Kripke’s Proof Is Ad Hominem not
Two-Dimensional
David Papineau
1. Kripke’s Argument
In the final pages of Naming and Necessity Kripke offers an
argument against mind-brain identity theories.
It runs like this.
Identity
theorists make claims like ‘pain = C-fibre stimulation’. These claims must be necessary if true, given
that terms like ‘pain’ and ‘C-fibre stimulation’ are rigid. Yet there is no doubt that such claims appear
contingent. It certainly seems that
there could have been C-fibre stimulation without pains or vice versa. So identity theorists owe us an explanation
of why such claims should appear contingent if they are in fact necessary.
One
model for such an explanation would be the story we use to explain why scientific
identity claims like ‘heat = molecular motion’ appear contingent. In that case we can say that this appearance of
contingency is due to our recognition of the genuine possibility that molecular
motion might not have been felt as heat.
However this won’t work in the mind-brain case. There is no corresponding possibility that
C-fibre stimulation might not have been felt as pain, if pain is in fact
C-fibre stimulation. Pains can’t be pulled
apart from their appearance, in the way that heat can. A situation in which C-fibre stimulation isn’t
felt as pain is a situation which lacks pain, not just the appearance of
pain. And this possibility—of C-fibre
stimulation without pain—is inconsistent with the identity theorist’s
insistence that pain is C-fibre stimulation.
Given
that no other explanation of the apparent contingency of ‘pain = C-fibre
stimulation’ offers itself, identity theorists have no option but to accept
that this claim is not necessary after all, and so not true.’
This paper will first focus
on the correct exegesis of this argument and will then consider what
substantial points it establishes.
2. The Orthodox Interpretation
There is currently a widespread consensus on the way to read this
argument. According to nearly all
contemporary commentators, when Kripke refers to the ‘appearance of
contingency’ displayed by mind-brain identity claims, he is in effect talking
about the a posteriority of these claims. Of course, these commentators do not think
that Kripke equates a posteriority with contingency itself—that would run
counter to his fundamental separation of epistemology and metaphysics. But they do think that Kripke takes a
posteriority to give rise to an appearance of contingency. (After all, the thought runs, a posteriority
leaves it open whether or not a claim is true.)
Furthermore, according to this contemporary consensus, Kripke holds that
such an appearance of contingency can only be explained if some term in the relevant
claim picks out its referent by contingent description, or at least can plausibly
be understood as having such a descriptive content. Thus, in the heat = molecular motion case,
the term ‘heat’ can be understood as equivalent to ‘the cause of our sensations
of heat’. This then enables us to
(mis)understand that claim ‘heat = molecular motion’ as ‘the cause of our
sensations of heat = molecular motion’.
Since this latter claim is genuinely contingent, this then offers an
account of the appearance of contingency associated with the former claim. However, no such descriptive reading is
available in mind-brain cases, which means there is no way to explain why they
too should appear contingent.
So the contemporary consensus reads Kripke as assuming that a posteriori
necessities are always due to the presence of descriptive content. In support of this interpretation, we might
suppose that Kripke was thinking along something like the following lines. When we think of some referent via a
description, we are thinking of it at second hand, thinking of it via some
associated property. So thinking of this
kind may well conceal necessities from us—the indirecteness of our thought may
prevent necessary properties from being a priori discernable. By contrast, if we are thinking of something
directly, without the mediation of some contingent description, then any
necessary properties will be available a priori. (And then, in the case of pain, where our
thoughts clearly aren’t mediated by any description, any necessary property of
pain, such as its putative identity with
C-fibre stimulation, ought to be available a priori. But pain’s identity with C-fibre stimulation
clearly isn’t available a priori. So
pain can’t in fact be identical with C-fibre stimulation.)
On this reading, then, Kripke’s argument hinges crucially on the
principle:
(I) If a necessarily true claim
is a posteriori, then at least one of its terms must be associated with a
descriptive content.
Some of those who hold that Kripke is committed to (I) also take him to
be implicitly committed to the doctrines of ‘two-dimensional semantics’ and in
particular to the claim that every a posteriori necessity has an ‘unstable primary
intension’ which is sensitive to the actual-world facts in the way that the
denotations of contingent descriptions are.[1] Other philosophers hesitate to attribute any such
systematic view to Kripke on the basis of the discursive and unmethodical
remarks in Naming and Necessity, but even so feel that there is enough
in that text, and in particular in the implicit logic of the anti-materialist
argument at the end of the book, to make it clear that Kripke must accept
something along the lines of principle (I).
Let me cite some examples of the orthodox interpretation from the recent
literature.
In his ‘Phenomenal States’ (1997) Brian Loar
asserts that Kripke’s anti-physicalist argument hinges on the assumption that
‘the only way to account for the a posteriori status of a true property
identity is this: one of the terms expresses a contingent mode of presentation’
(p 600).
Christopher Hill (1997) puts
the point in terms of inferences from conceivable distinctness to real
distinctness. The conceivable
distinctness of some commonsensical kind X (eg pain) from a theoretical kind
(C-fibre stimulation) means that their identity can at best be a
posteriori. So principle (I) would mean
that such conceivably distinct kinds can only be identical if at least one is
picked out descriptively. In line with
this, Hill attributes to Kripke the view that such conceivable distinctness
does imply real distinctness unless the commonsensical kind at issue is
associated with ‘a property that normally guides us in recognizing instances of
X, but that is only contingently connected with X’ (p63).
Again, Joseph Levine’s Purple
Haze (2001, pp 47-8) presents Kripke as assuming that conceptual
possibility implies metaphysical possibility save in cases where the claim at
issue can be reinterpreted in terms of some property we use to pick out the
relevant kind. And in his Thinking
about Consciousness (2002, pp 92-3) David Papineau attributes to Kripke the
‘transparency thesis’ that necessary identities will be a priori unless one of
the terms refers by contingent description.
In the Introduction to their
anthology on two-dimensional semantics (2006) Manuel Garcia-Carpintero and Josep
Macia allow that there is no explicit commitment to anything like (I) in Naming
and Necessity: they say that Kripke
avoids any general claim that any posteriori necessities must hinge on the
presence of a descriptive term that gives rise to ‘an appropriate corresponding
qualitative statement’. Nevertheless
they argue specifically that his argument against mind-body identity ‘depends
essentially’ on this assumption, and so that he must be implicitly committed to
it (p 2). Here they are following David
Chalmers line in The Conscious Mind (1996). Chalmers similarly maintains that Kripke’s
argument suggests an ‘implicit endorsement of the two-dimensional framework’ (p
149).
Stephen Yablo goes so far as
to give the name ‘Textbook Kripkeanism’ to the view we can move from
conceivable possibility to metaphysical possibility in cases where ‘no
obfuscating presentation can be found’ (‘Textbook Kripkeanism and the Open
Texture of Concepts’, 2000, p 100). It
should be noted, however, that Yablo himself is non-committal about this interpretation. (‘How well it corresponds to any actual
belief of Kripke’s is hard to say, and something I take no stand on.’)
3. Can Kripke Really Be A Textbook Kripkean?
Can Kripke really be committed to principle (I)? It seems to me that such a commitment would
run quite counter to the central doctrines of Naming and Necessity. After all, much of the book is devoted to
persuading readers that most ordinary proper names lack descriptive content,
but rather have their referents fixed causally.
Given this, it looks as if proper name identities are prima facie
counter-examples to principle (I). These
necessities are manifestly a posteriori, but arguably lack any descriptive
content. Yet Kripke seems quite
unconcerned about this. Maybe there is
more to say here—maybe the text leaves it open that Kripke thinks proper names
all carry some kind of associated descriptive content alongside their causal
referential ties. But the point remains
is that Kripke himself does not say anything more in Naming and Necessity,
as he surely would if his anti-materialist argument did hinge on principle (I).
Further, consider the epistemological implications of principle
(I). This principle says that necessary
claims can only be a posteriori if some of their terms refer by
description. It follows immediately that
when reference is direct, and not mediated by description, then all necessary
properties will be a priori—we will be acquainted directly with the objects,
and so their nature will be transparent to us.
I see no hint of this kind of Russellian epistemology in Kripke. He has plenty to say about reference that
isn’t mediated by description, but no suggestion that such reference renders
all necessary properties epistemologically transparent.
I shall have more to say against the orthodox reading of Kripke’s
anti-materialist argument below. But
first it will be helpful to show that there is an alternative way of
understanding Kripke that does not commit him to Principle (I).
4. Kripke’s Ad Hominem Argument
I think that Kripke’s argument is best read as an ad hominem challenge
to identity theorists, rather than as a manifestation of some implicit
Russellian epistemology. It is
instructive to see how Kripke formulates his argument From the beginning he presents the issue as a
problem arising specifically for ‘identity theorists’—those who embrace claims
like ‘pain = C-fibre stimulation’—and not as an instance of some general
constraint on a posteriori necessity.
Thus he says:
. . . the identity theorist is
committed to the view there could not be a C-fibre stimulation which was not a
pain nor a pain which was not a C-fibre stimulation. . . . Can he perhaps show that the apparent
possibility of pain not having turned out to be C-fibre stimulation . . . is an
illusion? . . . Now I do not think it likely that the
identity theorist will succeed in such an endeavour. (pp 149-50)
There is no suggestion here that there is some general problem about
posteriori necessities with no descriptive content. The problem isn’t that any such necessity
ought to be a priori available to every thinker. Rather Kripke is pointing out that there is a
specific issue facing thinkers who actually believe that pain = C-fibre stimulation. These people need to explain why this
identity still strikes them as possibly false. People who aren’t committed to the
identity will of course countenance worlds in which pain turns out not to be
C-fibre stimulation. But once you
embrace the identity it ceases to be at all clear how you can still have room
for the thought that pains might not have been C-fibre stimulation.
What are you supposed to be
thinking, when you think this? You think
that pain and C-fibre stimulation are one and the same thing. So how can you countenance a world in which
‘they’ come apart? Of course, if ‘pain’
(or ‘C-fibre stimulation’) had some descriptive content, they you could say
that you were imagining a world in which something else satisfied the
description is question, just as we can imagine a world in which molecular
motion doesn’t satisfy the description associated with ‘temperature’. But there isn’t any such descriptive content
available in the pain = C-fibre stimulation case. So you really can’t explain why this identity
strikes you as contingent, given your contention that it is actually true.
Consider this analogy. People who do not know that
Of course, in proper name
cases there may be some surrogate contingency in the offing, associated with
your attaching some descriptive content to ‘
On my reading, then, Kripke’s
challenge isn’t to explain how mind-brain identities are a posteriori—as it
were, to explain how they can appear possibly false to people who don’t yet believe
them. Rather his challenge is to explain
why they still appear possibly false, even to people who do
believe them. Given that there is no
descriptive content to hand, there seems no room for someone simultaneously to
think that pain is C-fibre stimulation, but that it might not have been.
So construed, Kripke’s
argument still hinges on the claim that ‘pain’ (and C-fibre stimulation’) lack
descriptive content, just as it does on the orthodox interpretation. But on my reading he isn’t committed to Principle
(I) at all. Rather his crucial premise
is:
(II) If a necessary truth still seems contingent
after it is believed, then it must have some descriptive content.
5. Kripke’s Actual Argument Can’t Be Blocked As
Easily As The Orthodox Interpretation
To further bring out the
difference between my reading of Kripke and the orthodox interpretation, note
that the standard physicalist rebuttal of the orthodox interpretation is
ineffective against Kripke as I am reading him.
On the orthodox
interpretation, as we have seen, Kripke’s argument is that mind-brain
identities lack the descriptive content that principle (I) says is
characteristic of any a posteriori identity.
To this the normal physicalist response is to deny principle (I). Contemporary physicalists simply insist that
there can be a posteriori identities even in the absence of any descriptive
content. Sometimes two terms can refer
directly and yet their co-reference not be a priori.[2]
However this is no help in
answering Kripke’s argument as I am construing it. Even after you deny principle (I), you will
still face Kripke’s ad hominem argument.
Denying principle (I) allows
materialists to hold that mind-brain identities are a posteriori, and therewith
to explain how these identities can appear possibly false to people who do not
yet believe them. But this doesn’t help
to explain how such description-free identities can still appear possibly false
even to people who do believe them.
Let us agree that pain = C-fibre stimulation is a posteriori and so
rests on empirical evidence. Still, how
come this claim still appears possibly false to us, even after we have
the evidence? Shouldn’t this appearance
simply disappear, once you accept that pain is C-fibre stimulation, and
have no descriptive content with which to create a contingent surrogate for
this claim? Kripke’s challenge as I am
construing it simply isn’t addressed by the standard physicalist insistence
that mind-brain identities can be a posteriori.
6. Further Evidence Against The Textbook
In section 3 I made two
initial exegetical points against the orthodox interpretation of Kripke:
(i) He doesn’t offer any explanation of why
proper name identities aren’t counterexamples to principle (I), as one would
expect if he endorsed that principle.
(ii) He doesn’t seem to embrace the kind of
Russellian epistemology—direct reference renders essential properties
epistemologically transparent—that is implied by principle (I).
Let me now add to the
exegetical case. For a start, note how Kripke
does not object directly to the idea that mind-brain identities are a
posteriori, as we might surely expect if he were arguing on the basis of
principle (I). After all, principle (I)
says that mind-brain identities must be a priori if true. So why doesn’t Kripke simply argue that it
absurd to suppose that ‘pain = C-fibre stimulation’ might be known a priori? However, this isn’t what he does. Rather he starts talking about the possible
ways in which ‘the identity theorist’ might explain the ‘appearance of
contingency’. This strongly suggests
that when he talks about the ‘appearance of contingency’ he is referring to
something more complicated than the mere a posteriority of mind-brain
identities.
It is also noteworthy that
Kripke here uses the specific phrase ‘appearance of contingency’. If the complaint were simply that mind-brain
identities can’t be a posteriori, then the problem would be to explain why they
should appear possibly false to anybody.
But Kripke doesn’t ask for an explanation for the appearance of possible
falsity, but more specifically for the appearance of contingency. I take ‘contingency’ in this context to mean
contingent truth, that is, the combination of truth in the actual world with
falsity in other possible worlds. It is
this specific combination that ‘the identity theorist’ seems to be committed
to, and which I say Kripke views as problematic. If Kripke were not arguing ad hominem against
‘identity theorists’ who are already committed to the truth of mind-brain identities,
then it would be mysterious that he asks about the appearance of ‘contingency’
rather than a mere appearance of possible falsity.[3]
In addition to all these
implicit considerations against the orthodox interpretation of Kripke, there is
also some direct textual evidence that he rejects principle (I).
Thus, just before he starts on
his anti-materialist argument, Kripke is concerned to defend the necessity of rigidly
framed identity claims, against such thoughts as that Hesperus might not have
turned out to be Phosphorus. Kripke
explains that such ‘might have turned out’ thoughts hinge on reinterpreting the
relevant claims in term of something like descriptive content. Thus there is a possible world in which the
star seen in the morning is not the star seen in the evening, even though the
actual Hesperus is necessarily the actual Phosporus. More generally, rigidly framed identity
claims can often be re-read as contingent claims about objects which satisfy
salient qualitative criteria. Here is
how Kripke puts the point.
Any
necessary truth, whether a priori or a posteriori, could not have turned out
otherwise. In the case of some necessary
a posteriori truths, however, we can say that under appropriate qualitatively
identical evidential situations, an appropriate corresponding qualitative
statement might have been false (p 142, my italics).
I have italicized the ‘some’
in the second sentence. This seems to
tell decisively against the view that Kripke’s anti-materialist argument
assumes principle (I). If he were about
to embark on an argument that hinged crucially on the premise that all a
posteriori necessities involve some kind of descriptive content, surely he
would not say, two pages earlier, that such contents are associated with ‘some’
a posteriori necessities. The clear
implication is that Kripke takes examples like ‘Hesperus’ and ‘Phosphorus’ to
be relatively special among proper names, and accepts that in other cases no
‘corresponding qualitative statement’ will be available. If this is right, then his argument against
mind-brain identities can’t possibly be that there are no brute
(non-descriptive) a posteriori necessities.
There is a similar passage a
few pages later, in the middle of the anti-materialist argument itself. Kripke is now considering the ways in which
the ‘identity theorist’ might explain the ‘appearance of contingency’.
What
was the strategy used above to handle the apparent contingency of certain cases
of the necessary a posteriori?
The strategy was to argue that although the statement itself is
necessary, someone could, qualitatively speaking, be in the same
epistemic situation as the original, and in such a situation a qualitatively
analogous statement could be false (p 150).
Kripke then proceeds to
explain how this strategy won’t work in the mind-brain case, for lack of any
suitable qualitatively analogous contingent statement.
Note how Kripke here refers
to the strategy used to ‘handle the apparent contingency of certain
cases of the necessary a posteriori’. Kripke
does not say that the qualitative strategy will be available in all
cases of the necessary a posteriori.
Rather the suggestion is that it ought to be available specifically in ‘certain’
cases—namely, those that display ‘apparent contingency’ by continuing to appear
possibly false even after they are believed to be true. This seems to make it quite clear that
Kripke’s argument isn’t that descriptive content is needed for a posteriori identities
per se, but rather that it is needed specifically for identities that continue
to seem possibly false after they are believed
7. Kripke Shows That Physicalists Don’t Fully
Believe Their Physicalism
In my view, Kripke’s actual
ad hominen argument poses a genuine problem for physicalism—a problem that
isn’t addressed by the physicalist response to the orthodox
interpretation. In effect, Kripke shows that
with mind-brain identities there’s no gap between an appearance of possible
falsity and a belief in actual falsity.
If you believe a mind-brain identity to be true, then it can’t appear
possibly false to you, due to the absence of qualitative surrogates. So its appearing possibly false is
incompatible with your believing it.
I think that Kripke is quite
right about this. In sections 9 and 10
below I shall aim to show that there are no gaps in his line of reasoning. But first, in this section and the next, I
want to clarify the larger dialectical situation.
I say that Kripke
demonstrates that, if a mind-brain identity appears possibly false to you, then
you can’t believe it. And I accept that mind-brain identities do appear
possibly false to physicalists. However,
I don’t think that this disproves physicalism.
For there is another option open to physicalists, apart from accepting
that physicalism is false. Instead they
can admit that they don’t fully believe their physicalism. That is, they can agree with Kripke that mind-brain
identities won’t appear possibly false to someone who fully believes them—and
so conclude, from the fact that such identities do strike them as possibly
false, that they don’t fully believe them.
This might seem tantamount to
abandoning physicalism itself. But this
need not follow. The idea would be to
disbelieve physicalism at an intuitive level, while continuing to be
committed to it at a theoretical level.
There are plenty of good models for this kind of doxastic split
personality. Consider the familiar
Muller-Lyer illusion. At a theoretical
level, we know that the lines are the same length. But at a more intuitive level of judgement
they strike us as of different lengths.
Nor is this kind of set-up restricted to cases involving perceptual
illusion. At a theoretical level, I am
entirely convinced that there is no moving present and the B-series description
of reality is complete. But at an
intuitive level I can’t stop myself thinking that I am moving through
time. At a theoretical level, I am
persuaded that reality splits into independent branches whenever a quantum
chance is actualized. But at an
intuitive level I can’t shake off the belief that there will be a fact of the
matter about whether the Geiger counter will click in the next two seconds or
not. And so on.
Physicalists can appeal to
the same model in the mind-brain case.
At a theoretical level, they fully accept mind-brain identities. But at an intuitive level something is
stopping them embracing these beliefs.
This will then allows them to
explain the appearance of possible falsity attaching to claims like ‘pain =
C-fibre stimulation’, consistently with their theoretical commitment to these identities. This appearance of possible falsity arises at
the intuitive level—at bottom it is simply the intuitive thought that
pains and C-fibre stimulation are actually distinct. That’s why it seems intuitively that there
could be beings with C-fibre stimulation but not pain. At the theoretical level, on the other hand,
possible falsity will be ruled out—once we keep it firmly in our theoretical
minds that pain is C-fibre stimulation, we will find no room for the
thought that ‘they’ might come apart, and so will dismiss the possibility of
C-fibre stimulation without pain.
At both levels, then, Kripke’s
argument is respected. There’s no
question of simultaneously believing pain is C-fibre stimulation and in the
same mode also thinking they mightn’t be identical. We only think they might be different at the
intuitive level where we think they are different; at the theoretical
level where we think they are the same we don’t think that they might be
different. (Note that at neither level
do we get an appearance of contingency—an appearance combining actual truth
and possibly falsity. That I take to be
ruled out by Kripke’s argument. Rather
we get an impression of actual falsity—a fortiori possible falsity—at the
intuitive level, and an impression of actual truth—and so no possible
falsity—at the theoretical level.)
I don’t take this line of
response to Kripke to weaken physicalism.
Indeed I think that it is positively advantageous for physicalists to
recognize that they are in the grip of a persistent dualist intuition. It is widely supposed that orthodox
physicalism leaves us with some kind of ‘explanatory gap’ and so that some kind
of extra resource—some new way of thinking about the physical world, perhaps—is
needed fully to vindicate physicalism.
The idea that we all are stuck with a persistent intuition of dualism
casts a new light on this issue. Perhaps
there is nothing more to the ‘explanatory gap’ than the strong intuition that
C-fibre stimulation is one thing and pain another—that in itself would make us
wonder why this extra feeling should be present whenever C-fibres are
stimulated. If this is right, then
physicalism won’t need any extra resources to bridge the ‘explanatory gap’. It will be enough to recognize that we are
all in the grip of an intuition that physicalism is false.
Of course, this analysis does
leave us with the question of why we should all be subject to this dualist
intuition. A fully satisfactory
physicalism would need to explain why physicalism is so hard to believe
intuitively. Still, there seems no
principled difficulty here. The current
literature contains a number of suggested explanations for such an intuition of
dualist distinctness. (Bloom 2004,
Melnyk 2003, Papineau 1993, 2002, 2006.)
However this is not the place to pursue this issue. My present concern is only to show that
there is an intuition of distinctness.
The further question of why this intuition should arise can be
left for another occasion.
8. A Final Exegetical Point
I say that Kripke has a good argument to show that even physicalists intuitively
disbelieve physicalism. Let me lay out
the argument explicitly.
Argument A
(1) If you fully believe that pain
= C-fibre stimulation, then this can’t appear possibly false to you.
(2) Pain = C-fibre stimulation
does appear possibly false, even to physicalists. So
(3) Even physicalists don’t fully
believe that pain = C-fibre stimulation.
In the final two sections I
am going to consider possible objections to this argument. But first I want to consider one final
exegetical point. It might be objected
that this can’t possibly be Kripke’s argument, because Kripke is arguing for
the metaphysical conclusion that physicalism is false, whereas this arguments
leads us only to the psychological conclusion that physicalists don’t believe
their physicalism.
This is a reasonable point. Even so, I would argue that my reconstruction
fits nearly everything that Kripke says in the last pages of Naming and
Necessity. After all, it would certainly
make a kind of sense for Kripke’s argument to have a psychological
conclusion. For there is no question, I
take it, but that his argument hinges crucially on a psychological premise,
namely, that mind-brain identities ‘appear contingent’. Remember, Kripke’s argument isn’t that these
identities are contingent—he concedes from the start that this would beg
the question against physicalism. Rather,
he immediately retreats to the claim that they ‘appear contingent’, and starts
arguing from there. Given this, it would
scarcely be surprising that his argument should end up with a psychological
conclusion. Indeed we might well wonder
how it could end up with anything else.
How could a psychological premise, that the reality appears a
certain way to people, possibly imply anything about reality itself?
Still, I concede that Kripke
does present his argument as an objection to physicalism, not just as a piece of
psychological diagnosis. In this
connection, it is interesting to ask how he manages to bring the psychological
analysis to bear on the metaphysical issue.
As far as I can see, this only happens in the last three sentences of
the book. Until then he restricts
himself to the ad hominem claim that ‘identity theorists’ have no good way of
explaining the ‘appearance of contingency’ consistently with their physicalism. But in the final paragraph he adds the
suggestion that this ‘tells heavily against the usual forms of
materialism. Materialism . . . must hold
that . . . any mental facts are “ontologically dependent” on physical facts in
the straightforward sense of following from them by necessity. No identity theorist seems to me to have made
a convincing argument against the intuitive view that this is not the case.’
We can construe Kripke as
here arguing along something like the following lines:
(i) intuitively it doesn’t
seem that the physical facts necessitate the mental facts
(ii) unless this intuition
can be explained away, it is evidence that physicalism is false
(iii) this intuition cannot
be explained away, so
(iv) (there is evidence that)
physicalism is false.
If this is how Kripke reasons
right at the end, then the rest of his anti-physicalist argument can be seen as
a defence of the claim (iii) that the anti-physicalist intuition cannot be
explained away on the ‘qualitative surrogate’ model—the appearance of
contingency attaching to mind-brain identities is not an illusion akin to that attaching
to ‘heat = molecular motion’.
Of course, if this is how
Kripke is arguing at the end, it is not a particularly strong form of
argument. As he himself allows, the
inapplicability of the ‘qualitative surrogate’ model doesn’t necessarily mean
that there is no other way of explaining away the anti-physicalist intuition.[4] For instance, consider the issue from the
perspective developed in the last section.
I there argued that the earlier part of Kripke’s argument already forces
physicalists to recognize that the so-called ‘appearance of contingency’ is
simply an intuition of actual falsity, and that a fully satisfactory
physicalism therefore needs an explanation of why this intuition should arise
even thought it is false. If
physicalists succeed in finding such an explanation—and, as I said, a number of
such explanations are on offer in the current literature—then they will
automatically block Kripke’s final metaphysical move, by showing that the
anti-physicalist appearance of contingency can indeed be ‘explained away’.
It might seem too easy to
respond to Kripke’s final move simply by agreeing that we intuitively
disbelieve physicalism, and then seeking to explain why this should be so. But in truth this is an entirely reasonable
move. At bottom, Kripke’s final
argumentative step is of this form: p
seems false, so p is false. This is not
a particularly strong form of argument.
Many true things seem false. The
natural response to any argument of this form is surely simply to explain why
we should be inclined to disbelieve p even though it is true.
In effect, Kripke restricts
himself to considering a different kind of explanation for the anti-physicalist
intuition. He considers the hypothesis
that claims of brain-mind necessitation strike us as counterintuitive because we
muddle them up with other claims (their ‘qualitative surrogates’) that really aren’t
necessary—and he quite rightly points out that this hypothesis won’t work in
mind-brain cases. But it’s scarcely as
if the only—or even the most natural—way to explain a mistaken counterintuition
about some truth is to show that we have muddled the truth up with some other claim
that really is mistaken. After all, we
are surely perfectly capable of directly disbelieving truths, even when we don’t
muddle them up with other claims.
9. Is Fully Belief In Mind-Brain Identities
Incompatible With An Appearance of Possible Falsity?
Let me now return to argument
A. This may not culminate in the
metaphysical anti- physicalist conclusion that Kripke was aiming at. But it is a striking argument for all
that. It is certainly worthy our
attention if Kripke’s analysis shows that physicalists must intuitively
disbelieve their physicalism. In the
rest of this paper I want to show that this really does follow from the
considerations that Kripke adduces.
I repeat Argument A for ease of reference.
(1) If you fully believe ‘pain =
C-fibre stimulation’, then this can’t appear possibly false to you.
(2) ‘Pain = C-fibre stimulation’
does appear possibly false, even to physicalists. So
(3) Even physicalists don’t fully
believe ‘pain = C-fibre stimulation’.
In this section I shall focus on premise (1). Is full belief in mind-brain identities
indeed incompatible with any appearance of their possible falsity? I shall consider four possible reasons for
doubting this premise. Then in the next
section I shall turn to some other responses to Argument A.
(i) Does Everybody Know About The
Necessity of Identity? Some might
wish to query premise (1) on the grounds that it presupposes that ordinary
thinkers are familiar with the necessity of identity. At bottom, the argument for premise (1) is
that anyone who thinks that pains and C-fibres are actually identical will be
in no doubt that they are necessarily so, given the absence of confusing
qualitative surrogates in mind-brain cases.
Still, what about people who don’t know about the necessity of identity? After all, it was a great achievement of
Kripke’s to convince philosophers of the necessity of identity in the
1970s. So surely it would be easy for
people who don’t know of Kripke’s work to believe fully that pain is C-fibre
stimulation and yet not think that this is necessary?
I find this suggestion unpersuasive.
After all, it is not as if Kripke changed the first-order modal thinking
of ordinary people, as oppose to changing the explicit theories about modality
articulated by philosophers. I would say
that ordinary people have always thought that if
Furthermore, we can consider the case of more sophisticated physicalists
who have read Kripke and do know about the necessity of identity. These thinkers at least will surely be
disposed to move from full belief that pain is C-fibre stimulation to the
conclusion that they can’t possibly come apart.
But, to run through the argument once more, it does appear even to these
more sophisticated thinkers that pain can come apart from C-fibre stimulation; so they can’t fully believe that pain is
C-fibre stimulation; so something must
be stopping them from believing their physicalism. And if something is stopping even
sophisticated Kripkean physicalists from believing their physicalism, we can
surely infer that the same barrier is also present in less sophisticated
thinkers.
(ii) Two Kinds Of Imagination. It is widely supposed that a physicalist
answer to Kripke’s challenge lies in the fact that we have two quite different
ways of imagining the brain states that physicalists identify with conscious
states.[5] We can imagine such brain states
perceptually, by imagining what it would be like to observe them (by imagining the
neurographic readings produced by C-fibre stimulation, perhaps). But we can also imagine these states
sympathetically, by imaginatively recreating the conscious states with which
they are putatively identical (by imagining what it is like to be in pain, so
to speak). Physicalists hold that what
is being imagined in such cases is one and the same state, imagined either
perceptually or sympathetically. But the
two faculties of imagination are clearly themselves independent, in the sense
that one can be exercised without the other.
The suggestion is then that this independence of faculties itself enough
to account for the ‘appearance of contingency’.
Because we can imagine the state perceptually without imagining it
sympathetically, this creates the (illusory) impression that the brain state
could exist without the conscious state.
I don’t this that this succeeds in explaining the appearance of possible
falsity. I accept, of course, that the
two kinds of imagination mean that it an a posteriori matter that pain is
C-fibre stimulation. Somebody might have
the abilities to imagine C-fibre stimulation perceptually and to imagine pain
sympathetically, but lack any evidence that the two states are the same. Moreover, this means that it is certainly conceivable
that pain not be C-fibre stimulation, in that somebody could well suppose this
without conceptual contradiction. Still,
a central point I have aimed to establish in this article is that such
conceivability does not by itself guarantee any appearance of contingency in people
who already believe the relevant identity claim. You can recognise that a ≠ b is
conceivable, and that other people can coherently believe this, but this
needn’t make you yourself think make the claim might have been false. Suppose you believe firmly that
Does it make a difference that that in the mind-brain case we have
perceptual and sympathetic imagination, rather than mere ‘symbolic’
imagination? Symbolically imagining
But I don’t see that the extra graphicness of perceptual and sympathetic
imagination makes a difference. Suppose
that I grow up thinking of ‘Presley’ visually, as the man with the pout and the
long black hair, and of ‘Elvis’ aurally, as the man with the arresting
voice. And then, later on, I discover
that they are one and the same man. Now,
I can still imagine someone with that pout and hair but without that
voice. However, this act of imagination
isn’t going to make me feel that Elvis mightn’t have been Presley,
once I know that they are in fact the same man,. (What am I suppose to think here? That Elvis might not have been himself?)
Of course there is a real possibility here—a world where the person who
sounds like that doesn’t look like that.
This may not be a world where Elvis isn’t Presley, but it is clearly a
genuine possibility for all that. So
mightn’t the impression of mind-brain contingency arise from an analogous
source? Maybe this impression come from
our awareness that the state with certain perceptual effects (producing those
neurographic readings) might not appear sympathetically as it does (feeling
like pain).
But this is where we came in, with Kripke’s original argument. Mind-brain identities don’t work like claims
involving descriptions. Elvis might have
cut his hair, but pain can’t be pulled apart from its appearance. So when physicalists consider a scenario in
which C-fibre stimulation does not appear sympathetically as it does, they
aren’t merely supposing that C-fibre stimulation (ie pain) might lack some
contingent feature that it actually displays—rather they are supposing that C-fibre
stimulation might appear without pain itself.
And Kripke’s original point was that this supposition can’t explain how physicalists
can think that pain and C-fibre stimulation might come apart, since it presupposes
that C-fibre stimulation and pain are distinct states, contrary to their
physicalist view.
(iii) Descriptive Content On The
Right-Hand Side Maybe pain can’t be pulled apart
from its appearance, but C-fibre stimulation arguably can. It is quite natural to think of this state,
and of brain states generally, as physical kinds which only contingently have
the characteristic causes and effects (including neurographic readings) by
which we identify them. This is a
familiar view of scientific kinds in general, associated with the Ramsey-Carnap
thesis that theoretical terms are implicitly defined as referring to those
entities that satisfy the assumptions of some surrounding theory.
This then suggests an alternative explanation for the appearance of
possible falsity associated with claims like ‘pain = C-fibre stimulation’. In this world the C-fibre role, so to speak,
is realized by some physical kind X, and pain is identical to this kind X. But it is perfectly possible that the C-fibre
role might have been be realized by some other kind Y, that is, by something
other than pain. True, this isn’t the
possibility that C-fibre stimulation is not pain, if ‘C-fibre
stimulation’ rigidly designates state X.
Even so, perhaps an awareness of this possibility, that something other
than pain might have realized the C-fibre role, could be responsible for the
impression that ‘pain = C-fibre stimulation’ is contingent, even among those
who firmly believe it.
I don’t think that this works either.
I am happy to agree that even those firmly committed to ‘pain = C-fibre
stimulation’ can recognize the possibility that something other pain might have
played the C-fibre role. But I don’t
think that this is enough to explain the familiar impression that C-fibre
stimulation might not have been pain.
When it strikes people—committed physicalists included—that C-fibre
stimulation might not have been pain, their thought isn’t just that some other
(pain-free) state might have realized the C-fibre role. Surely they also think that the C-fibre role
might have been realized just as it is, and yet pain might have been
absent. That’s surely the intuition that
Kripke draws our attention to:
everything in the brain could have been just as it is, all the way down,
and yet pain might still have been absent.
Committed physicalists still seem to me to lack the resources to explain
this intuition. It is
incompatible with their belief that the actual brain facts fix the mental
facts. True, they can suppose,
compatibly with that belief, that the actual brain facts might have been
different, and in particular that the C-fibre role might have been realized by
some pain-free physical kind. But this
supposition isn’t going to help think that the brain facts might have been just
the same, and yet pain have been absent.
Yet this is the intuition that Kripke challenges them to explain, and
which still seems quite incompatible with their physicalism.
(iv) Mightn’t The Identity Have
Turned Out To be False? Some readers
might still be wondering whether there isn’t a simple explanation for the
appearance of contingency. Even after
you come to believe that pain = C-fibre stimulation, won’t the mere fact that
this is an a posteriori truth mean that you will still be aware that it might
have turned out otherwise? Mightn’t
the evidence have shown that pain wasn’t C-fibre stimulation after all, but
some other brain state? And won’t this
in itself account for your persistent feeling that C-fibre stimulation mightn’t
have been pain, even though it actually is?
No. As I have been stressing for
some while, a posteriority by itself need not create an impression of
contingency, at least not in those who fully believe some necessary claim. Suppose you are now certain that
I say that, if you firmly believe some necessary truth, and there are no
qualitative surrogates around, then you won’t think that this truth might have
turned out otherwise. You can think that
people might have failed to discover this necessary truth, but that is
different. So if you firmly believe that
pain is C-fibre stimulation—that they are one and the same state—then you will
have no room left for the thought that they might have turned out to be different. Once more the persistence appearance of
possible falsity in physicalist believers resists explanation.
10. Is The Intuition of Possible
Falsity Simply Due to Ignorance?
Maybe the reason ‘pain = C-fibre stimulation’ appears possibly false to
physicalists is not that they think it might have turned out differently, but
rather that they don’t yet know how it will turn out.
After all, nobody really believes the specific claim that pain is
C-fibre stimulation. It is well-known
that this is not a good account of the complex psycho-physiological data on
pain. And the same could be said of most
other attempts to equate specific conscious states with specific brain
states. At best these equations offer hypotheses
for further investigation. That’s why philosophers
typically invoke an example of a mind-brain identity—pain = C-fibre
stimulation—that they know not to be adequately supported by the evidence. For in fact there aren’t any good examples of
well-established mind-brain identities to use instead.
Accordingly, most contemporary philosophical physicalists restrict
themselves to a generalized physicalism.
They admit that brain science is still in its infancy, and that we are
not yet in a position to assert any claims of the form ‘M = P’ where ‘M’ names
some specific mental state and ‘P’ some definite brain state. At best we can assert that for each mental
state there is some physical state to which it is identical, but which
we don’t yet know about.[6]
This then suggests a different explanation for the impression of
possible falsity that attaches to any specific mind-brain identity of the form
‘M = P’. Maybe any such claim strikes us
as possibly false simply because we don’t believe it, and so of course attach
some positive credence its being false.
If this is the right diagnosis, it suggests that Argument A is perfectly
sound, but quite unsurprising. Let us
agree that (1) if you fully believe ‘pain = C-fibre stimulation’, then this
can’t appear possibly false to you, and that (2) ‘pain = C-fibre stimulation’
does appear possibly false, even to physicalists. From this it certainly follows (3) that even
physicalists don’t fully believe ‘pain = C-fibre stimulation’. However, on the current suggestion, this
isn’t because some deep cognitive obstacle is preventing them from embracing
physicalist claims that are supported by overwhelming evidence. Rather, it’s simply because there isn’t yet
any good evidence for any specific such claims.
That’s the only reason why physicalists don’t believe them, and so of
course think them possibly false. Or so
at least the current suggestion goes.
However, I don’t think that this suggestion amounts to an adequate
response to argument A. It is perfectly
true that specific claims of the form ‘M = P’ are currently under-evidenced,
and that as a result few physicalists actually believe any specific such
claims. But even so I think that
reflection on Argument A gives us good reason to suppose that anti-physicalist
intuition goes deeper than this temporary ignorance.
For a start, note that the suggested response implies that, once
physicalists do have good evidence for specific claims of the form ‘M =
P’, such claims will cease to strike them as possibly false. I’m not persuaded of this. Imagine that physicalist scientists have now
uncovered overwhelming evidence that in the actual world pain always goes hand
in hand with some brain process K. Wouldn’t
it still strike them that there could be a possible being with process K yet no
pain? I would have thought that this
intuition would remain. But then
Argument A shows once more that something would be stopping these scientists
fully believing the physicalist claim ‘pain = process K’, in the face of
evidence that we would expect to persuade them of this—for, if they were fully to
believe the identity, then it wouldn’t still appear possibly false to them.
Effectively the same point can be made without any counterfactual
assumptions about how things would strike phsyicalists who had more
evidence. Instead of focusing on
specific mind-brain identities ‘M = P’, consider instead the generalized
physicalist belief referred to above, that for each mental state there is some
physical state to which it is identical.
Most contemporary physicalists take themselves already to have conclusive
evidence for this generalized claim, even if not for more specific identities
(cf Papineau 2002 ch 2). But if such
physicalists fully believe this generalized physicalism, then the Kripkean
considerations imply that ought to have no remaining room for any thought that
this generalized physicalism could be false.
But they do.
After all, most contemporary physicalists will readily admit that
zombies do at least appear possible, even if in reality they are not. It certainly seems at first pass that a being
could share all our physical properties and yet lack any conscious life. But this appearance itself is incompatible
with a commitment to a generalized physicalism.
If you believe that each of our mental properties is identical to one of
our physical properties, then surely you must think that a being with all
our physical properties must have all our mental properties too. So what are you doing thinking that such a
being might lack our mental properties?
(If you fully believe that your son is one of the children in that
group, even though you can’t see which, do you think that all those children could
go into the classroom but your son not be there?)
So now we have a generalized version of Argument A.
(1’) If you fully believe that
each mental state is identical to some physical state, then zombies
ought not to appear possible to you.
(2’) Zombies do appear possible,
even to physicalists. So
(3’) Even physicalists don’t
fully believe their generalized physicalism.
I conclude, once more, that physicalists are in the grip of an intuition
of dualism that prevents them from fully believing their generalized
physicalism.
Let me conclude with a concession.
Maybe the appearance of falsity attaching to physicalist doctrines will indeed
fade away one day. In a future where physicalist
views have become part of educated common sense, then perhaps it will no longer
seem obvious that brain facts cannot necessitate conscious facts. Thus zombies may cease to appear
possible. (Look—they’ve got all my
mind/brain states—how can they not be conscious?) And perhaps even specific mind-brain
identities may cease to appear possibly false. (How could there be process K without
pain?—that’s what pain is.)[7]
I have been arguing that the Kripkean considerations show that there is
some barrier to physicalist belief—something is stopping even those who profess
physicalism from fully embracing their own doctrines. But I haven’t said anything about the nature
of this barrier to belief. Maybe it lies
in some deep architectural feature of the mind, and so will indefinitely remain
a source of cognitive illusion (cf the Muller-Lyer lines). On the other hand, maybe it is a relatively
shallow phenomenon, due to nothing more than the unfamiliarity of genuine
physicalism, and so will disappear once this view ceases to seem so
strange. To decide between these options,
we would need to know more about the mechanism behind the resistance to
physicalism, and that is an issue I have avoided in this paper.
Still, I hope I have done enough to show that there is as least some current
resistance to physicalist doctrines, even among those who sincerely profess
physicalism. Contemporary physicalists
may not be fully committed to any specific mind-brain identities, but they do
at least profess to believe in a generalized physicalism. Yet zombies still strike them as intuitively
possible, even if this intuition is fated to fade away in the future. I say that they wouldn’t currently have this
zombie intuition, if they fully believed their generalized physicalism. So something is currently stopping them from fully
believing this.[8]
Bibliography
Chalmers, D. 1996 The
Conscious Mind
Chalmers, D. 2006 ‘The
Foundations of Two-Dimensional Semantics’ in García-Carpintero, M. and Macià,
J. (2006)
García-Carpintero, M. and
Macià, J. 2006 Two-Dimensional Semantics
Hill, C. 1997 ‘Imaginability,
Conceivability, Possibility and the Mind-Body Problem’ Philosophical Studies
87
Hill, C. and McLaughlin, B.
1999 ‘There are Fewer Things in Reality Than Are Dreamt of in Chalmers'
Philosophy’ Philosophy
and Phenomenological Research 59
Kripke, S. 1980 Naming and
Necessity
Levine, J. 2001 Purple
Haze
Loar, B. 1997 ‘Phenomenal
States: Second Version" in Block, N., Flanagan, O., and Guzeldere, G. eds Consciousness,
Melnyk, A. 2003 ‘Papineau on
the Intuition of Distinctness’ SWIF Forum on Thinking about Consciousness
http://lgxserver.uniba.it/lei/mind/forums/004_0003.htm
Nagel, T. 1974 ‘What is it
Like to be a Bat?’ Philosophical Review 83
Papineau, D. 1993
‘Physicalism, Consciousness, and the Antipathetic Fallacy’ Australasian
Journal of Philosophy 71
Papineau, D. 2002 Thinking
about Consciousness
Papineau, D. 2006 ‘Comments
on Strawson’s “Realistic Monism: Why Physicalism Entails Panpsychism”’ Journal
of Consciousness Studies 13
Yablo, S. 2000 ‘Textbook
Kripkeanism and the Open Texture of Concepts’ Pacific Philosophical
Quarterly 81
[1] See for example Chalmers
(2006). It should be noted that Chalmers
and some other advocates of two-dimensional semantics take ‘unstable primary
intension’ to be a more general notion than that of a contingent descriptive
content and would object on these grounds that my formulation of principle (I)
coarsens their interpretation of Kripke.
However, I don’t think that this point matters to any of the arguments
to follow—and it is much easier to explain things in terms of ‘descriptive
contents’ rather that ‘unstable primary
intensions’.
[2] I am here rehearsing the ‘a posteriori’ physicalist
line. Different a posteriori
physicalists have different views about the extent to which principle (I)
fails, but they all agree that mind-brain identities in particular constitute
exceptions. (Loar 1997, Hill 1997,
Levine 2001, Papineau 2002.)
[3] A couple of years ago I offered my colleague Keith Hossack the standard physicalist response to the two-dimensional reading of Kripke. He responded ‘That might explain an appearance of possible falsity, but it doesn’t explain the appearance of contingency’. It was this remark that got me thinking about the proper way to interpret Kripke.
[4] ‘That
the usual moves and analogies are not available to solve the problems of the
identity theorist is, of course, no proof that no moves are available’ p 155.
[5] See Nagel 1974, footnote 11, Hill 1997, Hill
and McLaughlin 1999.)
[6] True, many contemporary physicalists believe something yet weaker,
since they formulate physicalism in terms of supervenience rather than
identity. But let me continue to focus
on identity, in line with Kripke’s original discussion of physicalism. This is for expository convenience only—all
the points made in this paper apply equally to supervenience versions of
physicalism.
[7] Stephen Yablo suggests that future
discoveries may well undermine anti-physicalist intuitions: ‘Am I the only one who feels the
intuition of zombies to be vulnerable in this way?’ (op cit, p 119).
[8] A version of this paper was given at
a Pacific APA symposium in April 2007. I
would like to thank Tyler Doggett for his very helpful comments.