Mentality and Modality
John P. Burgess
Department of Philosophy
Princeton University
ii
for Ruby
iii
Contents
Preface
I Descartes and Dualism
II Physicalism: Neuralism versus Functionalism
iv
1
26
III Mental Causation
51
IV Conceivability versus Possibility
76
V Names versus Descriptions and Minds versus Bodies
VI The Bounds of Physicalistic Science
VII Realism versus Pragmatism about Modality
101
126
151
VIII Beyond the Supervenience Issue
176
References
201
Index
iv
Preface
This short book is on a topic familiar to all students of philosophy, and a
subtopic familiar at least to specialists in the philosophy of mind, but it is written
from an uncommon point of view. The topic is the venerable mind-body problem,
now almost four centuries old. The subtopic is the much-discussed thesis of the
supervenience of the mental on the physical, the claim that minds could not have
been different without bodies having been different. The point of view from which
these issues are approached is that of a logician who has come to the question from
modal logic rather than philosophical psychology. The overarching theses of the
work are two, one in the foreground, the other in the background.
First, I argue that there are several reasons to suspect that the supervenience
thesis deserves less attention than it has received, and that the real division of
philosophical importance over mind and body lies elsewhere. It lies in the
difference between those who — whether they call themselves dualists or
physicalists or neither, and whether they oppose or support or are indifferent to the
supervenience claim — recognize that there is something that a physicalistic
account of the world leaves out, and those who are in denial about such omissions.
Second, I try to demonstrate that attention to the historical source of our
problem remains of continuing relevance. Only if we remember that something
was deliberately left out of the conceptual range of physical science when the
subject was launched four hundred years ago can we fully appreciate the fact that it
is still left out today.
To keep the work to a manageable size I have set aside many subtopics with
no more than an oblique reference to external sources. Contrasting views that are
amply and ably represented in the literature I generally no more than summarize,
and not always impartially, in overview. I have also tried to avoid all unnecessary
v
technical jargon. Above all, I have avoided any elaborate apparatus of possible
worlds. I have used such apparatus in some of my work on other topics, but I do
not find it useful here.
This book is a kind of sequel to my Saul Kripke: Puzzles and Mysteries. As
such it shares some of the intellectual debts of that earlier book, but the debts to
other scholars connected with this one are mainly to their writings, rather than to
direct personal contact. The authors whose works have been most important to me
are listed at the end of the first section of the first chapter of the text, and the whole
list need not be repeated here. But I do have some more direct personal debts that it
is a pleasure to acknowledge in this place, to Ned Block, David Chalmers, Daniel
Garber, Stevan Harnad, Frank Jackson, Mark Johnston, Thomas Nagel, Neil
Tennant, Alexander Williams, and doubtless others I am forgetting. To some I owe
inspiring conversations before undertaking this project, to others opportunities to
present work in progress, and to yet others comments at greater or lesser length on
an earlier draft of this work. Besides these figures I should mention in gratitude my
late colleague David Lewis. It was a remark of his, to the effect that all positions in
contemporary philosophy of mind would probably have been considered forms of
materialism by Victorians, that first set me off thinking about these topics.
I
Descartes and Dualism
The problem of the relation of mind to body, or mental to physical, as it is
discussed and debated among philosophers and psychologists today, goes back
some four centuries to René Descartes, at the very beginning of modern philosophy
and modern science. To understand the terms of the problem, we need to take a
backward look at its source.
Mental versus Physical and Possible versus Necessary
The issue as we know it retains from its Cartesian origins a couple of
features that distinguish our mind-body problem from what have been viewed as
similar or related problems in other intellectual and cultural traditions: first, a focus
on the brain as the part of the body most relevant to the mind; and second, the
mathematical character of the science that informs the conception of the physical
that is contrasted with the mental. The topic of this book will be a third feature that
the contemporary mind-body problem shares with that of four hundred years ago,
the curious entanglement of the issue mentality and physicality with the notions of
necessity and possibility.
Descartes’ position as a whole has few if any defenders today, but there are
present-day defenders of two of his main theses: first, that mind and body are a
duo, two distinct things; and second, that there is no necessary connection between
the two. What exactly this latter claim means will emerge only gradually, but at the
outset it may be noted that since the first law of the logic of these notions is that
not necessarily is equivalent to possibly not, denial of the necessity of a connection
is equivalent to affirmation of the possibility of a disconnection.
This circumstance accounts for another curious feature of recent discussion
of mind and body, namely, that much of it is concerned with ghosts and zombies,
2
minds disconnected from bodies and bodies disconnected from minds. Perhaps a
century or so ago, in the days of William James and the Society for Psychical
Research, there may have been philosophers or psychologists who seriously
advocated for the actual existence of such beings; but not so to any noticeable
degree today. The fact that such beings are nonetheless mentioned in the
contemporary literature is indicative of how today, as in the days of Descartes,
discussion still, or again, often concerns possibility rather than actuality.
My aim in this book will be to argue for disentanglement, largely from a
point of view otherwise sympathetic to many Cartesian positions in philosophical
psychology, though not in metaphysics. To make a case for disentanglement is,
however, only my final destination, and I attach less importance to arriving at it
than to journeying towards it, enlarging our understanding of the mental and
physical, and of the necessary and possible, along the way. I know no better place
to begin that journey than with Descartes, and the remainder of this introductory
chapter will be devoted to contrasting Descartes’ position with the kind of
contemporary position that I especially wish to consider.
For reasons of space among others, in making my comparisons I will deal
with a picture of Descartes simplified in places to a caricature; and what I contrast
with him will not be a gallery of realistic portraits of his most distinguished
present-day heirs, either. What I contrast with Descartes will be a simplified
composite sketch, which to be frank will be partly a self-portrait, of a kind of
contemporary philosopher of mind who has been influenced by key works of Ned
Block, David Chalmers, Stevan Harnad, Frank Jackson, Saul Kripke, Joseph
Levine, Colin McGinn, Thomas Nagel, John Searle, and other such thinkers,
without becoming an entirely faithful disciple of any one of them, not that anyone
could be an entirely faithful disciple of more than one.
3
Notational Conventions
In the process of carrying out my comparison, I will gradually narrow my
focus in from the vast topic of the relation of mind and body, which there can be
no hope of treating adequately in a book of modest size, towards a more
manageable subtopic. Various other subtopics will be set aside one by one, each
the subject of a large literature of its own, though I will mention no more in any
case than a bit of the jargon characteristically used in that literature.
These mentions will appear inside double quotation marks (and indeed
inside such “scare-quotes” inside parentheses), where otherwise I use italics, rather
than quotation marks, as in much of the linguistics literature, when I wish to
mention rather than use a word — and dashes rather than parentheses to set off
material from the main text — while I use boldface, as in some mathematics
textbooks, rather than italics, to highlight a word that is going to be used repeatedly
as a term of art when it first appears so used, where I give at least an informal
explanation, if not a rigorous definition, of the special sense of term. Displayed
propositions that will be referred to repeatedly are numbered, but the numbering
begins anew from (1) in each new section of text.
The scare-quoting is not just a stylistic tic, and not just a signal that previous
familiarity on the part of the reader with the expression quoted is not being
assumed. It plays a more important role: if the scare-quoted expression(s) is/are
typed exactly as quoted into the most common search engine, together with
“stanford encyclopedia” as an additional search term, the reader interested in the
issue I am leaving aside will be led at once to the latest version of one or more
generally excellent on-line survey-expository articles — the versions of which
current at the time of this writing are among the references at the end of this
volume — touching on whatever issue is in question, with bibliographies leading
on to yet further material. The reader who carries out this exercise every time a bit
4
of technical jargon is parenthetically mentioned will probably never make it to the
end of this first chapter, but will certainly learn a great deal about the philosophy
of mind. I hope some readers will choose to defer enough of these intellectual
excursions to stay with me to the chapter’s end, for many as are the subissues I will
be setting aside, there will still be many left.
Descartes’ Context and Ours
Though Descartes and the kind of contemporary philosopher I want to
consider are both dualists, holding that the mental and physical are distinct, with
no connection between them necessary, or equivalently, with disconnection
between them possible, and though both are much concerned with modality, the
category to which the notions of necessity and possibility belong, still there are
inevitably many differences between philosophers four hundred years apart. Let
me begin the process of comparison with consideration of the overall intellectual
situation of Descartes on the one hand and of my hypothetical contemporary
dualist, or indeed any present-day philosopher of mind, on the other.
At the outset note must be taken of some major differences as regards the
two features of scientific thought about mind and body already mentioned as
Descartes’ most enduring legacy in this area: the neural or encephalic orientation,
or the emphasis on the nerves or brain, and the mathematicization of physical
science, its concentration on what is quantitative or measurable. The main point is
that while today both are so much taken for granted that it takes some effort for us
even to see them as substantive assumptions, in Descartes’ day both were
contentious novelties.
Descartes’ advocacy of them was part of a comprehensive campaign against
the still powerful and baleful influence of Aristotle and his Scholastic disciples in
physiology, physics, and beyond. In physiology Descartes championed not only
5
concentration on the brain as the bodily organ most relevant to mental life, but also
on the heart as the key organ of a system that circulated fluids around the body. On
both points he had to oppose the authority of Aristotle, who took the heart to be the
primary organ for thinking, and the brain to be an organ for cooling the blood. The
new approach to physics that Descartes favored involved not only
mathematicization but even more importantly avoidance of Scholastic-style
pseudo-explanations of events in terms of supposed purposes or goals
(“Aristotelian teleology”). Such would-be explanations, unsympathetically viewed,
seem to involve attributing human-like motives — and what could be less
mathematically representable than a human-like motive? — to inanimate objects,
as when heavy objects falling towards the center of the earth were said to be
seeking their natural place, as if they were travelers impatient to get home.
Closely related with the main point that Descartes’ ideas in physiology and
physics were new and disputed is the fact that there was very little established
science along the lines he advocated for him to make use of in his philosophizing.
He was thus obliged to fall back on his own resources, carrying out his own
investigations in various domains. It is hardly surprising that, though he scored
some partial successes, among other things with phantom limb pain in amputees
and with the refraction of light, many of his efforts were unsuccessful.
Having correctly focused on the brain, he notoriously then went wrong by
focusing on the pineal gland, the epiphysis cerebri or conarium of anatomists, as
supposedly the crucial part of the brain. Having developed his gloriously
successful application of algebra to geometry, he then went ingloriously wrong in
trying to apply such mathematics to physics; so that today, while Cartesian
coordinates are taught to middle school students the world over, Cartesian vortices
are forgotten by all but historians of the arcana of seventeenth-century
protoscience.
6
As regards the brain, the present-day philosopher of mind faces the opposite
of Descartes’ situation: there is not too little scientific information but too much.
There are so many new neurophysiological and neuropsychological results arriving
so rapidly that it is a large task to try to pick out from the vast quantity of
scientifically interesting material being generated the small fraction that may be
philosophically relevant.
As regards mathematicization, the present-day philosopher of mind is also in
a different situation. Today it goes without saying that physics is expected to be
formulated mathematically, that chemistry is being slowly absorbed by physics,
that biology is being slowly absorbed by chemistry, and so on. The fact that the
original project of mathematicization was controversial because it deliberately
excluded certain factors from consideration may have been lost sight of, as it never
could be for Descartes.
That is an issue to which we will have to return later. For the present, the
next thing that needs noticing is that in proportion as the influence of science was
smaller four centuries ago, the influence of religion was greater. To be sure, a
distinctive domain of philosophy, which unlike theology may make no appeal to
faith, revelation, tradition, authority, or the like, but only to reasoned argument and
the evidence of experience, had been recognized in principle since the middle ages.
But in practice very often argumentation that passed for philosophical was directed
towards offering proofs of conclusions already dictated in advance by theology;
and Descartes’ work on mind and body, as he himself presents it, is a case in point.
Descartes professed a faith according to which the human mind or soul survives
death and exists in a disembodied state pending eventual resurrection and
re-embodiment; and he maintained that this doctrine of survival after death is not
merely true and known by faith, but also something that could and should be
proved philosophically, as part of a defense of that faith.
7
By contrast, contemporary philosophical writers on the mind-body question
virtually never announce overt theological agendas behind their philosophical
theses and arguments. At most there may remain a suspicion in some quarters of
some kind of implicit or covert connection between dualism and theism, and
between antidualism and atheism. Such suspicions, even or especially when they
have no basis in fact, may engender tendencies on the part of the secularly oriented
to be more easily struck by weaknesses in the arguments of the dualist side than by
weakness in the counterarguments of the antidualist side, and to engender opposite
tendencies on the part of those oppositely oriented; but all this remains deep in the
background, and seldom openly discussed.
Zombies, the Zombie World, and Supervenience
Now despite his claim that human survival after death can be proved
philosophically, in his first and best-known philosophical treatise, the Meditations
on First Philosophy, what Descartes offers is not a purported philosophical proof
that the human mind actually does survive death, but only that it possibly could.
This is precisely where modality comes in for him. Admittedly, the fact that his
conclusion is a modal one is not conspicuously advertised.
On the contrary, the publisher’s title page to the first, Latin, edition
erroneously claims that the book offers a proof of the immortality of the soul,
though the text contains nothing that even purports to be such a proof. And the
corrected title page to the second, French, edition claims only that the book
contains a proof of the real distinction between the human soul and body, without
mentioning that the possibility of immortality is supposed to follow (that being
implicit in how Descartes distinguishes a “real distinction” from a “rational
distinction”). But conspicuously advertised or not, the modal aspect is there.
This brings us to a crucial difference between seventeenth-century and
8
twenty-first-century dualism in the nature of their modal claims. The possibility
that concerned Descartes was, in line with his religious commitments, that of
disembodied souls, while the contemporary dualist is more concerned with the
possibility of the opposite combination, soulless bodies. Other differences and
asymmetries are connected with this one. Descartes is interested in disembodied
souls and the possibility of an afterlife for their own sake, so to speak.
Contemporary dualists seem to be interested in soulless bodies mainly for the sake
of what their alleged possibility tells us about the nature of physical science and
the prospects for mental science.
Above all, Descartes claimed that minds could exist without bodies, but did
not claim that minds could exist just as they are, with all their mental properties, in
the absence of bodies. Rather, he held that some features of mental life were more
independent of embodiment than others. By contrast, contemporary dualism does
maintain not merely that bodies could exist without minds, but that bodies could
exist just as they are, with all their physical properties, without minds. For while
any claim at all about the existence of bodiless minds is a substantive one, a mere
claim of the existence of mindless bodies is a triviality, since there indisputably do
exist some bodies without minds, namely, those we call corpses or cadavers. There
is a broader sense of body in which we speak, for instance, of heavenly bodies and
fruiting bodies, and taken in this broader sense, meteors, which never were alive,
and mushrooms, which still are alive, join mummies, which once were but no
longer are alive, as examples of mindless bodies. What the contemporary dualist
maintains is that, though there do not actually exist any, there possibly could have
existed mindless bodies that, unlike mummies, let alone meteors or mushrooms,
moved around just as we do when we act in pursuit of our preferences and desires,
and emitted sounds just as we do when we speak in expression of our opinions and
beliefs.
9
Such are zombies, in the philosophical sense popularized by Chalmers, as
opposed to a folkloristic or horror-movie sense. The contemporary dualist indeed
maintains that there could have existed nothing but zombies, so to speak, that the
world could have been physically just as it actually is, without there being any
mental activity at all in it.
Philosophers nowadays often, instead of saying the world could possibly
have been thus-and-so, will say there is a possible world that is thus-and-so, and
instead of saying the world is actually thus-and-so, will say the actual world is
thus-and-so. This possible world talk was popularized by Kripke, who later
retreated from it somewhat, as having a potential to create confusions if taken too
seriously. Like Kripke, I will use it here only as a picturesque manner of speaking.
So speaking, the dualist hypothesis is that of a zombie world, a possible world that
is just like the actual world physically, but is devoid of mental life.
The characteristic contemporary antidualist claim is one that strongly
contradicts this hypothesis: the thesis of the so-called supervenience of the mental
on the physical. Here X is said to supervene on Y if some difference in Y would
be necessary for any difference in X to be possible. To make somewhat plainer
how necessary and possible are being understood here (pending future elaboration
on “metaphysical modality”), to claim supervenience of the mental on the physical
amounts to claiming that the mental could not have different from how it is,
without the physical having been in some way different from how it is; or
equivalently, that in order for the mental to have been different, the physical would
have to have been different. Or at least, this will do for the moment.
Being wholly absent is the most extreme way of being different, so the
supervenience thesis is stronger than denial of the zombie world hypothesis,
though generally those who deny that hypothesis do so by subscribing to
supervenience. I said earlier that the topic of this book will be the entanglement of
10
the mental with the modal; I can now say more specifically that the topic will be
the supervenience thesis.
Kinds of Minds
Further differences between Cartesian and contemporary thinking on the
mind-body problem pertain to what portion of that which I have heretofore been
nebulously calling the mental is of central concern. First under this heading comes
the question of whose mental life is at issue. A half-dozen types at least of real or
imaginary minds have been considered over the past four centuries, with
considerable disagreement as to which are real and which are imaginary:
human
living, embodied human beings
ghostly
deceased, disembodied ex-human beings
angelic
including demonic, good or evil spirits never embodied
animal
nonhuman terrestrial species, incapable of rational speech
robotic
artificially intelligent machines
alien
naturally intelligent extraterrestrials
Descartes, with his religious commitments, recognized human, ghostly, and
angelic minds. He surprises, even shocks, present-day readers by denying that mice
or even monkeys have mental lives in anything like the way human beings do. For
him, an ability to engage in rational discourse, to speak in a more serious sense
than that in which mynahs can speak, is the chief note of having a mind. He had no
hope that the mechanics or physics he was beginning to develop would ever be
able to explain rational speech, but he did believe that it would one day be able to
explain all the motions of animal bodies, and it is in this sense that he took animals
to be mere mechanisms or machines.
11
Descartes’ taking animals to be mindless was so closely bound up with his
taking them to be machines that there would be no room in his system for robotic
minds; inversely, his taking human beings to have minds was so closely bound up
with their ability to speak that he would have had to made room for alien minds,
had any language-using creatures from other planets shown up. They did so in
fiction a few years after Descartes’ death in Cyrano de Bergerac’s Comical History
of the States and Empires of the Moon.
Contemporary philosophers of mind typically have little to say about ghostly
or angelic minds, but are virtually unanimous in being more accepting than
Descartes of at least some animal minds, though there are, right at the time of this
writing, very active controversies over the nature of the mental lives if any of fish,
and in particular over whether they can feel pain. Being the product of an era with
a developed science-fiction genre literature, contemporary writing on the mindbody problem includes at least occasional mention of alien minds, often called
Martian minds, and sometimes heated discussion of the nearer-term or longer-term
future prospects for robotic minds.
Speculations on this latter topic (“strong AI [artificial intelligence]”), on
whether robotic minds are in principle possible and perhaps even in practice
feasible to create as technology advances, are a point of contention between certain
confidently antidualist philosophers and others who are dualists, or if not outright
dualists, antidualists with reservations. The former tend to argue the positive, and
the latter the negative (most famously Searle with his “Chinese room argument”
and Block with his “blockhead argument”).
For virtually all philosophers of mind from Descartes onwards, however,
whatever other kinds of minds they do or do not admit, the primary concern is with
the minds of living, embodied human beings. And here, as a major step towards
cutting down the vast issue of mind and body to more manageable size, I will
12
henceforth restrict attention to such minds, apart from some occasional asides. In
so doing I will be leaving out of account not only the vast literature on the debate
about the prospects for robotic minds (“can a machine think?”), but also some
annoying distractions over the proper formulation of the supervenience thesis.
The Proliferation of Varieties of Supervenience
For some philosophers who deny that there could have been zombies grant
that there could have been ghosts or angels, even or especially ones whose
presence had no effect whatsoever on the physical world, who were mere
watchers, as one might call them. And if the world had contained such watchers,
but had otherwise been as close as their presence would allow to being as it
actually is, then in one sense the mental would have been different, since there
would have existed more minds, those of the said watchers, but the physical would
have been the same, since watchers by hypothesis have no effect upon it; and so
the supervenience thesis in my initial rough formulation would be violated.
In the literature there has been no lack of discussion of candidate alternative
formulations. But I mention this matter only to dismiss it: since I am considering
only the minds of living, embodied human beings, for me the mental would have
been different only if such human minds had been different, and thus I can stick
with my original, rough formulation of supervenience, and will.
While I am at it, I should acknowledge that there has been what my late
colleague David Lewis aptly called an unlovely proliferation of versions of
supervenience. About this topic I will only mention that the formulation given
earlier, in order for the mental to have been different, the physical would
necessarily have to have been different, was a so-called global claim, pertaining to
the mental and the physical each as a whole; but most who adhere to it adhere also
to a so-called local claim, pertaining to the mental and physical subject by subject,
13
namely, the claim that in order for any aspect of the mental life of any given human
being to have been different, some physical aspects of his or her brain would have
to have been different.
To say more about the proliferation issue (for instance, about why not all
who accept “global supervenience” accept “local supervenience”, or about a
distinction between “weak supervenience” and “strong supervenience” that cuts
across the global-local distinction) would involve us in on-going disputes (for
instance, over so-called “internalism” versus “externalism” in semantics, and over
the apparatus of “possible worlds”) in the mazes of whose sprawling literatures we
might soon become irretrievably lost. I will therefore, keeping my eyes on the
prize, resolutely ignore such potentially distracting issues.
Thinking or Consciousness
So much for whose minds are in question. Other issues of focus remain,
since even restricting attention to human minds, there remain questions about
which aspects of mentality are to be considered central. Here there seems to be
little difference in the range of mental states and processes that are in principle of
concern to Descartes and the range of concern to contemporary dualists. Yet there
are differences over terminology for distinguishing that range, over the occurrence
and importance of mental states and process outside that range, and over the
emphasis in practice within that range. Let me take these up in turn.
First, with Descartes the basic term for the mind is thinking thing, or its
Latin or French equivalent, res cogitans or chose qui pense, and his basic term for
mental activity is thinking. But he is emphatically not restricting his attention
exclusively to thinking in its narrowest colloquial sense. On the contrary, he
explicitly counts as forms of so-called thinking not only affirming, denying, and
doubting, but also willing, remembering, imagining, sensing, and more. The
14
present-day dualist tends to speak rather of conscious states, where the most
popular characterization of consciousness among dualists today seems to be the
one offered years ago by Nagel: a state is conscious if there is something that it is
like to be in that state. This Harnad would amend to there is something that it feels
like to be in that state. I will use both formulas. A dualist’s list of conscious states
can be expected to agree roughly with Descartes’ list of thinking states.
Second, a major difference is that, unlike Descartes, today’s dualists per
force recognize that the mental in the broadest sense, as studied by psychology in
the broadest sense — represented, for instance, by the American Psychological
Association’s Encyclopedia of Psychology — has many aspects whose connection
with conscious experience is only indirect or simply nonexistent. The mind-body
problem, however, as we have inherited it from Descartes, is about the relation
specifically of the conscious mind to the body, and dualism being a position on the
mind-body problem, the dualist qua dualist, so to speak, is concerned with
conscious states, whereas the dualist qua psychologist or philosopher may be
interested in much more.
Third, a lesser difference is that for Descartes judging and willing were
supposed to be more independent of, and imagining and sensing more dependent
on embodiment; and though he had a genuine interest in other aspects of mentality,
to the extent of writing at length about vision, and composing a whole treatise on
the emotions, or passions of the soul as he called them, still the aspects least
dependent on embodiment, affirming, denying, doubting, and willing, were in
some sense for him the most quintessentially mental. By contrast, while a presentday dualist is in principle interested in all conscious phenomena, in practice a
dualist may tend to emphasize a certain subrange of the full range as the most
vividly conscious, and the list of them might roughly agree with Descartes’ list of
embodiment-dependent states, especially sensations and affects.
15
Consciousness and Self-Consciousness
In these connections, potential sources of confusion lurk in the literature. A
first is simply that understandings of the term conscious incompatible with the
dualist’s are in circulation. It should not be surprising that such conflicts of usage
arise. It is a commonplace that since the bulk of our language is directed towards
communication about the shared, external world, whenever we try to speak about
the private, internal world, try as we might, we are liable to be misunderstood. And
introduction of technical jargon to explain what one means often is
counterproductive, only providing more opportunities for misunderstanding.
On the formula I have adopted, to be conscious it is enough, even perhaps
more than enough, to be aware of something, while to be self-conscious, as I will
say by contrast, it is needful to be aware of one’s self or of one’s awareness. In
other terminology, what I am calling consciousness may be said to require only
sentience and not introspection; but as I have already said, technical jargon is
easily misunderstood. In principle the so-called second-order states of awarenessof-awareness are included along with the so-called first-order states of awareness
in the dualist’s class of conscious states; in practice, dualist attention may tend to
be focused on the first-order states as most vividly conscious. A deviant usage, by
contrast, counts only second-order and not first-order states as conscious.
A good example of this usage is provided by Julian Jaynes’s book The
Origin of Consciousness in the Breakdown of the Bicameral Mind. Jaynes makes
two spectacular claims. First, when Homer and his contemporaries faced a
decision, they did not think about it, but just heard voices in their head, coming
from the right side of the brain but interpreted as gods or goddesses, telling them
what to do. Second, Homer and his contemporaries were not conscious, because
consciousness arose only when the voices ceased. Most psychologists and
16
historians will boggle at the assertion that Homer and his contemporaries were
subject to constant auditory hallucinations, but the dualist will boggle even more at
Jaynes’s second assertion, that Homer and his contemporaries were not conscious.
If Homer was having auditory hallucinations, then he was having sensations; and if
he was having sensations, then he was conscious, as the dualist understands
consciousness. Evidently — and further reading in his book amply confirms this
interpretation — Jaynes is using conscious for self-conscious. I will not.
Occurrent versus Dispositional Belief
A second and more serious potential source of confusion is an ambiguity in
words like believe. If I say that Bill believes that Hill will be the next president on
the grounds that right this moment he is saying to himself, as the expression goes,
that Hill will be the next president — though not out loud — then I am speaking of
an occurrent belief. There is something that it is like to be saying to oneself Hill
will be the next president, and so occurrent beliefs are conscious. When Descartes,
at the opening of his Third Meditation, says that thinking includes among other
things affirming, denying, and doubting, he surely is speaking of conscious acts of
judging something true, judging something false, and suspending judgment, and
implying occurrent belief, occurrent disbelief, and occurrent doubt.
By contrast, when writers today speak of believing, disbelieving, or
doubting, they generally mean not acts but standing states. If I say that Bill
believes that Hill will be the next president on the grounds that he habitually says
such things to himself, even though right this moment he is deep in dreamless
sleep, I am speaking of a dispositional belief. A similar distinction between
occurrent and dispositional can be made for desire, and indeed believing and
desiring are only the two most prominent examples of a whole family of related
notions (discussed in philosophy of language under the rubric “propositional
17
attitudes”) for which one can make similar distinctions.
Now deep in dreamless sleep there is no feeling: there is nothing that it feels
like to be in this state. And so, given that one can have dispositional beliefs while
in this state, dispositional beliefs cannot properly be called conscious states. Still,
to have a dispositional belief is to have a disposition to have occurrent beliefs, and
hence a disposition to have certain conscious states. An entity entirely incapable of
having conscious states, such as a thermostat, to give one common example, is not
capable of having dispositional beliefs any more than occurrent ones.
Nevertheless, we sometimes use the language of belief in speaking of
gadgets like thermostats. If I ask why, in winter, when there is a fire in the hearth
in the living room, the bedrooms become cold, I may be told that the thermostat
turns down the furnace because, having been installed too near the hearth, it
believes the house is overheated. It is easy to fall thus into speaking of gadgets as
having beliefs, when they operate in certain respects as if they were conscious
subjects with beliefs (in jargon, this is the “intentional stance”). But note the as if
here, marking this usage as metaphorical rather than literal. The kind of dualism in
which I am primarily interested, which I will distinctively call hardcore dualism,
recognizes no literal notion of belief beyond occurrent and dispositional.
It seems to be an occupational hazard — or what the French call a
déformation professionelle — of those who habitually work with electronic
gadgets more sophisticated than thermostats, and of computer scientists in
particular, to become so habituated to applying belief talk metaphorically to
gadgets that they lose their grip on the literal meaning of belief. Or at any rate, that
is how the hardcore dualist would diagnose their situation. Some present-day
professedly dualistically-inclined writers, whom I will call softcore, make more
concessions on this matter than my hardcore dualist would, and may grant the
occurrence of so-called unconscious beliefs not merely among human beings, but
18
among electronic devices.
Belief and Desire versus Sensation and Affect
Virtually all contemporary dualists, I think, would hold that, whether or not
beliefs are examples of conscious states, they are far from being the best, or most
vivid examples. The clearest cases are provided by sensations on the one hand and
affects or moods on the other: there is very clearly something it is like to see a light
or hear a tone or smell a scent or feel a texture; there is clearly something it feels
like to be elated or depressed or anxious. These are the cases on which the dualist
typically would like to concentrate.
Indeed the dualist may by preference wish to concentrate on what may be
called raw sensation and affect, rather than sensory or affective states (states of
“perception” or “emotion”) that involve a lot of intellectual content: on seeing a
bright red light rather than on noticing that the traffic-signal has switched to stop,
on feeling free-floating anxiety rather than on worrying that one will be late for the
meeting. Often in the literature the focus is specifically on visual sensation, or
more specifically on color vision, or most specifically of all on visual experience
of the color red (as in the “knowledge argument” of Jackson).
I can frame a paradigm of the mental for myself in the following way. I
place on the table before me a ripe tomato. I look at it concentrating on the color
rather than the tomato, as a painter working on a Still Life with Love-Apple might,
trying to decide what pigments to put on the canvas, rather than as a horticulturalist
might, trying to identify the particular cultivar of which the tomato is a specimen. I
say to myself, with the object in normal lighting conditions and myself in my
normal state of health, to me an object that is red looks like THIS. The foregoing
long, italicized formulation I will abbreviate to red looks like THIS, and I will
sometimes say that what is paradigmatic is what colors look like.
19
The abbreviation leaves out the first-person pronoun me, and the restriction
to normal lighting and health, but retains what is most crucial of all, the
demonstrative pronoun THIS. Needless to say, the omitted elements are important,
and will need to be tacitly understood: an object that is red may look different to
me when presented in strongly yellow light or when I am suffering from jaundice,
and your paradigm of the mental will be of how red looks to you, while mine is of
how red looks to me, and one of us may have anomalous color vision.
If I have any readers who lack color vision altogether, they can frame
paradigms for themselves by substituting a different sensory channel: the smell of
ammonia is the traditional choice, and the taste of a ripe strawberry a more
pleasant alternative. I cannot promise in what follows to consider only examples as
near to totally raw as these, but I will take another step towards restricting our
subject matter by restricting attention primarily to sensations and secondarily to
affects, to the exclusion of beliefs or desires.
In discussing the supervenience thesis, then, the main point will be that if
there were a zombie physical duplicate of myself, it would emit the same sounds
that I do when I say THIS is what red looks like, but as a matter of fact red would
look like nothing to a zombie, there being nothing that it is like to be a zombie
looking at a ripe tomato. And the issue is whether the existence of such zombies, or
a whole world of them, is in principle possible.
Substance and Other Categories
So much for what aspects of human mental life are at issue. Before leaving
the topic of focus, there is a more abstract difference between Cartesian and
contemporary dualism to be noted. Looking back on Descartes today, we see by
hindsight that his escape from mediæval ways of thinking, and from Aristotle, was
inevitably only partial. In particular, a further contrast between Descartes and
20
ourselves is that he operates, as philosophers generally did in the seventeenth
century, with a notion of substance that to us must seem quasi-scholastic, since it
involves so much antique claptrappery (in particular, a contrast between “essence”
and “accident”, and another between “potency” and ”act”).
Descartes’ thesis was that he has a sufficiently clear and distinct conception
of mind and of body to recognize that these are distinct substances in his
seventeenth-century, quasi-scholastic sense; and the possibility of the existence of
one without the other was for him closely related to the assumption that where
substances in such a sense are of distinct natures, it lies within the power of God to
create or preserve one without the other (something that might not hold for
“modes” in place of “substances”).
But the notion of substance with which Descartes operated is now obsolete.
The usage of substance among many Anglophone philosophers today scarcely
differs from the usage of substance among nonphilosophers. It generally just
means a kind of stuff as opposed to a kind of thing, the denotation of what is called
a mass noun, such as water or gold, as opposed to a count noun, such as ice cube or
wedding ring. Substance in this sense takes up space, which is just what mind does
not, according to Descartes, who was therefore no believer in mental substance in a
present-day sense of substance. There have been thinkers who have believed in
mental substance in such a sense, but Descartes was not among them. They have,
rather, included those traditional thinkers, going back to antiquity if not to
prehistory, but opposed by Descartes, who took spirit or soul to be literally a kind
of breath or gas — and later, in the Jamesian era, those whose will-to-believe
allowed them to be duped by spirit mediums who pretended at their séances to
produce so-called ectoplasm, the quintessential example of a mental substance.
To say all this is not to suggest that there is any one single category that all
participants in present-day discussions are agreed should replace that of substance
21
in any sense. On the contrary, it is a conspicuous feature of the current literature in
philosophy of mind that different parts of it are focused on different categories:
parts of it on mental versus physical properties, parts of it on mental versus
physical states, parts of it on mental versus physical processes, parts of it on
mental versus physical events, and so on. Moreover, often the key word, property
or state or process or event or whatever, may be understood in a somewhat
artificial, specialized, or technical sense: its usage may be subjected to
regimentation, as is said; and moreover usage may be regimented in different
ways by different writers. Unfortunately some, besides regimenting existing terms,
have coined new ones (notably a grotesque philosophical usage of “trope”,
unrelated to the traditional rhetorical notion).
Comparing the different parts of the literature, we see a contrast between
two oppositions: first, an opposition between the standing or static and the
changing or dynamic, and second, cutting across it, and overall more important, an
opposition between universal types and particular tokens. Thus what are set
against each other as mental and physical may be sometimes static types, as with
being angry and having excess bile, sometimes dynamic types, as with becoming
angry and getting excess bile, sometimes static tokens, as with my being angry now
and my having excess bile now, sometimes dynamic tokens, as with my becoming
angry now and my getting excess bile now. I myself will adopt — indeed, have
already adopted, in speaking earlier of belief as a state — a chameleon-like policy
of following, so far as I am able, the preferred terminology of whatever part of the
literature I am discussing or alluding to at any given point.
The Correlation Principle
A further range of points of difference between Cartesian and contemporary
dualism pertains to what kinds of lawlike regularities or laws of nature are taken to
22
govern the relationship of the mind to the brain and the brain to the cosmos
(whatever exactly “lawlike” and “law of nature” mean).
To begin with mind and brain, whatever Descartes may have held about the
issue, nowadays it seems widely accepted, even by dualists, though not all are
explicit on the question, that every mental item of the kind we still have under
consideration, every kind of human sensation or affect, has a physical correlate,
with which it is associated in a lawlike way, and indeed a neural and in most cases
encephalic correlate, located somewhere in the nerves and in most cases in the
brain. This correlation principle seems a reasonable presumption, even in the
absence of any total success in exactly pinning down the location and nature of the
neural item associated with a given sensory or affectual item, by the many partial
successes in approximately pinning down such associations. I will accordingly take
the correlation principle for granted here, and include it in my conception of
hardcore dualism. It is to be understood that no claim is being made to the effect
that we already have in our current lay or technical language, in advance of future
neurophysiological and phenomenological work, all the mentalistic or physicalistic
vocabulary that would be needed for the ultimate statement of the relevant laws.
This may be the best point to mention that, for the sake of example, it is
often pretended in philosophical discussions that we have after all exactly pinned
down the location and nature of the associated neural process in at least one
important case, that of pain, and that in this particular case the correlate is the
firing of nerve-fibers of the C-bundle. This is pure pretence, among other reasons
because sharp so-called first pains as contrasted with lasting so-called second
pains or aches are associated with the firing less of C-fibers than of A-fibers, and
because the firing of some kinds of C-fibers, such as the histamine-selective, is
associated not with pain but with other sensations, such as itching; and anyhow, all
the fiber-firing in the world will not be accompanied by pain unless the right brain
23
structures are present and in working order. But for illustrative purposes it will be
convenient if I, too, sometimes below follow tradition, and pretend that the neural
correlate of pain is known to be C-fiber firing.
It will be noted that by taking the dualist view to include both the zombie
world hypothesis and the correlation principle, I am taking the dualist view to
imply that though the nerves and brain might possibly have existed just as they are
without conscious experience occurring at all, for the mental to be different in this
or any other way from how it actually is, without the nerves and brain being
different, would involve a violation of the laws of nature: a zombie world would
exhibit the same purely physical laws as the actual world, but not the same
psychophysical correlation laws, and so not the same total system of laws of
nature. Thus I am taking the dualist view to presuppose that the laws of nature
could have been different from what they actually are, and taking the
supervenience issue to be whether the laws of nature could have been different in a
certain specific way from what they actually are.
The Uniformity Principle
Turning to brain and cosmos, despite what was held by Descartes, nowadays
it also seems widely accepted, again even by dualists, though again not all are
explicit on the question, that matter and energy are everywhere subject to the same
physical laws, even in systems of which a human body or human brain or some
part of one is one of the components. Where the laws are statistical rather than
deterministic, this would mean that they apply inside and outside the body and
brain with no change in the statistics.
This uniformity principle seems a reasonable presumption in the face of the
failure of decades of work in parapsychology to come up with any reliably
reproducible results demonstrating transcendence through the mental of ordinary
24
limitations on the physical. No credible case has been made for, say, the ability to
transmit something from one brain to another without any physical signaling, or
the ability by thinking about them to alter the probabilities of random quantum
events. This last was the object of extended efforts at my own university’s
now-defunct Engineering Anomalies Research Laboratory. I share the sense of
some of my colleagues that the extended operation of this laboratory was an
embarrassment, and I will take the uniformity principle for granted here.
Descartes’ position was incompatible with uniformity: on the Cartesian
view, laws applicable to matter outside the pineal gland may break down in
application to the fluids supposed to be inside it. Notably, Cartesianism cannot
accept the uniform operation, across the whole material world, of the law of
conservation of momentum, the conservation of the directed quantity mass times
velocity, though there is a related weaker law that may be uniformly operative, the
conservation of the undirected quantity mass times speed. Conservation of
momentum may operate throughout the mineral and vegetable kingdoms, and
among the so-called lower animals, the applicability to nonhuman animal bodies of
the very same laws as apply to dead matter being essentially what Descartes means
by calling animals machines. Conservation of momentum may operate as well in
the human heart and liver and kidney and spleen, which other traditions and
cultures may have seen as the bodily seat of mind, and even in the human brain
outside the pineal gland. But it does not operate on the fluid there, since the
direction of motion of that fluid can be changed by the action of the immaterial
mind. To be sure, the further changes that then follow, for instance in the motions
of the tongue and lips in rational speech, need require no further violations of
conservation of momentum.
Or so the story goes. Descartes scholars will be quick to point out that the
foregoing cartoon, inspired by G. W. Leibniz’ impression of Cartesianism, is not
25
entirely accurate, and not just because Descartes lacked the Newtonian concept of
mass; but it will do for present purposes.
Parapsychology
Fifty years or so ago, when parapsychology had such distinguished
defenders among philosophers as C. D. Broad, it stood in higher repute in
philosophical circles than it seems to do today, and there was genuine
disagreement over its status. So long as there was a division among philosophers
on this issue, that disagreement was probably the most important philosophical
division over the mind-body problem. It was not the division between dualism and
antidualism: the division line ran through the dualist side. (It more or less
coincided with the distinction between “interactionist” dualism and its opposite.)
I am taking it that belief in the interference of immaterial minds with the
physical world — I am not speaking here of sporadic and miraculous interventions,
whose occurrence can never be disproved, the burden of proof being rather on the
other side, but of regular and lawful interference, such as would have to be taken
account of in a science of mind — is no longer what James called a live option for
present-day philosophers, as it was for James himself. I am assuming the twentyfirst century dualist views of interest stand on the opposite side of the line from
where Broad, and before him James, and before him Descartes stood. I am
supposing that the side on which these older thinkers stood has by now been pretty
much abandoned by philosophers and left to popular culture, which even today is
still very much taken with mind over matter.
If so, then today the most important remaining division among philosophers
must run elsewhere, and presumably must be either the division between dualism
and antidualism, or else some division whose dividing line runs through the
antidualist side. We must next look at what views there are over on that side.
26
II
Physicalism: Neuralism versus Functionalism
Descartes’ dualist view has traditionally been contrasted with what were
called idealism, which denied the existence of matter, and materialism, which
denied the existence of anything else. The views contrasting with and rivaling
contemporary dualism are rather different.
Isms in the Philosophy of Mind
Today, idealism is extinct, and the label materialist has largely been
superseded by physicalist, the most respectable reason for the switch being that
energy as well as matter, and influence through force fields rather than by direct
contact, are components of our enlarged present-day conception of the physical.
But today there are many views that are not direct descendants of any seventeenthcentury option.
Now the dualist thesis is a conjunction: the mental and physical are distinct,
and there is no necessary connection between them. The denial of this thesis is
therefore equivalent to a disjunction: either the mental and physical are not
distinct, or else the mental and physical are distinct but necessarily connected.
Antidualist views are correspondingly of two kinds, one maintaining the first
disjunct, and specifically that the mental is included in the physical, the other
maintaining the second, and specifically that the mental is distinct from but
supervenient on the physical.
To keep the length of my discussion within feasible bounds, I will give
extended consideration to just two antidualist views, one of each kind. I believe
that the two I plan to consider have been the two most popular views among
philosophers of mind over the past several decades — though this belief is
admittedly only an impression, and is not claimed to be based on conducting any
27
scientifically-designed opinion polls — and I believe further that much of what I
have to say about them would apply mutatis mutandis to many other views.
But no book of modest size can hope to examine every view that is out there,
and I simply will be taking no note of certain views, some of which were popular
within living memory but seem by now to have lost much of their following, and
others of which have some vigorous present-day sympathizers, though as yet not
many. (Such views would include among others the once-fashionable “anomalous
monism”, according to which my being angry now may be my having an excess of
bile now, while your being angry then was your having an excess of phlegm then,
and the speculative “panpsychism”, according to which even electrons are
conscious in some low-grade way.)
Since I will be considering just two views directly, I will keep my labels for
them simple. I will call them both forms of physicalism, distinguishing the one as
neuralism and the other as functionalism. The terminological situation in the
literature is more complex. (This is partly because of a lack of agreement as to
what it takes to qualify as “physicalism”: must one hold that everything mental is
identical to something physical, or is it enough to hold to that everything mental
supervenes on something physical?)
Neuralism is arguably the only view that can be called physicalist by right,
since it is the only view that holds everything mental to be physical: every mental
property or state is a physical property or state, every mental process or event is a
physical process or event, every mental type is a physical type, every mental token
is a physical token. What I call neuralism is only one version of what is called the
identity theory in the literature (one version of the central-state identity theory
a.k.a. the “mind-brain identity theory”, namely, the version distinctively called the
“type-type identity theory”). It can be easily described: it is the view that each
mental item is identical with — is quite literally one and the very same thing as —
28
its neural correlate.
I use the wordy phrase is quite literally one and the very same thing as here
rather than the simple is of identity, which to me means exactly the same, because
of the tendency of some writers to use the weasel-phrase is nothing over and
above. My interest is solely in whether a mental item is or is not its neural
correlate, and not in whether, supposing them to be distinct, the mental should be
described as being over and above, or besides and alongside, or perhaps under and
below the physical, whatever any of those spatial metaphors means, if anything.
Terminology aside, the identification immediately implies the supervenience
thesis: if every mental item is a physical item, then no mental item could have been
other than as it actually is without some physical item — namely, itself — having
been other than as it actually is.
Functionalism: Roles and Realizers
Functionalism I am calling physicalist by courtesy, since I take it most of its
present-day adherents consider the label an honorific and want it applied to
themselves. A generation ago, many adherents of functionalism would have
insisted their view was neither materialist nor dualist; today many insist that it is a
form of physicalism (and more specifically, of “nonreductive physicalism”, a
phrase they do not consider to be an oxymoron).
What I call functionalism is only one version of functionalism in the
literature (to which distinctive version both the distinctive labels of apriori
“analytic functionalism” and of causal “role functionalism” would apply, the latter
in contrast to “realizer functionalism” found, for instance, in Lewis, which
amounts to a special variety of or line of argument for neuralism). The view is not
so easily described.
It is often introduced by way of an analogy with computing, where a single
29
piece of software or program, say for addition, is capable of being implemented on
different kinds of hardware or computers, such as mechanical and electronic
adders. An abstract algorithmic process or role such as carrying one from the tens
to the hundreds column might be implemented by different concrete physical
processes or realizations in the two cases, describable as the turning of a certain
gear of the mechanical adder in one case and the transmission of an electronic
impulse between certain locations in the electronic adder in another.
But this kind of diversity of implementation (in jargon, “multiple
realizability”) is not really the crucial point for present purposes, but rather simply
the distinction between the implemented and its implementation. Even if we are
considering only just one single implementation of the operation of addition, say
the ringing up of some customer’s grocery bill on an old-fashioned mechanical
cash register, there still nonetheless remains an important distinction between the
role or abstract software process of one-carrying and its realization by the concrete
hardware process of gear-turning. The difference is as great as the difference, even
if we are speaking of just one single performance of just one single production,
between the role of Hamlet and the actor who plays it.
One-carrying and gear-turning are things of entirely different sorts, though
the one is supervenient upon the other, in the sense that there cannot be a
difference as to whether the machine is or is not carrying one without there being a
difference as to whether its gear is or is not turning. The one clear point about
functionalism in philosophy of mind is that it is a view according to which the
relationship of a mental process or state to the neural process or state correlated
with it is supposed to be analogous to the relationship of a software process or state
to a hardware process or state implementing it: it is supposed to be a relation of
role to realizer, and thus is supposed to involve supervenience without identity.
The less clear point is just how the analogy is supposed to go, but for present
30
purposes I will not need to enter deeply into that large and complex topic here. (I
defer to Janet Levin, whose entry on “functionalism” in the Stanford Encyclopedia
of Philosophy presents the view clearly and concisely and with more sympathy
than I am able to muster myself.)
Let me just emphasize that it is indeed today generally a question here of an
analogy, and not of an attempt at a direct, literal application of the softwarehardware distinction to the mind-brain or mental-encephalic distinction, as in early
work of Hilary Putnam, for one. There are several obstacles such to direct, literal
application of the distinction, including one much emphasized by Kripke: whereas
computers are artifacts designed to implement algorithms, so that the design
provides a criterion of what physical process realizes what role, by contrast the
brain is a piece of meat, open to many different attempts to construe it as an
implementation of an algorithm, with no clear criterion of rightness for such
construals. (And if not for this reason then for others, what seems to be the most
widely-held version of functionalism in philosophy of mind today is not “machinestate functionalism”, which Putnam himself eventually came to oppose.)
The Case of Pain
The differences among dualism or antiphysicalism, in the contemporary as
opposed to Cartesian form in which I am considering it, and the two antidualist or
physicalist views, neuralism and functionalism, may be and often are illustrated by
the case of pain, or better, the state we call being in pain.
The dualist view is that to be in pain is to feel a certain way, to have a
certain feeling. What feeling that is, anyone who has ever been in pain, and heard
the word pain, or the equivalent in some other language, applied to his or her
condition at the time, will know from personal experience. Inversely, it is only
from personal experience that one learns what it is like to be in pain. Or so the
31
dualist will insist.
For those who have never had such personal experience — including it is
claimed some 40% of the population of the village of Vittangi in Sweden, who are
said to exhibit the condition known as congenital analgesia, and in fiction one of
the villains in the thrillers of Stieg Larsson — nothing we could tell them about it
would enable them to imagine it, and they cannot know what it is like to be in pain.
Or again, that is what the dualist will insist.
The neuralist view is that to be in pain is quite literally one and the very
same thing as its neural correlate, which we are pretending is to have C-fibers that
are firing. It has to be admitted that there is this difference between the neural item
called pain and the neural item called a brain tumor, that while the way I would
come to know that I had a brain tumor would involve some medical specialist
examining my nervous system or a scan thereof, the way I ordinarily come to know
when I am in pain requires neither the services of a medical specialist nor any
specialist equipment for use in examination of my nerves or brain.
Come to think of it, however, there presumably is a less ordinary way I
could come to know my C-fibers are firing that would after all involve some
qualified person examining or testing my nerves. The terms pain and C-fiber firing
refer to the same item, according to the neuralist, but the two terms are associated
with two different routes to knowledge of that item. (This fact about the route to
knowledge is supposed to be characteristic of “phenomenal concepts” like pain in
contrast to ordinary neurological concepts like brain tumor.)
Now what I have just said can’t be literally true, since the neural correlate of
pain is not really C-fiber firing, and we’ve only been pretending it is. And here the
important point is that in order to determine what that neural correlate really is, we
will have to use the ability we already have now, even before the determination has
been made, to know when we are in pain. So, it seems to me, the candid neuralist
32
must admit.
First- and Second-Order States
The functionalist view turns on a subtle distinction that may be made within
the category of states, between first-order states or specifications of states, such as
a subject having excess bile, and second-order states or specifications of states,
such as there being a state of such-and-such a kind, and the subject being in it. The
jargon here is borrowed from logic, where a so-called second-order specification
of a state would be one involving so-called quantification over states, phrases like
there is a state.
Even when there is indeed a unique state of such-and-such a kind, and it is
precisely the state of having excess bile, there will be a conceptual or, as is said,
intensional distinction here between the second- and first-order, despite any
objectual, or as is said, extensional distinction being absent: absent in the sense that
the subjects who are in the state indicated by the shorter specification will be
exactly the same ones as the subjects who are in the state indicated by the longer
specification.
The functionalist view is then that for a human being to be in pain requires a
second-order rather than a first-order characterization: to be in pain is to be in a
state that plays a certain role in human life, just as for a mechanical adder to be
carrying one is for it to be executing a process that plays a certain role in
mechanized addition. Now while functionalists seem to wish to claim that we all
implicitly apply the characterization of pain as that which plays a certain role
whenever we make judgments about who is or isn’t in pain, no functionalist claims
to be able at present to give a complete explicit characterization of that role.
Just as we have been pretending that having C-fibers that are firing is the
neural correlate of pain, let us similarly pretend for a moment that the following
33
two specifications, one pertaining to the external stimuli that cause pain, the other
to the behavior responses that are effects of pain, together constitute the desired
characterization of the role of pain in human life: the state is typically caused by
bodily injury and the state typically causes avoidance behavior.
It hardly needs saying that this is pure pretence: a good many more clauses
would be required for a realistic candidate for a functional role (the “pain role”).
However exactly the role should be characterized, it is supposed to be played or
realized in human beings by the firing of C-fibers, just as the carrying-one role is
played or realized in mechanical adders by gear-turning. Nonetheless, there
remains the same kind of distinction between the second-order state of being in
pain and the first-order state of having C-fibers that are firing that there is between
being in the process of carrying one and having a gear that is turning.
If I temporarily relax my restriction to human mentality to allow
consideration of extraterrestrial mentality, of alien minds, and assume for the sake
of example that some quite different physical process, call it the jangling of
D-cables, plays a role in the lives of Martians, who lack C-fibers, analogous to the
role that the firing of C-fibers in lives of us earthlings, who lack D-cables, then the
difference between the neuralist and functionalist views becomes obvious: on the
neuralist view, pain — pain itself, as opposed to something somehow analogous to
pain — does not occur in C-fiberless Martians, while on the functionalist view,
pain does occur in Martians whenever their D-cables jangle. It is important,
however, that even when we do restrict attention to human as opposed to alien
minds, there is still a distinction between the second-order state of being in a
physical state that realizes a certain role, and being in any first-order physical state,
even the one that does as a matter of actual fact realize the role in question.
Much of the debate that goes on between neuralists and functionalists
pertains to alleged nonhuman minds: neuralists tend to think functionalism makes
34
it too easy, and functionalists tend to think neuralism makes it too hard, to
recognize robotic minds, for instance. Debate over such topics is being excluded
from consideration here by my restriction of attention to the human case. The
debate with dualists on one side, and neuralists and functionalists together under
the banner of physicalism on the other, will be enough to keep us busy.
The Coextensive States Psi and Phi and Sigma
Given our pretences, all three views I am considering may agree that the
states specified by the following conditions are coextensive, in the sense that for
any human being at any time, if one of them holds, then so do the others:
(1)
the individual is in pain
(2)
the individual has C-fibers that are firing
(3)
the individual is in a state realizing the following role P:
typically being caused by bodily injury
typically causing avoidance behavior
Let  or Psi denote the psychological state specified by (1), that of being in
pain. And let  or Phi denote, not the physical state specified by (2) as it stands,
but rather the physical state that would be specified by an amended version of (2)
in which the real rather than the pretended neural correlate of being in pain was
mentioned.
Let  or Pi denote the real characterization of the role of pain in contrast to
the pretend characterization P mentioned in (3), and let  or Sigma denote, not the
second-order state specified by (3) as it stands, but rather the one that would be
specified by an amended version of (3) in which P was replaced by this .
All should be able to agree that  and  and  are coextensive.
The dualist view maintains that the -state is that of having a certain feeling
35
knowable only by acquaintance, as THIS feeling. The coextensiveness of this state
with its neural correlate, the -state, is contingent, meaning something that
actually is so, but that did not necessarily have to have been so, that might possibly
have been otherwise: the feeling in question might have been correlated with some
quite different neural state, the neural state might have been correlated with some
quite different feeling; or there might have been neural states but no feelings at all,
as on the zombie world hypothesis; and similarly for the coextensiveness of the
-state with its functional correlate, the -state.
The neuralist view by contrast maintains that the - and -states are quite
literally one and the very same thing, and that this state is coextensive but not
identical with the -state. The functionalist view by contrast maintains that the and -states are quite literally one and the very same thing, and that this state is
coextensive but not identical with the -state.
Both the antidualist views hold that the -state supervenes upon -state,
though for different reasons: in the case of the neuralist view, because the -state
is the -state, and anything supervenes on itself; in the case of the functionalist
view, because the -state is the -state, the second-order state of being in a firstorder state that realizes the role , and it is the -state that realizes the role , and
roles supervene on realizers.
Armchair Methods
A major difference between the neuralist and functionalist views pertains to
how we are supposed to be able to know that the -state is identical with its
correlate. The identity of being in pain with such-and-such a neural state, as
asserted by the neuralist view, is supposed to be an aposteriori discovery, a result
of scientific observation and experiment, of specialists with special equipment
looking closely at what is going on in the nerves and brain when pain is present.
36
The identity of being pain with being in a state that realizes such-and-such a
role in our mental life, as asserted by the functionalist view, is supposed to be an
apriori discovery, something one can arrive at by so-called armchair methods, just
by sitting and thinking, and indeed it is supposed to be an analytic rather than a
synthetic truth, something one can arrive at just by sitting and thinking specifically
about the meanings of words, and in particular, of the word pain.
More precisely, if we had a genuine discovery in place of the pretence that is
all we actually do have at present, it is the assumption of the neuralist that the
discovery would be an aposteriori and synthetic one, and the assumption of the
functionalist that the discovery would be an apriori and analytic one: to eliminate
the need for pretence, the neuralist is waiting for a discovery on the part of
neuropsychological observation and experiment, while the functionalist is waiting
for a discovery on the part of armchair philosophical analysis.
Note that none of the three ’isms under discussion can claim, at least not at
present, to be in a position to tell us what being in pain, for instance, is. Dualism
offers for this inability the excuse that what being in pain is can only be learned by
personal experience, not by being told about it.
Neuralism offers the excuse that what being in pain is has been narrowed
down partially and approximately by scientific research, but that more such
research is needed before it can be pinned down fully and exactly, and such
research is costly and difficult, not to mention being constrained by ethical
prohibitions, which often oblige the scientist to rely on clinical rather than
experimental data, and wait for cases to occur naturally rather than induce them
artificially.
I have represented dualism and neuralism as granting that the -state is
coextensive with the -state. They do not, however, grant what is in a sense the
functionalist’s main claim, as it applies to pain, namely, that it is an apriori and
37
analytic truth that the -state is coextensive with the -state. Functionalism,
inasmuch and insofar as it makes such a claim, might seem to be in a more
embarrassing and vulnerable position than its rivals, since it claims the right
characterization of the role of pain can in principle be known apriori, by analysis of
meaning. What is the excuse, then, for not having arrived, after decades of
discussion, at the knowledge aspired to?
It will be important for future purposes to spell out why this kind of
objection cannot be taken to rule out functionalism immediately. I therefore beg
the reader’s patience while I stop to summarize the explanation of why even
analytic, apriori truths may be difficult to discover (a matter with which many
readers will already be familiar, since it is much discussed in the literature, under
the rubric “paradox of analysis”).
Trial and Error
To start far from contentious issues in semantics, all native English speakers
can effortlessly produce the plurals of English nouns in the regular case —
excluding talk of geese or oxen or the like. But most cannot state the rule for when
one adds an s-sound, as with gates, or a z-sound, as with gales, or a syllable
consisting of a neutral vowel followed by a z-sound, as with gazes. Most do not
even have in their vocabularies the technical terms in which phonologists would
formulate the rule: voiced, sibilant, schwa, and so forth. And two hundred years
ago there just were no phonologists, though there were English speakers and they
were forming plurals just as we do today.
I have said there is vastly more to psychology than issues about
consciousness. One instance is that linguists often hold that their rules are in some
sense what they call psychologically real, though native speakers do not
consciously follow them the way someone learning English as a foreign language
38
might. All they are aware of consciously, and the only so-called linguistic
intuitions they can supply to the linguist who consults them as so-called
informants, are case-by-case judgments about what is the plural of this word, or
that word, of some other word.
The linguist seeking to arrive at a rule can only proceed by looking at a body
of such judgments, forming a hypothesis that seems to fit them, testing this original
hypothesis against further such judgments as data points, revising the hypothesis if
discrepancies are found, testing again, revising again if discrepancies are found,
and so on, until the process stabilizes and a hypothesis that does not need revision
in the light of further examples is at last arrived at. The process is an armchair
method so long as the linguist is dealing with his or her native language and taking
him- or herself as the only informant.
The case seems entirely similar with semantic as contrasted with
phonological rules. Lexicographers seeking definitions and philosophers seeking
analyses start with candidates, check these against intuitions about particular
examples, revise, and repeat. The need to engage in such a trial-and-error dialectic
of analysis, as I will term it, is part of the reason why lexicography is as much art
as science. It is part of the reason why competence in the use of words does not
inevitably entail facility in producing definitions: though William Shakespeare was
an immeasurably greater writer than Samuel Johnson, we cannot be confident that
if he had set about to produce a dictionary he would have produced a better one.
The fact that semantic rules can only be got at indirectly through a dialectic of
analysis is also part of the reason why in philosophy correct analyses are not
immediately recognized as right, and why erroneous analyses are not immediately
recognized as wrong.
There is another reason why even apriori, analytic truths may be difficult to
discover, namely, that the category is supposed to be closed under logical
39
consequence — logical deduction being a clear case of an armchair method — and
even if one has definitions of the key terms before one, it may not be evident that
some thesis formulated in those terms follows logically from those definitions.
This book is not the place for any extended excursion into philosophy of
mathematics, but it may be mentioned that mathematical theorems have often been
claimed to be apriori and analytic, though mathematical proof involves such long
chains of logical deductions that there is no easy way to foresee what results it may
lead to. Results may languish as a conjectures for centuries before being turned
into theorems by the discovery of proofs. How far anything similar occurs outside
mathematics is more questionable. But even without bringing in this factor we
already have, in the nature of the dialectic of analysis, a sufficient excuse for
functionalism’s not having arrived sooner, or as yet at all, at a characterization of
the role of pain.
Behaviorism versus Functionalism
Nonetheless, it might be felt that the excuse must wear thinner and thinner,
and functionalism must seem less and less plausible, each year that goes by
without the specifications being produced. So it is important to note that my
description of the situation has been understated or oversimplified so far. For the
role of pain, or any other mental state, is not supposed by the functionalist to be
characterizable just in terms of the external stimuli that cause it or the external
responses that are its effects, as in my pretend characterization P of the role of pain
in human life.
On the contrary, the view that mental states can be characterized in such
stimulus-response terms was called analytic behaviorism, and that is a view now
discredited and defunct, with functionalism as its chief heir. Functionalism differs
from it by allowing that interaction with other internal states, including beliefs and
40
desires, may be involved. Mention of these would also be expected to figure in the
real characterization : we might have not just typically being caused by bodily
injury on our list of specifications, but also typically causing belief that the body
has been injured, for instance.
As a result, ultimately role analyses must be sought not one by one, but all at
once, for the whole range of interacting internal states mediating between external
stimuli and behavioral responses. This circumstance makes the task of analysis an
immense one, and so perhaps excuses further delays. But perhaps it also makes the
overall idea of functionalism less plausible. Do we really already have in us, in
some psychological real albeit implicit way, an enormous and complicated network
of interlocking role specifications for sensations and affects, for perceptions and
emotions, and for beliefs and desires alike? If so, how did we ever manage to
acquire it, or are we being asked to believe that it is innate?
Is it not simpler to suppose in the case of pain, say, that we directly learn
what it means as very young children by hearing the word when experiencing the
feeling, and associating the two, thereafter finding out various facts about pain,
beginning with various identifiable causes, by induction from our own experiences,
and after the very young age at which we become able to project ourselves
imaginatively into the situations of others, from their testimony as well? It is no
wonder the dualist and the neuralist are dubious, though I will not press this
objection to functionalism here.
In any case, the supervenience thesis marks a main feature of agreement
between the two rival antidualist views, and therewith the dividing line between
dualism and physicalism. This is one reason for the interest and importance the
supervenience thesis has seemed to many recent philosophers to have.
But my ultimate conclusion, which I described earlier as being that the
mind-body issue needs to be disentangled from issues about modality, may now be
41
more specifically described as being that the supervenience question does not after
all have the intrinsic interest and importance many recent philosophers seem to
have attributed to it. Still, again as I have said earlier, my ultimate conclusion or
final destination will be less important than the journey towards it, and I hope it
will be found along the way that the supervenience thesis has at least a
considerable extrinsic interest as a peg on which to hang the discussion of a series
of issues about mentality and modality.
The Dominance of Physicalism
For simplicity I have been and will be writing as if there are only three
positions on the mind-body problem that need to be taken into consideration:
dualism in its hardcore contemporary form, and two varieties of physicalism,
namely, neuralism and functionalism of the kinds I have sketched. If a solid case
could be made for any one of the three, the question of supervenience would be
settled, albeit indirectly: in the negative if dualism could be established, in the
affirmative if either variety of physicalism could be upheld. It is therefore natural
to begin the examination of the question of the status of supervenience by
considering the prospects for such an indirect approach to supervenience through
direct argumentation pro or con dualism or neuralism or functionalism.
But here it must be said that a quite striking feature of the mind-body
literature is a strong tendency for arguments that to one side seem decisive and
dispositive to seem to the other side feeble and flimsy. This is so to a degree that
makes some suspect that allegiance to dualism or physicalism may be more a
matter of personal temperament (“tender-minded” or “tough-minded” in Jamesian
terms) than anything else.
William Lycan, a physicalist who professes to believe in giving dualism its
due, as he says in a well-known note of that title, is one who notes this point from
42
the physicalist side. If there is no one equally prominent that I know of noting the
point as emphatically from the dualist side, that is perhaps simply because there are
so many fewer dualists than antidualists. Fewer, at any rate, among specialists in
philosophy of mind. For it may be that there is a self-selection phenomenon at
work here: it may be that many instinctive dualists simply avoid that area of
philosophy knowing how much it is dominated by physicalists.
Be all that as it may, physicalism is so much the dominant view that it is
often treated as the default view, as if the burden of proof were obviously on the
other side. The result is that many physicalists offer only very casual and cursory
arguments against dualism before getting down to the issue that really interests
them, that of which form of physicalism is the best. The casual and cursory
antidualist arguments, Lycan candidly observes, generally are not such as would
need to be taken very seriously by a committed dualist.
Arguments and Counterarguments: Parsimony
Let me just review a few before turning to what is by contrast perhaps the
one and only argument that the dualist has to take very seriously indeed, which will
occupy us for the whole of the next chapter. Two very general physicalist
arguments are premised on the correlation and uniformity principles, which I am
taking the contemporary dualist to concede, or rather, to endorse, plus distinctively
physicalistic versions of two alleged maxims of scientific method, pertaining to the
need for parsimony in making assumptions and for evidence before adopting them,
that dualists would surely accept in some version or other of their own.
A first physicalist argument is that since the neural and/or functional states
that are their physicalistic correlates have to be recognized anyway, it is a violation
of what may be called the principle of parsimony, the rule not to recognize more
entities or kinds of entities than one needs to — a rule sometimes called Occam’s
43
razor — to recognize those strange alleged entities, the irreducibly mental states
posited by dualists, as something additional.
To this the dualist may reply that while indeed one should not gratuitously
postulate entities that one does not need to, one does need to acknowledge entities
that are staring one in the face, and this includes mental states such as feeling pains
or seeing red — than which, by the way, nothing is more familiar, which is to say,
nothing is less strange.
What if it is suggested in rebuttal that still one would be recognizing fewer
entities if, without renouncing mental states as gratuitous, one doubled up, so to
speak, and while conceding the occurrence of sensations, identified them with their
physicalistic correlates? Well, the dualist might note in response that likewise one
would be recognizing fewer entities if, while conceding the existence of the letters
A-Z, one identified them with the numbers 1-26; yet no one takes this as a reason
to believe that the letter F is quite literally one and the very same thing as the
number 6. The dualist may simply deny that there is any general rule of scientific
method enjoining not merely the avoidance of needless posits, but the doubling up
of needed ones.
More Arguments and Counterarguments: Evidence
A second physicalistic argument is that the uniformity of the laws governing
the behavior of brainful and brainless physical systems guarantees that if mental
items are distinct from their physicalistic correlates, one can never in practice have
evidence that mental states are present — at least not in any case other than one’s
own. It is, therefore, the physicalist may continue, a violation of the what may be
called the principle of evidence, the rule not to accept beliefs without positive
reasons in their favor, to posit that other people besides oneself have mental lives.
For any supposed argument by analogy from one’s own case would be an
44
incredibly weak induction from a single case. Thus one of our most basic beliefs, a
virtual postulate of sanity, would be left without justification, if the distinctness of
the mental from the physical were granted.
To this the dualist may reply first of all that, even granting the physicalist’s
claim about dualism’s implying the lack of a certain justification, the circumstance
that a view would imply that we lack justification for some basic beliefs seems at
most a reason for hoping, and not a reason for believing, that the view is untrue.
And the existence of a reason for hoping that is not a reason for believing is first
and above all a reason for being especially careful about avoiding wishful thinking.
That aside, the dualist will grant that indeed one should not adopt new
beliefs without evidence, but insist that adopting new beliefs is not what is at issue
here. The belief that people around me are not zombies (belief in “other minds”) is
of a piece with the belief that the future will resemble the past (belief in the
reliability of “induction”) and that there is a physical universe around me and not
merely a play of my own sensations (belief in an “external world”). These beliefs
are not something being proposed as novelties just now. They arise early in
childhood — so early that they cannot plausibly be regarded as conclusions
inferred by any sort of intellectual argumentation whatsoever.
Indeed, they are really less articulated beliefs than spontaneous reactions to
experience: I spontaneously react to other people as if they had minds, in part by
projecting myself imaginatively into their situations; I spontaneously form
expectations of what will happen tomorrow on the basis of what is happening
today and what happened yesterday; I spontaneously react to my sensations as if
they were occasioned by things outside myself.
The question is whether I must give up these instinctive beliefs or
spontaneous reactions if I cannot provide an after-the-fact justification and
articulate positive reasons for them (“epistemic foundationalism”), or whether I
45
may retain them until someone articulates convincing reasons against them
(“epistemic conservatism”). Is positive evidence needed to justify retaining old
beliefs, or only to justify changing beliefs? The physicalist who offers the
argument under discussion seems to assume the former, without argument.
So did Descartes, whose project begins by discarding anything that seems
open to any kind of doubts however far-fetched, even those based on fantasies
about the evil machinations of demons, and then turns to the attempt to build up a
structure of belief from indubitable first principles by indubitable inferential steps.
The most important difference between contemporary dualism and Cartesian
dualism is one that I have not previously mentioned, namely, that present-day
dualists are not engaged in any such project.
Scratching a Representation
If these first two, general physicalist arguments make little impression on
dualists — it perhaps hardly needs saying that, reciprocally, dualist responses make
little impression on physicalists — other physicalist arguments, based on more
obviously contentious premises, make even less impression, and are even more
easily dismissed. Lycan considers several such arguments, and I will mention one
more here, for which Lycan himself seems to have a weakness — not that I would
attribute to him the crude formulation of the argument that I am about to present.
The contentious premise in the argument in question is that a sensation is
never anything more than a so-called representation, of a certain kind, of some part
of the world as being a some way (this claim being an aspect of a more general
“representational theory of consciousness”). Thus my having a sensation, say, my
feeling an itch on the tip of my nose, is nothing more than a representation, of a
certain kind, being put to me, presenting some part of the world, presumably the tip
of my nose, as being some way, presumably itchy.
46
This premise is then usually combined with — and from a point of view
concerned only with the mind-body problem is of scant interest unless combined
with — a claim that a physicalistic analysis of representation in general is
available, making every representation, itches included, into something
physicalistic, whether neural or functional or whatever.
To this the dualist may respond by asking whether my having a
representation, of the kind in question, of the tip of my nose as itchy, is or is not
supposed to involve my feeling a certain way, my having a certain feeling. If it is
not, then having such a representation cannot be what having an itch amounts to. If
it is, then however successful a physicalistic analysis may be with other kinds of
representations, it must fail in application to this kind, because physicalistic
analyses always leave out the way things feel.
There is another physicalist argument that the dualist must take and does
take more seriously. It is a specifically neuralist argument that seems on the face of
it to bear as strongly against functionalism as against dualism. But I will leave the
functionalists to defend themselves on this point, and consider only potential
dualist responses. Yet before turning to these matters and starting a new chapter, I
want to mention one further antifunctionalist argument that has perhaps been
slower to come to the fore than it ought to have been.
Dissociation of Suffering from Pain
The argument turns on a special difficulty concerning the analysis of pain,
suggesting that attempts at characterizations of its role that emphasize the supposed
behavioral effects or motivating force of pains are off-track, reflecting a limitation
of armchair methods in dealing with questions of sensation and affect.
Usually when pain is present there are present two elements: first, a
sensation, characterizable as sharp or dull, steady or intermittent, and referred to
47
some part of the body, real or phantom; second, an affect, a negative one, which
tends to be called, depending on the intensity of the pain, either being bothered by
it or suffering from it. It seems that hardly any philosophers have had the
experience of feeling the sensation but not the affect, though this condition of pain
dissociation occurs regularly, at least in the case of second pains, as the effect of
certain drugs used to treat chronic pains that persist past the stage of doing
anything useful in the way of signaling possible damage to the body. I myself have
experienced the effect of such drugs just once many years ago as a student, before I
had taken up philosophy in a serious way, and a close family member, not a
philosopher, has experienced it twice. (A related “pain asymbolia” results from
certain kinds of brain surgery, though there is a question whether it is the same
experience in these exotic cases as in the more common drug-induced case.)
The drugged state is unforgettable to those who have experienced it, but
probably unimaginable to those who have not. It seems to be, spontaneously and
noncollusively, described by those who have experienced it in similar terms,
namely as feeling pain but not being bothered by it or feeling pain and not
suffering from it. The ordinary childhood training in the use of the word pain
seems to leave the speaker disposed to describe the unusual experience in this way.
If so, then it seems the ordinary meaning of the word pain, as internalized
after the ordinary childhood training in the use of the word, is such that pain can
properly be said to be felt when only the sensation, and not the affect that usually
accompanies it, is present. If so, then it seems that pain denotes the sensation,
rather than the affect or the complex of sensation plus affect. But the sensation by
itself, without the affect that usually accompanies it, seems to have very few
behavioral or motivational effects. For instance, it seems an error to say that being
in pain, as contrasted with being bothered by or suffering from being in pain,
causes wincing or moaning.
48
How Armchair Methods Go Wrong
There is a point about meaning here that needs emphasis, connected with the
trial-and-error dialectic of analysis. The crucial point is that, when checking an
example, when considering whether such-and-such an expression would be
appropriate in such-and-such circumstances, one normally does not need to have
been in or need to get oneself into those circumstances. Usually it is enough just to
imagine them. But it is a peculiarity of sensory experiences that those who have
never had them generally cannot imagine them, either.
For this reason, when philosophers who have never experienced a situation
where the sensation is present but not the affect try to guess what would be the
correct use of pain in such a situation, they may go wrong; that is, they may arrive
at judgments differing from those displayed by ordinary speakers who are actually
experiencing such a situation. They may guess that without the affect the sensation
cannot properly be called pain, and so offer definitions or analyses according to
which pain is inherently unpleasant; and indeed such definitions and analyses are
quite common, and typically causing wincing or moaning is indeed often included
in functionalists’ accounts of what sort of thing might be part of the
characterization of the role of pain.
I am suggesting that such analyses are wrong, and with them many
assertions about the supposed effects of pain on behavior, put forward under such
labels as platitudes or truisms, and claimed to be deliverances of common sense,
and potential elements of a characterization of the functional role of pain. I am
suggesting that armchair methods go systematically wrong in treating as effects of
pain what are not effects of pain properly so called, but at most of the affect that
usually but not invariably accompanies the sensation of pain.
Of course, pain and suffering do typically go together. If we could know
49
apriori that it is typically a case of the sensation of pain causing the affect of being
bothered or suffering, which may then in turn cause wincing or moaning, rather
than a case of two distinct effects of a common cause, then the formulation
guarded by the word typically, as functionalist formulations typically are, would be
defensible. But the formulation about pain typically causing wincing and moaning
cannot as a matter of actual fact be defended in this way, since it is surely an
entirely empirical, aposteriori issue what causes what.
Against all this it may be urged that the expert authorities, scientific groups
devoted to the study of pain, have offered official definitions according to which
pain is an unpleasant sensory experience, building the affect in. To this argument
from authority it must be replied that, while specialist groups may have very good
reasons for framing their definitions as they do, it does not follow that those
technical definitions are in agreement with the ordinary meaning of the word,
which is to say, in agreement with what is internalized after the ordinary childhood
training in the use of the word. Nor does it follow that philosophers ought to follow
the specialist usage if there is a divergence.
A clear counterexample is provided by the case of abortion, where
professional medical terminology radically diverges from popular usage. The latter
distinguishes abortion from miscarriage, while the former calls both abortion and
distinguishes them as induced and spontaneous. Philosophers generally and rightly
follow the populace rather than the medicos, and though there is a vast
philosophical literature indexed under the term abortion, what is at issue in this
literature is abortion in the popular sense, which in medical terms would be called
induced abortion. There is almost no philosophical literature on abortion in the
medical sense, which in popular terms would be called disjunctively abortion-ormiscarriage, since abortion and miscarriage seem to have next to nothing of
philosophical, as opposed to medical, interest in common.
50
Other Dissociations
Now the case of pain dissociation is only one of many in the literature of
clinical and experimental psychology where aspects of conscious experience that
normally go together come apart. No one has experienced all of these dissociations
— of the ability to see from the ability to recognize faces, of the ability to
recognize faces from the ability to read their mood, and so on. Most seem
incredible and unimaginable when described to those who have not lived through
them, as any reader of the popular works of the late Oliver Sacks will be aware.
All this raises the gravest doubts about whether an assemblage of what seem
truisms or platitudes to those with little experience of abnormal dissociations can
ever amount to a specification of a role for an inner state that could justly be called
an apriori or analytic account, comparable to lexicographers’ definitions or
philosophers’ analyses in other areas where no similar limitations on the ability to
imagine hypothetical cases apply. Reflection on such matters can make the
functionalist approach, in the version I have been considering, or any version
involving similar claims of analyticity and aprioriness, seem simply a nonstarter.
I make no claim to have a knock-down argument here. I merely offer the
foregoing considerations as part of an excuse, along with the ever-present
consideration of limitations of space, for the intention I now announce to adopt a
policy of giving henceforth less attention to functionalism than to neuralism as an
alternative to dualism. In particular, as I have already hinted, I will in the next
chapter be considering a specifically neuralist argument against dualism, and
dualist replies that are to a degree replies to neuralism specifically.
51
III
Mental Causation
One last point of comparison between Cartesian and contemporary dualism
remains to be noted, concerning the degree and kind of commitment undertaken to
this concept of cause (or “causation”). A popular anti-dualist argument turns on it.
The Closure and Exclusion Principles
Descartes’ acceptance of the concept is unquestioning and uncritical. The
concept is, moreover, indispensable for Descartes’ view. For on that view, apart
from his knowledge of himself as a thinking thing, everything else depends on his
purported knowledge of the existence of God, while the purported proof thereof
depends on the concept of cause and a principle formulated in terms of it, the
principle that there is as much reality in the cause as there is in the effect (indeed,
as much “formal reality” or “eminent reality” in the cause as there is “objective
reality” in the effect). Take away the concept of cause and there is nothing left of
the Cartesian system.
Almost all contemporary philosophers of mind tend to be just as
unquestioning and uncritical in their acceptance of the concept of cause as was
Descartes, but the role the concept plays differs from theory to theory. For
antidualists the concept is indispensable, or at least useful, while for the dualist it is
only a source of trouble. It is indispensable for the functionalist, since specification
of the functional role of pain and other internal states is supposed ultimately to be
given in terms of the state’s place in a causal network, mediating with other
internal states between external stimuli and behavioral responses, as we have seen.
The concept of cause is useful for the neuralist, and a source of trouble for the
dualist, because what is perhaps the neuralist’s strongest argument against dualism
turns crucially on the notion of causation.
52
If one accepts the concept of cause, not merely for purposes of everyday life,
but for those of philosophical analysis and/or metaphysical speculation, it is natural
to attempt to rephrase our earlier discussion of the correlation and uniformity
principles, which I formulated in terms of laws of nature, in causal terms. If one
undertakes this task of reformuation, one is likely to be led pretty quickly to a
closure principle, to the effect that physical events have physical causes to the
extent that they have causes at all — in particular, where they have only statistical
causes with given statistics, they have statistical physical causes with those given
statistics.
From the closure principle it is not a long step to the exclusion principle, to
the effect that physical events have no nonphysical causes. (The route from closure
to “the exclusion problem” depends on philosophers’ pretty uniformly finding
repugnant the postulation of “overdetermination”, of massively coincidental
redundant nonphysical causes for physical events that already have fully sufficient
physical causes, a view uncomfortably closely akin to Leibniz’ doctrine of “preestablished harmony”.) And it is on the exclusion principle that what is perhaps the
most popular argument for neuralism rests.
The Neuralist Causal Argument
The argument I am alluding to is not the one, sometimes encountered, that
on dualist principles I can never know I have a sensation, because the sensation as
such plays no causal role in generating my belief that I have a sensation. That
argument would belong to the category of what I characterized earlier as easilydismissible arguments based on obviously contentious premises: in this case, the
thoroughly discredited causal theory of knowledge.
The argument I am alluding to is the one in which the claim made for
neuralism is that it alone can reconcile the exclusion principle with the robust
53
commonsense belief, present from time immemorial and not likely to go away any
time soon, that mental causes often underlie physical events, specifically those that
consist in changes in or movements of human bodies.
For even if one restricts attention, as I am doing here, primarily to sensations
and secondarily to affects, or primarily to perceptions and secondarily to emotions,
common sense will insist that they, in conjunction with other factors, do have
physiological and behavioral effects. Surely it is just common sense that the onset
of embarrassment may help cause the onset of blushing, and that visual sensations
of the traffic light turning red may help cause a motor act of pushing down on the
brake pedal. According to neuralism, there is perfect compatibility between there
being such mental causes and there being no non-physical causes, since according
to neuralism, everything mental is physical as well, and there is nothing mentalbut-not-physical left over to be the cause of anything.
It is only if one recognizes the event of beginning to feel embarrassed, or of
becoming aware of the light turning red, as something other than its neural
correlate, that any problem arises about how these mental events can have physical
effects. The dualist seems committed — I mean one who accepts the exclusion
principle, as I am assuming the contemporary dualist does — to maintaining that
common sense systematically errs by mistaking for the effects of mental events
what are really effects of neural events distinct from them, and merely contingently
correlated with them. Or so the neuralist argues.
Types of Responses
The kind of answers to the physicalist’s earlier arguments that I have put
into the dualist’s mouth in my discussion of Lycan constrain the answer the dualist
can give to this latest challenge: there can be no denying that the contemporary
dualist’s position is indeed contrary to common sense. The belief that conscious
54
experience influences bodily behavior is as much an instinctive, spontaneous
reaction, going back to early childhood, as the belief that other people have minds,
and so presumably, on my own principles, or those I have ascribed to the
contemporary as opposed to Cartesian dualist, it is something not to be given up
unless strong reasons can be put forward against it.
If the belief has indeed been given up by dualists between Descartes’ day
and our own, the reasons cited can only be those that could be cited for adopting
the correlation, uniformity, closure, and exclusion principles: the slow, steady
progress in discovering neural and encephalic structures and processes connected
with our mental life, and the utter failure to find any convincing example of
anything going on in the nerves or brain constituting an exception to the general
laws of physics and chemistry applicable to dead matter.
The principles that seem to leave no room to the dualist for recognition of
mental causation may, to be sure, be less firmly established than, say, the law of
the conservation of electrical charge. But the evidence does seem to establish a
strong, though defeasible, presumption in their favor. And are they not more or less
taken for granted in the methodology of on-going scientific research?
Meanwhile, the dualist’s refusal to double up, taking mental items to be a
special kind of physical item, prevents the dualist from saving mental causation by
making it a special case of physical causation. Nonetheless, I see four
distinguishable lines of potential dualist response to the challenge of the mental
causation problem, partly independent, partly overlapping and mutually
reinforcing.
First, the dualist may cite history, which shows that common sense has been
found grossly mistaken in the past, and that we have learned to re-educate it.
Second, the dualist may cite scientific psychology, as producing ever more
evidence of the unreliability of common sense judgments precisely about
55
conscious experiences as causes of behavior. Third, the dualist may cite a
distinguished philosophical tradition of questioning reliance on the concept of
cause beyond the realm of everyday practical affairs, rejecting employment of the
notion in philosophical analysis or metaphysical speculation. Fourth, the dualist
may make a tu quoque response, telling the neuralist you, too when it comes to
charges of violating common sense: the dualist may claim that the identification of
the mental with the neural is itself uncommonsensical, and indeed that doubling up
is a conceptual confusion.
With apologies to those impatient readers who find such points obvious and
not in need of elaboration beyond that already given them in an extensive literature
— a point of view with which I have no lack of sympathy — I will elaborate these
lines of argument one by one in turn.
Historical Precedents: Geocentrism
The view that embarrassment causes blushing may be as firmly entrenched
in naive, primitive, pre-scientific, folk belief as once was the belief that the
alternation of day and night is caused by the revolution of the sun about the earth,
and yet be just as wrong. So the dualist may contend. The wrongness of the former
commonsense belief in geocentrism has by now been generally recognized, and
there is no reason why the wrongness of the commonsense belief in mental
causation should not be generally recognized in the future. So the dualist may add.
Note first of all that recognition, as a matter of scientific and philosophical
principle, that it is the neural correlate of embarrassment, and not embarrassment
itself, that causes blushing need not lead to any change in the way we speak in
everyday life, any more than recognition, as a matter of scientific and
philosophical principle, that the earth goes around the sun and not vice versa led to
any change in the way we speak of sunrise and sunset.
56
If Jill asks Why did John’s face go all red? and Jack answers He was
embarrassed by what Joan was telling him, Jill will not respond with a lecture on
how it is the neural correlates and not embarrassment itself that causes blushing,
any more than, if Copernicus asked Why has it gotten so dark? and a neighbor
answered The sun has sunk below the horizon, Copernicus would have responded
with a lecture on how it is the earth turning on its axis and not the sun moving that
causes the evening dusk.
We very often speak as if something were the case that some or all of us,
after a greater or lesser amount of reflection, recognize is not. Where it is only a
few of the more reflective among us who see that it is not, philosophers will speak
of an error theory of the status of ordinary discourse, while where the error has
come to be more widely recognized, they may speak of fictionalism or figuralism,
but it is in either case the same sort of thing, or at most, different stages in the same
sort of process.
Copernicus may have spoken with crossed fingers, deliberately concealing
his true belief for fear of the inquisitors, but as geocentrism gradually gave way to
heliocentrism, talk of the rising and setting of the sun came to be quite generally
regarded as no more than a figure of speech, ellipsis for talk of the apparent rising
or setting of the sun. And as dead metaphors become in time secondary literal
senses, and perhaps ultimately primary literal senses, it is not surprising to find online dictionaries today defining sunset not as the sinking of the sun below the
horizon but rather as the apparent sinking of the sun below the horizon.
Common sense thus is capable of being educated. The dualist is claiming no
more than that such education is needed in a case where today even philosophers,
under the influence of false doctrines such as neuralism, have mostly thus far failed
to recognize the need, just as in the sixteenth century even astronomers, under the
influence of false Aristotelian and Ptolemaic doctrines, failed to recognize the need
57
to revise our understanding of why day follows night.
One difference between the cases of mental causation and geocentrism, a
difference that only strengthens the dualist’s case, is that Copernicus was able,
when writing for posthumous publication directed at philosophical and scientific
colleagues, to say what really is going on when the sun appears to rise or sink in
the east or west, whereas we do not — at any rate, I do not, and I expect the reader
does not, either — know how to describe in physiological terms the exact neural
correlate of the feeling embarrassment, which is supposed to be the real cause in
the case of the blushing. We are able to describe it only as the-neural-correlate-ofembrassment-whatever-that-may-turn-out-to-be.
If we actually took to so describing it, and constantly saying such things as
that John’s face flushed as an effect of the neural correlate of embarrassment, it
would be like constantly speaking of the sun’s apparent setting. The fixed phrase,
neural correlate of, like the fixed word apparent, being invariant and an object of
tedious constant repetition, would most likely soon be left to be understood tacitly.
We would then be back to the ordinary way of speaking, with a mental reservation
or purpose of evasion, a tacit understanding that it is not to be taken quite literally.
And then eventually it could be expected to follow that the meaning of the words
we use would shift to agree with what we really intend when we use them, as has
happened in the case of the dictionary definitions of sunset that are now current,
which take it to refer to a mere appearance of motion rather than a genuine motion.
Intolerable Thoughts and Incredible Theses
Against all this it may be urged that to speak of the dualist position as going
against common sense is to understate the case. The proposition that none of our
sensations or affects ever issue in vocalizations and bodily motions, but only ride
along atop the physiological we-know-not-what that really causes our lips and
58
tongues and our arms and legs to move, is not just a departure from naive ways of
thinking, like the proposition that it is the sun that stands still and the earth that
moves. It is, rather, a way of thinking of ourselves that may justly be called
intolerable.
The dualist position may be termed incredible in the literal sense that it is so
intolerable a thing to try to believe, that no one can succeed in believing it
seriously and sustainedly, 24-7. One may profess it in philosophical debate, but
will fall back into the ordinary way of thinking when one lets down one’s guard.
This may be less a sober argument than an emotional protest, a cri de cœur
as the French would say, but in any case it deserves a reply. I think the dualist’s
best reply would be that, even allowing the foregoing antidualist claim to pass for
the sake of argument, it does not constitute a disproof of dualism. Here a wellknown aphorism of Nietzsche’s, — item #121 in the third book of The Gay Science
— springs to memory. In opposition to the claim that no one could manage to live
without postulating this or that, it is answered that life is no argument: error might
be among the conditions of life.
The dualist’s reply might be strengthened if it were possible to offer an
analogy, an example of a proposition having the features of being literally
incredible in the sense that no one can believe it all the time, and yet being true.
Any discussion of the putative example will be a digression, and there can be no
question, therefore, of going into the matter at any length or in any depth. Yet I
will hazard a candidate, and tentatively cite the proposition that there is no such
thing as what I will call strong free will.
By strong free will I mean free will in a sense that is strongly a matter of
freedom, incompatible with determination by chains of antecedent causes
ultimately going back outside the agent, and equally strongly a matter of will,
incompatible with occurrence by mere random happenstance — and indeed I mean
59
free will in a sense that is incompatible with determination by a combination of
antecedent causes plus random happenstance. Let me attempt, in the limited space
of a digression no longer than one further section of the present chapter, not so
much to defend the claim that the nonexistence of strong free will is an incredible
truth, as simply to explain what I mean by claiming that it is, leaving evaluation to
the reader.
On the one hand, as to the denial being true, as to there being no such thing
as strong free will, we do in fact seem often to think of the behavior of third parties
as if it exhibited nothing of the sort. We seem to do so when we make forecasts
about the probabilities of various reactions on the part of masses of people to
proposed public-policy measures, for instance.
And anyone who has read attempts by philosophers to explain the concept of
strong free will (which in the literature goes by the label “agent causation”), who
for instance reads commentators on Immanuel Kant’s conception of transcendental
freedom behind empirical determinism, may very well soon be led to doubt that the
notion of strong free will is even genuinely intelligible.
Newcomb’s Problem
On the other hand, it may I think with some justice be said to be intolerable
emotionally and infeasible psychologically to think of ourselves — and our friends
and loved ones, when we are relating to them as an I to a you — as mere
concourses of atoms, whose motions and vocalizations are the result of a
combination of causal chains beginning long before the birth of anyone alive today
with mere random quantum fluctuations. What may seem unremarkable in third
person cases may feel unthinkable in the second- and above all first-person cases.
My suggestion would be that we inevitably keep falling back into thinking of
ourselves as having strong free will, and that in a sense do so every time we ask
60
ourselves What to do?
For it is, I would claim, difficult to the point of being infeasible to keep in
focus simultaneously the question What to do? and the thought that what is going
to be done is already determined by causal chains going back four hundred years,
or is going to rest on something as haphazard as whether a cosmic ray will happen
to strike the brain at a certain angle during the next hour. And I think a well-known
puzzle goes some way towards demonstrating the difficulty of my holding on
simultaneously to the question what to do and the thought that it is already
determined what I am going to do, though whether an analogous puzzle could be
developed showing the difficulty of holding on simultaneously to the question
what to do and the thought that what is going to be done will occur simply at
random I will leave to the reader.
The puzzle I am alluding to is that known as Newcomb’s problem, though
admittedly in that puzzle the additional perhaps problematic assumption is made
that what is determined not only can be deduced by a superhuman intelligence
from a description of the prior state of the universe together with deterministic
laws, but further can be deduced in real time, as is said, deduced ahead of its actual
occurrence, as a before-the-fact prediction rather than an after-the-fact explanation.
The superhuman being puts before me two boxes, one opaque, the other
transparent and containing $1000, and offers me the choice between taking only
the opaque box or taking both boxes, telling me that what has been put in the
opaque box depends on its prediction of what I will choose after the terms are
completely specified to me. The specification is completed by telling me that if the
prediction is that I will take one box, $1,000,000 has been put into the opaque box,
and if the prediction is that I will take two boxes, a cabbage has been put into the
opaque box.
Arguments can be put forward in favor of choosing one box, on the grounds
61
that, supposing it to be part of the conditions of the puzzle that I do completely
believe in the being’s predictive powers and promises, I must conclude that I will
be left with $1,000,000 if I take only the one box, but $1000 plus a cabbage if I
take both boxes. Arguments can be put forward in favor of choosing two boxes, on
the grounds that whatever is in the opaque box is already in it and can’t now be
changed, and that whatever it is I will be $1000 richer if I take both boxes.
While many philosophers regard the second argument as better than the first,
and some regard the first as better than the second, I myself am inclined to see a
genuine antinomy here, in a more or less Kantian sense, showing the
incompatibility of the assumption of the puzzle, that what I am going to do is
already determined, and the presuppositions of my asking myself what is to be
done.
Whatever one thinks of this suggested analogy, I would underscore, as
already constituting a sufficient reply to the original objection, the Nietzschean
point that a feeling that it is intolerable to think otherwise, or even the outright
psychological infeasibility of seriously and sustainedly thinking otherwise, is no
guarantee that what we are habitually inclined to think is true.
The Bearing of Scientific Psychology
A second line a dualist might take would reinforce this first, purely
defensive, line and go over to the offense, citing scientific psychology,
introspective and clinical and experimental alike, as indicating that common sense
is demonstrably at best extremely unreliable when it judges that this or that
conscious experience has played a causal role in generating this or that behavior.
The basic situation can be indicated schematically by arrow diagrams of
cause-and-effect relations, with causes to the left of and sometimes above their
effects, connecting an external stimulus with a conscious sensation and its neural
62
correlate as well as a behavioral response, plus perhaps some neural intermediate.
(1)
stimulus

sensation

response
(2)
stimulus

correlate

response
(3)
stimulus

correlate = sensation

response
(4)
stimulus

correlate

response

sensation
(5)
(6)
stimulus
stimulus


intermediate
intermediate

response

correlate = sensation

response

correlate

sensation
In many cases, common sense would posit pattern a pattern like (1), whereas
the exclusion principle would posit a pattern like (2). Now neuralism would
analyze the situation in terms of a pattern like (3), which is in agreement with both
(1) and (2). By contrast dualism would have to analyze the situation in terms of a
pattern like (4), which though it agrees with (2), disagrees with (1), the
commonsense picture.
Such is the neuralist’s causal argument, reduced to pictures: neuralism can
agree with common sense, dualism must take common sense to be making the
error of confusing a sequence of two successive or anyhow separate effects, the
sensation and the response, of a common cause, the correlate, for a sequence of
cause and effect — not an uncommon type of error generally speaking, but
unwelcome to have to recognize in this particular case.
63
The dualist line I am now suggesting would say that in many cases scientific
results indicate that the real pattern must be one that the neuralist and dualists
views would have to analyze in terms not of patterns (3) and (4), but rather of
patterns making room for an intermediate as the true cause of the response. The
neuralist version would look like (5), whereas the dualist version would look
like (6).
Since both (5) and (6) are incompatible with (1), the supposed advantage of
neuralism, that it agrees with common sense, where dualism differs, disappears.
Both views must say that common sense is in error, though the two views give
different accounts of just what the error amounts to.
Touching a Hot Stove: Reflexes
To begin with a simple instance of the situation depicted in the diagrams,
absolutely untutored common sense may assume that when a person touches a hot
stove, the resultant pain causes that person to jerk away the finger, indeed the
whole hand, an instance of the stimulus-sensation-reaction pattern with the
stimulus being contact with the stove, the sensation being pain, and the response
being jerking away the hand. I have even seen this untutored opinion asserted as
fact in print by a respectable philosopher, though I will name no names.
No more than introspection, if one engages in it while accidentally burning
one’s finger, is required to show that this is wrong: the hand is jerked away before
the pain is felt, and the temporal gap, though short, is noticeable, at least if one
knows to expect it.
What is going on neurophysiologically in this case seems to be well
understood by specialists, at least to a first approximation. The contact with the hot
stove leads to jerking away the hand by a reflex arc through the spinal cord,
whereas the pain is not felt until an impulse has traveled up the spinal cord to the
64
brain, which takes some fraction of a second not too small be noticed. The pattern
involves an intermediate, something going on in the spinal cord.
The prudent physicalist will therefore at most hold that pain is responsible,
not for the immediate reaction of withdrawing the hand, but only for a subsequent
attitude of wariness around hot stoves, in accordance with the proverb once burnt,
twice shy. Even this conclusion may be challenged if one distinguishes pain, the
sensation, from suffering, the affect that usually accompanies it, as already
discussed; and so the really prudent physicalist will at most hold that suffering
from pain is responsible for wariness.
The dualist will deny even that, citing the neural correlate of the suffering,
which seems to be some electrochemical goings-on in the part of the brain known
as the amygdala, and not the suffering itself, as the true cause; but nothing I have
said so far bears on that issue, except indirectly by suggesting that the testimony of
common sense against the dualist claim is essentially worthless, given the gross
unreliability of the witness.
Navigating a Crowded Room: Blindsight
I have so far mentioned only introspective evidence of the unreliability of
commonsense judgments about mental causation, and considered only errors on the
topic of pain. But clinical and experimental psychology suggest many more
substantial and interesting cases of the same kind of errors on the part of common
sense, about a variety of other topics. I will take space to mention just one.
(Block’s discussion of cases where there is what he calls “access
consciousness” but not “phenomenological consequence” is very much to the point
here, to such a degree that I do not feel a need myself to multiply examples; let me
just mention, though, that while dualists can welcome the substance of Block’s
discussion, they can hardly approve his terminology. From a dualist point of view,
65
phenomenal consciousness is a case of pleonasm or tautology, and access
consciousness a misnomer, since it is a case of something that would properly
called unconscious serving as an intermediary.)
One of the strange situations uncovered by neurology over the past decades
that has generated the most discussion among philosophers, so-called blindsight,
is the case I mean. Blindsight was first discovered in monkeys, who even after the
destruction of the part of the visual cortex peculiar to mammals showed a residual
ability to respond to visual stimuli, which ability was then verified to be present
also in human beings who had suffered damage to the corresponding area of the
brain, though they reported a complete absence of conscious visual sensation.
These discoveries may raise some doubts as to whether reptiles, for instance,
really have any such conscious sensation, since they do not have the corresponding
brain structure at all; but having set aside the question of animal minds, let me
concentrate only on the human case. In that case, limited ability to guess what is
going on in what would be the visual field — or one side of it in the case of
patients whose lesions affect only input from one eye — and even to reach
effectively for objects reportedly unseen, has been found, and one even more
curious phenomenon encountered.
A sighted person crossing a lounge full of scattered furniture, if asked how
he or she manages to do so without bumping into it, will doubtless answer, in
accordance with common sense, that the sight of the furniture leads to its
avoidance. The pattern here is the simple one of stimulus-sensation-response, with
the presence of the furniture as the stimulus, the sight of it the sensation, and
avoidance manœuvres the response. In support of this analysis, the fact that if the
lights are turned out, or the subject blindfolded, stumbling into furniture promptly
ensues, can be cited as evidence that vision is indeed involved.
Now the blind, too, as is well known, are to varying degrees — with some,
66
to quite astonishing degrees — able to navigate around obstacles in a similar way.
Only their ability is not, of course, affected by turning out the lights, clearly
indicating that for them vision is not involved, but some other sensory channel. It
is thought that something like location by analysis of echoes is at least a large part
of what is involved, and that those who exhibit the most astonishing abilities are
those who have been able, owing to the plasticity of the brain, to devote parts of it
that would ordinarily be used in processing input from the eyes to processing input
from the ears instead. This need not mean that the conscious experience involved is
auditory, like ordinary hearing; on the contrary, traditionally the presence of
obstacles up ahead has been described as being felt as a kind of pressure on the
face, whence the term facial vision.
The case of the blindsighted is different from the cases of both the sighted
and the ordinary blind. For instance, they report no conscious sensation, visual,
auditory, or tactile, but they do need light. In one celebrated case — of which
videos are available on-line as I write — reported by Beatrice Gelder and
coworkers, a blindsighted individual TN, who had lost conscious visual experience
after a couple of strokes, navigated around obstacles in a lighted corridor while
denying any kind of visual awareness of their presence. The preferred
interpretation here seems to be, to a first approximation, that input from the eyes is
processed through two different channels in the brain, one normally leading to
conscious visual experience, but the other having also important functions, in
particular it seems in navigation around obstacles.
Reflection on such cases must sap confidence in the commonsense beliefs
about the causal role of conscious visual and other sensation in explaining
behavior. For they suggest that in the case of the sighted person crossing the
lounge, the commonsense explanation is, if not wholly incorrect, perhaps
incomplete. There may be a component in the avoidance behavior that is caused by
67
neural states on the path of processing other than the one that leads to conscious
visual experience. Perhaps some substantial fraction of the subject’s evasive
motions are initiated by some stage of neural processing prior to, or anyhow
distinct from, that which leads to consciously seeing the furniture.
More Material in the Same Spirit
When the traffic light turns red as we are driving towards it, and we apply
the brakes, can we be quite sure that the activation of the motor neurons
responsible for the relevant contractions of muscles in the legs was initiated only
after visual sensation of the color change had reached consciousness? Perhaps the
structures involved in blindsight are indeed not responsible, but may not some
other intermediate neural process very shortly before and on the way to the
correlate of conscious sensation be at work? It is hard to exclude the conjecture
apriori, even in the absence of any positive evidence for it.
Dualism, of course, maintains that quite generally, if nothing earlier than the
neural correlate is the initiating cause, at any rate that correlate itself is something
distinct from the sensation proper that suffices to get the body moving. The study
of the limited range of abilities available to the blindsighted does not directly
support such a view, but it supports the view indirectly inasmuch and insofar as, in
common with a good deal of other material to be found in the literature of clinical
and experimental psychology, it undermines confidence in commonsense
explanations, and impugns the reliability of the stories we are inclined to offer
ourselves about why we act as we do.
Though I have been setting aside or ignoring issues about willing or volition,
I may just mention that in connection with those matters, too, there has been
scientific research tending to undercut the commonsense view that attributes causal
efficacy to conscious intentions (and more generally about “agency” and “personal
68
autonomy”). I am alluding to the work of Benjamin Libet and others which claims
to show that when a subject is asked to choose one of two voluntary motions to
make, the motor process is initiated before the subject becomes aware of having
made any decision.
But the interpretation of the experimental results of Libet and allies is, like
the interpretation of most results in experimental psychology only more so, a
matter of controversy.
The Critique of Causality in Hume and Kant
It may seem that the dualist has available only two positions: to reconsider
and reject the exclusion principle, a course I assume the contemporary dualist will
not want to take, or else to hold that the body acts causally on the mind but not vice
versa (“epiphenomenalism”), the position I have so far been implicitly assuming
the contemporary dualist to hold in my first two responses to the neuralist causal
argument. But there is a third alternative, to reject the use of the concept of cause
outside the realm of immediate practical affairs.
I am generally leaving functionalists to speak for themselves in response to
the neuralist causal argument, but I may just mention that functionalists doing so
have sometimes been known to claim that it is not the cause-effect relationship in
the strictest sense that is important, but some other, nearby relationship. The line I
am suggesting for the dualist at this point goes much further, not merely
downplaying but, so to speak, denouncing the notion of causation.
A line the dualist might take, I am saying, would be the bold one of
challenging the unquestioning and uncritical acceptance of the notion of causation,
in terms of which the neuralist’s objection, and the two dualist responses
considered so far have been couched. This third route seems, so far as the
published literature is concerned, generally a road not taken. Yet I want to put it on
69
record that the road is there, beckoning and inviting, for the dualist courageous or
quixotic enough to want to explore it.
Again the issue goes back to the days of Descartes, or almost. Descartes’
immediate successors were as unquestioning and uncritical in their acceptance of
the notion of causation as he was himself, and yet the principles they found evident
and compelling were very different from his. For Descartes, if I stub my toe, this
causes a motion in the fluid supposed to fill the nerves, leading from the toe to the
brain and ultimately the pineal gland, and the motion in this last key organ then in
some unexplained way causes a sensation of pain in my mind. For some of
Descartes’ successors (advocates of “occasionalism”), however, the events in my
body only provide the occasion for God, the one and only true causal agent, to
cause a sensation in my mind.
Such disagreements as that between Descartes and these successors of his
could only occur because, unlike the stone and the pain, causal connections are not
perceived, but only inferred. This last observation was the entering wedge for
David Hume in his critical approach to the whole topic of causation, most
conveniently available in the seventh essay in his Enquiry Concerning the Human
Understanding.
On Hume’s view the whole notion involves a kind of illusion or confusion.
We experience a classes of cases that to us seem similar, call them A-cases, each
followed by an experience of a case of another class that also seem to us similar,
call them the B-cases. After repeated exposure to the temporal succession and
constant conjunction of A-case followed by B-case, on next experiencing an
A-case there arises in us a certain impression, a sentiment of anticipation,
expecting another B-case, waiting for the other shoe to drop. The idea we derive
from this impression, of a connection between A-cases and B-cases, though it
originates in ourselves, we project onto the world around us, and erroneously come
70
to think of ourselves as having discovered that there is out there a secret power
within A’s to produce B’s.
Kant, in his Prolegomena to Any Future Metaphysics, characterized Hume’s
view as being that the notion of causation was not a child of reason but a bastard
of the imagination, impregnated by experience. And yet, in the preamble to the
Prolegomena, Kant defended Hume against some of his critics, noting that Hume
never questioned whether the notion of cause was right and even indispensable.
And indeed, Hume did not suggest that we should give up the notion, or even that
we could give up the notion if we wanted to, for everyday, practical purposes: we
will inevitably think in causal terms.
The issue, according to Kant, was over the origin of the notion, of which
Kant offered a very different and much more complex account, which fortunately
there will be no need to enter into here. In this connection Kant notes that the issue
of origin is closely connected with the issue of wider application of the concept;
and this is what was really concerned Hume.
For Kant’s remark is a delicate allusion to Hume’s criticism of the appeal to
the notion of cause outside everyday situations, in metaphysical speculation, a
criticism that becomes most prominent in Hume’s critique of natural theology,
adumbrated in the eleventh essay in the Enquiry, and elaborated in the
posthumously published Dialogues Concerning Natural Religion. And to come to
the crucial point, Kant, for all his differences with Hume, in the end largely agreed
with him about the limited applicability of the notion to topics that transcend
experience. (When it comes to arguments for the existence of God as a First Cause,
Kant is no more friendly to the “cosmological argument” than was Hume to the
“argument from design”.)
71
The Critique of Stone-Age Metaphysics in Russell and Norton
At a later date discussion of the limitations of the notion of causation was
revived by Bertrand Russell at one stage in the kaleidoscopic evolution of his
views, with sharply critical remarks on the notion of cause in an essay of that title,
where he characterizes it as a relic of a bygone age, and indeed of the stone age: it
is not merely that many of our beliefs about what causes what are naive, primitive,
pre-scientific, folk beliefs; it is that the very notion of cause itself is a naive,
primitive, pre-scientific, folk notion.
One can hardly deny that the notion of cause is as ubiquitously employed in
the softer sciences as in everyday life or in the law, but as for the harder sciences,
and above all the hardest of all, physics, Russell was prepared to dismiss the
notion, and this largely on the basis of nineteenth-century developments
concerning energy and fields, even apart from twentieth-century relativistic and
quantum revolutions. Causation simply is not a technical notion of physics. Indeed,
it is not even an everyday notion that has lent its name to a technical notion of
physical science, as with energy and force and power and action and work.
Perhaps the best-known latter-day response to Russell’s skepticism by a
philosopher of physics is John D. Norton’s discussion of what he calls causation as
folk science. After careful consideration, Norton ends with at best a very qualified
verdict and partial rehabilitation:
There is no caloric in the world; heat is not a material substance. However,
in many circumstances heat behaves just as if it were a material fluid, and it
can be very useful to think of heat this way. It is the same with cause and
effect.
Norton’s reconstruction of a restricted notion of causation in restricted areas
72
of physics conspicuously fails — or rather, makes no attempt — to reconstruct any
general notion of cause and effect that would apply between the mental and the
physical in either direction, let alone unify all the sciences by making the search
for relations of cause and effect in their different domains the central task of them
all. A notion of causation that is or would be relevant to our present discussion
simply is not considered in what has perhaps been the most influential qualified
defense of stone-age metaphysics.
And thus Hume, to a degree followed even by Kant, and with them Russell
in one phase or mood, with Norton only partly dissenting, form something of a
critical tradition, not concerned to advocate what is never going to come about, the
banning of causal talk from colloquial speech, but to advocate suspicion of the
notion in hard science and philosophical analysis.
The focus in such a tradition is not so much on rejection of commonsense
beliefs about what causes what, as with the first two dualist replies to the mental
causation argument for neuralism discussed above, but rather on rejection of
extrapolation beyond everyday use of the notion, employing it in metaphysical
speculations, whether in natural theology or in philosophical psychology. But as I
have said, the resources of the critical tradition have not yet been much tapped by
dualists.
Categorization and Identity Criteria
A fourth and quite different dualist reply to the physicalist’s argument from
mental causation is also available, one that turns around the neuralist’s accusation
that the dualist is departing from commonsense. It is a response answering the
claim that it is counterintuitive to deny mental causation of physical events with
the counterclaim that it is far more counterintuitive to identify the mental with the
physical. Physicalism, on this view, is not merely wrong but wrong-headed, a
73
conceptual confusion.
Ironically, I am inclined to apply the term category mistake to the confusion
in question — a term that was the invention of Gilbert Ryle, introduced by him in
connection with his advocacy of analytic behaviorism, a view now wholly
abandoned as being one of the zanier manifestations of old-fashioned materialism
in the last century, no more plausible than John Watson’s analysis of thought as
subvocalization, or Norman Malcolm’s analysis of having dreamt as having
acquired overnight a disposition to tell stories about things that never occurred.
The fourth line of response to which I am alluding is one that has left me
personally never able to take physicalism entirely seriously; and though it is
perhaps not so often met with in print, it has surely influenced the thinking of
many more dualists than just myself.
The nature of the alleged confusion in neuralism is best indicated by
analogy. And so I ask: what is the largest church in the world? An expert on
religious denominations will doubtless cite Catholicism; an expert on ecclesiastical
architecture, Saint Peter’s Basilica. Suppose a child were to overhear these
answers and conclude Catholicism is quite literally one and the very same thing as
Saint Peter’s Basilica. The conclusion is surely utter nonsense. Ultimately what I
will suggest is that neuralism is nonsense of the same kind.
The conclusion that Catholicism is Saint Peter’s Basilica is nonsense, but
why is it nonsense? How could one argue that it is nonsense? The first answer that
comes to mind may very well be one phrased in modal terms, involving the
hypothesis of something happening that hasn’t actually happened but presumably
possibly could happen: a terrorist bomb might possibly reduce the basilica to a pile
of rubble; but if this happened while it was empty, and the Catholic clergy and
laity from Pope Francis on down escaped, shaken a bit perhaps, but with their core
beliefs and practices intact, then Catholicism would survive the destruction of
74
Saint Peter’s, showing the two to be indeed two, not one and the same.
But let us not be too quick here. Granting the possibility of the bomb
scenario contemplated, the basilica reduced to rubble, pontiff and parishioners
alike escaping unchanged, just why would such an occurrence constitute the
destruction of Saint Peter’s but the survival of Catholicism? I do not see how this
question can be answered without bringing in the categorization or classification of
Catholicism and Saint Peter’s, their characterization each as a thing of a certain
sort, with certain identity criteria, telling us where one thing of the given sort is to
be counted as leaving off and perhaps another counted as beginning.
To anyone familiar with the names, Catholicism denotes an institution, but
Saint Peter’s Basilica an edifice. The identity criteria for institutions and for
edifices, telling us how much one can be reformed or remodeled and still itself
remain there, and not a replacement, may not be entirely clear. But surely it is at
least clear enough that reduction to a pile of rubble is enough to make for the
destruction of an edifice, while survival of all the people involved with beliefs and
practices unchanged is enough to make for the continued existence of an
institution.
But now comes the important point. Arguably, once we have these
judgments about classification and criteria in place, the modal claim, and with it
the bomb scenario, may become superfluous. Isn’t the fact that Catholicism and
Saint Peter’s Basilica are things with very different identity criteria, and therefore
very different sorts of things, already enough by itself to establish that they are
very different things? If so, any assertion to the effect that they are quite literally
one and the very same thing must be a piece of nonsense.
Now the neuralist holds that being in pain, as a type, is the same as having
C-fibers that are firing, and that my being in pain now, as a token, is the same as a
my having C-fibers that are firing now. Let me concentrate on the tokens, leaving
75
the reader to think through the case of the types. To the dualist, it seems clear that
my being in pain now is my having a feeling now, and that one doesn’t understand
what the word pain means unless one understands this, and also that my having
C-fibers that are firing now is my undergoing an electrochemical process now, and
that one doesn’t understand what the phrase C-fiber firing means unless one
understands this. These things are as clear as or clearer than that Catholicism is an
institution and that Saint Peter’s is an edifice.
To the dualist, it seems clear that the identity criteria for tokens of being in
pain involve who is having them, and when, and what it is like to have them — are
they dull or sharp? steady or intermittent? to what body-part, real or phantom, are
they referred? — but nothing about physical particles. Likewise it seems clear that
the identity criteria for tokens of undergoing C-fiber firings, involve the motions of
physical particles down to individual electrons, and nothing about what anything
feels like. Indeed, to the dualist these claims seem clearer than any claims one
might make about the identity criteria for institutions or for edifices.
To the kind of dualist I am imagining, the conclusion seems inevitable that
since the identity criteria are radically different, the things are of radically different
sorts, hence are radically different things, so that the neuralist’s assertion that they
are one and the same thing is absurd, like a childish confusion of church in the
sense of religious denomination with church in the sense of house of worship.
76
IV Conceivability versus Possibility
The indirect attack on the supervenience questions by a direct attack on the
mind-body problem fails, or anyhow bogs down, obliging us to take a more direct
approach to the issue of the possibility of a zombie world.
Sorts and Identity Criteria
The principle that different identity criteria make for different things will, to
dualists who accept it, make short work of either physicalistic view in philosophy
of mind. For what has just been said in the preceding chapter against neuralism
would apply a fortiori against the functionalist claim that to be in pain is to be in a
second-order state, since though it may be obscure what are the identity criteria for
being in a second-order state, it seems clear enough that they are if anything even
farther away than are the identity criteria for undergoing electrochemical processes
from the identity criteria for feeling sensations.
In the background here is a view including such components as these:
(1)
Things come in sorts.
(2)
Sorts come with identity criteria.
(3)
For things of the same sort, the identity criteria for things of that sort
determine whether they are — or count as — the same.
(4)
Things of different sorts never are — or count as — the same.
This view is not unquestioned in philosophy, but it can be found invoked in areas
far from philosophy of mind, so at any rate it is not an ad hoc invention of a certain
kind of dualist.
As I have already said earlier, this book is not the place for any extended
excursion into philosophy of mathematics, but it may be mentioned that a notable
77
example of appeal to a view about identity criteria similar to that just expressed in
(1)-(4) is found in Crispin Wright’s Frege’s Conception of Numbers as Objects,
put forward as a solution to the so-called Julius Caesar problem, the answer to
Gottlob Frege’s question how we know that the number two is not Julius Caesar. It
is because the identity criteria for numbers, as set forth by Frege in his Grundlagen
der Arithmetik and endorsed by Wright, are so very different from those for human
beings, Caesar included, that a human being can never count as a number.
Needless to say, antidualists of either kind, adherents of neuralism and
functionalism alike, can be expected to reject the whole foregoing line of thought.
But is there any debatable or discussable issue here? The situation is very largely
one of the dualist in effect saying that the antidualists have misunderstood the very
meaning of the word pain and for that matter of the word identity as well, and of
the antidualists maintaining that, no, they haven’t. And we seem to see here little
more than a mere slinging back and forth of intuitions.
Back to Supervenience
This brings me back to my official topic, the supervenience thesis, which has
been left pretty much in the background for some time now. One reason for interest
in the supervenience thesis is precisely that it does not seem to provoke in the
dualist the reaction that I have been describing in the case of identity claims: the
reaction of immediate dismissal as an absurdity. To claim identity between items
with different identity criteria is one thing, merely to claim necessary correlation is
another. Supervenience therefore seems to promise us a more debatable and
discussable issue to replace mere intuition-slinging. The modal turn seems to offer
a way around an impasse. How far it really does so remains to be examined.
Let us then set aside the issue of identity criteria and turn from the attempt to
settle the issue of supervenience indirectly, by a direct attack on the mind-body
78
problem, to the converse project of a direct attack on the supervenience question,
which might indirectly settle the mind-body problem, at least in part. For a
refutation of supervenience would amount to a refutation of physicalism and a
proof of dualism, while a proof of supervenience would amount to a refutation of
dualism and a proof of physicalism, though the mere statement of the thesis of
supervenience in itself does not favor one form of physicalism over another.
I said earlier that I am ultimately going to question the interest and
importance of the supervenience issue. I will ultimately express three different
grounds for suspicion on this point, and I can now say that the first suspicion I will
express is that the modal turn will not, after all and in the end, give us a way
around an impasse, but rather will prove to be a pointless detour, because in order
to settle the supervenience issue, we will need to come back to the very issues
about identity criteria that I have just temporarily set aside. But such issues will
reemerge only slowly, and there will be a good deal else to discuss before we can
usefully come back to them.
A Trio of Notions of Necessity
The supervenience thesis involves three concepts: the mental, the physical,
and the modal. Under the latter head fall both the concepts of the necessary and the
possible, but by the laws of modal logic these are interdefinable, necessarily being
equivalent to not possibly not and possibly to not necessarily not; and so the two
concepts may be counted as the two sides of one coin. Something has been said
about the scope and limits of the mental in my earliest chapter. Something will be
said about the scope and limits of the physical in a later chapter. I will devote the
present section of the present chapter to some discussion of various flavors of
modality, in order to distinguish the one involved in the issue before us.
It will be well to begin with what has been said about the matter of different
79
varieties of modality by linguists without ulterior philosophical motives. I rely
largely on the Cambridge Textbooks in Linguistic volume on Mood and Modality,
by Frank R. Palmer. The category of mood and modality refers to the distinction
between what is affirmed as actual or real and what is not, insofar as that
distinction is grammaticalized, as one says: represented in language by either
modifications of the main verb of a clause, or the use of special auxiliary verbs.
The term grammatical mood in a narrow sense refers to the first of these means of
expression, and in a broad sense to both.
As for distinctions in verb form, that between what are called the indicative
and subjunctive moods, which still flourishes in related languages such as German,
survives only vestigially in English. As for auxiliary verbs, in English those used to
express mood or modality overlap with those used to express tense or temporality.
They consist of the pairs will and would, shall and should, may and might, can and
could, plus a few others. Notoriously, there are differences as regards the first two
pairs between usage in the United States and usage in South Britain, with further
variation in other Anglophone regions; but fortunately these differences seem to
relate more to temporal than to modal uses.
The various flavors of modality recognized by linguists are expressed by
very different means in some world languages, so there is no question but that a
conceptual distinction can be drawn; but in English the same auxiliaries, and
especially the pair consisting of must for necessity and may for possibility, do duty
for all of them. Palmer recognizes a trio of flavors of modality that he calls deontic
and epistemic and dynamic.
Deontic modality is represented by the contrast between she may go and he
must stay meaning she is permitted to go and he is obliged to stay. Deontic
necessity and possibility thus amount to obligation and permission. This flavor is
not what will be at issue for us.
80
Epistemic modality is represented by the contrast between she must have
seen it and he may have seen meaning given what is known, she must have seen it
and for all that is known, he may have seen it. As knowledge varies from person to
person and time to time, so does epistemic modality. The epistemic possibility of
an afterlife — our not knowing for sure that there isn’t one — is the starting point
for Blaise Pascal in his famous wager. It is also as far as Descartes gets in his
Second Meditation. But it is far short of Descartes’ ultimate goal: this flavor is
again not what will be at issue for us.
Dynamic modality, the remaining flavor, pertains to what I or you or he or
she or they can do or could have done, meaning am/are able to do or was/were
able to do. For our purposes what will be of most interest is an impersonal variant:
what it could have happened that, or what it could have been the case that. Kripke,
uses possibility without distinguishing adjective — simpliciter, tout court, sans
phrase — for this kind of possibility, what could have been, or more precisely,
what is or isn’t but could have been, and necessity without distinguishing adjective
for the correlative notion what is and couldn’t have failed to be.
This omission of any distinguishing adjective is inconvenient when so many
others are using possibility and necessity in so many other senses, and inspired by
an off-hand remark of Kripke’s, in the literature the distinguishing adjective
metaphysical is commonly applied to these modalities. Here I will let
metaphysical be tacitly understood, as the default case, when no other
distinguishing adjective appears, but will sometimes add it explicitly for emphasis
or to avoid ambiguity.
A Trio of Notions of Firm Truth
So much for the deontic and epistemic and metaphysical flavors. Turning
fully from linguistics to philosophy now, there is a traditional trio of notions to be
81
considered there as well, of which [metaphysical] necessity is again one. For as W.
V. O. Quine, a notorious archenemy of modal thinking, writes in his best-known
book:
[P]hilosophical tradition hints of three nested categories of firm truth: the
analytic, the apriori, and the necessary. Whether the first exhausts the second,
and the second the third, are traditional matters of disagreement [Word and
Object, 59].
It is a commonplace that the three notions are of different character.
Necessary as opposed to contingent truth, in the sense intended, has as noted been
called metaphysical, pertaining to being. Aprioriness or apriori truth as opposed to
aposteriori truth may be called epistemological, pertaining to knowledge.
Analyticity or analytic truth as opposed to synthetic truth may be called semantical,
pertaining to meaning. While the epistemically possible was characterized earlier
as what may be for all that is known, the apriori possible or epistemologically
possible may be characterized as what may be for all that is knowable
independently of sense-experience, and the analytically possible or semantically
possible may be characterized as what may be for all that is knowable just from the
meanings of words.
As to matters of traditional disagreement, as one moves from Leibniz to
Kant to Frege to Rudolf Carnap and P. F. Strawson and other contemporaries of
Quine, there is a tendency for the [metaphysically] necessary to be conflated with
the apriori, and then the apriori with the analytic, explained in terms of so-called
semantic rules or linguistic conventions. Quine himself famously attempted to
persuade his colleagues do away even with the analytic-synthetic distinction,
leaving none of the traditional categories standing.
82
Quine did not persuade many, and on the contrary, one distinction formerly
collapsed has since Quine’s heyday come to be restored to respectability, largely
by examples of so-called aposteriori necessities in Kripke’s 1970 Princeton
lectures entitled Naming and Necessity. By contrast, there has been less tendency
to revive the notion of the synthetic apriori.
This leaves us with two degrees of firm truth, the [metaphysically] necessary
on the one hand, and the analytic or apriori — not distinguishing the two — on the
other hand. A notion of firm truth brings with it a correlative soft notion of not
being firmly untrue, and from some points of view it is more illuminating to start
with the soft rather than the firm notions.
A Quartet of Soft Notions
In fact, we will have to take at least brief note of no fewer than four soft
categories. Let me first list them, then explain and exemplify them. I will give each
a one-word label, among which possible will be reserved for the metaphysical
notion, along with an adverb that may be added for emphasis when contrasting one
of the four with another; but I will also give each a two-word label in which the
second word will be possible.
superficially imaginable
= imaginatively
possible
coherently
conceivable
= conceptually
possible
genuinely
possible
= metaphysically
possible
outright
noncontranomic
= nomologically
possible
The superficially imaginable or imaginatively possible is that which, unlike
a round square or married bachelor, we can at least begin to picture without
encountering anything obviously analytically or apriori false. The coherently
83
conceivable or conceptually possible is that whose falsehood is not analytic or
apriori, not even unobviously so. The genuinely possible or metaphysically
possible is what could have been, that whose falsehood is not a [metaphysical]
necessity, not even one discoverable only aposteriori and synthetically. The
outright noncontranomic or nomologically possible is what could have been
without violation of the laws of nature.
An example of something imaginable but not conceivable would be anything
that can be ruled out by analytic or apriori considerations, but only unobvious ones,
such as a definition arrived at only after a long and difficult dialectic of analysis, or
a theorem arrived at only after a long and difficult deductive proof. A case would
be a counterexample to the result conjectured by Descartes’ contemporary Pierre
de Fermat but only proved by our contemporary Andrew Wiles. A counterexample
is not coherently conceivable because of Wiles’s proof, but is superficially
imaginable because the proof is long and difficult. We can begin, but only begin, to
picture what it would be like for such a counterexample to exist.
The mathematical gadfly Doron Zeilberger once announced on his blog that
just such a counterexample had been found — supposedly involving numbers so
large they took several supercomputers to store — and complained about journal
editors who were refusing to publish an announcement about the discovery.
Reportedly a number of readers failed to note that the date of the posting was April
1, and were taken in for a time. One can envision a play or movie inspired by the
blog post, on the order of David Auburn’s Proof. The movie would show the
excitement and consternation in the mathematical community, and perhaps bits and
pieces of the calculation. What could not be shown would be a complete
calculation, since to show that would amount to showing a counterexample, and
there is none.
Examples of things conceivable but not possible are provided by Kripke’s
84
examples of aposteriori necessity, which we will be examining soon enough. More
precisely, if p is a necessary but aposteriori, not-p is impossible but conceivable,
while if q is impossible but conceivable, not-q is necessary but not knowable
apriori, and so either knowable but only aposteriori or else unknowable.
As I have already said twice earlier, this book is not the place for any
extended excursion into philosophy of mathematics, but mathematics does provide
a putative case. It is often held, on the one hand, that mathematical statements are
necessary if and only if they are true, and on the other hand, that they are apriori if
and only if they are provable. Work of Kurt Gödel, Alfred Tarski, and others on
the relationship of truth to proof in mathematics has suggested to many that there
may be truths that are not provable — not merely unprovable from or relative to
this or that accepted system of axioms, but in some sense absolutely unprovable.
Such an unprovable mathematical truth would be a necessary truth not knowable
apriori. It might simply be unknowable, or it might instead be well supported by
computational verification of special cases and various kinds of heuristic
argumentation, and so arguably knowable aposteriori.
It is even sometimes suggested that Goldbach’s conjecture — to the effect
that any even number greater than four is a sum of two primes — may be an
example of an aposteriori necessity of this kind. For illustrative purposes, let me
pretend it is. This is not one Kripke’s primary examples, but he briefly mentions it
in passing in Naming and Necessity as a case of a kind about which there had been
some amount of discussion even prior to his own intervention.
Examples of things possible but not noncontranomic are arguably provided
by many science-fiction fantasies, such as gravity shields or faster-than-light
spaceships. A zombie world would be another case according to dualists, who
claim it is possible, but accept that it would be contrary to the laws of nature.
The four possibility-like notions come with four correlative necessity-like
85
notions, a quartet of nested categories of firm truth to replace Quine’s trio, as well
as with four associated implication notions, all defined as follows:
p is [insert adverb] necessary if and only if
not-p is not [insert adverb] possible
p [insert adverb] implies q if and only if
if-p-then-q is [insert adverb] necessary
A little thought shows that conceptual necessity works out to be analyticity or
aprioriness, also called — as if there weren’t enough terminological clutter here
already — conceptual truth. The phrases nomologically impossible and
nomologically necessary may be abbreviated to contranomic and nomic.
This has been a lot of jargon, but far from all the jargon to be met with in
reading around on our topics. An unlimited number of other firm necessity-like and
soft possibility-like notions can be found mentioned in the literature, which
fortunately there will be no need to consider here. (In particular, the labels “logical
necessity” and “logical possibility” have been very widely and very loosely and
very diversely used in the literature, so much so as to make them at present quite
serious impediments to understanding. I will avoid them.)
The Status of the Zombie World
Four questions now arise about the existence of zombies, or rather, about the
zombie world hypothesis. Let me list them and what I take to be the answers
offered by dualists and neuralists and functionalists.
(1)
Is it superficially imaginable, like Fermat counterexamples?
Dualist: Yes
Neuralist: Yes
Functionalist: Yes
86
(2)
Is it coherently conceivable, like Goldbach counterexamples?
Dualist: Yes
(3)
Neuralist: Yes
Functionalist: No
Is it genuinely possible:
could it have existed even if only by violation of the laws of nature?
Dualist: Yes
(4)
Neuralist: No
Functionalist: No
Is it outright noncontranomic:
could it have existed without any violation of the laws of nature?
Dualist: No
Neuralist: No
Functionalist: No
I will not say anything more about (1) and (4), on which there appears to be
agreement; they are significant only as foils, for contrast with the disputed
questions. I will devote the remainder of this section and the next section to
question (2), whether the zombie world is conceivable. I will thereafter assume that
it is, as maintained by dualists and neuralists, and begin in the section after next an
extended discussion of question (3), whether the zombie world possible.
Arguments in the literature against the conceivability of zombies include two
kinds, one of which I will quickly dismiss.
By definition, a zombie world would contain an exact duplicate of each
physical object, including each human body, to be found in the world as it actually
is. Therefore there would be for each of us something in the zombie world moving
around and emitting sounds just as we do, and therefore moving around and
emitting sounds just as if it were a conscious subject with beliefs of various kinds.
What I am calling a hardcore dualist will insist that though there undeniably would
be beings with beliefs in this as if or metaphorical sense, there would be no beings
with beliefs in a literal sense. Some arguments against the conceivability of
zombies begin with the contrary assumption that zombies would indeed have
beliefs and knowledge, the same as we do in the world as it actually is. Such
87
arguments a hardcore dualist will dismiss out of hand and at once.
Their very existence is a warning. It seems that whenever we get drawn too
far away from sensation and affect, from red looking like or anger feeling like
THIS, and allow belief (or more generally “intentionality”) to be brought into the
argument, our attention is being slyly misdirected, so that some materialist sleight
of hand can make what is truly distinctive about the mental, conscious experience,
disappear. We have seen an instance of this kind of legerdemain already in talk
about itches as representations.
But I have to admit that though the arguments I am dismissing are not
cogent ad rem, as the traditional phrase has it, not cogent as directed to the issue,
they may be cogent ad hominem, as the companion phrase has it, cogent as
directed to certain opponents, namely, those I am calling softcore dualists. These,
while rejecting any physicalistic account of sensation and affect, incautiously
concede some kind of physicalistic account — perhaps neuralist but more likely
functionalist — of belief.
An Inconceivability Argument Reconstructed
More serious consideration, however, is due to those functionalist arguments
that do not rely on the softcore dualists’ incautious concessions. Let me try to
reconstruct what seems to me the main such argument. Recall now the definitions
of , , ,  associated respectively with pain, its neural correlate, its functional
role, and the second-order state of being in a first-order state that realizes that
functional role, so that functionalism maintains that  is , neuralism maintains
that  is , and dualism maintains that  is neither.
Let now  or Omega be the conjunction of all — or if one is worried about
infinity, don’t say all but rather enough — statements expressible in purely
physicalistic vocabulary that are true in the world as it actually is, and therefore
88
would be true also in a zombie world, if there were such. Let us not for the moment
inquire too closely into the scope and limits of the vocabulary in .
By the very definition of zombie, while it is true that there is something in the
-state in the actual world, there would be nothing in a -state in a zombie world,
if there were such a world. The conclusion the functionalist seeks, as a formal
counterpart of the conclusion that a zombie world is inconceivable, is then this:
(1)  conceptually implies that there is something in the -state.
Towards establishing this conclusion the functionalist now assumes three
things:
(2)  and  and therefore  are physicalistically specifiable.
(3)  conceptually implies every true conclusion about which physicalistically
specifiable states have something in them.
(4)  conceptually implies every true conclusion about which physicalistically
specifiable states realize which physicalistically specifiable roles.
There is room for doubt over (2), since we have neither explicitly formulated
 and  and , nor indicated the scope and limits of the vocabulary of . But let
this pass. (3) and (4) seem hard to question, nor is it easy to question the next steps:
(5)  conceptually implies that there is something in the -state.
(6)  conceptually implies that the -state realizes .
(7)  conceptually implies that the -state and -state are coextensive.
(8)  conceptually implies that there is something in the -state.
Here (5) follows from (2) and (3), (6) follows from (2) and (4), (7) follows from
89
(6) and the definition of , (8) follows from (5) and (7).
Next, the main thesis of functionalism — of the kind I have been considering
— applied to the pain case, is the following:
(9) It is a conceptual truth that the -state is coextensive with the -state.
From (8) and (9) we at once get (1), Q.E.D.
Thus the route from the functionalist’s main thesis — that there is a
conceptual connection between being in pain and being in a certain second-order
state, which is claimed to be physicalistically specifiable — to the inconceivability
of zombies is fairly straightforward, as things in this area go. Despite there being
some room for doubt elsewhere in the argument, the antifunctionalist will
presumably target first of all (9), the main thesis of functionalism.
I have already indicated some of the reasons why dualists and neuralists alike
reject that main thesis, and I have already announced a policy of giving less
attention to functionalism than to neuralism as an alternative to dualism. In
accordance with that policy, I will pursue the issue no further here, and assume in
what follows that, as agreed by neuralists and dualists alike, the zombie world is
conceivable. The issue now is whether, granting it is conceivable, it is possible.
Conceivable versus Possible and Nomic versus Necessary
When T. H. Huxley heard of Charles Darwin’s theory of evolution by natural
selection he is reported to have said, How very stupid of me not to have thought of
that! But, of course, Huxley was not stupid, Darwin was a genius. Often
discoveries of genius come to seem obvious by hindsight, once they have been
pointed out. Kripke’s examples of conceivable impossibilities have something of
this obvious-after-the-fact character for many. But the failure of figures from
90
Leibniz to Kant to Frege to Carnap and Strawson and beyond to notice such
examples was not mere stupidity, and successive figures in this tradition certainly
had an excellent motive in seeking to replace the notion of necessity ultimately by
a notion of truth by virtue of semantic rules or linguistic convention.
The problem they were in effect addressing is expressed by Kant early in the
second edition of the Critique of Pure Reason, by saying that experience teaches
us that a thing is so and so, but not that it cannot be otherwise. Experience is our
main source of knowledge, but it seems a mystery how it could tell us that some
truth is not merely true but necessarily so. I will come back to this mystery later,
and the attempts to demystify it.
For the moment it will be enough to say that the view that necessity reduces to
semantic rules or linguistic convention seemed to offer an answer, namely, that our
knowledge of necessity is acquired in the course of learning our language,
internalizing its rules and conventions. The view that necessity reduces to rules and
conventions we pick up in learning to speak, that as the saying goes there is no
necessity but verbal necessity, was attractive because it seemed to clear up a
mystery about modality.
When this demystifying collapse of [metaphysical] modal notions into
epistemological notions was undone by Kripke — for that was, indeed, the effect
of his lectures on Naming and Necessity — the question of how we can come to
know of something that is, that it could not have failed to be, and inversely, of how
we can come to know of something that isn’t, that it could have been, was
reinstated with full force.
Part of the question is how we can ever advance from knowledge of the
conceivability that p to knowledge of the possibility that p, and the question
connected with the mind-body problem that Kripke’s work opens up is of this kind,
and a case in point: how can we hope to advance from knowledge that zombies are
91
conceivable to the knowledge that they are possible? That is how the neuralist
would put the challenge of the supervenience question to the dualist.
The dualist has a corresponding question to put to the neuralist: how could we
hope to advance from knowledge that zombies would be contranomic to knowledge
that they are impossible, or how could we hope to advance from knowledge that
mental-neural correlation is nomic to knowledge that it is necessary? Note how
these questions evade direct engagement with the issue of identity, yet still seem to
get at the heart of what is at issue in the supervenience debate.
The Burden of Proof
Which question one finds most salient, how to get from conceivable to
possible or how to get from nomic to necessary, depends largely on which side one
thinks bears the burden of proof; but in the current climate of philosophical opinion
the onus probandi is very often assumed, with little argument or none at all, to lie
entirely on the dualist side.
Now even the coextensiveness of  with , to say nothing of their alleged
identity, is surely an empirical, aposteriori, synthetic discovery. It is not a
conceptual, apriori, analytic truth that everything mental has a physical correlate at
all, and that the soul of the sleeper is not capable of slipping away from the body
and engaging in independent travels to a spirit realm beyond the material, as for
ages many believed.
Nor is it a conceptual, apriori, analytic truth that if the mental has a physical
correlate, that correlate must be in the body, as opposed to the body’s being a
remote terminal, while thinking goes on in a cloud.
Nor is it a conceptual, apriori, analytic truth that if the mental has a bodily
correlate, it is in the brain rather than the heart or liver or spleen or kidney, as
thinkers as astute as Aristotle have thought and taught.
92
But these comparatively uncontentious observations only establish
conceivability of the failure of mind-body correlation, and for that matter more in
the minds-without-bodies direction than the bodies-without-minds-direction. I have
been taking for granted also since the end of the preceding section conceivability
also in that latter direction: conceivability that everything neural or encephalic, and
more generally all bodily states and processes, and more generally still the whole
material and physical world, could have been just as it is, yet mental life absent.
But once Kripke has convincingly argued that there can be things conceivable
yet impossible, the concession that the zombie world is thus coherently
conceivable is no longer enough for the purposes of the dualist, who wants to claim
it is genuinely possible. Two substantial bodies of work by authors writing from a
point of view at least partially sympathetic some kind or other of dualistical view
are available, both relevant to how the dualist might respond to the challenge of the
gap between conceivability and possibility.
On the one hand, there is the work of Kripke himself, especially towards the
end of the third lecture on naming and necessity, and in the supplementary material
added in the form of footnotes and an afterword when the transcript of the audio
tape of the lectures was published. The material in the lectures generated an
enormous amount of discussion; the supplementary material, a fair amount, but not
so much. The Saul Kripke Center at the City University of New York has many
further audiotapes in its files, awaiting transcription and editing for publication, but
whatever additional relevant material, if any, may be in them has as yet played no
role in debate over the mind-body problem. In the absence of substantial new
published material on that problem from Kripke, discussion of his work has
somewhat subsided, and the work itself has tended to be eclipsed for many by
newer work of others.
On the other hand, there is the work of the enormously prolific Chalmers, the
93
foremost among these others, going back to his influential book on the conscious
mind, a quarter-century after Kripke’s famous Princeton lectures. His personal
webpage [http://consc.net/chalmers/] is a hub for information on the topic, with
links not only to his own work published and forthcoming, but to that of many of
his most important critics, along with his replies.
Jackson’s work must also be mentioned. For his position as of around the time
of Chalmer’s book was dualist — there has since been a falling away on Jackson’s
part — and similar to that of Chalmers; not identical, but similar enough that in
short discussions in the literature often little distinction is made between the two.
One point of agreement between Kripke and Chalmers seems to be — and I
infer from personal communications that Jackson agrees, even now — that
however convincingly Kripke may have argued that there are examples of
conceivable impossibilities, his arguments do not provide any general license to
posit new instances of the phenomenon at will, wherever it would suit a
philosopher’s convenience or confirm a philosopher’s prejudices.
To say this is as much as to say that the burden of proof does not, after all, lie
entirely on the dualist side. And the efforts of all three of Jackson, Chalmers, and
Kripke can be viewed, and are perhaps best viewed, not as efforts to make a knockdown case for dualism, but rather as substantial attempts to shift the burden of
proof to the antidualist side.
How the Kripkean Examples Differ from the Zombie Example
From this point of view the approaches of all three all be viewed
schematically as each involving a pair of contrasting claims:
(1) Every Kripkean example p of an aposteriori necessity has a certain distinctive
feature K.
94
(2) Neuralism’s example p* of an alleged aposteriori necessity lacks the given
distinctive feature K.
Here for present purposes p* may be taken to be simply the claim that -state is
coextensive with the -state. The K feature, as I will unimaginatively call it,
differs with the different approaches.
The combined effect of (1) and (2) is to argue that the neuralist’s example is
unlike the Kripkean examples of aposteriori necessities, thus implicitly challenging
the neuralist to explain how it is nonetheless supposed to be a case of that curious
phenomenon, and thus putting the ball in the opposition’s court.
A complication here is that, while I have been and will be calling the
Kripkean examples cases of aposteriori necessity or conceivable impossibility,
Chalmers and Jackson in a sense only half-accept that the examples are correctly
characterized by such expressions: they make certain distinctions that Kripke does
not, and so recognize certain ambiguities that Kripke does not, resulting in a split
decision about just how the Kripkean examples would be best described. And
unfortunately, an enormous amount of highly contentious apparatus (“twodimensionalism”) is involved here, at just this point, and the expositor therefore
faces a dilemma.
On the one hand, an attempt to describe the Chalmersian and Jacksonian
arguments that ignores their fairly elaborate apparatus cannot be wholly faithful to
the viewpoint from which the arguments are given. On the other hand, any attempt
to deal with the contentious issue of the status of that apparatus in a book no longer
than this one must fall far short of complete adequacy.
The dilemma arises regardless of the expositor’s personal opinion about the
apparatus, though for one who like myself is skeptical, the first horn must be the
more attactive. And so I will impale myself on it: I will attempt an apparatus-free
95
sketch of an un-Kripkean dualistical argument, inspired by Chalmers and Jackson,
especially the latter; but I acknowledge now in advance that criticism of the
argument to be sketched cannot be taken to be directly criticism of the overall
position of either writer from whom I take inspiration.
Atomic Numbers
The un-Kripkean dualistical argument I want to sketch involves a pair of
claims about a K feature, as described above. One claim is that in every example
where an aposteriori or synthetic truth p is claimed by Kripke to be metaphysically
necessary, we find that though indeed p is not conceptually necessary, it is a
conceptual implication of the premise , the conjunction of the complete
physicalistic description of the world as it actually is.
Thus the K feature involved is supposed to be the following:
(1)  conceptually implies that p.
Most of the work will have to go into showing that (1) holds for the various
Kripkean examples of aposteriori necessities.
The other claim is that the neuralist’s key example, the statement the neuralist
wants to claim is also an aposteriori necessity, the coextensiveness of the -state
with the -state, is neither a conceptual truth nor a conceptual implication of . In
other words (1) does not hold for p = the neuralist’s key example p*. Actually, this
last we have in fact already been in effect assuming since we dismissed the
functionalist claim that zombies are inconceivable.
Putting the two claims together, it follows that the neuralist’s key example,
the coextensiveness of the -state with the -state, does not share the key feature
of the Kripke examples, which was to be proved.
96
Now, as will be seen when I start listing Kripkean examples below, none of
them involves overtly mentalistic vocabulary, while the neuralist’s alleged
example involves the -state or being in pain. Thus we know simply from
inspection of the examples that the neuralist’s example differs in one respect from
the Kripkean examples. The un-Kripkean argument just outlined will not tell us
any more than this if it is too generous about what kind of vocabulary it allows to
appear in , about which I have only said so far that it is supposed to be
physicalistic. For if we take that to mean simply, free from overtly mentalistic
terms, each Kripke p will count as a truth expressed in physicalistic terms, and
hence will be a conjunct of , making (1) trivial.
We will only get a substantive result out of the un-Kripkean argument if we
assume the vocabulary in  must be something that could rightly be called basic
physical vocabulary in some sense. Let us do so, and examine the main claim of
the un-Kripkean argument with this point in mind. To see what is at stake there, we
need to look at some more typical and illustrative Kripkean example than the
somewhat special case of Goldbach’s conjecture, the only case explicitly described
so far.
One class of examples much emphasized by Kripke and subsequent
discussions of his work is provided by the atomic numbers of chemical elements.
Kripke takes the following to be an aposteriori necessities, for instance:
(2) Tin or Sn is the element with atomic number 50.
(2') Zinc or Zn is the element with atomic number 30.
(2'') Copper or Cu is the element with atomic number 29.
The un-Kripkean agrees with Kripke that these are not conceptual truths, and wants
to claim they are, however, conceptual implications of . Let me take up these two
97
points in order.
First, as to (2), say, not being a conceptual or apriori or analytic truth, not
being a consequence or part of the very meaning of tin, one should not be misled
by the fact that a present-day dictionary might include the information (2) in its
entry for the word tin. For by so doing, the dictionary is stepping out of the role of
a lexicon and into the role of an encyclopedia, and providing nonlinguistic
information. Ignorance of (2) surely is not linguistic ignorance.
Mendeleyev published his version of the period table, from which his fame
and assignment of atomic numbers derives, in 1869, but the present atomic
numbering of elements, taking into account such noble gases as neon, dates only
from the 1890s, when they were discovered. Charles Dickens died in 1870, after a
career of several decades producing a corpus of literature that, according to on-line
concordances, uses the word tin five dozen times. If ignorance of (2) were
ignorance of the meaning of tin, we would have to conclude either that Dickens
didn’t know what the word meant, which is absurd, or that the meaning of the word
has shifted between his time and ours. But dictionaries themselves implicitly
testify against this latter assumption, since it is their custom, when a word changes
meaning over time, to list the successive meanings by separate numbers under the
word, and this is not done in the case of tin. No other characterization of tin based
on microscopic features can be a conceptual truth, either, for similar reasons:
Dickens didn’t know about them.
Second, as to whether (1) holds for p = the tin example (2), we may begin by
noting that what we are calling basic physical vocabulary may presumably be
taken to include talk about atomic numbers and other microscopic structure. It had
then better not be allowed to contain the word tin, lest (2) become a conjunct of ,
and (1) a triviality.
98
Tinny Stuff
The basic physical vocabulary may, by contrast, presumably be taken to
include also talk of the macroscopic spatiotemporal distribution of substances with
different microscopic structures. So the following, if true, should be a conjunct,
hence a conceptual implication, of :
(1) The element with atomic number 50 is the second ingredient in the alloy most
often produced on the surface of this planet in the period from 5000 to 3000
years before the present.
I take it (1) is indeed true, since tin is the second ingredient after copper in bronze,
and the Bronze Age began about 3000 BCE with the Iron Age following about
1000 BCE. For the atomic number of tin to be a conceptual consequence of , it
would thus be enough for the following to be a conceptual truth:
(2) Tin is the second ingredient in the alloy must often produced on the surface of
this planet in the period from 5000 to 3000 years before the present.
Any other characterization would do as well as (2), if only some such
characterization could be claimed to be a conceptual truth, or part of the meaning
of tin, something it seems hard to claim for (4) itself. The general form would be
this:
(3) Tin is the tinny stuff in the vicinity of here and now.
Here tinny stuff is a placeholder for some characterization in the vocabulary of ,
identificatory of tin. Just here, however, we come to the key disagreement between
Kripke on the one hand and un-Kripkean dualists on the other.
99
The Bearing of the Theory of Reference
The un-Kripkean needs there to be, at least for each speaker on each occasion
of use, some uniquely identifying description of tin, represented here by the
placeholder phrase tinny stuff, associated with the word tin as something very like
a meaning (in the jargon of the Chalmers-Jackson apparatus, this would be a
“primary intension” as opposed to a “secondary intension”). Kripke, by contrast,
denies there is any uniquely identifying description of tin associated with the word
tin as its meaning, or in any way that would require a speaker to associate the
description with the word in order to qualify as a competent user of that word.
The disagreement here extends to two large classes of expressions: first, such
phrases as Murray Gell-Mann or Steven Weinberg, generally called proper nouns
or proper names; second, such words as tin and zinc, or whale and tiger,
belonging to a subclass of what are generally called common nouns, which
subclass are variously called common names or natural-kind terms. And so we are
led from philosophy of mind deep into philosophy of language.
The relation that the names Murray Gell-Mann and Steven Weinberg and zinc
and tin bear to Murray Gell-Mann and Steven Weinberg and zinc and tin is
reference, a.k.a. denotation or designation. Since the nineteenth century, this
notion has been contrasted with what is called sense, a.k.a. connotation or
meaning. Two descriptions may have the same reference but different senses: they
may describe the same thing, but in different ways, as with this pair:
the main element used to fill lighter-than-air balloons
the element with atomic number 2
where the reference in both cases is to helium, but the senses convey different
information about it. The terms sense and reference come from English
100
translations of Frege, and Kripke’s main target in his famous lectures is a view,
descriptivism, associated with Frege, according to which names, both proper and
common, have uniquely identifying descriptive senses.
Some Kripke-inspired critics of the Chalmers-Jackson apparatus, such as
Alex Byrne and James Pryor, in effect take its advocates to task for failing to
appreciate the significance of Kripke’s arguments — especially those that turn
more on epistemic than modal considerations — against descriptivism, or of failing
to appreciate the similarity of the central notion of their apparatus to Frege’s notion
of sense. Defenders of the apparatus respond, attackers rebut, defenders rejoin, and
a large literature results, none of which will I directly discuss, though my own
discussion will be influenced by bits and pieces of it.
I have myself written a book-length account of and commentary on Kripke’s
work, Saul Kripke: Puzzles and Mysteries, with an extended discussion of Kripke’s
antidescriptivism. So I as much as anyone, and more so than many, am dauntingly
aware of how difficult it can be to try to isolate from a large body of material the
aspects most relevant to any particular philosophical issue. Daunted though I am, I
must make the attempt, and will do so next.
101
V
Names versus Descriptions and Minds versus Bodies
Kripke presents his antidescriptivism first for proper names, then for
common names, and then discusses applications to metaphysical questions
including the mind-body problem. I will follow suit in my summary here.
Modal Arguments against Descriptivism
Kripke has two kinds of arguments which may be called the modal argument
and the epistemic arguments, the latter including arguments from ignorance and
arguments from error. In connection with the epistemic arguments Kripke
acknowledges overlap with independent work of Keith Donnellan — and of
precursors back to the late 1950s, though these only saw a need to amend
descriptivism, not to abandon it. In connection with the extension from proper to
common names Kripke acknowledges overlap with independent work of Putnam.
Both the modal and epistemic arguments are in the first instance directed
against the thought that a proper name C might be associated with a uniquely
identifying description of the form the fulfiller of condition , say Aristotle with
the teacher of Alexander, or Murray Gell-Mann with the particle physicist who
developed the theory of strangeness.
The modal argument is that if the name and the description had the same
meaning, then this:
(1)
It could have happened that Aristotle did not teach Alexander.
would mean this:
(2)
It could have happened that the teacher of Alexander did not teach
Alexander.
102
which at best is ambiguous between a true and a false, indeed absurd, reading:
(3)
Someone taught Alexander but it could have happened that he did not teach
Alexander.
(4)
It could have happened that someone taught Alexander but did not teach
Alexander.
But in fact (1) is unambiguously true: Aristotle might have taken up his father’s
profession of medicine and never gone into teaching at all.
In speaking of an alternate course history might have taken, say one in
which some rival of Aristotle got the job of teaching Alexander, the description the
teacher of Alexander might be read as referring to the person who in the actual
course of history did teach Alexander, which is to say, Aristotle, giving reading
(3), or the person who in the alternate course of history would have taught
Alexander, which is to say, the rival, giving reading (4). But the name Aristotle
refers to Aristotle whether speaking of the actual or the alternate course of history:
as is said, the reference of the name is rigid, while the reference of the description
is flexible. And that should be enough to show they are not synonymous.
Epistemic Arguments against Descriptivism
The epistemic argument is that one can be able to use the name Gell-Mann
to refer to Gell-Mann even if one can provide no true uniquely identifying
description of him. All X may be able to say is that
(1)
Gell-Mann is a famous physicist.
which does not distinguish him from Steven Weinberg, while all Y may be able to
say is that
103
(2)
Gell-Mann is the physicist who unified the electromagnetic and weak forces.
which is false of Gell-Mann but true of Weinberg. Yet X and Y, for all their
ignorance and error, are both referring to Gell-Mann, one saying something true
but uninformative about him, the other saying something false about him.
So even if we allow that the meaning of a name may vary from person to
person and time to time, no uniquely identifying descriptive meaning need be
associated with the name by a given person at a given time. Nor is there any reason
to suppose X or Y has some implicit knowledge that might be brought out by a
dialectic of analysis: (1) and (2) may be all they have ever heard about Gell-Mann.
Exactly similar considerations apply to common names: X may only be able to say
that tin is a grayish metal or semimetal often used for plating or in alloys, which
does not distinguish it from zinc. Y may only be able to say that tin is the second
ingredient after copper in the alloy brass, which is false of tin but true of zinc.
If an associated uniquely identifying descriptive sense does not determine
what person or substance or other item a proper or common name refers to, what
does? Kripke offers a picture of an historical chain of communication, as he calls
it, on which an item is picked out — say by pointing at a person, or at a sample of
a substance, or whatever — and a name assigned it by a so-called initial baptist,
and then the name is passed from speaker to speaker, each adopting it with the
intention of using it to refer to what the speaker from whom the name was adopted
was using it to refer to. No uniquely true noncircularly identifying information
need be passed on along with the name.
Descriptivist Responses
Against the modal argument it is often suggested that we could take Aristotle
104
to mean, not the teacher of Alexander but the person who actually did teach
Alexander, by inserting actually — or rather, by making the verb explicit, and
putting it in the indicative mood — and rigidifying the description, as it is called.
This knocks out one reading of It could have happened that Aristotle did not teach
Alexander leaving only the reading on which Alexander denotes the same person in
the hypothetical contrary-to-fact situation as in the actual, factual situation.
Against the epistemic argument it is sometimes suggested that there is one
thing X and Y could say that would be uniquely true of Gell-Mann.
(1)
Gell-Mann is the bearer of the name Gell-Mann.
One really should add here as used in the present discussion, or something of the
sort, since presumably Gell-Mann has relatives who share the family name; but let
the phrase as used in the present discussion be tacitly understood, and grant that
(1) is uniquely true of Gell-Man.
Still it does not identify him. Simply knowing that everyone associates with
the name Gell-Mann the description (1) cannot suffice to identify who Gell-Mann
is, anymore than the inhabitants of an island can support themselves by, as the
phrase goes, taking in each other’s washing. The adherent of the metalinguistic
theory, as the view that a name N means that which bears the name N is called,
will need some account of what determines the reference of names, if not a
descriptive sense associated with them. But there is no obvious reason why such an
adherent cannot simply appropriate Kripke’s historical picture.
A little thought shows that rigidifying by inserting actually does not help
with the epistemic problem: if one has an identifying description, rigidification
may solve the problem about its denoting the wrong thing when speaking of other
worlds; but for this one cannot be so far sunk in ignorance and error as to lack an
105
identifying description to rigidify. A little thought also shows that going
metalinguistic does not help with the modal problem, since a different person
might have borne a given name than the one who actually does bear it.
If we put on both patches at once, however, we are left with one surviving
candidate as the potential descriptive meaning of Aristotle or Gell-Mann or tin,
thus:
the individual who actually bears the proper name Aristotle
the individual who actually bears the proper name Gell-Mann
the substance that actually bears the common name tin
There are still problems, but I will end my summary of Kripke and his critics
on reference here, and return to our zombies.
The Bearing of These Issues on Zombies
At the end of the last chapter we were left with the question whether tin
admits a definition that, when added to a complete description  of the world in
basic physicalistic terms, would imply that tin has atomic number 50. Our recent
discussion leaves only one candidate definition standing: that tin is the kind that
actually bears the common name tin, or more simply:
(1)
Tin is what tin refers to.
What we now want to know is whether, when (1) is added to , the
conclusion is conceptually implied that tin is the element with atomic number 50.
The answer depends on whether, when conceptual truths about reference are added
to , the following is implied:
106
(2)
Tin refers to the element with atomic number 50.
We may suppose  to include or imply information about what kinds of
sounds are emitted by what kinds of organisms where, and so on. The question
thus is whether when conceptual truths about reference are added to such
information, (2) follows.
That depends ultimately and crucially on whether some kind of physicalistic
analysis, perhaps neuralistic but more likely functionalistic, of the notion of
reference can be produced — or can plausibly be supposed to exist already,
implicit but psychologically real, in speakers.
Kripke himself is adamant that his picture is only a picture, and certainly not
a reductive analysis permitting reference to be defined in physicalistic terms.
Could something more be made of the historical chain of communication picture
than Kripke himself does or thinks can be done?
If so, there may be some hope of sustaining a physicalistic analysis of
reference; if not, the un-Kripkean line of dualistical argument seems to falter. That
is perhaps as far as the issue can be followed without getting deeper into
contentious issues about theoretical apparatus, Kripkean and ChalmersianJacksonian.
I will only remark that, if the un-Kripkean wants to adopt, say, a
functionalist analysis of reference, it will probably not be easy to reject a
functionalist analysis of belief. To exhibit any weakness towards functionalist
analyses of belief — what I call being softcore — is to provide an opening and
even an invitation to counterarguments turning on the assumption that zombies can
have beliefs, as I have in effect noted earlier. There is one well-known such
challenge from Katalin Balog, but I will not pursue it, and possible softcore
responses to it, here.
107
Back to Kripke
Rather than pursue further objections to the un-Kripkean dualistical approach,
I will turn now to Kripke’s own discussion, which as I said has been somewhat
eclipsed by such newer developments.
But first I should review a wider range of his examples of aposteriori
necessities:
(1) Bronstein and Trotsky are the same person.
(2) Gorse and furze are the same plant.
(3) Tungsten and wolfram are the same substance.
(4) Nero was descended from Augustus.
(5) Whales are descended from therapsids.
(6) The Colossus of Rhodes was made of copper among other ingredients.
(7) Ethanol is made or consists of carbon among other ingredients.
(8) Light consists of electromagnetic radiation.
Here (1), (4), and (6) involve proper names of individuals, the rest common
names of kinds of creatures or substances. (1)-(3) involve identity, (4)-(5)
pedigree, (6)-(8) constitution. The tin example can be construed as another
constitution example, the atomic number of an element being now understood as
the number of protons among the component particles of each of its atoms.
All of (1)-(8) are widely accepted to be nonconceptual, synthetic, aposteriori,
empirical discoveries; there is some dissent about the identity examples on the part
of Scott Soames, Nathan Salmon, and others, but I will leave that issue aside, being
more interested in the other cases.
The general form of the examples is C is R-related to D, where C and D are
108
names, proper or common, and R the predicate for some relation: identity or being
the same as in some, being descended from in others, being in part made of in yet
others.
It has been granted that it is aposteriori, an empirical discovery, that C is
R-related to D, and Kripke wants to claim that it is also necessary:
(9) It could not have failed to be the case that C was R-related to D.
Kripke claims that on his views about naming, (9) reduces to the following, which
does not involve names or descriptions at all:
(10) If two items are R-related, they could not have failed to be so.
And he claims that (10) is true in the case of the identical-with and descended-from
and made-in-part-of relations.
For identity this is indeed almost immediate. To say two items are identical is
to say that they are not two but one, and to ask whether they could have failed to
identical is thus to ask whether that one item could have failed to be identical with
itself. But no item can fail to be identical with itself. For the other relations,
additional intuitions have to be evoked and invoked, to which I will return in due
course.
It is significant that descriptivists replying to Kripke generally accept the
reduction of the necessity of (9) to (10) and the truth of (10) and hence the
necessity as well as the aposterioriness of all the Kripkean examples. What they
claim is that Kripke’s views about naming are not needed to achieve this result, but
that all this can also be explained on their descriptivist principles. This is an issue
that need not concern us here.
For us, the important point is that Kripke’s necessity claims are not widely
109
disputed, at least not compared to what is for us his crucial nonnecessity claim, that
the - and -states could have failed to be correlated, or to revert to the usual
pretence, that being in pain and having C-fibers that are firing could have failed to
be correlated.
The neuralist wants to add to the list aposteriori necessities the following,
which Kripke insists is not like his examples:
Being in pain consists of having C-fibers that are firing.
And how are the Kripkean examples supposed to differ from this new,
antizombie example? What K feature is supposed to be present in the former and
absent in the latter?
There are two answers in Kripke, one explicit, the other implicit. The
Kripkean examples all concern impossibilities with a tendency, if looked at the
wrong way, to appear possible. In his third Naming and Necessity lecture, Kripke
explicitly offers an explanation for the illusory appearance of possibility in his
examples, an explanation he explicitly denies would be available in the zombie
case.
In his notes and addenda one can find implicit hints towards an explanation of
the necessity in Kripke’s examples, one that again would not be applicable in the
zombie case. I will take these both up in turn.
Our Illusion of Possibility
Kripke in the Naming and Necessity lectures tries to explain how the necessity
of examples of the kind he discusses could have been overlooked for so long, and
why when he presented such examples in seminars in the mid-to-late 1960s there
was usually some initial resistance to them. The kind of example he most closely
110
considers is perhaps best represented by light consists of electromagnetic radiation
— never mind what electromagnetic radiation itself consists in, whether waves or
particles or both or neither.
The reaction he wants to diagnose and explain is that represented by those
who on first hearing him protested: but surely light could have turned out to be, or
could have been discovered to be, something else. Kripke’s view is in effect that
one can only discover what is there to be discovered, and since light does not
consist of anything but electromagnetic radiation, it could not after all have been
discovered to consist of something else.
He is not content, however, simply to say that could have turned out is an
idiomatic expression of some epistemic rather than metaphysical modality, or that
because English uses very similar forms of words to express epistemic and
metaphysical modalities, it is easy to confuse them.
Rather, he attempts to explain what he considers the mistaken feeling that
light could have consisted of something other than what it does, by positing that
the world really could have been such that there were creatures in it in exactly the
same epistemic position in which human beings found themselves before certain
discoveries — creatures having like us eyes and vision and something needed for
vision they called light, streaming down to their planet from the star around which
it revolves — but in which what was called light was not electromagnetic
radiation.
The Mistake to Be Avoided
The easy-to-make mistake here would be to conclude that the world described
is one in which light is not electromagnetic radiation, while the correct assessment,
according to Kripke, is that it is one in which light is used to denote something
other than light. The subtle point about reference here is that with a word like light,
111
even when we are speaking of a nonactual possible world, the word as we use it
denotes the very same stuff it denotes when speaking of the actual world: even
when we are speaking of a way the world isn’t but could have been, the word as
we use it denotes the very same stuff it denotes when speaking of the world the
way it actually is —and not the stuff it would be used to denote by the inhabitants
of the world being spoken of. That is what so-called rigidity as opposed to
flexibility amounts to in this instance.
What Kripke claims is that in each of his examples, where some aposteriori p
is necessary, the following condition is met:
(1) The world could have been such that there were in it creatures in an epistemic
state just like the one we were actually in prior to certain discoveries, in
whose mouths the form of words we use to express the truth that p would
have expressed a falsehood instead, because some key term would have
denoted something different in their mouths from what it denotes in ours.
The illusion that p could have been false is explained as a confusion over
what the form of words we use to express that p when speaking of how things
actually are expresses when we use in it speaking of some counterfactual situation.
What the form of words expresses when we are speaking of such a situation is
confused with what it would express when creatures like us were speaking in the
counterfactual situation in question.
The condition or candidate K feature (1) is rather a mouthful, but its content
should become clearer on thinking through the example already given in slow
motion: the world could have been such that there were in it creatures in an
epistemic state just like the one we were actually in prior to modern scientific
discoveries about the nature of light, in whose mouths the form of words we use to
112
express that light consists of electromagnetic radiation, namely, the words light
consists of electromagnetic radiation, would have expressed something false —
because though the creatures would have used light in the same visually observable
circumstances in which we actually use it, what they were using it for would have
been something other than electromagnetic radiation, something that played the
role in their visual observations that electromagnetic radiation plays in ours.
The illusion that it is possible that light could have consisted of something
other than electromagnetic radiation is then explained as a result of misdescribing
the envisioned possibility, using the key term light as if it were flexible when as a
matter of actual fact it is rigid.
To sum up, Kripke’s story is as follows:
Being in an epistemic situation like ours, the hypothetical creatures would be
aware of something playing the role in their lives that light plays in ours, and
would call it light, but in so doing they would be using light to denote
something other than light, which is to say, something other than
electromagnetic radiation of certain frequencies.
It may take some work to think through what the claim that Kripke’s
candidate K feature (1) is not present in the zombie example amounts to.
How the Case of Pain Differs
The neuralist wants to claim that it only seems to be possible that C-fibers
could have been firing without pain being felt, though really it is impossible, owing
to the supervenience of pain upon C-fiber firing. The possibility of C-fiber firing
without pain is, according to the neuralist, an illusion.
How would one attempt to explain this alleged illusion along the lines on
113
which Kripke explains the illusion that light could have consisted of something
other than electromagnetic radiation? The claim would have to be that the world
could have been such that it contained creatures in an epistemic situation just like
the one we were in prior to certain discoveries, in whose mouths C-fiber firing
does not occur without pain was false, because for them, pain did not refer to what
it refers to for us.
The story would have to be, by analogy with that for light, as follows:
Being in an epistemic situation like ours, the hypothetical creatures would be
aware of something playing the role in their lives that pain plays in ours, and
would call it pain, but in so doing they would be using pain to denote
something other than pain, which is to say C-fiber firing.
And this is the claim that Kripke is at pains to deny.
He denies the genuine metaphysical possibility, or indeed even the coherent
conceivability, of a world epistemically indistinguishable from our world, but in
which what was called pain was not really pain. The reason for the denial is that
anything that could legitimately or plausibly be called epistemically
indistinguishable from being in pain would have to be something that could be said
to be felt as being in pain; but anything felt as being in pain is being in pain and
nothing else.
That is because being in pain is having a feeling, and it displays ignorance of
the meaning of the word pain to think it is anything else. To seem to be in pain is
to be in pain. All that glitters like gold is not gold, but all that hurts like pain is
pain. As Kripke has sometimes liked to put it, there can be fool’s gold, but no
fool’s pain.
We have here, however, I’m very much afraid, an instance of what I hinted
114
earlier would be coming: an argument over supervenience that ends up turning on
just the same sorts of considerations — in this case, considerations about
classification, and in particular, about the classification of being in pain as having a
feeling — as the direct dualist argument for the conclusion that neuralism is a
conceptual confusion.
If the direct arguments breaks down in a mere slinging back and forth of
intuitions, plus perhaps some tossing around of contrary assumptions about the
locus of the burden of proof, the indirect argument seems here to be threatened
with a similar breakdown, with the added complication that it involves — or would
involve, in a fuller account that I have been offering — a good deal of contentious
apparatus.
Our Knowledge of Necessity
It is fairly common for metaphysicians today, especially of the school who go
in for complex views about the fancy notions they call essence and grounding, to
be rather unconcerned with questions about how we can know this or that
metaphysical claim is true (questions about the “epistemology of modality”). One
well-known figure has even been quoted as coming out with the slogan
epistemology last if ever, though as I have not been able to find that phrase in print,
I will name no names. For all one could tell from the Naming and Necessity
lectures, Kripke’s attitude might have been similar.
The addenda provided when the lectures were published show otherwise, and
by contrast do take the question how do we know? seriously, though Kripke’s
formulations in the direction of an answer are extremely cautious and heavily
guarded: it is not a matter of presenting a solution, but of presenting what he says
are hints as to what may be a solution.
A less cagey writer might have put the matter differently, directly suggesting
115
that in all the Kripkean examples of aposteriori necessities, it is a conceptual truth
that if p is true, then it is necessarily true. That p is true is an aposteriori, empirical
discovery, but that necessity comes with truth is knowable in advance of scientific
investigation.
The Goldbach Example and Others
Kripke hints towards such a suggestion in connection with the Goldbach
example. Here the conceptual truth would be that mathematical facts always hold
of necessity. But is it indeed plausible to regard this as a conceptual truth? Is it
plausible to regard this as something reflecting a semantic rule or linguistic
convention?
It seems pretty clear what the linguistic rule would have to be: distinctions of
grammatical mood are not applicable to purely mathematical statements. In other
words, purely mathematical facts count as necessary because, or in the sense that,
there is no significant distinction marked by the contrast between could have been
and is and could not have failed to be in the mathematical realm.
Well, this formulation somewhat overstates the rule, which would rather be
that distinctions of mood are not applicable if taken in a metaphysical rather than
an epistemic sense. The best evidence for the existence of such an implicit rule at
work in our language would seem to be the fact that when modal distinctions are
applied to purely mathematical matters, as does sometimes happen, they are
generally interpreted without hesitation as being meant epistemically rather than
metaphysically.
Consider:
(1) Theorem (Euclid): There is no largest prime.
(2) Proof (by reductio): If there were a largest prime N, then…
116
Clearly what the mathematician is saying in (1) and (2) here is not something
metaphysical like this:
You know what the actual mathematical facts about primes are, and in
particular that there is no largest one, since I have just told you; but imagine
an alternate reality in which the mathematical facts are different.
Rather it is something epistemic like this:
I’ve just asserted something, but you don’t know it’s true, because I haven’t
yet proved it; so imagine that, despite my having just asserted it, and my
being generally reliable on this subject matter, it is false.
Similarly, if we are told that possibly Euler’s constant is rational and possibly
Euler’s constant is irrational, we take understand this to mean, not that there are
alternate mathematical realties, in one of which the constant is rational and in the
other of which it is irrational, but merely that it hasn’t been proved to be irrational,
and hasn’t been proved to be rational, either.
Another case, not discussed by Kripke, where a supervenience claim is widely
granted is that summed up in the slogan the moral supervenes on the natural. But a
slight preliminary clarification is needed here. For most who advance this slogan
do not really mean it. Taken literally the slogan would seem or threaten to imply
that supernaturalistic theories of morals — so-called divine command theories and
the like — are not just false but impossible; but while the sloganeers may or may
not believe this, even if they do believe it, it pretty clearly isn’t part of what they
intend to assert in advancing the slogan. Rather, the slogan should be something
like this: the evaluative supervenes on the non-evaluative.
The thought is that whether, say, some act is right or wrong, is completely
117
determined by the features of the act and its situation and consequences and so
forth that would be describable in neutral, value-free terms. If the act and its
situation and consequences and so forth were unchanged in every respect so
describable, its moral status would necessarily remain the same.
Can this be regarded as a conceptual truth, extrapolating from Kripke’s hints?
Perhaps it could be excogitated from close consideration of the very idea of an
evaluation, which seems to presuppose that there is first something there to be
evaluated, before one then evaluates it. But I only put the question, as a test case,
by way of illustration. If this book is not the place for any extended excursion into
philosophy of mathematics, still less is it the place for any extended excursion into
metaethics.
What we really need to consider is what sorts of conceptual truths might be
the ultimate source of the necessity — or anyhow, of our knowledge of the
necessity — in the kinds of examples listed in the preceding section, and whether a
similar truth could serve as a source of necessity in the case of the correlation of
pain with C-fiber firing.
Possible Worlds Jargon
Kripke popularized world talk in his early work on the formal logic of
modality, but had distanced himself from it somewhat by the time of the Naming
and Necessity lectures, though he never switched over entirely from speaking of
multiple worlds to speaking exclusively of multiple states that the world could
have been in — the one and only world there is, setting aside talk of a world to
come as not relevant to the present discussion.
Possible-world talk involves dropping the kinds of distinctions of grammatical
mood we usually make, so that instead of saying such-and-such could have been
[subjunctive] the case if the world had been other than as it actually is, one says
118
such-and-such is [indicative] the case in some world other than the actual one.
The switch raises certain questions that don’t arise without it. For instance, if
instead of saying, I could have taken today off, I say, there is a possible world in
which I have taken today off, a question arises about the relationship between the I
in this actual world, who has not taken the day off, and the I of this other world,
who is taking the day off.
As Ambrose Bierce says somewhere, conception of two myselfs is difficult.
Can I myself really be present both in this actual world and in some other world?
Or must I think that what is there over there is not really me but only a so-called
counterpart or Doppelgänger of me? Or that I am stretched out in a modal
dimension in addition to the three spatial dimensions, with only a slice of me here
and a slice of me there, as when I stand on a bridge over the Delaware River
between Lambertville, New Jersey and New Hope, Pennsylvania, looking down
the river, one side of me may be in one state and the other side in the other?
To take possible worlds talk seriously, as did Lewis, is to take such questions
seriously. Such issues do not arise if one sticks to saying that the way the world is,
I did not take [indicative] today off, but the world could have been such that I took
[subjunctive] the day off. With that usage, there is no question of there being more
than one I, one taking the day off and one not; there is only one I, who actually is
not taking the day off, but who could have done so.
Engaging like Kripke in possible worlds talk only as a picturesque manner of
speaking, without taking it seriously, amounts to being willing to fall back into the
usage that restores distinctions of grammatical mood, and concerns different ways
the world could have been, different states the world could have been in, rather
than different worlds, whenever the possible-worlds jargon threatens to raise
puzzles. I will here, as I claim Kripke in effect does, always consider a question of
the form, is such-and-such an individual who exists in this or that possible world
119
the same as thus-and-so an individual who exists in the actual world? to amount to
would such-and-such an individual who would have existed if the world had been
this way or that way have been the same as thus-and-so an individual who exists in
the world as it actually is? My suggestion will ultimately be that the source of
Kripke’s pedigree and constitution examples must be sought in such questions and
the criteria used to answer them.
Criteria of Transworld Identification: Pedigree
In publishing the transcript of his lectures, Kripke added some footnotes, and
we must turn from the addenda to these if we are to get any help from Kripke with
the questions we were left with at the end of the preceding section. Kripke says
that what appears in some of these notes are words actually spoken in his talks, as
asides. He does not tell us which notes these are, though any that cite literature
from after 1970 must surely be later additions and not asides. The one footnote that
will most interest me (number 57, from which I quote below) is one of these.
Philosophers tend to leap at once to questions about the world having been
different from the beginning of time, and I suppose that one must do so in
connection with the examples about constitution. But with the examples about
pedigree, and others as well, it is enough to imagine running the engine of history
in reverse back to a certain time, and then running it forward again along a
different track. Apropos of this matter, Kripke says:
…ordinarily, when we ask intuitively whether something might have
happened to a given object, we ask whether the universe could have gone on
as it actually did up to a certain time, but diverge in its history from that point
forwards so that the vicissitudes of the object would have been different from
that time forth.
120
This is the method Kripke uses when arguing that the name Aristotle cannot
be an abbreviation for some description of one of the Stagirite’s famous deeds,
such as the teacher of Alexander. We backtrack to Aristotle’s youth and imagine
him going into the family business, rather than pedagogy, and conclude he need
not have become the teacher of Alexander.
In pedigree examples, Kripke notes, in the same footnote from which I quoted
earlier, that the time at which the divergence from actual history occurs may be
sometime before the object itself is actually created. This is what would be at issue
if we asked whether Nero could have had a different mother from the one he did,
or whether that mother, Agrippina the Younger, could have had a different mother,
or whether that grandmother, Agrippina the Elder, could have had a different
mother, or whether that great-grandmother, Julia, could have had a different father
from Augustus. We imagine a situation some time before the scandal-shadowed
Julia’s birth or conception developing in a different direction from the one the
course of history actually took, and resulting in the birth of a certain girl, and we
ask whether she, that baby girl, would have been Julia.
Kripke suggests that really the only obvious things we have to go by is the
gametes from which the zygote that became Julia were formed. If the same egg and
sperm had united, we will say it would have been the same Julia, and if not, not. In
a later work on first-person pronouns, Kripke adds an anti-Cartesian gloss:
If we had a clear idea of the soul or the mind as an independent, subsistent,
spiritual entity, why should it have to have any necessary connection with
particular material objects such as a particular sperm or a particular egg? A
convinced dualist may think that my views on sperms and eggs beg the
question against Descartes. I would tend to argue the other way; the fact that it
121
is hard to imagine me coming from a sperm of egg different from my actual
origins seems to me to indicate that we have no such clear conception of the
soul or self.
When we start at some early stage in the actual life of someone like Aristotle,
and imagine various lines on which history might have moved forward from that
point, we have available criteria to tell us whether this or that person later in the
story, such as the one who becomes teacher of Alexander, would or would not
have been Aristotle. They are the same criteria for identifying persons over time
that we employ when we imagine possible courses events may take over the next
year or decade, and ask whether this or that person later in the story, such as the
one who holds the office of U.S. president in 2020, is or is not Donald Trump.
When we start before the birth or conception of a given individual, criteria are
harder to come up with, and Kripke finds none that go back very far before
conception. The main point for present purposes is that it is on criteria of identity
that he is ultimately relying in his judgments of possibility.
Though it is extrapolating beyond anything Kripke says explicitly about his
examples, if there is a conceptual truth in the background behind those that turn on
genealogical or ontogenic pedigrees of individuals and the evolutionary or
phylogenic pedigrees of species, I would guess it must be something like this:
If an individual or group that actually exists actually had a certain ancestry,
then any individual or group there is or could have been that has or would
have had a different ancestry, does not count or would not have counted as
the same individual or group.
So much, then, for pedigree; it is time to take up what seems the most obscure and
122
difficult case.
Criteria of Transworld Identification: Constitution
Kripke has less to say about what kind of conceptual truth might be behind
aposteriori necessities about constitution, and I am not in a position to make any
exegetical claims about his intentions. So I am extrapolating further than in the
pedigree cases when I suggest that the principle at work may be something like the
following, and that I for one can think of nothing else that it might be:
(1) If an substance that actually does exist actually has a certain constitution, then
any substance there is or could have been that has or would have had a
different constitution, does not count or would not have counted as the same
substance.
Thus if alcohol — I mean ethyl alcohol or ethanol — actually is a compound
of carbon and hydrogen and oxygen in a ratio of two to six to one, then a substance
actually found on some other planet, or a substance that would have been found on
this one if the world had been very different from the way it actually is, is not or
would not have been alcohol if it has or would have had any different chemical
constitution, not even if it induced exactly the same degrees of intoxication with
exactly the same degrees of consumption in the people — or whatever was there in
place of people — who consumed it.
A possible substance in a possible world that had a different chemical
constitution would not be alcohol. Contraposing, turning this around, in any
possible world, a possible substance there that was alcohol would not have a
different chemical constitution from what alcohol has in the actual world. Or in
other words, whatever chemical constitution alcohol actually has — supposing it
123
does actually have a uniform constitution — it has in any possible world where it is
present: it could not possibly have had a different one, it has the one that it has of
necessity. And this is knowable in advance of empirical investigation, simply by
understanding and internalizing that something like (1) is a semantical rule or
linguistic convention of our speech.
To return to the issue of a K feature, supposing the Kripkean examples
generally work as I have suggested, with something on the order of (1) in the
background as a conceptual truth adding an element of necessity to whatever is
empirically discovered, boosting aposteriori truth to aposteriori necessity, the
Kripkean claim would be that nothing like this would be available to raise the
correlation of pain and C-fiber firing to the status of a necessary truth. For the
analogue of (1) in this case would have to be something like this:
(2) If a psychological state that actually occurs actually has a certain neural
correlate, then any psychological state there is or could have been that has or
would have had a different neural correlate does not count and would not
count as the same psychological state.
But no such principle can plausibly be claimed to be a conceptual truth. What
makes psychological states, or anyhow conscious ones, and especially sensations,
count as the same are such factors as who is the subject of them and when the time
of their occurrence and what is their so-called phenomenal character, what it is
like or feels like to be subject to them.
One would have to be grossly mistaken about what makes for sameness of
sensations to think that something like (2) held as a semantical rule or linguistic
convention, making the identity of a state as one of being in pain dependent on
what underlies it physically or neurally — or, I would add, what role it plays
124
functionally as an intermediary between stimulus and response, though having
more or less dismissed functionalism from further consideration some time back
now, let me not elaborate on this point.
Grounds for Disappointment
It is only too obvious, however, that if our last candidate for a K feature is
really what is at work, and this consideration has to be appealed to in order to
resolve the supervenience question, then what has to be appealed to in order to
resolve the supervenience question are exactly the same considerations that would
be involved in a direct confrontation between dualism and neuralism, with the
dualist diagnosing conceptual confusion on the part of the neuralist. If that debate
was at an impasse, the detour through modal considerations has not after all taken
us around it, but rather has brought us right back to it after forcing us to make a
long circuit.
The promise of the supervenience issue to provide a way around or workaround has proved false. Worse, the detour has exposed us to certain dangers. For
we have been exposed all along the way to the threat of becoming bogged down en
route in obscurities about reference or especially about modal apparatus.
The latter threat, at least, I have tried to ward off here mainly by keeping well
back from the quicksand that surrounds the modal route on either side. I have
played down the role of any theory of possible worlds that takes them seriously
and not as mere manners of speaking, resolutely ignoring the debate over the
nature of possible-worlds apparatus between the un-Kripkeans Chalmers and
Jackson on the one hand, and Kripke and supporters like Soames on the other
hand.
This has, indeed, made my account something very much less than a full-scale
survey of the status quæstionis, or what the cognoscenti might have been led to
125
expect from the title of this book; but I hope I have shown enough of what is going
on to explain why I do not want to show more.
The complaint about the supervenience issue that I have just been voicing, to
the effect that it is a pointless detour, is as I hinted earlier only one of three that I
have, and perhaps of the three it is the least serious.
For I will next be suggesting that the supervenience issue has been serving as
a pernicious distraction from the issues that really matter. In the process I will be
suggesting that the most important division among philosophers today on the
mind-body problem is not the one marked by the supervenience thesis, which is the
division between dualists and physicalists.
Rather, it is a division whose dividing line runs through the physicalist side;
only it is not the one division over on that side that I have mentioned so far, the
division between neuralists and functionalists, but another. But to say just what it is
will take some background preparation, filling in a gap in my exposition thus far.
For it is a defect of many discussions of the mind-body problem that they go on
about the distinction or alleged lack thereof between mental and physical without
ever spelling out just what the adjective physical connotes, and my own account
has suffered from this deficiency up to this point.
126
VI The Bounds of Physicalistic Science
In order to arrive at a more definite view of what it is that we are contrasting,
as physical, with sensation and affect, we will need to revisit Descartes.
Academic Administration
The only worse procedure than offering no preliminary account whatsoever
of the nature of the physical is to dispose of the issue quickly and cheaply by
saying that the physical is what is studied by physics, which is to say, by the
science done in physics departments. To take that line is to hold an important
philosophical distinction hostage to the convenience or the whims of academic
administrators. Indeed, it is to do worse with the question than academic
administrators typically do, since they at least often place their physics
departments inside schools of physical science, along with chemistry, astronomy,
geology, and more, thus implicitly recognizing that there is something that may be
called a physicalistic character to a range of sciences extending beyond what is
conventionally recognized as physics proper.
But this character, I would submit, extends well beyond schools of physical
science into the life sciences, embracing not only molecular biology but even
evolutionary biology as well. To know where to draw the line, we need to go back
to the beginning of the period when natural science was separating from natural
philosophy, and physics was ceasing to be the name of a branch of philosophical
speculation, and becoming the name of a branch of scientific research, and to a
certain distinction made not only by Descartes but in one way or another by his
older and younger seventeenth-century philosophical and scientific contemporaries
and successors, notably Galileo Galilei, Robert Boyle, and John Locke.
All in one way or another make a distinction between a list of properties or
127
qualities of things they considered in some sense more objective or real, and a list
of those they considered more subjective or apparent. The distinctive labels, if any,
used for the first list and the second list varied from writer to writer (though
“primary and secondary qualities”, introduced by Boyle and popularized in a
modified sense by Locke, has won out in subsequent discussion). So did the
account, more extended in some writers than others, of the principle of division.
And so did the lists of examples of the two kinds, though generally size and
shape would be mentioned at the top of the first list, and color high up on the
second. The features on the first list were those on which the new science was to
focus, and it is notable that they are all such as to lend themselves pretty
straightforwardly to mathematical treatment. Thus making the division was
important to the new direction science and especially physics was to take over the
centuries to come, a direction that has been mentioned from early on in this book,
but whose full significance we have yet to ponder.
Speaking of Color
No philosopher today can be quite happy with the diction of the various
seventeenth-century worthies in marking their division. Perhaps the language of
color would raise the most objections.
In ordinary language we distinguish between appearance and reality when
using color terms. Thus we may say that what appears a uniform orange patch on
the picture just now coming out of the printer is really made up of tiny red and
yellow dots, as can be seen with a magnifying glass. More generally, we apply the
same color vocabulary when speaking of purely subjective phenomena such as
what we see when we close our eyes in a lighted room — a changing view we have
no inclination at all to take to be a view of anything in front of our eyes, be it a gas
in the room before us, or the back of our eyelids — and when speaking of
128
something objective present even in objects never seen — as with our belief that
our gallbladders are greenish, though most of us never have been nor will be
opened up to show that organ.
This looseness in ordinary language can be a source of confusion, since
when a dualist claims that one cannot really know what the state of seeing red is
like without having been in the state, this may be misunderstood as meaning that
one cannot really know what the state of seeing red is like without having seen an
external red object, whereas it would suffice to have seen an internal red image.
Early modern thinkers tended to regiment the language of color in the
direction of restricting it to the subjective. To do so was a mark of following strict
scientific rather than loose lay usage — or in the terminology of the period, of
following the learned rather than the vulgar. The classic expression of this attitude
is found in Newton’s Opticks (First Book, Part II, Experiment VI, Definition):
And if at any time I speak of light and rays as coloured or endowed with
colours, I would be understood to speak not philosophically and properly,
but grossly, and accordingly to such conceptions as vulgar people in seeing
all these experiments would be apt to frame. For the rays to speak properly
are not coloured. In them there is nothing else then a certain power and
disposition to stir up a sensation of this or that colour.
In putting color on the second list, it was not intended that whatever it is in
external things that is responsible for — to use Newton’s language — their power
or disposition to stir up the sensations of the relevant kind in us should be placed
outside the boundaries of physics. The intent was at most that it was to play no
immediate role, the hope and belief being that it would prove to be some
combination of features on the first list, and as such measurable and
129
mathematically treatable through them.
What was being excluded — and the word is crucially important, for this
was not some inadvertent oversight but indeed a deliberate exclusion, and one
central to the program and project of the new science — when color is set down on
the second list, is the sort of thing I have been referring to when speaking of red
looking like THIS, or what colors look like. That is how the regimented usage must
be understood, despite the fact that many philosophers today, I have said, would
disapprove of such regimentation of color terms towards subjectivity, and some
even prefer a regimentation in the opposite direction.
Speaking of Size and Shape
As for the first list, contemporary critics already noted that with size and
shape, too, we in ordinary language distinguish real from apparent, though with
these features there are not one but two sensory channels through which they
appear, visual and tactile, and the link between sight and touch is not directly
sensed. (This fact gave rise to the seventeenth thought experiment known as the
“Molyneux problem”, only recently apparently empirically resolved.) Now what a
shape looks like or what a shape feels like are really intended to be excluded quite
as much as what a color looks like, and in this respect the placement of shape on
one list and color on another is misleading.
But again to use Newtonian language, whatever it is in external things that is
responsible their power or disposition to stir up visual and tactile sensations of
shape was intended to be directly admitted in the new science in a way that
whatever it is in external things that is responsible their power or disposition to stir
up visual sensations of color or auditory sensations of pitch was not. Behind the
distinction was the perhaps inchoate thought — on which the criticism of the
immaterialist George Berkeley focused — that in the one case but not the other
130
whatever it is in external things could be said to resemble the sensation stirred up.
Descartes’ first or good list consisted of geometric and kinematic features:
size and shape, as already mentioned, and motion, that is, speed and direction of
displacement, at least in a relative sense. Important physical quantities, such as
Newton’s notion of mass, that must be considered along with the geometric in
statics, and along with the kinematic in dynamics were not given their due.
But even Newton’s list would have failed to note some features any presentday physicist would be sure to include. For a version of the first list today would
have to admit many features not recognized until a century or two or three after
Newton, and not accessible through any sensory channel, down to the so-called
charm and strangeness of quarks, all in the end posited to explain what is sensible.
But the subsequent development of physics, though it has admitted many
such unobservable theoretical properties, as philosophers call them, has not
undone the original exclusion of qualities accessible through a single sensory
channel: what colors look like and what pitches sound like, or in jargon, the
phenomenal character of visual and auditory experience.
A science may be called physicalistic, to give this heretofore rather loosely
used expression a more precise sense, to the extent that it resembles physics in this
respect, and has no truck with color or pitch or the like understood as subjective,
phenomenal qualities. And indeed, as I said, not only chemistry and astronomy and
geology are, at least in their theoretical cores, physicalistic in this sense, but so
likewise are molecular and evolutionary biology.
Reduction Not Needed
So far as we know, subjective, phenomenal qualities are experienced only by
creatures with brains. Restricting science to the physicalistic, and ignoring these
qualities, is therefore leaving out something, and indeed something very important
131
— for is it not arguably the locus of all noninstrumental value? — that
distinguishes at least some brainful bodies from brainless ones. It was not at all
unreasonable to expect that, by leaving out something important, one would be
limiting one’s ability to explain even the physicalistically describable motions and
changes in such bodies. Or in other words, belief that there would be exceptions to
what I have called the uniformity principle was not unreasonable.
Indeed, the discovery, if so it may be called, of the uniformity principle, or if
one wishes, the closure principle, must be considered in historical terms a great
surprise, whose implications did not really begin to be appreciated fully until the
late nineteenth century, and are perhaps not fully appreciated even today, as our
earlier discussion of the clash between such principles and what passes for
common sense suggests.
It cannot be emphasized strongly enough that the physicalistic character of a
science in no way depends on its being reducible to physics (in any of the many
senses of “scientific reduction” in circulation). Astronomy and geology, for
instance, are inherently irreducible to physics and chemistry, because they deal
with specific objects or places in spacetime, and not exclusively with universal
laws applicable across the whole sweep of the universe.
Also, chemistry has been physicalistic all along, though right through the
nineteenth century it remained largely a science unreduced to physics. For
instance, the chemical case for atomism was much stronger circa 1900 than the
physical case. The phenomena of tautomerism and optical polarization, for
instance, simply demand to be understood in terms of real molecules composed of
real atoms in a real geometric arrangement. By contrast, in 1900 there were still
eminent physicists, Max Planck for one, who doubted the atomic hypothesis.
Statistical mechanics, and the relationship of the laws of thermodynamics to the
time-symmetric general laws of physics were poorly understood.
132
The real integration of chemistry with physics began only when subatomic
structure began to be explored, eventually to lead to the quantum mechanical
account of chemical bonding, as in Linus Pauling’s famous book. Even today,
however, it is only to an extremely limited degree that the energies of reactions,
say, can be calculated from quantum mechanical first principles, and not merely
measured empirically. Such ab initio calculations, as they are called, require
exponentially more steps of computation as the number of atoms involved
increases, and will likely be forever infeasible for really large systems.
Despite infeasibility in practice, there is nowadays a general conviction of
possibility in principle, but all this is historically speaking a quite recent
development. The integration of biology with chemistry is if anything even less far
along than the integration of chemistry with physics, but this in no way means
biology as presently conducted is any less physicalistic, in the sense of still
excluding what was put outside the bounds of science in the seventeenth century,
what colors look like and what pitches sound like and what odors smell like and
what textures feel like and so on.
Evolution and Insect Color Vision
This physicalistic character extends to evolutionary and not just molecular
biology, since evolutionary theory has always aspired, whether the aspiration was
articulated or not, to maintain a purely physicalistic character, so as to keep open,
as a distant horizon, the possibility in principle of eventual absorption into a
unified physicalistic science. But what physicalistic science excludes from even
being mentioned it cannot take even the first step towards explaining, and
physicalistic aspirations therefore go hand in hand with explanatory limitations on
evolutionary biology. A simple hypothetical example should make this clear, if I
may be allowed for a moment to turn our attention to animal minds, and
133
specifically insect minds.
Suppose we have two closely related species of insects with overlapping
territory, almost indistinguishable in juvenile stages, that as adults develop
different patterns of pigmentation. If the pigmentation tends to make them more
conspicuous to predators, what can be the evolutionary explanation for its
development? Probably that it enables members of the same species to recognize
each other, and so avoid wasting reproductive effort on matings that would
produce only sterile offspring like mules.
But what do the insects look like to each other as they go about trying to find
mates? Not as they look to us, presumably, if as we are told of some other insects,
they are unable to distinguish red and black as we can, but able to distinguish
ultraviolet and black as we cannot. Is their whole spectrum shifted, so orange looks
red, yellow orange, violet blue, and ultraviolet violet? Or does orange look orange,
blue blue, violet violet, and ultraviolet red? Or does ultraviolet look like something
we have never experienced and cannot imagine? Or is it perhaps that, insect brains
being so unlike our own, the little bugs have no visual experience at all?
The answer is that for purposes of the evolutionary explanation, none of this
matters. All that matters is that the different patterns of pigmentation have different
effects on the insects’ visual organs and nervous systems, and ultimately on their
behavior, including their reproductive behavior. The explanation is an entirely
physicalistic one, such as it has always been the implicit ambition of evolutionary
theory to provide. The other side of the coin is then that if there is something the
colors look like to the insects, evolutionary biology cannot tell us why it is there or
what it is like or why it is like that. To sum up in a single slogan: there can be no
evolutionary explanation of consciousness.
134
Nonphysicalistic Science
It should not be necessary to say, and yet I fear I had better say nonetheless,
that to deny there is an evolutionary explanation is not to suggest that there is some
other explanation, let alone a so-called creationist explanation. It is true that
Descartes and other pioneers of his and the next century almost all were or
professed to be theists or deists and would as such presumably have subscribed to
the view that whatever the laws of nature are, they are so because a creator so wills
them to be. But I think we may credit them all with the wit to see that a general
salute to divine providence is not going to provide an explanation of any specific
law, and this very much includes psychophysical correlation laws.
When I say there is no evolutionary explanation of consciousness, I must add
that there is no explanation, and certainly no scientific explanation, at all at
present. But to see this clearly one must consider what kinds of nonphysicalistic
science is currently available. I see three varieties. One is concerned with relations
between conscious states and internal physical states, neural or encephalic, as
represented by the portion of the work going on under the medical label of
neurology and scientific label of neuropsychology studying brain and
consciousness, rather than just brain and behavior — the work, in short, on which
the case for the correlation principle rests.
Another variety of nonphysicalistic science is concerned with relations
between different conscious states, as represented by work in introspective
psychology, a flourishing field a hundred years ago, which fell out of favor for a
time, but in which work has picked up since. Along with it should be mentioned
the related work that goes on in philosophical circles, sometimes in rather cryptic
language, under the label of phenomenology.
Yet another variety is concerned with relations between conscious states and
external physical stimuli. This work includes the oldest branch of experimental
135
psychology, called, psychophysics.
Psychophysics
With telepathy or extrasensory perception ruled out, to communicate anything
at all, even pure mathematics, some sensory channel is needed. Arguably, to
understand physics one must have an understanding of oneself as physically
embodied and located in a physical universe, and this also requires some kind of
sensation. But any of several senses, especially but not exclusively the so-called
proprioceptive, might suffice for this purpose.
To communicate physicalistic science, there is no one channel such that it
specifically is needed: not sight, not hearing, not even their disjunction. It must be
acknowledged that great determination and skilled assistance is needed to
overcome the impediments presented by a lack of sight and hearing, but it is no
surprise to read in the autobiographical writings of Helen Keller that she studied
physics. Nor should it be a great surprise to find that the only difficulty she
complained of was that of getting books Brailled in a timely fashion. For
physicalistic science cannot say anything that cannot be said in Braille, and cannot
teach us anything that it cannot teach her.
By contrast, psychophysics is explicitly concerned with a variety of specific
sensory channels. The most basic result, applicable to several such channels, is
Weber’s law, to the effect that the amount of increase in stimulus needed to
produce a noticeable increase in sensation is proportional to the amount of stimulus
already present: a bright light must be increased more than a dim one to be seen as
having got brighter, and a heavy weight must be increased more than a light one to
be felt has having got heavier.
Here we have two clear points of contrast with physicalistic science. First,
reported sensations are treated as data on a par with measured stimuli. Second, the
136
results will not be fully intelligible to subjects lacking the relevant sensory
channels. Keller, never having had the experience of seeing one light look brighter
than another — or having forgotten any such experience she may have had in
infancy before losing her sight — will not be able to understand what Weber’s law
for brightness is even about, in the way we can understand what Weber’s law for
brightness is about, or she can understand what Weber’s law for heaviness is about.
During the period, lasting for far too much of the first half of the last century,
when the understanding of psychologists was darkened by the scientistic ideology
of behaviorism, Weber’s law was reinterpreted physicalistically, as being about the
minimal increase in stimulus needed, not to produce a noticeable increase in
sensation, but to produce a different behavioral response, to begin with, the verbal
behavior of reporting a noticeable increase in sensation. That transformation at first
appears an empty verbal trick, or simply a bad joke, since the only reason to be
interested in the difference in reporting behavior is belief that those doing the
reporting are being truthful, and that their reports reflect their actual experiences.
But the behaviorization of the subject does extend its scope to organisms incapable
of reporting and perhaps incapable of having subjective experiences.
Physicalistic Pure Theory and Psychophysical Practical Lore
Much psychophysical science, however, escapes behavioristic
reinterpretation; indeed, physicalistic theoretical science always comes wrapped in
psychophysicalistic practical lore that would entirely lose its point if subjected to
such reinterpretation. Thus in pure physics, the core theory of black-body radiation
concerns the relationship between temperature and frequency of emitted
electromagnetic radiation, but this material is never taught to students without
adding something not present in the mathematical formulation of the pure theory,
without informing or reminding them, in the ordinary color vocabulary known
137
since childhood, of what the various frequencies of light look like.
Equipped with this practical lore in addition to the pure theory, the physics
student presented with a distant hot body will be able to give a rough estimate of
its temperature simply by looking whether it is black or red or yellow or white or
blue. Students who lack normal color vision, and have to rely on a spectroscope,
will be at a disadvantage, though perhaps not a very great one, since everyone will
have to rely on instruments — and for that matter, calculations involving the
Doppler effect, if the light source is moving — to get more than a rough estimate.
Behavioristic reinterpretation is out of the question here. For what interests
astronomers, say, is how the stars will look to themselves, not how their behavior
on peering into their telescopes will look to any rat psychologists who may have
sneaked into the observatory to spy on them.
The wrapping in psychophysical practical lore becomes thicker with
chemistry and geology, with their bead tests and flame tests. All the senses are
drawn in: not only do acids turn litmus paper red, but they taste sour. The
importance of the wrapping for applications should not be allowed, however, to
obscure the physicalistic character of the theoretical core.
Gloucester’s Principle
In the category of work concerned with relations between different sensory
states, special interest attaches to studies going back to the nineteenth century that
led to the various so-called color solids. Consideration of these can help to
underscore just what is and what is not knowable about color without actual
experience of color vision, a topic given philosophical prominence by Jackson.
In Henry VI, Part II, Act II, Scene I, Humphrey Duke of Gloucester exposes
the fraud of one Saunder Simpcox, a knave claiming to have been born blind and
to have had his sight restored by miraculous intercession of Saint Alban.
138
Pretending to doubt that Simpcox can see well even after the supposed miracle,
Gloucester asks him to name the colors of various articles of clothing worn by
some of those present, which he successfully does. He then asks Simpcox to name
several of the persons present, which he cannot. The Duke then argues thus:
If thou hadst been born blind, thou mightest as well have known all our names
as thus to name the several colours we do wear. Sight may distinguish of
colours, but suddenly to nominate them all, it is impossible.
But is this last principle correct? Let us suppose that, rather like Jackson’s
Mary, Simpcox had learned all that can be learned of color theory by a blind man,
and retained it all in memory. Gloucester concedes that, upon receiving his sight,
Simpcox should be able to distinguish colors, and presumably this ability would
extend to noting various relations among them, such as one being more similar to a
second than to a third, or one looking like a mixture of a second and a third, and
the like. Given the chips, and given the time, he might by putting similarly-colored
chips next to similarly-colored chips, assemble a whole color solid. Yet Gloucester
would doubt he could name any of the colors in it.
Simpcox’s Defense
The conclusion is perhaps hasty. Simpcox will have heard, while still blind, of
a regular alternation of periods of daylight, when distinguishing objects by sight is
easy, and periods of nighttime, when it is difficult. And unless he is in the land of
the midnight sun, or indoors, or in a place brightly lit by modern artificial means
— none of which applies to the streets of fifteenth-century Saint Albans — he will
soon enough experience nighttime. When daylight returns, he will be able to
distinguish one color at least on the color solid, since he will have heard of black as
139
the color to which all things tend at night.
Next white can be identified as the color most opposite, or least similar to
black, and the gray scale as the colors that look like mixtures of black with white,
and the hues as the colors that seem to have no admixture of gray to them. These,
with most similar adjacent to most similar, form a ring, in which some are brighter,
or more similar to white, and others are darker, or more similar to black. Most will
look like mixtures of the colors on either side, but for four pure colors this will not
be so, and yellow and blue can be recognized as the lightest and darkest,
respectively, of these. It only remains to determine which direction around the ring
from blue to yellow is the path through red and which through green.
If, as some suppose, the synæsthetic perception of some colors as warm and
others as cool is innate, and not the result of association of red and yellow with
fire, and blue and green with lakes and seas, red and green can be distinguished as
warm and cool. If not, one may have to resort to counting the number of
distinguishable shades, there being more between blue and green than between
blue and red, where lie the nonspectral purples. But with a half-dozen points on the
solid identified, all other colors can be described as mixtures thereof, and it seems
all colors named, on the basis of information acquired while still blind.
And so it seems Gloucester may have been wrong. But note carefully what
our hypothetical Simpcox could and could not come to know while still blind:
everything about the structure, one might say, of the color solid or color space,
nothing about the content; how each color is related to each other, but not how any
of them looks. Study of the physical properties of the color chips, what frequencies
of light they reflect and how well, and study of the portion of the nervous system
concerned with color vision, together might be able to explain — modulo the great
unexplained fact of the correlation principle, elaborated with the assumption that
similars are correlated with similars — how Simpcox is able to make the
140
distinctions and similarity and other judgments that he makes; and all this could be
communicated in physicalistic terms. But what the colors look like, that could not
have been communicated to Simpcox while still blind, and lies outside the bounds
of physicalistic science, and beyond what evolutionary biology can explain.
Pessimistic Induction
McGinn is one who notoriously would place it beyond what any science of
ours, present or future, could hope to explain. And certainly, looking at the three
kinds of nonphysicalistic science we actually have, it is hard to imagine how more
of the same of any of these could help explain why the sensation correlated with
any particular stimulus or neural activity is as it is, and why the system of relations
among sensations is filled in with the content that is there, or for that matter with
any content at all. There are ample grounds for a so-called pessimistic induction
here: nothing we have so far been able to come up with in the way of a scientific
investigation seems of any help, and work has gone on long and hard enough to
suggest that our failure to date is no accident, but can be expected to continue.
Still, I would distinguish: that there will never be a physicalistic explanation is
arguably something that might be called a conceptual truth for anyone who really
understands the post-renaissance conception of the physical with which we have
been operating since the days of Descartes at least. That there will never be a
scientific explanation of any kind is at least for the present only what I called it, a
pessimistic induction. The main obstacle to elevating this conclusion to the status
of conceptual truth is the lack of any very clear analysis of what science in general,
as opposed to physicalistic science specifically, may or must mean.
The most important division on the mind-body problem among philosophers
today, I submit, is one whose dividing line runs through the nondualist side. The
line divides those who, like Block or Levine, acknowledge something not unlike
141
the sort limitations of scientific explanation I have just been describing (who
acknowledge an “explanatory gap” which they characterize in various ways that I
will not attempt to summarize, not wanting to put words in their mouths), from
those who are in denial about the matter. The latter sometimes write big books
purporting to explain consciousness while in fact ignoring it.
Philosophical Denialism
Denial that there is anything missing in a purely physicalistic account of the
world remains common. It may be less popular among philosophers than it once
was, to the point of being wholly out of fashion in the rising generation of graduate
students, but it still has a large following among the scientific and lay public,
where the best-known denialist philosophers still have many fans. Denialism
comes in two main forms, one explicit and extreme, the other implicit and
moderate by comparison.
The extreme version, known as eliminativism (or “eliminativist
materialism”), can be traced back at least as far as the early career of my late excolleague Richard Rorty, in the days before he renounced philosophy for
comparative literature. It is a more radical philosophical materialism than either of
the two physicalist views I have taken into account so far, neuralism and
functionalism. It does not identify a conscious state with a neural or functional
state, but simply denies that there are any conscious states. This may seem too
crazy a view to be held even by a philosopher, and I have not wanted to say much
about it, there being something ridiculous about debating someone who is
pretending to be a zombie. But it must be mentioned.
Eliminativists may be one shade less crazy than my description of them so far
may suggest. For they do not say that we should claim we do not have sensations,
and are blind and deaf and subject to congenital analgesia, but only that we should
142
refrain from claiming that we do have sensations. Their view is that phrases such
as color-sensations or pitch-sensations or pain-sensations really denote nothing
and should be banned from serious discussion.
This claim is supposed to follow from the fact that many commonsense
beliefs about such sensations are confused. How it is supposed to follow is not
clear, since after all in Homer’s day many commonsense beliefs about stars and
planets were confused, and yet stars and planets did and do exist for all that. That
there are some commonsense confusions can hardly be denied; on the contrary, it
is a point I have emphasized. But why should such errors be taken to prove
anything more than the need for more and better introspective psychological and
phenomenological research, alongside on-going neuroscientific work?
How much commonsense confusion there is seems in any case to be
exaggerated by eliminativists. Sometimes the exaggeration is accomplished using a
kind of philosophical prestidigitation: first ordinary vocabulary is replaced by a
technical jargon (“qualia”), then the technical jargon is dissuasively defined to
include a whole list of implausible features, such as infallibility.
Scientific Denialism
If denialism consisted only of such philosophical hocus pocus, it might be of
little moment. But there is second form the implausibility of whose assumptions is
less conspicuous, simply because the assumption itself is less conspicuous. Among
scientists it may be embodied in a claim that some latest discovery or conjecture
about the brain promises a final, thoroughly materialist, solution to the mind-body
problem. At its simplest, this form of denialism just assumes, perhaps without even
mentioning or even noticing the assumption, that once we have pinned down for
each mental state what physical state is identical with it, then as physics and
physiology advance to give complete explanations of the physical states in
143
question, it will be giving us explanations of everything there is to be explained.
Professed physicalists or antidualists who reject this assumption do so for
various reasons. One common line holds that though being in pain and having
C-fibers that are firing are two descriptions of the same state, a physical one,
nonetheless they are associated with different routes, which may be contrastingly
called first-person and third-person, to knowledge that one is in that state; and
why this, that, and the other first-person route is associated with this, that, or the
other third-person route has not been explained.
Once this sort of gap or lapse has been pointed out, it may seem difficult to
ignore. Yet denialism that implicitly ignores it remains common. It is enough so
that every year or two the New York Review of Books feels obliged to take note of
some new denialist work, and publishes a critical review by one of the journal’s
regular dualistically-inclined reviewers. Sometimes it is a work of an eliminativist
philosopher at issue, as when Searle takes on Daniel Dennett, or McGinn deals
with Patricia Churchland; but more often it is a work by a scientist.
The works by nonphilosophers discussed in the NYRB typically contain a
good deal of science popularization, sometimes so much that a surprisingly small
number of pages are left to be devoted to what is supposed to be the main event,
the solution or dissolution of the centuried mind-body problem. The merits of the
nonphilosophical parts of such books vary, but generally make simple dismissal of
these volumes difficult, though the few overtly philosophical pages may be, and
according the reviewers generally are, quite weak.
It is no surprise that the solutions to or dissolutions of the mind-body problem
offered in such works fail to satisfy the reviewers already known for their dualistic
tendencies. In general they fail to satisfy other denialists as well, since the denialist
books keep coming out, and the later ones seldom cite the earlier ones, and never
as having already solved the problem; but that is so because a later writers will be
144
less impressed by whatever new discoveries or conjectures interest the earlier ones.
More denialist positions are to be found in places where they may not be soon
noticed by professional philosophers: in incidental passages in publications or
presentations in other fields, and in personal weblogs. One psychologist at my own
university, for instance, has in public talks I have attended repeatedly compared the
belief that we are conscious to the delusions of a psychiatric patient who maintains
that there is a squirrel inside his skull.
That only the deluded could imagine that anyone is ever conscious is a more
radical claim, I think, than one is likely to find in philosophers, even ones
influenced by Rorty or Dennett or Churchland. For philosophers are unlikely to fail
to recall Descartes’s argument that, since a delusion involves thinking that
something is some way that it isn’t, only conscious subjects, capable of thinking,
are capable of being deluded.
The supervenience thesis, as the issue marking the boundary between dualism
and physicalism, cuts through and divides and to that extent weakens the
antidenialist camp, putting Chalmers on one side and Block and Levine on the
other. To the extent that the supervenience issue has this effect, it constitutes a
potentially pernicious distraction.
I have no doubt that those who popularized the zombie idea introduced it with
a sincere desire to illustrate or dramatize, and thus call attention to, the limitations
of science as we know it; but in practice the zombie issue has tended to divert
attention from that matter into issues about, among other things, an apparatus of
possible worlds. The issue whether beings conceded to be conceptually possible
and nomologically impossible are possible or impossible in some intermediate
distinctively metaphysical sense threatens to draw attention away from the critique
of denialism.
How serious that matter is depends on what one thinks of denialism and its
145
influence, especially outside professional philosophical circles, a matter about
which there may be legitimate differences of opinion. I myself would say that
denialism now occupies the position once occupied by credulity about the
paranormal, that of being the most serious widespread philosophical error about
mind, and that it more seriously obscures intellectual vision and obstructs
intellectual progress than any view in philosophy of mind since behaviorism a halfcentury and more ago. The thought that there can be an evolutionary explanation of
everything that needs explaining, especially, clouds contemporary discussion of
some important issues, as I will try to illustrate in my last chapter.
Special Quantifiers versus Special Objects
There is another distraction, closely related to the zombie issue. Philosophers
may be diverted from discussing whether the physicalistic account of the world is
incomplete into debating whether more specifically there are facts — or if not
facts, items of some other specified category — as one or another metaphysician
understands them, that the physicalistic account of the world leaves out. How such
diversion can occur calls for comment. Dualists and nondenialist physicalists
maintain that there is something the physicalistic description of the world leaves
out, but we need to focus for a moment on this phrase there is something.
Such an expression, labeled by Friederike Moltmann a special quantifier, has
under one label of another attracted a good deal of attention recently from
philosophers of language and linguistic semanticists. The behavior of such
expressions contrasts with that of such quantifier-expressions as there is an
institution or there is an edifice or there is an electrochemical process or whatever,
involving there being a thing of some more or less definite sort or other. The
peculiar feature of special quantifiers comes out when someone makes a specialquantifier assertion, there is something that does such-and-such, and someone asks
146
for specifics, well, what is it?
If I say there is something blocking the road and you ask me what it is, I may
answer a piece of construction equipment. In this case the expression I utter in
reply is what linguists call, or used to call, a noun phrase, and as such it is an
expression that at least purports to denote something, whether I am right and there
is really a piece of construction equipment blocking the road, or I am wrong and it
is something else that is blocking the road, or I am even more wrong and there is
nothing blocking road, and the slow-down in traffic up ahead is due to rubbernecking. But if I say there is something the physicalistic description of the world
leaves out, and you ask what it is, I may answer that red looks like THIS. Here my
response is not a noun phrase, but a that-clause, and as such a linguistic item that
does not have to be regarded as even purporting to denote a person, place, or thing.
And thus there appears to be a gap between granting that there is something left out
and holding that there is some THING left out.
The gap becomes wider if one insists on a thing of a specified sort: a property
or an event or a fact. It becomes a veritable yawning chasm if one asks not just for
a property, but for a property as understood in Professor X’s metaphysical theory
of the ultimate nature of properties, or not just for an event, but for an event as
understood in Professor Y’s metaphysical theory of the ultimate nature of events,
or not just for a fact, but a fact as understood in Professor Z’s metaphysical theory
of the ultimate nature of facts.
Now an articulated dualist view is likely to claim not merely that physicalism
leaves something out, but that it leaves some THING out, and is likely moreover to
be specific about what sort or sorts of thing is or are left out, and it will often be at
least loosely tied to some view of the nature of those things. This leaves room for
views that are not denialist, but rather hold that there is indeed something left out,
but that are physicalist at least to the extent of being dissatisfied with all available
147
articulated dualist views. If I am not mistaken, such nondenialist physicalism —
granting, perhaps, that the identity of being in pain with having C-fibers firing does
not explain why being in pain feels as it does, without granting that the phrase the
way pain feels denotes some THING — is becoming increasingly common.
Descriptive or Preservationist versus Revisionary Metaphysics
The issue between dualism and nondenialist physicalism is one of
metaphysics, and in confronting it we must confront the deepest division there is or
can be on metaphysical topics, the opposition between what Strawson called
descriptive metaphysics and its opposite. Strawson introduced his term in
connection with a specific project, but with a view to more general issues; for
present purposes it may be defined as the attempt to articulate the most general
categories of our thought, as embodied in our language. In its original form,
manifested at Oxford in the 1950s (for Strawson was a prominent figure in Oxford
“ordinary language philosophy”) it has been criticized for its antitheoretical bias,
not that in the 1950s there was much that theoretical linguistics, still mired in
behaviorism, could have contributed. It has also been criticized for parochialism,
and a lack of interest in, if not an outright denial of, significant historical or
cultural variation. Its closest present-day successor, as seen for instance in
Moltmann’s work, differs from the original version on just such points, being
closely integrated with linguistic semantics and making use of cross-linguistic data.
Strawson contrasts descriptive with revisionary metaphysics, which
terminologically is odd. The natural contrast would be between descriptive, which
only describes our concepts as they are, and prescriptive, which would prescribe
what our concepts ought to be. Within prescriptive metaphysics one could then
distinguish prescribing that we preserve things as they are, and prescribing that we
revise things. In fact, the distinction between descriptive metaphysics, which only
148
characterizes, and what I will call preservationist metaphysics, which both
characterizes and endorses, seems to have meant little to Strawson, and in practice
I will not myself make much of the distinction that exists in principle.
Perhaps the most significant division would be within the revisionary
category, over what is supposed to be accomplished by the prescribed revision.
Here we may distinguish a modest variety, which aims to create a new system of
concepts that will do better for certain human purposes, and an ambitious variety,
which aspires to transcend, to get behind or outside or above or beyond, all merely
human representations to ultimate reality as it is in itself, and revise our concepts
until they are in alignment with it. Typically when I speak of revisionist
metaphysics I will be thinking of the latter, transcendent variety.
Though achieving alignment with ultimate reality seems more like the
traditional aim of metaphysicians, it by now is almost equally traditional to
complain that what such alignment would amount to is less than clear. For
seventeenth- or eighteenth-century philosophers, who generally were or professed
to be theists or deists — and ones who seem not to have been very worried about
the dangers of anthropomorphism in theology — it might mean alignment of our
concepts with God’s concepts. Today it can at least be said that if we were to
achieve such alignment, then any intelligent extraterrestrials whose modes of
thought and communication were not aligned with ours would be doing something
wrong that we would be doing right.
Facts of the Matter, Facts of the Mind
Now if one looks at the issue of whether there is, say, something that might be
called a fact that the physicalistic description of the world is leaving out, then
descriptive metaphysics must answer in the affirmative, but revisionary
metaphysics may answer in the negative. For if we look at the ordinary usage of
149
fact, we find a conception that makes facts both abundant and fine-grained, while
revisionists may posit a realm of facts that are sparse and/or coarse-grained.
Facts as descriptive metaphysics would have to describe them are abundant in
the sense that whenever I may affirm that p I may with equal right begin to speak
of the fact that p. If there are no unicorns, then ordinary language immediately
allows this assertion to be pleonastically expanded to the assertion that it is a fact
that there are no unicorns. Revisionary metaphysicians, by contrast, often claim
there are no negative existential facts, no facts about the nonexistence of kinds of
things, unicorns included.
Facts as descriptive metaphysics would have to describe them are finegrained in the sense that, except where p and q are pretty obviously equivalent, the
fact that p is a distinct fact from the fact that q. For unless p and q are pretty
obviously equivalent, some person might be in a state ordinary language would
describe as one of being aware that p and unaware that q, and then ordinary
language immediately allows this assertion to be pleonastically expanded to the
assertion that this person is aware of the fact that p but not of the fact that q. But if
the fact that p is one of which this person is aware, and the fact that q is one of
which this person is unaware, then these must be two distinct facts. Revisionary
metaphysicians, by contrast, often claim that facts are coarse-grained, and that
anyone aware of the fact that p must be ipso facto aware of the fact that q, since
these are quite literally one and the very same fact. This difference is directly
relevant to the dualist-physicalist debate in cases where p is about conscious
experience while q is about some physicalistic correlate.
To give an example from personal experience, some years ago I had for ten
minutes or so a striking hallucinatory experience, as of water running down the
walls of a small room where I happened to be. There had previously been problems
with leakages from the plumbing above this particular room, and I at first believed
150
that the plumbing problems had recurred, until I touched the wall and found it bone
dry. Then I did not know what to think. The experience was not disturbing enough
to send me to a doctor, but when I did in the normal course of events have to go to
various doctors, I did ask about it, but received no enlightenment until I mentioned
it to my ophthalmologist, Dr Samuel Liu.
He was able, after asking a very few apt questions, to give a diagnosis of
optical migraine. Prior to consulting with Dr Liu, I had been aware of having had a
strange experience, but not of what my kind of experience is called; after
consultation, I was aware of having had an optical migraine. I was not, however,
and still am not, aware of what goes on in the nerves or brain during such an event.
Dr Liu, who presumably has knowledge of the neurology of these events, could
therefore say that JB is aware of the fact that he has had such-and-such an
experience but JB is unaware of the fact that such-and-such went on in his nervous
system. And from the point of view of descriptive metaphysics, the dualist
conclusion now follows that the physicalistically-indescribable fact that I have had
such-and-such an experience is distinct from any physicalistically-describable fact
to the effect that such-and-such was going on in my nervous system. But the
bearing of the descriptive-revisionary divide does not end with such examples.
151
VII Realism versus Pragmatism about Modality
The distinction between modest descriptive metaphysics, which is only
concerned to work out explicitly what our implicit linguistic conventions are, and
more ambitious revisionist metaphysics is closely connected with a distinction
between two conceptions of the nature of modality, one of which leads to
suspicions, not directly about the correctness, but rather about the significance of
the supervenience thesis.
Modal Knowledge
Dorothy grew up in Kansas, where everything is gray, until one day a
cyclone picked her up and set her down in a technicolor land where she
encountered a rainbow of hues, and she was able at last to say to herself, so this is
what emerald green looks like. According to Jackson’s former view, even if
Dorothy back home had learned everything there is to be learned about color and
color vision from black and white books and black and white newsreels, still she
would now be learning something new. And I have just argued towards the end of
the preceding chapter that from the standpoint of descriptive metaphysics, she was
indeed becoming aware of a new fact.
A revisionary metaphysician, as we have seen, may deny this kind of claim.
Such a metaphysician might perhaps want to claim that though Dorothy may,
looking at an appropriate sample, say emerald green looks like THIS, still there is,
owing to the sparsity of facts, no fact about what emerald green looks like for her
to find out. Or such a metaphysician might more likely want to say that the fact of
what emerald green looks like, which Dorothy supposes she has only just now
learned, is on the contrary, owing to the coarse-grainedness of facts, quite literally
one and the very same fact as some physicalistically-describable fact of which she
152
knew all along from her book-reading and newsreel-viewing back on the gray
prairies of the American Midwest.
Anyone who has read any substantial amount of the contemporary literature
on the mind-body problem will have encountered metaphysicians engaging in such
manœuvres. Such shenanigans cannot make the notion of an evolutionary
explanation of consciousness anything but absurd, but given enough of them,
enough dust may be kicked up that the absurdity is no longer clearly visible.
My first suspicion about the supervenience issue was that it was a pointless
detour, temporarily taking us away from direct confrontation with issues about
identity criteria, but in the end taking us right back to them. My second suspicion
about the supervenience thesis is that it has something of the same dust-kicking
character as a revisionary metaphysics of facts, so that it tends to produce a
pernicious distraction from what we ought to be focusing on, the limitations of
evolutionary biology and indeed all current science in dealing with consciousness.
I said earlier on that I would voice three suspicions about supervenience. I
am now done with two, and must start on the third. That third suspicion about the
supervenience issue is connected with its being an issue about metaphysical
modality specifically, rather than any other kind. My suspicion is that metaphysical
modality — the only kind of modality at issue in the remaining pages of this book
— may involve a popular delusion in the same sense in which Hume may without
too much distortion be described as holding that causality involves a popular
delusion (as for him do a good many other things that are supposed to result from
the mind’s “propensity to spread itself on external objects” and its penchant for
“gilding and staining” all natural objects with colors borrowed from internal
sentiments).
The relevant issue is sometimes discussed under the rubric of the origin of
necessity, but the first thing that must be said when such a phrase is used is that it
153
cannot be taken literally. It is for this reason perhaps best to begin with a distinct
but related question that can be taken literally, namely, the issue of origin, not of
necessity itself, but of our knowledge of necessity — an issue that has already been
briefly touched upon.
Knowledge of Necessity through Kant
Here is how the question arises. We think that when things are some way, in
some cases they could have been otherwise, and in other cases they couldn't. That
is what we call the modal distinction between the contingent and the necessary.
And we not only think that there is such a distinction, we also think that we know
some examples.
For instance, though up to the time of this writing there has been no female U.
S. president, there could have been, whereas the number 29 not only has no
nontrivial divisors but couldn't have had any. Thus we may speculate
counterfactually about What if Hillary Clinton had been president in 2001? but
won't speculate counterfactually about What if 5 had been a divisor of 29? And
what is called the question of modal epistemology is just this: How do we know
that the examples in question genuinely are examples of that of which they are
supposed to be examples?
And why should this question be considered a difficult problem, a kind of
mystery? This is the point already touched upon, with a quotation of a line of Kant
from the introduction to the B edition of the first Critique, a line that by the way is
already quoted in Kripke’s addenda, about experience being unable to teach that
something could not have been otherwise. Sense-experience, with which our
knowledge begins, seems able only to provide knowledge about what is or isn't, not
about what could have been or couldn't have been. How do we bridge the gap
between is and could?
154
Kant does allow that experience can teach us that a necessary truth is true;
what it is not supposed to be able to teach is that it is necessary. For example,
experience with what they call manipulatives in pre-school, trying to arrange 29
blocks in various patterns, may convince one that 29 is a prime number, one that
has no nontrivial factorization; but granted that it doesn't actually have one, how do
we know that it couldn't possibly have had one, and is necessarily a prime number?
How can experience teach us that extra fact?
The problem becomes more vivid if one adopts the language that was once
used by Leibniz and much later repopularized by Kripke, the usage according to
which the necessary is that which is true in all possible worlds. In these terms the
problem is that the senses only show us this world — imagine here a sweeping
gesture — or the actual world as it is called, whereas when we claim to know
about what could or couldn't have been, we are claiming knowledge of what is
going on in some or all other worlds. For that kind of knowledge, it seems, we
would need a kind of sixth sense, or extrasensory perception, or nonperceptual
mode of apprehension, to see beyond the world in which we live to these various
other worlds. That is the problem.
Let me now rapidly review some of the highlights of the history of attempted
solution prior to Kripke’s intervention, beginning with Kant himself. Kant
concludes, in the very next sentence after the one I quoted earlier, that our
knowledge of necessity must be apriori knowledge or knowledge that is
independent of experience, rather than aposteriori knowledge or knowledge that is
dependant on experience. And so the problem of the origin of our knowledge of
necessity becomes for Kant the problem of the origin of our apriori knowledge.
Well, that is not quite the right way to describe Kant's position, since he
doesn't think all a priori knowledge is mysterious, but only most of it: there is one
special class of cases where Kant thinks it isn't really so hard to understand how we
155
can have apriori knowledge.
We can generate examples from Kant's special class of unproblematic apriori
truths by the following three-step process. First, take a simple logical truth of the
form Anything that is both an A and a B is a B, for instance, Anyone who is both a
man and unmarried is unmarried. Second, find a synonym C for the phrase thing
that is both an A and a B, for instance, bachelor for one who is both a man and
unmarried. Third, substitute the shorter synonym for the longer phrase in the
original logical truth to get the truth Any C is a B, or in our example, the truth Any
bachelor is unmarried. Such a truth will be an example of what Kant calls an
analytic as opposed to a synthetic judgment. Such truths are always apriori, but
according to Kant our knowledge of them is unproblematic, because it’s not really
knowledge of external objects, but only of the content of our own concepts, or as
later philosophers have preferred to say, the meanings of our own words.
So the problem for Kant is not exactly how apriori knowledge is possible, but
more precisely how synthetic apriori knowledge is possible. Kant thought we do
have examples of such knowledge, and indeed the example I have been using, that
29 is a prime number, is one of them. Arithmetic, according to Kant, was supposed
to be synthetic a priori, and geometry, too — all of pure mathematics. In his
Prolegomena to Any Future Metaphysics, Kant listed How is pure mathematics
possible? as the first question for metaphysics, for the branch of philosophy
concerned with space, time, substance, cause, and other grand general concepts —
including modality.
Knowledge of Necessity since Kant
Later philosophers questioned whether there really were any examples of the
synthetic apriori. Geometry, so far as it is about the physical space in which we
live and move — and that is how Kant and his contemporaries conceived it —
156
came to be seen as, not synthetic a priori, but rather a posteriori. The great
mathematician Carl Friedrich Gauss had come to suspect that geometry is a
posteriori, like the rest of physics, by some time in the first half of the nineteenth
century. Since the time of Albert Einstein in the early twentieth century the a
posteriori character of physical geometry has been the received view. As for
arithmetic, it has since the time of the great logician Frege in the late nineteenth
century come to be thought by many philosophers to be, not synthetic apriori, but
analytic — of the same status as Any bachelor is unmarried, except that to obtain
something like 29 is a prime number one needs to substitute synonyms in a logical
truth of a form much more complicated than Anything that is both an A and a B is
a B.
Once Kant's synthetic apriori has been rejected, the question of how we have
knowledge of necessity reduces to the question of how we have knowledge of
analyticity, which in turn resolves into a pair of questions: On the one hand, how
do we have knowledge of synonymy, which is to say, how do we have knowledge
of meaning? On the other hand how do we have knowledge of logical truths?
As to the first question, presumably we acquire knowledge, explicit or
implicit, of meaning as we learn to speak, so that by the time we are able to ask the
question whether this is a synonym of that, we have the answer, though it may take
a long dialectic of analysis to tease it out. But what about knowledge of logic? That
question didn't loom large in Kant's day, when only a very rudimentary logic
existed, but after Frege vastly expanded the realm of logic — only by doing so
could he find any prospect of reducing arithmetic to logic — the question loomed
larger.
Many philosophers, however, convinced themselves that knowledge of logic
also reduces to knowledge of meaning, namely, of the meanings of logical
particles, words like not, and, or, all, some, and a few more. To be sure, there are
157
infinitely many logical truths, in Frege's expanded logic. But they all follow from
or are generated by a finite list of logical rules, and philosophers were tempted to
identify knowledge of the meanings of logical particles with knowledge of rules
for using them: Knowing the meaning of or, for instance, might be knowing that
A-or-B follows from A and follows from B, and that anything that follows both
from A and from B follows from A-or-B. So in the end, knowledge of necessity
reduces to implicit or explicit knowledge of semantical rules or linguistics
conventions.
On this picture we know that five isn't a divisor of 29 in any possible world,
not because we have some kind of parapsychological faculty that permits us to
engage in so-called remote viewing of other possible worlds while remaining in
this one, but rather because, just as we have implicitly learned conventions or rules
that tell us no one who is married may properly be described as a bachelor, so also
we have implicitly learned conventions or rules that tell us, or from which it
follows, that one may not describe any world as a world in which 5 is a divisor
of 29.
For instance, a world in which whenever one lays out five rows of six blocks
and then counts the total, one finds 29, is not properly described as a world in
which five times six is 29. Rather it is a world where, by some curious physical
principle, whenever 30 blocks are brought together in five rows of six, one of them
disappears; or where, by some curious psychological principle, whenever we try to
count 30 blocks arranged in five rows of six, we always miss one; or something of
the sort.
Such is the sort of picture that had become the received wisdom in philosophy
departments in the English speaking world by the middle decades of the last
century. A. J. Ayer, the notorious logical positivist, and Strawson, the notorious
ordinary-language philosopher, much as they disagreed with each other across a
158
whole range of issues, both put forward a picture quite like this, Ayer in his
Language Truth and Logic, Strawson in his Introduction to Logical Theory.
Ayer and Strawson on the Origin of Necessity
And people like Ayer and Strawson would sometimes go on to speak as if
linguistic convention were the source not only of our knowledge of modality, but
of modality itself, and go on further to speak of the source of modality lying in
ourselves. For while individually, as children growing up in a linguistic
community, or foreigners seeking to enter one, we must internalize the pre-existing
implicit rules of the communal language, collectively, as a speech community, we
do not so much learn as create the language with its rules. If the origin of modality,
of necessity and its distinction from contingency, lies in language, it therefore lies
in a creation of ours, and so in us.
Thus Ayer, in his chapter on the apriori says this:
The principles of logic and mathematics are true universally simply because
we never allow them to be anything else. (p. 41)
And Strawson early on (p. 5), speaking of why some pairs of predicates are
incompatible, says this:
It is we, the makers of language, who make predicates incompatible … It is
we who decide where the boundaries are to be drawn.
The common element here is the pronoun we. The picture is that the necessary,
identified with the apriori, identified with the analytic, is made so by us, that we
are the source.
Such was the climate of opinion before January, 1970, when Kripke delivered
159
his Princeton lectures on Naming and Necessity. Since then it has become clear
that there are several things wrong with the conclusion that we are the originators
of necessity, and the line of thought from Kant onwards that led to this sort of
conclusion. I won't try to say in advance just how many main things are wrong
with this line of thought, lest I end up like Cardinal Ximenes in the Monty Python
sketch about the Spanish Inquisition, and find myself repeatedly having to revise
my estimate upwards. I'll just mention several difficulties one by one and we can
tote them up afterwards if we wish.
To begin with, there is a difficulty discussed by Kripke in his lectures, to be
found right at the beginning of the whole line of argument, with Kant, and indeed
at the very first step, Kant's inference from the premise that experience does not
teach us that a thing cannot be otherwise to the conclusion that our knowledge of
necessity is apriori. Strawson was less dismissive of Kant than was Ayer, and
wrote a whole book about him, The Bounds of Sense, but in it he famously
described (p. 137) one argument of Kant’s as a non sequitur of numbing grossness.
I don't know whether the argument to which this phrase was first applied really
deserved words so harsh, but Kant’s argument from experience isn't sufficient to
teach us that a thing couldn’t have been otherwise to experience isn’t needed to
learn that a thing couldn’t have been otherwise certainly does deserve the harsh
phrase. It’s like arguing from sunshine isn't sufficient to grow corn, rainfall is also
needed to sunshine isn't needed to grow corn, rainfall by itself is sufficient. It really
is a non sequitur of numbing grossness.
Kripke’s achievement was not just to point to this logical lapse on the part of
a famous philosopher — finding logical lapses in the works of famous
philosophers is something it is all too easy to do, I’m afraid, and one doesn’t
become famous just for doing that — but rather, his achievement was to offer
plausible counterexamples to Kant’s claim, plausible cases of aposteriori
160
necessities, of the kind we have examined in the preceding chapter.
Such examples show that what philosophers of the period of Ayer and
Strawson tended to claim about our knowledge of necessary truths is wrong: it
does not just reduce to knowledge of linguistic rules. This purely negative
conclusion is what Kripke left us at the end of the oral delivery of his lectures, but
in the addenda to the published version he offered a partial repair, of a kind we
have also examined in the preceding chapter. On the picture in the addenda, or the
one I arrived at by extrapolating a bit from Kripke’s hints there, our knowledge of
aposteriori necessities reduces to a combination of aposteriori knowledge gained in
the ordinary way involving sense sense-experience, with analytic knowledge of
semantic rules or linguistic conventions.
If this is so, then the formulation that our knowledge of necessity derives from
our knowledge of semantic rules and linguistic conventions would be defensible if
understood as meaning only this, that though our knowledge of the truth of a
necessary truth may depend on sense-experience, our knowledge of the necessity
of that truth, given that it is true, derives from knowledge of rules and conventions
alone. The main point is that the epistemology of modality is still demystified:
there is no need for a third eye and a second sight to peer into a hypercosmos of
potential universes, a pluriverse of possible worlds. And perhaps that is all the
logical positivists and ordinary language philosophers really cared about. Kripke’s
addenda can come to appear but a friendly amendment to the theses of logical
positivist and ordinary language theorists.
A Digression on Skeptical Paradox
I must digress for a moment to acknowledge that Kripke found another
problem, beyond any discussed so far, with the line of thought leading to the
conclusion that our knowledge of necessity reduces to knowledge of semantic rules
161
and linguistic conventions. It is a problem with the notions of rule and convention
themselves, one that according to Kripke casts a shadow not only over semantics
and linguistics but over large further tracts of cognitive studies.
Trouble had already been spotted by Quine in the mid-1930s, towards the end
of a famous paper, directed mainly against Carnap, on so-called truth by
convention. There are supposed to be infinitely many necessary truths, but the
human race has only had time to establish finitely many conventions, Quine
observes. It seems then that some necessary truths cannot have been established by
convention. It will not do to say that some finitely many key conventions are
directly established by conventions, while the rest follow from these. This will not
do because to take such a line only raises the question of what sort of a truth it is
that such-and-such a necessary truth follows from such-and-such a convention:
presumably it must be a necessary truth, yet on pain of infinite regress it cannot be
one established by convention.
The conventions in question are often supposed, as I mentioned in passing
earlier, to take the form of rules for the various logical particles such as or,
including the rule that A-or-B follows from A. To derive all the infinitely many
logical truths we need to make infinitely many applications of such rule,
recognizing that the viola is sharp or it is flat follows from the viola is sharp, and
the wine is sweet or it is sour follows from the wine is sweet, and so on. And what
seem necessary truths not directly conventional are such assumptions as the rule is
applied correctly in the viola case and the rule is applied correctly in the wine case
and so on.
This Quinean difficulty expands into a so-called skeptical paradox and rulefollowing in Kripke’ second book, the much-discussed Wittgenstein on Rules and
Private Language. There Kripke follows up the paradox with remarks about a socalled skeptical solution, but many find that a cure almost worse than the disease.
162
Kripke himself hints that he cannot accept it, and also hints that he favors some
alternative, nonskeptical solution that he is not expounding. But he has not
expounded it in the period of more than a quarter-century since his book, either.
Perhaps some exposition of it is there in the vaults of the Kripke Center; but I have
not seen it. Nonetheless, I will presume there is a solution, and even one that would
leave the conclusion that modal knowledge reduces to linguistic knowledge still
defensible if properly understood, however tricky may be the question of what the
proper understanding of it should be. But these are deep waters in which it is easy
to drown. And so I will close the present brief digression and leave this topic aside.
Nonsense about Necessity
The point we have reached is this, that Kripke has at least taught us that we
must be very careful about how we understand such a formulation as the claim that
our knowledge of necessity derives from knowledge of linguistic conventions, but
he has not really shown, or even really claimed very strenuously to have shown,
that such formulations are wholly off the mark.
There remains, however, to be evaluated the step from the intermediate
conclusion that knowledge of modal distinctions derives knowledge of linguistic
rules — presumably rules about using modal auxiliary verbs, such as could and
would and might — to the further conclusion that modal distinctions themselves,
and not just our knowledge of them, derive from linguistic rules, and the final
conclusion that the origin of necessity lies in ourselves, qua makers of language.
One immediate difficulty with these further steps is that the conclusion just
enunciated, along with all talk about the origin of necessity, is simply nonsense:
the problem of the origin of necessity is not a mystery but a muddle. For if one
takes the word origin at all literally, then to speak of the origin of necessity is to
speak of a time when necessity came into being, and before which there was none.
163
And that certainly seems to be nonsense. The number 29 not only is today but
always has been necessarily prime.
Even if we substitute ground for origin, to get rid of the temporal
connotations, in asking after the ground of something we still seem to be asking
what is it without which that thing would not have been, whereas the necessary as
opposed to the contingent is precisely that which would have been no matter what
as opposed to that which might not have been if only. So we seem to be asking,
What is it, without which that which could not have failed to be would have failed
to be? And surely such a question is nonsense.
As for our role in the matter, as makers of language, certainly any suggestion
that if only we had acted differently, if only we had adopted different semantical
rules or linguistic conventions, different things would have been necessary,
involves a kind of confusion, very much akin to the speaking-of-speaking-in
confusion.
Suppose, for instance, we had adopted the conventions of a base-thirteen
numeration system rather than our actual decimal or base-ten system. Then the
number that we wrote as a two followed by a nine, and perhaps even pronounced
twenty-nine, would indeed have been a number divisible by five, since it would
have been two times thirteen plus nine or 35 that we would have been calling
twenty-nine, and that number is five times seven. But our adopting a different
numeration system would not have changed any facts about the number 29, except
the fact that we call it twenty-nine and not twenty-three as we might if we used
base thirteen.
So in speaking of ourselves as the originators of modality, the philosophers of
the mid-twentieth-century will have been speaking nonsense if they were speaking
literally. But surely they, or the best of them, must be credited with having had the
wit to see this, and so we must take them to be speaking figuratively, even if they
164
make no particularly strenuous efforts to warn against literalism or to spell out in
terms that can be taken literally what they express figuratively by speaking of
ourselves, the makers of language, as being the creators of necessity. Let us grant
this, and let us allow ourselves to go on speaking figuratively for a bit. Even so, the
formulation according to which we are the creators of necessity is open to at least
one further challenge.
Even granting that what is needed to get from the kind of knowledge senseexperience can give us to knowledge of the necessity of necessary truths is
knowledge of linguistic rules, and even granting also that despite skeptical
paradoxes we do have the requisite kind of knowledge of linguistic rules, still it is
premature to conclude, even speaking figuratively, that necessity itself derives
from linguistic rules, and so from us.
For another interpretation of the situation remains possible, and logical
positivists and ordinary language philosophers have given no real argument against
this rival interpretation, but seem merely to have assumed that the rival view, to the
extent the were aware of it, is too metaphysical, in a pejorative sense, to be taken
seriously by analytic philosophers. The real problem or mystery of the so-called
origin of necessity lies in the clash between these two interpretations, and my next
goal must be to try to bring out what the conflict here consists in.
Theological Metaphor
Let me now try to describe the two rival views, and after that try to explain
the significance of their rivalry. I will give first the view that people like Ayer and
Strawson want to reject, and then the view they want to accept.
Here is the first view. Though the fact is not mentioned in Genesis, the first
thing God said on the first day of creation was Let there be necessity. And there
was necessity. And God saw necessity, that it was good, and God divided necessity
165
from contingency. And only then did He say Let there be light. Several days later,
Adam and Eve, in addition to introducing names for the animals, introduced modal
auxiliary verbs and rules for their use, rules according to which it can be said of
some things that they could have been otherwise, and of other things that they
could not. In so doing they were merely putting labels on a distinction that was no
more their creation than were the fishes of the sea or the beasts of the field or the
birds of the air that they were naming.
And here is the rival view. The failure of Genesis to mention any command
Let there be necessity is to be explained simply by the fact that no such command
was issued. We have no reason to suppose that the language in which God speaks
to the angels contains modal auxiliary verbs. Sometime after the Tower of Babel
some of the seventy-two nations found that their purposes would be better served
by introducing into their language certain modal auxiliary verbs, and fixing certain
rules for their use. When we say that such-and-such is necessary while such-andsuch is contingent, we are applying such rules, rules that are products of human,
not divine intelligence.
I have been allowing myself here to use theological language in describing
the distinction I want to draw. That might have been the natural way for
seventeenth or eighteenth century philosophers, nearly all professed theists or
deists, to discuss the matter. For many today, such language cannot be literally
accepted, and if it is only taken metaphorically, then I have not really done better
than those who speak figuratively and frame the question as that of whether the socalled origin of necessity lies outside us or within us. So let me drop the
theological language, and try again without it to describe without it the division
between two views that I have in mind.
166
Optical Metaphor
Well, here the first view: Ultimately reality as it is in itself, independently of
our attempts to conceptualize and comprehend it, contains both facts about what is,
and superfacts about what not only is but had to have been. Our modal usages, for
instance, the distinction between the simple indicative is and the construction had
to have been, simply reflect this fundamental distinction in the world, a distinction
that is and from the beginning always was there, independently of us and our
concerns.
And here is the second view: We have reasons, connected with our various
purposes in life, to use certain words, including would and might, in certain ways,
and thereby to make certain distinctions. The distinction between those things in
the world that we say would have been no matter what and those we say might
have failed to be if only is a projection of the distinctions made in our language.
Our saying there were necessities there before us must be called a retroactive
application to the prehuman world of a way of speaking invented and created by
human beings in order to solve human problems.
Well, that's my second try. With it I have gotten rid of theology, but
unfortunately I have not gotten rid of all metaphors. The key remaining metaphor
is the optical one: reflection versus projection. Perhaps I should give up the
attempt to get rid of all metaphors, and admit that the two views I am discussing
should be labeled not so much theses or doctrines as attitudes or orientations: a
stance that finds the reflection metaphor congenial, and the stance that finds the
projection metaphor congenial. But let me try a third time to describe the
distinction between the two outlooks in literal terms, avoiding optics as well as
theology.
167
The Order of Explanation
To begin with, both sides grant that there is a correspondence or parallelism
between two items or lists of items. On the one hand, there are facts about the
contrast between what is necessary and what is contingent, for instance, between
29 being a prime number and 29 being the number of years it takes for Saturn to
orbit the sun. On the other hand, there are facts about our usage of modal auxiliary
verbs such as would and might, for instance, the fact that we have no use for
questions of the form Would 29 still have been a prime number if such-and-such?
but may have use for questions of the form Would 29 still have been the number of
years it takes for Saturn to orbit the sun if such-and-such? The difference between
the sides concerns the order of explanation of the relation between the two parallel
ranges of facts.
And what do I mean by that? Well, both sides grant that 29 is necessarily
prime, for instance, is a proper thing to say, but they differ in the explanation why
it is a proper thing to say. Asked why, the first side will say that ultimately it is
simply because 29 is necessarily prime. That makes it proper to say that it is
necessarily prime, to assert the proposition that it is necessarily prime, and since
the sentence 29 is necessarily prime expresses that proposition, it is proper to utter
that sentence assertively. The second side will say instead that there is a rule of our
language according to which, or from which it follows, that 29 is necessarily prime
is a proper thing to say, and that is why it is a proper thing to say.
Note that the adherents of the second view need not deny that 29 is
necessarily prime. On the contrary, having said that the sentence 29 is necessarily
prime is, per rules of our language, a proper thing to say, they will go on to say it.
Nor need the adherents of the first view deny that recognition of the propriety of
saying 29 is necessarily prime is enshrined in a rule of our language. The adherents
of the first view need not even deny that proximately, as individuals, we learn that
168
29 is necessarily prime is a proper thing to say by picking up the pertinent rule in
the course of learning our language.
But the adherents of the first view will maintain that the rule itself is only
proper because collectively, as the creators of the language, we or our remote
answers have, in setting up the rule, managed to achieve correspondence with a
pre-existing fact, or rather, a pre-existing superfact, the superfact that 29 is
necessarily prime. The difference between the two views is, as I said, in the order
of explanation.
I will want labels for the two sides, and rather than invent new ones, I will
simply take two of the most overworked terms in the philosophical lexicon and
give them one more job to do. I will call the reflection view realism about
modality, and the projection view pragmatism. That at least will be easy to
remember, since realism and reflection begin with the same first two letters, as do
pragmatism and projection. And now, having introduced the realist-pragmatist
distinction, I need to say something about its significance.
To begin with, the two sides will, or ought to, recognize quite different
questions as the central unsolved problems in the theory of modality. For those on
the realist side, the old problem of the ultimate source of our knowledge of
modality remains, even if it is granted that the proximate source lies in knowledge
of linguistic conventions. For knowledge of linguistic conventions constitutes
knowledge of a reality independent of us only insofar as our linguistic conventions
reflect, at least to some degree, such an ultimate reality. So for the realist the
problem remains of explaining how such degree of correspondence as there is
between distinctions in language and distinctions in the world comes about.
If the distinction in the world is something primary and independent, and not
a mere projection of the distinction in language, then how the distinction in
language comes to be even imperfectly aligned with the distinction in the world
169
remains to be explained. For it cannot be said that we have faculties responsive to
modal facts independent of us — not in any sense of responsive implying that if
the facts had been different, then our language would have been different, since the
facts presumably couldn't have been different, these being modal facts that we are
speaking of. What then is the explanation? This is the problem of the
epistemology of modality as it confronts the realist, and addressing it is or ought to
be at the top of the realist agenda.
As for the pragmatist side, a chief argument of thinkers from Kant to Ayer
and Strawson and beyond for their antirealist stance has been precisely that if the
distinction we perceive in reality is taken to be merely a projection of a distinction
created by ourselves, then the epistemological problem dissolves, though to be
frank, that seems more like a reason for hoping the Kantian or Ayerite or
Strawsonian view is the right one, than for believing that it is.
In any case, the pragmatist side has an important unanswered question of its
own to address. The pragmatist account, as I formulated it just a moment ago,
begins by saying that we have certain reasons, connected with our various purposes
in life, to use certain words, including would and might in certain ways, and
thereby to make certain distinctions. What the pragmatist owes us is an account of
what these purposes are, and how the rules of our language help us to achieve
them. Addressing that issue is or ought to be at the top of the pragmatists' to-do
list.
Realism versus Pragmatism and Revisionism versus Preservationism
And this brings me to another important difference between realists and
pragmatists, a methodological difference over how much weight is to be given to
ordinary ways of speaking and commonsense ways of thinking in addressing
modal issues. For the realist-pragmatist distinction can be expected to track more
170
or less faithfully the revisionary-descriptive distinction discussed earlier. It is no
accident that Strawson’s name has appeared both as a representative of what I am
calling the pragmatist approach and of what he called descriptive metaphysics —
though to be sure, Ayer, whom I have also and equally set down on the pragmatist
side, had no use for any kind of metaphysics at all.
My suggestion is that pragmatists or projectionists about modality will or
ought to tend to be descriptivists, while realists or reflectionists will or ought to
tend to be revisionists. But let me use the alternate label preservationist for
descriptivist, so as to maintain the pattern of the sameness of the first two letters of
the relevant labels. I say tend to be because what I have in mind is a connection
that is not logically inevitable, but psychologically likely.
The root of the connection lies in a feature of our commonsense thinking
about modality and all metaphysical matters and indeed virtually all matters
whatsoever that cries out for description in Kantian terms. I gave earlier a famous
quotation from Kant, and at this point let me recall another still more famous, from
his Idea for a Universal History. I mean the line according to which From the
crooked timber of humanity, no straight thing was ever built. It is a feature of
anything created by us, including our systems of linguistic rules and conventions as
much as anything else, that it should be full of irregularities and anomalies. Our
rules and conventions may be called lawlike not in the sense in which philosophers
use that term, but in the contrary sense of being literally like the law.
And what is the law like? It is a hodge-podge of elements derived from
different sources: in the case of American law, ancestral customs of the Angles
and Saxons and surviving fragments of Roman law, principles that came into use
in courts of law and principles that came into use in courts of equity, bodies of
statute law and masses of case law, all several times at different periods halfoverhauled and semi-recodified by reform movements that achieved partial but not
171
total success before running out of steam.
In law hardly any principle is truly universal, like conservation of electrical
charge, say, in physics; rather, one has general rules with particular exceptions.
Specialists will think of the principle that so-called mens rea is required for
criminal culpability or civil liability, which has very broad scope, and yet
encounters exceptions in different varieties of so-called strict liability, with lots of
messy case law at the points were the two come into conflict.
Something as messy as the law is what a completely successful description
of our existing patterns of modal speech and thought would surely reveal, and no
one with the psychological temperament of a genuine realist is likely accept that
ultimate metaphysical reality is as messy as that. Our use of would and might and
the rest may reflect an ultimate metaphysical distinction between facts and
superfacts, but it will reflect it as a fun-house mirror reflects, subject to all sorts of
crooked human distortions. Hence revision will be needed if a truer, fuller
correspondence with extra- and pre-human realities is to be achieved. Hence the
realist will tend to be a revisionary.
A pragmatist might in principle be a reformer, since whatever purpose our
linguistic conventions may be thought to serve, they doubtless serve them
imperfectly and in a way susceptible to improvement. But in practice, until we
understand much better than we currently do just what purposes they are that our
use of modal auxiliary verbs is serving, we will hardly be in a position to suggest
how those purposes might be better served by a revised usage. And hence prudent
pragmatists — which admittedly need not include all pragmatists — should
provisionally at least be a preservationists.
So we now have two large-scale differences between the opposing outlooks.
Realists should tend to be revisionary, pragmatists to be preservationist. For
realists, perennial epistemological problems demand attention, whether or not they
172
are at present prepared to give them the attention they deserve, while pragmatists
less familiar questions about the utility of modal thinking come to the fore, or
ought to. Beyond all these very general differences, the realist and the pragmatist
will surely differ in their approaches to more specific puzzles and problems. Yet
the greatest difference may be less over what solution is proposed to this or that
problem or puzzle, than over the assessment of how much is at stake in such issues.
What the Stakes Are
From the premise that in some nonliteral sense modal distinctions are
ultimately being made up by ourselves, and could be altered if we found that doing
so served our purposes better, it seems a psychologically likely inference, though
admittedly not a logically inevitable implication, that how the modal distinctions as
we now make them apply in this or that case is not of earth-shaking importance.
Doubt as to the ultimate intrinsic interest of modal questions in general, and
the supervenience question in particular, will remain what I have called it, a
suspicion, so long as it remains only a suspicion that the pragmatist stance towards
modality is the right one, and not the realist. And the realism-pragmatism debate is
not one that can be expected to be settled any time soon. I do myself very strongly
suspect that the epistemological problems concerning modality are insoluble on a
realist basis, but this is very much a minority point of view. At present many and
perhaps most of those analytic metaphysicians seem to tend to hold that necessity
and possibility derive from features of essence and grounding, and seem equally to
tend to be willing to postpone indefinitely epistemological questions about our
supposed knowledge of essences and grounds.
While waiting for the pendulum to swing back towards taking seriously
worries about how we could ever come to know about such matters, it may be
observed that a divide parallel to that between realists and pragmatists about
173
modality is to be found also in the domain of so-called ontology, the general
theory of what sorts of things there are. In that domain I have already contrasted, in
the case of talk of facts, two stances. One may be called descriptive ontics, the
other revisionary ontosophy. The former is but a kind of glorified taxonomy, that
does not aim to go beyond listing all the sorts that are explicitly or implicitly
recognized in current scientific theorizing and especially in commonsense thought
— as in the linguistically-oriented studies of Moltmann on what she calls minor
entities. The latter is often willing and eager to discard, in the name of parsimony,
many of those sorts, a willingness and eagerness that manifests itself in
nominalism, physicalism, and other such ’isms.
Behind this descriptive-revisionary divide it is tempting to see a more
fundamental pragmatist-realist divide over the source of sortal classifications and
ontological categories. Here the realist seems to assume there is a — literally or
metaphorically — God-given list of genuine sorts of genuine objects, which our
classifications and identity criteria can at most hope to reflect. And the realist
seems to combine this assumption with another, to the effect that God, in creating
genuine objects, seeks always to economize on the number of sorts created. For it
seems to me that, for whatever reason — I do not pretend to understand the
connection fully myself — realism tends to go with a strong belief in the opposite
of Hamlet’s famous dictum, with a strong belief that there are fewer rather than
more things in heaven and earth than are dreamt of in our philosophy — or our
science, or our common sense.
Carnapianism
The pragmatist is more likely to follow here Carnap, in his famous discussion
of empiricism, semantics, and ontology in a paper of that title, in holding that
existence questions make sense only within a framework that we must set up for
174
ourselves, there being no one else to set it up for us, and that in setting up such
frameworks we need consult nothing beyond our own convenience.
If we choose to recognize feelings, distinguished by who has them and when
and what it is like to have them, then we may legitimately say there are feelings.
And a philosopher who seeks first merely to describe our existing categories before
presuming to criticize them must recognize that we have indeed already long since
chosen to recognize feelings. Which feelings there are, of course, is not for us to
stipulate, but rather for us to observe, which we do by having them, or rather, by
attending to our having them as we have them — by self-consciously
introspecting, as it is called. But that there are to be recognized any feelings at all
— any THINGS that we feel whenever we feel — depends on our implicitly
deciding, or rather, having long ago in remote prehistory decided, to recognize
them by setting up a framework for speaking of them.
For Carnapians it is not a legitimate objection to such a framework of
feelings, or to a framework of facts, or to a framework of functions in mathematics,
to say but such things as feelings, or facts, or functions don’t really exist, as if God
had already on the first day of creation chosen a framework, and all that was left to
us is to conform to it and put labels on its categories. A proposed framework may
be questioned on pragmatic grounds, by asking what work it will do for us, but not
on theoretical grounds, not by asking do such items exist in ultimate metaphysical
reality as it is in itself, independent of us and our concerns? Such forms of words
do not succeed in posing sensible queries, since existence questions make sense
only within the context of a framework already adopted.
Adopting a framework may be all right even though it is not uniquely right
even for human beings, let alone for extraterrestrial intelligences whose cognitive
faculties may operate in a very different way from our own. So-called external
questions as opposed to internal questions, questions about whether the categories
175
posited by a form of language or kind of theory are components of ultimate
metaphysical reality independently of us and our concerns, are to be rejected if
understood as anything more than infelicitously disguised formulations of merely
pragmatic questions about whether a framework is profitable to adopt.
Carnapian rejection of such questions is in the end based on a rejection of
metaphysics in anything like a traditional sense. And that rejection is based on a
perception that the subject is futile. Debate goes on and on without ever being
settled — I have compared it elsewhere to the endless lawsuit in Bleak House —
and this is because it takes place outside any agreed framework that would tell the
debaters what does and what does not count in favor of or against a given position.
For what it is worth, my own reluctance to engage in or with any metaphysics
but the unambitious kind that aims only to describe the categories we have is based
on a similar perception, as I have struggled with the literature, that the debaters in
other forms of metaphysics seem to be making up the rules as they go along; in this
sense my own position is Carnapian.
The parallelism between the realist-pragmatist divide over necessity and the
corresponding divide over existence will, I hope, be clear enough even from the
foregoing rough remarks. Indeed, to the extent that modality or our knowledge of it
derives ultimately from identity criteria, as I suggested earlier on in this discussion,
the two questions very nearly become one: not two separate though parallel strings,
but two strands in one braid. Taking both strands together, to the extent that the
pragmatist position is plausible, the interest of the supervenience thesis and
therewith the dualist-physicalist divide is, so to speak, doubly suspect.
176
VIII Beyond the Supervenience Issue
If one begins to doubt the interest and importance of the supervenience
issue, or equivalently, of the dualist-physicalist issue, what other issues would
survive giving it up? Let me take stock of where we are, and then address this new
question.
Tentative Conclusions
Are there facts about the sensations and affects of human beings that are not
physicalistic even in the broadest sense, that are neither neural nor functional nor
anything of the sort? Assuming denialism is to be rejected, it seems that descriptive
metaphysics must return to this question the affirmative answer that dualism wants.
For to establish the existence of a kind of thing by the lights of descriptive
metaphysics little more is required than to establish the existence of a kind of noun
phrase. And this is generally very easy to do, because our language abounds in
devices of nominalization, constructions that convert words and phrases belonging
to some other part of speech into noun phrases. And one of these is precisely the
construction consisting of prefacing a clause with the words the fact that.
Thus, if one denies denialism and grants that there is something that is left
unexplained, say, why pain feels the way it does, then it is very hard to avoid the
conclusion that there is some THING that is left unexplained, and specifically, the
fact that pain feels the way it does.
But a more important point may be that victory medals awarded by
descriptive metaphysics are never gold, but only brass. If descriptive metaphysics
tells us there are psychological facts distinct from any neural or functional facts,
this is so only because our usage of fact works the way it does.
Similarly with the direct argument for dualism based on identity criteria,
177
which makes being in pain a state distinct from any neural or functional state, but
only because we choose to count it so. In either case, dualists may score a victory,
but a victory that depends on little or nothing more than how we choose to use
words may seem a hollow one.
To the extent that there are factors, especially epistemological worries, that
may lead one to suspect that the only good metaphysics is descriptive metaphysics,
it may seem to matter little which side wins metaphysical debates over
supervenience or identity. This is the conclusion to which the line of thought
pursued in the preceding chapter seems to lead.
With arrival at that conclusion, I have finished giving voice or vent to my
suspicions that the supervenience issue is a pointless detour and pernicious
distraction, based on a popular delusion. Along the way I have also said something
about why officially I do not claim anything stronger than that there are grounds
for suspicion on all these points.
What remains before closing is to try to indicate, at least in rough sketch,
what might follow if these suspicions were taken to be ultimately correct, and were
taken seriously to heart. While the other chapters of this book have been negative,
devoted to building a case against becoming too involved with supervenience
issue, this last chapter will be more positive, containing suggestions as to what one
might become involved with instead.
Dualist-Physicalist Cooperation: Obstacles and a Proposal
The one consequence of taking suspicions about supervenience to heart that
I have hinted at already is that dualists and physicalists — by which I mean
hardcore dualists and nondenialist physicalists, the only kinds of dualists and
physicalists I want to consider further — should be able to agree to disagree about
the metaphysics of facts and identity so forth, and yet cooperate and collaborate in
178
pursuing more significant and substantial questions instead. And in a sense, all my
positive suggestions to follow can be subsumed under this head. But disagreement,
alas, even if both sides sincerely wish to put it aside, tends to hinder collaboration,
even if both sides sincerely wish to engage in it. This is because one result of
disagreement in doctrine is disagreement in diction: dualists and physicalists speak
in different ways, which can hardly fail to impede communication in any joint
enterprise.
To illustrate, and to revert to a hot topic I mentioned in passing early on, the
dualist will ask the following:
(1)
Do fish, in addition to having functional and neural states, have feelings?
To this question, any physicalist worth the name must answer in the negative, since
the physicalist view is that any feeling is quite literally one and the very same thing
as some functional state or neural state, not something additional.
Yet the nondenialist does recognize another question in the area:
(2)
Once the functional and neural states of fish have been explained, is there
anything left unexplained about them?
Here question (2) provides what may be called a physicalist’s surrogate or
substitute for the dualist’s question (1). But it would be tedious to have to go on
thus stating every question that comes up twice, and presumably each proposed
answer twice as well, in different forms, one suitable to the dualist, the other to the
physicalist — nor am I the best person to speak for the physicalist. There are two
solutions I can see to this problem, ways to work around persistent disagreements.
The first solution would involve the physicalist’s pretending to agree with
the dualist, using dualist diction as a convenient manner of speaking, but not taking
179
it too seriously. A parallel might be with my practice of pretending to agree that
there are such things as possible worlds, and falling in with the usage of true
believers, while not really taking it seriously. For I while I believe there are states
the world could have been in other than the one it is, I don’t really believe there are
any other worlds than this one, worlds that are in the states that this world is not in
but only could have been in.
Not taking entirely seriously the manner of speaking one adopts for
convenience means being willing to drop it and go back to what one considers the
more accurate manner of speaking whenever directly asked whether what one has
been saying is literally true, and whenever confusion threatens. In particular, one
must be ready to drop the pretence whenever a question is raised in the new
language that has no obvious counterpart in the old, and whose meaningfulness
must therefore be suspect — like the question whether the I in this world and the I
in that world are both the same myself.
An Alternative Proposal
The second solution would involve the physicalist’s in a sense actually
agreeing with at least the letter, if not the spirit, of dualism. The parallel would be
with a development that has taken place in philosophy of mathematics, to which I
turn for an analogy for one last time.
Philosophers of mathematics who call themselves nominalists used to deny
that there really are such items as, among other things, mathematical functions.
They then faced the problem of what to say about the fact that their doctrine was in
apparent conflict with the expert opinion of mathematicians, who accept many
existence theorems concerning functions. More recently some, without giving up
the nominalist label, have conceded that functions do exist after all, while denying
— and it is in this denial that the residue of nominalism in their view consists —
180
that functions are fundamental. This term fundamental is then further explained
and elaborated (perhaps in terms of “metaphysical grounding”).
The switch from claims of nonexistence to claims of nonfundamentality
rescues these soi-disant nominalists from being in the position of appearing to
claim that they know better than the mathematicians about mathematical questions,
of appearing to be engaged in what might be called an enterprise of philosophical
Besserwissenschaft. For while mathematicians do prove results they call existence
theorems that traditional nominalists appear to contradict, they do not prove any
results they call fundamentality theorems for philosophers to gainsay.
The switch marks and labels the issue of nominalism as belonging to the
kind of high metaphysics in which philosophers bat around terms like
fundamentality and grounding, which in contrast to terms like function and
existence, do not occur in mathematics. And so it becomes clear to mathematicians
— and to those philosophers of mathematics most interested in issues about
mathematical practice — that the issues being debated are ones with which they
need not concern themselves.
Physicalists might wish to ponder whether, without giving up the physicalist
label, they could make an analogous move, and recognize that after all sensations
and affects, as items distinct from any neural or functional correlates, do exist,
denying them only a certain fundamentality, insisting only that they are grounded
in something else.
I will write in the pages to follow as if one or the other of these proposals
has been accepted. That is to say, while still implicitly or tacitly claiming to speak
to both hardcore dualists and nondenialist physicalists, I will not state every
question twice, as I did with the issue about fish in (1) and (2), but rather will state
each question only once, and in dualist rather than physicalist form, giving the
analogue of (1) and leaving it to the reader to work out the analogue of (2). I will
181
directly address a dualist position, and leave to nondenialist physicalists the task of
working out a rewording in conformity with their principles.
Residual Questions Related to the Mind-Body Problem
The philosophical logician who has come to the supervenience issue from the
side of modal logic, as I have, may upon giving up the supervenience question
simply return to the theory of modality, as I expect I will, attempting to arrive at
some more definite conclusions in that domain than the suggestions and suspicions
and speculations I have been dealing in, but leaving philosophical psychology
behind.
What I wish to address here, however, is where the philosopher of mind who
has come to the supervenience issue instead from the direction of the mind-body
problem might turn upon coming to regard supervenience in the way that I have
been suggesting it should be regarded. As I have said, I will address this question
from a dualist point of view, or at least, in dualist language.
One conclusion a dualist might draw, upon losing interest in supervenience, is
that the mind-body problem itself is less interesting than had been previously
supposed. That is to say, it may be concluded that there is not much more to be
said on the topic once one has emphatically said two things.
First, there is simply no serious evidence of systematic interference of
immaterial minds with the physical world. Second, there is no serious hope that
any science at all resembling any of the sciences we actually have at present will
be able to offer any explanation of why there are conscious experiences associated
with certain physicalistic states, or of why the experiences there are are as they are.
Though the problem, or rather the vast complex of problems, involved in
locating the neural correlates of conscious states are enough to keep generation
after generation of scientists busy, and cannot in any serious senses be considered
182
easy, it is the issue of why there are any conscious correlates at all of neural states,
or at least of some of them, that is conventionally called the hard problem. Like
many I suspect it may well be an impossible problem.
After acknowledging that something is going on, something with which we
are all intimately familiar, but that we appear to be powerless to explain, there
seems nothing doable left to do. Or there is nothing except perhaps to warn that the
intellectual situation would not be improved by turning to nonscientific so-called
spiritual or occult inquiries — while at the same time urging that the kind of
intellectual arrogance involved in various ultimately denialist and ultimately failed
attempts to explain consciousness, as in various books reviewed in the NYRB, is to
be rejected, and that a certain intellectual humility in the face of a universe whose
ultimate mysteries are beyond us may be in order.
This acknowledged, the philosopher of mind may simply turn to other, less
intractable, issues. After all, since psychology is concerned with much more than
the conscious mind, philosophy of mind as well is presented with a range of topics
quite distinct from the traditional mind-body problem.
Two traditional aims of philosophy are what may be called speculation and
synthesis: on the one hand, the attempt to think in an organized and disciplined
way about issues not yet ripe for fully scientific treatment; and on the other hand,
the attempt to connect the technical results emerging from on-going scientific work
with our everyday, commonsense picture of the world. And the more actively an
area of science is advancing, the more scope there is for both kinds of
philosophical activity. And hardly any area of science at any period has been as
active as is the consortium known as cognitive studies, and such disciplines as
neuropsychology, at the present time.
And besides, there is the whole huge field of introspection and
phenomenological research.
183
One might, for instance, while denying that is an illusion that we are
conscious, hold on introspective — as well as experimental — grounds that
consciousness, though not an illusion in itself, is full of illusions. Continuity of
consciousness, or the existence of a single, unified, enduring self as the continuing
subject of consciousness, may well be an illusion. This is an issue at least related to
the traditional mind-body problem.
There are a number issues more directly related that remain as well, after
one gives up on supervenience. In earlier chapters I have touched on several, only
to set them aside after brief remarks, or even briefer parenthetical notice. I have, in
particular, taken a number of background assumptions for granted after only the
briefest acknowledgment if any that they could be critically questioned, and these
would be open to discussion even by discussants committed to remaining silent, if
not neutral, on supervenience and other modal matters.
Here I will mention just two background presuppositions that might be
doubted, a pair of assumptions about which critical queries can be and at least
implicitly have been raised. One faces doubts brought forward in various ways in
mostly unpublished work of Kripke, in middle-to-late period work of Putnam, and
elsewhere; the other confronts questions pressed in a couple of places by Chalmers.
The Status of Function-Talk
The challenge I associate with the names of Kripke and Putnam is to the
assumption I have made throughout that a certain three states are coextensive:
those of being in pain, of having certain neural events going on in one’s body, and
of being in a state that plays a certain role in our lives.
Dualists and physicalists of the two kinds I have considered, neuralist and
functionalist, disagree over which, if any, of these correlations are outright
identities, and over whether whatever correlations or identities obtain are apriori,
184
or aposteriori but necessary, or contingent; but these are the sorts of questions I am
putting aside when putting aside the supervenience issue. There remains the
background assumption, which thus far I have never questioned, that there is
indeed associated with being in pain or having C-fibers that are firing a certain
well-defined functional state.
What may be questioned here is whether when dealing, not with an artifact
manufactured to carry out certain functions, but with an organism that just grew
naturally, one may meaningfully speak of the function of various states obtaining
and processes proceeding within it. That is to say, may one meaningful speak of
the function, with a definite article? May one ascribe functions in an absolute
sense? Or must functional language as applied to a chunk of flesh and blood like a
human brain or a human body always be understood as relative to some particular
observer’s or interpreter’s way of analogizing the brain or body to a machine, or of
otherwise theorizing about brain and behavior, as the language of velocity must
always be understood as relative to a frame of reference?
Back in the days of Descartes, talk of goals, purposes, functions, and the like
was already being banished from physics. It persisted for a long time afterwards in
biology, until the theory of evolution by natural selection seemed to show how to
explain away the appearance of goals, purposes, functions, and the like on the part
of anything but human beings and any animals sufficient like them in relevant
respects. Talk of goals, purposes, functions, and so on, has come to be regarded in
biology as no more than a metaphorical shorthand for a more elaborate but in
principle purely physicalistic description of what has been going on.
By contrast, the current popularity of talk of function in philosophy of mind
may very well come to seem to the critical observer like backsliding from Darwin
towards Aristotle, or at best like systematic failure to make clear that certain
usages are not to be taken literally.
185
Whether I am in pain does not depend on how you choose to regard me; and
reciprocally, whether you are in pain does not depend on how I choose to regard
you. Thus being in pain cannot be a functional state if ascriptions of functional
states have always to be implicitly relative to some choice among analogies by
third-person observers making the ascriptions.
And therefore in taking the state of being in pain as at least having a welldefined functional correlate that is not thus relative, I have been making an
enormous concession to the functionalist position, for the sake of argument. If the
issue that was being argued, supervenience or the like, is now given up or set aside,
there is no reason to maintain the concession, and the question whether there is any
reason to make it may be reopened, or opened to serious inquiry for the first time.
There are practical obstacles to pursuing this line of inquiry, one of the more
important being the unavailability at this time of Kripke’s most important
contributions to the subject, since all that exists in print is one lengthy footnote in
his Wittgenstein book, and an illuminating but inevitably second-hand account
from Jeff Buechner of materials from the vaults of the Kripke Center.
Moreover, the discussion so far has been intricately intertwined with the
deep issue of skeptical paradox, and thus ultimately with the exegesis of the cryptic
writings of a cult figure, with all the distracting controversies and bitter polemics
such exegetical questions bring with them.
More importantly, perhaps, in the present context, the relevant writings
address explicitly only one form of functionalism, and that not the currently most
popular, namely, the kind that wants to apply the language of software and
hardware directly and explicitly to minds and brains. The same is largely true of
the later, antifunctionalist work of Putnam as well.
All I mean to claim is that we have in this area a serious question about the
ascriptions of function, that still makes sense even if one is determined to avoid
186
modal entanglements. Investigating this question is one item on the list of things
one might do instead if one gave up on the supervenience issue.
Metaphysics and Quantum Physics
The challenge I associate with the name of Chalmers is to another assumption
that I have made throughout, this one not quite so far in the background, for it is
closely related to the uniformity and closure principles. Namely, I have assumed
not only that anything like idealism is to be dismissed — since even if, as Hume
said, Berkeley’s arguments admit of no answer, still as Hume also said, such
arguments are produce no conviction, and thus are, as Hume called them, merely
skeptical — but also that the failure of parapsychology makes warranted the
presumption that there is no interference of immaterial minds and aphysical
consciousnesses with the material and physical world.
But Chalmers reminds us that Eugene Wigner and others, extrapolating from
the kind of rhetoric indulged in by Neils Bohr, and under his influence parroted by
a whole generation of physicists, have suggested that on the contrary only the
involvement of consciousness ever collapses so-called quantum superpositions and
puts the physical world into a definite physical state.
Here we confront a bizarre feature of the sociology of theoretical physics. The
formalism developed by the pioneers of quantum mechanics permit the calculation,
in some cases to many significant digits, of predictions of certain experimental
outcomes, which predictions have proved remarkably accurate. This is a great
intellectual achievement.
Yet quantum mechanics would clearly be a pretty idle luxury if it did nothing
more than to permit such calculations, if it told us about nothing more than
experiments going on in physics laboratories. And nonetheless many physicists
have been prepared to echo Bohr’s rhetoric about the theory applying only to
187
specified experimental situations.
For all that, one finds these same physicists, Wigner very much included,
citing quantum mechanics in connection with innumerable phenomena occurring
outside physics laboratories, from the spectrum of frequencies of solar radiation to
the lengths and strengths of chemical bonds. Unless one is an outright idealist,
these processes must be recognized to have been going on since long before the
appearance of physics laboratories, or of the human species, or of any known form
of life at all.
There is an obvious difficulty here, analogous to one that Berkeley faced in
the third of his dialogues, in the form of the objection that his view seemed to
imply that nothing was there before Adam was there to perceive it, though
according to scripture Adam was created only on the sixth day. And Berkeley’s
response, that God was there to keep an eye on things even during the first five
days, was hardly available to twentieth-century physicists, being incompatible with
the quantum perspective, since constant divine surveillance would presumably
keep all so-called wave packets collapsed.
There are practical obstacles to pursuing this line of inquiry, delving into
whether physical principles really do somehow rehabilitate something like a
Berkeleyan view minus God, with all its problems, at least to the extent of making
the physical universe’s being in any definite state dependent on their being
conscious observers present to take note of it.
For to explore the issue thoroughly, the philosopher of mind would have to
become a philosopher of physics, and that is an almost full-time occupation,
especially when one considers that it is not only the quantum mechanics of the first
half of the last century that would need to be explored, but quantum field theory as
well, and ultimately the so-called theory of everything, incorporating also
gravitation — a theory that does not yet even exist — since only that theory would
188
have a serious claim to being regarded as literally true, and not just a useful
approximation to something much more complicated.
Again all I mean to claim here is that we do have in this area a serious
candidate to be put down on the list questions that attention might turn to if it
turned away, as I have argued it should, from the supervenience issue.
Practical versus Theoretical
The two issues that survive when supervenience questions are set aside may
by some be considered rather abstruse, and the same must be said, I fear, about
some other such questions (for instance, those arising out of Chalmers’ intriguing
discussions of “dancing qualia” and “failing qualia”, fascinating as these may be).
In this feature such questions sharply contrast with another range of issues, that if
not themselves practical or moral rather than theoretical or speculative, are at any
rate directly relevant to practical and moral questions.
The practical and moral questions I have in mind are those pertaining to the
treatment of nonhuman entities. And there are a great variety of issues here. At one
extreme, there is widely thought to be value in the mere existence of biological
diversity, even of species of microorganisms, fungi, plants, or very simple animals
such as sponges that few are tempted to view as having mental lives. These are
issues that arise even for paramecia.
At the other extreme, it is widely thought that certain select nonhuman animal
species, beginning with but not limited to our nearest relations, may be endowed
not only with consciousness but with self-consciousness of a kind that is often
thought to bring with it a right to life that might be violated, not merely by
slaughter in an abattoir, but even by the kind of euthanasia to which pet owners are
often led to subject seriously ill dogs or cats. These are issues that arise, probably,
for porpoises.
189
I will consider only cases in the middle, and the case of fish, perches and
pikes and so on, issues about humane treatment that arise for creatures in that
range, rather than diatoms on the one hand or dolphins on the other. A large range
of questions arise even about such creatures, including issues about the
permissibility of using their flesh for pet food or their bone as plant fertilizer, of
keeping them in captivity for display in aquariums, of various modes of catching
them from angling to trawling, and above all of eating them.
There are questions falling within the domain of philosophy of mind, and by
no means unrelated to the traditional mind-body problem, that are clearly relevant
to these issues, and to which one might turn after becoming disillusioned with and
giving up on supervenience.
Methodological Issues
Inevitably, however, there is as Hume taught us long ago no immediate
inference from is to ought, leaving a gap between the theoretical and the practical.
Among the issues in the gap are complicated policy questions about decision
methods.
Standard Bayesian decision methods require weighting utilities and disutilities
by probabilities and improbabilities to calculate expectations, but there are no
agreed methods of aggregating utilities and disutilities of different subjects, let
alone of subjects of different species, primate and teleost. Moreover, the
probabilities generally have to be subjective credences, which troubles some,
though it troubles others no more than the fact that the values, the utilities and
disutilities, seem likewise subjective.
And in the literature, decision making by maximizing expected utility is often
challenged by proponents of other methods. In particular, it is often challenged by
an assumption cited as the precautionary principle, or rather, by several different,
190
incompatible assumptions to which such a label is applied.
For in the general literature on risk management, and the label is often applied
to a conservative principle warning against departures from established practice,
while in the specific literature on animal welfare the same label is often applied to
the principle that if there is any reason to suspect a practice causes harm, it should
be assumed to do so. Since there may be reasons to suspect that established
practices do cause harm, the one interpretation of precaution conflicts with the
other.
As my present focus is on philosophy of mind, I will take no stand on these
various issues, insisting that one policy is right and another wrong. The only thing
I will insist is that discussion of relevant theoretical questions should not be
corrupted or compromised by looking nervously over our shoulders at potential
practical and policy implications. The probabilities of various hypothesis of harm
or benefit should be assessed so far as possible from a value-free point of view, the
proper place for values to be brought in being at a later stage in the decision
process.
Misplaced Concern and Moral Hazards
Discussion of fish-related questions requires, of course, lifting the restriction I
have imposed in the body of this book to the case of human mentality, of human
feelings and thoughts; and if that restriction is lifted, questions not only about
animals aquatic or terrestrial may arise, but also issues about robots and aliens.
But sufficient unto the day is the evil thereof. We may wait to consider the
case of extraterrestrials until we are a little bit closer to having serious evidence for
believing there are any such entities. And the same may be said about artificial
intelligence, since what has been achieved in that direction so far, though
impressive as technology if no large philosophical are made for it, is not such as to
191
make anyone who has not wholly lost grip on the distinction between literal and
metaphorical language think there is any need for immediate moral concern over
our treatment of gadgets.
Universities all have committees concerned with the ethics of animal
experimentation and the welfare of experimental animals, but even those with the
largest robotics programs have no corresponding committees for the robots
involved, any more than for experimentation on plants — for while there are
ethical considerations pertaining to plant experimentation, relating for instance to
the danger of genetically manipulated organisms escaping from the laboratory or
greenhouse or experimental farm plot into the wild, these are not questions about
the welfare of the plants being experimented upon.
Nonetheless, it may not be premature to mention what may be called a moral
hazard in connection with the more exaggerated claims one sometimes hears being
made for automata, if not for the present, then for the not-too-distant future. For
while the faculty of sympathy, of concern for other entities, is capable of being
cultivated and expanded, it is not feasible, except perhaps in the case of a few truly
extraordinary individuals, to expand it without limit. It follows that eventually
wasting any concern on unworthy objects is going to mean some worthy ones will
be getting less concern than they deserve.
If crash-test dummies are made more sophisticated, so that they take
themselves off to the repair shop when damaged, or if so damaged to be unable to
do so, set of sirens calling for others to pick them up, some, especially among
functionalists, may perhaps want to claim they experience pain. But any time
diverted to committee meetings and conferences of ethicists devoted to setting
guidelines for the treatment of such dummies must in the end make less time and
energy available for considering the interests of animals — and even of human
beings.
192
But let me not myself be drawn for too long into the discussion of dummies
and robots. Let me rather return to the animals, fish — and if and where it matters,
more specifically bony rather than cartilaginous fishes — that have become the
topic of so much recent discussion. Here again there may be a certain amount of
misplaced concern, in the sense that the theoretical question that — on account of
its potential practical bearings — has become a focus of discussion, namely the
question whether fish feel pain, is perhaps not quite the right one.
Benthem’s Principle
This indeed seems not quite the right question for more than one reason. To
begin with, though the classical utilitarians often and wrote as if pleasure and pain
were simply the positive and negative ends of a single, unified, so-called felicific or
hedonic scale, we know they are not. For pain and pleasure can coexist.
We do not need to enter into consideration sexual masochism to appreciate
this fact. I vividly recall, for instance, taking a kind of perverse pleasure as a child
in the soreness of a wiggling a baby tooth not quite ready to come out, and I think
there was more to it than how good it felt when I stopped.
Even at a constant zero or nonzero level of pain there can be differences in
level of pleasure, and presumably these are of moral significance.
More importantly, it is not sensations of pain as such that count, or so I will
assume. For I will here assume just one premise as axiomatic, the principle
famously enunciated by Jeremy Benthem at the end of a long footnote in his
Introduction to the Principles of Morals and Legislation:
[T]he question is not, Can they reason? nor, Can they talk? but, Can they
suffer? (chapter 17, note 122)
193
As there are forms of suffering other than suffering from pain, even a creature
with total congenital analgesia may be capable of suffering, as such a creature may
be of feeling pleasure or the lack of it.
And even apart from all that there is the crucial point of which I have already
made much, that pain can occur in dissociation from suffering, and that in the
drugged state where pain is felt without the suffering that usually accompanies it,
the pain seems of no importance at all. If analogies between fish brains and human
brains are relevant, the most relevant parts of the human brain should be those
connected with suffering, rather than those connected with pain.
Discussion focuses on pain in fish perhaps only because, as elusive as it may
prove to be, it is more accessible than the question of suffering, and to that extent
may serve as a surrogate or substitute for it, though it must always be remembered
that any conclusions drawn as to whether fish feel pain must be regarded as only
provisional and suggestive and not conclusive or dispositive as to whether they
experience suffering.
At any rate, in the question whether there are creatures in pain around us we
have an issue that clearly does not just depend on or come down to issues about
how we choose to use words, or about what boundaries we choose to draw, or
anything of the sort, as I argued the supervenience question threatens to do, and at
several points and in several ways. The issue of pain in fish and other nonhuman
animals is, moreover, an issue where nondenialists, if I am not mistaken, and
especially those who emphasize the limitations of evolutionary explanation, should
be able to speak with something like a united voice, not perhaps about the ultimate
conclusion, but at any rate about what kinds of considerations are or are not
pertinent.
194
Animal Sentience
An immensely encouraging sign as to the intellectual health of the mental
sciences has been the launching of a new journal, Animal Sentience. It is edited by
Harnad, founder and for decades editor-in-chief of Behavior and Brain Science,
and the new journal follows the same format as the old: a so-called target article is
followed by a dozen or score of more of briefer comments by workers from a wide
range of disciplines, with brief replies by the original author.
Comparison of the titles of the two journals shows — though as anyone
acquainted with academic journals knows, the range of content in such periodicals
often is rather different from what their official titles might suggest — the
historical progress of psychology’s recovery from the intellectual disaster of
behaviorism. The older journal, founded in 1978, already breaks with the dogma
that anything inside the skin must be treated as a black box, and is prepared to treat
the brain as something more than an organ for cooling the blood; but it does not
depart so far as to mention consciousness as a topic separate from
neurophysiology. The newer periodical, founded in 2015, also avoids the word
consciousness in its title, to be sure, and this is perhaps a wise choice, given the
word’s ambiguities and its popularity with occultists; but the title openly and
boldly mentions sentience, a dirty word during the behaviorist period.
The known concerns of the main editor, and the institutional affiliations of the
associate editors, both from the Humane Society, which is also involved in
sponsorship of the enterprise, suggest a certain orientation, but the inaugural issue,
with a target article denying, rather than affirming, the reality of pain in fishes, and
the wide range of commentators, who differ not only in discipline but in views on
the target question, with most but by no means all disagreeing with the target
author, demonstrates the breadth of its focus, and its sincere, and so far as I am
competent to judge successful, desire to represent the full spectrum of scientific
195
and scholarly opinion on the target question.
I have had the privilege of seeing the whole contents of the issue prior to
publication, though it would be out of place for me to take advantage by
commenting in any specific detail while writing at a time prior to the journal’s
official appearance and general availability to the scholarly public. That the first
target article is by Brian Key is no secret, however, nor will I be revealing anything
knowledgeable readers couldn’t guess for themselves if I say that the line he takes
is not strikingly different from that taken in his previous publications on the same
and related topics.
The back-and-forth between the target author and critics illustrates the tension
between two main ways of arriving at judgments about the private, inner, mental
states of animals that cannot speak to tell us in human language whether they are
thinking or feeling, and if so what.
Correcting Overgeneralization
The tendency to view nonhuman entities as having feelings, sensations and
affects, as well as thoughts, beliefs and desires, emerges at too young an age for it
to be the product of any sort of logical argumentation, representable as a
conclusion drawn from certain premises. In this sense, the tendency is
spontaneous, though it is not unprompted.
It is cued in part simply by the look of the entities in question, and whether
they have something similar in gross pattern to a human face, but the main prompts
and cues are behavioral.
There can be little doubt, from an adult standpoint, that young children’s
spontaneous reactions tend to overgeneralize, to attribute feeling where none is
really present, and perhaps to attribute thought where none is present though
feeling may be.
196
Childhood memories are notoriously unreliable, but for what it is worth, I
seem to be able to remember, from around age three or four, at a time when I still
half-believed there were little people living inside the television set, my first
encounter with an automatic door at the grocery store. Such gadgets were novelties
at the time, the early fifties, at least in the part of the world where I grew up. And I
seem to remember being disappointed when I was disabused of the idea that the
door could, in any real sense, be said to see people coming.
It is conceivable that naive, childish, spontaneous tendencies undergeneralize
in some directions, but that they overgeneralize in others seems indisputable. If
there weren’t any costs involved, in the form of the danger of sacrificing the
genuine interests of beings that really are sentient to the supposed interests of
beings that are falsely imagined to be so, as well as the general dangers of
misplaced concern, it would not matter much how or whether the tendency to
overgeneralization was corrected.
But at least the grossest errors do get corrected. Children generally do learn to
think as typical adults do, that the cat really does hurt if its tail is pulled, while the
teddy bear, even if it is a fancy kind that can be said to talk in the sense of playing
back recorded messages, really does not feel anything, whatever one pretends
while indulging in make-believe.
Speaking in very general terms, there are two routes to critical retraction of
spontaneous overgeneralization: looking more closely at behavior, which often
involves imputing functions; and looking beneath external behavior to internal
mechanisms. The latter route only becomes available when one has some
substantial amount of knowledge of internal mechanisms in the one case where
there is no serious disagreement among sane persons as to whether feeling and
thought are present, namely, our own, human case.
The difference between the routes can be illustrated by a hypothetical case
197
suggested by the finger-on-the-hot-stove example. So suppose we have some type
of small aquatic worm, swimming in an aquarium with bristles in it here and there.
Those on one side have no particular effect, and the stimulus of being touched by
them produces no detectable response. Those on the other administer some kind of
electric shock, and if the worm is touched by one, it will very quickly twist way the
body segment in contact with it. The naive reaction is to suppose that the bristles
on the second side hurt the worm.
One way one might be brought to doubt this initial reaction would be by a
closer, more extended look at the worms’ behavior. Suppose it were found, for
instance, that the worms, despite sharply withdrawing from the bristles of the
second kind when touched by them, never come to avoid the side of the aquarium
where those bristles are present, as surely porpoises would if we scaled up the
whole experiment and substituted a more intelligent organism. The absence of
learning-behavior might — though then again it might not — lead to questioning
whether any pain was really felt.
Another way one might be brought to doubt the initial reaction would be by
internal, anatomical examination of the worms — supposing this can be done in a
way that would not be disapproved by university animal-welfare committees. I
mean more specifically, anatomical examination comparing what goes on in the
worm when it touches and electrified bristle and what goes on in a human being
when one touches a hot stove.
Suppose it were found that, though the worm has ganglia like those of our
spinal cords, there is nothing at the anterior or head end comparable to our brains,
and that the withdrawal upon being shocked only involves nerves connecting the
skin and muscles to the nearest ganglia, and is not accompanied by any movement
of an impulse up towards the head. The absence of any involvement of a brain
might not only explain the absence of learning behavior, but also lead to more
198
serious questioning whether any pain was really felt, or even practical certainty
that none was.
Piscene Pain
The most conspicuous difference between Key and his critics is that many of
the latter consider, as do many other writers on pain in fish, only behavioral
criteria, or at most behavioral criteria with some supplementary consideration
given to whether administering drugs that act as analgesics in human beings affects
the relevant behavior in fish in a way analogous to that in which it affects the
behavior in human beings. There seems to be hardly any question that if behavioral
and quasi-behavioral features were all that counted, it would have to be accepted
that fish feel pain.
Key’s argument that they do not, though it is complicated in a way I do not
have space or competence to discuss by his views on avian pain, is based on
examination of the underlying neurophysiology. To a first approximation, what is
found is that fish are lacking some of the neural and encephalic structures that
seem to be crucial to pain in the case of human beings, or of mammals whose
capacity to feel pain is not seriously doubted.
Though it is only a pretence that in human beings the firing of C-fibers is the
exact correlate of pain, it is not a pretence based on nothing. Reportedly, the
Swedes with congenital analgesia have far fewer of those fibers than do the rest of
us, for instance. Fish seem to have even fewer still, though Key’s main emphasis is
not on nerve fibers but on the parts of the brain known or suspected to be crucial in
the case of human pain.
These structures, or ones parallel in origin and operation — structures that are
homologous (as the term is understood in embryology and “developmental
biology”) to those involved in human pain — simply are not found in fish, it is
199
claimed. If, for the sake of argument, we provisionally grant that reliance on purely
behavioral criteria would pose a strong threat of overgeneralizing, and grant
structures homologous to human pain-related structures are absent in fish,
inevitably attention will turn to whether structures that can at any rate be called
analogous to human pain-related structures are present.
How much analogy, or similarity or function without similarity of form. is
present, turns out, unfortunately, to be an area where the neurological facts
themselves seem to be in some dispute, or subject to some divergent
interpretations.
A conclusion looms that was argued for carefully and at length by Marian
Stamp Dawkins in her recent book, to the effect that even those firmly convinced
of animal sentience would be more likely to achieve their practical goals and
purposes if they based their public advocacy on considerations that may appeal
even to those who don’t share that belief. This is so because we aren’t going to
settle the question of sentience any time soon.
Envoi
When it becomes clear that there are disagreements about the
neurophysiological facts, it straightaway becomes clear that there are limits to how
much a philosopher qua philosopher can contribute in this area. Perhaps the
philosopher can contribute nothing but occasional reminders.
For I do find that some of the writers in the new journal, the many who appeal
to considerations of so-called convergent evolution, which produces structures that
are analogous but not homologous, seem to have forgotten a crucial point involved
in the denial of denialism. They seem to forget — something they must surely have
read in the works and correspondence of the new journal’s editor, if nowhere else
— that feeling is not a product of evolution in the sense of being something that
200
was selected for.
It is only a by-product, through psychophysical correlation laws we are unable
to explain, of neural correlates that were selected for. Pain as such has no survival
value, but is only correlated with something else that that does, and unfortunately
by laws of correlation whose ultimate sources are unknown to us at present and
perhaps unknowable to us for all time.
It therefore is not obvious that neural states that are analogous but not
homologous to ones that in humans are the neural correlates of mental states such
as being in pain themselves have any mental states correlated with them. Nor, of
course, is it obvious that they don’t.
But once again, and for the last time, let me say that my present concern
with this problem area, as with others before it, is not a concern to try to settle the
issues in it, but simply to call it to the attention of philosophers of mind as
providing a collection of issues they might turn to if they turned away from the
issue of supervenience.
That is something which, as I have argued, there are serious reasons to
suspect that they ought to do.
201
References
Entry…