Conceivability And Qualia – A Response To Block’s China Argument










Johnathan Nightingale

Professor John Vervaeke



            Functionalism stands at the end of a long list of theories.  This is true for two reasons: first, it has had many predecessors, it is fairly new, and the product of many failed theories before it; and second, there does not seem to be anything following it, waiting to take up the fight, should it falter like its parents.[1]  Its materialist precursors have largely fallen victim to a tendency to be too chauvinistic, or in any case, too extreme, in their attempts to subject mental phenomena to analysis into non-mental terminology.[2]  Functionalism, in fact, finds its roots in an attempt to alleviate this problem by removing the neurological bias, and thus developing a theory that offers the ability to analytically examine the mind without a reliance on human-like beings as sole possessors thereof.  So it naturally comes as a bit of a painful surprise, when someone offers an argument that, at least prima facie, would seem to refute functionalist doctrine in one sweeping motion.  Block’s homunculi-headed simulations are a proposed demonstration that functionalism is too liberal in its ascriptions of mentality and further, that this liberalism cannot be adequately fixed within the current theoretical framework – in short, functionalism is false.

            Block’s argument, as I will show, is rather unfortunately valid – unfortunately, that is, for proponents of functionalism.  What the argument is not, or perhaps, to be more precise, what the argument may not be is sound.  This paper will concern itself with a detailed analysis of precisely what it is that Block says in his argument, followed by a discussion of the possible criticisms thereof (there are distressingly few) and finally an attack at what appears to be the argument’s only weak point.  My argument, much like Block’s, cannot claim to have made a wholly compelling case.  Whereas I will show that Block’s argument is basically an argument from inconceivability, my own argument must then, to adequately counter Block’s, be an argument from conceivability.  Neither of these is binding, the argument is a question of shifting the burden of proof.[3]

            Block is concerned with showing that a functional description of a mind is insufficient for capturing all of the mind’s qualities. To do this, he proposes a thought experiment designed to play our intuitions against functionalism, by realizing a functionalist system in a very bizarre form: China (Block 1975, 276-278).  What he proposes we imagine is that, in the name of philosophical progress, China embarks on an attempt to realise a human mind by having its citizens play the roles of individual neurons,[4] interconnected by radio link. (Block 1975, 276).  Oddly enough, this is actually Block’s attempt to propose a more realistic thought experiment, extending from another experiment of similar intent. It will be useful to consider the experiment from which the China example arose (although the China example is perhaps more persuasive) which dealt instead with a single body,


… externally like a human body, say yours, but internally quite different.  The neurons from sensory organs are connected to a bank of lights in a hollow cavity in the head.  A set of buttons connects to the motor output neurons.  Inside the cavity resides a group of little men. Each has a very simple task: to implement a ‘square’ of a reasonably accurate machine table that describes you.  (Block 1975, 276)


The idea of both experiments is to provide a correct, if somewhat bizarre, implementation of a mind, as defined by functionalists.  The idea that the little men (and by extension, we as well) are guided by a machine table is taken from Putnam’s definition of functionalist theory, whose first principle is “All organisms are probabilistic automata”[5] (Putnam 1967, 227).  All Block has done then, is to extend this principle in some rather unusual ways; the first in the fantastic notion of a homunculi-headed copy of yourself, and the second, a homunculi-writ-large version encompassing an entire nation.  In both cases, they are doing exactly what Putnam has deemed they ought to do, they are implementing a machine table which we assume presents an accurate picture of your mind.  According to Functionalism then, argues Block, China and the homunculi-head both, should be (or have, the distinction is unclear) minds.

            This notion, however, is squared against the common sense intuition we have, that China[6] does not have, cannot have, that it is almost nonsensical to discuss the idea that China has a mind.  In Block’s words,


What makes the homunculi-headed system (count the two systems as variants of a single system) just described a prima facie counterexample to (machine) functionalism is that there is a prima facie doubt whether it has any mental states at all—especially whether it has what philosophers have variously called ‘qualitative states’, ‘raw feels’, or ‘immediate phenomenological qualities.  (1976, 278).


What we have then, in essence, is an argument against functionalism by Modus Tollens:


·        If machine functionalism as Putnam describes it is true then China has a mind.

·        However, by intuition, China does not have a mind.

\     Machine functionalism as Putnam defines it is false.


This reduction to a basic logical truth adds a great deal of strength to Block’s argument, or rather, makes clear just how powerful Block’s argument is.  The fundamental structure is logically valid, unlike so many philosophical arguments which, in the end, reduce themselves to an inductive appeal at best.  Now, in truth, this is not a perfect modus tollens, because Block freely admits that the second premise, that China does not have a mind, seems self-evident, but is not certain.  In fact, he points out that “Appeal to intuitions when judging possession of mentality…is especially suspicious” (Block 1976, 281).  Thus the argument remains logically valid, but rests on the assumption that our intuitions hold.

            Having outlined the argument in what I think is a fair and accurate way, I will examine some potential criticisms of Block, and why I believe they fail.  While it is my eventual goal to present an argument against Block, it is equally important to establish what my argument is not, to spare some objections that might otherwise be made.  The analysis will also demonstrate again that Block’s argument is really quite strong, for the most part.

            The first, instinctive method of defense is to deny Block’s claim that China satisfies the functionalist criteria of mind.  To me, this seems contradictory, but a functionalist might be disposed to do so, when confronted with the fact that her philosophy licenses such a conclusion.  The only avenue I can see from which to launch such an attack would be in the idea that somehow, the argument was a poor analogy, but I don’t think that such an argument can run.  Block’s argument is not, in truth, an argument from analogy at all.  He is not saying that China is like a computer, or other more common functionalist tools, he’s saying that China, by Putnam’s own criteria, is a satisfactory realization. The idea that functionalism implies a machine-table theory of mind realizable in any properly designed physical system is Putnam’s, not Block’s, and any attempt to attack the argument from this direction would be to attack one’s own theory.

            A second, more appealing defense might be to suggest that the argument simply points out a failure of current functionalist theory, and that a modification of the theory would save us from Block’s conclusion.  What the nature of this modification could be, I am not sure, but it doesn’t seem to matter—Block has already answered this approach quite effectively, in addressing the chauvinism/liberalism distinction (1976, 291-293). What Block argues is that, as functionalism now stands, it is “hopelessly liberal” (1976, 292), and further, that any attempts to tighten functionalist theory, through restricting the systems classified as minds, leads immediately to human chauvinism.  This is the source of his Functionalism/Psychofunctionalism distinction.  In contrast to the traditional functionalists, which allow a large variety of systems into the category of mental systems, Psychofunctionalism would be a doctrine that restricted entry to only those systems that shared a psychology with us.  This might rescue functionalism from some arguments, but Block goes on to point out that this eliminates the possibility of other minds, equally possessed of mentality, but opposite in psychological makeup (1976, 291-292).  At the risk of getting sidetracked, I will leave this argument where it stands, but suffice it to say that this too is not my objection – I do not see that functionalism can be successfully mended in this way.

            So what then, is my argument against Block?  An examination of the previous two arguments is illuminating – both proceeded from the basic notion that Block was correct in assuming that China did not have mind.  They were arguments framed as attempts to save functionalism from the seemingly contradictory conclusion that China did have mentality.  Going back to my analysis of Block’s argument structure, then, they were attempts to falsify the first premise.  I have already indicated my own opinion though, which is that this first premise is sound.  My task then, is to examine the second premise, in hopes of weakening it sufficiently that Block’s argument becomes questionable, at best.  At this point, we must return to the idea of inconceivability arguments, why this is such an argument, and what can be said about such arguments in general.

            To see that this is a conceivability argument is not difficult, Block himself points this out frequently in the argument, but uses the phrase prima facie instead.  Essentially, the argument rests on the appeal to the reader’s intuition that China cannot in fact have mental states, especially not qualia (Block 1976, 278,280).  In attempting to strengthen this appeal to intuition, Block notes the more general principle that without sufficient justification, there is good intuitive evidence against any mechanism possessing qualia,


No physical mechanism seems very intuitively plausible as a seat of qualia, least of all a brain.  Is a hunk of quivering gray stuff more intuitively appropriate as a seat of qualia than a covey of little men?  If not, perhaps there is a prima facie doubt about the qualities of brains systems too?

            However, there is a very important difference between brain-headed and homunculi-headed systems… we are brain-headed systems, and we have qualia.  (Block 1976, 281)


While he is right, in the sense that our experience of qualia helps to reduce doubts we might otherwise have about our own mentality, the statement ends there in its utility.  It does not give us any more reason to believe his argument, if anything, it just peremptorily answers a potential objection that may have been raised.  Still, in shifting the issue to one of allaying doubt, an important fact is brought to light.  Block’s argument is critically dependant on doubt.  That is, we have no solid reason to suspect that China has qualia, and so in the absence of such a reason, we doubt that it does.  This doubt is the nature of the second premise in the argument as outlined above, and it also presents a possibility to counter the argument.  To do se we must show that there is reason to doubt the doubt, as it were, reason at least in principle, to accept the possibility that China might experience qualia.  If a good argument can be given for the possibility of China’s having qualia, then we will have as much reason to accept China’s mentality as we do to accept our own, and Block’s argument will be significantly weakened, if not overturned.  The argument need not be entirely incontrovertible either[7], but must simply provide an acceptable reason to suspend judgement about China’s mentality.  To put it into legal jargon then, it is sufficient to show that there is ‘reasonable doubt’ about the impossibility of China’s mentality.[8]  It shall be my focus in the remainder of this paper to demonstrate that such reasonable doubt exists.

            First of all, let us establish what exactly we mean when we talk about “having mentality”.  As an earlier quote demonstrates[9], Block seems at least partly willing to accept that China would be capable of some mental-like operations, though not the elusive qualia.  This point is perhaps a fine one, but it limits the scope of Block’s objection.  There is no reason to reject China as a mind, where mentality consists of calculations, or even linguistic use, for example.  These are things that can be accomplished by a sufficiently well-designed computer, and China as Block describes it is indeed a sort of organic computer.  This is only to say that Block’s contention is not, or should not be, that China is capable of none of the things that a mind is, certainly a great deal of mental function can be accurately implemented by China without conflicting with our intuitions.  The real sticking that Block perceives in Functionalism is the issue of qualia, there does not seem to be anything that it is like to be China.[10]

            What is important to note here, is that Block, like the rest of us, are presupposing a psychological framework, under which we do our philosophy.  An admittedly difficult thing to overcome, we are restricted by our conception of qualia – our conception of mind – to considering only minds like our own.  What we are saying, when we say that we cannot conceive of China having a mind, is exactly that – we cannot conceive it.  Our folk-psychological theories of mind extend only to minds that in some very important ways, are identical to our own, and it must be acknowledged that this may not be an objectively valid premise from which to reason.  Accordingly, any theory of qualia that was sufficiently divergent from Block’s (or what I will call the Intuitive theory of Qualia), might be enough to allow such things as China to have minds after all.  Do such theories exist as a matter of anything but theoretical possibility?  Indeed, they do, and I shall attempt to demonstrate that these theories (there are two I will consider specifically) are both philosophically strong, and sufficiently divergent from the Intuitive theory that they cast doubt on Block’s conclusion.  Remember that in an argument of burdens of proof, reasonable doubt is all that is required.

            The first alternate theory I will consider comes from recent writings by Fred Dretske.  In Naturalizing the Mind, Dretske develops an alternative theory of Qualia that treats Qualia as representational in nature, that is, the subjective feel of something, the elusive aspect of mentality, can be reduced to talk about representation.[11] More precisely, qualia is a sort of metarepresentational effect,[12] that arises out of the different ways our representational systems are structured.  To illustrate this, Dretske uses a pair of speedometers, 


Imagine two instruments, J, and K.  J is a precision device manufactured to measure speed in hundredths of miles per hour.  When it registers 78.00 that (as the digits after the decimal suggest) means not 77.99 (and below) and not 78.01 (and above)… K is a less expensive device, designed to provide rough information about speeds.  Its registration of 78 means not 77 (and below) and not 79 (and above).  (Dretske 1997, 75).


Dretske argues that even when travelling at the same speed, J and K will ‘experience’ different things, because they represent the information in different ways (see especially Dretske 1997, 75, 78-80).  Further, he argues that this is at the root of all the subjective experience that we encounter.  His theory describes qualia as simply a product of each individual’s unique representational system.  Naturally, this example is simplified, and it is important to note that he is by no means making the claim that when two people differ, it is because one is better than the other, only that their representational schemes differ in some measurable, naturalizable way. The implication of such a theory then is obvious – qualia can be naturalized.  And if qualia can be naturalized, then they can be mechanized, which means that they no longer present a problem for China, nor any other properly implemented probabilistic automaton. While a full treatment of his theory is beyond the scope of this paper, it is clear that Dretske presents a coherent prima facie alternative to the Intuitive theory.

            A second escape from the Intuitive theory lies in its potential to contradict itself.  This argument is not so much an alternative to the Intuitive theory, but more a good reason why we should consider rejecting it, even if there were no acceptable alternative.  Stated simply, the objection is: 


·        If qualia cannot be captured by a probabilistic automaton, it is because they play no role in that which a probabilistic automaton considers, i.e.: they do not result in behaviour, or a change in state – they are not mentally causal.

·        But if they are not causal, they cannot cause memories, nor learning, nor any other mental phenomena.[13]

·        And yet we retain sufficient memory of qualia that we can describe them (however awkwardly) and even develop philosophical theories about their natures, therefore they must be causal in some way.

·        Thus, by another application of modus tollens, the claim that qualia cannot be represented in a probabilistic automaton must be false.[14]


This objection is not an easy one for Block to circumvent if he wishes to maintain his rejection of functionalism.  He cannot challenge the first claim - that by giving up its place in the probabilistic automaton, qualia gives up mental causality - without inventing some new form of causality not expressed by said automata.  If he accepts their lack of causal role, then the second claim is a logical necessity.  Finally, to reject the third premise, to state that we do not in fact have a memory of qualia per se would be to give up entirely, since his objection rests on qualia being very real.  It is possible that with sufficient argumentation, Block could develop a defense to this argument, but remember that all we require is prima facie doubt.  Clearly, if his assumptions lead as quickly to contradiction as those he presumes to dismiss, there is reason to doubt his conclusions.

            Block is correct in saying that conceivability arguments should always be regarded with a great deal of skepticism.  The danger of appealing to our intuitions is that if they were all we relied upon, we could never develop the truly counterintuitive theories that are frequently quite valuable, if not true.  There is every reason to suspect that if China did have mentality, it would be of such a vastly different nature that it is doubtful we would ever comprehend it, let alone intuitively predict it.  Contrariwise it could have mentality just like our own, perhaps mind is consistent over vast differences of scale and implementation.  More importantly though, this paper is not about proving that China has mentality at all, it is simply about establishing reasonable doubt.  Block’s argument is predicated on the assumption that what functionalism predicts about China is false, but he has only burden-of-proof arguments upon which to base that conclusion.  Through analysis of the argument, and consideration of some persuasive alternatives, it becomes clear that there is at least as much argument for China’s mentality as there is against it.  Functionalism, in other words, is not refuted by Block’s argument.  But as I stated at the beginning of this paper, all I can do is shift the burden of proof back on to his shoulders – there may be other good reasons why functionalism fails.  The China argument is simply not one of them.


Works Referenced


Block, Ned.  ‘Troubles with Functionalism’, in Block, 1980.  pp. 268-305.  1975.

Block, Ned, ed. Readings in the Philosophy of Psychology, vol. 1.  Cambridge, MA: Harvard University Press, 1980.

Dretske, Fred.  Naturalizing the Mind.  Cambridge, MA: MIT Press, 1997.

Nagel, Thomas.  ‘What is it like to be a bat?’  Philosophical Review.  4 (1974).  pp. 435-450.

Putnam, Hilary.  ‘The Nature of Mental States’, in Block, 1980. pp. 223-231. 1967.

Putnam, Hilary.  ‘The Mental Life of Some Machines’, in Intentionality, Minds, and Perception, ed. Hector-Neri Castañeda, Detroit, 1967, pp. 177-200.  1967a.

Shoemaker, Sydney.  ‘Functionalism and Qualia’, in Block 1980.  pp. 251-267.  1975

Shoemaker, Sydney.  ‘Absent Qualia are Impossible’, Philosophical Review. vol. 90(4).  pp. 581-599


[1] Of course, I do not mean to imply by this that functionalism is the only available theory of mind.  Rather, that in the traditional program of philosophy of mind, (that is, to the exclusion of recent connectionist and dynamical systems models) there are few other materialist philosophies with the strength that functionalism possesses.

[2] Consider the original type-identity theorist as an example of the former, and Churchland’s eliminativism for the latter.

[3] Or perhaps, to borrow Block’s terminology, the argument between establishing a prima facie case for, and a prima facie case against.

[4] Or better, functional units.

[5] For an explanation as to how Turing machine tables can be extended to become probabilistic, see Putnam 1967a.

[6] From this point on, I will assume that all arguments apply equally well to both thought experiments since one is just the other writ large, they are functionally equivalent.

[7] As Block is no doubt aware, a truly compelling, objective reason to accept our own qualia as valid is exceedingly difficult - hence his appeal to the rather elusive Anthropic Principle.

[8] Or, to untangle the double negative, that there is reason to accept the possibility of China’s mentality.

[9] (Block 1976, 278)

[10] Block’s reference, and my own, are to Nagel, 1974.

[11] Admittedly, this brings along its own set of problems, the mechanics of representation is not a topic that has been solved, by any stretch of the imagination.  Naturally, a proper treatment of the debate is beyond the scope of this paper, but is also unnecessary, as we need only demonstrate prima facie that adequate alternative theories exist.

[12] This is my terminology, not his, to the best of my knowledge.

[13] This is clearly an argument similar to that levelled against epiphenomenalism.

[14] I arrived at this argument independently, although a similar argument has been made in Shoemaker 1975 and again in Shoemaker 1981.