The Dreaming Argument

Versions of the dreaming argument are a fixture in

philosophy, East and W est. One shows up in Plato, in the
Theaetetus, another in Chuang Tzu. The version which has been
most influential in the W est is found in the first of Descartes' six
meditations in the work entitled Meditations on First Philosophy.
Descartes does a good job not only of giving the argument but of
setting the stage so as to make it plausible. That's one of the
problems. Though the problems seems to arise in connection with
examples of dreams and dreamlike or uncertain moments, it turns
out they arise out of an abstract, tempting line of philosophical
thought.
The purpose of the dreaming argument is not to make
anyone think they are now dreaming. W e cannot take that
seriously, and that's not how it is meant. Instead, it is about
knowledge; the punch line is that we do not know. W e could put
the punch line, the conclusion or claim of the argument, more
strongly or less strongly. W e could say we don't know anything.
Or we could be more tentative and it's still pretty strong, that it is
not based on our perceptions that we know anything about the
world around us. We do not really know that we are awake right
now, and in the absence of that knowledge then any knowledge is
a problem and needs us, philosophers with current licenses in our
back pockets, to account for it. Further, this is the opening
wedge in the argument from illusion, basically the same
argument with the artificial and more broad term illusion
substituted for dream.
One version, then, of the argument goes basically like
this: the way things seem to me now could be the same whether
based on reality or based on a dream. Therefore I cannot tell the
difference. Therefore I know nothing based on my senses.
Further, we are not dealing with this argument just
because it is a piece of our intellectual history, in our cultures
canon, but because it shapes our thinking. We are trying to get a
grip on ourselves. If we are tempted to think any of the items on
the following list, then we are almost certainly Cartesian dualists,
and the argument from illusion and the dreaming argument are
part of what has set us up to think them:
I am not identical to my body; I am identical to my
body; my mind is not really connected to my body; my mind is
really connected to my body; my knowledge of reality is certain;
my knowledge of reality is not certain; my knowledge of reality
cannot be certain; reality is the same no matter what people think
about it; reality is different for each person; reality is reality;
reality is really only appearances/perceptions/seemings; reality is
physical; reality really is physical and mental; reality is really
only mental; it doesn't matter whether reality is as we think it is
because we have to live as though it is reality; human values are
separate from the world; human values cannot be separated from
the world; words are the means by which human beings can find
it possible to communicate their ideas, feelings, thoughts,
memories, interpretations, perceptions, experience, plans, hopes,
meanings, desires, loves and hatreds; words cannot really capture
the life of the mind.
This can be put in the first person as follows---If I am
inclined to agree or to disagree with any of the above, then I am
almost certainly (unless I am dreaming) a dualist. And if I take
seriously the parenthesis, then I am certainly a dualist. Therefore
I am a dualist. I need help.

and what the world is) depends on how we answer the dreaming
argument and the argument from illusion, then we can try to get
out of the dreaming argument. Here are some suggestions for
answers to the question, How do I know I am not now (right
now, as I read these words) dreaming?
In waking life, I can control my actions.

In waking life, shocking, non-ordinary things

don't happen.

In waking life, science works (and not in

dreaming).

In waking life, things are more vivid and clear,

and what is weird is different from what counts as weird
in
dreams.

In waking life, you can really feel pain, but not

in dreams.

In waking life, we discuss dreams, but not in

dreams.

In waking life, we can feel the difference

between dreams and waking life, and in dreams we cannot.

In waking life, life is boring, but dreams are

not (and compare daydreams, when often we are
deliberately constructing something less boring).

In waking life, you have to go through routines,

go through space to get from one place to another.

In waking life, existence is not just in flashes or

episodes, and I know what's ahead.

In dreams, time is different.

None of these, though, will work. Those of us who are
tempted by any of them need only consider a few responses.
Claims about the characteristics of either waking life or dreaming
have to be exceptionless for the argument to work, and
possibilities of exceptions are not hard to raise. Things which
one thinks never show up in dreams may be conceived or
imagined as dreamt--though you may never have pain in your
dreams, and though most people do not, yet arguing that it is
conceptually impossible to dream that you hurt is insupportable.
It would be like arguing against the possibility that you dream
that you never have pain in your dreams. (Further, there are
plenty of reports of dreaming that one is in pain.) Things which
always show up in dreams which are not in waking life are not
there by necessity--e.g., what one student called the typical jump
cuts of her dreams do not have to show up in all dreams, so that
the present moment, of your reading or listening to these very
words, may not be interrupted by jump cuts whether it is dreamed
or waking. Finally, some have offered suggestions that the
existence of particular things in dreams might do the work of
letting us know something, but what these things purport to let us
know is that we are dreaming, not that we are not dreaming.
That is, they are on the wrong issue.
Attempting to answer the dreaming argument on its own
terms will not work. W hatever certain signs we offer to show
that we are not now dreaming can be dreamt. Descartes reminds
us of his dreams; he often dreams that he is doing things which in
fact he does do, including the writing he is doing as he composes
the Meditations. You may not dream about reading philosophy,
but you can imagine that you could.

Possible Positions
In the face of that problem of knowledge and realizing
that a tempting way of thinking (about what human beings are

I mean here to sketch or comment on some of what I

think are the most tempting responses to the dreaming argument.

They come under the following headings: a materialist response
such as many current writers on cognition (Pinker, Tye, Searle,
Dennett, Fodor) might give if we could get them to address it, an
idealist response consistent with Berkeley, Bradley, Collingwood
and with important similarities to phenomenologist approaches
such as Husserl's; an empiricist logical monism such as John
Cook attributes to W ittgenstein and as was endorsed by Russell
in his middle years; a logical positivist view such as one might
expect from Ayer or Carnap (though in fact they give an account
more like the materialist account).
(And a prefatory comment on how all the possible
positions seem consistent with a picture of human beings)
I keep thinking that all the responses I list here as
possible positions may be thought of as generated by a simple
sketch of a person looking out at the world, a person receiving
internally information from that outside world in the form of
appearances or perceptions and basing her view of the world on
those appearances. On this view, each of us is dependent on our
senses to tell us of the world, and what the senses do is provide
us with the product of some kind of causal chain originating in
the world, but the product provided by that process is a mental
product, a way that things seem or look or appear or are
perceived by each of us in each and every moment. In the face of
such a sketch, then the problems (Ayer's book makes it singular
in his title, The Problem of Knowledge) can be seen as problems
of, on one hand, accounting for how the causal chains work
within the sensory organs, and then accounting for the relations
of the mental products of those causal chains to the world outside
us. Dreams become a wildly intransigent problem, since they are
not caused in the same way as perceptions, but they present us
with similar products--somethings, whatnots, like perceptions,
but whatnots whose relation to the external world is not the same
relation, we hope, as the relation of perceptions to the external
world. Thinking of it this way suggests that it might be possible
to call the whole sketch into question. But calling the sketch into
question is difficult because it has a profound, simple and
powerful pull to it. And in fact, none of the standard positions
with the possible exception of a purist logical positivist position
really can call the sketch into question--and even the logical
positivists' response has disturbing similarities to McGeorge
Bundy's recommendation for strategy in the VietNam W ar:
"Declare victory and pull out."
Behaviorism: One response then might be based in
deep suspicion of talk about the mental. No one is really a
behaviorist any more, that position having been very publicly
slaughtered and its still-beating heart held up by Chomsky in his
review of B.F. Skinner's book Verbal Behavior. But there were
things about the view which one may still descry in later
functionalist and cognitivist views, things which have some
appeal still. Some of this is brilliantly laid out in Gilbert Ryle's
Concept of Mind. W e decide on this view that mental terms are
(often? usually? always? you can pick the severity of your
position anywhere along a long continuum from an Amish kind
of behaviorism like Skinner's to a Unitarian view like Dretske's)
a kind of illicit shorthand for describing things we might think
are accessible only through subjective introspection but which
really are events which are public and observable and objective.
That is, we might think that a clown's clowning is intelligent (one
of Ryle's examples) and think we are thereby describing
something about the mind of the clown, but in fact we are

describing an ability the clown has to do things in a certain way,

with originality and prompting surprise in us the observers of a
certain kind, all of which is describable in the way events or
behavior are describable. Norman Malcolm's reading of dreams
flirts with this when he decides that having a dream is no more
nor less than waking up with a story to tell. The story, after all,
is not something hidden and invisible but is rather something
anyone could hear, something the dreamer can repeat for you or
me or over the phone. The idea that a dream is separate from the
report of the dream is thus thought to be a kind of caving in to
loose talk. Dreams are a kind of behavior we engage in, a telling
of stories with which we awake. An investigation of dreams will
not be philosophy but will be psychology, the psychology of
dreams, in principle no more difficult than an investigation of
digestion or preferences regarding tastes (tastes being thought too
to be about something hidden and internal but the fact that that is
mistaken can be seen by reminding ourselves that tastes are made
manifest or are real to the extent they show in public choices we
make). Mental things are an embarrassment of which we can rid
ourselves by insisting on public accounts involving confirmation
by other people, and in the insistence that dreams are the waking
with a story to tell we find we have already made the crucial
move before the problem arises, a kind of safe sex approach to
the problem in which the power of dreams to puzzle us as to what
is real is taken care of by insisting they be presented only on the
public stage and there we have previously agreed to what is real,
namely you've woken with a story to tell us and all of us are not
now dreaming.
Chomsky's unanesthetized vivisection of behaviorism is
well worth reading, but I want to make a couple of other, simple
remarks about materialistic and behavioristic accounts which are
also relevant to later versions of materialism and physicalism-and perhaps relevant to others of our possible positions on the
dreaming argument. These are abbreviated remarks about
motivation, about a kind of diagnosis which helps me make sense
of all the physicalists as dualists in denial. The remarks come out
of taking a more distant stand on the issues, and asking how
Ryle, how Malcolm, and how Skinner know what needs to be
explained--in all their cases the answer is they know what needs
to be explained away. I think once the question is asked, the
answer is easy to see: they feel they need to give an explanation
for exactly those things which Descartes would classify as
mental. That may not have much bite, but it could also be put
this way--they feel they need to explain away those things we
might be tempted to classify as mental. They know what those
things are by having as a conceptual tool going into their work a
particular dichotomy, the dichotomy between mental and
material (or physical), and their work is basically work to show
that one half of that dichotomy is unreal but the other is real.
They work within a dualistic framework to try to deny the
existence of the framework. How we know what material means
depends on the intelligibility of its opposition, namely mental,
but they wish to deny that mental means anything. Having done
away with mental they will find they have no longer any meaning
for material. Therefore, their view is incoherent.
Idealism: Another response is that of an idealist, for
whom the dreaming argument only represents one aspect of a
more general problem, the problem of getting past perceptions to
anything which is not perceptions, including getting past
perceptions to physical objects. On this view, dreams do show
an important aspect of perceptions, namely the lack of necessity,
the lack of a reliable, logical connection, in their relationships
with the objects of perception. Dreams are dreams of something,

granted, but the something does not exist in the way physical
objects exist, and therefore there is a disconnect between
perceptions and physical objects. The dreaming argument has an
important part to play in the construction of idealist epistemology
because the dreaming argument provides or can provide an
intermediate step toward the view that there are no such things as
physical objects. One could caricature this slightly by saying that
if there are perceptions in the way all psychologists think there
are, then Berkeley is simply and irretrievably right-epistemology has to have a different agenda from the agenda of
accounting for knowledge of physical objects in the external
world, because there is no certainty to be had regarding physical
objects and further there is no need to include them in an account
anyway. In a mirror image of the behaviorist/physicalist view,
we can look at our talk of tables and chairs and all the other
medium-sized dry goods (to use J.L. Austin's phrase) surrounding
us and realize that categorizing them as physical objects is a
mistake since all our knowledge of those alleged objects and all
our talk of them is really knowledge of and talk about our ideas
of them and they, the ideas, have no physical existence. W hat
the dreaming argument does for us is contribute toward the
insight that all sight is insight. The world, as Schopenhauer says,
is my idea. All is interpretation, perception, representation.
Husserl's advance on this is to realize how it undermines itself
unless one can isolate interpretations from the things of which
they are accounts, by a bracketing of experience in which we
notice our own perceiving and experiencing in the moment and
separate that from the content of our experience. Our ability to
do that, he thinks, saves the idealist view from the kind of
incoherence I mentioned above which one can realize underlies
behaviorism. There is still a contrast, which we need in order for
the terms we use to be intelligible, but it is a contrast between
experience of the self and experience of other--experience of
experiencing and the thing that fills out the phrase
"consciousness of . . ." or perception of . . . or dream of (or heft
of, etc.) when we fill it out with accounts of things which are
other than ourselves.
W hat the idealist position does, then, is reduce all reality
to something like dreams. The difference between when I'm
driving home and when I'm dreaming I'm driving home is even
more difficult to articulate for an idealist than it is for Descartes
or for the skeptic, and I'm tempted to say that idealism is where
one ends up if one believes in the skeptic's dreaming argument.
Im further tempted to agree that if there are such things as
perceptions, then the argument is a good one and Berkeleys
idealism is the inevitable position.
Logical M onism: Another position we might take with
regard to the dreaming argument is to regard the appearances of
things as neither mental nor physical but rather as the atoms of
which both the mental and the physical are made when we add
the hamburger helper of linguistic conventions to them. That is,
we might regard the things which come to us in dreams as just
like the things which come to us in waking life, which are neither
our ideas nor the physical world which causes ideas but rather the
moments of contact between the two or between each of us and
whatever else there is, of which the present moment is also made.
Not only that, the flow of these logical monads or atoms, which
we might as well call the flow of empirical atoms or the flow of
experiences--this flow is the only thing which is real (this is what
makes this a monism), and our ability to make sense of the
experiences and then to talk of all the things we do when
prompted by our experiences is what needs explaining. This is
the cartoon view of logical monism, the view John Cook

attributes to Wittgenstein in his series of powerful books

beginning with W ittgenstein's Metaphysics. It is the most
fundamental of empiricist views, a view of reality as purely
empirical, purely the appearances of things manifested in the
flow of human experience.
Logical Positivism: Finally, the verificationist or
logical positivist approach also has a lot of pull. The way this
would go is as follows: like Russell's five minute old universe
hypothesis (in which we ask why we should not suppose that the
world was created five minutes ago with our memories and the
records and geological stratigraphy and libraries which we have
hitherto taken as evidence of a past that lasted longer than five
minutes), the dreaming argument's punch line is logic- and
evidence-proof, but this is its weakness as well as its strength.
That is, nothing can tell against it, but that means nothing can tell
for it either. Its problem is not that we cannot find a way to
refute it but rather that there is in principle no way either to refute
or to verify it and so it is either unscientific (this is the mild form)
or meaningless (the stronger, positivist form). The argument is
built on the possibility that anything one can perceive or sense
one can also dream, and so my present moment's perception can
possibly be a dream. I might suppose that I can publicly verify
that this moment's perception of the purple finch on the bird
feeder is not a dream by asking you to verify that there's a purple
finch on the feeder. But I would be wrong to suppose this is
foolproof since I could also dream that I am asking you to verify
there's a purple finch on the feeder. Because of the possibility
that any verification could itself be dreamed, there is no way to
verify that one's present perception is or is not dream, and so the
claim that I do or do not know that I am not now dreaming is
beyond the reach of verification. Since there is no way to verify
the claim or the denial, one reasonable way to respond would be
to dismiss the claim and its denial as senseless or at the very least
as unscientific.
A main objection to the positivist approach is that it
seriously undermines itself once one asks on what grounds we
should decide that what is unverifiable is senseless--what's the
argument? Further, if it is true, how does one verify the claim
that what is unverifiable is senseless? The argument for the
equivalence of unverifiability and senselessness seems to be only
that without it we have no way to cope with (which means,
dismiss) the five-minute-old universe hypothesis, the dreaming
argument, and other similar positionsa tactic a little like
claiming that we must have lost our keys over under the
streetlight because otherwise it's too dark to find them. W orse,
suppose it is true that claims which are unverifiable are senseless.
How shall we verify that? The answer is that the claim, that
claims which are unverifiable are senseless, is not verifiable, and
so it is senseless. The only way to save it from being senseless is
to suppose that it is false. Either way the positivist is in trouble.
This positivistic response can still be defended. Its main
appeal was that it does give us some insight into a problem with
the five-minute-old universe hypothesis which seems analogous
to something wrong with the dreaming argument. I expect that
working on the five-minute-old universe hypothesis might be
helpful to us on the dreaming argument, and that in fact there are
other responses, but we'll have to find a way to do that without
relying on a verificationist theory of meaning like the logical
positivists'. Some of the attack on the five-minute old world
hypothesis can be made without leaning on verification even
though it's the verificationists which pointed out the way to us-for instance, if the hypothesis is right then a great many terms
about time do not mean what we take them to mean and indeed a

great many terms will become hopelessly confused nonsense,

such as "geological time" and "six minutes" and "history" and
"the seventeenth century," with the result that the term "five
minutes ago" threatens to dissolve in the same acidic swamp.
O.K. Bouwsma's essay on the Evil Genius makes a
related point about Descartes' attempt to conceive of a godlike
creature who deceives us about everything, the point being that
the meaning of deception depends on their being an alternative to
deception, namely getting it right, and that if Descartes tries to
take that away from us then we no longer have any way to make
sense of deception. I doubt that such a point can be made in
connection with the dreaming argument because Descartes claim
there (as opposed to the in the Evil Genius argument) is not ever
categorical in such a way that the contrast between dreaming and
waking is threatened. Instead, it's now that I might be dreaming,
and now that I have no certain signs, rather than some such claim
as one finds in Chuangzhi or in Russell, that my life might be a
dream or dreaming might be the real thing.

Pragmatism and the Question, W hat's at Stake?

Pragmatism: The Pragmatist position is importantly
founded on the idea that nothing turns on the dreaming argument.
W hile it may be that I can find no certain signs by which I may
tell I am not now dreaming, nevertheless I m ust live as if I can
find those signs easily. That this is so shows even beyond the
dramatic cases or those like Hume mentions, of my leaving the
philosopher's study wondering about whether this or that might
be a dream and stepping into the street as a coach and four
approaches--whether it's dream or not, I had better step back.
That might be a pedestrian case, but others are even more
mundane. I may be dreaming I am pecking away on my laptop
and the purple finch is at the feeder outside the window, but I
have to accept that those things are not dreams in order to go on
living as I do. The result is that if I am now dreaming, it makes
no difference. The pragmatist goes on to suppose that she should
turn her attention to things which do make a difference. Rorty,
for instance, hopes to get intellectuals engaged with issues
(poverty, compassion, oppression, socialized health care,
educating for imagination) regarding which the social
consequences are more clear, and to get them to put the
traditional philosophical issues behind them as of no
consequence. For Rorty, paying attention to the dreaming
argument is indulging in a kind of sterile scholasticism.
Nothing's at stake. We should leave it alone.
This notion of nothing being at stake is one of the main
blunders made by the pragmatists. W hat's at stake is subtle and
articulating it may be difficult, but it is not small. Part of what is
at stake is our ability to mean what we say when we say we
know, when we deny knowing, when we inquire about knowing,
at least if epistemology is partly an account of those things.
Further, part of what is at stake is a susceptibility to Cartesian
dualism and all the alienation and isolation that makes us heir to.
Part of what is at stake is our ability to tell when we are still
saying things rather than gabbling parrotwise or deluding
ourselves into thinking we are making sense. The dreaming
argument is in part a test case for whether philosophy is doable
without engaging in nonsense, and so the stakes involved are
stakes not just for the dreaming argument but for the profession.
One way to see that pragmatists may have
misunderstood the stakes is to consider their claim that if we are
dreaming, it makes no difference. We might be tempted by this
idea, but it is an important reminder that we are tempted within a

particular context, namely the context in which the dreaming

argument presents itself as a threat to our ability to judge what to
do, considered abstractly and philosophically. Because we wish
to defuse the strength of the argument, responses which deny that
strength are tempting. One problem is that the response does not
deny the argument--indeed, saying that [even] if we are dreaming
it makes no difference allows for the possibility that we might be
dreaming, which means the pragmatist allows that the argument's
conclusion may be right. Descartes would cackle in
triumphOkay, you can live as if you know, but nanny nanny
boo boo, its just living as if. Another problem is that the
pragmatist has only a pragmatic view of what matters, namely
whether the argument results in anything which will change what
we do. Their claim that if I am dreaming I am writing this paper
or dreaming that I see an oncoming coach and four heading my
way as I step into the street it does not matter because I will go
on writing the paper and I will step back out of the street leaves
out other possible ways it might matter whether I am dreaming. I
dream my roommate says shes pregnant and think its true, tell
her mom or sister later up at the coffee shop, and all Hell breaks
loose. It does make a differencecheck against examples.
How might that go? The pragmatist's notion that it does
not matter whether I am dreaming or not is grounded in an
abstract notion of dreaming, and conspicuously is not grounded
in any examples of dreaming--the quickest survey of possible
cases will suggest that sometimes it does not matter and
sometimes it might matter. The suppositions, that I might now
be dreaming or that I am dreaming or that I am not dreaming, can
sometimes affect what I do. Suppose I have a history of
dreaming that the phone is ringing or someone is at my door, and
when I have this dream I often start to get out of bed before I
catch myself, and tell myself it's only that dream again (or some
such). These dreams come and go in their frequency, sometimes
occurring a couple of times a month, sometimes once a year. I've
had them now for about ten years, spent some time trying to
figure them out, now just tolerate them. This morning, after a
late night grading papers, I sleep in, but wake to a knock at the
door, pull back the covers but realize that I was dreaming and fall
back onto my pillow. Then I hear a knock at the door, and in my
befuddled state I feel some alarm because after all I'm more
awake now but there it is, still the knock--this is a new and more
insistent version of my recurrent dream, I think, and I pull the
pillow over my head in time that I avoid hearing Ed M cMahan
call my name.
Or you, to take another case, sometimes have vivid
dreams of interactions with your siblings and parents and confuse
yourself into thinking they really happened. W ith some of these
of course it won't matter--you ask your sister whether she got her
car fixed and find it was never sick, and all that turns on it is a bit
of embarrassment--but with some it may involve
misunderstandings for higher stakes, as when, as up above, you
ask your sister whether she's told the father yet that she's
pregnant or you let it slip to him yourself.
W hat accounts for the pragmatist's failure to
acknowledge these commonsensical stakes? W hat the pragmatist
was thinking is that if all our experiences are dreams or all our
experiences are on the same level as regards dreaming versus
waking, then we will have to treat them all the same. She was
thinking of my looking up and seeing what seems to be a coach
and four bearing down on me as something I might do whether I
am dreaming or awake because seeing what seems to be a coach
and four bearing down on me will always be the same whether I
am dreaming or awake--and given that, then I will always if I am
reasonable react in the same way. The dreaming argument,

though, does not attempt to establish any such general claim. In

both Plato's version (in Theaetetus 157b and following) and in
Descartes' version) there is a carefully thought-out emphasis on
the present moment, on the issue of whether there are any certain
signs that I may tell that I am not now dreaming. Cases like
those above show that the stakes change with the content of the
moment, and that sometimes it does matter. W hether Descartes
is sitting by the fire in his robe pecking on his laptop or is
sleeping naked in his bed may not matter much, but whether he is
taking the reddish glow of firelight on his curtains as a dream
rather than as a neighbor's house burning--that might matter.
Thinking of dreams and appearances and reality in a
schematic way which emphasizes how all these are the same
across various cases obscures how the cases can be or are, in fact,
different. The phenomenologist and the logical monist are at
about the same point in this labyrinth as the pragmatist, having
granted more than the dreaming argument tries to establish in
their treatment of all moments as the same in this way--that since
whatever we see presents us with an appearance which remains
the same whether it is dream or real, then only those appearances
are truly real.
Part of the charm of the pragmatist position is that it
bites the hand that feeds it--Rorty's ditching philosophy for
humanities and literature because philosophy does nothing to
help cope with oppression, injustice, bigotry, human and global
catastrophes, alienation and ennui, besides falsely implying that
literature and the humanities have a better record as helping
professions in humane causes, ignores the source of his ability to
articulate the causes, which source is philosophy.
The result, once again, is that the question about what is
at stake with the dreaming argument can be seen as closely akin
to the question about what is at stake with philosophy. Rorty
takes the dreaming argument to typify what is wrong with
philosophy, but curiously he takes it to do this by showing what
philosophy does best and then he rejects this as otherworldly.
W hat the dreaming argument does is show that certainty is not
present in our ordinary lives and that the attempt to derive
certainty is an attempt to run down a mythological beast. Much
better to stick with real possibilities and to treat certainty as if the
kinds of certainty which can be achieved were all that should be
sought. In other words, it is always possible that we could be
wrong, but we cannot practically eliminate that possibility and so
must live as though it is not a real possibility.
I've suggested some stakes involved with accepting
dualism. Here's more. Consider some of the pathologies of
dualism. We absorb dualism so deeply and so pervasively in this
culture that it is like the air we breathe, so permeating our world
that we cannot see it. Yet it divides and alienates for all that.
Before we ever get to the mental aspects of human beings, the
basic taken-for-granted part about our physical existence is an
account that emphasizes separation, individuality, boundaries,
countability and position and extension as though that were what
we were. Human beings' bodies are bounded by their skins.
Humans weigh a certain amount, should never be confused one
with another, always have a particular location in space which
cannot be the location of another human being at the same time
no matter how much we might try, are individuals and separate.
Separate from everything. Separate from the landscape, separate
from the earth, separate from the universe, separate from our
neighborhood and neighbors to the extent that loving them seems
a wildly paradoxical notion because it is flatly impossible.
Nothing in the story about dualism overlaps with the other stories
we might tell about being together or being a part of something,

being in a group or family or couple or discipline or situation or

church or pickle or love or neighborhood so thought I would stop
by. Instead it is all about being alone and separate and divided
by space from every other human being and we haven't even got
to the place where we really live up there behind our eyes where
we are pulling levers and shouting to the engine room. That's the
other side, the mental.
The mental, especially to the extent that it is private and
inaccessible to others, provides us with yet another grand set of
dissociations and divisions and alienations, all in a realm that is
already spectacularly separate, one that threatens to float isolated
in Laputa or a Never-never land beyond the reach of our friends
and family. Characterizing the mental and the issues which arise
in connection with the mental calls forth an array of horribly
mixed metaphors. The relation of the mind to the body, however
we would like to settle the philosophical problem (even mindbrain identity stories, founded on heroic denial, leave intact the
problems about how my hopes that you will call can interact with
my leaping to the phone when it rings), winds up being an
account of how the captain on the bridge can understand the
messages from the lookout and can convey orders to the engine
room, an account which leaves me connected to my own feet by
the most fragile and mysterious lines of communication--that is,
if I am connected at all. Besides being characterized as a captain
of a ship, the mind sometimes is conceived as an inner theater, or
as a windowless room within which we live and on whose walls
we see portrayed for us the world as the senses paint it, along
with the world as we dream and imagine and hope and dread and
hallucinate it, a room one could rent out if one were a brain in a
vat. I pass over the part of this story about rationality as the one
thing we can trust, rationality conceived as that faculty which
evaluates arguments impartially, and impartiality conceived as
the deliberate dissociation (because they are never a help and
often a hindrance) from our families, our bodies, our hopes and
memories and histories of victories and defeats, our desires and
our appreciations. But one clear result of the standard accounts
of mind is that within the mind is where we exist with our
thoughts and hopes and significances and associations and
feelings and memories and dreams, and that in there each of us is
irredeemably alone.
W ittgenstein has at both pieces of this account. It is part
of his accomplishment that he increases our awareness of how
many of our philosophical lines of thought rely on fragments of
dualism, as though those lines of thoughts are carriers of this
recessive gene. He examines, for instance, the temptation to say
"Only I can have my pain." He winds up on the same side as
Josiah Royce. Royce castigates people who would say, "Your
pain is not the same as my pain, for your pain is much easier for
me to bear" with the remark that pain is pain, after all, and this
puts Royce on W ittgenstein's side in this battle against our own
philosophical ontogeny. After all, it is not always the case that
your pain is much easier for me to bear, not when I am speaking
to my children or lover or when I am compassionate.
Suffering with others is one clue that dismantling
dualism might have a therapeutic effect on more than academic
philosophy. Outside philosophy we easily acknowledge things
which are of the greatest puzzlement to philosophers. "This is
the same headache the previous tenants used to get." "She's a
great one for stoicism, but I know she's in pain." "W e're in this
together." "M adame Chairman, I've not met the previous
speaker, but I am of the same mind on this matter as he."
This therapeutic work might lead in part to a simple
recognition. In this way it is reminiscent of some literature of
enlightenment. It's a recognition which would prompt us to

express humility on behalf of the profession of academic

philosophy and on behalf of all of the academic world who have
swallowed and now in their work express those unacknowledged
bits of philosophical legislation. The recognition would be that
in philosophy and in the academic world we are accounting for
things which are not our property--we are accounting for
knowledge, justice, language, art, science, argument--and that
what we are giving accounts of are things we need to keep our
eyes on. W e have not done this. If we give, and we have given,
an account of knowledge which has nothing to do with the cases
in which people claim to know or ask whether someone knows or
deny knowing or correct claims to know or confirm that they do
know, then this is a clue that perhaps we have gone astray. And
to the extent that we are working on accounts of what it is to be
human beings, we might be led to a recognition that sometimes
we are not alone, not cut off from others, in the way philosophy
so quickly and deeply assumes and even requires that we are.
W e'll need to keep our eyes on human beings too. Doing so, we
might find that philosophy could rejoin the human race, be its
ally and its friend.
Objections (pointing out that dualism is poisonous doesnt
show its wrong, only that it would be nice if it were wrong)
I.
It is crucial to the dreaming argument that it succeed in
bringing us to doubt that the senses can provide us with
knowledge, and that it does this by bringing us to the point that
we can entertain doubts about whether we are now awake or
dreaming. The first objection is that for Descartes' argument to
work we have to treat doubt as a performative, a term for which
if we fulfill the constitutive rules for engaging in the practice
then when we say we do it we in fact thereby do it.
Consider the following version of the dreaming argument. How
do I know I am not now dreaming? The way things look to me
could be the same whether they are real or are dreamed.
Therefore there are no certain signs by which I can know that I
am not now dreaming. Therefore I do doubt that I am not now
dreaming.
One problem with this argument is with whether we can
take the issue seriously. That is, if I agree that I can doubt
whether I am now awake, then Descartes' dreaming argument
leaves me gasping and flopping around on the dock, having
swallowed hook line and sinker--BUT I have to agree that I can
doubt whether I am now awake, or doubt that I am not now
dreaming, and in order to do that it seems I have to follow
Descartes' lead and agree to doubt just that.
But the word "doubt" is not a performative. That is, it is
not like "promise" or "warn" or "guarantee," those words where
if I say the words in the right kinds of circumstances then I have
done the thing that the words say--If I say "I promise" in the right
circumstances, then I have promised, no matter even if I intended
to promise or not, no matter if I intend to fulfill my promise or
not. Likewise for the wedding vows--I may intend to sleep
around or ditch my wife the next night, but if I say "I do" at the
altar in front of witnesses and all that, then I did. This is in
contrast with other words--run, know, understand, remember,
appreciate, own, bite--where saying the words does not make it
so that we have understood or bitten. I might say, "I hereby bite
you," but this is unlike when I say in the proper and standard
circumstances, "I hereby swear that the testimony I am about to
give is the truth, the whole truth, and nothing but the truth,"
because I have not bitten by saying I have bitten as I have sworn

by saying I swear. This can be investigated by those who are

standing around, for whom my saying "I hereby bite you" does
not establish that I have bitten you, though my saying "I hereby
swear that the testimony . . .[etc.]" does establish that I have
sworn.
Doubt does not work like that. It does work, though,
even with dreaming and being awake. It is possible to doubt we
are awake. Nightmarish events may make us doubt whether we
are dreaming. A sudden appearance of an old lover in the hall
outside class might make us think we are dreaming. Hitting a
bear while driving home on 101, or being involved in a multi-car
pileup on a busy freeway, or hearing a doorbell in that transition
period between sleep and waking--any of these might prompt us
to ask ourselves whether we are dreaming. But march through
Descartes' dreaming argument and then see whether you can say,
"I doubt whether I am not now dreaming." Toss in the word
"hereby," even, to make it more official. That is, check and see if
you do doubt. You don't. The issue is a non-starter, because
doubt requires more; it requires the kinds of circumstances and
contexts we see in examples, the examples of which we can
remind ourselves in which we do doubt whether we are dreaming
or not. It's not a performative, not something we make so by
saying it's so. W e might be tempted to go on and say what it is,
tempted perhaps to say that it is a report of a psychological state
which either obtains or does not obtain. Following up on this we
will be tempted to say further that Descartes' insight in the
_cogito_ that even if I doubt I exist I must existto do the
doubting depends on an assumption that the law of the excluded
middle holds, that there are only two possibilities, that I do not
doubt I exist and so I do exist, or that I doubt I exist and therefore
I do exist (or I could not be doubting), --and perhaps these
temptations would lead us to the insight that Descartes has left
out the possibility which in fact does obtain for almost all of us
almost all the time. That possibility is that we neither doubt nor
do we not doubt that we are dreaming (or that we exist). It is the
possibility then that we need more to make sense of this way of
talking than the dreaming argument has given us.
II.
There is a related objection to the dreaming argument
which can be made into a more general objection to some
common philosophical methods. The relevance of examples and
the existence of the issues in examples of which most
philosophical problems are made are the key ingredients in this
objection. To get to this, let's look again at part of Descartes'
account which greatly aids him in establishing the plausibility or
legitimacy of the question. That question is, how do I know I am
not now dreaming?
Descartes is in search of a method by which he can sort
out the things to which he can give assent with confidence. That
he hits on the knowledge he derives from the senses as a category
probably deserves attention, but I'll pass that by for now. Let us
grant that it is a category of our knowledge (we sometimes do
reply to the question, "How do you know?" by saying things like
"I saw it myself.") Descartes also shrewdly acknowledges that
doubting what one sees or hears or touches seems crazy, as
though the philosopher were (as indeed she is) like those persons
who maintain they are made of glass or have heads made of
pumpkins or that they are dressed in royal robes though they are
in rags. That is, questioning one's senses seems something we
can do on pain of suspicion that we are crazy or deluded.
But Descartes dreams. He does this, he says, because he
is a human being. (For a bit there he was apparently in danger of
forgetting that he is.) Of course most of us dream. But Descartes

has a certain kind of dream which helps make his asking the
question, how does he know he is not now dreaming? a quite
reasonable asking. Here he is, writing for us across a gap of
three hundred sixty years, sitting by his stove wearing his robe.
But after he quits, the battery runs down and the laptop starts
beeping and he shuts it down, brushes his teeth, takes off his
clothes and crawls under the quilts, he dreams that he is writing,
dreams that he is sitting by his stove wearing his robes, and
sometimes he awakens and realizes he was dreaming that he is
doing the same thing he was doing before the laptop battery gave
out.
Some of us have dreams of doing the things which we
do when we are not dreaming, and for most of the rest of us it is
easy to conceive of doing this. W hen I first got the crane
operator's job in the mill, I went through a spell of dreaming that
I was walking around the mill with the hook and the cargo belts
following behind me. Several of us have similar accounts. A
few, perhaps, have dreams of coming to philosophy classes or
seminars and sitting listening to professors read papers about the
dreaming argument. W e will expect the result is a kind of
vertigo, a sense of deja vu which threatens to disorient us, as
philosophy often does.
W hen and where do we feel this disorientation, and how
much hamburger helper will it take before the disorientation
takes over all our senses?
Suppose that you belong to a group, Food and Thoughts
Not Bombs, that meets at Powell's house on Tuesday evenings,
that the group has been discussing the dreaming argument for
some weeks, and suppose that you dream on a Sunday morning
that you are listening to Powell read on Descartes' dreaming
argument.
W hen you awaken will you doubt that the dream is a
dream, or doubt that you have awakened? Further, when you go
to Powells' and the group finishes eating dessert and turns to the
dreaming argument, will you doubt whether the discussion might
be a dream? You will not--or, stronger, it is hard to know what
doubting such a thing would be so that we can say you will or
will not--despite telling the group your dream and despite
Descartes' supposition that the fact that you have dreamed the
same thing you are doing now not dreaming would or at least
could lead you to doubt whether you are not now dreaming.
It is only in the face of the dreaming argument that it
seems we are called on to account for the fact that we do not
doubt.
Something odd has happened. The dreaming argument
has the effect of making us feel we need to account for
something, something for which otherwise we should feel no
need to give an account. The something is equivocal, too--on
one hand, there is a need to claim and to defend the claim that I
am not now dreaming, and on the other there is a need to claim
and to defend the claim that I know I am not now dreaming.
That there is such a distinction and that Descartes does not see
the distinction needs some comment. Perhaps we need the kind
of case in which we could be brought to doubt whether we are
not dreaming.
Suppose Pat is working two jobs, sixteen hours a day in
two different factories on two different production lines. He
cannot get the hang of sleeping during the day when he is not
working. He gradually over several weeks becomes exhausted to
the extent that he is beginning to fall asleep in the lunch room
and sometimes has to stop on the shoulder of the highway to take
a nap driving home. He begins occasionally to hallucinate, and
the hallucinations seem to him dreamlike intrusions into his
waking life. Sometimes his dreams and the hallucinations seem

the same--twice while driving home he has hallucinated

someone, a largish stranger, standing up on the shoulder of the
highway in the predawn darkness and throwing a football at him
in his car. And he has had this same thing happen while he was
lying in his bed dreaming--he dreams he is driving home,
exhausted, and sees someone rise up in the weeds along the road
and throw a football at him. The first time this happened while
he was driving he veered, badly frightened, though the apparition
vanished as it should have gotten clear. The second time, over a
week later, he was startled but steady and tried to notice more
detail--the stranger seemed to rise up off one knee and stood a
second, cocked his arm and threw, leading the car enough that
Pat felt that, had the car been a convertible, he would have had a
chance to catch it. Suppose Pat next makes a terrible mistake.
He tells a couple of his friends from his college football days
about this hallucination/dream. One, Roger, is a bit of a sadist.
After Pat has told his friends about it, Roger puts together a
prank--he gets his older brother, whom Pat has never met, to wait
in the weeds etc. Pat is driving home slowly, absolutely wasted,
and is hardly fazed when he sees the man rise up, until the
football hits his windshield.
Now Pat could doubt the football thrower was not a
dream or hallucination, though perhaps the fog might clear
enough that he will realize right away that it was a mistake to tell
Roger. In case he does doubt, what accounts for his doubting?
The answer is that the example has provided us with the account.
W e can summarize that, but the summary has the story we have
told as its basis. Pat is exhausted and has been dreaming or
hallucinating the same thing that just took place, and it is
important for the account that "the same thing" get spelled out,
along with the extreme nature of his fatigue and the sadistic joke
that is being pulled on him. That means that it is this story, this
example, which provides the account making sense of his wail.
The particular story, the particular example, is the account which
explains in case Pat should wail in bewilderment to a passenger,
"I don't know if I am dreaming or hallucinating or not." If his
passenger is not in on the prank, has seen the football thrower,
then the passenger will need to know more of the story before she
can understand his wail, and when she understands enough to pat
him on the arm and say, oh, no, I saw it too, that will let him
know it was a major mistake to tell Roger.
I expect other examples could be put together which
would make intelligible a character in the example saying, "Am I
dreaming?" or "I don't know if I am dreaming or not." These
examples might not have to include extreme sleep deprivation,
though we expect such things as fatigue, hallucinogenic drugs,
_delirium tremens_, wildly unlikely events, odd juxtapositions of
the ordinary and the bizarre, a history of having dreamt the
particular thing which is at issue, and perhaps other factors. We
can construct cases such that someone might say, "I don't know
whether I'm dreaming or not," or "I wonder if I'm dreaming," or
"I might be dreaming," or "I know [or, I don't know] I'm not
dreaming." W ith enough details and weirdnesses we might be
able to provide cases in which the speaker will think she does
know but in fact is wrong, but that will not matter. The truth of
the speaker's claim, whatever it is, may be a separate matter from
whether we have the material in the example to understand the
claim. That is, its intelligibility can sometimes be distinguished
from its truth.
It might seem we have helped Descartes out, since we
have succeeded in providing an account for how someone might
be led to say he doubts he is not dreaming. W e have not. W hat
we have provided is an account _for the particular person in the

example_, and not for anyone who is not in the example--not

even for anyone else who is in the example (for instance Roger or
the passenger in Pat's car). Most importantly, we have not
provided an account which would make sense of your or my
saying that this present moment, my typing these words, my
reading these words aloud, your reading or your hearing these
words, might be a dream. Instead the examples support the idea
that if such an account as I have given makes intelligible saying
this present moment might be a dream, then unless you and I
have had some similar, strikingly unusual things going on we are
lacking just such an account. In other words, absent such a
particular account our saying such a thing would be without
basis, without sense. Descartes thought his example would
justify a categorical claim for all of us that we might now be
dreaming (or that we do not know we are not now dreaming).
But in fact an examination of examples shows that the examples
are what justify the claims which arise in and for those examples,
and that those examples are particular, that is, they are not all
examples. In particular and most pointedly, they are not
examples which include either the present moment you are in or
the present moment I am in. His argument fails. It fails to
establish the truth of his claims, that I might now be dreaming
and that I do not know I am not now dreaming, but more
importantly it fails to provide intelligibility for that claim.
In the face of any question we might raise about why we
should think that this moment [my pecking away, your reading or
listening] might be a dream, Descartes only has the insight that
some people have dreamed or could have dreamed such moments
as we are in. A quick look at examples shows that won't be
enough to make sense of such a question, won't be enough that
someone would ask or even understand his question. His own
case is a halfway house case--he dreams often of doing such
things as we are doing now, of pecking away or of listening to a
professor read a paper on the dreaming argument. Does he then
wonder whether he might now be dreaming? No, he does not do
that--instead he wonders how he knows he's not dreaming and
fails to come up with an answer. Knowing, after all, has been his
main interest, not dreaming. He concentrates on the
epistemological question. But the reason he fails to come up
with an answer is not that he does not know but that he has failed
to ask an intelligible question. And there is an explanation at
hand for how he could have been led to ask his question, an
explanation which casts doubt on the legitimacy of his enterprise
and doubt on the legitimacy of our temptations to the extent we
were happy to follow his lead.
That explanation can be abbreviated in different ways.
One way would be in terms of assumptions: he assumes that
either we know or we do not know whether we are dreaming; he
assumes that dreams provide us with appearances just as our
perceptions do; he assumes that the possibility of someone
dreaming what we are doing now implies that we might be
dreaming now without looking at what kind of examples make
sense of that possibility. The first assumption, the application of
the law of the excluded middle restricting the answers to the
question, do you know you are not now dreaming? leaves out the
possibility that the question gets no traction, makes no sense,
absent the kinds of details that can be provided in examples. The
second assumption is the assumption of dualism, assuming
mental appearances and physical reality, an assumption which
has to be projected into examples rather than one that emerges in
them. The third assumption confuses logical possibility with
possibility in just the way that an examination of examples can
help to disentangle.

II, Continued: a Comment on M ethod

Finally, some comments which I hope will help clarify
the method involved in this kind of objection. The objection
does not hold up what we would say in examples of
conversations and examples of thinking from when we are not
doing philosophy as a criterion. It only holds up what we would
say as a way to raise suspicions about sloppy work. W hen we
are absolutely sure of ourselves and our work, it is absolutely
unlikely we will take such reminders about what we would say
seriously. To the extent that those reminders do show something
bothersome or unexpected or jarring, they may reveal hidden or
covert sloppiness, but if we can examine that possibility and
dispose of it we are back to business, can proceed reassured, and
need not worry about our work. To the extent that those
reminders reveal problematic assumptions or unconscious
sketches guiding our formulation of the philosophical problems,
then we cannot go back to business but must settle whether the
problems are legitimate problems, whether they make sense in
the way we thought.
Descartes' dreaming argument cannot survive this
objection. Am I dreaming? or How do I know I am not now
dreaming? is a question I can ask in some examples but not now
and it is not a question Descartes can ask based only on the fact
that he has dreamed of doing what he is doing now. W e may be
tempted to think that one need not establish the intelligibility of
questions one asks but can rely on the question making sense in
terms of the philosophical problem, but the objection in terms of
appealing to examples of what we would say and think when we
are not doing philosophy--that objection makes this temptation
into a live issue. If it's a live issue, we need arguments. The
arguments for the legitimacy of asking "might I now be
dreaming?" or "how do I know I am not now dreaming?" have
not been forthcoming, unless we do as Descartes does and rely
on assumptions which have the effect of legislating the
legitimacy of the question.
jwp
-|-