Open Journal of Philosophy
2011. Vol.1, No.1, 1-10
Copyright © 2011 SciRes. DOI:10.4236/ojpp.2011.11001
Three Philosophical Problems about Consciousness and Their
Possible Resolution
Nichol as M axw ell
Emeritus Reader in Philosophy of Science at University College London, London, Britain.
Email: nicholas.maxwell@ucl.ac.uk
Received July 29th, 2011; revised August 12th, 2011; accepted August 15th, 2011.
Three big philosophical problems about consciousness are: Why does it exist? How do we explain and under-
stand it? How can we explain brain-consciousness correlations? If functionalism were true, all three problems
would be solved. But it is false, and that means all three problems remain unsolved (in that there is no other ob-
vious cand id at e for a s olu ti on ). Here, it is argu ed tha t th e firs t prob lem c ann ot ha ve a s olu tion ; thi s is inheren t in
the natu re of exp la na tion . Th e sec ond prob lem i s solved b y recogn izin g tha t (a ) there i s an exp lan at ion as t o why
science cannot explain consciousness, and (b) consciousness can be explained by a different kind of explanation,
empathic or “personalistic” explanation, compatible with, but not reducible to, scientific explanation. The third
problem is solved by exploiting David Chalmers’ “principle of structural coherence”, and involves postulating
that sensations experienced by us—visual, auditory, tactile, and so on—amount to minute scattered regions in a
vast, multi-dimensional “space” of all possible sensations, which vary smoothly, and in a linear way, throughout
th e sp ace. Th ere i s als o th e s pac e of all p o ss ib le s en t i ent br ai n p roc ess es. Th ere i s ju s t on e, un iq u e one-on e ma p -
ping between these two spaces that preserves continuity and linearity. It is this which provides the explanation as
to why brain processes and sensations are correlated as they are. I consider objections to this unique-matching
theory, and consider how the theory might be empirically confirmed.
Keywords: Consciousness, Mind-Body Problem, Brain Processes, Explaining Consciousness, Functionalism,
Experiential Functionalism, Physical Explanation, Empathic Understanding, Sensation-Brain
Correlations, Unique Matching Theory, Hard Problem of Consciousness
The Three Problems
I am inclined to t hink that t here are three basi c philosophical1
problems that arise in connection with consciousness.
(1) The Problem of Existence. Why does sentience or con-
sciousness exist at al l ? Why are we not zo mbi es2?
(2) The Problem of Intelligibility. Granted that consciousness
exists, what is it? How is it to be explained and understood? On
the face of it, t here cou ld be no greater myst ery than t hat brain s
should somehow produce, or be, our states of awareness, our
thoughts, feelings, perceptions and desires. What is so baffling
and mysterious about consciousness is that each one of us
knows it exists, and knows what it is, because we possess it,
indeed we are it, in a certain sense; and yet, if we examine a
conscious brain, we find such things as neurons and synaptic
junctions, but nothing remotely like consciousness as we ex-
perience it. Consciousness is wholly apparent to the owner of
the conscious brain, but bafflingly invisible and ineffable to
everyone else. How is this familiar and utterly inexplicable
stuff of consciousness to be explained and understood?
(3) The Problem of Explaining Brain-Mind Correlations.
What possible explanation could there be for the way brain
processes and sensati ons are correlated?
In wha t fo l lo ws I s uggest solutions to two of these problems,
and indicate why, in my view, the other problem has no solution,
and thus does not de s e r ve to be r egar de d as a legitima te problem.
Functionalism, If Correct, Solves the Three
Fundamental Problems of Consciousness
If functio nalism is correct, all th ree problems are solved at a
stroke. According to functionalism—as I think it ought to be
formulated—the mental aspect of brain processes is simply
what ma y be call ed t he “co n trol ” aspect , t hat aspect i nvo lved in
guiding the animal or person to act in the way that they do3.
Viewed from a Darwinian perspective, the function of the brain
is to control the animal to act in ways conducive to survival and
reproductive success in the given environment. In referring to
sensations, perceptions, feelings, desires, states of awareness,
imaginings, thoughts, decisions to act, we are referring to neu-
rological processes going on in the brain from the standpoint of
their role in guiding or controlling action: detecting bodily
3Some functionalists take the analogy with the computer and the Turing
Universal Machine very seriously, and hold that a major task is to give
definitions to interconnected psychological terms: see (Putnam, 1960; Lewis,
1972; Rey, 1997). The latter is a characteristic obsession of analytic phi-
losophy: for a corrective see (Popper, 1962: pp. 9-21). I am inclined to see
functionalism as a minor modification of the kind of brain process theory
defended by (Place, 1956; Smart, 1963; Armstrong, 1968)—a modification
whic h stress es tha t the mind is to be id entified with the fun ctional or control
aspect of the brain, the stuff of the brain being irrelevant. This viewpoint takes
the basic task to be to solve the philosophical mind-brain problem, it being
important to put this problem into the context of biology and evolution: see
(Maxwell, 1984: pp. 174-181, 269-273); and (Maxwell, 1985). The argument
of the present paper does not, however, require that functionalism be
interpr eted specifical l y in th is “control”, Darwinian fashion.
1By a philosophical problem I mean a conceptual problem so baffling that
we don’t know whether or not it is a serious problem of knowledge and
understanding, there being no agreement as to what the problem is or what
would count as a solution. One basic task for philosophy is to try to clarify
serious philosophical problems so that they turn into fruitful empirical, or
solvable, problems: see final section.
2See (Ca mpbell, 1970 ; Kirk, 1974).
N. MAXWE LL
2
changes or aspects of the environment (sensation and percep-
tion), assessing significance and prompting appropriate kind of
response (feeling), determining or influencing choice of goals
(desire), registering the current environmental situation (aware-
ness), or exploring possibilities (imagining); and so on. Ac-
cording to functionalism, the mental aspect of brain processes is
nothing more than this kind of control aspect.
This means that any brain, of whatever constitution or struc-
ture, that is sufficiently sophisticated to produce action just like
the actions of a conscious person, thereby has a conscious,
mental aspect just like the conscious, mental aspect of our
brains, the brains of conscious persons. A zombie who behaves
like a conscious person is a conscious person. Philosophical
zombies d o not, and cannot, exist.
One slight qualification to this conclusi on can be recognized
by functionalism. It is just about conceivable that a robot with-
out a brain transmits radio signals to a vast, very rapid com-
puter, which calculates what the robot would do, on the basis of
received in formati on , were it to have su ch an d such a brain , and
then transmits instructions to the robot as to how it should act.
The robot acts as if con sciou s, as if it had a co nscio us brai n, bu t
in this case no such brain exists, but only a model of it in the
computer, and so the robot is not conscious. It is a zombie.
For functionalism, then, there is no philosophical or concep-
tual problem concerning the existence of sentience or conscious-
ness—or rather, in so far as there is a problem, functionalism
solves it.
Functionalism also solves the problem of intelligibility, the
problem of understanding what the nature of consciousness is.
Sentience and consciousness are no more than the relevant
control aspects of brains sufficiently sophisticated to produce
action that we would describe as “sentient” and “conscious” in
character.
And functionalism also solves the third problem, the problem
of what possible explanation there can be for the way brain
processes an d sensations are correlated. The mental aspect of a
brain process is given by the role that process plays in guiding
the animal or person to act in the way he or she does (possibly
taking counterfactual situations into account). Correlations are
between the neurological processes, described as neurological
processes, and these processes described in terms of the control
role they play in producing actual and potential actions. What
the control role of a neurological process is will depend on such
things as its physical or neurological character, how it is situ-
ated in the brain, what the overall functioning structure of the
brain is, what other functionally described brain processes the
given process can, in part, cause to occur, when the rest of the
brain is in this or that state. There is, in short, according to
functionalism, no big mystery, no philosophical or conceptual
problem, about why the neurological and mental aspects of
brain processes are correlated in the ways that they are. There
are, of course, immense and highly intractable empirical prob-
lems about how precisely neurological and mental (or control)
aspects are correl at ed made al l the more difficult to solve by the
complexity of the conscious brain, and by the moral objections
to investigating the conscious brain in an invasive manner.
Functionalism highlights the importance and intractability of
these empirical problems4, but disposes of the problem of how
there could possibly be an explanation for brain-mind correla-
tions. Given functionalism, there is no such problem.
Thus functionalism, if correct, disposes of the three funda-
mental philosophical problems of consciousness at a stroke. No
wonder it is a popular view.
Functionalism Is Not Correct
Functionalism is, however, untenable. A simple, well known
argument shows decisively that functionalism cannot be correct.
The argu ment goes like this.
Functionalism is put forward as a part of the reductionist
programme of natural science, and can legitimately be assessed
in that light. What functionalism achieves, if correct, is to show
that there i s nothing asso ciated with consci ous b rains whi ch lies
irredeemably beyond the scope of scientific explanation. The
mental aspect of brain processes is no more than the control
aspect which will, one day, be explained and understood in
neurological terms, in terms of brain structure and functioning,
which in turn will be explained and understood in biological,
chemical and molecular terms and, ultimately, in principle (if
not in practice), in physical terms. We are physical systems put
together by evolution to function in extraordinary ways, but
nevertheless in ways that are ultimately, in principle at least,
fully explicable physically. The brain is just another organ, with
its speci fic function, li ke the heart, the lungs, the stomach o r the
liver: one day science will give us just as good an explanation
of the structure and functional aspect of the brain as it does at
present of the other organs5.
But physics, and that part of natural science in principle re-
ducible to physics, cannot conceivably predict and explain fully
the mental, or experiential, aspect of brain processes. Being
blind from birth—or being deprived of ever having oneself
experienced visual sensations—cannot in itself prevent one
from understanding any part of physics. It cannot prevent one
from understanding the physics of colour, light, physiology of
colour perception and discrimination, just as well as any nor-
mally sighted person. In order to understand physical concepts,
such as mass, force, wavelength, energy, spin, charge, it is not
necessary to have had the experience of any particular kind of
sensation, such as the visual sensation of colour. All predictions
of physics must also have this feature. In order to understand
what it is for a poppy to be red, however, it is necessar y to have
experienced a special kind of sensation at some time in one’s
life, namely the visual sensation of redness. A person blind
from birth, who has never experienced any visual sensation,
cannot know what redness is, where redness is the perceptual
property, what we (normally sighted) see and experience, and
not some physical correlate of this, light of such and wave-
lengths, or the molecular structure of the surface of an object
which causes it to absorb and reflect light of such and such
wavelengths. It follows that no set of physical statements,
however comprehensive, can predict that a poppy is red, or that
a person has the visual experience of redness. Associated with
neurological processes going on in our brains, there are mental
or experiential features which lie irredeemably beyond the
scope of physical description and explanation. Functionalism is
4One of the gre at virtue s of funct ionalism i s that it transforms the apparently
utterly inexplicable philosophical mind-body problem into a problem that is,
in principle, an empirical problem, however intractable it may be. This virtue
is reta i ned by the mo difie d “experien tial func tio na l ist” view I defend below .
5Some have argued, however, that functionalist phenomena are not reducible
to physics: see, for e xampl e, ( Block, 1990 : p. 446; M a xwell, 2001 ) .
N. MAXWELL 3
thus shown to be false6.
I might mention in passing that this argument, usually attrib-
uted to (Nagel, 1974) and (Jackson, 1982, 1986), was actually
first put forward by me several years before Thomas Nagel and
Frank Jackson, in two papers published in 1966 and 19687.
There are two other arguments, in addition to the above col-
our-blindness argument, regularly employed by philosophers to
establish the incompleteness of physics, the falsity of function-
alism. There is, first, the inverted spectrum argumen t: it is con-
ceivable that a person might make the same colour discrimina-
tions as I do, and might have a similar physiology, but might
experience an inverted spectrum, seeing red when I see blue
and blue when I see red8. In the two cases, th e physics is essen-
tially the same, but the experience is different; hence physics
cannot be complete. Second there is the zombie argument: it is
conceivab le that I h ave a twin, pr ecisely the sa me as me ph ysi-
cally, but de v oid of consc i ousne ss : henc e phy s ics is inc om ple te9.
The three arguments are related to one another: the col-
our-blindness argument considers sensorially deprived persons;
the inverted spectrum argument considers the possibility of
people with different sensory experiences; and the zombie ar-
gument considers the possibility of complete sensory depriva-
tion. The great advantage of the colour-blindness argument
over the other two, however, is that it alone does not rely
merely on what is conceivable or possible: we know there are
people who are colour-blind; we can consider what happens
when people who are blind from birth have their sight restored.
These are actualities, not mere possibilities.
The three arguments have been much discussed. Some hold
onto functionalism and reject the arguments10; others hold that
the arguments are valid, and reject functionalism11. Here I as-
sume that one or more argument is valid, and functionalism has
been shown to be false12.
At once we are confronted again by the three philosophical
problems of consciousness with which we began (given that
there is no other obvious candidate for the solution). In what
follows I sketch a two-aspect theory of consciousness which, I
claim, can be deployed to solve the second of the above three
problems, the problem of intelligibility. I then put forward a
unique-matching theory which is able to solve the third of the
above three problems, the problem of explaining brain-mind
correlations.
Ex peri ential Funct ionalis m
Before us there is, let us suppose, another conscious or sen-
tient being, whether person, animal, alien, or even, possibly,
robot or android. What is this utterly mysterious sentience or
consciousness, associated with the brain processes of the other
being? Why does it resist scientific explanation? How is it to be
explained and understood?
Senti ence or con sci ousn ess, acco rdin g to th e two-aspect vi e w
I wish to d efend, is that aspect or feature of a brain pro cess that
we can only get to know about as a result of having a suffi-
ciently similar brain process occurs in our own brain. It is what
it is to have that kind of process occur in one’s own brain. It is
just that, and nothing more.
This thesis, note, does justice to the baffling privacy of con-
sciousness, expressed above in problem (2). If the mental as-
pect of a brain process is just what we get to know about in
having that process occur in our own brain, then of course we
cannot discern the mental aspect if the process occurs in an-
other’s brain. In order to discern the mental aspect it is neces-
sary and sufficient to ensure that a sufficiently similar process
occurs in our brain (assuming our brain is sufficiently similar to
the other brain). However hard we peer at another person’s
brain, and however probing and thorough our investigation, we
will never, in that way, detect the faintest hint of sentience or
consciousness.
But if I want to know what the other being is experiencing in
having a brain process, N, occur in his brain, how “sufficiently
similar” a brain process, M, must occur in my brain (and how
“sufficiently similar” must my brain be)? There are at least six
possibilities.
6The nub of this argument is that there are real features in the world (sensory
qualities as we experience them) which are such that, if F is such a feature,
then it is necessary oneself to have experienced a specific kind of F-sensati on
if one is to know what F is, which means (according to the view to be
defended below) that it is necessary to have had a specific kind of brain
process, functionally specified, occur in one’s own brain. No physical pro-
perty is like this. Hence physics cannot predict F-type features, if they exist.
On the face of it, they do exist: colours, sounds, smells as we experience them
(whether interpreted as being without or within us) appear to be just such
features. Some rebuttals of the argument, such as (Mulhauser’s, 1998), fail to
come to grips wit h the arg ume nt w hen formulated as abov e.
7See (Maxwell, 1966, es pecially pp. 303-308 ); and (Maxwell, 1968b , espe-
cially pp. 127, 134-137, 140-141). When I recently drew Thomas Nagel’s
attention to these publications, he remarked in a letter, with great generosity:
“There is no justice. No, I was unaware of your papers, which made the central
point before anyone else”. Frank Jackson acknowledged, however, that he had
read my 19 68 paper.
8This possibility was first discussed by (Locke, 1961). For discussion see
(Dennett, 1 991: p p. 389-398; Cha l m er s , 1 99 6: pp. 99-101) .
9Versions of this argument have been discussed by (Kirk, 1974) and
(Campbell, 1970): the argument is rejected by (Dennett, 1991), and endorsed
by (Cha lme rs, 1996: pp. 94-99) .
10(Nemi row, 1 990 ; Lewis , 1990 ; De nn et t, 1991 : pp . 398 -406 ; Cla rk, 2 000,
p
p.
30-35).
11(Maxwell, 1966, 1968b, 1984: pp. 259-273; 2001, ch. 5; Nage l, 1974, 1986;
Chalmers, 1996).
12There is also the so-called “argument of the explanatory gap”—see (Levine,
1983)—designed to show that physics cannot explain the experiential. But it
seems to me that the main reason for believing in the “explanatory gap” is the
“col our- blindne s s” argum e nt, sket ched in the text.
(1) N and M are precisely the same physically, even if the
two brains are not pr eci sely the same.
(2) N and M are precisely the same neurologically (i.e. the
same pattern of neurons fire in the same way), even though
there are otherwise differences between the physical states of
the neurons.
(3) Neurons may be quite different physically (e.g. in one
case neurons are biological, in the other case made out of mi-
crochips), but the pattern of firing of the neurons, and the in-
terconnections between the neurons, is the same.
(4) “Strength of signal” may be coded in quite different ways
at the neuronal level (so that in one case this is related to rapid-
ity of firing of neurons, while in the other case it is related to
strength of electric current, let us suppose); once these differ-
ences are igno red , ho wever, the p att ern o f signal s is th e same in
the two cases.
(5) The functional or control role of the neurological proc-
esses, N and M , are ident ical in the two brai ns, even thou gh the
pattern of signals, the “code” at the neuronal level, and the
physical structure and functioning of the neurons, is entirely
different.
(6) The behaviour of the two beings is similar, even though
the cont rol architecture of the two brains is entirely different so
N. MAXWE LL
4
that, from a functional or control standpoint, the neurological
processes, N and M, work in quite different ways.
(1) and (2) require such a high level of similarity between N
and M that they probably imply that we never ourselves have
the same kind of experien ce on different occasi ons. (6) requires
such a low level of similarity between N and M that it is indis-
tinguishable from behaviourism. The robot, considered above,
that has no brain but is controlled by a computer to act as if it is
conscious, satisfies (6); but even functionalism, let alone the
two-aspect view being considered here, can give reasons for
holding the robot is not conscious. We are left with (3), (4) and
(5). It is not eas y to see how, even in princi ple, we could ob tain
evidence to decide between these options. Here, without argu-
ment, I plump for option (5). Sensations are to be correlated
with the control aspect of brain processes—brain processes
functionally described. This version of the two-aspect view
might be called “experiential functionalism” 13.
But if mental features, correlated with brain processes de-
scribed in control or functional terms, really do exist, why do
such mental features lie beyond the scope of physics?
Physics does not, even in principle, predict and explain such
a mental feature because physics is concerned only with those
features of things that need to be referred to in order to predict
how states of affairs evolve with the passage of time. Physics,
in other words, is concerned exclusively with what may be
called t he “causally efficacious” asp ect of things: see (Maxwell,
1968a, 1998: pp. 141-155). Features of things which do not
need to be referred to in order to predict future states of physi-
cal systems, will not be referred to by physics.
Suppose that the world is such that there is a yet-to-be-dis-
covered, unified, explanatory, true physical “theory of every-
thing”. Suppose further, to keep the argument simple, that this
theory is deterministic and classical in character. Such a theory is
compreh ensive an d complet e—a th eory of ev erythi ng—because,
given any isolated system, the theory, together with a precise
specification of the instantaneous physical state of the system
(formulated in the highly specialized, restricted vocabulary of
the theory), predicts future states of the system, described in
terms of the same cau sally efficaci ous (i. e. physical) prop erties.
But to say this is not to say that the theory predicts everything
about the system, all facts about the system. If the system in-
cludes a conscious being, the comprehensive physical descrip-
tion of the system will include a precise specification of the
physical state of the being’s brain. But in order to carry out the
predictive task of physics there will be no need to refer to the
mental aspect of the brain, what it is to have that kind of proc-
ess occur in one’s own brain. As a result, the “theory-
of-everything” will make no mention of such a mental feature.
But could not the physical “theory-of-everything” be ex-
tended so that it includ es reference to mental features, an d thus
becomes a genuine theory of everything? Let it be conceded
that thi s can be done. The crucial poin t to appreciat e is that the
new, amplified theory would be so horribly complex and ad hoc
that it would entirely cease to be explanatory. Given the vast
richness and complexity of the experiential world, and given
the mind-boggling complexity of the manner in which even the
most elementary of mental features, such as the visual sensati on
of redness, are correlated with physical states of affairs, the
unity and explanatory power of the physical “theory-of-every-
thing” would be entirely lost in the amplified theory. The am-
plified theory would consist of millions, possibly billions, of
distinct postulates linking physical and mental features, each
postulate itself of incredible complexity. All this would be in
striking contrast to the fundamental simplicity, unity and ex-
planatory power of the physical “theory-of-everything”.
There is, in short, an explanation as to why physics does not,
and cannot, include the mental, the experiential. If it did, the
extraordinary explanatory power of physical theory would van-
ish. Excluding the experiential is the price we pay for having the
marvellously explanatory theories that we do have in physics14.
It should be noted that the silence of physics about the expe-
riential provides no grounds whatsoever for holding that the
experiential does not exist. It is only if we hold that the only
properties that exist are causally efficacious properties (or as-
pects of such properties) that this conclusion follows. But the
thesis that only causally efficacious (or physical) features exist
seems decisively refuted by our experience of colours, sounds,
smells and other sensational features—features which can only
be known as a result of oneself having certain kinds of sensa-
tions, and hence certain kinds of (functionally described) brain
processes occur in one’s own brain.
It is rather natural to suppose that the stubborn resistance to
scientific explanation exhibited by sentience and consciousness
is due to, and is a sign of, their inherent mysteriousness and
inexplicability. This is, indeed, a major reason for supposing
that the mental is inherently inexplicable: the mental is so mys-
terious that it resists above all our very best kind of explanation,
namely scientific explanation. But the supposition is wrong.
Sentience and consciousness (like perceptual properties out
there in the world, such as co lours and sound s) evade scientific
explanation, not because of any stubborn inexplicability, but
because th ey are not required for the ki nd of causal exp lanati on
that science provides, and cannot be incorporated into science
because, if they are, science (or at least that part in principle
reducible to physics) ceases to be explanat ory15.
But if sentience and consciousness cannot be explained and
understood scientifically, how are they to be understood?
Elsewhere I have argued at some length that mental features
can be explained and understood by what may be termed “em-
pathic” or “personalistic” explanation and understanding, a kind
of explanation different from, compatible with, but not reduci-
ble to, scientific explanation16.
To understand another empathically or personalistically is to
know what it would be like to be the other person (or sentient
being), experiencing, feeling, thinking, believing, desiring,
planning and deci ding what that other person experiences, feels,
etc. This involves arranging to occur in one’s own brain neuro-
logical processes that are sufficiently similar (in control terms)
to the processes that are occurring in the other person’s brain,
without this leading to one actually doing what the other is
doing. We imagine that we are the other being. Imagination,
quite generally, is arranging to occur in one’s own brain proc-
esses sufficiently similar to those that would occur were one
actually doing what one imagines one is doing.
Personalistic explanation is fundamentally anthropomorphic
in character, and thus fundamentally distinct from scientific
explanation, which is not anthropomorphic. Scientific under-
13Versions of this view have been defended by (Maxwell, 1966, 1968b,
1984: pp. 174-181; 2001) and (Nagel, 1974, 1986). For a particularly lucid
and detailed exposition of the view see (Chalmers, 1996). 14See (Max well, 2001) for a fuller exposition of this argument .
N. MAXWELL 5
standing never involves relating what is to be understood to
oneself, in an essential way. If, however, I want to understand
another conscious being, an alien let us suppose, as a person,
then it is essential that I bring myself into the picture, and relate
the other to myself. If I want to know what the other, the alien,
is experiencing, p erceiving, feeli ng, etc., what I want to know is
what it would be like for me to be the other, having processes
occur in my brain that are sufficiently similar, in the relevant
respects, to what occurs in the alien’s brain. If I want to under-
stand what the alien says or thinks, then I must discover how to
translate the alien’s language into mine. In every case, mental
or personalistic features of the alien, which lie beyond the
scope of science, are known, and can only be known, by bring-
ing oneself into the picture and relating the other, the alien, to
oneself—by understanding the other anthropomorphically, in
other words17.
According to the psycho-functionalist version of the two-
aspect vie w, indicated abo ve, the mental features o f brain proc-
esses are precisely the kind of features to be explained and
understood personalistically. The mental feature of a brain
process is what we know about in having a sufficiently similar
process occur in our own brain; personalistic explanation and
understanding is a kind of understanding quite specifically de-
signed to enable u s to understand just such a feature.
My solution to the second philosophical problem of con-
sciousness, then, comes in two parts. First, a major reason why
consciousness seems inh erently inexplicable i s because it seems
to be inherently beyond the scope of even our best kind of ex-
planation, namely scientific explanation. There is, however, an
explanation for the incapacity of physics, and science reducible
to physics, to explain consciousness: physics is concerned only
with the causally efficacious aspect of things, and if physical
theory is amplified to include the experiential, it might, in prin-
ciple, be predictive, but it would cease entirely to be explana-
tory—for reasons that can be entirely explained and understood.
Consciousness resists scientific explanation, not because it is
inherently mysterious and inexplicable, but because it is the
kind of thing which science can ignore, given its predictive task,
and must ignore, if it is to be explanatory. Second, conscious-
ness, the mental asp ect of br ai n processes, can be explained and
understood, namely by means of personalistic explanation, a
kind of explanation that is compatible with, but not reducible to,
scientific explanation. The mental aspect of that kind of brain
process that is the visual sensation of redness cannot be under-
stood scientifically, but it is wholly understandable personalis-
tically, for those of us with normal colour vision: we under-
stand what it is, personalistically, in having that kind of process
occur in our own brain.
This argument requires that personalistic explanation, even
though not reducible to scientific explanation is an intellectu-
ally authentic mode of explanation in its own right. Elsewhere18
I have put forward arguments in support of this thesis. I have
argued that it is the evolution of our human capacity for per-
sonalistic understanding that has transformed mere sentience
into consciousness, and made our human language and culture
possible (construed in personalistic terms). Even science pre-
sumes personalistic understanding in that scientists, in order to
understand each other’s ideas, problems and theories, must see
things imaginatively from each other’s perspective, thus ac-
quiring a kind of etiolated personalistic understanding of each
other (one that emphasizes beliefs about aspects of the envi-
ronment and ignores most of the personal dimension). Science
is thus, in a sense, based on personalistic understanding. Per-
sonalistic understanding is not folk psychology—construed to
be a pre-scientific psychology which will be replaced by a
genuinely scientific psychology as knowledge advances.
15An anonymous referee has raised two questions about this argument that
there is an explanation as to why science cannot explain consciousness. First,
the very question “Why does consciousness exist?” is ambiguous. It could
mean either “What is the cause of consciousness?” or “What is consciousness
for?”. I have argued that there is an explanation as to why there cannot be an
explanation for the existence of consciousness when the question is inter-
preted in the first way. As for the second interpretation, it is clear physics, and
that part of science in principle reducible to physics, cannot answer “What is
consciousness for?”, because physics does not deal with function and purpose.
But evolutionary biology, a part of science, does. Perhaps Darwinism can
supply an explanation when the question is interpreted in this fashion?
Orth odox Darwini sm seems incapable, however, o f providin g an answer. For
any such answer would, it seems, have to appeal to the survival value, or
reproductive fecundity, of consciousness, and any such explanation could not,
it seems, explain why zo mbie consciou sness woul d not serve t hese en ds just as
well as real consciousness. I have however, elsewhere, distinguished nine
differe nt ver sions of Darwi nism, which progres sively give greater and grea ter
roles to purposiveness, sentience and consciousness, to the theory: see
(Maxwell, 2010). The last three of these appeal to sentience, and to what may
be called “sentient explanations”, and the last two to consciousness and
“personalistic” explanations, but this hardly amounts to a Darwinian expla-
nation for the evolutionary emergence of sentience and consciousness, as the
context makes clear. These e xpla nations appeal to se ntie nce and con scious ness
when they may be presumed to have evolved, but do not explain why they
evolved in the first place (a non-zombie version, that is). Secondly, the referee
rai ses the question of wheth er physics can even explain biological phenomena,
let alone phenomena associated with sentience and consciousness. There is,
however, in my view, a fundamental difference between the incapacity of
physics to explain biological phenomena, conceived of in zombie terms, and
phenomena associated with consciousness. Nothing going on in connection
with biology, conceived of in zombie terms, cannot in principle (we are
entitled to assume) be explained physically. It is just that physics cannot
supply us w ith c ertai n k in ds of e xpl a natio n – purposive a n d pe rsona l i sti c —as I
go on to make clear. But the situation seems to be very different when we
come to sentience and consciousness. Here there seem to be features of brain
p
rocesses, our inner experiences, which seem to be utterly beyond the scope of
physics, and of science, unless some part of science—psychology, for
example—is interpreted in a special way from the outset so as to include inner
experiences and consciousness.
16See (Maxwell, 1984 : especially pp. 181-189, 264-275) and (Maxwell, 2001).
See also (Maxwell, 2 010).
17See also (Maxwell, 2001: pp. 114-115).
What may be called “teleological” or “purposive” explana-
tion constitutes a thir d kind of explanation. This explains by
interpreting the actions, or growth, of something as being de-
signed to achieve goals. It is a watered-down version of per-
sonalistic explanation, in that actions are explained as being
designed to realize aims, but no attempt is made to enable one
to know what it would be like to be the thing in question: all
reference to the mental or experiential is thus excluded (unless
construed in purely functional terms). It applies equally to
thermostats, robots, guided missiles, and all living things. In my
view, everything can in principle be understood physically; all
living things, and all purposive things created by us can, in
addition, be understood purposively; and sentient things are
open, in addition, to being understood personalistically. If func-
tion al i sm were correct, then all that we are, as persons, could be
explained and understood purposively, and personalistic under-
standing would be redundant. It is the incompleteness of func-
tionalism, its failure to encompass fully the mental, which re-
quires t hat purposi ve explanation be enriched so that it becomes
18See note 1 6.
N. MAXWE LL
6
empathic or personalistic explanation. (Experiential functional-
ism enriches, and does not just reject functionalism as tradi-
tionally conceived.)
Explaining Correlations and the
Unique-Matching Theory
So much for my solution to the second of the three philoso-
phical problems about consciousness, with which we began.
But what about the third problem? How could it be possible to
explain correlations between brain processes and sensations?
Given existing correlations which we may presume hold be-
tween brain processes and visual sensations of colour, why
should not the spectrum of colour sensations be reversed, so
that the sensation of redness is now correlated with that brain
process that was correlated with the sensation of blueness, and
so on?
We have seen, in effect, that no amplification of scientific
theory could explain why correlations between brain processes
and sensations are as they are; any physical theory-of-every-
thing amplified to include the experiential will be just as
non-explanatory whether one considers colour sensations as
they are, or as they would be given a reversed colour spec-
trum.
Furthermore, it seems that no amplification of personalistic
explanation could do the trick either. Suppose there is a
God-like brain, which can experience all possible sensations;
suppose further that He knows all there is to know about the
way His brain processes and sensations are correlated. Despite
this vast store of experience and knowledge, it would seem that
the God-like being is in no better position to explain why brain
processes and sensati ons are correlated as they are than we are.
Until fairly recently, I found this argument convincing. And
then i t stru ck me that th ere is j ust o n e circumstance i n which an
explanation for brain-sensation correlations does exist.
Supp ose that a God-l ike brain is indeed po ssible; it is able to
experience all possible sensations. Suppose, further, that our
visual, auditory, tactile, olfactory and other sensations form
disjoint regions within the continent of sensations experienced
by the God-like being. To us, the different modes of experience
seem utterly distinct in kind. Visual sensations seem utterly
different from auditory ones, which again seem utterly different
from tactile and olfactory sensations. If one had only experi-
enced visual sensations, one could never guess what auditory,
olfactory or tactile sensations would be like. But let us suppose
that the God-like being is able to experience endlessly many
sorts of sensations that lie between our visual, auditory, olfac-
tory and tactile sensations. To Him, as he moves from the vis-
ual to wards the auditory, there is a con tinuous, slight change in
the quality of the sensations experienced, much as there is for
us, within the auditory, when a tone goes continuously up in
pitch. Moving, by means of continuous changes in the quality
of sensations, from visual to auditory, the auditory comes as no
surprise: it emerges as a result of smooth transitions, as in the
case of the tone rising in pitch. And the same goes for transi-
tions between other modes of sensations: the visual, the audi-
tory, the olfactory, the tactile and so on. For the God-like being,
all possible sensations lie in a multi-dimensional “space” of
possible sensations, the experienced quality of sensations vary-
ing smoothly, continuously, indeed in a steady, linear way, as
one moves around in the multi-dimensional “space” (one kind
of smooth variation in sensation, such as changes in pitch of a
sound of specific timbre and loudness, corresponding to one
dimensio n in th e “space” of sens ation s). To the God- like bein g,
this vast realm of possible sensations has a kind of overall co-
herence, a u nity, an overall stru cture, based on the fact that the
quality of sensations varies smoothly, from sensation to sensa-
tion. There is, we may suppose, just one way in which all these
possible sensations can be ordered and “placed” in the multi-
dimensional space, so that continuity is preserved throughout
the space, so that the closer together any two sensations are in
this sp ace then the more n earl y al i ke t hey are experientially.
And let us suppose, further, in accordance with David
Chalmers’ “principle of structural coherence” 19, that all this is
mirrored in the “space” of the functionally described brain
processes that are the sensations. As the God-like being moves
smoothly from experiencing one sensation to experiencing a
slightly different sensation in a neighbouring “place” in the
multidimensional space of all possible sensations, so the brain
process, that is the first sensation, becomes the slightly different
brain process that is the second sensation, slightly different, that
is, when described in functional or control terms. The smooth
coherence and unity of sensation space is matched by a corre-
sponding smooth coherence and unity of brain process space,
when brain processes are characterized in control terms. There
is, as mathematicians would say, something like an isomor-
phism (a common structure) between the space of sensations
and th e space of brain processes. This matchi ng of structure is,
we are to suppose, unique. It can only be done in one way.
Change the way sensations are matched up with (functionally
described) brain processes, and the common structure between
the realm of sensations and the realm of brain processes is lost.
Even a rigid “rotation” of the two realms, the two spaces, with
respect to each other, cannot be done20.
If all this is the case, as a possibility, then an explanation is
possible as to why sensations and brain processes are correlated
in the way that they are. They have to be correlated in this way,
because if they are not, the matching of structure, of unity,
based on continuity, between sensations and brain processes,
would be lost.
If this proposal holds up, then the third philosophical prob-
lem of consciousness has been solved.
Objections
I now consider objections to the above unique-matching ex-
planation for brain-sensation correlations, considered as a pos-
sibility.
A first reaction might be that the theory is much too wild and
crazy to be a serious candidate for truth (even possible truth). I
am almost inclined to agree with this assessment. One should
note, however, that again and again in the history of thought,
ideas that initially seemed wild and crazy have subsequently
become solid, almost prosaic items of knowledge.
19Chalmers’ “principle of structural coherence” asserts that, as far as human
b
rains and states of consciousness are concerned, structural features of brain
p
rocess space match structural features of conscious experience space: see
(Chalmers, 1996: pp. 222-225). (Lockwood, 1989: pp. 109-210) indicates a
similar idea. The unique-matching theory is, in effect, a particular application
of Chalmers’ principle.
20I first put t hi s idea f orwa rd in (Ma xwell, 200 1: pp. 12 6-129) .
N. MAXWELL 7
On similar lines, the idea may be held to be far too vague and
speculative to be taken seriously as a potential contribution to
knowledge. But, again, ideas that were initially vague and spe-
culative, have subsequently become important contributions to
knowled ge: atomis m is an example.
Elsewhere (Maxwell, 1998: especially p. 7 and pp. 80-89) I
have argued that “blueprints”—vague ideas for future theo-
ries—play a vital role in physics. In view of its vague,
open-ended character, the unique-matching idea ought perhaps
to be called a blueprint rather than a theory.
It may be objected that the above unique-matching explana-
tory theory (or blueprint) is extremely limited in scope, in that it
applies only to sensations shorn of all intellectual content, and
does not apply to thoughts, to feelings, to perceptions, imbued
with intellectual content of one kind or another.
But this objection misses the point. What is hardest of all to
explain and understand is why sensations that are stripped of
intellectual content, “raw feels” as they have been called, are
correlated in the way they are with brain processes. This is the
fundamental mystery. When it comes to thoughts, feelings and
perceptions imbued with intellectual content one expects corre-
lations to exist between head processes described “mentalisti-
cally” or personalistically, on the one hand, and described in
functional or control terms, on the other hand. An explanatory
theory, of the kind indicated above, would be all the better for
targeting the hardest part of the problem.
It may be objected, again, that the above matching-structure
explanatory theory, even if in some sense correct, would nev-
ertheless be largely incomprehensible to us human beings since
we, unlike the purely notional God-like being, have not experi-
enced, and cannot experience, all the sensations that intervene
between the isolated patches of sensation, visual, auditory and
so on, th at we do experience. The theory “explain s” at the pri ce
of being incomprehensible.
I have five remarks to make in response to this objection.
First, a theory of sentience or consciousness, couched in
purely functional or control terms, omitting all reference to the
experiential or personalistic, would predict that the notional
God-like being (if it existed) would find it increasingly difficult
to discriminate between sensations (functionally described), as
these become closer together in the “space” of all possible
(functionally described) sensations. (I am assuming, here, that
the uni que-matchin g th eory is correct .) The fun ctio n al theo ry of
sentience would tell us, in effect, that for the God-like being,
sensations are continuously arrayed in the “space” of all possi-
ble sensations, but would not tell us what the God-like being
would experience, where we have not ourselves had the corre-
sponding sensations21. Such a functional theory of conscious-
ness would, of course, be entirely understandable to us.
Second, the matching structure-theory, if correct, would ap-
ply to the isolated patches of sensation that we do experience,
visual, auditory, etc. We can now discern structure in these
distinct modes of sensation, and the unique-matching theory
would predict the existence of this structure, and match it up
with corresponding structure in the distinct spaces of possible
brain processes of a kind that occur in (normally experiencing)
human brains. We would be able fully to understand this part of
the un ique-matching t heory, applicab le to us, even if we wou ld
not be able to have personalistic understanding of that part
which applies to the God-like being but not to us.
Third, it is inherent in the very idea of an explanatory theory
that it should predict new phenomena. The more it does this, so
the more powerfully explanatory it is, other things being equal.
In the case o f the un ique-matchin g theory, predicti ng new phe-
nomena means predicting the potential existence of sensations
of which we have had no experience, and probably cannot ex-
perience. That the theory does predict the potential existence of
these “new” sensations, smoothly joining up visual, auditory,
olfactory sensations and so on, is, it should be noted, essential
to its explanatory capacity. Thus this feature of the theory, far
from bein g a defect, is intrin sic to the explan atory power of the
theory, an inevitab le and even d esirable feature.
Fourth, it deserves to be noted that when one views the mat-
ter from a Darwinian perspective, it would seem not implausi-
ble that there should be a continent of potential sensations
smoothly varying from visual to auditory to olfactory, etc. It is
important to survival that a sentient animal does not confuse
vision with hearing, with smell or with touch; natural selection
would favour changes in brain structure that make distinct
modes of sensation as distinct as possible, and different sensa-
tions within a mode as distinct as possible. It is not unreason-
able to suppose, in other words, that our brains have been de-
signed by Darwinian evolution to deliver to us distinct patches
of sensation as different from one another as possible. Darwin-
ian Theory makes the matching structure-theory seem not un-
reasonable.
Fifth, the solution to the philosophical problem of how it is
possible to explain correlations between brain processes and
sensations being offered here does not require that our isolated,
distinct modes of sensation really are connected up by smoothly
varying, but unknown sensations in a vast space of possible
sensations: all that is required is that this is a possibility, not
that brains really could exist that would experience all these
sensations unknown to us. But it is, of course, always possible
that the unique-matching theory, or something like it, really is
the correct explanation for mind-brain correlations.
It may be objected, yet again, that the above matching-
structu re explan ation does no t explain , and canno t explain, why
this particular brain process correlates with (or is) thi s particu-
lar sensation, the visual sensation of greenness, let us say, or the
sound of a flute softly playing middle C. But this is just what
the theory would explain. Such and such a kind of brain process
(described in functional or control terms) cannot be other than
the sound of a flute softly playing middle C because, if it is
anything else, the isomorphism between the space of (linearly
varying) brain processes, and the space of (linearly varying)
sensations, is lost. Matching of smooth structure requires that
this particular brain process be correlated with this particular
sensation, and not some other sensation.
It may be objected that the unique-matching theory, if true,
must be true analytically; it cannot, therefore, be explanatory.
Suppose the theory is true to the extent that the continuously
varying space of all possible sentient brain processes exists.
The closer to gether two po ints in this sp ace are, the more d iffi-
cult the God-like being finds it to discriminate between the
correspo nding sen sations. This mu ch a correct, comprehensive,
but purely functional theory of the brain will be able to predict,
without any appeal being made to the sentient or phenomenal.
21(Chalmers, 1996: pp. 233-242) makes a related point in connection with
his “principle of structural coherence”.
N. MAXWE LL
8
But if this is the case, then it must follow, analytically, that as
points become close in brain process space, so corresponding
poin ts become close in sensation space. On e simply cannot have
dramatically different sensations, and yet find it very difficult to
discriminate b etween them. So if the sp ace of all sentient brain
processes (functionally described) is continuous, then it follows,
analytically (all but logically) that the same must be true of the
matchin g, correlated space of all possib le sensations.
But what this shows is that the unique-matching theory, if
true, is powerfully explanatory; it does not establish that the
theory itself is analytic. The theory postulates that all (function-
ally described) brain processes that are sentient can be arrayed
in a space which is such that brain processes vary smoothly,
linearly, everywhere, in the manner indicated. This is a factual
postulate which may well be false. The unique-matching theory
is thus also factual and not true analytically; it may well be
false. But if the space of all possible sentient brain processes
varying linearly everywhere exists, then the corresponding
space of all possible sensations must be phenomenologically
linearly varying everywhere as well and, to this extent, there
must be an explanation as to why brain processes and sensa-
tion s are correlated in th e way that they are.
It may be ob jected t hat even if the uniqu e-matching th eory is
true of the God-like being that experiences all possible sensa-
tions, it would still be possible for us human beings to have
brain processes and sensations correlated in a way that is dif-
ferent from the way they are actually correlated; indeed, each
one of us may, for all we know, have quite different sensations
even though we have roughly similar brains and make the same
discriminations. None of this would refute the unique-matching
theory, interpreted as being about the God-like brain; hence the
theory cannot explain brain-sensation correlations in us.
My reply is that a basic presupposition of the unique-
matching theory is that experiential functionalism is correct.
The mental or experiential aspect of a brain process occurring
in another’s brain is what it is like to have that kind of process
(described in functional terms) occur in your own brain
(granted that your brain is sufficiently similar functionally to
the other’s brain). The presupposition is, in other words, that
functionally sufficiently similar brains, that are conscious, ex-
perience the same kind of sensations. This is a presupposition
of the unique-matching theory, and hence something that the
theory cannot explain. What the theory can explain (if true) is
why all functionally sufficiently similar conscious brains all
have the particular sensations that they do have, and not sensa-
tions that are, one and all, locally different (so that all experi-
ence colours reversed, for example).
It is important to appreciate that there are limitations to what
the unique-matching theory can explain. It cannot explain why
functionally sufficiently similar conscious brains all have the
same kind of sensations, as opposed to quite different sensa-
tion s. Nor can it exp l ain why the God-like brain exp eri ences the
sensations He does, and not sensations that are all systemati-
cally different. And nor can the theory explain why there are
sensations at all, and not just zombie-like functional or control
aspects of br ains.
Another objection that may be made is that the God-like
brain, able to experience all sensations, may well be quite im-
possible, even in principle. It is important to appreciate that,
given a more or less specific, localized kind of brain process, this
may well have quite different control and experiential features
associated with it, depending on how it connects up with the
rest of the brain, and depending on what the character of the
rest of the brain is. Given this, there is a problem as to h ow the
God-like brain could possibly have sensations similar to ours –
our brains being very different from and much smaller than, the
God-like brain. It may be, in other words, that it is quite impos-
sible for a single brain to exist that can experience all possible
sensations. But if this is the case, what can it mean to say of the
space of all possible sensations that, in it, sensations vary con-
tinuously as one move through the space? Must we not, rather,
think of the sp ace of all po ssible sensations as being made u p of
isolated islands rather than a single continent, each island cor-
responding to a particular kind of conscious brain, comparisons
of sensations from different islands being meaningless?
Even if the isolated island hypothesis is true, the unique-
matching theory could still have an explanatory role to play. It
could be that the “isolated island” that encompasses all sensa-
tions experienced by human beings is such that, sensations vary
continuously as one moves through the “isolated island” space.
A brain is possible that can experience all these sensations. In
this case, the unique-matching theory would explain why we
cannot experience sensations different from the ones we do
experience, taken fro m our “isol ated isl and” sp ace o f sensatio n s.
The theory would not explain why we cannot have sensations
taken from another “isolated island” space—all the sensations
of the two islands being swapped. It could be, however, that
some other explanation can be found as to why sensations can-
not be swapped wholesale between “isolated islands”.
There is another possibility. Even if the space of all possible
sensations does split up into islands, each island corresponding
to a specific kind of functionally described possible sentient
brain, it could still be that these islands overlap. Overlapping
would make it meaningful to speak of sensations varying con-
tinuo usl y through out the space o f all po ssib le sen sation s. In th is
case, even though the God-like brain is not possible, the unique-
matching theory nevertheless applies to all possible sensations.
Another objection, made in a helpful spirit by David Chalmers
during an email conversation, is that symmetries that obtain in
the space o f human sensation s would have th eir counterparts in
the space of all possible sensations, and hence the unique-
matching theory cannot be correct. An example of symmetry in
our sensation space is the inverted colour spectrum: colour
sensations are inverted but colour discriminations remain the
same.
My reply to this objection comes in three parts.
First, the space of all possible sensations, postulated by the
unique-matching theory, does have a kind of symmetry—a kind
of position invariance: wherever you are, sensations vary line-
arly in all directions. But this is of course different from the
kind of discrete symmetry that Chalmers has in mind, involving
something like a local reflection or rotation. Ignoring the space
of all human sensations for the moment, are there grounds for
holding that the space of all possible sensations must possess
discrete symmetries? What prevents rigid rotations of the whole
space, for e xample? This cou ld be prevented by the spa ce hav-
ing a jagged enclosing boundary, excluding the possibility of
rigid rotation. If the jagged enclosing boundary is modified to
form an enclosing surface of a hypersphere, which would make
rigid rotations possible, then sensations no longer vary at the
N. MAXWELL 9
same rate throughout the space in all directions. But suppose
the boundary is fixed, and suppose points in the space are
moved so as to preserve continuity: what would prevent that?
The answer would have to be that, after the continu-
ity-preserving transition, the rate of change of brain processes
with changes of position would no longer keep pace, through-
out, with the rate of change of sensations with position. Linear-
ity would be violated.
Second, Chalmers’ objection can be interpreted to be that it
is not logically possible for the space of human sensations to
exhibit discrete symmetries (such as colour reversal), and for
such d iscrete symmetries to b e absen t in the space of all p ossi-
ble sensat io ns. But this i s surel y wrong. Postul ate th at the sp ace
of all possible sensations exists, sensations changing in a con-
tinuous and steady (i.e. linear) way with changes in position.
Postulate that there are no local or global discrete symmetries.
Such a space seems logically possible. Given this space, it is
clearly going to be possible to select out widely separated
sub-spaces to constitute all possible human sensations. Fur-
thermore, this can be done in such a way that, given this collec-
tion of sub-spaces, one can easily imagine the possibility of
discrete symmetries, such as colour reversal, obtaining, without
this d estroying continui ty (linearity) in the sp ace of all possible
sensations. No contradiction seems to result from postulating a
continuous space of all possible sensations, in which there are
no discrete symmetries, and a collection of sub-spaces, which
forms all possible human sensations, within which discrete
symmetries are possibl e.
Third, Chalmers’ objection may be interpreted to be, more
modestly, that it is implausible to suppose that the space of all
possible sensations does not exhibit discrete symmetries, espe-
ciall y in vie w of the fa ct that th e space of all human sen sations
does exhibit such symmetries. My reply to this is that at present
we know next to not hing about the character of the sp ace of all
possible sensations, and certainly not enough to be able to de-
clare that the hypothesis of continuity, or linearity, is implausi-
ble. Furthermore, if we accepted that the space of all possible
sensations is globally linear, then it is entirely to be expected
that the space of human sensations will not have this ch aracter,
and will in fact exhibit discrete symmetries. As I have already
pointed out, it is vital for survival that sensations are such that
different sensations can be quickly and easily distinguished,
both within a sensory mode, and between modes. Fruit eating
creatures need to be able to distinguish red from green easily,
so that ripe fruit may be readily seen among green leaves. It
would be disastrous if an animal got confused as to whether
sensations were visual, auditory or tactile. It is entirely to be
expected, in other words, that natural selection will select out
sub-spaces from the space of all possible sensations, to form a
creature’s sensations, which are widely separated in the con-
tinuous, or rather linear, space of all possible sensations. To
form an animal’s sensations from some small, homogeneous
region in the space of all possible sensations would be unhelp-
ful, even disastrous, from the standpoint of survival. Darwinism,
in other words, ensures that, IF the linear space of all possible
sensations exists, THEN the space of all possible sensations of
this or that creature that is a product of evolution, will not be
linear or continuous, and will exhibit discrete symmetries. That
some kind space of all possible sensatio ns exists, l arger than the
space of all possible human sensations, larger even than the
space of all possible sensations of sentient creatures on earth,
seems, on the face of it, entirely reasonable. It is the further
requirement that this space is linear throughout that must be
regarded as u ncertain and speculative.
Finally, it may be asked: what evidence can be gained in
support of the conjecture that the space of all possible sensa-
tions exists, with the crucial feature of linearity required for the
matching structure-theory to be true? It is difficult to see how
we can explore this conjecture experientially, without submitting
ourselves to thoroughly objectionable and radical brain surgery.
It is possible, however, that we may one day develop a func-
tional theory of sentience and consciousness, which meets with
empirical success as far as human brains are concerned (and per-
haps also other mammalian brains), and which makes predic-
tion s about the ch aracter of the space o f all possibl e sensations.
Even hard-nosed functionalists ought, perhaps, to be interested i n
the speculative idea put forward here, because it has implica-
tions for even a purely functionalist theory of consciousness.
This concludes my discussion of objections to the unique-
matching explanation as to why brain processes and sensations
are correlated as they are. If this theory is at least possibly true,
then this solves the third of the above three philosophical prob-
lems of consciousness, with which we began.
Can the Existence of Consciousness Be
Explained?
So far I have suggested solutions to the second and third
philosophical problems of consciousness, but hardly anything
has been said about the first problem, the problem of explaining
the existence of consciousness.
My claim is that this problem has no solution, and is thus not
really a problem at all. Any explanation can only explain X by
showing that it is a manifestation of something else, Y, in terms
of which X can be explained. In asking for an explanation as to
why anything whatsoever exists, rather than nothing, one de-
prives oneself of anything, any Y, in terms of which the expla-
nation is to be couched: this problem is a non-problem. In the
physical realm, we cannot reasonably expect to be able to ex-
plain why the physical universe exists, rather than nothing,
because th is provides u s with nothing, no Y, in terms of which
the existence of the universe may be explained. (This remains
true even if a version of inflationary big bang cosmology is true,
which asserts that the cosmos emerged from the vacuum as a
runaway quantum fluctuation. The physical vacuum or pre-
vacuum is not nothing.) So, too, in the experiential realm, we
cannot explain why the experiential exists, rather than there
being nothing experiential. In this case, we cannot appeal to the
physical, and explain the experiential as emerging from the
physical for, as we have seen, the experiential cannot be de-
rived from the physical. We can, perhaps, explain the way the
physical and the experiential are correlated; but it is inherent in
the very concept of explanation that neither the existence of the
realm of the physical, nor the existence of the realm of the ex-
periential, is capable of being explained. In both cases, the
problem is a non-problem.
Conclusion
Here are a few lessons which, in my view, can be drawn
N. MAXWE LL
10
Clark, A. (2000). A theory of sentience. Oxford: Oxford University
Press.
from the above discussion.
One big division in the community of those who seek to un-
derstand consciousness is between those who hold, roughly,
that functionalism will one day make sense of consciousness,
and those who hold that functionalism is false or incomplete,
there being what David Chalmers and others have called “the
hard problem of consciousness”, the problem of explaining
consciousness in a sense which goes beyond functionalism. The
view that emerges from the above discussion differs somewhat
from both these orthodox positions. Traditional functionalism is
indeed radically incomplete. Nevertheless, understanding what
functionalism leaves out is not quite as “hard” as it is some-
times taken to be. We do already have a mode of explana-
tion-personalistic explanation—which does enable us to explain
and understand features of consciousness, which we experience
directly, and which functionalism leaves out. The really “hard”
problems of consciousness, on this view, are the functionalist
problems: the problems of linking up brain processes described
in neurological terms, and in control or functionalist terms-
terms which we can relate to our personalistic understanding of
ourselves and each other. As these “hard” functionalist prob-
lems are solved, this will undoubtedly have major implications
for personalistic understanding. The result will be the enrich-
ment and improvement of personalistic understanding, not its
replacement by something better (as some proponents of folk
psychology have claimed).
Dennett, D. (1991). Consciousness explained. London: Allen Lane.
Jackson, F. (1982). Epiphenomenal qualia. Philosoph ical Quarterly, 32,
127-136. doi:10.2307/2960077
Jackson, F. (1986). What mary didn’t know. Journal of Philosophy, 83,
291-295. doi:10.2307/2026143
Kirk, R. (1974). Zombies versus materialists. Aristotelian Society, 48
(Supplement), 1 35-15 2.
Lakatos, I. (1976). Proofs and refutations. Cambridge: Cambridge
University Press.
Levin e, J. (198 3). Mat eriali sm and quali a: The exp lan atory gap. Pacific
Philosophical Quarterly, 64 , 354-361.
Lewis, D. (1972). Psychophysical and theoretical identifications. Aus-
tralasian Jo urnal of Philosoph y, 50, 249-258.
doi:10.1080/00048407212341301
Lewis, D. (1990). What experience teaches. In W. Lycan (Ed.), Mind
and cognition (pp. 499-51 9). Oxford: Blackwell.
Locke, J. (1961). An essay concerning human understanding. London: J.
M. Dent & Sons.
Lockwood, M. (1989). Mind, brain and the quantum. Oxford: Black-
well.
Maxwell, N. (1966). Physics and common sense. Bri tish Jo urnal for the
Philosophy of Science, 16, 295-3 1 1. doi:10.1093/bjps/XVI.64.295
Maxwell, N. (1968a). Can there be necessary connections between
succ essive even ts? B ritish Jou rnal for the Philo sophy of Sc ience, 19,
1-25. Reprinted in R. Swinburne (Ed.) (1974), The justification of
induction (pp. 149-174). Oxford: Oxford University Press.
Maxwell, N. (1968b). Understanding sensations. Australasian Journal
of Phil osophy , 46, 127-146. doi:10.1080/00048406812341111
If the unique-matching theory, sketched above, cannot be
shown to be untenable, then it deserves to be taken very seri-
ously, as demonstrating that it is possible to explain brain-mind
correlations22. If the theory turns out to be untenable then the
way forward, in my view, is to take seriously the arguments
designed to show that no scientific, indeed no, explanation of
brain-mind correlations is possible, and search for counter-
examples23. This is the strategy that I have somewhat blindly
put into practice; it has led me to stumble across the
unique-matching idea indicated here. Finally, in tackling the
philosophical mind-body problem, and indeed all philosophical
problems, we should try to come up with solutions which have
fruitful consequences. We should keep before us the great ex-
ample of Darwin, who succeeded in solving a profound phi-
losophical problem—the problem of how purposive living
things can have proliferated in a purposeless universe—in a
way which has had endless fruitful implications for the whole
of biology, and for our understanding of ourselves24.
Maxwell, N. (1984). From knowledge to wisdom: A revolution in the
aims and method s of sc ience. Oxford: Bla ckwell
Maxwell, N. (1985). Methodological problems of neuroscience. In D.
Rose, & V. G. Dobson (Eds.), Models of the visual cortex (pp. 11-21).
Chichester: John Wiley.
Maxwell, N. (1998). The comprehensibility of the universe: A new
conception of science. Oxford: Oxford University Press.
Maxwell, N. (2001). The human world in the physical universe: Con-
sciousness, free will and evolution. Lanham, MD: Rowman and Lit-
tlefield.
Maxwell, N. (2010). Cutting god in half—And putting the pieces to-
geth er again: A new approach to phil os o phy. London: Pentire Press.
Mulhauser, G. (1998). Mind out of matter. Dordrecht: Kluwer.
doi:10.1007/978-94-011-5104-7
Nagel, T. (1974). What is it like to be a bat? The Philo sophi cal Re view ,
83, 43 5- 4 5 0. doi:10.23 07/2183914
Nagel, T. (1986). The view from nowhere. Oxford: Oxford University
Press.
Nemirow, L. (1990). Physicalism and the cognitive role of acquaint-
ance. In W. Lycan (Ed.), Mind and cognit ion (pp. 490-499). Oxford:
Blackwell.
Place, U. T. (1956). Is consciousness a brain process? British Jou rnal of
Psychology, 46, 44-50. doi:10.1111/j.2044-8295.1956.tb00560.x
References Popper, K. R. (1962). The open society and its enemies volume II. Lon -
don: Routledge and Kegan Paul.
Armstrong, D. M. (1968). A materialist theory of the mind. London:
Routledge and Kegan Paul. Putna m, H. (1960 ). Mind s and machin es. In S. Hook (Ed.), Dimensions
of mind (pp. 138 -1 64). London: Collier-Macm illan.
Block, N. (1990). Troubles with functionalism. In W. Lycan (Ed.),
Mind and cognition (pp. 444-468). Oxford: Blackwell. Rey, G. (1997). Contemporary philosophy of mind. Oxford: Blackwell.
Smart, J. J. C. (1963). Philosophy and scientific realism. London:
Routledge and Kegan Paul.
Campbell, K. (1970). Body and mind. Garden City, NY: Doubleday.
Chalmers, D. J. (1996). T he co ns c ious min d. Oxford: Oxford University
Press.
22Science quite generally, I have argued (Maxwell, 1984, 1998), must make
metaphysical assumptions concerning comprehensibility or explainability. A
“science” of conscio usnes s mus t do lik ew ise.
23(Lakatos, 1976) has argued, brilliantly, that one use of a proof is to aid the
search for co unterexamples.
24See (M axw e l l , 2 00 1: p p. 168-179) f or a dev e l opment o f this po int.