Is the Brain a Digital Computer?

Author(s): John R. Searle

Source: Proceedings and Addresses of the American Philosophical Association , Nov.,
1990, Vol. 64, No. 3 (Nov., 1990), pp. 21-37
Published by: American Philosophical Association

Stable URL: https://www.jstor.org/stable/3130074

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Philosophical Association is collaborating with JSTOR to digitize, preserve and extend
access to Proceedings and Addresses of the American Philosophical Association

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
 IS THE BRAIN A DIGITAL COMPUTER?

John R. Searle
University of California/Berkeley

Presidential Address delivered before the Sixty-fourth Annual Pacific Division Meeting of the
American Philosophical Association in Los Angeles, California, March 30, 1990.

I. Introduction, Strong AI, Weak AI and Cognitivism.

There are different ways to present a Presidential Address to the APA; the one I have
chosen is simply to report on work that I am doing right now, on work in progress. I am
going to present some of my further explorations into the computational model of the
mind. [1]
The basic idea of the computer model of the mind is that the mind is the program and
the brain the hardware of a computational system. A slogan one often sees is "the mind
is to the brain as the program is to the hardware." [2]
Let us begin our investigation of this claim by distinguishing three questions:

1. Is the brain a digital computer?

2. Is the mind a computer program?
3. Can the operations of the brain be simulated on a digital computer?

I will be addressing 1 and not 2 or 3. I think 2 can be decisively answered in the

negative. Since programs are defined purely formally or syntactically and since minds have
an intrinsic mental content, it follows immediately that the program by itself cannot
constitute the mind. The formal syntax of the program does not by itself guarantee the
presence of mental contents. I showed this a decade ago in the Chinese Room Argument
(Searle, 1980). A computer, me for example, could run the steps of the program for some
mental capacity, such as understanding Chinese, without understanding a word of Chinese.
The argument rests on the simple logical truth that syntax is not the same as, nor is it by
itself sufficient for, semantics. So the answer to the second question is obviously "No".
The answer to 3. seems to me equally obviously "Yes", at least on a natural
interpretation. That is, naturally interpreted, the question means: Is there some description
of the brain such that under that description you could do a computational simulation of
the operations of the brain. But since according to Church's thesis, anything that can be
given a precise enough characterization as a set of steps can be simulated on a digital
computer, it follows trivially that the question has an affirmative answer. The operations
of the brain can be simulated on a digital computer in the same sense in which weather
systems, the behavior of the New York stock market or the pattern of airline flights over
Latin America can. So our question is not, "Is the mind a program?" The answer to that

21

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
22 APA PROCEEDINGS, VOL. 64, NO.3

is, "No". Nor is it, "Can the brain be simu

question is, "Is the brain a digital computer?
taking that question as equivalent to: "Are b
One might think that this question would
receives a negative answer. That is, one migh
there is no interest to the question whether t
the case. Even for those who agree that pro
mental phenomena, there is still an importan
mind than the syntactical operations of the d
case that mental states are at least comp
computational processes operating over the f
in fact, seems to me the position taken by a
I am not saying that the view is fully cle
some level of description brain processes are
in the head". These need not be sentences
"Language of Thought" (Fodor, 1975). Now, l
structure and a semantics or meaning, and th
the problem of semantics. The problem of s
head get their meanings? But that questi
question: How does the brain work in process
latter question is: The brain works as a di
operations over the syntactical structure of
Just to keep the terminology straight, I cal
is having a program, Strong AI, the view tha
be simulated computationally, Weak AI, and
Cognitivism.
This paper is about Cognitivism, and I had better say at the beginning what motivates
it. If you read books about the brain (say Shepherd (1983) or Kuffler and Nicholls (1976))
you get a certain picture of what is going on in the brain. If you then turn to books about
computation (say Boolos and Jeffrey, 1989) you get a picture of the logical structure of the
theory of computation. If you then turn to books about cognitive science, (say Pylyshyn,
1985) they tell you that what the brain books describe is really the same as what the
computability books were describing. Philosophically speaking, this does not smell right to
me and I have learned, at least at the beginning of an investigation, to follow my sense of
smell.

II. The Primal Story

I want to begin the discussion by trying to state as strongly as I can why Cognitivism
has seemed intuitively appealing. There is a story about the relation of human intelligence
to computation that goes back at least to Turing's classic paper (1950), and I believe it is
the foundation of the Cognitivist view. I will call it the Primal Story:

We begin with two results in mathematical logic, the Church-Turing thesis (sometimes
called Church's thesis) and Turing's theorem. For our purposes, the Church-Turing
thesis states that for any algorithm there is some Turing machine that can implement

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
PRESIDENTIAL ADDRESSES 23

that algorithm. Turing's theorem

can simulate any Turing Machi
result that a Universal Turing

But now, what made this result s

spines of a whole generation of y
thought: Suppose the brain is a U
Well, are there any good reasons
Machine? Let us continue with t

It is clear that at least some hu

can consciously b do long division
long division problems. It is fur
and Turing's theorem that any
a Universal Turing Machine. I ca
that I use for long division on
Turing (1950), both I, the hu
implementing the same algorith
nonconsciously. Now it seems r
of mental processes going on in
And if so, we could find out ho
on a digital computer. Just as w
long division, so we could get a co
language, visual perception, cate

"But what about semantics? After

of logico-mathematical results co

The development of proof theor

semantic relations between pro
relations between the sentences
mental contents in the head are
need to account for mental pro
syntactical elements in the head
take care of itself; and that is w

We thus have a well defined rese

implemented in the brain by pro
We do this in turn by getting th
human computer (i.e. to pass the
for evidence that the internal pr
Now in what follows I would lik
especially Turing's contrast betwe
human computer and the noncons
or by the mechanical computer; n

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
24 APA PROCEEDINGS, VOL. 64, NO.3

programs running in nature, the very same

computers.
If one looks at the books and articles suppor
assumptions, often unstated, but nonetheless
only alternative to the view that the brain is
The idea is that unless you believe in the exis
believe that the brain is a computer. Indee
question whether the brain is a physical mec
whether the brain is a digital computer are t
idea is to bully the reader into thinking tha
some kind of computer, he is committed to s
field has opened up a bit to allow that the
Neumann style digital computer, but rather a
computational equipment. Still, to deny that
your membership in the scientific community
Second, it is also assumed that the question
is just a plain empirical question. It is to be s
way that such questions as whether the h
photosynthesis were settled as matters of fa
conceptual analysis, since we are talking abou
think many people who work in this field wo
appropriate philosophic question at all. "Is the
a philosophical question than "Is the neurotra
acetylcholene?"
Even people who are unsympathetic to Cogn
to treat it as a straightforward factual issue
question what sort of claim it might be that
question: What sort of fact about the brain c
Third, another stylistic feature of this lit
carelessness with which the foundational ques
anatomical and physiological features of brai
a digital computer? And how are the answers
The usual procedure in these books and artic
l's, give a popular summary of the Church-Tu
exciting things such as computer achievement
literature I have found that there seems to b
hand, we have a very elegant set of mathema
to Church's thesis to recursive function theor
set of electronic devices which we use ev
mathematics and such good electronics, we a
done the basic philosophical work of connect
as far as I can tell that is not the case. On th
where there is little theoretical agreement a
fundamental questions as, What exactly is a d
What exactly is a computational process? Und
systems implementing the same program?

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
PRESIDENTIAL ADDRESSES 25

m. The Definition of

Since there is no universal agree

to go back to the sources, back to
According to Turing, a Turing m
It can rewrite a 0 on its tape as
tape 1 square to the left, or it ca
a program of instruction and eac
carried out if the condition is sat
That is the standard definition o
misleading. If you open up your h
1's or even a tape. But this does
object is really a digital computer
and l's, etc.; rather we just have t
or could be used to function as O's
it turns out that this machine cou
says, "It could be made out of cog
it could be made out of a hydraul
out of transistors etched into a s
could even be carried out by the
to represent binary symbols--the
level of the voltage and perhaps
Similar remarks are made by mo
Ned Block (Block, 1990), shows h
are assigned to voltage levels of
we should go and look for voltage
assigned to a certain voltage level
further that we did not need to u
system of cats and mice and che
strain at the leash and pull open
The point, as Block is anxious to
computational description. These
computationally equivalent" (p. 26
al sequence could be realized by "a
(Pylyshn, 1985, p. 57)
But now if we are trying to take
we get the uncomfortable result t
does out of pretty much anythin
a "brain" that functions just like
or water pipes or pigeons or anyt
"computationally equivalent". You
waterpipes, or whatever it might
sheer and unconcealed delight. Bu
going to try to show that it is jus

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
26 APA PROCEEDINGS, VOL. 64, NO.3

IV. First Difficulty: Syntax is not In

Why are the defenders of computationalism n

realizability? The answer is that they think
same function admits of multiple realization
carburetors and thermostats. Just as carb
computers can be made of an indefinite range
But there is a difference: The classes of car
terms of the production of certain physical ef
you can make carburetors out of pigeons
syntactically in terms of the assignment of
consequence not of the fact that the same p
physical substances, but that the relevant pro
irrelevant except insofar as it admits of t
transitions between them.
But this has two consequences which might

1. The same principle that implies mult

universal realizability. If computation is
syntax then everything would be a digita
could have syntactical ascriptions made t
terms of O's and l's.

2. Worse yet, syntax is not intrinsic to

properties is always relative to an agent o
phenomena as syntactical.

Now why exactly would these consequence

Well, we wanted to know how the brain w
phenomena. And it would not answer that q
computer in the sense in which stomach, liver
are all digital computers. The model we had
the operation of the brain which would show
if there was not some sense in which brains
that green leaves intrinsically perform photo
It is not a matter of us arbitrarily or "conven
or "photosynthesis" to leaves. There is an ac
asking is, "Is there in that way a fact of the
digital computers?" It does not answer that
computers because everything is a digital com
On the standard textbook definition of com

1. For any object there is some descript

description the object is a digital comput

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
PRESIDENTIAL ADDRESSES 27

2. For any program there is so

description of the object unde
example the wall behind my
program, because there is s
isomorphic with the formal str
Wordstar then if it is a big en
any program implemented in

I think the main reason that t

realizability is a problem is that
point, namely that the "syntax"
On the contrary they talk of "sy
talk were like that of gasoline en
of fact that the brain or anythin
I think it is probably possible to
up our definition of computation
and engineers regard it as a quirk
of computation. Unpublished w
suggest that a more realistic defi
causal relations among program
mechanism, and situatedness in
definition of computation are n
problem is that syntax is essenti
of computationally equivalent pro
the processes were abstract, but
depended on an interpretation fr
which would make brain process
computation, there never could
hand, say that anything is a dig
suppose there is a factual questi
natural system such as the brain
And if the word "syntax" seem
That is, someone might claim tha
of speaking and that what we ar
discrete physical phenomena and
really need O's and I's; they are
no help. A physical state of a
assignment to that state of some
problem arises without O's and 1
program do not name intrinsic p
discovered within the physics, the
This is a different argument fr
it ten years ago but I did not. Th
intrinsic to syntax. I am now ma
intrinsic to physics. For the purp
the syntactical characterization o

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
 28 APA PROCEEDINGS, VOL. 64, NO.3

There is no way you could discover that som

because the characterization of it as a digital
who assigns a syntactical interpretation to th
applied to the Language of Thought hypothe
is incoherent. There is no way you could disc
sentences in your head because something is a
who uses it as a sentence. As applied to t
characterization of a process as computationa
from outside; and the identification of the pr
intrinsic feature of the physics, it is essential
This point has to be understood precisely. I
the patterns we could discover in nature. We
in my brain that was isomorphic to the im
computer. But to say that something is func
something more than that a pattern of ph
assignment of a computational interpretatio
discover in nature objects which had the sam
therefore be used as chairs; but we could
functioning as chairs, except relative to some
chairs.

V. Second Difficulty: The Homunculus Fallacy

is Endemic to Cognitivism

So far, we seem to have arrived at a problem. Syntax is not part of physics. This has
the consequence that if computation is defined syntactically then nothing is intrinsically
a digital computer solely in virtue of its physical properties. Is there any way out of this
problem? Yes, there is, and it is a way standardly taken in cognitive science, but it is out
of the frying pan and into the fire. Most of the works I have seen in the computational
theory of the mind commit some variation on the homunculus fallacy. The idea always
is to treat the brain as if there were some agent inside it using it to compute with. A
typical case is David Marr (1982) who describes the task of vision as proceeding from a
two-dimensional visual array on the retina to a three-dimensional description of the
external world as output of the visual system. The difficulty is: Who is reading the
description? Indeed, it looks throughout Marr's book, and in other standard works on the
subject, as if we have to invoke a homunculus inside the system in order to treat its
operations as genuinely computational.
Many writers feel that the homunculus fallacy is not really a problem, because, with
Dennett (1978), they feel that the homunculus can be "discharged". The idea is this: Since
the computational operations of the computer can be analyzed into progressively simpler
units, until eventually we reach simple flip-flop, "yes-no", "1-0" patterns, it seems that the
higher-level homunculi can be discharged with progressively stupider homunculi, until
finally we reach the bottom level of a simple flip-flop that involves no real homunculus at
all. The idea, in short, is that recursive decomposition will eliminate the homunculi.
It took me a long time to figure out what these people were driving at, so in case
someone else is similarly puzzled I will explain an example in detail: Suppose that we have

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
PRESIDENTIAL ADDRESSES 29

a computer that multiplies six tim

it?" Well, the answer might be t
"How does it add six to itself seve
of thenumerals into binary notati
on binary notation until finally w
are of the form, "Print a zero
intelligent homunculus says "I kn
But at the next lower-level he is
actually know how to do multi
stupider ones who say 'We do not
we know how to convert decima
do not know anything about any
symbols." At the bottom level ar
zero one". All of the higher leve
really exists; the top levels are al
Various authors (e.g. Haugeland
say that the system is a syntact
face the question we had before:
What facts about the bottom leve
ones? Without a homunculus that s
have a syntax to operate with. Th
recursive decomposition fails, be
physics is to put a homunculus in
There is a fascinating feature to
higher levels of computation, e.g. "
really there that corresponds
homunculus/beholder. But they
electronic circuit, they admit, d
manipulate O's and l's and these m
to concede that the higher levels
to concede that the lower levels a
with us.

For real computers of the kind y

each user is the homunculus in
digital computer, we are still fac
homunculus questions in cognitiv
system compute shape from shad
retinal image?" A parallel question
to travel in the board from the im
the answer is the same in both sor
intrinsically neither nails nor vis
might describe them computatio
understand hammering by suppo
hammering algorithms and you
implementing, e.g., the shape fro

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
30 APA PROCEEDINGS, VOL. 64, NO.3

VI. Third Difficulty: Syntax Has No

Certain sorts of explanations in the natu

function causally in the production of the ph
common in the biological sciences. Think of t
photosynthesis, the DNA theory of inherited
natural selection. In each case a causal mech
specification gives an explanation of the outp
and look at the Primal Story it seems clear tha
Cognitivism. The mechanisms by which brain
to be computational, and by specifying the pr
cognition. One beauty of this research program
to know the details of brain functioning in o
provide only the hardware implementation of th
is where the real cognitive explanations are g
Newell for example, there are three levels
intentionality (Newell calls this last level,
contribution of cognitive science is made at th
But if what I have said so far is correct, then
project. I used to believe that as a causal acc
false; but I now am having difficulty formula
the point where it could be an empirical thes
lot of symbols being manipulated in the brai
lightning speed and invisible not only to the
electron microscope, and it is these which caus
O's and 1's as such have no causal powers at al
the eyes of the beholder. The implemented pro
of the implementing medium because the pro
beyond that of the implementing medium. Ph
a separate "program level".
You can see this if you go back to the Pr
difference between the mechanical computer
human computer there really is a program level
causally at that level to convert input to outpu
following the rules for doing a certain com
performance. But when we program the m
computation, the assignment of a computation
outside homunculi. And there is no longer a le
system. The human computer is consciously f
behavior, but the mechanical computer is not
designed to behave exactly as if it were follow
purposes it does not matter. Now Cognitivism
commercial computer and this causes cog
commercial computer and brain have only patt
in addition to those of the implementing med
could give a causal account of cognition.

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
 PRESIDENTIAL ADDRESSES 31

However there is a puzzle for

casually knows that we often
program. For example, we can s
because the machine is implemen
looks like an ordinary causal exp
that syntax, as such, has no
explanations that appeal to pr
explanations provide an appropri
Could we for example rescue the
notion "thermostat" figures in c
physics of its implementation, so
independent of the physics.
To explore this puzzle let us tr
Primal Story Co to show how the
practice. The idea, typically, is
some cognitive capacity, such as
that gives us at least Turing e
running the same program as the
for indirect psychological evid
causally explain the behavior of
same sense in which we can expl
is wrong with that? Doesn't it sou
We know that the commercial c
a program, and in the brain w
explanation.
Two things ought to worry us
accept this mode of explanation fo
how it worked at the neurobiolo
of system that we can simulate c
example, the famous account of
al. 1959 in McCulloch, 1965). The
physiology of the frog's nervou
this:

1. Sustained Contrast Detectors

An unmyelinated axon of this group does not respond when the general illumination
is turned on or off. If the sharp edge of an object either lighter or darker than the
background moves into its field and stops, it discharges promptly and continues
discharging, no matter what the shape of the edge or whether the object is smaller or
larger than the receptive field. (p. 239).

I have never heard anyone say that all this is just the hardware implementation, and
that they should have figured out which program the frog was implementing. I do not
doubt that you could do a computer simulation of the frog's "bug detectors". Perhaps

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
32 APA PROCEEDINGS, VOL. 64, NO.3

someone has done it. But we all know that o

system actually works, the "computational lev
To illustrate the second point, consider sim
for example, typing these words on a mac
fashioned mechanical typewriter. [4] As sim
simulates a typewriter better than any Al pr
sane person thinks: "At long last we unde
implementations of word processing program
computational simulations provide causal exp
So what is going on? We do not in general
of brain processes give us any explanations i
accounts of how the brain actually work
computational simulation of Y" to name a
suppose that because the computer simulates
simulates a computer. We do not suppose tha
hurricane, that the causal explanation of the
program. So why should we make an excepti
processes are concerned? Are there any good
what kind of a causal explanation is an expla
Here, I believe, is the solution to our puzzle
the system, you are left only with a pattern
could attach a computational interpretation. N
of the pattern by itself provides a causal ex
pattern exists in a system you know that some
So you can, for example, predict later stages
already know that the system has been prog
give explanations that make reference to th
say, e.g., this machine behaves the way it
explaining that this book begins with a bit ab
long passages about a bunch of brothers, bec
Dostoevsky's The Brothers Karamazov. But y
a typewriter or a brain by identifying a pat
simulation, because the existence of the patte
works as a physical system. In the case of cog
of abstraction to explain such concrete ment
occurrence of a visual perception or the unde
Now, I think it is obvious that we cannot ex
by pointing to formal patterns they share w
it not obvious in the case of the brain?
Here we come to the second part of our so
for Cognitivism we were tacitly supposing that
for cognition, in the same sense that Tur
computer implement algorithms. But it is pr
to be mistaken. To see this, ask yourself wh
algorithm. In the human computer the system
algorithm, so the process is both causal and l

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
PRESIDENTIAL ADDRESSES 33

a set of rules for deriving the o

agent is making a conscious eff
mechanical computer, the whol
homunculus the system is bo
provides an interpretation to
hardware of the machine caus
cannot be met by the brute, blin
In the brain computer there is n
as there is in the mechanical com
to attach a human computer, bu
is in the computational interpre
the brain is a pattern of events
the mechanical computer, but t
and hence explains nothing.
In sum, the fact that the attri
fatal to the claim that programs
consequences of this, let us remi
like. Explanations such as Choms
account of vision proceed by sta
converted into a symbolic outpu
S, is converted into any one of a
application of a set of syntac
dimensional visual array are con
in accordance with certain algor
tationaltask, the algorithmic sol
algorithm has (like Newell's dist
pattern of the explanation.
If you take these explanations
that it is just as if a man alone i
rules to generate English senten
us ask what facts in the real wo
applied to the brain. In Chomsk
the agent consciously goes thro
supposed to think that he is un
Rather the rules are "computatio
what does that mean? Well, we
computer. The sort of thing tha
a commercial computer is suppo
brain. But we have seen that in t
relative, the ascription is mad
interpretations to the hardware
just an electronic circuit. So h
homunculus? As far as I know, n
even thought there was such
explanatory power to the post

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
 34 APA PROCEEDINGS, VOL. 64, NO.3

mechanism, the brain, with its various real

description.

VII. Fourth Difficulty: The Brain Does

Not Do Information Processing

In this section I turn finally to what I think is, in some ways, the central issue in all
of this, the issue of information processing. Many people in the "cognitive science"
scientific paradigm will feel that much of my discussion is simply irrelevant and they will
argue against it as follows:

There is a difference between the brain and all of these other systems you have been
describing, and this difference explains why a computational simulation in the case of
the other systems is a mere simulation, whereas in the case of the brain a computa-
tional simulation is actually duplicating and not merely modeling the functional
properties of the brain. The reason is that the brain, unlike these other systems, is an
information processing system. And this fact about the brain is, in your words,
"intrinsic". It is just a fact about biology that the brain functions to process
information, and since we can also process the same information computationally,
computational models of brain processes have a different role altogether from
computational models of, for example, the weather. So there is a well defined research
question: "Are the computational procedures by which the brain processes information
the same as the procedures by which computers process the same information?

What I just imagined an opponent saying embodies one of the worst mistakes in
cognitive science. The mistake is to suppose that in the sense in which computers are used
to process information, brains also process information. To see that that is a mistake,
contrast what goes on in the computer with what goes on in the brain. In the case of the
computer, an outside agent encodes some information in a form that can be processed by
the circuitry of the computer. That is, he or she provides a syntactical realization of the
information that the computer can implement in, for example, different voltage levels.
The computer then goes through a series of electrical stages that the outside agent can
interpret both syntactically and semantically even though, of course, the hardware has no
intrinsic syntax or semantics: It is all in the eye of the beholder. And the physics does not
matter provided only that you can get it to implement the algorithm. Finally, an output
is produced in the form of physical phenomena which an observer can interpret as symbols
with a syntax and a semantics.
But now contrast that with the brain. In the case of the brain, none of the relevant
neurobiological processes are observer relative (though of course, like anything, they can
be described from an observer relative point of view) and the specificity of the neurophysi-
ology matters desperately. To make this difference clear, let us go through an example.
Suppose I see a car coming toward me. A standard computational model of vision will
take in information about the visual array on my retina and eventually print out the
sentence, "There is a car coming toward me". But that is not what happens in the actual
biology. In the biology a concrete and specific series of electro-chemical reactions are set
up by the assault of the photons on the photo receptor cells of my retina, and this entire

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
PRESIDENTIAL ADDRESSES 35

process eventually results in a co

of a bunch of words or symbols
of a concrete specific consciou
concrete visual event is as specif
meal. We can, with the compute
of its production, as we can do an
or any other phenomenon, but
processing systems.
In short, the sense of informa
much too high a level of abstrac
intentionality. The "information"
It is specific to thought, or vi
information processing which is
of cognition, on the other hand,
in response to a set of symbols a
We are blinded to this differe
coming toward me", can be used
of the computational model of v
the visual experience is a concret
chemical biological processes. To
manipulation is to confuse the r
discussion is that in the sense of
to say that the brain is an infor

VI. Summary of the A

This brief argument has a simp

1. On the standard textbook def

of symbol manipulation.

2. But syntax and symbols ar

tokens are always physical tok
terms of physical features. Sy

3. This has the consequence tha

assigned to it. Certain physic
or interpreted syntactically. S

4. It follows that you could

intrinsically a digital comp
interpretation to it as you cou
"The brain is a digital compu
of falsehood. It does not have
account if you think that I am
digital computer. The question

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
36 APA PROCEEDINGS, VOL. 64, NO.3

as the questions "Is it an abacus?", "Is it a

a set of mathematical formulae?"

5. Some physical systems facilitate the com

That is why we build, program, and
homunculus in the system interpreting th
terms.

6. But the causal explanations we then give do not cite causal properties d
from the physics of the implementation and the intentionality of the hom

7. The standard, though tacit, way out of this is to commit the homunculu
The homunculus fallacy is endemic to computational models of cognit
cannot be removed by the standard recursive decomposition arguments
addressed to a different question.

8. We cannot avoid the foregoing results by supposing that the brai

"information processing". The brain, as far as its intrinsic operat
concerned, does no information processing. It is a specific biological org
specific neurobiological processes cause specific forms of intentionalit
brain, intrinsically, there are neurobiological processes and sometimes t
consciousness. But that is the end of the story. [5]

Footnotes

[1] For earlier explorations see Searle (1980) and Searle (1984).
[2] This view is announced and defended in a large number of books and articles
many of which appear to have more or less the same title, e.g. Computers and Though
(Feigenbaum and Feldman, eds. 1963), Computers and Thought (Sharples et al, 1988), Th
Computer and the Mind (Johnson-Laird, 1988), Computation and Cognition (Pylyshyn, 198
"The Computer Model of the Mind" (Block, 1990, forthcoming), and of course,
"Computing Machinery and Intelligence" (Turing, 1950).
[3] People sometimes say that it would have to add six to itself eight times. But tha
is bad arithmetic. Six added to itself eight times is fifty-four, because six added to itsel
zero times is still six. It is amazing how often this mistake is made.
[4] The example was suggested by John Batali.
[5] I am indebted to a remarkably large number of people for discussions of the issue
in this paper. I cannot thank them all, but special thanks are due to John Batali, Vino
Goel, Ivan Havel, Kirk Ludwig, Dagmar Searle and Klaus Strelau.

References

Block, Ned (1990), forthcoming "The Computer Model of the Mind".

Boolos, George S. and Jeffrey, Richard C. (1989). Computability and Logic. Cambridge
University Press, Cambridge.

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms
 PRESIDENTIAL ADDRESSES 37

Dennett, Daniel C. (1978). Brainst

Press, Cambridge, Mass.
Feigenbaum, E.A. and Feldman
Company, New York and San F
Fodor, J. (1975). The Language o
Haugeland, John, ed. (1981). M
Johnson-Laird, P.N. (1988). Th
Cambridge, Mass.
Kuffler, Stephen W. and Nich
Associates, Sunderland, Mass.
Lettvin, J.Y., Maturana, H.R
Frog's Eye Tells the Frog's Brai
1940-1951, reprinted in McCu
Marr, David (1982). Vision. W
McCulloch, Warren S. (1965). Th
Pylyshyn, Z. (1985). Computatio
Searle, John R. (1980). "Minds,
3, pp. 417-424.
Searle, John R. (1984). Minds, Brains and Science. Harvard University Press, Cambridge,
Mass.
Sharples, M., Hogg, D., Hutchinson, C., Torrance, S., and Young, D. (1988). Computers
and Thought. MIT Press, Cambridge, Mass. and London.
Shepherd, Gordon M. (1983). Neurobiology. Oxford University Press, New York and
Oxford.
Turing, Alan (1950). "Computing Machinery and Intelligence." Mind, 59, 433-460.

This content downloaded from

177.38.180.22 on Tue, 13 Jul 2021 03:21:55 UTC
All use subject to https://about.jstor.org/terms