Presented at the Italian Institute for Philosophical Studies, International School of Biophysics study program "From Neuronal Coding to Consciousness", Ischia (Naples), 12-17 October 1998.
In Werner Backhaus, (ed), Neuronal Coding of Perceptual Systems. New Jersey: World Scientific, Series on Biophysics and Biocybernetics, vol 9, 2001, ISBN 981-02-4164-X, pp 405-422.
Department of Philosophy U-54
University of Connecticut
Storrs, CT 06269-2054
"Consciousness" is a multiply ambiguous word, and if our goal is to explain perceptual consciousness we had better be clear about which of the many senses of the word we are endorsing when we sign on to the project. I describe some of the relatively standard distinctions made in the philosophical literature about different meanings of the word "conscious". Then I consider some of the arguments of David Chalmers and of Ned Block that states of "phenomenal consciousness" pose special and intractable problems for the scientific understanding of perception. I argue that many of these problems are introduced by obscurities in the term itself, and propose a distinction between epistemic and non-epistemic senses of the term "phenomenal consciousness". That distinction helps explain why phenomenal consciousness seems so mysterious to so many people. States of "phenomenal consciousness" are not states of one, elemental (and inexplicable) kind; they are a ragtag lot, of differing levels of complexity, corralled under one heading by a regrettable ambiguity in our terminology.
Suppose a cruise line offered you a ticket that, like this conference, promised to take you from "neuronal coding to consciousness". Do you have a clear idea of what the destination of that cruise would be? Could you distinguish it from other distinct islands of mental phenomena towards which the pilot might mistakenly steer? And, once you had disembarked, do you know how you would set about determining whether or not you had actually landed on your target?
The fact that these questions are non-experimental, abstract, and intractable has not stopped philosophers from trying to answer them, and in fact those efforts have clarified a number of important distinctions between various senses of the word "consciousness". As these provide various distinct destinations towards which our pilot might steer, I think it might actually be useful to experimenters to have some acquaintance with them. They can help sort out the different itineraries that all advertise themselves under the umbrella term "consciousness," and perhaps help prevent a few travelers from being shanghaied.
My plan of attack is as follows. Recently a number of philosophers have proposed a notion of what they call "phenomenal consciousness", and they have argued that this notion does not map onto any of the current categories available in cognitive science or philosophy. This turns out to be a good way to test the adequacy of those categories: produce some new (or neglected) phenomenon and ask whether it can be adequately characterized in the terms of that taxonomy. Is this slime mold a fungus? Is it a protist? Perhaps neither answer is quite right, and we need to revise our categories. Here phenomenal consciousness is our slime mold, and the question is whether the conventional categories can adequately characterize it. So first I will sketch the now relatively conventional distinctions, and next introduce the putative newcomer: phenomenal consciousness. Then the fun begins.
The first trick in the philosophical tool kit for clarifying a multiply ambiguous word like "conscious" is to ask: to what sorts of things do we ascribe this predicate? This immediately yields our first distinction. Sometimes we speak of a person or other creature as being conscious or not; other times the subject of the attribution is not the creature per se but a particular mental state of that creature. The former is often called "creature consciousness", the latter "state consciousness". Creature consciousness is a characteristic that a creature has when it is sentient and awake, and lacks when it is in a dreamless sleep (Rosenthal 1997, 730). Another way to put it: a creature is conscious in this sense if and only if there is some mental activity currently underway in it (Armstrong 1980, 58). If a creature is awake and senses something, it is in this sense conscious. When we arrive at some adequate account of what it is for a creature to sense its surroundings, and what it is to be awake, we will thereby have produced an account of creature consciousness.
State consciousness, in contrast, is a characteristic ascribed to or withheld from particular mental states. Was that state of mind conscious or not? For example, we might ask whether the witness consciously intended to say something false, or was honestly confused by the complicated definitions. Another example: we might have good grounds for thinking that our subject sees an X in such and such a portion of visual perimetry, but still significantly ask whether the seeing is conscious or not. In raising such questions we are not simply asking whether there was any mental activity going on, since we already know that there was some. We are asking whether a particular state of mind was conscious or not. This is state consciousness, and it is where the big problems lie.
We can next divide those problems into two, by relying on a distinction found in surface grammar. In one construction we speak of states of being conscious of something: as, variously, being conscious of one's intention to lie, conscious of the X in visual perimetry, or conscious of seeing the X. One might be conscious of objects, events, or states of affairs (Dretske 1993); but in any case the construction uses a preposition with an associated (grammatical) object: that of which one is conscious. States that are states of being "conscious of" something in this way are called states of "transitive consciousness". In another construction we use the word "conscious" attributively, and speak of "conscious mental states" as opposed to "unconscious mental states". Lacking an object, these constructions, naturally enough, denote "intransitive state consciousness".
Philosophers have toiled mightily to clarify these different senses in which one might have conscious mental states; and thanks in particular to the work of David Rosenthal (1997 and forthcoming), David Armstrong (1980), and William Lycan (1996), it is fair to report that some progress has been made. First, we have some idea of how one might account for transitive state consciousness-states of being conscious of something. Hypothesis: to be conscious of something in this sense is to have a mental representation of it. States of transitive consciousness are representational states; the object of which one is conscious is the object one mentally represents. As Rosenthal puts it, "One is transitively conscious of something if one is in a mental state whose content pertains to that thing-a thought about the thing, or a sensation of it" (Rosenthal 1997, 737). This makes transitive state consciousness a species of what philosophers call "intentionality"; the state is "directed at an object", à la Brentano, in the sense that it is directed at an intentional object: it represents that object. To be conscious of something is to represent it. The state of being conscious of something is a state that has a content pertaining to that thing; it refers to it or is about it.
This may not seem to be much progress, but in fact there has been an enormous amount of work in recent decades in linguistics, philosophy, and cognitive science on the nature of mental representation (see Stich & Warfield 1994); and if the proposed equation holds true, then all that work can be brought to bear on this seemingly distinct problem of consciousness. On this line an answer to the question "how does a mental state represent a thing?" would also be an answer to the question "how can one be conscious of a thing?". Transitive state consciousness would fall out, as a lemma, from a solution to the problem of mental representation.
I noted that there are at least two prima facie different ways in which mental states might have contents that pertain to an object: one can have a thought about the object, or one can have a sensation of it. Some have challenged the legitimacy of the second disjunct, denying that sensations have any representational character. It is indeed true that a sensation of a red apple differs from a thought about it, and that the way in which thought hooks up with its objects is not the same way that sensation hooks up with its. But it is relatively uncontroversial to claim that sensory states are invested with at least some representational content. Consider two cases:
(1) A stimulus is F and is perceived to be F.
(2) A stimulus appears to be F, but is not.
The second is a case of "non-veridical appearance": something appears to have some feature that in fact it does not. Cases (1) and (2) may present the same appearance, but in one case the appearance is less than veridical. How do we explain these "facts of sensible appearance"? (See Broad 1927, 234-37.) What is common to those two situations is not necessarily any shared property of the distal stimuli; instead it is a characteristic found only in the sensing of those stimuli. Both appear to be F. One potent explanatory strategy is to hypothesize that the stimulus that appears to be F but is not causes a sensory state with the same qualitative character as one caused by stimuli that are F (see Clark 1993). We posit that in both cases just mentioned the stimulus configuration causes sensory states with the same qualitative character. Both situations cause the subject to represent the stimulus as if it were F. In one situation that representing goes wrong. Notice that we cannot explain the latter, non-veridical appearance in terms of anything currently present in the scene; the explanation lies in how that scene is represented. We posit a state with a qualitative character whose content is not currently satisfied.
These "qualitative characters" give us the classical understanding of the now infamous "qualia". Qualia in their original acceptation were cited in explanations of the facts of sensible appearance (see Broad 1927; Lycan 1996, 69-89). They were properties of sensory states in virtue of which the things causing those sensory states appear as they do. The term has since acquired many other connotations, often contradictory, so one never knows exactly what someone else might mean by it. Today it sometimes betokens a commitment to intrinsic, non-intentional mental properties of sensory states. Yet even the proponents in those debates accept that sensory states have some representational content; so they would accept the thesis that both thought and sensation can subserve transitive state consciousness. Both have contents that pertain to something, and so provide different ways in which one can be conscious of that thing.
Transitive state consciousness may in this way reduce to mental representation. A second proposed reduction is that intransitive state consciousness can be explained in terms of transitive state consciousness. That is, what it means to say that M is a conscious mental state is just that one is conscious of M. The attribution does not add some ghostly intrinsic property to M; it simply indicates that M is represented in the appropriate way by another of one's mental states. Typically that other mental state must occur with M and arise from M in some relatively direct and unmediated fashion. Other conditions are added until one is satisfied that if M is represented in that way, then one is conscious of M.
If we also accept the first thesis we get one or another of the two current "higher order" theories of consciousness. One way of being conscious of something is to have a thought about it. This is the basis of the "higher order thought" model of consciousness.
Higher order thought: M is a conscious mental state if and only if it is accompanied by the appropriate kind of thought about M.
That thought is "higher order" because it is a mental state about another mental state. Not all thoughts about one's mental states suffice to make those states conscious mental states: the details of what makes a thought to be a thought of the appropriate kind are quite important. But if having a sensation of an object can also suffice to make one conscious of it, there is a second possibility, called the "higher order perception" or "inner sense" model.
Higher order perception (inner sense): M is a conscious mental state if and only if one is conscious of it, and to be conscious of it one must sense it. The latter requires the operation of a quasi-perceptual and higher order sensory modality, variously called an "internal scanner" or "inner sense", whose objects are one's own mental states.
On both accounts, that M is a conscious mental state is a relational fact, requiring some additional mental state representing M in a suitably immediate way.
This complicated structure gives both theories the wherewithal to distinguish between state consciousness and more complicated varieties. For example, both theories treat introspection-the intentional and deliberate scrutiny of one's own mental states-as a special case. Typically the higher order thought in virtue of which one is conscious of M is not itself a conscious mental state. One is conscious of M, but not also conscious of thinking of M. But the deliberate character of introspection can change this: in that setting one may not only be conscious of one's own mental states, but also conscious that one is conscious of them. In introspection one might be "consciously aware" of one's own mental states, as we say. The higher-order theories can accommodate introspection by allowing that the higher order mental state can itself become a conscious mental state; one simply tacks on a yet higher order thought, about it. But at some point, perhaps when working memory is filled, the cascade of being conscious of being conscious of being conscious of ... something must come to an end.
Similar complications arise when providing an account of Thomas Nagel's famous "what it is like to be a bat" locution. Nagel claimed that
The fact that an organism has conscious experience at all means, basically, that there is something it is like to be that organism. ... an organism has conscious mental states if and only if there is something that it is like to be that organism-something it is like for the organism. (Nagel 1979, 166)
He calls the properties that would characterize what it is like to be the organism the "subjective character of experience", and this is a useful technical term. As Nagel himself insists, the question "what is it like?" arises only for conscious mental states. Even if we accept that DB, the blindsight subject, sees the X in his scotoma, he is not conscious of seeing the X. What is it like to be DB seeing the X in his scotoma? There is nothing it is like. What is like to see an X in your blindspot? Not much. Similarly, perhaps you have an ache or pain that lasts all afternoon, but thanks to anesthesia or momentary distraction, there is an interval during which you are not conscious of it. During that interval, what is it like to have that pain? Again the answer is: it is like nothing at all. (Or, one might say: it is just like having an unconscious pleasure.) Only conscious mental states have a subjective character. Nagel would agree.
It is important for later developments to notice that such subjective character is therefore distinct from the qualitative character earlier ascribed to sensory states. Sensory states are not invariably conscious mental states (Rosenthal 1991). Witness DB seeing the X in his scotoma, or the pain one still has during those intervals one is not conscious of it. These states would have qualitative, but not subjective, character. Perhaps, as Lycan (1996, 54-68) argues, "subjective character" can be treated as a kind of second-order qualitative character. Subjective properties are properties of the states of the internal scanner that help to explain how one's other mental states appear to one. Lycan argues that this account can explain the appearances of private, perspectival, and incommunicable subjective facts. That it can provide such an explanation is one attraction of the "inner scanner" model. But the important point is that subjective character, however it is handled, is not co-extensive with qualitative character. The latter can occur in first-order sensory states of which one is not conscious.
A higher order theory of consciousness is, in a certain sense, a reductive theory of consciousness. Suppose some day we are granted an understanding of sentience and sensory states with their qualitative character. These provide one variety of transitive state consciousness. Suppose we also come to understand mental representation, or at least how thoughts represent. We would then understand two kinds of transitive state consciousness, or two ways of being conscious of something: by sensing it, or by thinking about it. What other resources do we need to give a full account of conscious mental states? It may seem that the latter question is a further and much harder question, requiring new and mysterious resources for its solution. But the bold answer of the higher order theorist is: nothing more is needed. Solving those problems would solve the problem of consciousness. With an understanding of sensory qualities and of how thought represents, you can give a full account not only of mental states but of the difference between conscious mental states and unconscious ones.
At this point, if not earlier, certain philosophers are likely to erupt in protest. They claim that there is a kind of consciousness entirely missing from our catalog: a kind that cannot be identified with or analyzed in terms of the functional and psychological notions so far listed. A zombie, they claim, might be endowed with all the machinery we have listed so far, yet in the most important sense remain totally unconscious. So our catalog is missing the most important kind of consciousness: phenomenal consciousness.
Ned Block introduced the technical sense of the term "phenomenal consciousness" (or P-consciousness) in the course of contrasting it with what he called "access consciousness". Of course since it cannot be analyzed in terms of functional or psychological notions, it is (regrettably) impossible to give a definition, but one can at least list some synonyms and point to examples. Block says:
P-consciousness is experience. P-conscious properties are experiential ones. P-conscious states are experiential, that is, a state is P-conscious if it has experiential properties. The totality of the experiential properties of a state are "what it is like" to have it. Moving from synonyms to examples, we have P-conscious states when we see, hear, smell, taste, and have pains. P-conscious properties include the experiential properties of sensations, feelings, and perceptions, but I would also include thoughts, wants, and emotions. (Block 1995, 230)
Suppose you experience the sight of a red patch or the smell of something musty. The seeing and the smelling are then states of phenomenal consciousness. Those states have a special kind of property: experiential properties. The totality of those properties define "what it is like" to have the experience-the seeing or the smelling, respectively.
Experiential properties are also called "phenomenal qualities" by Block: their sum defines how something looks, feels, smells, or (generically) appears. Block says "it is in virtue of its phenomenal content or the phenomenal aspect of its content that a state is P-conscious" (Block 1995, 232). What is phenomenal content? Etymologically, phenomena are appearances. A characterization of phenomenal content is a characterization of sensory appearance: how things look, seem, feel, or (generically) appear. So a state of phenomenal consciousness is one in which something appears somehow-presents some kind of appearance-to someone. The qualities that characterize these appearances, are, aptly enough, phenomenal qualities. To characterize the content of a state of phenomenal consciousness, you must somehow identify that quality: the exact quality of how things look, smell, or feel. Block (1995, 232) says "there is such a thing as a P-conscious type or kind of state. For example, the feel of pain is a P-conscious type-every pain must have that feel." But various "appears" words might be pressed into service: something looks red, smells musty, or feels smooth. Sadly, these words do not and cannot convey the exact qualitative content of the experience. According to Block (1990, 55), looks red and all other public language words have public meanings, and express only the intentional contents of those experiences-what the experiences represent. The sought-after qualitative contents are closer to a particular person's "mode of presentation" or concept of the look of red (Block 1990, 55). Such qualitative contents may (allegedly) differ in two people even though something looks red to both of them.
Block's position is an intricate one, and here there is neither time nor space to do it justice. I have no quarrel, for example, with the useful and important distinction between phenomenal consciousness and access consciousness. I will focus instead on the very narrow question of whether P-consciousness is a kind of consciousness that cannot be analyzed in terms of the notions introduced above. Block explicitly enters such a claim on behalf of P-consciousness (Block 1995, 230, 235), but the evidence for it is unconvincing. The various examples and synonyms can be handled perfectly well with extant categories; we do not need to enter a new kind into our taxonomy. But before making that argument, it will be helpful to describe another recent and related account of phenomenal consciousness. This is found in David Chalmers' book The Conscious Mind. Its subject matter turns out to be precisely: phenomenal consciousness. Chalmers says
a mental state is conscious if it has a qualitative feel-an associated quality of experience. These qualitative feels are also known as phenomenal properties, or qualia for short. The problem of explaining these phenomenal properties is just the problem of explaining consciousness. This is the really hard part of the mind-body problem. (Chalmers 1996, 4)
Phenomenal consciousness, he says, is "the really difficult problem for a science of the mind"; it "poses the worrying problem of consciousness" (Chalmers 1996, 29-30). So what are phenomenal properties?
According to Chalmers, many of our terms for states of mind have both a "psychological sense" and a "phenomenal sense". The psychological concept characterizes causal or functional roles. The phenomenal concept characterizes appearances or what it is like to have the state in question. Chalmers says
the phenomenal concept of mind ... is the concept of mind as conscious experience, and of a mental state as a consciously experienced mental state. ... On the phenomenal concept, mind is characterized by the way it feels... (Chalmers 1996,11)
He says that "what it means for a state to be phenomenal is for it to feel a certain way" (Chalmers 1996, 12). By "feel a certain way" Chalmers means not just tactile experience, but sensory appearances of any kind, including visual, auditory, and so on. So conscious mental states are states that have a "phenomenal feel".
Colour sensations provide the "paradigm examples of conscious experience" (Chalmers 1996, 6). Other examples of states of phenomenal consciousness include sensory experiences such as hearing music, smelling freshly baked bread, or feeling pain, and also feeling emotions or having mental imagery. Chalmers also equates the phenomenal features of a mental state with "what it is like" to have that state. We get the following juxtaposition:
what it means for a state to be phenomenal is for it to feel a certain way. ... in general, a phenomenal feature of mind is characterized by what it is like for a subject to have that feature ... (Chalmers 1996, 12)
Presumably, then, characterizing the phenomenal feel of a mental state characterizes what it is like for a subject to have that state. We can link all the terms together: there is something it is like to have a state if and only if it is a conscious state; if and only if it has a phenomenal feel (or a qualitative feel or a phenomenal property or a qualitative property); if and only if, finally, it feels a certain way.
Chalmers serves admirably as an exemplar of the view that states of phenomenal consciousness are states of an elemental kind: they serve as primitive and irreducible notions in the science, which can be connected to the rest only at the level of fundamental laws. He says these things outright. The notion of conscious experience-phenomenal consciousness-should be taken as a primitive term (Chalmers 1996, 4, 23). Just like Block, Chalmers explicitly denies that states of phenomenal consciousness can be functionally defined (Chalmers 1996, 105). Phenomenal consciousness is irreducible, and an explanation of phenomenal properties will require a new set of fundamental laws, analogous to new fundamental laws of physics (Chalmers 1996, 277).
Although Chalmers' formulations are considerably more exuberant than Block's, they do share some conclusions. States of phenomenal consciousness involve a special kind of quality: phenomenal qualities. A state of phenomenal consciousness is a state in which something appears somehow to someone, and phenomenal qualities characterize that appearance. They characterize, for example, how something looks, feels, tastes, smells, etc. On both accounts, states of phenomenal consciousness, to the extent that they can be identified using any words in a natural language, are best identified using various "appears" words. The feel of pain provides a paradigm example for both authors (see Block 1995, 232). Looks red provides another.
I will argue that the examples and formulations of Block and Chalmers do not suffice to identify any one kind of mental state, but instead introduce several distinct kinds, of differing levels of complexity, intermixed and named with the same name. Each of these various kinds already has a place in our taxonomy. So states of phenomenal consciousness are not mental states of some new and previously unrecognized kind. What is new about the notion is just the way it mixes together several old kinds. These investigators have not produced a slime mold, which might challenge the very categories of fungi and protist. Instead the sample they bring forth is just a mix of well known lichens and algae.
The argument starts by posing a dilemma. "Appears" words are notoriously ambiguous. Looks red has what philosophers call epistemic and non-epistemic senses (see Chisholm 1957; Dretske 1969; Jackson 1977). An "epistemic" sense characterizes evidentiary relations: the role of that sensory experience in the justification of belief. Obviously it is of some practical significance to notice which judgements a sensory experience would tend to validate or refute. "Looks" and the other "appears" words used epistemically characterize sensory experiences as being of kinds that would tend to confirm some claims, and disconfirm other ones. In contrast, the "non-epistemic" sense is devoid of implications about what one might or might not judge.
When we propose that a state of phenomenal consciousness is a state of sensing something that looks red, which of these senses of "looks" is meant? The first possibility is that we are using "looks" in its epistemic sense. Consider, for example, Wilfrid Sellars' account of what it means to say "that looks red". Sellars argued that "looks red" is logically more complicated than "is red"; he says "being red is logically prior, is a logically simpler notion, than looking red" (Sellars 1963, 142). To illustrate the point Sellars tells a story about a necktie shop, whose owner John has never used electric lighting, and has never reported the colour of anything except under "standard conditions" of northern daylight. John has no locutions for reporting appearances or how colours look; he just reports what the colour is. Then at last one day John installs electric lighting and turns it on for the first time. Under the new lights a blue necktie (as we would say) looks green. His friend Jim demonstrates how the apparent colour of the necktie changes depending on the illumination. Initially John is befuddled, since he doesn't want to say that the blue necktie is green when inside the shop, or that the lighting changes its colour. But Sellars tutors him, when inside the lit shop, to stifle his otherwise natural report that the necktie is green. He is taught to say first "It is as though I were seeing the necktie to be green" (Sellars 1963, 143). Finally he acquires a new locution: the necktie looks green.
When I say 'X looks green to me now' I am reporting the fact that my experience is, so to speak, intrinsically, as an experience, indistinguishable from a veridical one of seeing that x is green. Involved in the report is the ascription to my experience of the claim 'x is green'; and the fact that I make this report rather than the simple report 'X is green' indicates that certain considerations have operated to raise, so to speak in a higher court, the question 'to endorse or not to endorse'. I may have reason to think that x may not after all be green. (Sellars 1963, 145)
He also says
To say that 'X looks green to S at t' is, in effect, to say that S has that kind of experience which, if one were prepared to endorse the propositional claim that it involves, one would characterize as seeing x to be green at t. (Sellars 1963, 146)
So "that looks green" means something like
It causes in me the same kind of visual state as would normally lead me to judge it to be green, but something is amiss, and I wish to withhold judgement.
Sellars notes that the experiences that would lead one to report "it looks green" or "it is green" might as experiences be indistinguishable from one another. "Two experiences may be identical as experiences, and yet one be properly referred to as a seeing that something is green, and the other merely as a case of something's looking green" (Sellars 1963, 145). The only difference is epistemic: in one the content of the experience is endorsed, and in the other it is not. "It looks green" might also be phrased "Visually it is just as if I were seeing something green, but I do not endorse the claim that it is green". I am visually representing something to be green, but I do not endorse that representation.
The interesting implication is that reports in the epistemic sense of "that looks green" are themselves reports of a higher order thought. Unlike reports of the form "that is green", a report "that looks green" expresses a thought about one's own mental state. Suppose it means "I am in the sensory state that would normally lead me to judge that thing to be green, but something is amiss, and I wish to withhold judgement". It follows that to be in a state in which something looks green to me, I must be in a state whose content is "I am in a sensory state that would normally lead me to judge that thing to be green, but something is amiss, and I wish to withhold judgement." This content includes a higher-order comment about one's own visual state: that it is of a kind that would normally lead me to judge the thing in question to be green. Or: that I am in the same kind of visual state that I would be in if the thing causing it were green.
Any time one issues the epistemic report "that looks green" one is making a report in part about one's own visual state. One is reporting that visual state to be of kind G, where G is the kind of visual state that would normally lead one to judge the stimulus causing it to be green. If I can report "I am having a visual state of kind G" then I am expressing a thought about one of my visual states. It follows that the visual state in question is a state of which I am conscious.
In short, on this, the epistemic horn, "that looks green" is not a simple report about the visible features of something in front of the eyes; instead it is a much higher order report, in part characterizing one's own visual state. If Sellars is right, its truth conditions require one to be conscious of one's visual state-and to be conscious of it as a state of a particular kind. It is not true unless I withhold endorsement from one of my own visual representations; and I cannot do that unless I am conscious of it as a representation. It requires a rather stunningly sophisticated judgement about the judgements one would make on the basis of the sensory state one is in.
This helps to explain why so many people think that phenomenal properties must be conscious. To report "that looks green" in the epistemic sense in fact does require one to be conscious of one's visual state. We will never teach John the locution "that looks green" unless John is conscious of his own visual state while he scrutinizes the necktie, and is conscious of it as being of a particular kind. If therefore we think of "looks green" as the prototypical phenomenal property, we will find it only in episodes featuring a conscious visual state. Phenomenal properties become properties exclusively of conscious states.
But there is an illusion involved in thinking of "looks green" as the prototypical phenomenal property. "Looks green" is not the simplest, bedrock level; in its epistemic settings it reports a higher order thought. Notice also that the vaunted first person perspective here contributes to the illusion. It invariably confirms the natural prejudice that phenomenal states are, necessarily, conscious mental states. If we start with the first person perspective, we start by confining ourselves to those mental states of which we are conscious. Given that sample, it will be impossible to disconfirm the hypothesis that states of being appeared to are necessarily conscious states. After all, the first person perspective cannot generate any counterexamples. Any example you produce from your first person perspective will be a state of which you are conscious. Failing to produce a first-person example of a phenomenal state of which one is not conscious, one naturally thinks that counterexamples are impossible. That states with phenomenal content are necessarily conscious mental states may acquire the lofty appearance of a conceptual truth. But it is simply a confusion to equate the properties of having phenomenal character and of being a conscious mental state (see Rosenthal 1991). That the equation must hold for that which falls within the ambit of the first person perspective is due to the limitations of the perspective, not to the identity of properties. These are two independent characterizations, not one.
The non-epistemic sense of "looks red" is logically simpler, and in many ways it provides a more natural interpretation for what proponents of phenomenal consciousness propose. So let us see how things stand on this side of the dilemma. To get at the non-epistemic sense of "looks red" we need to scrape away all logical connections between the characterization of the visual state and the sorts of judgements it might tend to confirm or disconfirm about the stimulus causing it. That is, one must sever all logical connections between something "looking red" in this non-epistemic sense and any belief, any judgement, or any propensity to judge that something is red or even appears to be red (see Dretske 1969, 4-18).
The surest way to sunder all such connections is to interpret "looking red" as designating a non-relational or intrinsic property of the visual state itself. The idea is that the visual state intrinsically either is or is not of this kind, the non-epistemic-looking-red kind; it either does or does not have a "looks red" character. It can have that character independently of whatever stimulus caused it and whatever thoughts and beliefs it later causes. The claim that a visual state has this character has very modest logical implications: it does not even require that a stimulus exist, or that the visual state engender any thoughts. To use another terminology, it requires only that the creature hosting be "appeared red to" (see Chisholm 1957, 62). It is sometimes put adverbially: the creature is "appeared-to redly". The odd terminology is useful: it reminds us that non-epistemic talk about visual states arises only in very odd contexts. To talk about these properties we probably need such new terminology, since it is doubtful that any ordinary words pick them out.
The non-epistemic interpretation of states of phenomenal consciousness has a logical simplicity sorely lacking from "looks red" interpreted epistemically. But just how simple are they? One must not confuse the non-epistemic qualitative character looks red (or "being appeared-to redly") with the content that one might express by saying "that looks red." The former states with their qualitative character presumably occur in humans (and in other animals) that lack language. It is entirely possible that states with the same qualitative character occur in members of different species. As I understand it, macaque monkeys have colour vision more or less equivalent to that of human beings. If my visual system can enter a state with the qualitative character "being appeared-to redly", there seems no reason to deny that a macaque monkey, whose colour vision is so similar, might entertain a state of the same kind. Those kinds are defined, recall, so as to exclude implicating much of anything outside of basic seeing, which the monkey assuredly has. A human might be appeared-to redly without knowing that (or being able to judge that) he or she is being appeared-to redly. Perhaps our subject is a profoundly retarded human being, who lacks the capacity to acquire a natural language.
It is helpful to think of such states as occurring in a creature that cannot formulate any judgements: a creature, for example, that can sense red things, but lacks any capacity to judge that the things it sees are red. Perhaps this is any animal that cannot learn a natural language. To fix our ideas, imagine that the visual state in question is occurring in a macaque monkey, hunting for red berries in the green foliage. Jack the macaque monkey is being appeared-to redly. In the sea of green he spies a red berry, and proceeds to pluck it. If you believe that macaque monkeys can formulate a judgement that we would translate as "that looks red", then substitute an animal that you believe lacks that capacity, but nevertheless can still see things. Perhaps Jack is a bird (or an insect) hunting for a red flower in the green foliage. Jack the honey bee is appeared-to redly.
States of phenomenal consciousness, so understood, have the requisite simplicity; but that very simplicity leaves proponents without the resources needed to summon up their mysteries. The macaque monkey is appeared-to redly, and perhaps is thereby conscious of the cluster of berries, but it does not follow that it is thereby conscious of being appeared-to redly. Its state of being appeared-to redly is not necessarily a conscious state. Furthermore, the monkey has a state with a qualitative character, but unless that state is a conscious state (in the intransitive sense), it is hard to make out the claim that there is something it is like to be the monkey being appeared-to redly. Recall the day-long headache. It has an unfortunate and distressing qualitative character, but in the intervals during which you are not conscious of it, there is nothing it is like for you to have that headache.
There is no compelling reason to conclude that states of being appeared-to redly, in this non-epistemic sense, are necessarily conscious states. Granted, they are sensory states: they are states of a creature that is awake and sentient, and they are states in virtue of which the creature is conscious of the berry or flower in front of it. But sensory states are already well ensconced in our catalog. Not all such states are necessarily conscious states. Being appeared-to redly might occur without one's being conscious of being appeared-to redly. Perhaps that state makes one conscious of something-something in front of the sense organs, or appearing to be there-but still one may not be conscious of that state.
Phenomenal properties-states of being appeared to, or states in virtue of which something looks or feels or appears a certain way-must be distinguished from conscious phenomenal properties. A state of "phenomenal consciousness" might mean "phenomenal state of which one is conscious", but even then both such terms are already in our catalog. Indeed there is already a special entry for such states. Nagel (1979) insisted that his "what is it like" locution can apply only to conscious mental states. Subjective character ("what it is like") is already a heading, distinct from qualitative character. So states of phenomenal consciousness once again fail to be a new kind of consciousness.
If we treat "being appeared-to redly" in the non-epistemic sense, then it names a property with very modest connections to anything outside of basic seeing. It is an intrinsic property of one particular visual state. But intrinsic properties of one particular cannot entail facts about other particulars. Those further facts are separate existences. That one is (or is not) conscious of the state in question, for example, is such a further fact. Just as no intrinsic property of a Honda could tell us who owns the Honda, so no intrinsic property of the state itself could tell us who is conscious of it, or, for that matter, whether anyone is. Hence, if we adopt the non-epistemic interpretation, and plunk for such modest implications, it must be possible for a subject to have a state with the "being appeared-to redly" property without (also) being conscious of that state. Similarly, qualitative character is claimed to be an intrinsic property of a visual state. But then it must be possible for a subject to have a state with a green qualitative character without being conscious of that state. The subject may be a macaque monkey, spotting a red berry.
Allow me to enter a bold conjecture. States of "phenomenal consciousness", so understood, are just ordinary first order sensory states: the very same states that psychologists and neuroscientists study when they study the neuronal coding of perceptual systems. Suppose for example we are curious about a visual state with the qualitative character looks red. One must carefully distinguish these states from their more sophisticated cousins, and crop away the mysteries born of confusing them. But after that exercise, I believe you will find that the proper chapter in which to look up looks red is the chapter on colour perception, not the chapter on consciousness. The story you read about colour perception, and about the mechanisms that enable the macaque monkey to discriminate the coloured fruit from the green and dappled background, is a story about a state of phenomenal consciousness. It is a state of a creature that is sentient and awake (i.e., a state of creature consciousness); and it is a state that make its host (transitively) conscious of something. We need not stir in some additional mysterious property to elevate it to the status of phenomenal consciousness. It already has that status. So when we understand colour perception, we will thereby understand some states of phenomenal consciousness-those involved, for example, in our humble episode of Jack being appeared-to redly, as he hunts for a berry in the bush.
The problem with the notion of "phenomenal consciousness" is that the term is so laden with ambiguities that we cannot trust it to pick out any one kind of mental state. It points ambiguously at a number of different kinds, of differing levels of complexity. One would naturally suppose, for example, that states of phenomenal consciousness are both phenomenal and conscious; but there is no good reason to think that states with phenomenal properties are necessarily states of which one is conscious. The best available hypothesis is that there are two distinct processes involved. First we have sensory processes, generating mental states with first order qualitative character, invoked to explain how things out there appear. Second, there is an additional and optional process, in virtue of which one is sometimes conscious of one's sensory states. This is a higher order process that has a different subject matter than do sensory processes. It is optional in the sense that it need not always accompany them. Granted, when one is awake and alert, the two kinds of process are almost invariably conjoined. But if they are two, it is possible that one occur without the other. "States with phenomenal properties" pick out one set. "States of which one is conscious" pick out a different set. Definitions should not fuse them.
So we hunt for some other interpretation for what a "state of phenomenal consciousness" could be. Unfortunately, given the ambiguities in the words "phenomenal" and "conscious", there are lots of possibilities. One is simply that we are talking about a "state of consciousness" in the sense of creature consciousness: a state of a creature that is conscious. It is the state of a creature that is awake and sentient; our macaque monkey, for example, seeing the red berry. The monkey is being appeared-to in just the way that we would be if we saw the red berry. So the state has phenomenal properties. Voilà: a state of consciousness with phenomenal properties, a.k.a. a state of phenomenal consciousness. Alternatively, one might interpret "consciousness" as transitive state consciousness: a state of being conscious of something. By seeing the red berry our monkey is conscious of it. So its seeing the red berry is a state of transitive consciousness. Once again, voilà, a state of phenomenal consciousness: this time a state with phenomenal properties in virtue of which one is conscious of something (see Dretske 1993). In neither case should the portentous terminology be allowed to mask the fact that we are just talking about sensory states: the same states that researchers in colour perception are busy researching. Little did they know that they are at work on the problem of phenomenal consciousness. Let the word go out from this place! You can if you like attach this impressive label to their problems, but I do not see how it helps matters.
The more dangerous conflation is to mix together and confuse states of differing orders of complexity. If we have terms that are applied both to states that are higher-order and to states that are not, then this conflation can become difficult to detect. It might tempt us into believing that some states can somehow fuse into their substance mutually incompatible orders of complexity. The notion of phenomenal consciousness can lead us further into such temptations. The words we use to characterize sensory appearance, such as "looks", "appears", "seems", "feels" and so on, have precisely such ambiguities. The report "that looks red" in its epistemic sense requires stunning sophistication: one must note that one is in a visual state of a certain kind-the kind typically produced by things one would judge to be red-but also that this time the visual state is produced by a thing about which such a judgement must be tentative. So epistemic use of "appears" words requires one to be conscious of one's own sensory states; it requires higher-order states. Yet in the non-epistemic sense, "that looks red" allegedly characterizes an intrinsic property of the simplest kind of mental state: the qualitative character of a sensation. Conflating these two uses of "looks red", one might naturally think that a state in which something looks red is both phenomenally elemental and at the same time a state of which one must be conscious. It is a simple qualitative state that somehow, at the same time, carries within it consciousness of itself. Perhaps this is a state of phenomenal consciousness. It would indeed be a new, and utterly mysterious-nay, paradoxical-kind of consciousness.
At least some of these mysteries are of our own making. I approve here of the Wittgensteinian diagnosis: our intellect is bewitched by language. Conflating the different senses of "looks red", and attempting to fuse the differing orders of complexity, will load metaphysically impossible demands onto states of consciousness. States of "phenomenal consciousness" are not states of one, elemental (and inexplicable) kind; they are a ragtag lot, of differing levels of complexity, corralled under one heading by a regrettable ambiguity in our terminology. Further signs confirm this diagnosis. Other terms have the same ambiguity as "looks red", and they too are prominent in the discussions of phenomenal consciousness. Take the words "feel" and "feeling". They are applied both to the activity of sensing and to the conditions or things sensed. Consider the scope for confusion if we used "see" both for the activity of seeing and for the things seen. For the latter we have the word "sight", in its first acceptation, "seeing the sights"; but imagine the scope for confusion if we talked about "seeing the sees" instead of "seeing the sights". Suppose you see your sees. If you see something white, are you seeing a see that is white? If that is possible, why can't you just have a white see? Is the seeing of the white see also white? Are the two whites the same? And so on. It would become very confusing to sort out which are properties of the things seen and which are properties of the seeings of them.
Yet these are precisely the problems that "feeling" forces upon us. If you feel an ache in your knee, is it the knee that aches, or the feeling of the knee that aches? How could the feeling of the knee be located out there, in all that cartilage, where the ache appears to be? And so on. Furthermore, "feeling" as a verb has a sense in which it implies attentive regard, probing and exploring, testing and withdrawing, and so implies awareness of the things touched and felt. Such probing and exploring is conscious. In this sense "feeling for the keys" implies that one is conscious of the things one feels. We had already decided that what you feel are feelings, so naturally enough we get: conscious feelings. Mixing together all three senses, we arrive at "phenomenal feels". These would be objects that entail their own apprehension, that have phenomenally simple properties, yet at the same time are conscious states. Mysterious indeed!
"Experience" has the same ambiguities. The things you experience are, of course, experiences. "Experiential properties", like "phenomenal feels", come to characterize both the activity of experiencing and the things experienced. You cannot very well experience something without being conscious of it. So, just like phenomenal feels, experiential properties acquire the capacity to necessitate their own apprehension. Haven't you heard? Any self respecting conscious state can do that.
Armstrong, David M. (1980) The Nature of Mind and Other Essays. Ithaca, New York: Cornell University Press.
Block, Ned, Flanagan, Owen, and Güzeldere, Güven (1997) (eds.) The Nature of Consciousness: Philosophical Debates. Cambridge, Mass.: MIT Press.
Block, Ned (1990) Inverted earth. In James E. Tomberlin (ed), Philosophical Perspectives, 4. Action Theory and Philosophy of Mind, 1990. Atascadero, California: Ridgeview Press. Reprinted in Block, Flanagan, and Güzeldere (1997), 677-94.
Block, Ned (1995) On a confusion about a function of consciousness. Behavioral and Brain Sciences 18 (2): 227-88. Reprinted in Block, Flanagan, and Güzeldere (1997), 375-416.
Broad, C. D. (1927). Scientific Thought. London: Kegan Paul.
Chalmers, David J. (1996) The Conscious Mind. New York: Oxford University Press.
Chisholm, Roderick M. (1957) Perceiving: A Philosophical Study. Ithaca, New York: Cornell University Press.
Clark, Austen (1993) Sensory Qualities. Oxford: Clarendon Press.
Dretske, Fred (1969) Seeing and Knowing. Chicago: University of Chicago Press.
Dretske, Fred (1993) Conscious experience. Mind 102: 263-283.
Jackson, Frank (1977) Perception: A Representative Theory. Cambridge: Cambridge University Press.
Lycan, William G. (1996) Consciousness and Experience. Cambridge, Massachusetts: MIT Press.
Nagel, Thomas (1979) What is it like to be a bat? In Mortal Questions, by Thomas Nagel. Cambridge: Cambridge University Press, 165-80. Reprinted in Block, Flanagan, and Güzeldere (1997), 519-28.
Rosenthal, David (1991) The independence of consciousness and sensory quality. In Consciousness, ed. by Enrique Villanueva. (Philosophical Issues, vol. 1). Atascadero, California: Ridgeview Publishing Co., 15-36.
Rosenthal, David (1997) A theory of consciousness. In Block, Flanagan, and Güzeldere (1997), 729-53.
Rosenthal, David (forthcoming) Consciousness and Mind. Oxford: Clarendon Press.
Sellars, Wilfrid (1963) Science, Perception, and Reality. London: Routledge & Kegan Paul.
Stich, Stephen & Warfield, Ted (1994) (eds.) Mental Representation: A Reader. Oxford: Blackwell.
Return to Austen Clark's online papers .
Return to the Philosophy Department home page.