Published in Mind, 108 (1999), pp. 705-725.
Phenomenal Consciousness: the Explanatory Gap as Cognitive Illusion
Woody Allen once remarked, “What a wonderful thing, to be conscious! I wonder what the people of New Jersey do?” However things are in New Jersey, what is wonderful about consciousness or at least most immediately striking is the subjective or phenomenal character of such states as the visual experience of bright red, the feeling of elation, the sensation of being tickled. Our grasp of what it is like to undergo these and other experiential states is supplied to us by introspection. We also have an admittedly incomplete grasp of what goes on objectively in the brain and the body. But there is, it seems, a vast chasm between the two. Presented with the current physical and functional story of the objective changes that occur when such-and-such subjective feelings are experienced, we have the strong sense that the former does not fully explain the latter, that the phenomenology has been left out. We naturally ask: What is so special about those physical or functional goings-on? Why do they feel like that? Indeed, why do they feel any way at all?
Compare this case with that of solidity or digestion, say. Once one learns that in solid things, the molecules are not free to move around as they are in liquids, one immediately grasps that solid things do not pour easily, that they tend to retain their shape and volume. Having been told the physical story, to ask: “Yes, but why are things with molecules that are fixed in place solid? Why shouldn’t such things not be solid?” is to show a conceptual confusion. One who responds in this way simply does not understood the ordinary notion of solidity. What it is for something to be solid is for it to be disposed to retain its shape and volume (roughly). Once the molecules are fixed, the shape and volume are fixed and thereby automatically the disposition to retain shape and volume, i.e., the solidity.
Similar points apply in the case of digestion. Upon learning that there are enzymes in the alimentary canals of human-beings that break down food and convert it into energy, only a failure to grasp that the word ‘digestion’ means (roughly) internal process whose function is to convert food into energy could lead one to ask: “Why does the action of these enzymes in humans generate digestion? Why shouldn’t the enzymes turn food into energy in the absence of digestion?”
In the case of phenomenal consciousness, however, the corresponding questions remain even for those who understand full well the relevant phenomenal terms and who know the underlying physical and functional story. One who has a complete understanding of the term ‘pain’, for example, and who is fully apprised of the physical facts, as we now know them, can still coherently ask why so-and-so brain processes or functional states feel the way pains do or why these processes feel any way at all. In this case, it seems that as far as our understanding goes something important is missing. Herein lies the famous “explanatory gap” for consciousness.1
Some say that the explanatory gap is unbridgeable and that the proper conclusion to draw from it is that there is a corresponding gap in the world. Experiences and feelings have irreducibly subjective, non-physical qualities over and above whatever physical qualities they have. The physical (and functional) story is incomplete (QUALOPHILIA: Jackson 1982, 1993, Chalmers 1996). Others take essentially the same position on the gap while urging that being objective is not a necessary condition of being physical. Thus, it is claimed, there is nothing in the gap that detracts from a purely physicalist view of experiences and feelings. The introspectible, phenomenal qualities of experiences and feelings are indeed irreducibly subjective, but this is compatible with their being physical (CLOSET QUALOPHILIA: Searle 1992). Others hold that the explanatory gap may one day be bridged but we currently lack the concepts to bring the subjective and objective perspectives together. On this view, it may turn out that phenomenal states are physical, but we currently have no clear conception as to how they could be (PHYSICALISM, FINGERS CROSSED, Nagel 1974). Still others adamantly insist that experiences and feelings are as much a part of the physical, natural world as life, photosynthesis, DNA, or lightning. It is just that with the concepts we have and the concepts we are capable of forming, we are cognitively closed to a full, bridging explanation by the very structure of our minds. There is such an explanation, but it is necessarily beyond our cognitive grasp (PHYSICALISM, DEEPLY PESSIMISTIC: McGinn 1991).2
I reject all of these positions. What they have in common is the idea that if experiences are indeed fully physical, in the traditional sense of the term ‘physical’ (opposed by Searle), then an explanation is needed, but has not yet been found, for why the relevant physical states and qualities feel on the inside as they do. Where the proponents of these positions differ is over whether to affirm the antecedent of this conditional and thus to accept that the phenomenology of the appropriate physical entities needs explaining but is, as yet, unexplained or to deny that experiences are fully physical in the traditional sense, thereby removing the need for an explanation of the kind specified in the consequent. Jackson, Chalmers, and Searle do the latter; McGinn does the former, adding that the phenomenology will never be properly explained by us or by creatures like us. Nagel does not unequivocally fall into either camp, but overall he seems best taken to favor the former strategy (without the thesis of cognitive closure).
I deny the conditional. I accept that experiences are fully, robustly physical but I maintain that there is no explanatory gap posed by their phenomenology.3 The gap, I claim, is unreal. It is a cognitive illusion to which we only too easily fall prey. As such, it has no consequences for the nature of consciousness and physicalist or functionalist theories thereof. There is nothing in the alleged gap that should lead us to any bifurcation in the world between experiences and feelings on the one hand and physical or functional phenomena on the other. There aren’t two sorts of natural phenomena: the irreducibly subjective and the objective. The so-called “explanatory gap” derives largely from a failure to recognize the special features of phenomenal concepts. These concepts, I maintain, have a character that not only explains why we have the intuition that something important is left out by the physical (and/or functional) story but also explains why this intuition is not to be trusted.
My discussion is divided into four parts. I begin by making some remarks about the perspectival subjectivity of phenomenal states and I explain why I reject the familiar view that phenomenal concepts are simply indexical concepts that are applied to phenomenal states (see, e.g., Horgan 1984, Loar 1990, Rey 1991). In Section II, I go on to elaborate an account of phenomenal concepts. The nature of the alleged explanatory gap is addressed in Section III. Remaining worries are addressed in Section IV.
I. Perspectival Subjectivity
One generally agreed upon fact about phenomenal states is that they are perspectivally subjective. Consider the case of pain. It seems highly plausible to suppose that fully comprehending the character of the feeling of pain requires knowing what it is like to feel pain. And knowing what it is like to feel pain requires one to have a certain experiential point of view or perspective, namely the one conferred upon one by being the subject of pain. This is why a person born without the capacity to feel pain and kept alive in a very carefully controlled environment could never come to know what it is like to experience pain. Such a person could never herself adopt the relevant perspective. And lacking that perspective, she could never comprehend fully what that type of feeling was, no matter how much information was supplied about the firing patterns in the brains of people who are experiencing pain, the biochemical processes, the chemical changes, the disturbed bodily states.
Phenomenally conscious states, then, are perspectivally subjective in the following way: each phenomenal state, S, is such that fully comprehending S, as it is essentially in itself, requires adopting one particular point of view or perspective, namely that provided by undergoing S. The perspectival character of these states is, I believe, a reflection of the concepts that are deployed in thinking about or understanding these states from the inside. Without such an understanding, phenomenal states intuitively cannot be fully understood.
Consider again the case of pain. Fully understanding pain requires grasping how it feels, its distinctive phenomenal character, what it is like to undergo pain. That, in turn, requires applying to pain the concept that is typically applied when people introspect pain and pay attention to what it is like subjectively. This concept is a phenomenal concept. A person who lacks the phenomenal concept PAIN thereby is prevented from possessing the kind of understanding of pain provided by introspection. For such a person, there is a way of understanding pain that lies beyond his or her grasp. Consequently, such a person does not fully understand pain, as it is essentially in itself. And this is so whether or not the state of pain is itself physical.
My suggestion, then, is that the perspectival subjectivity of phenomenal states goes hand in hand with the perspectival character of phenomenal concepts, where phenomenal concepts are the concepts that are utilized when a person introspects his or her phenomenal state and forms a conception of what it is like for him or her at that time. Given the perspectival nature of indexical concepts, it may now be tempting to suppose that phenomenal concepts just are indexical concepts applied via introspection to phenomenal states.4 This temptation is one that a physicalist (or functionalist) should resist, however, even though it is certainly true that we do often conceive of our phenomenal states in a manner that brings to bear indexicals as well as phenomenal concepts.
For one thing, the perspective in the case of indexical concepts is very different from that which is relevant to phenomenal concepts. What is characteristic of indexical concepts is that they all involve an egocentric perspective. Each subject of psychological states, in thinking of something under an indexical concept, is thinking of it via a mode of presentation that bears an a priori connection to the first person concept I. For example, in thinking of the place I am in as here, I exercise the indexical concept here and this concept is such that I can know a priori that I am here.5 Similarly, in thinking of something as this object, I exercise the concept this and this concept is such that I can know a priori that this thing is the object I am attending to (if I am attending to anything at all). Likewise, in thinking of the time as now, I exercise the concept now and this concept is such that I can know a priori that I am here now. Each indexical concept, then, is a priori linked with the concept I and thereby each indexical concept incorporates a certain perspective, namely the very special, first-person perspective each person has on himself. In the case of phenomenal concepts, however, the relevant perspective is not the generic first-person one. Grasping the phenomenal concept PAIN, for example, requires more than having a first-person perspective on oneself. As noted earlier, the relevant perspective or point of view is that conferred upon one by one’s undergoing an experience of pain. Each phenomenal concept is thus tied to a particular experience-specific perspective occupied by the possessor of the concept. As the experiences vary, so too do the phenomenal concepts.
Another objection to identifying phenomenal concepts with indexical concepts is that if the phenomenal aspect of pain is physical (or functional), then an indexical conception of it can be formed from an external perspective by the person who is incapable of feeling pain. Such a person might think of the phenomenal character I experience on some given occasion when I feel pain as that quality. Clearly, this person would not have the same cognitive ‘fix’ on the quality as I have on the inside. Intuitively, her demonstrative conception of it would be very different from my phenomenal one. Were this not so, she would not be able to make a significant discovery about the way pain feels if a neurological operation subsequently permitted her to experience pain. This does not show, of course, that no demonstrative is at play in the phenomenal conception. But it does indicate that there is some additional concept being exercised that is not operative in the external conception, a general phenomenal concept.
What, then, is distinctive about our phenomenal concepts? What marks them out as special? It is to these questions that I now turn.
II. Phenomenal Concepts
Phenomenal concepts, I have suggested, are not third-person concepts; and neither are they just indexical concepts. They are conceptually irreducible. That is to say, no a priori analysis can be given of them in non-phenomenal terms. Phenomenal concepts, I have also suggested, are perspectival.
Consider, for example, the concept PAIN and suppose, for the sake of simplicity only, that this concept is purely phenomenal. The person who lacks the capacity to feel pain does not possess the phenomenal concept PAIN though she may acquire a non-phenomenal concept of pain. Intuitively, possessing the phenomenal concept requires knowing what it is like to experience pain. Likewise, in the case of the phenomenal concept RED, a concept that is exercised when one becomes aware via introspection of what it is like to experience red.6 Possession of the concept requires that one know what it is like to experience red.7 To appreciate the significance of these points to understanding phenomenal concepts, it is necessary to make some remarks upon what is involved in knowing what it is like.
David Lewis (1990) has claimed that knowing what it is like is a matter of having certain abilities. First, there is the ability to remember the experience in question. Suppose you smell a skunk for the first time, and you thereby learn what it is like to smell a skunk. Afterwards, you can remember the experience.8 Moreover, by remembering it, you can imaginatively recreate it. This will be the case, even if, as Lewis notes, you eventually forget the occasion on which you had the experience. By having the experience of smelling a skunk, then, you gain new abilities to remember and imagine.
Included within the ability to imagine is more than just the ability to imagine the experience you underwent earlier. After seeing something red, for example, and seeing something yellow, you are able to imagine something red with yellow spots, even if you have never seen anything red with yellow spots. By imagining certain situations you could not imagine before, you also gain the ability to predict with a fair degree of confidence what you would do were the situations to arise. For example, having seen the color purple, you can now imagine how you would likely react, if you were offered a purple shirt to wear.
Another important ability you gain is the ability to recognize the experience when it comes again directly and immediately on the basis of having the experience. Lewis says:
If you taste Vegemite on another day [your second encounter with it], you will probably know that you have met the taste once before. And if, while tasting Vegemite, you know that it is Vegemite that you are tasting, then you will be able to put the name to the experience if you have it again. (1990, p. 515)
These abilities — to remember, imagine, and recognize in an immediate and direct way — constitute knowing what it is like, on Lewis’ view. There is no claim that you could not possibly have these abilities without having the relevant experiences. After all, you might acquire them by some possible future neurophysiology or by magic. The point is that, given how the world actually works, lessons alone won’t do the trick, no matter how complicated they become.
Some philosophers contest Lewis’ claim that knowing what is it like is no more than the possession of certain abilities.9 And I myself reject Lewis’ thesis in an unqualified form.10 But intuitively, Lewis is surely right to draw a conceptual connection between knowing what it is like and the abilities specified above.11 Given that phenomenal concept possession itself conceptually requires knowing what it is like, there thus emerges an a priori connection between phenomenal concept possession and the Lewis abilities. This connection finds direct support in the view that our concept of a phenomenal concept is that of a concept which plays a characteristic functional role — a role whose a priori specification entails that any possessor of the concept have the pertinent phenomenal abilities.
Phenomenal concepts, I maintain, are conceptually irreducible concepts that function in the right sort of way. To possess the phenomenal concept RED, for example, is to possess a simple concept that has been acquired by undergoing experiences of red (barring neurosurgery to induce the state or a miracle) and that not only disposes one to form a visual image of red in response to a range of cognitive tasks pertaining to red but also is brought to bear in discriminating the experience of red from other color experiences in a direct and immediate manner via introspection. The functional role that the concept plays is what makes it perspectival. A person who is blind from birth or who is always restricted to an environment of things with achromatic colors cannot possess a concept with the requisite role and hence cannot possess the phenomenal concept RED.12 Let me now try to motivate this approach to phenomenal concepts further.
Consider first the following example of a phenomenal-physical identity claim:
The visual experience of red = brain state B.13
One reaction some philosophers have to claims of this sort is that they must be mistaken, since the the phenomenology isn’t captured by the right-hand side. From the present perspective, this reaction involves a sense/reference confusion. When we think of the referent of the designator on the left-hand side in a phenomenal way, we bring it under a concept that has a distinctive functional role. In reflecting on the identity claim and what is puzzling about it, the phenomenal concept we deploy is apt to trigger in us a visual image of red.14 In this event, if the identity is true, our brain actually goes into brain state B. But, of course, when we think of the referent of the designator on the right hand side as brain state B, nothing like that happens. Exercising the neurophysiological concept is not apt to trigger a visual image of red. It may then be tempting to infer that the right-hand side has left out the phenomenology of the left, that there is a huge gap that the physicalist has failed to close. This conclusion clearly does not follow, however. There is indeed a striking difference in the roles that the concepts play, in their functioning, but not (so far as is shown here) in their referents.15
Secondly, it seems to us that we have a direct and immediate cognitive access to certain qualities when we introspect our experiences and feelings16. These qualities — phenomenal qualities, as we might call them — are ones that intuitively we know in a specially intimate way. The intuition that there is a direct and immediate access to such qualities (under normal conditions) needs to be preserved. The proposal I have made does that in a natural and straightforward manner. In being cognitively aware of our own phenomenal states, we subsume those states under phenomenal concepts. These concepts are simple. They are also, in part, direct recognitional concepts. For it is part of their characteristic functional role, qua phenomenal concepts, that they enable us to discriminate phenomenal qualities and states directly on the basis of introspection. In having the phenomenal concept PAIN, for example, I have a simple way of classifying pain that enables me to recognize it via introspection without the use of any descriptive, reference-fixing intermediaries. Thus, it is guaranteed by the fact that the concept I am applying is phenomenal that I do not know introspectively that I am in pain by knowing something else connected to pain. My knowledge is direct and immediate. Introspection of the phenomenal character of pain causally triggers in me the application of the concept PAIN (under conditions of normal functioning of the introspective mechanism). Thereby, via the operation of a reliable process, I know that I am in pain.
Thirdly, the thesis that phenomenal concepts are conceptually irreducible in the elucidated way explains why Mary, the brilliant color scientist who has always been locked in a black and white room (Jackson 1982), cannot deduce what it is like to experience the chromatic colors while she is imprisoned. No matter how hard she thinks about the exhaustive physical and functional information at her disposal, it won’t suffice for her to possess the phenomenal color concepts of people with normal color vision. Lacking such concepts, Mary has no idea how such-and-such physical states visually feel. So, she certainly cannot deduce how they feel from the physical and functional facts.
Relatedly, the proposal I have made also explains why intuitively the phenomenal facts cannot be deduced a priori from the objective, physical and functional facts even by someone who possesses the requisite phenomenal concepts and who reflects upon them a priori in conjunction with the physical and functional truths. Suppose, for example, that I am presented with all the objective, physical and functional facts about what is going on in your brain and body at some time t. Suppose moreover that, within my repertoire of phenomenal concepts, I have the concepts necessary to conceptualize correctly what it is like for you at t. Intuitively, armchair reflection upon my phenomenal concepts and the physical and functional truths will not enable me to deduce how your state ‘feels’ on the inside. Why should this be? After all, such a deduction is possible in the case of, say, the fact that there is water in some given place p on earth and the underlying physical facts.17 Given my knowledge of an a priori truth of the form
1) Water = the F (or an F),
where ‘F’ is an objective, physical or functional predicate, (e.g., ‘bearer of enough of the following features: being a liquid, filling lakes and oceans, coming out of taps, being called ‘water’ by English-speaking experts, being necessary for life on the planet, falling from the sky’), together with knowledge of the physical/functional truth
2) H2O = the F,
and the physical truth
3)There is H2O is in place p,
I may deduce
4) There is water is in place p.18
Why is it that intuitively no corresponding deduction is possible in the phenomenal case?
The answer, I suggest, is that, where phenomenal concepts are concerned, there is no requisite a priori truth of the sort found in (1). This needs a little explanation. In the case of a natural kind concept such as the concept WATER, it seems plausible to suppose that the reference is fixed via a description. One who thinks of water thinks of it under some such description as “the liquid that comes out of taps, fills lakes, falls from the sky, etc”. The description fixes the reference of the term ‘water’, but the complex property it expresses is only contingently associated with the kind (H2O) ‘water’ rigidly designates. ‘Water’ designates, in each possible world (in which it denotes), the natural kind that has a certain cluster of manifest features in our world. Here, then, there is indeed a suitable a priori truth about water for a deduction of the above sort, namely, “Water is the bearer of such-and-such manifest properties.” This truth is contingent, but it is worth noting that there is a corresponding necessary a priori truth which would have done as well, namely, “Water is the actual bearer of such-and-such manifest properties.”
Appropriate a priori truths paralleling (1) are available for other third-person, objective concepts. So, similar deductions can be constructed in these cases.19 In the case of phenomenal concepts, however, matters are different. As noted above, one who has the phenomenal concept PAIN, for example, has a simple, recognitional concept. It functions in such a way that one can tell directly via introspection whether the distinctive phenomenal quality of pain is present without relying upon any additional clues. The phenomenal concept PAIN does not reach out to its referent via an associated description. More generally, no amount of a priori reflection on phenomenal concepts alone will reveal phenomenal-physical or phenomenal-functional connections, even of a contingent type.
This is evidenced by the widespread agreement that actual qualia inversions are epistemically possible. A priori reflection upon the phenomenal concept needed to conceptualize correctly what it is like for you at time t (call your phenomenal state ‘S’ and the relevant phenomenal concept ‘PC(S)’) does not rule out a priori the epistemic possibility that, in some other actual people, some phenomenal state other than S occupies the objective, functional role associated with S in you. Intuitively, cases of absent qualia are also epistemically possible. It could conceivably turn out that, in some actual people, no phenomenal state occupies the role S occupies in you. Likewise, intuitively, there is nothing in the character of phenomenal concepts that guarantees a priori that there are no actual cases of physical duplicates with inverted or absent qualia. It surely is not a priori true, then, that S is the F, where ‘F’ uses only objective, physical or functional vocabulary and ‘S’ expresses a purely phenomenal concept for the phenomenal character of your state at t.
It does not help, I might add, to propose, as one’s candidate a priori truth for the deduction, the weaker “S is an F” (where ‘S’ and ‘F’ are interpreted as above). Once again, no suitable a priori truth (contingent or necessary) is available. In this case, although the a priori possibility of actual cases of absent and inverted qualia is no longer relevant (for even granting such cases, S is still an occupant of the S role, for example), other a priori possibilities may be brought to bear. Thus, it could conceivably turn out that phenomenal state, S, does not actually have any of its standard physical causes and effects (and so does not occupy the S role). Epiphenomenalism with respect to the physical, for example, is surely an a priori possibility for phenomenal state S; as is the thesis that some evil demon is the real cause of its occurrence in the actual world. It is even a priori consistent with our phenomenal concepts that we, in fact, have no physical bodies, that the phenomenal appearances are radically in error.
To illustrate further the point just made about about causes and effects, suppose that you are introspecting your current experience and in so doing you bring to bear PC(S). Were you to find out that your experience actually has no physical effects — that some other state of you is the real cause of those physical goings-on you had previously taken S to produce — or were you to become convinced that some of your experiences, of which your present experience is one, actually have supernatural causes, you surely would not thereby have a license to infer that your experience does not ‘feel’ to you on the inside as you judge from introspection, that PC(S) does not apply. This is not to say that if initially you take yourself to be feeling pain, say, you would automatically be in error to revise your view and to declare now that you are not feeling pain. For my point concerns phenomenal concepts. And arguably, there is a concept of pain that is functional rather than phenomenal.
We are now ready to turn our attention to the alleged explanatory gap.
III. The Gap Examined
Consider the ‘gap’ question “Why does physical state P feel like this?” Just how is this question to be understood? Evidently, it is not a question about the causation of phenomenal states. For causation is a diachronic relation, and the question is posed with respect to a felt state (or quality) and a simultaneously occurring physical state. Nor is the question intended to presuppose that state P is the object of the feeling (the answer to the question “What is felt?”). What the question is really asking is why it is that to be in physical state P is thereby to have a feeling with this phenomenal character.
So interpreted, the question need not be taken as a question about identity. For even though the explanatory gap is one that supposedly confronts the physicalist and that, according to those who press the gap, prima facie indicates that something important is missing from the physicalist’s account, the question still arises even if it is denied that this feeling is one and the same as P. Consider, by way of comparison, the question “Why is it that to have no hairs on one’s head is thereby to be bald?” This question is not asking why having no head hairs is one and the same as being bald. For evidently one can be bald without having zero hairs on one’s head. The question is best taken to ask why having no hairs suffices for being bald; and the answer, of course, is that being bald is one and the same as having sufficiently few head hairs and having no head hairs is a realization of having sufficiently few.
Understood in a parallel way, the physicalist can respond that the explanatory gap question has a straightforward answer: this feeling is one and the same as a certain higher-level physical state Q and P realizes Q (just as in the earlier case of digestion, the action of enzymes in the alimentary canal realizes the process of digestion). To be in P , then, is thereby to have a feeling with this phenomenal character rather than to have a feeling with some other phenomenal character or to be in a state which is not a feeling at all, since being in a state with this phenomenal character = being in Q and P realizes Q rather than a physical state with which some other feeling (or no feeling) is identical.
No doubt it will be said that this reply by the physicalist merely shifts the focus of the puzzle from P to whatever higher-level physical state Q is chosen. For why is Q identical with this feeling? Why shouldn’t Q be another feeling or no feeling?
One interpretation of the question about identity is purely referential. Take the referent of the term ‘Q’ and the referent of the term ‘this feeling’ —- conceive of those referents as you will — why is the former the same as the latter? If this is how the question is understood, then there is no significant question here for the physicalist. Only one state exists, conceived of in two ways, and that state must be self-identical. On this interpretation, then, there is no need for an answer and no explanatory gap.
On an alternative concept-dependent reading, the question may be understood to ask why the physical concept expressed by ‘Q’ picks out the same state as the appropriate phenomenal concept (whichever concept is utilized in introspecting this feeling and forming a conception of its phenomenal character). Now, one natural way to take the force of the term ‘why’ here is as a request for cogent empirical reasons to believe that the concepts have the same referent. Taken this way, asking the question “Why is Q this feeling?” is like asking why Jones is the candidate most likely to be elected. In the latter case, evidently what is wanted are empirical reasons in support of the de dicto belief that Jones is the candidate most likely to be elected, e.g., that Jones has the backing of Smith, that Smith can raise more money for Jones than can be raised for any other candidate, and that campaign money is the most important factor in getting elected.20 Here, then, the given hypothesis is deduced from (or perhaps inductively corroborated by) a posteriori claims which express the empirical evidence for the hypothesis.
Understood in the above way, the question “Why is physical state Q this feeling?” has an answer of the following general sort: this feeling is physical (for how else are we to account for its causal efficacy with respect to behavior, given the very plausible empirical hypothesis that there are no non-physical causes of the physical?) and among the physical states, Q is the best candidate for identification with this feeling, all things considered (i.e., Q does the best job at explaining the range of facts concerning phenomenal consciousness we want explained21). So here there is a justification of the hypothesis that physical state Q = this feeling from empirical premises, and there is no explanatory gap between the presence of Q and the presence of that particular type of feeling.
Suppose it is now denied that what is wanted is a straightforward wholly empirical justification of the sort just provided. Suppose it is said that it is not enough to appeal to the physical facts and some appropriate a posteriori premises about experiences or feelings. What is wanted rather is an a priori demonstration or deduction of the given identity from physical facts about Q and no further information of an empirical sort. With such a demonstration, we will have an answer to the question along the following lines: Q = this feeling (e.g., the feeling of pain) since (a) it is an empirical, physical truth that Q = the F, where ‘F’ is physical, and (b) it is also a priori true that this feeling (e.g., the feeling of pain) = the F22. Understood in this way, however, as shown in the last section, it is conceptually guaranteed by the character of phenomenal concepts and the way they differ from third-person concepts that the question has no answer.23 But if it is a conceptual truth that the question can’t be answered, then there can’t be an explanation of the relevant sort, whatever the future brings. Since an explanatory gap exists only if there is something unexplained that needs explaining, and something needs explaining only if it can be explained (whether or not it lies within the power of human-beings to explain it), there is again no gap.
Whichever interpretation is adopted, then, the explanatory gap is unreal. In supposing otherwise, in insisting that there must be an answer to the question, “Why does so-and-so physical state feel such-and-such way?” that lies beyond our current grasp — an answer as complete and satisfying as our answers to the counterpart questions for solidity and digestion — we find ourselves the victims of a cognitive illusion, induced by a failure to recognize the special character of phenomenal concepts. There is no such answer; and neither is there any threat to physicalism or functionalism. What needs explaining is only why philosophers have been under the illusion that there is a troublesome explanatory gap. And that explanation I have provided.
IV. Remaining Worries
Suppose that the question,”Why is Q identical with this feeling?” is viewed as requesting cogent, empirical reasons in support of the hypothesis that Q is this feeling. Then, it may be argued, contrary to what I supposed earlier, the question cannot be answered by saying that this feeling is physical (given its causal efficacy) and, among the physical states, Q is the best candidate for identification with it, all things considered. The reason is that, given what we currently know about phenomenal consciousness, it would be quite arbitrary to settle on Q rather than some other physical state. As things now stand, we have no non-arbitrary way of selecting the privileged physical state. Here, it may be said, lies the real explanatory gap for consciousness.
My reply begins with the observation that there is a clear range of commonsense facts that any theory of phenomenal consciousness needs to explain. For example, there is the fact that I cannot experience your pains, itches, tickles, etc. There is the fact that pains, itches, tickles cannot exist unowned. There is the fact that phenomenal character is causally efficacious with respect to behavior. There is the fact that experience is transparent24. There is the fact that something can look F, in the phenomenal sense of the term ‘look’, without looking G, even if ‘F’ and ‘G’ are co-extensive. There is the fact that people can feel sensations in phantom limbs. There is the fact that I can feel a pain in a finger, when my finger is in my mouth, without thereby feeling a pain in my mouth. There is the fact that an after-image can be green, say, without anything in the brain being green.
This list is only partial, of course, even with respect to the commonsense facts. And I certainly do not mean to deny that further relevant facts may be discovered in the future. My point is that the known facts are sufficiently varied and rich that we are justified in accepting whatever philosophical theory of the phenomenal best explains them. Such a theory, I maintain, will inevitably restrict itself to the physical, given that there are no non-physical causes. And that there is a suitable theory that wins out over others, all things considered, seems to me clear.25 In my view, then, the claim that we have no way of non-arbitrarily selecting the privileged physical state is without foundation.
Another very different worry that may be raised with respect to the position I have taken focuses upon phenomenal concepts. If the special features of phenomenal concepts are largely responsible for the supposition that there is a perplexing explanatory gap between phenomenal feels and the underlying physical goings-on, then, it may be suggested, a new gap now arises at the level of concepts. For, letting the phenomenal concept of pain be PC(pain), the proposed view naturally leads us to ask, “Why does the application of PC(pain) to physical state P ensure that P feels like this?” This gap, it may be urged, is just as challenging as the original one. Thus, the problem hasn’t really been solved but instead kicked upstairs, from the level of reference to that of sense.
What is needed to answer this worry is a full appreciation of the relationship between phenomenal concepts and phenomenal ‘feels’. Phenomenal concepts are exercised (in the first person case) in our awareness of our phenomenal states via introspection. They enable us to become aware of the felt character of our phenomenal states. Without such concepts, we would be ‘blind’ to our ‘feels’. We would be in much the same state as the distracted driver who is thinking hard about philosophy, say, as he drives along the highway.26 The driver is unaware of how the road ahead looks to him, of the visual experiences he is undergoing; for his attention is focused elsewhere. But the experiences are there alright. He still sees the road in front of him. How else does he keep the car on the road?
Now the explanatory gap for consciousness is supposedly a gap between the ‘feels’ and the underlying physical/functional states. It arises, however, only for creatures sophisticated enough to be able to introspect their phenomenal states and reflect upon them. So, the explanatory gap would not arise for a creature that lacked any phenomenal concepts. Once the concepts are in the creature’s repertoire, the creature can raise and be perplexed by the explanatory gap question. The awareness of the gap, the appreciation of the supposed problem, demands phenomenal concepts; but the gap itself concerns the ‘feels’, not the concepts.
Cognitive awareness of our own feelings itself feels no special way at all. Phenomenal character attaches to experiences and feelings (including images), and not, I maintain, to our cognitive responses to them.27 Admittedly, as I noted earlier, phenomenal concepts are concepts that dispose their possessors to form images or phenomenal memories of the relevant experiences (among other things); but the concepts themselves do not have an experiential character.
In my view, then, the question “Why does the application of PC(pain) to physical state P ensure that P feels like this?” is based upon a mistaken presupposition. The application of PC(pain) does not ensure that P feels any way. Nor does it ensure that the awareness of P feels like this. For the awareness of P does not feel any way. So, there really isn’t a genuine counterpart question for phenomenal concepts. And without such a question, there certainly isn’t a higher level explanatory gap.
The final worry I want to address is one that arises out of my supposition that there is an important disanalogy between the case of natural kind concepts and that of phenomenal concepts. Natural kind concepts are rigid: each such concept picks out the same kind in every possible world in which it picks out anything. But, as noted earlier, reference is fixed to a given kind in the actual world (at least in part) via an associated description citing characteristic effects or manifest features of the kind. The concept HEAT, for example, gets its reference fixed to the natural kind it rigidly denotes, namely, molecular motion, via some such associated description as “the cause of certain characteristic bodily reactions (e.g., temperature increase, sweating, dehydration).” The property attributed by this reference-fixing description (“the ‘R”, for short), of course, is not a priori associated with the theoretical concept MOLECULAR MOTION. We discover by empirical investigation that molecular motion is the R, and thereby, given the a priori truth that heat is the R, we establish by a posteriori means that heat is molecular motion. Since the kind concept MOLECULAR MOTION, like other such concepts, is rigid, the identity between heat and molecular motion is both necessary and a posteriori.
In the case of phenomenal concepts, however, I have denied that there are any descriptive reference-fixers. Phenomenal concepts, in my view, are rigid designators, but they do not refer to phenomenal qualities (or the states whose essences are such qualities) via other qualities that users of the concepts a priori associate with them. One who thinks of the phenomenal state of pain under the phenomenal concept PAIN thinks of pain directly as pain. I have also claimed that there is no gap in the world between the phenomenal and the objective, physical (or functional) goings-on. I am thus committed to the thesis that certain phenomenal-physical (or functional) identity claims are necessary a posteriori truths. Some (e.g., Braddon-Mitchell and Jackson 1996, Chalmers 1996) will say that this is an unhappy, indeed unstable, combination of views..
This criticism gets its force from the supposition (S) that a sentence of the type “D1=D2” expresses a necessary truth that is also a posteriori only if one or other of the rigid concepts expressed by ‘D1 ‘ and ‘D2 ‘ picks out its referent via an associated property distinct from that referent to which the other concept cannot be linked just by a priori reflection. It should be noted, however, that this supposition does not require that the relevant concept always be a non-theoretical one. Indeed, we know that this cannot be the case, given the existence of wholly theoretical or scientific identity statements. Consider, for example,
Hydrogen = so-and-so quantum-mechanical system,
where ‘so-and-so quantum mechanical system’ is fleshed out so as to form a suitable rigid description. That claim, if true, is necessarily true. But it is evidently a posteriori. So, by (S), one or more of the two theoretical concepts expressed by the designators flanking the identity sign must pick out its referent via a distinct property to which the other bears only an a posteriori connection. And that condition is met. For it is surely no less reasonable to suppose that the concept HYDROGEN refers via a description, e.g., “Whatever passes the chemical tests for the stuff called ‘hydrogen’” (or something to that effect) than it is to suppose that the concept WATER has its referent fixed via a description. The complex property expressed by the description of hydrogen is not a priori linked to the relevant complex quantum-mechanical concept.
This point can be put to use by the advocate of phenomenal/physical (or functional) identities. Consider, for example, the identity claim, “Pain = B”, where ‘B’ is a rigid designator for a brain state, say.28 The concept B, like the concept HYDROGEN, can plausibly be held to have an a priori associated reference-fixing description citing appropriate laboratory tests, or perhaps the appearance of brain-state B under a cerebroscope. Since the connection between the property expressed by this description and the phenomenal concept PAIN is evidently a posteriori, there is no difficulty in holding that “Pain = B” is a necessary a posteriori truth.
Suppose, however, that the identity claim uses a rigid description on the right hand side, e.g., “the firing of C-fibres”, to take an old philosophical favorite, instead of a rigid name. Once again there is no problem. Since the phenomenal concept PAIN is not a priori connected to the property attributed by the description, the identity between pain and the firing of C-fibres will be a posteriori.
. To summarize: phenomenal concepts are very special concepts in some ways like indexical concepts. But they are not one and the same as indexical concepts. A failure to appreciate the special and a priori irreducible character of phenomenal concepts misleads us into thinking that there is a deep and puzzling explanatory gap for phenomenal consciousness. But this is an illusion. There is no such gap. Those who see in the alleged gap a reason for supposing that phenomenal qualities are special qualities, different in kind from anything physical or functional are doubly mistaken. Consciousness is indeed a wonderful thing, as Woody Allen said, but it is not so wonderful that it is magical. Why, it is, I suppose, even found in New Jersey!29
References
Armstrong, D. 1968 A Materialist Theory of Mind, London: Routledge and Kegan Paul.
Block, N. and Stalnaker, S. forthcoming “Conceptual Analysis, Dualism, and the Explanatory Gap,” The Philosophical Review.
Chalmers, D. 1996 The Conscious Mind, (Oxford: Oxford University Press).
Dretske, F. 1995 Naturalizing the Mind (Cambridge, Mass: The MIT Press, Bradford Books).
DuBois-Reymond, E. 1885-7 Reden, Leipzig.
Harman, G. 1990 “The Intrinsic Quality of Experience,” in Philosophical Perspectives, 4, J. Tomberlin, ed., (Northridge: Ridgeview Publishing Company).
Horgan, T. 1984 “Jackson on Physical Information and Qualia,” Philosophical Quarterly, 34, 147-183.
Jackson, F. 1982 “Epiphenomenal Qualia,” Philosophical Quarterly, 32, 127-136.
Jackson, F. 1993 “Armchair Metaphysics,” in Philosophy of Mind, ed. by J. O’Leary-Hawthorne and M. Michael, (Kluwer Books).
Levine, J. 1983 “Materialism and Qualia: The Explanatory Gap,” Pacific Philosophical Quarterly, 64, 354-361.
Loar, B. 1990 “Phenomenal States,” in Philosophical Perspectives, 4, J. Tomberlin, ed., (Northridge : Ridgeview Publishing Company).
Lycan, W. 1996 Consciousness and Experience, Cambridge, Mass: The MIT Press, Bradford Books. Armstrong, D. 19xx McGinn, C. 1991 The Problem of Consciousness, (Oxford: Blackwells).
Nagel, T. 1974 “What is it like to be a Bat?” Philosophical Review, 83, 435-456.
Papineau, D. 1994 Philosophical Naturalism (Oxford: Blackwells).
Rey, G. 1991 “Sensations in a Language of Thought,” Philosophical Issues, ed. by E. Villeneuva.
Searle, J. 1992 The Rediscovery of Mind (Cambridge, Mass : The MIT Press, Bradford Books).
Sturgeon, S. 1994 “The Epistemic View of Subjectivity,” Journal of Philosophy, 91, 221-235.
Tye, M. 1995 Ten Problems of Consciousness (Cambridge, Mass : The MIT Press, Bradford Books).
Tye, M. 2000 forthcoming Consciousness, Color, and Content (Cambridge, Mass : The MIT Press, Bradford Books).
Notes
1 The expression “explanatory gap” was coined by Joseph Levine. See his 1983. A referee for Mind notes that the first philosopher to have used the term ‘gap’ in connection with consciousness was Du Bois-Reymond (1885-7). Return
2 This view seems to trace back to Du Bois-Reymond 1885-7. Return
3 In some places in this chapter, I distinguish ‘functional’ from ‘physical’. Elsewhere, however, (as, e.g., in the last sentence of the text), for ease of exposition, I use the term ‘physical’ more broadly so that a functional type with a physical realization counts as physical. Likewise for a similarly realized representational type. Nothing of substance hangs upon this usage. Return
4 I take an indexical concept to be one that changes its reference from context of exercise to context of exercise. So understood, the demonstrative concept THIS is indexical. Return
5 What if I am a disembodied spirit? Arguably I am still here (in this place) even though my body is merely apparent. An alternative view is that what is a priori is that I am here, if I am in any place at all. Return
6 On my view, one becomes aware of what it is like to experience red by becoming aware (via introspection) that one is experiencing red. Those who take the position that there can be phenomenally inverted color experiences, each of which is nonetheless properly classified as the experience of red, will want to deny that there is such a single phenomenal concept as RED. Instead, there will be any number of general concepts, each of which refers to a different phenomenal quality (whichever quality that in the user of the concept is associated with experiencing red). This difference makes no difference for present purposes. The points I make below with respect to the phenomenal concept RED can be restated with respect to the appropriate preferred person-specific concepts. I use the example of RED here in part because of my own theory about the nature of phenomenal character (see my 1995 and Chapter 3 below) and in part for simplicity. Return
7 Let me stress that this is true only for the phenomenal concept RED. It is not true of the non-phenomenal concept of red, a concept normally sighted people share with blind people, for example. Return
8 The sort of memory Lewis has in mind here is phenomenal memory, not propositional memory that the experience occurred. Return
9 See, e.g., Loar 1990, Lycan 1996. Return
10 See here Tye 2000 forthcoming, Chapter 1. I should add that I think that many of the objections leveled against Lewis are unpersuasive. Return
11 This connection, in my view, is as follows: one knows what it is like to experience P only if he/she has the Lewis P-involving abilities, if P is no longer present. Return
12 Cp. Harman 1990, Tye 1995. Return
13 I myself do not accept identities of this sort. In my view, the objective states with which phenomenal states should be identified are complex representational states (Tye 1995, Tye 2000 forthcoming. For present purposes, however, this does not matter. Return
14 I do not mean to suggest here that the exercise of a phenomenal concept is always apt to trigger an image. In the case of the concept PAIN, for example, a phenomenal response may well be triggered in the form of a phenomenal memory of the feeling of pain (with the result that one shudders or grimaces a bit in reflecting upon the relevant identity claim) . But this memory need not take the form of an image (unless the term ‘image’ is understood very broadly indeed). Return
15 Cp. Papineau 1994. Return
16 In my view, the qualities of which we are directly aware when we introspect our experiences are not qualities of the experiences but rather qualities the experiences represent. This does not matter for present purposes. Return
17 Not everyone accepts this claim. See note 18. Return
18 18. See here Braddon-Mitchell and Jackson 1996, p. 133. Some dispute that such a deduction is possible even in the water/H2O case. See, e.g., Block and Stalnaker, forthcoming. Their argument seems to me unpersuasive. Block and Stalnaker mistakenly suppose that if (2) is to count as an appropriate truth for use in the deduction, it must be a priori entailed by the microphysical truths alone, and they then challenge that claim by noting that it is a priori consistent with the microphysical facts that in the actual world there is a nonphysical entity, ghost water, as well as H2O (so that while H2O is an F, it is not the F). However, the obvious reply is that what is required for (2)’s inclusion in the premises is that (2) be a physical truth, where a physical truth is not one that is a priori entailed by the microphysical truths simpliciter but rather one that is a priori entailed by the conjunction of the microphysical truths and the claim that the actual world is a minimal (micro)physical duplicate of itself. In the a priori scenario Block and Stanaker describe, the actual world is not a minimal (micro)physical duplicate of itself. So, that scenario is simply irrelevant. I owe this point to Brian McLaughlin. For a definition of the term ‘minimal physical duplicate’, see Braddon-Mitchell and Jackson, ibid., Chapter 1. It is also worth noting that in (1) the definite description ‘the F’ is not required — an indefinite description will suffice. Return
19 These deductions will not always use an identity claim at step (2). For example, where Q and R are states, the deduction will often take the following general form: Q = the state of having a state that occupies the F role; R (in Xs) occupies the F role; there is an X in state R at place p; therefore, there is an X in state Q at place p. Return
20 This example involves a descriptive concept. But the use of a descriptive concept is not essential to my point. If a similar question were raised in terms of Hesperus and Phospherus, the absence of a descriptive analysis of the names would not jeopardize an answer that cited straightforwardly astronomical reasons. Return
21 For more here, see below, pp. 000 Return
22 The term ‘feeling’ in ‘this feeling’ itself expresses a phenomenal concept, as (I am supposing) does the term ‘pain’ here. Return
23This claim is also made by Scott Sturgeon in his 1994. In this essay, which was brought to my attention by a referee for Mind, Sturgeon constructs a taxonomy of explanatory strategies and argues that none of them is applicable to solving what he calls “the problem of qualia.” His explanation of their inapplicability appeals to the special epistemic features of qualia concepts. Sturgeon’s conclusion is that “it is conceptually impossible to explain subjective experience in nonsubjective terms” (p. 235). That blanket claim seems to me too strong, given the different ways of interpreting the ‘gap’ question. But, on the current interpretation, I agree with Sturgeon. Return
24 Some philosophers (e.g., Block, 1990) deny that this is a fact. But properly explicated (see Tye 2000 forthcoming, Chapter 3), it seems to me an undeniable datum which any philosophical theory worth its salt needs to preserve. Return
25 The theory is representationalism. See Tye 1995 and Chapters 3-6 below. Once again, I use the term ‘physical’ broadly (see note 3). Return
26 This case is due to David Armstrong. See his 1968. Return
27 I do not deny that some cognitive responses have associated linguistic, auditory images. I should add that in saying here that phenomenal character attaches to experiences and feelings, I do not mean to commit myself to the view that phenomenal character is a quality of experiences and feelings. Indeed, I reject that view. See Tye 2000 forthcoming. Return
28 Again, this is not my own view. I use this identity for illustrative purposes only. Return
29 Earlier versions of this paper were read at a week-long seminar at the University of Bielefeld on Tye 1995, at a symposium on the explanatory gap at London University, at another symposium on the gap at the Society for Philosophy and Psychology, at a conference on consciousness at the University of Siena and also at colloquia at the University of Rochester, Western Michigan University, and the University of Mexico. I would like to thank Kent Bach, Ansgar Beckermann,, Earl Conee, Frank Hofmann, Terry Horgan, Keith Hossack, Brian McLaughlin, Christian Nimtz, John O’Leary Hawthorne, David Papineau, Chris Peacocke, Krista Saporiti, Wade Savage, Gianfranco Soldati, and Bob Van Gulick for helpful discussion. I am especially indebted to Martin Davies, Mark Sainsbury, and several anonymous referees for Mind. Return