Consciousness and the Introspection of Apparent Qualitative Simples

Presenter: Paul Churchland, University of California San Diego



  1. William S. Robinson
    Iowa State University

    Professor Churchland, you ask (p. 17; see also p. 14) “Why should we think that their inner analogues [i.e., inner analogues of sensible qualities of things] are any different?” I take this to be asking why the account of qualitative features of our own conscious states should not be expected to be parallel to the account of sensible qualities of things that has been accepted in outline at least since the days of Locke. I think you mean the question to be rhetorical, inviting the answer “no reason”. But it seems to me that there is a straightforward reason for rejecting the expectation of parallelism.

    To wit: The full Lockean account of external sensible qualities rests on the principle that F in bodies is whatever causes F sensations. For example, warmth in bodies is what causes warmth sensations – which leaves it open for science to discover that it’s mean kinetic energy that satisfies this causal role. Red in bodies is whatever causes red sensations, and science discovers that what plays this causal role is various surface molecular constitutions that result in the production of members of a certain class of reflectance profiles.

    If we try to impose a parallel account for the qualities of sensations themselves, we will have to regard sensations as having natures that are causes of further sensations (or causes of something referred to by a new term that has a role analogous to that of “sensation” in the Lockean account). If we try to account for the effect term of this new causal relation by reapplying a parallel strategy, we will be off on an infinite regress.

    The usual anti-dualist way of avoiding this regress – and the view that I’m inclined to think you mean to accept – is to claim identity of neural event properties and sensation qualities. This view is not a parallel to the Lockean account of sensible qualities of things, and it would be a confusion to invoke the success of the Lockean account (with, of course, updated science) in support of it.

    I’m not quite certain that you do mean to accept an identity claim. (1) On p. 16, you seem to be approving common sense when you say that it “ascribes both kinds of aspects/properties to one and the same internal states . . .” (emphasis in original). Both kinds of properties? That is property dualism. (2) On p. 19, we find that our “native, internal discriminatory mechanisms are also cognitively blind to whatever ontological complexities might happen to underlie these internal qualitative characters . . .” (emphasis in original). But underlying is an asymmetrical relation, and is thus incompatible with identity.

    The key phrases in (1) and (2) seem to be ways of avoiding the starkness of the denial of property dualism. If there are not two properties, there is only one. So, for example, if property dualism is false then the redness that’s in your sensation – that color itself, not its cause – is the very same property as being a set of neural firings of such and such a kind; and blueness, warmth, and so on are the very same properties as being neural firing sets of such and such other kinds.

    I confess to some surprise at seeing you appeal for support to Folk Psychology (pp. 16-17). But if we do consider such an appeal, I believe that Folk Psychology will not be found to accept the identity of sensation qualities and neural activation set properties – once that view is clearly distinguished, as it must be, from a claim that’s perfectly compatible with property dualism; namely, that neural events are bearers of “both kinds of aspects/properties”.

    The argument remains that a property monist view would be preferable to its main rival, which you rightly identify as epiphenomenalism. Apart from the appeal to folk psychology, you give two arguments against epiphenomenalism. The first is a version of the self-stultification objection, the second is based on the alleged need for a ‘someone’ to ‘host’ qualitative features. These arguments are too complex to comment on here, but replies already exist. The 2006 paper below counters the popular self-stultification point, and skeptical views about the need for subjects of experiences are explained and defended in the two books.


    Robinson, W. S. (2004) Understanding Phenomenal Consciousness (Cambridge: Cambridge University Press)

    Robinson, W. S. (2006) “Knowing Epiphenomena”, The Journal of Consciousness Studies 13:85-100.

    Robinson, W. S. (2010) Your Brain and You: What Neuroscience Means for Us (New York: Goshawk Books). Description available at

  2. Hi Paul, thanks for this excellent paper. I am largely sympathetic to the line that you are taking here. I think that the point about light is especially good. Because of the way our sensory systems work a think of light as essentially the thing which makes thing visible, when that is just a small portion of light. Kripke makes a similar point when he talks about possible worlds where something other than electromagnetic radiation produces visual experiences akin to ours. This is a world where there is no light since there is no electromagnetic radiation. We get tricked, though, because of the intimate connection that visible light has in the way we pick out the concept. Kripke calls this the illusion of necessity.

    I think that something like this goes on in the qualitative/phenomenal cases as well. The painfulness of pain seems to be essentially a part of the pain, but phenomenana like Pain Asymbolia push us to separate the two in much the same way that modern physics forced the separation of light from ‘that which makes things visible’. Pain and painfulness seem necessarily connected by this may just be an illusion fostered by the way we pick pain out (by its painfulness).

    But I do have some reservations with the argument that you develop. First I wonder about your arguments trying to blur the objective/subjective distinction. All of the examples from interoception, for instance, seem to conflate the two ways of knowing. I agree that what you know about is one’s body position, motion, etc but one knows it via a subjectively mediated state, as you admit. So one can have knowledge of an objective fact subjectively, but in so doing one also knows what it is like for one and that does, at least at first, seem like knowledge of something different, something that you can’t know. You may be able to agree with me that my stomach is full but only I know what it is like for me to have a full stomach. One might even go on to argue that what one knows directly is that one’s stomach feels full, and then infers that it is (indeed, it takes abut half an hour for one to become conscious of one’s stomach as full).

    The same can be said about the converse claim. I may know about the state of your stomach, but I cannot know what it is like for you to have a full stomach.

    I think, just as you do, that the above issues are more pressing than the bat issues. If we accept, as I do, that in order to fully posses a phenomenal concept one must have had the experience then physicalism is only committed to our being able to know what it is like for creatures that have the experiences that we do (or maybe, if one like Hume’s missing shade of blue, ones that are sufficiently similar to ones like ours). I take it that this is your point when you talk about being a ‘bat-style cognizer’.

    But all of this means that once Mary has seen red, and therefore has the phenomenal concept ‘red,’ she would be in a position to deduce which brain state ‘red’ just was a certain brain state in just the same way that we are now able to deduce that water is H20. But we do this while taking the subjective/objective distinction seriously.

    The second concern I have has to do with the argument against epiphenomenalism. I share your distrust of epiphenomenalism but we need to develop arguments against it that take it seriously. the reason for this is because without that the epiphenomenalist has a ready reply to all of the physical data objections. OF course knowing the brain structure will help us predict conscious experience since there is postulated to be a law which relates brain structure to conscious experience. The two realms of properties are thought to share structural similarities in their overall ‘space shape’.

    So, when you ask ‘in what does their apprehension consist?’ the answer given is that it consists in a hybrid physical/non-physical state. When I apprehend my conscious experience of red, on this view, the conscious experience of red is itself a part of my introspective state, but not in a causal way. It is partially constitutive of my knowledge. So when I know that I am seeing red my knowledge partially consists in the non-physical red quale. I agree that there are issues about how a non-physical state can be ‘bound’ to a physical state in this way without causally effecting it (which is what I take your second objection to be getting at), but it is an answer to your question.

    These kinds of consideration may be appealing to those of us already skeptical of epiphenomenalism but it seems to me that the best way to get at the people who actually go in for epiphenomenalism is to push on cases of pain inversion. People generally come to accept that colors may be inverted but can we really make sense of pain and pleasure being inverted without any functional difference? This seems highly unlikely and suggests that epiphenomenalism is faulty on its own terms.

  3. First, thanks to Prof. Robinson for his lucid and penetrating commentary. It is focused on the issue of whether the ontological/epistemological status of external color (i.e., as a feature of objects) is the same as or different from the ontological/epistemological status of ‘internal’ color (i.e., as a feature of sensations). I claim that, while the two features are indeed quite distinct (specifically, sensations are not literally colored in the same way as external objects), the two cases are entirely parallel, ontologically and epistemologically. Prof. Robinson suggests some historical (i.e., the views of Locke) and analytical (the potential infinite regress of sensations-of-sensations) reasons for claiming that the two cases are quite different.

    To begin, I reject Locke’s analysis of “is red” as “whatever power it is, in objects, that typically causes a sensation-of-red (here construed as a subjective phenomenological characteristic) in normal humans.” Partly because I reject ‘analyses’ in general (on Quinean grounds). And partly, but even more importantly, because the undoubted fact that objective redness tends to produce certain kinds of visual sensations in normal humans is among the least important (ontologically) and spectacularly contingent facts about objective redness that I can think of. Redness has been a robust feature of things in the Universe (e.g., stars, planets, rocks, etc.) for many billions of years before we humans ever arrived on the scene to react to it in our fragile and profoundly idiosyncratic neurobiological ways. To get an objective, non-species-relative grip on what redness is, we need to go to the physics of color, specifically, to the electromagnetic reflectance-profiles and emission-profiles of physical objects. That, if anywhere, is where the ontological meat resides. And, however inadequately we may currently be conceiving of them, those very power-spectrum profiles are what we are perceiving (however opaquely) whenever we perceive colors, for that is what the colors are. Our current conceptual responses to the relevant class of sensations (in terms of our familiar color vocabulary) are currently inadequate to the more complex objective reality that lies behind them. We need a new vocabulary for describing objective colors, one drawn from our matured Physics. Once we have become practiced in using that new vocabulary in spontaneous reaction to our (unchanged) sensations, our perception of the chromatic world will have become much more penetrating.

    Second, I reject Robinson’s idea that the successful introspection of a sensation S would require a secondary sensation-of that sensation S. What successful introspection requires is only that S cause a conceptual response of some sort, a cognitive response that classifies S as being some specific sort of sensation. Here also, our current common-sense conceptual resources and responses (in terms of sensations of-red, of-blue, etc.) are inadequate to the more complex (neuronal) reality that lies behind them. Parallel to the case of Physics and objective colors, it is the Neurobiology of color-coding that provides the real nature of our diverse sensations-of-color (presumably the Hurvich-Jameson ‘opponent-process’ account of our chromatic processing), and once we have become practiced in using its ‘neural-activation triplet’ vocabulary in spontaneous cognitive reaction to our visual sensations, our introspection of our sensational world will have become much more penetrating, for activation-triplets are ultimately what our sensations-of-color are. I freely confess, therefore, to being an Identity-Theorist, at least where color-sensations are concerned.

  4. Response to Paul Churchland’s post on 2/21 at 14:13

    Churchland, first paragraph: “the two features [i.e., ‘external color’, a feature of objects, and ‘internal color’, a feature of sensations] are indeed quite distinct (specifically, sensations are not literally colored in the same way as external objects) . . . .”

    Robinson: We agree. I’d just add that we could equally well make the following parenthetical remark, either in addition to or in place of the one in the quote: (specifically, external objects are not literally colored in the same way as sensations).

    Regarding the second paragraph of your response, I pretty much agree, provided it’s clear that the feature referred to by “redness” in this paragraph is the feature of objects. (So understood, Locke can agree too.)

    Coming to your third paragraph, I first want to note that I didn’t say anything about introspection. Of course, “introspection” may plausibly be used in such a way that if I know something about my own sensations, it’s guaranteed that I know that by introspecting. But still, I was talking about the sensations known, not about the conditions for getting to know them.

    More importantly, you say that “our current common-sense conceptual resources and responses (in terms of sensations of-red, of-blue, etc.) are inadequate to the more complex (neuronal) reality that lies behind them”. There are two points I want to make in response to this.

    (1) *Lies behind* is an asymmetrical relation. Thus it’s incompatible with the identity that, in this response, you clearly espouse.

    (2) To say that our current common-sense conceptual responses in terms of color sensations are inadequate to the more complex neuronal reality is a quasi formal-mode reflection of something more fundamental, which stated in material mode is this: the internal colors of sensations are not complex in the way that neural events are complex. (If they had had that kind of complexity, neuroscience would have been a lot easier to come by that it has been.) If the features that are the internal colors of sensations are less complex than the properties of neural events that lie behind them, then those features are not identical with those neural properties.

    I think there are only three choices here. (A) Accept the distinctness of features of sensations from neural properties and become a property dualist. (B) Hold that features of sensations just are neural properties (and explain how, after all, a single property could both have and lack a certain kind of complexity). Or (C) Hold that our common-sense concepts of features of color sensations are not concepts of any property that is actually instantiated at all (and explain how we could be under the illusion that they are).

    It’s clear you’re rejecting (A), but I’m not so sure about your choice on (B) vs (C). Or do you think I’ve missed a fourth possibility?

  5. Tom Clark
    Center for Naturalism

    In response to Paul Churchland:

    Many thanks to you (and the conveners) for making this paper available for comment. You argue, correctly I think, that subjective knowledge isn’t of an ontologically special class of qualitative simples (what I will call qualia), but of physical states of affairs, for instance of states of the body as represented by neural goings-on in the brain and nervous system. But this doesn’t necessarily mean that qualia are reducible to neural goings-on, what you hope to show. It might mean that qualia are not themselves objects of knowledge to which we stand in an epistemic relation, but rather *the terms in which* knowledge is available to conscious subjects. It’s these qualitative terms, present only to a system that instantiates the true theory of consciousness, that make the system a locus of *subjective* knowledge (your example: we subjectively know about neural color system fatigue *in terms of* qualitative after-images) .

    I doubt we’re in a direct observational or representational relation to our own experience – we don’t have a literal perspective on it.[1] If so, we can perhaps agree that “the case for ‘nonphysical facts’ evaporates.” Facts are what can be specified about objects of knowledge; but things like pain, red, and other experiential simples are not themselves things we know things about: we can’t specify a further experiential fact about red – what it is *about* red that makes it red, or what it is that makes pain painful. Rather, qualia are the basic terms in which states of affairs *in the world*, including the body, are characterized in consciousness.

    You aim to achieve an explanatory reduction of qualia “in terms of the physical dimensions of brain activity,” such that there is nothing *really* present, *really* existent, but the physical brain – a physicalist ontological monism. At the same time, however, you want to maintain, along with commonsense and folk psychology, that the “qualitative features” of brain states – qualia – play causal roles and thus aren’t epiphenomenal. Qualia therefore aren’t eliminated on your account. Since you think substance and property dualisms are explanatory non-starters (since physical systems are “dynamically closed”), the causal potency of qualitative character requires they be identical with something physical, and indeed you say “They [qualitative characters] are, all of them, complex neural and physiological states…”

    So the question about reductive explanation boils down to the plausibility of this identity claim (although as William S. Robinson points out in his post of February 18, 2011 at 14:20 you seem to vacillate between an identity claim and property dualism). In arguing for identity (what Robinson calls property monism), you try to cast doubt on the idea that qualia *are* ontologically simple and non-physical just because they *appear* simple: by dint of training we can often learn to discriminate further qualities in what seemed non-decomposable. But any discriminated component will itself be *experientially simple* – a quale – which is the phenomenon needing explanation. The simplicity is an *unequivocal fact about the phenomenal appearance*, not something that can be shown to be an illusion. Properly understood as *really simple appearances*, qualia survive as real subjective existents, no matter how far we go in decomposing experiences into distinct sub-components.

    You further suggest that subjective simples will likely fall to the same reductive analysis as what you call “*external* qualitative sensory characters,” such as the warmth of air in an oven, which have been shown to be more or less complex physical phenomena (e.g., warmth ends up being mean molecular kinetic energy). You ask “Why
    should being located *inside* the skin introduce such an enormous ontological contrast with qualitative states that are located *outside* the skin?” But the qualitative states at issue – qualia – are originally and primarily individuated as elements of subjective *experience*, in contrast to external qualities individuated as features of the physical *world*. This, not qualia being located inside the skin (if indeed it makes sense to assign qualia a spatial location), is what introduces the prima facie ontological contrast. So the fact that apparently simple external qualities turn out to be complex and physical doesn’t raise the plausibility of the claim that internal qualities (really simple phenomenal appearances) will fall to reductive analysis. The physicality of external qualities was never in question and their composite complexity is straightforwardly observable; neither is true of qualia.

    I agree with you that epiphenomenalism leaves it “an absolute mystery what these ‘ontological simples’ are and why they should exist at all.” The epiphenomenalist has precisely no story of how qualia, conceived as non-identical to brain states, are produced or generated by those states. But again, this doesn’t mean qualia thus conceived don’t exist, only that a causal account of them might be off track. They might be *non-causally* entailed by one’s *being* a representational system that instantiates the true theory of consciousness [2]. Phenomenal experience, categorically private and subjective as it is, hence not available to 3rd person observation and measurement, seems logically barred from playing a causal role in 3rd person explanations of behavior. This suggests that from the perspective of science *consciousness isn’t even epiphenomenal*. But from our perspective as conscious subjects it’s counter-intuitive, and practically speaking unnecessary, to deny experience a role in causing behavior [3].


    [1] Clark, T.W. (2005) “Killing the observer,” Journal of Consciousness Studies,

    [2] Clark, T. W. “The appearance of reality,” draft,

    [3] Clark, T. W. “Respecting privacy: why consciousness isn’t even epiphenomenal” draft,

  6. Thanks to Paul for an interesting and enjoyable paper. Forgive me for narcissistically just commenting on the parts that are about me, and also for making the boring response that the views attributed to me are not my own.

    In particular, section 4 (and the key paragraph on pp. 12-13) seems to hallucinate a putative argument of mine out of thin air. Paul says that I say that consciousness is ontologically simple and argue to dualism from there. I’m pretty sure I’ve never made any such argument. My own work on consciousness has very often stressed the complexity and the structure within conscious states, just as Paul’s does. And just as Paul does, I think that this structure opens the way to certain forms of explanation of consciousness. We differ on whether those explanations will be reductive or nonreductive, of course. But we don’t differ about that because of some prior difference about whether consciousness it ontologically simple.

    It’s true that I at least entertain the possibility that some phenomenal properties are fundamental properties, and then are presumably simple properties. But I don’t know that this is true, and I’m equally open to a view on which all phenomenal properties are complex and derive from fundamental protophenomenal properties. Furthermore, even if some aspect of consciousness is simple, I’m pretty sure it wouldn’t be something as gross as the sensation of redness (the view Paul attributes to me). And in “Moving Forward on the Problem of Consciousness” I take seriously the idea that consciousness has “structure all the way down”. Insofar as I entertain the view that phenomenal properties are fundamental, it’s as the conclusion of an argument against physicalism and not as a premise. I’m sure that Paul doesn’t buy the arguments that I actually give, but there’s nothing about them in the paper.

    As for the final section on epiphenomenalism, a view Paul says I “explicitly embrace as descriptive of his position”: My 1996 book, although not unsympathetic to epiphenomenalism, was neutral between that view and non-epiphenomenalist positions such as panpsychism and Russellian monism. My 2002 paper “Consciousness and its Place in Nature” was much less sympathetic and is neutral between those views and interactionism. It’s true that I think that epiphenomenalism is a contender, but it’s not my view (it’s probably my least favorite of the views I think are contenders). Of course I think the argument Paul gives against it (basically the argument from knowledge and the argument from commonsense) can be rebutted: I’ll leave it to Bill (and Chalmers 1996) to rebut them.

    I’m inclined to think that the best argument against epiphenomenalism is the argument from coincidence: it makes it a lucky accident that the laws of nature are such as to keep consciousness and behavior in sync with each other. Maybe there is some response to this argument by e.g. showing that in-sync laws are somehow simpler and more elegant or more cohesive with the rest of nature than out-of-sync laws, but I haven’t yet seen a plausible case for this.

  7. Hi Dave, you say that panpsychism and Russelian monism are not forms of epiphenomenalism but I don’t see how that could be the case. Isn’t any view that accepts that the zombie world is (2-)possible committed to the causal inefficacy of qualia? Granted that these views may not be forms of property dualism but they still can’t allow a casual role for qualia can they?

  8. Hi Richard — Russellian monism in both panpsychist and non-panpsychist variants certainly allows a causal role for consciousness. On this view, fundamental phenomenal or protophenomenal properties are the categorical basis of fundamental microphysical dispositions. As such they are the causes of the relevant microphysical effects — in fact all microphysical causation can be seen as connecting these properties. So insofar as our consciousness is constituted by and metaphysically supervenient on these properties, there’s no bar to it having physical effects too, or at least no bar any worse than the ordinary exclusion problem for metaphysically supervenient high-level properties (whereas standard property dualism is faced with a much worse problem due to having only nomological supervenience).

    Re zombie worlds, a Russellian monist will regarde these as either worlds where (i) microphysical dispositons have a different categorical basis (one with nothing to do with consciousness) or (ii) they have no categorical basis at all. Importantly the possibility of (i) or (ii) does nothing to show that the original categorical bases isn’t causally relevant in our own world. Inicdentally I think that interactionism can also allow the possibility of a zombie world (perhaps with causal gaps) but interactionism certainly allows a causal role for consciousness.

  9. Thanks for the follow up Dave.

    I think I am a little confused at this point. I don’t see how you can say that a conscious experience of pain, say, is the cause of my wincing but then go on to say that we can remove the conscious experience of pain without altering the causal system at all (I still wince). We remove the fundamental properties (the ‘insrutables’ as Barbara Montero calls them) and the lights go out but everything else goes on the same. So what were the insrutables causing? Maybe they we causing epiphenomenal qualia but how could they be causing the microphysical effects? It seems to me that this is more than the problem of how high level supervenient properties can be causally efficacious, it is the question of how the insrutables themselves could have any physical causal powers. But maybe I am missing something?

    As for the interactionism point, I wonder why you think that the physicallist can’t make use of the same strategy?

  10. I don’t want to hijack the discussion of Paul’s article here. But just quickly:
    possibility (i) was that the zombie world involves new inscrutables taking over the microphysical causal role previously played by (proto)phenomenal inscrutables. Obviously it’s a fallacy to say that A doesn’t cause B just because B could have occurred without A (instead caused by something else, C). Possibility (ii) was that the zombie world lacks inscrutables altogether and is a pure structure world in which fundamental dispositions lack categorical bases. It’s not at all clear that this is coherent, but if it is, I think one should allow that the mere fact that a disposition could lack a categorical bases does not entail that its categorical basis is inefficacious when present. There are roughly similar options for interactionism.

  11. re: “consciousness is ontologically simple” vs “the complexity and the structure within conscious states.” Could it be simple as a fundamental property and the conscious states themselves be complex? For instance, some argue that meaning is virtual, that it is not present at an instant. And likewise, if conscious drops of experience are fundamental, the complex feeling is built up from the connections in the network — in the network they add up to a substantial thought with certain effects and inferential relations. I’m not sure if property dualism is needed as much as a broader notion of logically supervene. It seems that if the metaphysical implications of propositional attitude predicates are the same as those of physical measure ones, then we could talk about one inter-connected causal system (which I assume is close to the idea that the physical world is closed). Perhaps a vector describes the impulse of a germ of thought, and following its path arrives at a certain destination. But I worry about enjoying the ride. There was a discussion at the Cuny CogSci group one time between Strawson and Fodor where Strawson mentioned “what makes life interesting” and Fodor seemed more concerned about the “causal significance.” Fodor insisted that the output of a memory dump was the right kind of comparison to thought. I think, though, that even if you idealize away from any warming or shaking of an old computer, the memory dump leaves out the feel of thinking: that there is at least an “ah-ha” feel (which may be the same for many different thoughts and so is not necessarily connected to this token thought but rather thinking itself or grasping a rule etc). I’m worried that claims about intrinsic mass relations and the non-existence of numbers may be unhelpful when applied to the context of eliminating consciousness. I don’t know how we will discover (through stipulation maybe?) a class of novel fundamental properties from which phenomenal properties are derived or find the essence of consciousness or the physical. Perhaps ostension is used here, but there would be worries about what kind of theoretical or conceptual setup is behind fixing the reference. Is anyone in agreement on whether pain’s feel is essential to it, or whether there is a function of consciousness, or what a non-revolutionary extension of a science is? It seems that the right direction is an explanatory pluralism that allows for inter-theoretic reduction but difficult issues remain about essence and identity. More on those soon.

  12. Prof. Churchland provides robust and effective responses to some of the many varieties of dualism. There are though, I feel, serious difficulties with the materialist outlook he appears to propose as an alternative; one that would provide, using a “physics of color”, a “non-species-relative grip on what redness is”.
    The question remains whether it could provide a complete description of “Life, the Universe and Everything”, a question the thought experiments involving monochrome-Mary, zombies, Nagel’s Bat, and many others attempt to highlight. I agree with Prof. Churchland that it is misguided to use these in support of a dualist position, however the difficulty remains unless the logical framework underpinning the materialist stance is clarified. We need not just a scientific description, we need to understand the philosophy of science as well. This, I feel, Prof. Churchland fails to achieve.
    In very general terms, science consists of hypothetical assertions; the ability to infer, from such hypotheses, to assertions of actual situations being the case; and the ability “to see” (in a very broad sense of “in some sense, being aware”) whether such situations are in fact the case. In other words, science presupposes that we can think, and that we can see. And that we do so in the same way; in a broadly human way, allowing for individual differences such as colour-blindness. It does not provide a non-species-relative grip (though neither is it species-relative).
    Some disciplines, neuroscience for example, provide descriptions of how it is we are able to think and to see. There is though, a danger of circularity here, in that neuroscience presupposes that we can think and see in order to describe/explain how we are able to. As long as we are clear that it is a person that sees, and that particular neural activity is necessary in order that a person can see, the circularity is avoided. It is the person that sees, not the neural activity. [1]
    To make this distinction we need our everyday language, which Prof. Churchland and others refer to, perhaps derisorily, as folk psychology, but which is perfectly adequate to convey this distinction. Without it we can’t. The spectrum-inversion thought-experiment helps clarify this claim.
    Logically it is possible that someone may see what I see as red, as blue; and yet there be nothing that can be taken that is evidence of this (unlike with, say, colour-blindness). The important point is that for a scientist this is of no pertinent relevance; as long as two observers agree as to what they are seeing, this is all that matters, If we infer, using a scientific theory, that something will change from red to blue, it doesn’t matter that one observer may actually see the colours in inversion to another observer; as long as they agree as to what they are seeing (cf Wittgenstein’s Beetle). This is all that is meant as seeing and thinking in the same way.
    Science presupposes this agreement; it doesn’t itself explain or describe it. To do so requires non-scientific, everyday, “folk psychology”, language. To emasculate language of this, would deprive science of its presuppositions and render it meaningless.
    [1]Bennett M. R. & Hacker P. M. S. (2004), Philosophical Foundations of Neuroscience, Oxford, Blackwell Publishing Ltd

  13. Fantastic and beautifully clear paper! Don’t agree with a word of it though.

    I am also dubious of the general appeal to armchair intuitions as a guide to the nature of reality. It is plausible to think that armchair reflection tells us about our concepts, but difficult to see how we can ever know from the armchair whether or not those concepts are satisfied. But when it comes to consciousness, we are in a unique position of having a set of concepts we know for certain are satisfied. Suppose you’re in agony, you focus on your pain and form the concept ‘feels like that’; there’s no epistemic possibility that that concept is not satisfied.

    Of course, just because you know a concept is satisfied, it doesn’t mean we know anything about the nature of its referent, as it may not reveal its conditions of satisfaction a priori – call a concept that doesn’t reveal its satisfaction conditions a priori ‘opaque’. You seem to hold in later sections that phenomenal concepts are opaque (even though you don’t explicitly talk about concepts) as you hold that phenomenal qualities can turn out to be very different to our armchair conception of them (In this you seem close to embracing the phenomenal concepts strategy; I would be interested in knowing what you think of this form of physicalism).

    However, there are strong reasons to think that phenomenal concepts are not opaque. We have rich a priori knowledge of our internal relations our conscious states bear to each other, e.g. phenomenal red is similar to phenomenal orange. It is difficult to see how we could know a priori that these two qualities have an intrinsic nature such that they necessarily resemble each other, if we cannot know a priori anything about their intrinsic nature.

    If phenomenal concepts (i) are transparent, i.e. reveal their satisfaction conditions a priori, (ii) are certainly satisfied – and I think we have good reason to believe both of these things – then we are justified in taking armchair reflection on consciousness as a guide to the nature of reality: we can know a priori what it takes for the concept to be satisfied, and we know for certain that it is satisfied.

    Different point regarding your concerns about the causal efficacy of phenomenal qualities: Russellian monism – the view Chalmers calls ‘type-F monism’ in his 2002 paper ‘Consciousness and its place in nature’ – can make sense of the causal efficacy of conscious states in a way that is consistent with the causal closure of the physical. (Chalmers is open to this view and to substance dualism, so it is not quite right to label him a property dualist). Also, if we believe in phenomenal intentionality, a once ridiculed view which is increasing in popularity, it may be that consciousness can ground thought about itself, without there needing to be causal interaction between the phenomenal and physical. Chalmers spends a lot of time in ‘The Conscious Mind’ trying to make epiphenomanlism consistent with thought about consciousness; you provide no objections to the account he gives there.

Comments are closed.