Саморепрезентация, феноменология и Трудная проблема сознания

14 Июня, 2018

Интервью с Кеннетом Уиллифордом, профессором Техасского университета в Арлингтоне, редактором сборника «Self-Representational Approaches to Consciousness» (2006).

Представляем вашему вниманию одно из самых красивых интервью, которое было снято нами летним днём на фоне главного здания МГУ — интервью с Кеннетом Уиллифордом, профессором Техасского университета в Арлингтоне, редактором сборника «Self-Representational Approaches to Consciousness» (2006) и другом нашего Центра. Взгляды Уиллифорда, в прошлом изучавшего философию Беркли, близки репрезентативным подходам к объяснению сознания, и именно они обсуждаются в данном интервью.

Кроме того, ниже вы найдёте развёрнутый комментарий на данное интервью самого Кеннета Уиллифорда.

Self-Representationalism, Phenomenology, and the Hard Problem:
Comments on an Interview


In the summer of 2014 I was in Moscow visiting my wife’s family. The April before I had met Artem Besedin, Anton Kuznetsov, and Dimitriy Volkov at the 2014 “Toward a Science of Consciousness” conference in Tucson, Arizona. Artem and Anton proposed interviewing me during my visit to Russia. I would like to thank them for their hospitality and interest. I would also like to thank Dimitriy Volkov for making the Moscow Center for Consciousness Studies a reality. The Center has supported some truly interesting and valuable endeavors and exchanges and has very generously made them freely available to the public. I would also like to thank Dean Vladimir Mironov and Deputy Dean Anna Kostikova, both in the philosophy faculty of Moscow State University, for their hospitality and generous intellectual interest and good will. Огpомное спасибо!

What follows is a sort of re-presentation of my answers to Artem’s questions and some references to relevant literature including some papers of mine where I develop the ideas expressed (sometimes rather laboriously, I must admit) in my responses.

Question 1: The Influence of Bishop Berkeley

In graduate school I spent a fair amount of time working on George Berkeley’s philosophy under the direction of Phillip Cummins (University of Iowa) and Bertil Belfrage (Lund University). I worked mostly on Berkeley’s philosophy of language. I also worked on Hume’s Dialogues Concerning Natural Religion under Cummins’ direction. Artem asks me if the work on Berkeley had any influence on my subsequent work on consciousness. Two things came to mind.

First, I tended to agree with Hume that Berkeley’s attack on the concept of material substance can be extended to the concept of mental or “spiritual” substance (see Berkeley’s Principles of Human Knowledge, e.g., §§ 7, 16-17, 135; Hume Treatise of Human Nature, I.iv.v, “Of the immateriality of the soul”). Berkeley held that only minds are substances and have causal powers (Berkeley, PHK, §§ 7, 25-27). Notoriously, he had some difficulty with the fact that we lack “ideas” of the mind just as we lack them for material substance. This quandary led him to claim that while we lack ideas of the mind, we do have a “notion” of the mind, just as we have notions, but not ideas, of relations (Berkeley, PHK, §§ 89, 135-140). Hume, who tended to be a little more ruthlessly consistent in his empiricism, seized upon this lack of positive ideas for the mind and concluded, equally notoriously, that the mind is just a “bundle” of ideas (Hume, THN, I.iv.vi, “Of personal identity”).

The lesson I draw from this dialectic (one that, in the early modern period, extends from, roughly, Descartes to Kant — and not just to Thomas Reid, as I say in my response) is that we cannot determine what ontological category consciousness belongs to from introspection (or phenomenology) alone. Is it a substance? A relation? A process? A bundle of property instances? We can’t answer this question just on the basis of reflection alone, since the data can be made to fit any particular ontological framework one prefers. One might associate this lesson, more or less, with Kant (especially the “Paralogisms” in the Critique of Pure Reason), but I am suggesting that the early modern debates on this issue (mostly carried out as debates about mental substance) can in part be explained by the fact that consciousness tells us so little about its ontology.

The second “influence” of Berkeley on my thinking about consciousness has to do with Berkeley’s view that there can be no ideas without minds (Berkeley, PHK, e.g., §§ 3, 7). For example, pains and visual images do not just float around without belonging to some mind or other. They are always the objects of someone’s conscious awareness. I do not, however, interpret this metaphysically in the way that Berkeley does. Nor would I deny that in a certain sense there may be unconscious mental states. I take this point to be, rather, that conscious mental states always include an appearance of __ to__ structure; or, to put it in phenomenological jargon, there is a “dative of manifestation” in addition to a “genitive of manifestation”. When anything whatsoever appears in consciousness it appears to someone. Conscious pains, for example, don’t just “appear” simpliciter. They appear to the conscious being having the pains. The grain of truth in Berkeley’s “no ideas without minds” claim is a phenomenological point: consciousness seems to have a subject-object structure. When Hume denies that he can find himself upon introspection, one way to interpret him is as denying that consciousness has this structure (see Butchvarov 1959). For Hume, there could be ideas just appearing that appear to no one. I am on Berkeley’s side on this question, construed phenomenologically; but I am on Hume’s side if we construe the issue metaphysically. In other words, I agree (with Berkeley) that there is a phenomenological subject of consciousness, but I disagree with him that this is an enduring substance. I agree with Hume that there is no such enduring substance.
I hasten to add that accepting this phenomenological point does not commit one to homuncularism or some reified or robust notion of Dennett’s Cartesian Theater (see Dennett 1991). On my view, every episode of consciousness appears to itself; every episode is its own subject. This is a sort of ever-repeated, reflexive structure; it is not an entity or transcendental ego. To say that every conscious pain appears to someone, is, on my view, just to say that every episode of pain-manifesting consciousness is also self-manifesting. So there is something that is itself phenomenologically manifest to which the pain is manifest, namely the conscious episode itself. If we don’t say this, we will either have to accept a Humean view according to which pains could be manifest but manifest to nothing or a view according to which that to which pains are manifest is itself phenomenologically hidden and perhaps even inaccessible. But I take it that it is a phenomenological datum that we are aware of that to which pains (etc.) are manifest. If one also thinks that this “that to which…” is not a homunculus or substance, then it is natural to conclude that it is just the episode or stream of consciousness itself. This will mean that consciousness has a reflexive or self-manifesting structure (see Williford 2006, 2011a, 2011b, 2015 and Williford, Rudrauf, and Landini 2012 for more on how self-manifestation allows us to make sense of subjective character without homuncularism).

Questions 2 & 3: Representational Theories of Consciousness & Replicas

Artem asks me to say a bit about Self-Representationalism (see Kriegel & Williford 2006). This is the view that what separates conscious mental states from non-conscious ones is that the former represent themselves in some way, in addition to whatever else they might represent. Self-Representationalism (also sometimes called Same-Order Representationalism) is often taken to be an odd species of Higher-Order Representationalism. According to Higher-Order Representationalism, the conscious mental states are those that are the object of some other mental state (a thought or a perception, depending on the theory).

Higher-Order Representationalism is subject to a number of objections. The most damning objection is the “Targetless HOT” objection (see, e.g., Mandik 2009). Since the higher-order states just represent the lower-order state, it is conceivable that the lower-order state might fail to exist, just as one can represent Santa Claus or Baba Yaga. In such a case would there be any consciousness? If you say yes, then the lower-order state is not required in the explanation of consciousness — consciousness resides, so to speak, in the higher-order state alone, and the idea of reducing consciousness to a certain combination of representational states (which are themselves unconscious) falls apart. If you say no, then one is left wondering why exactly the existence of the lower-order state makes any difference. My representation of Santa Claus doesn’t do anything to Santa Claus — even if it turns out that he does in fact exist. Why should representing a mental state “do” anything to it, modify it in some way? In this case, why should it “make” it conscious? There seems to be nothing to motivate this claim. This particular horn of the dilemma is also sometimes presented as a separate problem for Higher-Order Representation theories — the so-called Problem of the Rock. The name comes from a particular example due to Leopold Stubenberg (1998): Thinking about or perceiving a rock does not make the rock conscious; why should it be any different with mental states?

Variant versions of the Higher-Order theory have been developed to escape from these objections. One version, due to Rocco Gennaro (2012), has it that the lower-order and higher-order mental states are really just parts of one integrated whole state. The integration is such that the parts cannot exist without each other. This might solve the Targetless HOT problem and the Problem of the Rock at one go — though consciousness would then have as much to do with “integration” as it would with higher-order representation. But one can wonder if the appeal to integration here is not just an ad hoc move. Of course, if one assigns a very high probability to the Higher-Order paradigm, one will assign a high probability to this integration thesis (or perhaps some other thesis that allows one to escape the objections). And it is true that an appeal to integration is attractive, since, no doubt, some sort of integration is surely indispensible to any account of consciousness. However, the claim that it is a matter of integrating higher-order and lower-order representational parts is entirely driven by the commitment to the paradigm. So while this appeal does allow one to escape from the objections, it certainly does not thereby constitute further evidence for the account. We would need, rather, some independent neuroscientific evidence that the integration of higher-order and lower-order representations indeed occurs and correlates with consciousness (I don’t deny that there may be some such evidence, see, e.g., Kriegel 2009, Gennaro 2012).

Self-Representationalism, as I define it, is the view that, so to speak, the higher-order mental state and the lower-order mental state are identical. The state representing the world also represents itself (see Williford 2006, 2009, and Kriegel 2003). Evidently, this view does not face the Targetless HOT problem, since if it is true that it represents itself, it is true that it exists (barring, some sort of Meinongianism, of course). It does, however, face other problems plaguing the Higher-Order family of theories. There is, for example, the “Division of Phenomenal Labor” problem (see Neander 1998). If, say, the lower-order representation is of a (seen) green ball, and the higher-order representation is of oneself seeing a red ball, what color does the ball consciously seem to have? All answers seem equally unsatisfactory. Self-Representationalism does not escape from this problem, in my view anyway (see Wiesberg 2008).

I think, however, that Self-Representationalism better corresponds to the phenomenology of consciousness. Consciousness, as scores of philosophers, Ancient and Modern, Western and Eastern, have maintained, is self-manifesting in some way — is aware of itself, is “reflexively” aware (see, e.g., Caston 2002, Frank 2004, Thiel 2011, Williams 2000, Yao 2012, Zahavi 1999). Self-Representationalism is as close as one can get to this intuition within the framework of Representationalism. However, that very framework, which is common to First-Order and Higher-Order theories, is inadequate, in my view. (For a paradigmatic First-Order Representationalist view, see Tye 2002. For classic papers on the Higher-Order Representationalist (specifically Higher-Order Thought) approach, see Rosenthal 2006).

According to the representationalist framework, we first develop a theory of representation that is naturalistically acceptable. Of course, we want any such theory to be no more ontologically promiscuous than contemporary science. But above all, the theory must not take intentionality or representation to be an irreducible feature of reality. If we end up holding views like Descartes, Spinoza, Leibniz, or C.S. Peirce, views that, in one way or another, takes intentionality to be bedrock (whether in a dualist or panpsychist way), then we will have failed. The idea then is to take a naturalistic theory of representation and build a theory of consciousness out of it. One does this by finding the “extra ingredient” that separates unconscious mental states from conscious ones. First-Order representationalists typically claim that it is non-conceptual content, on the one hand, and “poise” (or “global availability”) on the other that will distinguish the conscious mental states. Higher-Order representationalists, as noted, claim that it is being represented by a higher-order thought or perception. Self-representationalists claim that the reflexive representations are the conscious ones.

Any of these views can be wedded to any naturalistic theory of representation (except that self-representationalism cannot be wedded to a simple causal theory, but such theories are very implausible in any case). The most interesting naturalistic theory of representation, on my view anyway, is Millikan’s teleosemantics (see Ryder, Kingsbury, and Williford (eds.) 2012). It has some very nice features. In particular, it comes closest to solving the “disjunction problem” and other problems about the determinacy of content.

However, there is a serious problem plaguing any representationalist theory of consciousness wedded to teleosemantics. On that theory, the fact that some entity (vehicle) has the content it does is a contingent matter. Take an atom-for-atom replica of yourself and give it a non-evolutionary origin, or even just a different evolutionary history from the one it actually has. That replica could have very different representational contents or even no representational contents at all, if, for example, it came about entirely by accident (this latter is the so-called “Swamp Man” scenario). Never mind that these scenarios might be utterly improbable, the point is just that the theory (teleosemantics-cum-representationalism about consciousness) allows that this is possible. This means that some replica of yours might not even be conscious, so long as consciousness depends, in part, on having a certain type of representational content. So if you think that it is impossible for an atom-for-atom replica of yourself (whom we assume to be conscious) to be unconscious, then you will have to reject this sort of theory. I do think that is metaphysically impossible. So I reject these theories.

It does not help much to draw a wide vs. narrow content distinction here. We can accept that the atom-for-atom replica of you has different wide-contents, if, for example, we want to accept a Kripke-Putnam account of natural kind terms. But the narrow content will be the same. And that, for understanding consciousness, is where the action is. So much the worse then for theories of consciousness that rest on accounts of representation like Millikan’s, accounts that basically entirely dispense with the notion of narrow content. This might make one want to transition to a representationalism that is based entirely on a notion of narrow content or something close enough.
I take it that a self-representationalism based on a notion of “phenomenal intentionality” would fit such a description. “Phenomenal Intentionality” is basically a kind of representation that gets its content determinacy from its phenomenality (see Kreigel 2011 and Kriegel (ed.) 2013). And the notion also is more in accordance with intentionality as conceived of by Brentano, Husserl, and the Phenomenological tradition. To them, contemporary naturalistic representational theories of consciousness would seem like non-starters if not arrant nonsense. For them, the paradigmatic type of intentionality is conscious intentionality. And, indeed, one could take Brentano’s account of consciousness to be a type of self-representationalism based on a “phenomenal intentionality” view of representation. On that view, an episode of consciousness has itself as its “secondary object” and whatever else it is a consciousness of as its “primary object”. Though many have rejected Brentano’s view as an implausible “reflection theory”, it is important to understand that his use of “object” here is totally minimalistic. He is not saying that the episode of consciousness apprehends itself the way it apprehends a physical object (with as yet unseen profiles). He is merely say, and correctly in my view, that an episode of consciousness appears to itself. Other things appear as well. While different things appear in different ways, they all, generically speaking, appear to consciousness.

To me, this sort of view is really not distinguishable from the view that consciousness is always self-acquainted (which I defend in Williford 2015). Brentano did not enshrine the quasi-Russellian distinction between acquaintance and representation in his vocabulary. Husserl does effectively make the distinction. For him it comes down to an issue of transcendence vs. immanence — this maps onto the representation (description) vs. acquaintance distinction. Consciousness directly experiences itself, and its sensory hyle; and these are immanent to consciousness (see Williford 2013). Objects of (as we would say) representation are transcendent or “transphenomenal”. They transcend any particular experience of them; and no object of representation gives itself all at once — rather, we can always multiply profiles on any such object, turn it over, view it in a different light, etc. Moreover, in principle, any transcendent object may fail to exist — this is something characteristic of representation. By contrast, anything experienced immanently cannot fail to exist.

I would rather just say that all episodes of consciousness are self-manifesting or self-acquainted. Further, I am happy to accept the idea that there exists something like sensory hyle with which we are directly acquainted (a thesis I cannot defend here, see Williford 2013, 2015). I also hold that if you are acquainted with X, then X exists. This differentiates acquaintance from representation. Again, I cannot infer from the fact that I think of Baba Yaga that there exists some object I am thinking of. Baba Yaga does not exist — well, I hope she doesn’t! If you think she does, just think of the round square in your living room instead. All this is to say that it may well be the case that a self-representationalism wedded to a phenomenal intentionality view of representation might not be that different from a self-acquaintance theory of the “reflexive” structure of consciousness. I hold the latter. This immediately raises the question: What is acquaintance? I say a little about this in my answer to Artem, but it will make more sense if I put my answer at the end of my comments on his fourth question.

Question 4: Phenomenology — Who cares?

Artem asks me about why theorists of consciousness should be interested in phenomenology (including, I was adding in my mind, the Phenomenological tradition that includes Brentano (as precursor), Husserl, Scheler, Stein, Heidegger, Sartre, Merleau-Ponty, and many others). For simplicity’s sake, I won’t make a careful distinction here between the phenomenology of consciousness and disciplined introspection. One can, of course, engage in the phenomenology of practically anything. For example, one might be interested in the phenomenology of the animate vs. inanimate worlds. One would then look for something like general laws that animate things vs. inanimate things seem to obey in their manner of appearance to us. This is not what is ordinarily meant by “introspection”, but it does seem that to the extent that the “phenomenology of X” is at least in part about how X appears in/to consciousness, there will be an introspective component. One will, that is, be reflecting on how the object appears to/in one’s consciousness, rather than merely engaged with the object pre-reflectively.

The phenomenology of consciousness, then, will be “introspective” in two ways. In addition to the way in which all phenomenological efforts are “introspective”, in this case the very subject matter of the effort is consciousness itself. Hence, one will be paying attention to and reflecting on the manner in which consciousness appears to consciousness, that is, how it appears to itself. In principle, other attempts to describe the structures of consciousness that have not gone under the label of “phenomenology” (e.g., the introspective psychology of James, Buddhist psychology) are basically similar in attitude and method to classical Phenomenology, though certainly with much less methodological obsessiveness than that displayed by Husserl.

Why not just throw phenomenology (and introspective methods generally) out the window? If we do, I answer, the problem of consciousness will simply vanish. All we will be left with is the need to offer a theory of our verbal behavior involving words like “consciousness”, “experience”, etc., or our judgments involving the concept of consciousness. The theory must not, of course, take our descriptions to be accurate or claim that the major terms or concepts figuring in our judgments have real referents or instances. Instead, it would have the same structure that error theories of moral judgment have. According to such theories, we are never to explain our use of words (and concepts) like “moral”, “immoral”, “[morally] good”, “[morally] ought” in terms of a real relation we bear to real instances of moral goodness or real obligations. Whatever one might think of an error theory or moral judgment, I think most people would regard an analogous theory of consciousness as truly absurd. I, for one, can’t shake the “Cartesian” intuition that I am a conscious being and am consciously aware of the world, my body, and myself. I get this “Cartesian” intuition from introspection (or phenomenology). I surely can’t throw the intuition out, as much as I might like to some days.

One might nevertheless properly ask about the scope and limits of phenomenological methods. What can we know about consciousness via phenomenology? Can we have erroneous phenomenological intuitions? Can it tell us everything interesting there is to know about consciousness? My view is that phenomenology is not infallible. We can be mistaken especially in our general claims about putatively invariant features of experience. One might, for example, think that one must always experience oneself as distinct from, say, an object one is sitting on, standing on, leaning against, or floating in (e.g., a chair, a bed, a flotation tank). However, there is evidence from psychopharmacology, the study of mystical experience, as well as sensory-deprivation tank studies that one can indeed experience oneself as being not distinct, or not fully distinct, from one’s surroundings. This can range from merely feeling one’s body to be continuous with one’s “external milieu” to actually feeling oneself to be literally identical with some object one is resting on or against, or even an object that one merely sees. The study of abnormal and altered states of consciousness is thus very important for the overall project of phenomenology. It can allow us to see that some feature we thought to be an invariant feature of consciousness might not be after all.
The goal of phenomenology, as I see it anyway, ought to isolate and describe the invariant features of consciousness (see Williford 2005, 2006, and Williford, Rudrauf, and Landini 2012). This is our only guide to the essential features of consciousness, at least those that we have access to. I see the classical Phenomenological tradition as having great value for this project. One can interpret Husserl, Sartre, Merleau-Ponty and others as being engaged in this project as well, whatever other projects they were also pursuing. One can learn a great deal from reading them. One need never take what they say about the structure of consciousness as sacrosanct or immune to criticism. And it is facile to imagine that they did not realize that constructing an adequate phenomenology of consciousness was a difficult business.

As I see it, the phenomenology of consciousness codifies the data we want to explain. And we should proceed in this way — this is a version of so-called “neurophenomenology” (see Varela 1996): First, we isolate the invariant structures of consciousness. Second, we recast this structure at the appropriate level of abstraction as a mathematical model. Third, we seek to understand how this model could be implemented neurobiologically. Fourth, using the model and the implantation theory, we try to isolate the neural correlates of consciousness. This last step will involve testing a lot of predictions one can derive from the model and the implementation theory. If we succeeded in doing that, we will have solved the consciousness-brain problem to the extent that it is a scientific (as opposed to a metaphysical) problem. Moreover, I would add, we will have reduced the metaphysical component of the problem to the general metaphysical problem of how forms or structures are concretely implemented in the physical world. That is an ancient and general problem faced by all realistically interpreted scientific theories; it is not specific to the philosophy of mind. Reducing the consciousness-brain problem to that problem would, I believe, count as great intellectual progress, even if the latter problem is insoluble (see Williford 2015).

What are the invariant features of consciousness as revealed by the phenomenology? I would propose that the following are the serious candidates: intentionality, synchronic unity, diachronic unity (temporality), spatial projectivity, and pre-reflective self-manifestation (“reflexivity”) (see Williford 2005, 2006, and Williford, Rudrauf and Landini 2012). Theorists of consciousness typically take one of these features and develop a framework around it. I think we need to be looking at frameworks that can integrate them all. And even if it should turn out that they are not all essential, it is still the case that they characterize normal human consciousness as we know it. That fact alone is important enough to indicate that pursuit of these models is not a waste of time, even if one’s favorite feature turns out, after a successful empirical theory of consciousness is found, not to be essential.

I noted earlier that I maintain that consciousness is always self-manifesting, always acquainted with itself. If one holds this sort of view, one will be attracted to mathematical models that allow one to capture this “reflexivity” in one way or another. There are several such modeling strategies that I mention in the interview: Douglas Hofstadter’s Gödel-inspired model (Hofstadter 1979, 2007), models based on hypersets (sets that can be in their own membership chains) (Miranker and Zuckerman 2009, cf. also Varela 1979 and Khromov 2001), and projective geometrical models. I have myself developed just a bare sketch of a hyperset model (Williford 2006). And I am currently collaborating on projective geometrical models with the neuroscientist David Rudrauf, the mathematician Daniel Bennequin, and the logician Gregory Landini. The projective geometrical model is the brainchild of David Rudrauf (a preliminary sketch containing several oversimplifications can be found in Williford, Rudrauf, and Landini 2012).

Now, what about the “acquaintance relation” and “self-acquaintance”? What is acquaintance? On my view, we cannot say anything more informative about it by relying on phenomenology (see Williford 2015). It is simply a phenomenon that involves a certain structure. To know more about it, we need to develop the mathematical models just discussed and find the neural correlates of consciousness. Then we will have a rich mathematical (and probably computational) picture of the “deep” structure of acquaintance and a rich understanding of how it is implemented in at least one sort of real, concrete biological system. Explaining that implementation and understanding the structure at the mathematical level is, in my view, all there is to solving the “consciousness-brain” problem. This is better than a “brute” identification because one will be able to see how the neural correlates of consciousness have the same structure consciousness seems to have at the phenomenological level. If it turns out that all the structure consciousness seems to have is entirely illusory, then there can be no such solution. If it turns out that we have no way of getting enough significantly phenomenology-independent evidence that this structure (as described by the mathematical and implementation theories) is in fact implemented in the brain at the right level, this solution will fail for a different reason, even if its strategy is, in principle, correct. But if we never try, we will surely never know.
To sum up: We want to steer between the extremes of rejecting all introspective data as illusory or entirely untrustworthy, on the one hand, and imagining that introspection is infallible (or easy) on the other. Contemporary theorists of consciousness can learn much from reading the classical Phenomenologists, but they can safely dispense with the “anti-neuro” and ant-naturalistic stance one finds in many of them (Husserl and Sartre, in particular). This brings us right up to the last question about the Hard Problem.

Question 5: The Hard Problem

Artem asks me about my views on the Hard Problem and whether I think it is a real problem. My answer is “yes and no”. Yes, I think it is a real problem in the sense that it is very difficult to understand how it is that some collection of neurons and neurobiological processes, no matter how well integrated and complex they are, can literally constitute consciousness. How can consciousness just be this or that process running on a brain (or a computer for that matter)? It is totally mind- boggling. However, no, I do not think it is an insoluble problem in the final analysis.

The idea that the Hard Problem is truly insoluble is often supported by modal and conceivability arguments, Chalmers’ Zombie thought experiment being perhaps the most famous (see., e.g., Chalmers 1995, 1997). According to that thought experiment, zombies are conceivable. Zombies are beings just like us in every physical and functional respect except that they are not conscious. They say they are, of course! Well, the story goes, if they are conceivable, they are metaphysically possible. And if they are metaphysically possible, then the property of being conscious cannot be identical with any physical of functional property their brains. Ergo, some sort of dualism or perhaps pan-proto-psychism must be true.

My response amounts to an attack on the claim that conceivability in this case implies metaphysical possibility. If consciousness is, in its very nature, identical with the implementation or realization of a certain structure (perhaps computational) on some sort of hardware (or range of hardwares), then, supposing my brain implements the structure, my supposed zombie twin will be conscious after all. It would, of course, be rather question begging for the neo-dualist to tell me that it is just not conceivable that consciousness could be identical with such an implemented structure because, after all, zombies are conceivable. I could just reply that they are not conceivable because it is conceivable that consciousness is identical with said structure, and therefore possible; and therefore actual; and we would find ourselves at a perfect impasse. I want to say, rather, that both situations are conceivable. This means that I have to deny that they are both metaphysically possible (or hold a very strange view about the brute, transworld complete resemblance between non-identical properties, an oddity we cannot go into here). This will mean that I deny that conceivability is a guide to metaphysical possibility.

This does not mean that I have to reject conceivability as a guide to metaphysical possibility in all domains. But the domains have to meet some special demands. In particular, we need to be in a position to know that we have isolated the essential features of the sort of object in question, that we have found the laws of the “regional ontology” to which the sort of object belongs. We can only know that we have done this if we have very good reason to believe that all of the essential properties of the “region” are accessible to us. In physics this is achieved by a combination of theory and experiment. Only in reference to a fairly mature physics can we say, with any confidence, that some types of physical event (e.g., atomic explosions) are possible and others (e.g., perpetual motion machines) are not. But, one may say, this is mere “physical” possibility.

In mathematics we codify our delineation of the essential laws of a domain by axiomatization. And when axiomatization fails to decide every well-formulated question about our domain (e.g., the Continuum Hypothesis (CH) in Set Theory), we are forced to seek further axioms. Typically those further axioms are, for most people anyway, quite a bit less intuitive than the ones we began with. To the extent that one is nonetheless a realist about the domain (or region) in question, one will think that there is a definite fact of the matter about whether or not the new axioms are correct. If, for example, one’s new set theoretical axiom decides the CH one way, and another’s decides it another, it is not the case that both axioms can be correct — if one is a realist about sets. They can’t both describe the set theoretical universe correctly. And if the facts about the universe of sets are not metaphysically contingent, then one of the new axioms will be (metaphysically) necessarily false. However, we may never be able to generate a contradiction from it. Indeed, even an “ideal” conceiver may never be able to generate a contradiction from the axioms. This is so even if such a conceiver could just somehow “see” by direct inspection of the universe of sets that one of the competing axioms is correct. As long as we are imagining godlike cognitive powers, we might as well point out that the ideal conceiver might, likewise, be able to tell that the different axiom systems are both consistent. If this is so, then we may say that both possibilities are conceivable, even though it is not the case that both are metaphysically possible. And given our far less than godlike cognitive abilities, we must all admit that these competing possibilities are conceivable for all we can tell. Appeal to godlike or ideal conceivers in this context does nothing, unless one secretly imagines than one is an ideal conceiver.

If this sort of situation holds in the rarefied domain of pure set theory, where the objects in question are, at least to begin with, quite clearly defined, how much less should we be surprised if it holds in something concrete, complicated, and difficult to circumscribe adequately like consciousness? If we cannot even be sure that we have access to all the essential laws of the set theoretical universe, how much less can we be confident that we have isolated all the essential properties of consciousness? And if, in the case of set theory, realistically interpreted, we should not presume that conceivability implies metaphysical possibility, how much less should we make such a presumption in the case of consciousness?

In fact, there is really no good reason to think that we know what all the essential properties of consciousness are. Moreover, there are plenty of reasons for thinking that it would be entirely surprising if we did. Why would we need to have access to the metaphysical nature of consciousness or its categorial status? (See Williford 2004, 2007) Having access to is intentional content is a much more useful thing.

But now if we are in the habit of taking the absence of access to a property of consciousness as indicative of the absence of the property itself (and we are), then we will think that all kinds of things are possible with respect to consciousness that may not be at all. Suppose consciousness is a certain type of brain process. That does not appear upon introspection. To the extent that our concept of consciousness is determined by what does appear upon introspection, we will be able to conceive of consciousness in the absence of brain processes. (This will also help us to conceive of those brain processes in the absence of consciousness — we would find this much more difficult or even impossible if we had direct phenomenological evidence of the internal connection (identity) of those processes and consciousness.) It is much harder to conceive of consciousness without unity or consciousness without temporality or intentionality (though all of these have been attempted), since these features seem to be invariant. Brain and complicated computational processes do not appear on the phenomenological map at all. Hence, we find it so easy to imagine consciousness without them; and them without consciousness (see Williford 2007).

But if such essential properties are not given phenomenologically, then we should not trust the inference from conceivability to possibility in this domain. We do not know the “laws of the region”. We cannot be sure at all that our concept of consciousness tracks all of the essential features of consciousness. And if that is the case, we cannot accept inferences from conceivability to metaphysical possibility here. Finally, if it is the case that there are essential properties of consciousness that are hidden from the phenomenological viewpoint, this will mean that any correct theory of consciousness will have a certain irreducible a posteriori element to it. This will mean that we will have to find out some important facts about what consciousness essentially is by empirical investigation. And we will just have to get used to that. We got used to it for water and H2O; we’ll get used to it for consciousness and the fully fleshed out theoretico-empirical description of the physical (or implemented computational) process that is consciousness; formulating and verifying this description is the goal of the neurophenomenological project described above. With any luck, we will achieve it.


At some point in the interview I say something like “…to which it belongs to” or something unseemly like that. And I think I also say (in Russian) “M. GU” instead of “M. G. U.” (the Russian abbreviation for Moscow State University, Московский государственный университет). I apologize to English speakers, on the one hand, and Russian speakers, on the other, for these infelicities.

References and Relevant Literature

I have listed my own works, including co-authored and co-edited ones, first, followed by other works referred to above (in standard, alphabetical order). It should be noted that the listing of other works is idiosyncratic and by no means complete. A few works are listed more than once, since they fit into multiple sections. Since this was an interview primarily about my own views and work, I felt that this was not too ugly and narcissistic a thing to do. If I was wrong about that, again, I apologize.

Berkeley and Hume

Williford, K. (2003a) “Berkeley’s Theory of Operative Language in the Manuscript Introduction” The British Journal for the History of Philosophy, 11, 2, 271-301.

Williford, K. (2003b) “Demea’s a priori Theistic Proof” Hume Studies, 29, 1, 99-123.

Williford, K. & R. Jakapi (2009) “Berkeley’s Theory of Meaning in Alciphron VII”, The British Journal for the History of Philosophy, 17, 1, 99-118.

Butchvarov, P. (1959) “The Self and Perceptions; A Study in Humean Philosophy.” The Philosophical Quarterly, 9, 35, 97-115.

Self-Representationalism, Self-Manifestation, and Subjective Character

Williford, K. (2006) “The Self-Representational Structure of Consciousness” in Self-Representational Approaches to Consciousness, Uriah Kriegel and Kenneth Williford (eds.), MIT Press, pp. 111-142.

Williford, K. (2009) “Self-Representational Theories of Consciousness” In The Oxford Companion to Consciousness, Patrick Wilken, Tim Bayne, and Axel Cleeremans (eds.), pp. 583-585.

Williford, K. (2011a) “Auto-representacionalismo y los problemas de la subjetividad”. In Cuadernos de Epistemología, número 5. Reflexiones en torno a la filosofía de la ciencia y la epistemología, J. Aguirre y L. Jaramillo (eds.). Popayán: Universidad del Cauca, pp. 39-51.

Williford, K. (2011b) “Pre-Reflective Self-Consciousness and the Autobiographical Ego” in Reading Sartre: On Phenomenology and Existentialism edited by Jonathan Webber, Routledge, pp. 195-210.

Williford, K. (2015) “Representationalisms, Subjective Character, and Self-Acquaintance” in Thomas Metzinger & Jennifer M. Windt (Eds). Open MIND: 39(T). Frankfurt am Main: MIND Group.


Williford, K., Rudrauf, D., & Landini, G. (2012) “The Paradoxes of Subjectivity and the Projective Structure of Consciousness” in Consciousness and Subjectivity, S. Miguens and G. Preyer (eds.), Ontos Verlag, pp. 321-353.

Kriegel, U., & Williford, K. (eds.). (2006) Self-Representational Approaches to Consciousness. Cambridge, MA: MIT Press.

Philippi, C., Feinstein, J.S., Khalsa, S.S., Damasio, A., Tranel, D., Landini, G., Williford. K., Rudrauf, D. (2012) “Preserved self-awareness following extensive bilateral Brain damage” PLoS ONE, 7, 8 http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0038413

Caston, V. (2002) “Aristotle on Consciousness” Mind, 111(444), 751-815.

Frank, M. (2004) “Fragments of a History of the Theory of Self-Consciousness from Kant to Kierkegaard” Critical Horizons: A Journal of Philosophy & Social Theory, 5(1), 53-136.

Kriegel, U. (2003) “Consciousness as Intransitive Self-Consciousness: Two Views and an Argument” Canadian Journal of Philosophy, 33, 1, 103-132.

Kriegel, U. (2009) Subjective Consciousness: A Self-Representational Theory. Oxford University Press.

Thiel, U. (2011) The Early Modern Subject: Self-consciousness and Personal Identity from Descartes to Hume. Oxford: Oxford University Press.

Williams, P. (2000) The Reflexive Nature of Awareness: A Tibetan Madhyamaka Defence. New Delhi: Motilal Banarsidass Publishers.

Yao, Z. (2012) The Buddhist Theory of Self-Cognition. London: Routledge.

Zahavi, D. (1999) Self-awareness and Alterity: A Phenomenological Investigation. Evanston: Northwestern University Press.

Representationalisms and Millikan’s Teleosemantics

Williford, K. (2015) “Representationalisms, Subjective Character, and Self-Acquaintance” in Thomas Metzinger & Jennifer M. Windt (Eds). Open MIND: 39(T). Frankfurt am Main: MIND Group.


Ryder, D., Kingsbury, J., & Williford, K. (eds.). (2012) Millikan and her Critics. John Wiley & Sons.

Gennaro, R. J. (2012) The Consciousness Paradox: Consciousness, Concepts, and Higher-Order Thoughts. MIT Press.

Kriegel, U. (2009) Subjective Consciousness: A Self-Representational Theory. Oxford University Press.

Kriegel, U. (2011) The Sources of Intentionality. Oxford University Press.

Kriegel, U. (ed.). (2013) Phenomenal Intentionality. Oxford University Press.

Mandik, P. (2009) “Beware of the Unicorn: Consciousness as Being Represented and Other Things that don’t Exist.” Journal of Consciousness Studies, 16, 1, 5-36.

Millikan, R. G. (1984) Language, Thought, and Other Biological Categories: New Foundations for Realism. MIT press.

Millikan, R. G. (1995) White Queen Psychology and Other Essays for Alice. MIT Press.

Neander, K. (1998) “The Division of Phenomenal Labor: A problem for Representational Theories of Consciousness.” Nous, 32, S12, 411-434.

Rosenthal, D.M. (2006) Consciousness and Mind. Oxford University Press.

Stubenberg, L. (1998) Consciousness and Qualia. John Benjamins Publishing.

Tye, M. (2002). Consciousness, Color, and Content. MIT Press.

Weisberg, J. (2008) “Same Old, Same Old: The Same-Order Representation Theory of Consciousness and the Division of Phenomenal Labor.” Synthese, 160, 2, 161-181.

Phenomenology, Introspection, and the Hard Problem

Williford, K. (2004) “Moore, the Diaphanousness of Consciousness, and Physicalism” Metaphysica, 5, 2, 133-153

Williford, K. (2005) “The Intentionality of Consciousness and Consciousness of Intentionality” in Intentionality: Past and Future, Gabor Forrai and Gyorgy Kampis (eds.), Rodopi, pp. 143-156.

Williford, K. (2007) “The Logic of Phenomenal Transparency” Soochow Journal of Philosophical Studies, 16, 181-195 (Special Issue)

Williford, K. (2013) “Husserl’s Hyletic Data and Phenomenal Consciousness” Phenomenology and the Cognitive Sciences, 12, 501-519.

Williford, K. (forthcoming) “Degrees of Self-Presence: Rehabilitating Sartre’s Accounts of Reflection and Pre-Reflective Self-Consciousness” in G. Preyer, S. Miguens & C. Morando (eds.) Pre-Reflective Consciousness: Sartre and Contemporary Philosophy of Mind. Routledge.

Sartre, J.-P. (2012) The Imagination, translated and introduced by K. Williford and D. Rudrauf, Routledge.

Chalmers, D. J. (1995) “Facing up to the Problem of Consciousness.” Journal of Consciousness Studies, 2, 3, 200-219.

Chalmers, D. J. (1997). The Conscious Mind: In Search of a Fundamental Theory. Oxford University Press.

Dennett, D.C. (1991) Consciousness Explained. Little, Brown & Co.

Varela, F. J. (1996) “Neurophenomenology: A Methodological Remedy for the Hard Problem” Journal of Consciousness Studies, 3, 4, 330-349.

Mathematical Models of “Reflexive” Consciousness

Williford, K. (2006) “The Self-Representational Structure of Consciousness” in Self-Representational Approaches to Consciousness, Uriah Kriegel and Kenneth Williford (eds.), MIT Press, pp. 111-142.

Williford, K, (2011) Review of I am a Strange Loop, by Douglas Hofstadter. Philosophical Psychology, 24, 6, 861-865

Williford, K., Rudrauf, D., & Landini, G. (2012) “The Paradoxes of Subjectivity and the Projective Structure of Consciousness” in Consciousness and Subjectivity, S. Miguens and G. Preyer (eds.), Ontos Verlag, pp. 321-353.

Hofstadter, D. R. (1979/1999) Gödel, Escher, Bach. New York: Basic Books.

Hofstadter, D.R. (2007) I Am a Strange Loop. New York: Basic Books.

Khromov, A. G. (2001) “Logical Self-reference as a Model for Conscious Experience.” Journal of mathematical psychology, 45, 5, 720-731.

Miranker, W. L., & Zuckerman, G. J. (2009) “Mathematical Foundations of Consciousness.” Journal of Applied Logic, 7, 4, 421-440.

Varela, F. J. (1979) Principles of Biological Autonomy. New York: Elsevier North Holland.

Ещё на эту тему

Научные семинары

Сознание, мозг и свобода воли. Научный семинар Центра исследования сознания

Семинар был посвящен книге известного оксфордского философа Ричарда Суинберна «Сознание, мозг и свобода воли». Суинберн по современным меркам занимает непопулярную позицию, отстаивая субстанциальный д...
Май 2018

Конференция о проблеме тождества личности. Записи докладов

Проблема тождества личности является одной из самых обсуждаемых в современной философии. Она включает в себя вопросы о Я, личности, ее тождестве во времени и единстве. 17 мая 2018 года на философском ...
Май 2018