Deconstructing Anti-Physicalism

The arguments against physicalism with respect to consciousness, e.g., the knowledge argument, and the challenges to them present a familiar story. But it is seldom asked whether anti-physicalism is in itself a tenable view in terms of maintaining an internal consistency. Below I argue that on close inspection there are reasons to doubt that the view is tenable in this sense. There is, I contend, an internal tension to anti-physicalism that centres on a conflict between its principal conclusion, i.e., experiential properties are non-physical, and the fact that no one, including the anti-physicalist, doubts that others are conscious. It is a tension, I argue, that the anti-physicalists have overlooked, but which they may not have the resources to defuse. What follows is deconstructive analysis of the seminal arguments (texts) of anti-physicalism. It reveals a problematic ambiguity in their use of the terms relating to conscious states that, I submit, is the source of this tension.

  1. The tension

The various anti-physicalist arguments aim to show that experiential properties, specifically, are non-physical. Let us assume the truth of their conclusion. A corollary of this is that we cannot ever observe that others are conscious. Nothing physical about someone else, S, in other words, justifies my believing that S is conscious. That includes of course S’s verbal behaviour such as S’s report that she is conscious.

Now, the consciousness of others is undeniable. So I ought to believe S is conscious regardless. But on what grounds is this belief warranted? A result of the anti-physicalist conclusion is that it is impossible for me to obtain evidence of the consciousness of others. That follows if one supposes, quite reasonably, that the only avenue by which I can grasp contingent truths about others is sensory experience, i.e., observation; and being conscious is such a truth one assumes. In other words, that others are conscious is a posteriori true or false, but we cannot use experience to determine which it is.

In summary, the tension is registered by the following line of reasoning:

  1. Corollary: There is no observable difference between the presence and absence of experiential properties ( i.e., qualia).
  1. But someone’s experiences’ having such properties would be contingently true a posteriori.
  1. We cannot justify the presence of experiential properties in others a priori.


  1. Anti-physicalism implies that we are never justified in believing that others are conscious, even though all agree that others are conscious.

We can understand this tension, then, as a problem posed by the following question: How can the anti-physicalist believe that others are conscious given that the non-physicality of consciousness implies that it is impossible to justify this belief? It is not the problem of other minds as it is often understood, namely the challenge of our knowing that others are conscious, i.e., a sceptical problem. Here I take a sceptical problem in general as consisting in there being a hypothesis that undermines our presumed justification for holding some set of beliefs as true, e.g., that if I am a brain in a vat then all my beliefs about the external world are false. Rather, our problem is transcendental, one might say, in the sense that assuming consciousness is not physical there is no possibility of understanding others as being conscious to begin with. Anti-physicalism rules out the conditions for the possibility of understanding that others are conscious.

  1. Can the Tension Be Defused?

The anti-physicalists appears have one way of defusing this tension, that is, by appeal to some sort of inductive inference, e.g., inference to the best explanation. Induction at least offers the possibility of bridging the gap between the a posteriori nature of knowledge about others generally and the impossibility of gaining such knowledge a priori. We might be able to show by reason, that only appeals to observations indirectly, how others are (very probably) conscious.

Consider the knowledge argument against physicalism (Jackson 1982). If experiential properties are physical, then we can in principle come to know them from our knowledge of the physical processes correlated with them. However, as the early Frank Jackson argues, no amout of physical information tells us about these properties, e.g., what it is like to see something as green. Therefore, modus tollens, experiential properties cannot be physical. Now here it is assumed there is a law underwriting this correlation so that anyone who realizes the salient physical processes also has experiences with these properties.

However, as Herbert Feigl famously described it, any such laws would ‘dangle’, that is, there could be no way of confirming or disconfirming them by observation (Feigl 1958). They would be untestable. This is an aspect of the problem we are highlighting, that led Feigl to endorse mind-brain identity. Yet anti-physicalism needs these laws, short of collapsing into substance dualism. Experiential properties, though non-physical, are thought to be properties of physical entities, i.e., they are dependent on the physical in some sense. Jackson, for example, hypothesized that they supervene on physical properties, that there are supervenient psychophysical laws.

My realizing a pain quale when I have a pain experience, for example, is not miraculous. That is, the occurrence of pain qualia each time I have a pain experience, in terms of behaviour and physiology, must be caused by or supervene on certain physical properties. This regularity is best explained by its being underwritten by some set of psychophysical laws. Moreover, laws are invariant – they hold anywhere at any time. Therefore, so long as others have the same kind of physical properties as myself I can safely infer that they are conscious too.

This suggestion might succeed if we can identify which physical properties cause or subvene specific qualia, i.e., experiential properties. But the dangling nature of any such purported psychophysical laws rules out this possibility. Someone’s first-person reports of realizing a pain quale while exhibiting the physical symptoms of toothache on various occasions is no evidence of the existence of the property. They are compatible with its absence.

Crucially, the anti-physicalist holds that no particular physical properties are logically correlated with qualia. As suggested, for example, by Kripke’s thesis of the contingent correlation between them, and by Chalmers’s assertion that qualia do not logically supervene on any physical properties. Accordingly, the pain quale I realize at any instant might supervene naturally, instead, on any of an infinite array of neuronal activities. None can be ruled out.

Given that everyone realizes a distinct set of physical properties that are causally responsible for pain behaviour – Let that set be K, in my case – it is possible that the pain quale depends on a complex set of such properties, L, unique to me, where L is a subset of K. To rule this out we would have to isolate L from the rest of K. But there is no method available to distinguish L’s members from K’s. Imagine using Feigl’s autocerebroscope. You aim to excite by trial and error those neurons responsible for realising the pain quale alone. But if you hit upon them you will not register the fact. That would be a form of behaviour.

Appeal to inference to the best explanation fails not because there are better rival explanations of the regularity between physical and experiential properties. We have considered none after all. Rather in the end it fails because any physical and experiential properties can only ever be correlated from the first-person viewpoint, and these correlations are therefore not fit for generalization. It is not, to be clear, a matter of enumerative induction, where one could argue that it is very probable that I am not unique in this respect. The trouble is that we have no independent method of testing such a law. We can never know if any regularities generalize to others. Still, then, there is no way of justifying the belief that others are conscious.

Again, no one seriously doubts that others are conscious. The problem for anti-physicalism, however, is that it rules out the possibility of determining the truth-value of the generic proposition ‘Others are phenomenally conscious’ (OC), e.g., have such-and-such sensations.

To be clear, the worry is not that OC is meaningless, say by appeal to the verification principle. I am not arguing that anti-physicalism, if true, must lead us to conclude that OC is disguised nonsense. Everyone agrees that OC is meaningful. We are able to judge whether most things are conscious or not. The dinner plate before me is not conscious. Undoubtedly this assertion is true. The woman sitting alongside me on the bus is conscious, and that is clearly true. The worry is that anti-physicalism robs me of the possibility of justifying such assertions at all, and that is a bad result for anti-physicalism.

  1. What say the anti-physicalists?

Do anti-physicalists themselves see this tension? Thomas Nagel dismisses the general epistemological problem of other minds. The problem as Nagel views it is a conceptual problem – we must explain how we have a concept of mind that includes others (1986: 19-22). The tension for Nagel, therefore, does not exist; at least, he does not appear to see any.

David Chalmers, on the other hand, embraces the problem of other minds insofar as it follows from the zombie hypothesis. It shows, Chalmers argues, your zombie twin is a genuine possibility (1996: 74). That said, he offers a solution to it by appeal to inference to the best explanation, along the lines spelled out above. I observe a regularity between my own functional states and my various experiences and postulate laws that best explain this regularity, laws that can be generalized to others (ibid: 246). I have already pointed out that this solution does not work. Imagine, for example, a conscious subject, A, and his close functional duplicate, B, who is not conscious. Both, let us assume, are behaviourally indistinguishable in relevant respects, so that both report feeling pain and so on. Let A’s functional description be F and B’s be F*. Clearly some difference between F and F* accounts for A’s being conscious and B’s not being conscious. But we are never in a position to determine the relevant difference. The only way we have of determining the difference is A’s and B’s first-person reports and these do not differ. B invariably offers the same first-person reports for some stimulus input as A. At least this is always a possibility. Any regularities I identify between function and experience cannot be confirmed or disconfirmed. Chalmers, it seems, sees no real tension. He does not appear to think that anti-physicalism itself makes the problem of other minds vis-à-vis consciousness really intractable.

Jackson, using the voice of his imaginary Mary, acknowledges the problem. He tells us that Mary worries that the colour qualia she has come to learn about after her release from her black-and-white world may not be shared by others. Her generalization might seem to be ‘wild’, as he describes it. But, Jackson writes that Mary “decides she does know [that other have such qualia] and that skepticism is mistaken (even if, like so many of us, she is not sure how to demonstrate its errors).” (1986: 294) Like Chalmers, Jackson invokes the problem to support anti-physicalism. Mary is right to worry about this but if, following Laurence Nemirow (1980) and David Lewis (1988), Mary gains know-how – a new representational ability – rather than factual knowledge about colour experiences on her release, then the problem of other minds ought not to apply. The problem does present itself and that is a strike against Nemirow’s and Lewis’s representational ability hypothesis. Jackson appears not to see the problem as especially urgent for anti-physicalism either.

None of these anti-physicalists see any urgent tension in their position. A likely reason is that they assume their arguments are sound, and we must as a consequence bite the bullet and acknowledge the problem of other minds. If we cannot overcome it, too bad. However, the trouble for anti-physicalism suggested by this tension runs deeper still. The anti-physicalist conclusion, I suspect, retrospectively affects their anti-physicalist arguments for it, i.e., render the conclusion itself implausible. Let us see how.

  1. The arguments

There are three classic arguments against physicalism, namely the knowledge, conceivability and modal arguments. If each argument has as a premiss OC, or one derived from it, then OC’s indeterminability would prove to be problematic. We cannot tolerate a premiss that is indeterminable since any argument only goes through when its premisses are determinably true. Our reasoning, after all, is based on strict logical principles that can be likened to arithmetical laws. The anti-physicalist arguments is fatally ambiguous in the same sense. Again, OC seems unequivocally to be determinably true, reflected by our confidence in asserting a friend, say, is conscious, but the anti-physicalist conclusion undermines this confidence given that it makes it impossible to justify OC, hence the tension.

In Nagel’s version of the knowledge argument (1974) – here we treat his argument as anti-physicalist, though Nagel in his own presentation of it is agnostic about the falsity of physicalism – we ought to conclude experiential properties are non-physical given that for relatively alien creatures like bats no amount of physical information about them can tell us what it is like to be a bat. It does not allow us to know these properties of a bat’s echolocatory experiences for example. This conclusion is predicated on our holding that a bat is conscious, which is a version of OC. Bats have experiential properties and they are knowable in principle, according to Nagel. His argument, therefore, faces the difficulty outlined above.

In the case of Jackson’s version (1982), although Mary knows everything physical about other people’s colour experiences, she learns something new about them once she leaves her black-and-white world and experiences colours, namely their subjective experiential properties or qualia. Therefore, these properties cannot be physical. If, however, Mary cannot hold that other people’s experiences realize these colour qualia, then she cannot claim to have gained any knowledge about them. That said, Mary at least learns something new about her own experiences. Is not that enough for the argument to go through? No, because as Jackson admits himself, no one denies that Mary learns something new about her own experiences on her release, given that these experiences are novel to her (1986: 292). What she learns, by Jackson’s measure, must be something about other people’s colour experiences, i.e., their having these experiential properties. That is exactly what we are denying that she is justified in supposing. This version, therefore, faces the difficulty too.

The conceivability argument is most often presented in terms of the zombie hypothesis. Chalmers argues that since it is possible to conceive of a person physically indistinguishable from a conscious person who is not conscious, e.g., my zombie twin, it cannot be the case that qualia are identical with any physical properties (1996). The argument is predicated on being able to hold that one’s zombie twin could be non-conscious, that is, to hold that OC could be false. It is to assume that we are in a position to determine whether or not someone else is conscious, even though it is also assumed that S is conscious if and only if S’s experiences have qualia, the presence of which is unobservable, that is, indeterminable from the third-person viewpoint.  Again, we could never be in that position if qualia are non-physical, hence the conceivability argument also faces this difficulty.

In his modal argument Saul Kripke points out that if experiential properties are identical with physical properties, then they are necessarily correlated (1972). He argues that it is always possible that pain qualia are not correlated with physical properties – we can never rule out this possibility. Therefore, the correlation cannot be one of identity which, again, is necessary. To generalize, qualia cannot be identical with any physical properties, i.e, be physical. Here Kripke must assume that the term ‘pain quale’, for example, refers to a particular property. However, reference presupposes that speakers share access to the referent. ‘Obama’ refers to the man described as the current US president insofar as you and I pick out the same object by the name. In the case of ‘pain quale’ I cannot assume others refer to that property given that as a non-physical property I have no means of ever determining this. I cannot know what others pick out by it. The identity statement ‘pain is C-fibre stimulation’, for example, therefore cannot have a determinable truth-value. The modal argument, then, does not escape this difficulty either.

The anti-physicalist conclusion, then, fatally undermines the very arguments for it. Now, one might object that one only needs to assume that OC is true for each of these arguments to be viable. In reply, the challenge to these arguments is formal. The difficulty is that OC’s truth-value is indeterminate, making OC in principle unjustifiable. It is by that measure logically inoperative, thereby ruling out the arguments’ viability. One can only assume OC’s truth if we can in principle assign it a truth-value at all, and that, it seems, cannot be done.

But, the counter objection might go, OC, as it is understood in phenomenological terms, is so plausible no one can deny its truth all the same. It is most reasonable, for example, to believe that the world was not created five minutes ago by a deceiving God who gives us false memories etc., despite there perhaps being no way to show how it is true. Analogously it is reasonable to assume OC is true despite our being unable to justify it. The reply in turn is that the process of inference by which arguments operate demands that their premisses be truth-evaluable. To hold OC true because psychologically it is difficult, or even impossible, for us to deny it is logically irrelevant. That is to commit the psychological fallacy, to suppose that one’s subjective psychological states qualify as objective evidence for some claim – the cardinal sin of psychologism, for example, as pointed out by Gottlob Frege (1980: 33-38). Moreover, OC is not analogous to the belief that the world is more than five minutes old. That is because all the evidence supports this latter belief. That the world was created five minutes ago by a deceiving God is a sceptical hypothesis, and as noted earlier, the problem underlying the tension is not a sceptical one in this way. The problem is that anti-physicalism denies us any grounds for holding OC to begin with.

  1. The Diagnosis

What has gone wrong? How has the urgency of this tension gone unnoticed? The anti-physicalist knowingly works with a phenomenological notion of conscious states such as sensations, but then ‘inadvertently’ switches to a folk-psychological notion when asserting that others have conscious states. In other words, she equivocates with respect to two notions of conscious states, one phenomenologically based the other cause-based.

This first notion derives from the intuitive idea that at bottom we are acquainted with being conscious by introspection. The classic articulation of it is offered by Nagel in terms of our coming to know about qualia (what he calls ‘phenomenal features of experience’) from our own case, from the first-person point of view (1974). Consciousness is thus understood as an essentially subjective phenonenon. This is obviously the notion the anti-physicalist privileges.

The folk psychological notion has been helpfully spelled by David Lewis (1999: 298-300). Folk psychology is used to predict and explain our behaviour in causal terms, e.g., if someone expresses being in pain because of X then they are very likely to act with the aim to end X. In general we could construct a set of Ramsified sentences that spell out each of our mental terms as it relates to behaviour in this way, describing its typical causal role. That furnishes us with an understanding of conscious states that does not depend on thinking of them in phenomenological terms. The independence of such notions of conscious states from their phenomenological counterparts is illustrated, for example, by the following remark by Wittgenstein: “Just try – in real cases – to doubt someone else’s fear or pain.” (1958: §303) Wittgenstein is pointing out that often when we believe that someone is in pain, e.g., on seeing her exhibiting pain behaviour, we are not inclined toward doubt because we cannot ‘observe’ its qualia. Rather, our belief is secured by the behaviour alone.

An example of this sliding or switching between notions of conscious states is suggested by Neil Campbell with respect to Jackson’s knowledge argument. Campbell observes that in his original knowledge argument Jackson sometimes operates with an epiphenomenal understanding of qualia, i.e., as non-phyiscal properties of experience, and at other times operates with a causal notion of such properties (2012: 143-44). In the first case Jackson argues that in her black-and-white world Mary cannot come to know colour qualia from her cause-based understanding of vision. In the second case it is imagined that Fred’s behavioural differences from the normally sighted person – in terms of his ability to discern two types of red when tested – is evidence of a difference in qualia. As Campbell remarks: “Were he [Fred] to fail such tests we would have no reason to think he has a richer or more complex phenomenal life than the rest of us.” (ibid: 143) By this measure Jackson presupposes that Fred’s peculiar red qualia make a causal difference, contrary to his epiphenomal notion of qualia.

Now we have diagnosed the disease, the symptoms are more easily spotted. For example, Chalmers imagines his zombie twin, someone physically identical with him but who is ‘all dark inside’, adding that this is likely empirically impossible (1996: 96). Here Chalmers must presume that there is some way of empirically testing for zombies, but the very definition of them makes this impossible. There can be no observable difference between oneself and one’s purported zombie twin; that is the whole point. Like Jackson, he thinks of being conscious both in phenomenological terms, of course, and cause-based terms by which he can imagine testing for the existence of phenomenal zombies. This latter notion he explicitly repudiates, however. Consciousness, he insists, is entirely distinct from any such cause-based psychological phenomena (ibid: 11).

  1. Conclusion

I suspect that anti-physicalist sees the problem of others minds vis-à-vis consciousness as a general problem, and as such it is not something that she in particular needs to address. However, we have seen that it is a problem that she owns since it is the anti-physicalist conclusion that makes it properly intractable. The tension, recall, centres on the indeterminability of OC. It has two aspects to it:

(1) Its indeterminability undermines the anti-physicalists’ arguments themselves. If, as they conclude, consciousness is non-physical, then we can never determine whether others are conscious. But the conclusion rests on OC being determinable.

(2)  OC is undeniably true, and as such it is determinable. To the extent that anti-physicalism makes OC indeterminable, the view is untenable.

The anti-physicalist could attempt to defuse the tension by assuming OC is determinable somehow. That is, she could insist that OC is determinately true even though the epistemological problem of other minds vis-à-vis consciousness seems intractable. But by what measure can OC be judged to be determinable? Certainly we seem psychologically compelled to think of others as being conscious, but that cannot count as evidence of the truth of OC. The difficulty comes down to this: OC is contingently true – we cannot logically rule out its falsity – yet as a truth about others the only means we have of acquiring even mildly conclusive evidence for or against it is by observation, and as we have seen, that is ruled out by the non-physicality of its characteristic properties.

The anti-physicalists are in a bind. If they assume OC is indeterminable, then the problem of other minds vis-à-vis consciousness appears to be intractable in principle. Consequently, any of their arguments for the conclusion that consciousness is non-physical are undermined. If, in the interest of salvaging their arguments, they insist OC is determinable, then they face the seemingly insuperable difficulty of explaining how it is. As things stand there is good reason to think that anti-physicalism is untenable. It cannot maintain internal consistency.

© Hugh Alcock


Campbell, N. 2012. Reply to Nagasawa on the Inconsistency Objection to the Knowledge Argument. Erkenntnis 76: 137-45.

Chalmers, D. 1996. The Conscious Mind, Oxford: Oxford Univ. Press.

Feigl, H. 1958. The “Mental” and the “Physical”: The Essay and Postscript. Minnneapolis, MS: University of Minnesota Press.

Frege, G. 1980. The Foundations of Arithmetic, 2nd edition. Austin, J.L., trans. Evanston, IL: Northwestern Univ. Press.

Jackson, F. 1982. Epiphenomenal Qualia. Philosophical Quarterly 32: 127-136.

Jackson, F. 1986. What Mary Didn’t Know. Journal of Philosophy 83: 291-95.

Kripke, S. 1972. Naming and Necessity, Oxford: Blackwell Pub.

Lewis, D. 1988. What Experience Teaches. Proceedings of the Russellian Society (University of Sydney). In Philosophy of Mind: Classical and Contemporary Readings. Chalmers, D., ed. NY: Oxford Univ. Press (2002): 281-94.

Lewis, D. 1999. Reduction of Mind. David Lewis: Papers in Metaphysics and Epistemology. Cambridge: Cambridge Univ. Press.

Nagel, T. 1974. What Is It Like to Be a Bat? The Philosophical Review, 58(4): 435-50.

Nagel, T. 1986. The View From Nowhere. NY: Oxford Univ. Press.

Nemirow, L. 1980. Review of Thomas Nagel. Mortal Questions, Philosophical Review, 89: 473-77.

Wittgenstein, L. (1958) Philosophical Investigations, Anscombe, G.E.M. (trans), third edition. London: MacMillan Co.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s