dog jumps up on the counter because he
has egocentric information about the location of the steak. The hope is that we can appeal to egocentric
information to explain a creature’s behavior without having to ascribe beliefs
to it (at least if having a belief presupposes robust conceptual capacities
that the creature does not have).
Young children may present
a further complication. Unlike infants and animals, young children clearly have
beliefs. They might be thought, however, to lack egocentric beliefs. Some of
the concepts involved in having egocentric beliefs are quite
sophisticated. Children often do not
master the concepts of left and right, for example, until pre-school or
kindergarten. So here we have a
potential example of perceivers who are capable of having beliefs, but who lack
egocentric beliefs.[18] But perhaps we have reason to be skeptical
that children have no egocentric beliefs.
I assume that young children can point to the objects around them and
say, for example, “That’s a doggie” (or perhaps just “Doggie”). If they can do that, then it still seems
reasonable to ascribe rudimentary egocentric beliefs to these children about
the locations of the relevant objects in relation to their bodies, even though
they have yet to master the concepts of left, right, front and behind. The truth of beliefs expressed by utterances
such as “That’s a doggie” does, after all, depend on where the relevant objects
are with respect to the child who is making the utterance. If there is no dog where the child is
pointing, then the belief expressed by “That’s a doggie” is false.
Before I move on, I would like to make a
few brief points about proprioceptive awareness and its relation to
perception. These points will become
relevant at a later stage of the argument.
Although the issues here can be murky, I think we can say that
proprioception counts as a form of perception, and that the clairvoyant
therefore lacks any proprioceptive awareness of her own body. In the first place, proprioception is
representational: in proprioception, we represent our bodies or limbs to be in
a particular posture or position. I
think it is also plausible that, like other modes of perception, proprioceptive
awareness is perspectival. In cases of proprioception, we represent the
position of our bodies from a particular perspective or point of view. This kind of perspective is admittedly
different than the kind of perspective we have in visual perception. We cannot pick out a specific point on the
perceiver’s body as the point of view from which her body is represented to
her, since it is her body itself that is being represented in the first
place. Nevertheless, there is still a
kind of perspective at work in the sense that we are proprioceptively aware of
our bodies, so to speak, “from the inside.” There is a crucial difference
between the way my body is represented to me when I am proprioceptively aware
of my legs being crossed and the way in which my body is represented in a mere
belief that my legs are crossed.
Although this difference may be difficult to characterize, it does seem
to be related to the perspectival way in which my body is represented to me
when I am proprioceptively aware of it.
If what I have said about the relation
between perception and proprioception is correct, it seems reasonable to allow
the category of egocentric beliefs to include, in addition to beliefs about
where things are in relation to one’s body, beliefs about the position and
movement of one’s body itself. This is
a rather modest extension of the notion of egocentricity. My belief that my arms are at my sides is
not all that different than my belief that there is table in front of me. The difference is merely that the latter
belief is about a relation between my body and an object that is distinct from
it, while the former is about a relation between one part of my body and
another. The important point is that
proprioceptive awareness provides perceivers with beliefs about the position
and movement of various parts of their bodies.
With proprioception, I can acquire all sorts of beliefs about the
movement of my body or the position of my limbs.
I have argued so far that
perception is essentially perspectival and that any perceiver capable of having
beliefs will therefore have some egocentric beliefs. These egocentric beliefs turn out to be very important. I will argue in the next section that if we
suppose that a person has no egocentric beliefs whatsoever, it becomes
difficult to see her as having any beliefs about the world at all. This is because our supposition that the
person fails to have any egocentric beliefs precludes understanding anything
she does as an intentional action. We
cannot take someone who has no egocentric beliefs to be an agent. I will then show that if we cannot see a
person as an agent, it becomes more difficult to think of her mental states as beliefs, and not merely as causal
correlates of events in the world. A
subject with no egocentric beliefs would begin to look more like a complicated
thermometer than a thinking being. Her
mental states would at best serve as reliable indicators of particular states
of affairs; we would not be able to understand them as beliefs about those states of affairs.
As
it stands, this conclusion would still fall short of an answer to our original
question: namely, whether we could coherently imagine someone to be globally
clairvoyant. Due to the perspectival
character of perception, we expect anyone who is able to perceive (and who is
capable of having any beliefs at all) to have some egocentric beliefs. It remains to be seen, however, whether
perceivers are the only subjects who
could possibly have egocentric beliefs.
Is it legitimate to assume that the global clairvoyant has no egocentric
beliefs? If so, then the idea of global
clairvoyance will turn out to be unfathomable: we cannot suppose that a person
lacks the faculty of perception without making her seem more like a thermometer
than a rational agent with beliefs about the world. If not, then although I will identify something important about
the way in which perception causes beliefs, it will still be a contingent
matter of fact that a person’s beliefs are caused through perception. Beliefs could be caused in other ways as
long as the believer ends up with a significant body of egocentric beliefs.
I think it will prove
difficult to ascribe egocentric beliefs[19]
to the global clairvoyant. But since my argument for this claim is rather
involved, I will postpone that discussion until section four. Meanwhile, I propose that we try to imagine
a clairvoyant who neither perceives nor has any egocentric thoughts or beliefs.[20] I will then go on to argue in section four
that this is how we must understand
the global clairvoyant: someone who lacks perception cannot have any egocentric
beliefs.
Here is the scenario I am
trying to imagine: while the clairvoyant cannot perceive anything, she
nevertheless has largely true beliefs about certain states of affairs in the
world. There is some non-perceptual
causal mechanism that is responsible for these beliefs; her beliefs are
generally caused by the events her beliefs are about. But the subject has no thoughts or beliefs about where anything
is located or oriented with respect to her body or about the position of her
body itself. Though she has various
beliefs about how things are in the world, none of these beliefs are about how
things are in relation to her. We might therefore describe the
clairvoyant’s perspectiveless worldview as—to use Nagel’s suggestive phrase—a
“view from nowhere.”[21] Her conception of the world includes
nothing about where she is located
within it. In that way she is
completely detached from the world her beliefs are about.[22]
In
this section, we’ll see that this picture of the clairvoyant is not something
we can coherently imagine. If we
suppose that none of the clairvoyant’s beliefs are egocentric, her
“perspectiveless world view” becomes unrecognizable as a body of beliefs. I will try to motivate this position by
arguing, first, that we must be able to ascribe egocentric beliefs to a person
in order to understand her bodily movements as intentional actions on her
part. I will then make a similar point
about the clairvoyant’s utterances: to the extent that the clairvoyant lacks
egocentric beliefs, her utterances cannot be understood as intentional. Her utterances are therefore more akin to
the readings on a thermometer than to meaningful speech. Finally, we will see that on a reasonable
understanding of what it is to be a mental state of a given kind, the
clairvoyant’s mental states have so little in common with genuine beliefs that
it becomes difficult to think of them as such.
My argument for the claim
that the clairvoyant’s bodily movements cannot be intentional actions draws on
John Perry’s paper “The Problem of the Essential Indexical.” In that paper, Perry makes a persuasive
case for the idea that we must use indexicals to characterize certain beliefs
that are required to give a rational explanation of a subject’s bodily actions.
“These indexicals are essential,” claims Perry, “in that replacement of them by
other terms destroys the force of the explanation, or at least requires certain
assumptions to be made to preserve it.”[23] Although Perry does not speak in these
terms, the kinds of beliefs Perry speaks of are clearly related to what I have
been calling “egocentric beliefs.”
Admittedly, some beliefs expressed by the use of indexicals are not
strictly speaking egocentric (at least as I’ve been using the term). I express my belief that I am happy, for example, with the
indexical, ‘I’, but this belief isn’t egocentric in the sense that, in
expressing it, I do not refer to any object in terms of its location or
orientation with respect to me.[24] But many other beliefs expressed by
indexicals are, in fact, egocentric: I have in mind beliefs that we
characterize with terms such as ‘here’, ‘there’, ‘left’, or ‘right.’ Such beliefs supply what we need to explain
the agent’s behavior in the examples that Perry considers.
One of these examples
concerns a lost hiker in the wilderness.
The hiker studies his map at a trailhead by a lake. He believes that the best way back to the
road from Gilmore Lake is to follow Mt. Tallac trail, but he’s not sure whether
he’s standing beside Gilmore Lake or Clyde Lake. He eventually makes up his mind and proceeds to walk along the
trail. If asked to explain why he moved
in the direction he did, he would say something like, “I came to believe that this is the Mt. Tallac trail and that is Gilmore lake.”[25] These indexical terms are crucial to the
explanation. If he said, instead, that
he came to believe that Gilmore Lake is located at 37.08° N, 122.93° W, it
would cease to have any explanatory force.
He could restore the explanatory force if he added, “And 37.08° N,
122.93° W is here at this very
spot.” But that, of course, is to
add other beliefs expressed by indexical terms. In other words, to understand why the hiker proceeded to walk
down the trail, we must ascribe an egocentric
belief to him. We must take him to
believe that the Mt. Tallac trail is there
in front of him.
Perry
is merely trying to argue that there are some
cases in which understanding why a person acts the way she does requires citing
beliefs that are can only be expressed using indexical terms. But I think that we can extend Perry’s point
to bodily actions generally. Any
rational explanation of a bodily action must involve the ascription of
egocentric beliefs. How, for example,
can we understand why a subject reaches over to her left to grasp her coffee
mug unless we take her to believe that the mug is on her left? Naturally we must also take her to have
particular desires or intentions: presumably she would not have reached for her
mug unless she wanted to drink from it (or put it in the sink, or refill it,
etc.). But the ascription of desires
and intentions alone does not suffice to explain bodily action. My desire to open my window does not by
itself explain why I stand up and walk across the room; we must say, in
addition, that I believe that the window is over
there. Agents generally perform
bodily actions so that they can effect changes in the environment in order to
satisfy some desire.[26] A proper explanation of such actions should
therefore specify not only the desire the agent is trying to satisfy, but also
where the agent believes the relevant objects are located in relation to
her. Otherwise, it is inexplicable why
the agent moves in one direction rather than another. Her behavior is presumably the function of the intentions she has
and her understanding of how she must move her body in order to fulfill those
intentions. That understanding must
involve information about where the relevant things are located with respect to
her. Clearly, then, the subject’s
egocentric beliefs play an essential role in making sense of her behavior. Without them, her behavior is inexplicable
to us.
It could be objected that
the need to cite egocentric beliefs to explain bodily action depends on how the
action—as well as the agent’s motivation for performing it—is described. The same action, for example, could be
alternatively described as walking in
such and such a direction, lifting one’s arms, and pushing the window up from
the sill or, more simply, as opening
the window. To explain why the
agent performs the action as it is characterized in the first description, we
clearly need to cite some of the agent’s egocentric beliefs. We cannot explain why she walks in the
direction that she does unless we ascribe to her beliefs about the location of
the window with respect to her body.
But it might seem that such egocentric beliefs are unnecessary if we are
instead explaining why the agent performs the action as it is characterized in
the second description: as opening the
window. She opens the window, we
might say, simply because she wants the
window to be open. In giving such
an explanation, we do not appear to cite any of her beliefs about the
respective locations of the window and her body.[27] However, I do not think this objection
threatens my claim that without egocentric beliefs, a person’s behavior is
inexplicable. It would be very strange
to say that a person intentionally opens a window even though she has no
beliefs about where the window is located in relation to her. This suggests that the explanation, “She
opened the window because she wanted the window to be open,” proceeds upon the assumption that the
agent has an egocentric belief about the location of the window even though it
doesn’t make an explicit reference to that belief. If we therefore suppose that a person has no egocentric beliefs
at all, explanations such as “She opened the window because she wanted the
window to be open” no longer seem very helpful. In response, we just want to ask, “But how could she do that
without any beliefs about where the window is located?”
It is reasonable to assume
that a person’s behavior cannot count as an intentional action unless there is
some rational explanation of why she performs it. Part of what distinguishes intentional actions from mere bodily
movements is that intentional actions are explicable in terms of the agent’s
beliefs and desires. That is to say,
intentional actions—unlike mere bodily movements—are performed for
reasons. If this understanding of
intentional action is correct, then if it is not possible, even in principle,
to explain why a person acts the way
she does, it becomes difficult to see what she does as an action and not just
purposeless movement on her part. I
think this view about action is plausible and, as far as I can tell, it is
widely held.[28]
I do not mean to say here
that the agent must always be explicitly aware of her reasons for acting. Although we often act after conscious
deliberation, we also often act without first contemplating our reasons for
doing so. When driving from New Jersey
to Massachusetts, I often consciously deliberate about whether I should use the
George Washington or the Tapanzee Bridge.
I weigh the time I expect to be stuck in traffic on the George
Washington Bridge against the time it would take to drive the extra distance to
reach the Tapanzee. But not all
intentional actions fit this paradigm case of practical reasoning. When a driver slams on the brakes to avoid
an accident, for example, she probably does not first engage in conscious
deliberation about the most reasonable course of action.[29] This does not mean, however, that these
kinds of actions are not performed for reasons. The driver still has a reason for stepping on the brake: she
doesn’t want to hit the other car and believes that braking could prevent that
from happening.[30] She need not consciously think these things
to herself before she acts in order to have a reason for acting that way.[31]
So far I have concentrated
exclusively on descriptions of actions that make reference to objects outside
of the agent’s body. I have spoken of
actions such as opening a window or grasping a coffee mug. In these cases, the agent acts on objects in her environment, and so it
is not surprising that she must be understood to have beliefs about the
location and orientation of the relevant objects in relation to her. One could object, however, that I have
overlooked a whole other class of intentional actions whose descriptions do not
make reference to any objects in the vicinity of the agent; they are simply
descriptions of intentional bodily movements qua bodily movements. I
have in mind actions such as raising one’s arm,[32]
tapping one’s toes, snapping one’s fingers, dancing, and so on. In performing these actions, the agent moves
her body, but not for the sake of affecting any changes on the things in her
environment. It might appear, then,
that explanations of such actions do not require any appeal to the agent’s
egocentric beliefs. Why, then, couldn’t
the clairvoyant perform these kinds of actions?
I believe that the
clairvoyant cannot, in fact, be understood as performing such actions. Even if we do not understand an agent’s
bodily movement as an act on any of
the objects around her, we still cannot understand it as an intentional action unless she at least
has the relevant intention to move
her body in that way. And if we look
more closely at the examples in question, it becomes clear how strange it would
be for the clairvoyant to have the kinds of intentions or desires that could
motivate these actions. We have
understood the class of egocentric beliefs to include beliefs about the
position or movement of one’s body. In
lacking a point of view, the clairvoyant lacks not only beliefs about where
things are in relation to her body, but also beliefs about the relative
positions of various parts of her
body: limbs, fingers, etc. Just as she
doesn’t have any beliefs about whether a given table is in front of or behind
her, she fails to have any beliefs about whether her arm is raised in the air
or down by her side. She has no beliefs
about the position or movement of her limbs, whether she is walking or standing
still, whether she is standing up or sitting down, and so on.
Once we suppose that the
clairvoyant has no beliefs about the movement or position of her body, it
becomes difficult to see how the clairvoyant could have an intention to move in
some particular way. The clairvoyant
cannot have an intention to move her body in a certain way if she has no
thoughts or beliefs about how her body is moving. It seems very odd that the clairvoyant could intend to raise her
arm above her head—even if it is only for the sake of raising it—without any
beliefs about the position or motion of her arm with respect to the rest of her
body. A person cannot intelligibly have
an intention to do something without at least having some thoughts or beliefs
that pertain to whether she is currently performing the act in question. This is not to say that the person must know whether she is performing the act
in question. I can intend to impress
someone, for example, without knowing whether I really am impressing that
person. Similarly, one might think that
the clairvoyant could intend to raise her arm even though she does not know what position her arm is currently
in. All I am trying to argue is that
the agent must at least have some thoughts or beliefs about whether she is
performing the action in question. If
the clairvoyant has no thoughts about the position or movement of her arm at
all—if she isn’t even wondering whether
she is currently moving her arm—she cannot intelligibly intend to move her arm
in some particular way. We can
conclude, then, that without any beliefs about the position or motion of her
body, the clairvoyant cannot be understood to have any intentions to move her
body. [33] And without any such intentions, none of the
clairvoyant’s bodily movements can be viewed as intentional actions.
I therefore think it is
safe to conclude that we cannot find any agency in the clairvoyant’s bodily
movements. Once we assume that the
clairvoyant has no egocentric beliefs, we no longer have the resources to
explain why she behaves the way she does.
Without any beliefs about the location of surrounding objects with
respect to her body or even any beliefs about the position of her body itself,
there is no way to explain why the clairvoyant moves in one direction and not
another. It is unfathomable why she would,
for example, walk across the room and open a window if she has no beliefs about
where that window is located with respect to her. Nor could we make sense of her behavior by appealing to an
intention to move her body in a particular way, since she lacks the background
beliefs required to have such an intention.
And if it is not possible to give a rational explanation of her
behavior, we cannot understand anything she does as an intentional action. That is because what distinguishes
intentional actions from mere bodily movements is the fact that intentional
actions are performed for reasons.
So
far, I have concentrated on the clairvoyant’s bodily behavior. I would now like to turn my attention to the
clairvoyant’s verbal behavior. Not surprisingly, the points I have been
making about agency also hold here.
Verbal behavior is, after all, a kind of bodily behavior. When a person
speaks she moves her lips and whatever else (tongue, vocal chords) is required
in the production of sound. But unlike
a lot of our bodily movements (with the exception of sign language, hand
signals, etc.), the utterances we make often have meanings. For that reason, by applying what I have
argued so far to the clairvoyant’s verbal behavior, we get some important
results: it turns out that we cannot take her utterances to be meaningful
expressions of any thoughts or beliefs.
At best, they are merely “symptoms” of various states of affairs. Let us suppose, then, that our global clairvoyant
is able to speak. Suppose that when an
event causes her to adopt a belief, she then utters a sentence that corresponds
to the event in question. So whenever
there is an earthquake in California, for example, she utters the sentence
“California is having another earthquake,” whenever another species of
butterfly goes extinct, she utters, “Another species of butterfly is now
extinct,” and so on. Would we be able
to make sense of her utterances as actions?
In other words, could we see her utterances as intentional, as something she intends
to perform? I think it would be a
stretch to understand her utterances in that way. Like other bodily movements, a person’s utterances can be
understood as intentional actions only if we assume that she has egocentric
beliefs.
This point becomes clear
if we reflect again on what it takes to understand something as an intentional
action. We cannot see something as an
intentional action unless it is at least possible to explain why the person performs it. To apply this idea to a person’s utterances,
we cannot see these utterances as intentional unless it is possible to explain
why she makes them or—to put the point another way—what the subject is doing in making those utterances. What we need to see, then, is whether the
clairvoyant’s lack of egocentric beliefs prevents us from understanding what
she’s doing when she speaks. There are
several things people typically do when they make utterances. They speak to impart information, make
promises, make requests or commands, etc.
The examples I have listed so far all involve having some sort of effect
on another person. When I speak to
impart information my purpose is generally to get someone to believe something.
When I make a request, my purpose is often to get someone to do something. This is just to say that, in many cases, communication is the
purpose of speech. And this involves
having an effect on other people by means of the words we utter.
I
want to argue that we cannot understand a speaker as intending to communicate
anything unless we suppose that she has at least some egocentric beliefs. It is inexplicable how a speaker could
intend to have an effect on a hearer by means of her utterances if she has no
beliefs about where the hearer is located with respect to the source of these
utterances. The source of a speaker’s
utterances is, of course, her body (more specifically, her mouth and vocal
chords). This is true, at least, with
face-to-face communication. There are,
of course, less direct methods of communication where the speaker may not know
the exact location of her hearer. I can
talk to someone on the telephone without knowing exactly where she is calling
from. But even in these cases,
egocentric beliefs play an essential role.
It is not clear how the speaker could intend to have an effect on her
audience by means of her utterances if the speaker had no understanding of
where her telephone receiver is located with respect to her own body.[34] In this way, communicative speech is more
akin to other bodily actions than one might think. Both kinds of acts involve using one’s body to have some effect
on the environment. It is just that in
communication, what we’re trying to affect in our environment is another person.
Moreover, the intended effect is often psychological: we want the other
person to believe something, to feel a particular emotion, etc. These differences, however, do not change
the fact that we attempt to produce such effects by acting on our
surroundings. We make noises that we
hope the other person will hear and understand. This involves having some beliefs about where the other person is
in relation to us, or perhaps—as in the case of telephone calls—where our body
is in relation to the receiver that will carry our voice to the person on the
other end of the line.
Since
the global clairvoyant has no egocentric beliefs, she has no beliefs about
where other people (or telecommunication devices) are located in relation to
her body. This prevents her from
understanding how the utterances she
makes could affect another person. As a
result, I think it is difficult to view her utterances as attempts on her part
to affect anyone, and it is therefore puzzling how she could be acting
intentionally when she makes those utterances.
I do not mean to insist that a speaker must have very specific beliefs
about the physical processes by means of which communication takes place. I do not need to understand completely how
telephones work in order to intend for my utterances into the receiver to have
an effect on the person I’m calling.
Nor must I have any detailed beliefs about the physiology of
speech. My point is merely that if
something a person does is to count as an intentional act, the person must have
at least some beliefs, however vague, about how her action might accomplish
what she intends to do. Otherwise there
is no rational explanation of why the person does one thing rather than
another. And this is exactly what is
lacking in a person who has no egocentric beliefs. If a person has no conception of where anything is in relation to
her body, then she has no beliefs about how what she does with her body—whether
it is moving her limbs or her lips—affects anything around her. In that case, it becomes quite a stretch to
understand any of her physical or verbal
behavior as an intentional action.
Intentional actions are distinguished from unintentional bodily movement
precisely because actions are performed for reasons. The global clairvoyant is missing the beliefs necessary for her
to have any such reasons.
I
have argued so far that we cannot take the clairvoyant’s utterances as attempts
on her part to affect anyone. In other
words, we cannot view her utterances as acts of communication. But perhaps one could object that not all intentional
utterances are acts of communication.
After all, people sometimes talk to themselves when no one else is
around, and we would hesitate to conclude that such speech is
unintentional. So the question remains
as to whether we could still see the clairvoyant’s utterances as intentional without
supposing that she intends to affect anyone by means of those utterances. One reason for thinking that we cannot has
to do with the idea that communication is the primary way in which utterances
get their meaning. On this view, the
utterances we make when we talk to ourselves have meaning only derivatively, by
means of their relation to the utterances we make when we communicate with one
another. The thought here is that we
cannot make sense of a person who speaks only to herself and who never communicates
with anyone else. I think this view of
language is probably correct, though I am not in a position to argue for that
here. And even if it is right, someone
could still insist that there is no reason why the utterances the clairvoyant
makes when she talks to herself can’t be intentional, despite the fact that
they aren’t meaningful.
Nevertheless, it is
difficult to make sense of the clairvoyant intentionally talking to
herself. The clairvoyant cannot
intentionally speak, even if it is only to herself, if she lacks the intention
to do so. And without any egocentric
beliefs, the clairvoyant cannot have any such intention. In this way, talking to oneself is akin to
other kinds of actions that we can describe without making reference to any
objects (in this case, other people) in the agent’s environment. We saw before
that having an intention requires some thoughts or beliefs about whether one is
performing the act in question. A
person cannot have an intention to raise her hand in the absence of any beliefs
about the position or movement of her hand with respect to the rest of her
body. It is not surprising, then, that
the clairvoyant cannot have an intention to speak unless she has at least some
beliefs about whether she is currently speaking, whether her lips are moving,
whether there are sounds coming out of her mouth, etc. As before, I think it is fair to assume that
in lacking egocentric beliefs, the clairvoyant lacks these kinds of beliefs as
well. For that reason, I do not see how
we could view her as intentionally making any
of her utterances.
I
have been trying to show how difficult it is to find bodily agency in a person
who has no egocentric beliefs. This
goes not only for bodily movements such as walking across the room or raising
one’s hand, but also for the verbal utterances one makes. In fact, my purpose has been to bring out
the fact that speaking is a bodily activity like any other and, like other
bodily actions, requires having some egocentric beliefs. In the absence of such beliefs, it is
inexplicable why the clairvoyant makes any of her utterances, and so her
utterances cannot be understood as intentional acts on her part. This has an interesting effect on how we
should understand the meaning of the
clairvoyant’s utterances. It will be
helpful in this context to apply Grice’s distinction between natural and
non-natural meaning. To use one of
Grice’s examples: when the doctor says, “Those spots mean measles,” she is
using “mean” in the natural sense.
Nobody means anything by the
spots. It’s just that the spots on the
patient’s skin are a reliable sign that the patient has measles; there is a
causal connection between the appearance of spots and the presence of
measles. On the other hand, “That
remark, ‘Smith couldn’t get on without his trouble and strife,’ meant that
Smith found his wife indispensable,” is an example of non-natural meaning. What makes it non-natural is the fact that someone meant something by that remark;
someone made that remark in order to convey that Smith could not get by without
his wife.
Grice characterizes
non-natural meaning as follows. In
order for a speaker to mean something by an utterance, he must intend to induce
a particular belief in his audience by means of that utterance. He must also intend for his audience to
recognize that he intends to induce that belief by making the utterance; the
audience’s recognition of his intentions is supposed to play a part in inducing
the belief in question. For example,
when Smith’s colleague says, “Smith couldn’t get on without his trouble and
strife,” he intends his audience to come to believe that Smith finds his wife
indispensable, and he thinks that his audience will come to adopt this belief
by recognizing his intention to induce that belief.[35] Philosophers of language argue about the adequacy
of this characterization of non-natural meaning, and it is not my purpose to
enter the details of that debate here.
But I think that the general idea is very plausible: what distinguishes
natural from non-natural meaning has to do with an intention on the part of the speaker. To say that someone meant
something by a remark is to imply, at the very least, that the person intended to make that remark. The key difference between the man who says
“It’s hot outside” and the level of mercury on the thermometer is that the
man’s remark is intentional. The
thermometer does not intend to do anything.
And we can accept this basic point without committing ourselves to the
details of Grice’s account.
It
should now be clear that the clairvoyant’s utterances seem more like the level
of mercury on the thermometer or the spots on the patient’s skin than the man’s
remark about Smith and his wife. On the
assumption that there are causal regularities between the clairvoyant’s
utterances and events in the world, we can use her remarks as reliable
indicators of various occurrences in the way that the level of mercury in the
thermometer is a reliable indicator of the temperature. But we cannot say that she means anything by the remarks she
makes. We are barred from
understanding her utterances in that way precisely because we are barred from
thinking of them as intentional actions.
To mean something by a remark requires, at the very least, that one
intends to make it.
Our global clairvoyant is
beginning to look quite impoverished.
Nothing she does can be understood as an action. There is no rational explanation for any of
her bodily movements. Any utterances
she happens to make could have only natural meaning, assuming they have any
meaning at all. Even if we find causal
regularities between her utterances and events in the world, we cannot say that
she means anything by her
remarks. In this sense, her utterances
seem more like the readings on a thermometer than expressions of beliefs. Can we still hold on to the idea that the
clairvoyant has beliefs? In light of
all of the things we are barred from ascribing to the clairvoyant, I think it
is a difficult idea to maintain. Whatever
mental states she has are entirely unconnected to anything else about her. Her beliefs are like unconnected cogs in a
system of gears—they have no recognizable rational role to play at all. All we have here are mental events that are
regularly caused by various events in the world. It is tempting to say that these occurrences in the clairvoyant’s
mind are not all that different from the utterances that accompany them. Her mental states have “content” in the
sense that they indicate certain
states of affairs just as her utterances have “meaning” in the sense that they
are symptoms of the events they accompany.
Whether we still call
these mental occurrences beliefs
depends, of course, on what sorts of things we take beliefs to be. I take the following view to be plausible:
what it is to be a belief is defined, at least in part, by the rational role
that beliefs play in a person’s psychology.
This is partly how beliefs are distinguished from other kinds of mental
states such as desires, experiences, or mental images. And the role of belief is intimately
connected with action. Our beliefs, in conjunction with our desires, give us
reasons for action.[36] Clearly, the clairvoyant’s mental states
cannot play this role. None of her
mental states could help explain anything she does. I see no reason, then, for calling them beliefs.[37]
To maintain otherwise would commit us to a conception of mental states that we
might not be comfortable embracing. I
think we would have to think of mental states as completely self-standing
mental items that could be the types of states that they are without having any
rational connections to anything the subject does. Their character would have to be totally independent from the
role they normally play in our rational psychology. Someone could perhaps insist on such a view about the nature of
mental states. This would, of course,
require some account of what distinguishes beliefs from other mental states
(e.g., emotions, desires, etc.) and from things that aren’t mental states at
all (e.g., readings on a thermometer) that makes no reference to any rational
connection between beliefs and action.[38] I suspect that this will be a difficult
task. I will not defend that suspicion
here[39], but I think it is reasonable to assume—as
many philosophers do—that one cannot count as a believer unless those beliefs
could potentially serve as reasons for action.
And on that assumption, the clairvoyant more closely resembles a
thermometer than a thinking subject.[40]
In
this section, I have tried to argue that if we take seriously the idea that the
clairvoyant has no beliefs about the location and orientation of things with
respect to her body, we are hard-pressed to understand her as having any
beliefs at all. This is because without
such egocentric beliefs, we cannot see any agency in the clairvoyant’s
behavior. If nothing she does can be
understood as intentional, the clairvoyant’s mental states have no role to play
in her rational psychology. For this
reason, it seems forced to think of her mental states as beliefs, as opposed to
mere symptoms of the states of affairs in the world with which they are
correlated. The clairvoyant just seems
too impoverished to have a genuine mental life.
I began by asking about
the significance of the fact that it is typically through perceptual channels
that we are caused to form beliefs. We have seen that perception is significant
because it provides us with egocentric beliefs: as perceivers, we do not merely
have beliefs about how things are in the world; we also have beliefs about how
things are in relation to us. And it now looks as if we must possess such
egocentric beliefs in order to have any beliefs at all. At this point, someone might object that I
have merely identified something we
happen to get through perception that turns out to be essential for having
beliefs. Unless it can be shown that only perceivers can have egocentric
beliefs, then while I may have identified something important about the way in
which perception causes beliefs, I have yet to show that our beliefs must be caused through perception. They could be caused in other ways as long
as we end up with a sufficiently robust body of beliefs about the location of
various objects with respect to our bodies.
What we need to do now, then, is to revisit the connection between
egocentric beliefs and perception and the question of whether it is in fact
legitimate to assume that the clairvoyant has no egocentric beliefs.
Let us try to imagine a
clairvoyant who has no ability to perceive anything around her, but who is
nevertheless caused through an alternative mechanism to have many accurate
egocentric beliefs about the location and orientation of objects in her
immediate environment. Is this scenario
coherent?[41] We will see that it is quite difficult to
imagine. But first we need to look
briefly at one bad reason for
thinking that egocentric beliefs are impossible without perception. To have an egocentric belief about some
object, on this view, the believer must identify or refer to that object
demonstratively. I believe, for
example, that this table is two feet
away from me; in having this belief, I refer to the table by
demonstration. Now it is commonly
thought that genuine cases of demonstrative reference essentially involve
perception: pure demonstrative terms only succeed in picking out objects that
the speaker is perceiving.[42] If I say, “That table is two feet away,” I only succeed in referring to the
table, and thereby having a thought about it at all, insofar as I am now
perceiving it.[43] In that case, it would seem that the
clairvoyant could not have any egocentric beliefs since she would be unable to
refer demonstratively to any of the objects around her.
An obvious problem with
this line of argument is that it seems the clairvoyant could get along by
referring to the objects around her by
description. She could, for
example, refer to a table as the table in
front of her, provided that she has a way of distinguishing her front from
her back. To do that she needs only to
have some sort of bodily asymmetry: if she looks anything like us, she could
easily identify her front descriptively as the side of her body that her face
is on.[44] And even if we supposed that the
clairvoyant’s body was symmetrical in all respects—even if she were a perfect
sphere—this still would not prevent her from having the kinds of beliefs she
would need in order for us to see her as an agent. We can imagine that she could identify geographic locations by
means of a fine-grained system of coordinates; perhaps she has an internal
global positioning system. It seems
that she would then be capable of referring to objects by description in terms
of their latitude and longitude. She
could believe, for example, that there is a table at x degrees north and y
degrees west. If she also had beliefs
about her own location, this might be enough to make her actions
explicable. We could explain why she
walks in a particular direction by appealing to a belief that she is in one
particular location and a belief that an object she wants is in another. While I am not sure if these strictly count as
egocentric beliefs, they seem to resemble them closely enough to allow us to
find agency in what the clairvoyant does.
It might seem, then, that the clairvoyant could get by with just one
indexical: I. She could refer to the objects around her by
means of descriptions using that indexical.
At the very least,
however, she must be capable of using the first person indexical to refer to
some person as herself. This, I will argue, is something she cannot
do without perception. In this way, we
will see that the clairvoyant is incapable of meeting a critical requirement
for having an egocentric belief. And by
“egocentric belief,” I now mean any belief about the location or orientation of
an object with respect to one’s body (regardless of whether the object in
question is identified demonstratively or descriptively) or any belief about
the position or location of one’s body itself.
The requirement is this: a person who has an egocentric belief has a
belief about where things are in relation to someone she believes to be herself,
not merely about where things are in relation to some person who may in fact be her. My belief that there’s a table in front of
Cheryl Chen counts as an egocentric belief only if I realize that I am Cheryl Chen. This is exactly the sort of thing, we will
see, that the clairvoyant lacks the resources to understand. Even if we suppose that it is possible for
her to have beliefs about the location and orientation of things with respect
to some person who may in fact be her, I will argue that she would not be able
to believe that she is the person in question. For that reason, it is impossible for the clairvoyant to have
any egocentric beliefs.
Anscombe
has argued that one does not need to perceive anything in order to have a first
person thought. I could be anesthetized
and suspended in a pitch black tank of tepid water and still exclaim, “I won’t
let this happen again!”[45] But even if Anscombe is correct, it would
not affect the point I am trying to make.
For a subject’s belief to count as egocentric, it is not enough for her
merely to be capable of speaking or thinking in the first person. Rather, she must identify herself with the
sort of thing that could stand in spatial relationships to three-dimensional
objects. The entity she takes herself
to be must be an embodied person, not a mere incorporeal Cartesian
subject. In other words, I am
interested not simply in self-identification, but in bodily self-identification.
And nothing in Anscombe’s discussion indicates that we do not need to be
perceivers in order to think of ourselves as identical with embodied
persons. In fact, the whole point of
her example is that when I use the word ‘I’ in the sensory deprivation chamber,
it cannot mean “this body” or “this person.”[46]
Anscombe thinks this suggests that ‘I’ must either refer to a Cartesian Ego or
to nothing at all (in the end, she concludes that ‘I’ is not a referring
expression).
To see why bodily
self-identification is impossible without perception, it is helpful to understand
why, as normal perceivers, it hardly makes sense to question whether we are
identical with the subjects of our egocentric beliefs. The way in which we arrive at our egocentric
beliefs through perception gives rise to a phenomenon that some philosophers[47]
have called “immunity to error through misidentification.” To understand this phenomenon more clearly,
let us first consider a belief that is not
immune to error through misidentification.
Suppose I hear the faucet running and come to believe that my brother is
in the kitchen. I could be mistaken in
two different ways. First, I could be
wrong that my brother is in the kitchen,
as opposed to, say, the dining room.
(He could have accidentally left the water running.) Second, I could be correct in thinking that someone is in the kitchen, but wrong in
thinking that it’s my brother
(perhaps an intruder has broken into my kitchen to wash his hands). This second kind of mistake is what’s meant
by “error through misidentification”: I have mistaken someone else for my
brother.
By contrast, take my
egocentric belief that there’s a table in front of me. I can, of course, be mistaken that there is,
in fact, a table in front of me. Maybe
I have mistaken a chair for a table, or maybe there is nothing in front of me
at all. But if I believe that there is
a table in front of me because I see
a table (or what looks like one) there before me,[48]
I am immune to an error of another sort: the source of my error in thinking
that there is a table in front of me cannot be that I have mistaken someone
else for myself. While it makes perfect
sense for me to say, “Someone’s in the kitchen, but is it my brother who’s in
the kitchen?” it doesn’t make sense to say, “There’s a table in front of
someone, but is the table in front of me?”[49] That question presupposes that I have reason to believe that the table is in
front of someone. But when I see the
table there before me, I come to believe that the table is in front of someone by means of coming to believe that it’s
in front of me. If I am wrong that the
table is in front of me, then I do not have any reason to believe that it’s in
front of anyone else either. [50]
Of
course, there are other ways of coming to believe that a table is in front of
me that aren’t immune to error through
misidentification. Suppose, for
example, that I am blindfolded and I hear someone say, “That woman behind the
table is wearing a blindfold.” If I
assume that the person is speaking about me, I conclude that there is a table
in front of me. As a matter of fact, I
am not facing a table at all; unbeknownst to me, there’s another blindfolded
woman in the room who is facing a
table. It was she to whom the speaker was referring, not me. In this case, I am correct in believing that
there’s a table in front of someone, but wrong in thinking that it is in front
of me. While I do arrive at my belief
through perception insofar as I hear
someone talking about the blindfolded woman behind the table, the belief is not
immune to error through misidentification since I did not myself see or feel
the table in front of me. In order for
an egocentric belief to be immune to error through misidentification, the
subject must herself perceive the relevant objects to be in the appropriate
relation to her body.[51] While it is possible to have
egocentric beliefs that are not immune to error through misidentification, we
nevertheless arrive at a large portion of our egocentric beliefs in a way that
does guarantee immunity to error. In
the majority of cases, it is hardly appropriate to ask whether the person who
has an egocentric belief realizes that the belief is about her as opposed to someone else.
That is because she cannot possibly misidentify the subject of the
belief: there’s no room for her to mistake another person for herself in this
context.
What is not worth
questioning in the case of a normal perceiver becomes a live issue for the
global clairvoyant. The clairvoyant
supposedly has many beliefs about how things are in some region of the world. Some of these beliefs are about the location
and orientation of objects with respect to various people. She might believe, for example, that there
is a person sitting on a bench outside of a building called Moses Hall, that
there is someone else standing in the doorway of that building, that there is a
table in front of a third person in Howison Library, and so on. In order to have egocentric beliefs, she
must believe that one of these people is herself. She must understand, for example, that she is identical with the
person facing a table in Howison Library.
How, then, could the clairvoyant identify that person as herself as
opposed to someone else? If she had
arrived at this belief the way a perceiver would—by seeing the table in front
of her and the walls of the library around her—there would be no question about
whether it is she who is facing the
table in Howison Library. We have
supposed, however, that this is not how the clairvoyant comes to have beliefs;
she is caused to have beliefs by the events they are about, but the causation
does not take place through perceptual channels.
What would it be, then, for the
clairvoyant to believe that a particular person in the world is herself? To put the question another way: what
exactly is it that the clairvoyant could mean when she thinks to herself, “I’m the person behind the table in
Howison Library”? To begin with, the
clairvoyant’s belief must be quite different in character than what an amnesiac
learns when he recovers his identity and figures out, for example, that he is John Smith. The amnesiac has already identified himself
with a particular embodied person; what he discovers are just further facts
about the person that he has already identified as himself (what his name is, what
he did for a living, who he’s married to, etc.). Instead, if the clairvoyant believes anything when she believes
that she is identical to the person by the table in Howison Library, it must
have more to do with what we might call “body-ownership.” In believing that she is the person in Howison Library, the clairvoyant must believe,
in effect, that the body standing in Howison Library belongs to her. What, then, would it be for the clairvoyant
to believe that a particular body is hers?
We will see that it is difficult to imagine what the content of such a
belief could be in the case of the clairvoyant.
It seems plausible that a
notion of body-ownership is connected with either perception or action (or
both). After all, I have a special kind of awareness of my body that I do not
have of the bodies around me, I perceive the world from the point of view of my
body and not some other, and I have a kind of control over my body that I don’t
have over anything else. These seem
like natural facts to exploit in an account of what it is for a body to belong
to me. But we obviously cannot appeal
to any connections between body-ownership and perception in the case of the clairvoyant. The clairvoyant cannot believe that a
particular body belongs to her in the sense that she has the kind of special
awareness that we generally have of our own bodies[52];
nor can some body belong to her by means of being the origin of her perceptual
point of view. While we, as perceivers,
are sensationally or proprioceptively aware of our bodies, so to speak, “from
the inside,” and while it is from the perspective of our bodies that we view
the world around us, this basis for body-ownership is clearly unavailable to
the clairvoyant. Instead, if the
clairvoyant has any conception of body-ownership at all, it must be in some way
linked with action: or, more specifically, with bodily-control. If there is
anything that the clairvoyant could realize, it must be that a particular body
belongs to her solely in virtue of the fact that she can control its
movements.
I doubt that any other
notion of body-ownership would give us the resources to understand her as
performing any intentional actions. The
only alternative would be for the clairvoyant to identify her body as the body
that is the causal basis of her
thinking: what makes a particular body her own, on this view, is that she would
cease of have any mental states if that body were destroyed.[53] The problem with this view, however, is that
we can imagine scenarios in which the causal basis of a person’s thoughts is
displaced from the rest of her body.
This kind of scenario is the subject of Daniel Dennett’s paper, “Where
am I?” Dennett imagines that his brain
is extracted from his body, but through a sophisticated system of transmitters,
he continues to receive perceptual information from his sense organs and to
control his bodily movements.[54] Although Dennett’s body (sans brain) would
not be the causal basis of his thoughts (his thinking would continue even if
his body was destroyed), it would not for that reason cease to be his body. This suggests that body-ownership has more to do with perception
and body-control than with the causal basis of thought. If we are trying to explain intentional
action by appealing to a subject’s egocentric beliefs, we are interested in
beliefs about where things are with respect to the body that is performing the
act in question; beliefs about where things are with respect to the causal
basis of the subject’s thoughts seem beside the point. For that reason, the clairvoyant’s beliefs
about herself cannot play a role in explaining her bodily behavior unless she
identifies herself as the sort of thing that is capable of bodily action, and this seems to involve some notion of
bodily control.
The question, then, is
whether the clairvoyant has the capacity to believe that she has control over
some body. I will argue that without
perception, the clairvoyant is incapable of having such a belief. To see why, we first need to get clear on
what kind of control it is that we have over our own bodies. Obviously, not every thing that a person
controls counts as her body. A
puppeteer controls the movements of a puppet, but that doesn’t mean that the
puppet is part of her body, though skilled puppeteers might use such metaphors
as a manner of speaking. The mere ability to control the movements of an object
does not make that object identical with (or a part of) one’s body. In order to characterize a notion of
body-ownership in terms of bodily control, we must therefore distinguish
between the kind of control we have over our own bodies and the kind of control
we have over other objects. To use
Descartes’ famous metaphor, we are not in our bodies the way a pilot is in a
ship. As Arthur Danto puts it, “We do
not turn, as it were, an inner wheel in order, through some elaborate
transmission of impulse, to cause an external rudder to shift, and in so doing,
get our boat to turn.” [55] We are more intimately connected with our
bodies when we act.
The problem for the
clairvoyant, however, is that it is difficult to characterize this kind of
intimate connection without some appeal to proprioceptive awareness. When I raise my arm, I feel myself raising my arm.
The special kind of control we have over our bodies is a control that is
in some way governed by proprioception and bodily awareness. Because of this, there is generally no
question about whether it is me who is raising my arm. My beliefs about whether I am controlling my
body are usually immune to error through misidentification: in most cases, it
does not make sense for me to ask, “Someone is raising my arm, but is it me who
is raising my arm?” This seems to be
part of what distinguishes the kind of control we have over our bodies from the
control the pilot has over his ship.
The pilot can always wonder, “Someone is controlling the ship, but is it
me?” The pilot finds correlations
between the way he turns the ship’s wheel and the subsequent movement of the
vessel. He assumes that he is the one
controlling the ship, but it is always possible that these correlations are
merely accidental. He might, for
example, be turning a fake wheel, while the real pilot of the ship controls the
rudder from below the deck.
Lacking
perception, the clairvoyant cannot feel herself exert control over her
body. So once again, what is not worth
doubting in the case of the normal perceiver becomes an open question for the
clairvoyant. In what sense, then, could
the clairvoyant believe that she is the cause of the movements of a given
body? It seems that this could only amount
to a belief that there are regularities between her “volitions” or “acts of
will” and the way the body moves. The
clairvoyant would indeed be related to her body (assuming we can say it is
hers) like a pilot in a ship. She would
somehow have to turn an “inner wheel” in her mind and then come to believe that
her body moves in a corresponding way.
The result is this: if the clairvoyant could believe that a particular
body is hers, it would be in virtue of the fact that she believes that there
are correlations between particular kinds of volitions and various bodily
movements. She would have to believe,
for example, that whenever she has a volition to move her arm, the arm on the
body in question moves. This must be
the only sense in which the clairvoyant could possibly believe that she has
control over a given body.
I
fail to see how the clairvoyant could have such a belief. To begin with, whether there are such things
as “volitions” or “acts of will” is a contentious issue: many philosophers question
whether an intentional action can be separated into two distinct events: the
bodily movement and some inner mental episode that is supposed to cause
it. But even if we grant for the sake
of argument that such mental events exist, we will see that the clairvoyant
does not have the resources to correlate types of volitions with types of
bodily movements. We are trying to
understand how the clairvoyant could believe that whenever she has a volition
to move her arm, the arm on a particular body moves. It seems that in order for the clairvoyant to have such a belief,
she must at least be able to identify some of her mental states as volitions to
move her arm. In other words, she must
be capable of believing, of a given mental state, that it is volition of that
kind. Now it would be very odd for her
to believe that a mental state of hers is a volition to move her arm unless she
has some understanding of what it is for a mental state to count as a volition
to move her arm as opposed to a volition of some other kind. Otherwise it is difficult to see what the
content of her belief that a mental state is a volition to move her arm could
possibly be. The problem for the
clairvoyant, however, is that without perception, she cannot distinguish
between different kinds of volitions in a way that would make it possible for
her to have a non-trivial belief that her volitions are correlated with bodily
movements.
This will become clear if
we look more closely at what it could be for a mental event to count as a volition
of a particular kind. What, for
example, makes a mental event count as a volition to move my arm as opposed to
a volition to touch my toes? There seem
to be two ways to respond to this question.
The first is to define volitions functionally in terms of the type of
bodily movements they typically cause.
On this view, a mental state counts as a volition to move my arm just in
case, under normal circumstances,[56]
it causes my arm to move. The fact that
it causes that kind of bodily movement and not some other is what distinguishes
this volition from a volition to cross my fingers or a volition to touch my
toes. The second response resists this
functionalist characterization of volitions by insisting that volitions are
distinguished, not in terms of the bodily movements they generally cause, but
by their intrinsic qualitative features. Each kind of volition, on this view,
has a distinctive qualitative character in virtue of which it counts as the
kind of volition it is—regardless of what type of bodily movement it typically
causes. I do not intend here to argue
against either of these ways of understanding volitions. Instead, I will show that neither way will
be of much help to the clairvoyant.
First of all, it is
clearly unhelpful to suppose that the clairvoyant distinguishes her volitions
merely in terms of the bodily movements they typically cause. If she believes that a mental state counts
as a volition to move her arm solely in virtue of the fact that it typically
causes her arm to move, then her belief that her volitions to move her arm are
correlated with certain arm movements turns out to be trivial. Her belief would amount to nothing more than
the belief that certain kinds of arm movements are correlated with the mental events that typically cause them.[57] Although this is true, it is not the sort of
substantive belief that could support a notion of body-ownership or bodily
control. It would be as if a scientist
announced that he has discovered a correlation between the presence of X and
cancer. But when pressed for details about what X is, he can say nothing more
than “Well, it’s the cause of cancer.”
This would not be much of a discovery at all; we do not need any
scientists to tell us that the cause of cancer is correlated with cancer.
What the clairvoyant
needs, then, is a way of understanding what makes a given mental event count as
a volition to move her body in a certain way that does not consist solely in
the fact that events of that kind typically cause such bodily movements. This leads us to the suggestion that each
kind of volition has a unique qualitative feature in virtue of which it counts
as the kind of volition it is. The
following example makes this suggestion sound plausible. Suppose that a mad neuroscientist has
re-wired my nervous system so that whenever I try to move my left leg, my right
leg moves instead.[58] Insofar as this is imaginable, it might seem
that my volition to move my left leg has some kind of intrinsic qualitative
feature. It is in virtue of this
feature that my mental state counts as an attempt to move my left leg, even
though it is correlated with movement in the right. The difficulty, however, is to specify what these qualitative
features could be in a way that makes no appeal to any perceptual or proprioceptive
awareness. As we saw before, the kind
of control we have over our bodies when we act seems to be intimately connected
with proprioceptive awareness. If we
attempt to isolate a “pure act of will” from that awareness, it is not clear
that we are left with something that has any qualitative features
whatsoever. When I try to move my leg,
I do not find some separate mental act of will (an “inner turning of the
wheel,” as Danto might put it) that precedes my bodily movement; I just feel
myself moving my leg. And if my leg
does not move, or if the wrong leg moves, I assume I will still have some kind of proprioceptive awareness of
physical exertion. Alternatively,
suppose that someone challenges me to make my hair stand on end—not, of course,
by moving it with my hands or by sticking my finger into an electrical
outlet—but by “sheer force of will.”
Without any proprioceptive awareness of my hair, it is difficult to know
even how to attempt to meet that
challenge. I might find myself
squinting my eyes, concentrating on my hair, and tightening the muscles in the
back of my neck. But in doing this, I
not only fail to make my hair stand on end; I fail even to have a volition to make my hair stand on end.[59] I cannot imagine what it would be like to
have such a volition.
It is therefore completely puzzling to me
how volitions could have qualitative differences that would be available to the
clairvoyant. How could she distinguish
a volition to raise her arm from a volition to move her legs if she cannot even
feel her arms or legs move? If the
qualitative difference between these two kinds of volitions is to be of any use
to her, it would have to be a feature that such mental states could have
completely independently of any proprioceptive awareness. But it is difficult to imagine what such a
difference could be, for any qualitative difference between those two volitions
(assuming there is any) seems inextricably tied to the proprioceptive awareness
an agent normally has in performing such actions. We can conclude, then, that the clairvoyant cannot identify any
of her mental states as volitions, and so cannot have any beliefs to the effect
that her volitions are correlated with bodily movements. She therefore lacks the resources to believe
that she has control over some particular body in the world.
Let me try to summarize the argument I
have put forward in this section. I
have been considering whether it is possible for the clairvoyant to have any
egocentric beliefs. I began with the
following uncontroversial premise: in order to have an egocentric belief, the
clairvoyant must be able to believe that a particular person in the world is
herself. She must, for example, believe
that she is the person standing by
the table in Howison Library. This led
to the question of what the content of this belief could possibly be. I suggested that such a belief must have
something to do with body-ownership: if the clairvoyant believes anything at
all, it must be that a given body belongs to her. Since the clairvoyant lacks perception, there is only one
understanding of body-ownership that is available to her: the clairvoyant could
only believe that a given body belongs to her in the sense that she has control
over it. And without perception, her
belief that she has control over a particular body could only be a belief to
the effect that there are correlations between her volitions and various bodily
movements. We have just seen, however,
that it is unfathomable how the clairvoyant could have such a belief.[60] To believe that such correlations hold, the
clairvoyant would have to be capable of believing that a given mental event is
a volition of a particular kind, and that in turn requires that she have some
non-trivial conception of what makes a mental event count as one kind of
volition as opposed to another. But no
such conception is available to the clairvoyant in the absence of
perception. She cannot simply
distinguish her volitions in terms of the kinds of bodily movements they
typically cause or else her belief that her volitions are correlated with those
bodily movements would be completely vacuous.
Nor can the clairvoyant distinguish volitions in terms of their
qualitative features, for if volitions have any qualitative features at all,
these features must be inextricably linked to proprioceptive awareness.
For these reasons, I think it is
quite difficult to maintain that someone could have egocentric beliefs without
perception. As we have seen in section
three, we could not understand someone who lacks all egocentric beliefs as a
genuine agent. And if we cannot view a
subject as an agent, we cannot make sense of her as having any beliefs at
all. Putting these theses together, we can
now see why it is difficult to comprehend the possibility of global clairvoyance. The global clairvoyant was supposed to
acquire all of her most basic beliefs through some non-perceptual causal
mechanism. But now it is clear that it
is not enough for a person’s beliefs to be caused by the things they are about;
it also matters how those beliefs are
caused. Unless a person’s beliefs[61]
are caused through perception, they
are merely symptoms of various conditions or events in the world; they no
longer count as beliefs. Perhaps the
global clairvoyant would be very useful as a tool for us to determine what is
going on in the world. We could consult
her instead of watching the evening news.
But however useful she may be, she would still be more like a thermometer
than a genuine person with beliefs about the world. The problem is that perception is closely linked to many other
capacities that we have. If we take
away a person’s ability to perceive, we are forced to take away so many other
things that she ceases to resemble a person.
[1] An exception to this might be
clairvoyants who can accurately predict future
events. It would be very odd to say
that an event causes someone to form
a belief if the event occurs after
the belief is formed. But these are not
the kinds of cases I will consider here.
[2] This may not be completely correct in
situations where I already believe that it’s raining before I actually perceive
that it is.
[3] The distinction between these two theses
was pointed out to me by Daniel Warren.
[4] Shoemaker, Self-Knowledge and Self-Identity, 173.
[5] Ibid., 174.
[6] Ibid.
[7] Ibid., 175.
[8] I don’t mean to say here that we must
see things as standing in a very exact spatial relation to us (e.g., 3 feet 4
inches away, 34 degrees to the left).
[9] See, e.g., Searle, Intentionality, chap. 2.
[10] These kinds of examples are often cited
to support a causal theory of perception.
This might raise the question of how my view is in fact related to the
causal theory. As it happens, I don’t
think my view is incompatible with that theory. I’m saying only that it’s part
of the conditions of satisfaction of a perceptual state that the objects in
question are in the appropriate location with respect to the perceiver; I’m not
saying that this condition is sufficient for
the veridicality of a perceptual state.
In addition, one could insist that a particular causal relation must
hold between the subject and the things she perceives.
[11] This test will also apply to known,
though unusual, sense modalities such as echolocation in dolphins or bats.
[12] I owe this counter-example to Jason
Bridges.
[13] We might even be able to say that smells
are often represented as coming from a particular place (e.g., the smell of
apple pie coming from the kitchen). The difficulty, however, is to distinguish
between what is included in the representational
content of the olfactory experience itself and what is inferred from the experience.
Is it part of the content of the experience that the apple pie smell is
coming from the kitchen, or do I infer that the smell originates from the
kitchen from the fact that it gets stronger as I approach the kitchen
door?
[14] I owe this response to Daniel Warren.
[15] This raises the difficult question of
what to call these states, if they aren’t cases of perception. We might call them “sensations” or “raw
feels,” but (as Gil Harman has pointed out to me) even sensations are typically
presented as having a location in the subject’s body.
[16] We will see in section 4 that this definition doesn’t quite capture what I mean by an egocentric belief. It is important that the beliefs are about the locations of things with respect to someone the believer takes to be herself. To count as egocentric, it is not enough for a belief to be about the location of something in relation to someone who, unbeknownst to the believer, happens to be identical with herself. For example, Oedipus’ belief that there is a table to the left of the slayer of Laius does not count as egocentric as long as he fails to realize that he is the slayer of Laius.
[17] According to Evans’s “generality
constraint,” for example, someone cannot have a belief that a is F unless she’s
capable of believing, first, that a is G, that a is H, etc., and second, that b
is F, that c is F, etc. (Evan, Varieties
of Reference, 100-105). In that
case, a creature would have to be quite sophisticated in order to have any
beliefs.
[18] I owe this point to Daniel Warren.
[19] In fact, we’ll see that it’s difficult to ascribe any egocentric thoughts to the clairvoyant at all. It will be hard to see how she could even wonder whether, say, there is a table in front of her.
[20] From now on, I will generally speak of
egocentric beliefs, though it should
be understood that the clairvoyant doesn’t entertain any egocentric thoughts at all.
[21] Nagel, View from Nowhere.
[22] As I have been describing her, the clairvoyant
bears some resemblance to the two gods in David Lewis’s paper, “Attitudes De Dicto and De Se.” (We can think of
the clairvoyant as having attitudes de
dicto, but no attitudes that are irreducibly de se.) According to Lewis,
these gods have all sorts of propositional knowledge about the world they
inhabit (they are omniscient in that respect), but neither of them knows which
of the two gods he is. They both know
that one of the gods is on top of the tallest mountain and throws down manna
and that the other lives on the coldest mountain and throws thunderbolts, but
neither god knows whether he is the
god on top of the tallest mountain or the god on top of the coldest
mountain. Although I do not have room
to discuss Lewis’s paper here, I suspect that a consequence of my argument in
this chapter is that the example of the two gods is actually more difficult to
imagine than one might initially suppose.
[23]
Perry, “The Problem of the Essential Indexical,” 84.
[24]
Gil Harman has pointed out to me that even in these cases, there’s a
sense in which something is located
in relation to the believer. A belief
that one is happy locates an emotion—happiness—within the believer.
[25]
Perry, “Essential Indexical,” 85.
[26] We’ll see that there are some exceptions
to this claim: some intentional actions are not performed on objects in the agent’s surroundings. I will argue, however, that an agent still cannot intelligibly
perform such actions unless she has some egocentric beliefs.
[27] I owe this objection, as well as part of my response to it, to Daniel Warren.
[28] See, for example, Anscombe, Intention, 9, Davidson, Essays on Actions and Events, 264.
[29] Even more mundane actions such as
reaching over to grasp a coffee mug probably do not involve much, if any,
conscious deliberation. When I reach
for my mug, I don’t first think to myself, “I’d like a sip of that coffee. The
mug is on the table about two feet away from my right, so the most efficient
way to get to it is to reach out with my right hand, close my fingers around
the handle, etc.” Instead, I usually
just reach over and grab the mug without thinking about it.
[30] Alternatively, we could say that her
reason is simply that there’s a stopped car in front of her. (I owe this point to Mark Greenberg.)
[31] Guttenplan makes a similar point in “An
Essay on Mind,” 66.
[32] I’m assuming here that the agent is not
raising her arm in order to reach something, to attract someone’s attention,
etc.
[33] In section IV, we will have another
reason for thinking that the clairvoyant cannot have any desires to move her
body in a particular way. It seems to be a minimal condition on having a desire
to move one’s body that one is able to identify a particular body as one’s
own. But we’ll see that, without
perception, the clairvoyant does not have the resources to believe that a
particular body belongs to her.
[34] Daniel Warren has pointed out to me that
we can imagine telecommunication devices that do not even require that one
speak into a receiver: the speaker needs only to begin talking and a receiver
implanted in her head transmits her voice to another person. I don’t think that this example shows that
speaking intentionally doesn’t always require the possession of egocentric
beliefs. Even in this case, the speaker
would need to have the intention to speak out loud, believing that in doing so
she could communicate something to the person on the other line. And we’ll see shortly that we cannot
understand someone to have such an intention in the absence of any egocentric
beliefs.
[35] Grice, “Meaning,” 377-388.
[36] This view is similar to Stalnaker’s
“pragmatic picture” of mental representation.
Stalnaker says that “Beliefs are beliefs
rather than some other representational state, because of their connection,
through desire, with action” (Stalnaker, Inquiry,
19). It also bears an obvious
resemblance to functionalism. But
although my view takes up the functionalist insight that the notion of a belief
is a functional concept to be defined in terms of the role beliefs play in our
psychology, I depart from the functionalists in at least three ways. First,
unlike many functionalists, I am concerned with the rational –not merely the causal—role
that beliefs play in a person’s psychology.
Second, functionalism—as I understand it—holds that beliefs are constituted by their functional
relations to other mental states, to sensory stimuli and to behavioral
responses (See, e.g., Ned Block’s entry on functionalism in A Companion to the Philosophy of Mind,
323). But I do not aspire to give an
account of how beliefs are constituted;
I need only to give a necessary
condition for mental states to count as beliefs. Third, unlike many functionalists, I am not giving conditions for
distinguishing between beliefs with different contents (e.g., between a belief
that grass is green and a belief that tigers are dangerous). I am trying, rather, to give one feature
that distinguishes beliefs from other kinds of mental states (emotions,
desires, etc.) and from things that aren’t mental states at all (readings on a
thermometer).
[37] At this point, someone might raise the
following worry: if my reason for thinking that the clairvoyant lacks beliefs
is that her mental states cannot play the role that beliefs are supposed to
play in a person’s rational psychology, then the argument I’ve just given in
this section might seem superfluous.
That is because if the clairvoyant doesn’t perceive anything, we are
already in a position to say that none of her empirical beliefs are reasonable. If it is part of the role of beliefs to be rational responses to
what the subject is perceiving, we can already say that the clairvoyant’s
mental states do not play the role in her rational psychology that they would
need to play in order to count as genuine beliefs. We would not need, in that case, to bring in any further claims
about her inability to perform intentional actions. There are a few things I could say in response to this
objection. To begin with, it is
controversial whether the clairvoyant’s purported beliefs would in fact be unreasonable. Reliabilists, for example, think that for
her beliefs to be justified, it is enough that they are the products of a
belief forming mechanism that reliably produces true beliefs. And even if we grant that the clairvoyant’s
beliefs are not reasonable, it is not
clear how this fact on its own would
entail that they do not play the appropriate role in her rational
psychology. If we suppose that her
beliefs are reliably caused by events in the world and that they play a role
in motivating intentional action, it becomes much less obvious to me that they
would not be beliefs. In fact, this seems to be exactly what some philosophers
understand beliefs to be. Stalnaker, for example, says that beliefs “represent
what they represent not only because of the behavior they tend to cause, but
also because of the events and states that tend to cause them” (Stalnaker, Inquiry, 18). He does not say here that beliefs must be caused through perception; it is apparently enough that
they indicate states of affairs in
the world, in addition to motivating action.
[38] Nor can it make any reference to
connections between beliefs and perception.
Galen Strawson, for example, attempts to show that there is no essential
connection between thought and action by imagining creatures called “Weather
Watchers.” The Weather Watchers are
rooted to the ground and constitutionally incapable of action, but they are
nevertheless supposed to have a rich mental life: they have sensations,
thoughts, emotions, desires, and beliefs (Strawson, Mental Reality, 251). But to the extent that we are inclined to
accept that a Weather Watcher has beliefs, it is because they are far less
impoverished than the global clairvoyant.
Unlike the clairvoyant, the Weather Watchers have perception and egocentric
beliefs. And if we restore their
capacity for bodily action, we can appeal to those egocentric beliefs to
explain their subsequent behavior.
[39] A more sustained defense would need to
address the question of whether beliefs need not be connected up with reasons
for action if they are connected with each
other, by means of inference, and with other mental states such as desires
and emotions. Would it be enough for
the clairvoyant to make inferences from one belief to another, to have
emotional reactions to some of the things she believes, or to desire certain
things to be the case? (I owe this
objection to Mark Greenberg.) I think more needs to be said here, but I doubt
that these connections on their own are rich enough to support the idea that
the subject has a genuine mental life.
I am inclined to make the same point about desire that I made above
about belief: it is difficult to ascribe desires or emotions to a subject if
those mental states cannot play any role in explaining action. Nor do I think it
will help to insist that the clairvoyant makes inferences from one belief to
another. We can easily imagine a
thermometer that computes the “heat index” by combining data about the
temperature and humidity. We would then need a way to distinguish between
genuine inferences and the thermometer’s computations without making any appeal
to action or perception. And that seems
to leave us with the same problem we started with.
[40] Regardless of whether we make this
assumption or not, I should point out that we are still left with the epistemological point that we have no
grounds for ascribing any beliefs to
the global clairvoyant. We ascribe
beliefs to other people on the basis of what they say and do; the point of
belief ascription is to make the best sense we can of a person’s verbal and
bodily behavior. So if we are barred
from understanding the clairvoyant as performing any intentional actions, we no
longer have any grounds for ascribing beliefs to her. The resources we are left
with are simply too thin to give us much reason to distinguish between the
clairvoyant and a complicated thermometer. I should point out, however, that in
making this epistemological point, I do not wish to deny that there is a
distinction between what it is for a subject to have beliefs and what evidence
we must have in order to ascribe beliefs
to a subject. (Though some philosophers seem to think that there is no such
distinction.)
[41] One initially plausible, but ultimately
inconclusive, way to argue that this scenario is not coherent is to question
whether it is really a case of clairvoyance
and not perception. If the clairvoyant is regularly caused to
have egocentric beliefs by the events those beliefs are about, who’s to say
that she is not perceiving those events?
From an observer’s perspective she is almost indistinguishable from a
normal perceiver: she has many of the same beliefs and she’s able to act
purposively on the objects in her environment.
There’s a causal mechanism through which she is systematically able to
acquire largely accurate information about her environment. It’s tempting to say that this mechanism
counts as a form of perception. At
least, the burden of proof is on those who think otherwise to specify what
exactly the clairvoyant is missing that makes it the case that she is not
perceiving anything. The problem,
however, is that it’s also tempting to say that perception involves more than
just the systematic acquisition of egocentric beliefs about one’s environment. To count as perception, according to this
line of thinking, the acquisition of such information must be mediated by conscious experience. The reason the clairvoyant doesn’t perceive
anything is that she has no experiences; although she has many beliefs about
the location and orientation of things in her immediate environment, she
doesn’t believe what she does because these things look a certain way to her.
I must admit I’m at a loss about how to proceed when it comes to the
question of whether experience is an essential element of perception. Our intuitions can go both ways on this
issue. This is especially apparent in
the way cognitive scientists react to cases of “blindsight.” A blindsighted person claims to have no
visual experiences, though she’s able to make accurate “guesses” in forced choice situations about visual
features of the objects in front of her.
When asked, for example, whether the picture in front of her is a cross
or a circle, she’ll “guess” correctly more often than not. All the while, she insists that she’s
guessing randomly: she cannot see anything in front of her. The fact that scientists use the paradoxical
term, blindsight, to characterize
this phenomenon suggests that there are conflicting intuitions about what
counts as perception. On the one hand,
we want to say that the patient is “blind” insofar as she has no visual
experiences of the pictures in front of her.
On the other hand, there’s a temptation to say she’s “sighted” in the
sense that there is some causal mechanism through which she acquires
information, however subconsciously, about visual features of her
environment. For these reasons, it will
prove difficult to establish that perception is nothing more than a causal
mechanism through which the perceiver systematically acquires accurate
egocentric beliefs about her immediate surroundings. I don’t think this will be the most conclusive way to show that
the clairvoyant cannot have egocentric beliefs without perception.
[42] See Evans, Varieties of Reference, 141-145.
[43] Potential exceptions immediately come to
mind. For example, if I utter the
sentence, “That table was in front of me,” in the context of a conversation
about a table that I remember seeing at IKEA the other day, then I don’t need
to be currently perceiving the table in order to refer to it. Those who insist that demonstrative
reference requires perception will probably say that this isn’t a genuine case
of demonstrative reference, but rather a disguised description. What I’m really saying is that the table I saw in IKEA the other day was
in front of me.
[44] I owe this point to Daniel Warren.
[45] Anscombe, “The First Person,” 152.
[46] Ibid.
[47] See, e.g., Shoemaker, “Self-reference
and Self-awareness,” 80-93, Evans, Varieties
of Reference, chap. 7 and Cassam,
Self and World, 61-68.
[48] I am assuming here that I do not see in
a mirror that a table is in front of someone I take to be me. See footnote 51.
[49] See Evans, Varieties of Reference, 216-217 and Wittgenstein, Blue and Brown Books, 66-67.
[50] I think Evans is making a similar point
in Varieties of Reference, 221.
[51] Even this requirement must be stated
with care. It’s possible for me to look
in a mirror and see what I mistakenly take to be myself standing behind a
table. Strictly speaking, I myself see the table in front of (what
I take to be) me, but my resulting belief is not immune to error through
misidentification. (This example was suggested to me by Gil Harman.) Perhaps we could take care of such examples
by adding that in order for an egocentric belief to be immune to error through
misidentification, the subject must directly
perceive the relevant object without the aid of mirrors, video cameras,
etc.
[52] This conception of body-ownership is
what Quassim Cassam calls the “idealist” view of body-ownership. Cassam traces this view to Locke, who writes
“Thus the limbs of his body are to everyone a part of himself; he sympathizes
and is concerned for them. Cut off an
hand, and thereby separate it from that consciousness we had of its heat, cold,
and other affections; and it is then no longer a part of himself, and more than the remotest part of matter” (Locke, Essay, 336-7, quoted in Cassam, Self and World, 64). As Cassam points out, the idealist view has
its difficulties. We are not aware of many of our body parts: although my
kidneys are certainly a part of me, I’m not aware of them in the same way that
I’m aware of my arms and legs. And anesthetizing one of my limbs does not make
it cease to be a part of me. But Cassam’s
alternative “materialist” view of body-ownership is no more helpful in this
context. According to the “materialist”
view, “for a limb or body-part to be part of one, it is necessary and
sufficient that one is materially united with it” (Cassam, Self and World, 66). This
view is obviously limited to one’s ownership of limbs or body-parts. It does not make sense as an answer to the
question of what makes a given body one’s own: a person cannot be materially
united with her entire body in the way that she is materially united with her
arm. So clearly, then, whatever it is
that the clairvoyant believes, it cannot be that she’s “materially united” with
her own body. In any case, my purpose
is merely to show that a particular way of understanding body-ownership is
unavailable to the clairvoyant. I am
not trying to give a satisfactory account of the necessary and sufficient
conditions for body-ownership.
[53] Mark Greenberg suggested this
alternative to me.
[54] Dennett, “Where am I?” 310-313.
[55] Danto, “Basic Actions,” 141-142.
[56] In other words, it causes my arm to go
up as long as no one’s holding it down, I’m not paralyzed, etc.
[57] The argument I have just given resembles
what is sometimes called the “logical relatedness argument.”[57] The logical relatedness argument proceeds
from the assumption that volitions (or desires or reasons) are defined in terms
of their bodily effects, and concludes that a volition cannot be the cause of a
bodily movement. This is because, on
this view, “logical relations preclude causal relations”: two events cannot
stand in a causal relation unless it is possible to identify one event without
making reference to the other. (See, e.g., Brand, Intending and Acting, 13.)
But unlike the proponents of that argument, I do not aspire here to
argue against the idea that volitions cause bodily movements. All I am trying
to say is that the clairvoyant’s belief
that her volitions cause bodily movements is trivial if she distinguishes between volitions solely in terms of
the bodily movements they typically cause.
For that reason, I don’t believe that the objections commonly raised
against the “logical relatedness argument” (see, e.g., Davidson, “Actions,
Reasons, and Causes”) apply to me.
[58] I owe this example to Daniel Warren.
[59] Richard Taylor uses this example to
argue that there are no isolable acts of will or “tryings” apart from the
“bodily acts themselves” or the “exertion of actual physical effort” (Taylor, Action and Purpose, 81).
[60] Let me briefly explain how my claim that
the clairvoyant cannot believe that she has control over a particular body and
my argument in section three that the clairvoyant cannot be understood as an
agent. My claim that the clairvoyant
cannot believe that she has control over some body is, in effect, the claim that,
without perception, she cannot understand herself
as a bodily agent. In section three, I
am not so much concerned with how the clairvoyant understands herself; rather,
I am trying to argue that she cannot bean
agent unless she has some egocentric beliefs.
My overall argument is, in effect, the following: without perception,
the clairvoyant cannot understand herself as an agent, and so cannot have any
egocentric beliefs. But without any
egocentric beliefs, the clairvoyant cannot bean
agent, and so cannot have any beliefs at all.
Therefore, without perception, the clairvoyant cannot have any beliefs.
[61] Or at least some of them.