Inner Sense until Proven Guilty
Can one sense ones own mind, as one senses nonmental entities in ones environment and body? According to many contemporary philosophers of mind, the fraudulent commonsense idea of a "minds eye" obstructs clearheaded attempts to explain introspection and consciousness. I concede that inner sense cannot directly explain consciousness and introspection in all their forms, but I do think a carefully specified kind of inner sense can account for one very special kind of introspective consciousness. It is special because it is the key to explaining the most puzzling kind of consciousness, phenomenal consciousnessthere being "something it is like" to have certain mental states. My aim in this paper is to defend this view against accusationstwenty-two in all!rather than to argue positively for the view. However, I begin by indicating some of the motivation for the account I defend.
Postcartesian philosophers in almost every major tradition have turned to inner sense for a description or explanation of consciousness. The interesting regularity, historically speaking, is that key progenitors of these traditions appeal to inner sense. It is as if inner sense is the persistent, instinctual, null hypothesis about consciousness. At the dawn of British empiricism, Locke writes that "[c]onsciousness is the perception of what passes in a Mans own mind" (1689, 115; II, I, 19); he holds that this faculty of reflection, "though it be not Sense, as having nothing to do with external Objects; yet it is very like it, and might properly enough be calld internal Sense" (1689, 105; II, I, 4). The rationalist who is in many ways Lockes counterpart, Leibniz, recommends that "it is well to make a distinction between perception, which is the inner state of the monad representing external things, and apperception, which is consciousness or the reflective knowledge of this inner state itself" (1714, 637). German idealism arrives with Kant, who refers to "empirical apperception" of a "flux of inner appearances"mentioning that "[s]uch consciousness is usually named inner sense" (1787, 136; A 107). Phenomenology and experimental psychology find common ancestry in Brentano, who maintains that mental states are known most fundamentally through "inner perception" (1874). Likewise David Armstrongperhaps the first analytic philosopher to attempt a systematic theory of mind in the contemporary aftermath of behaviorismidentifies introspective consciousness with "inner sense" (1968).
How is inner sense supposed to be distinctively analogous to outer sense? Is there an illuminating analogy? Armstrong explains inner sense as being, like outer sense, "selective" (incomplete, about the mind or environment), "fallible," and "causal" (1980). This is not much of an analogy, since probably all cognitive processes display these features. Even ones most theoretical scientific beliefs, for example those about quantum-mechanics or cosmology, are selective, fallible and causal, without being sensory in any interesting sense. Another initially tempting idea is that inner sense is a causal but noninferential source of evidence about mental states (or, more cautiously, that it is as low as outer sense on flexible, all-things-considered inference). But this wouldnt distinguish inner "sense" from inner "reactions" of other sorts. Also, many sworn enemies of inner sense favor views according to which introspection produces beliefs noninferentially, and without the involvement of any sensory states. Lack of inference is insufficient to qualify a process as "sensory." In my view, the products of inner sense are sensory due to their use, even if not due to their origin. They are literally states of the various outer-sensory systems themselves, rather than being states of a distinctive inner faculty. This makes available a strong sense in which the products of inner sense count as visual, auditory or other sensory states, with minimal violence to the ordinary use of these terms. I will describe some consequences of this view at the end of this section, after detailing the inner-sense theory of phenomenal consciousness.
An inner-sense theory of phenomenality would be flatly circular if sensing, itself, presupposed phenomenality. This point is so important to the view and to criticisms of it that it deserves a name. With apologies to Merleau-Ponty:
"Perception" is nonphenomenal in the sense that neither the concept of a perceptual state nor the property of being a perceptual state presupposes that the state is phenomenal. Of course, the concept and property do not preclude phenomenalityindividual states can be both perceptual and phenomenal. I reserve the word "experience" for phenomenal states, whether perceptual or otherwise. (I also use "sense" and "perception" synonymouslylikewise for their respective cognates. And "having a mental state" broadly covers having or being in or undergoing mental states, events, processes, structures, objects, acts, etc.) Since I think the Nonphenomenology of Perception is best supported by actual cases, I will not delay over what senses of "can" should be involved in the principleconceptual possibility, logical possibility, biological possibility, etc. In cases of subliminal visual perception (Dixon, 1987) and "blindsight" (Weiskrantz, 1988), subjects act on the basis of information about the visual features of objects, despite denyingsincerely and without hypochondriathat they have relevant visual experiences. It is plausible, though somewhat controversial, that in at least some cases there is nothing it is like to have subliminal and blindsight states. Also, the early layers of processing in vision and other sensory modalities seem not to constitute experience; further processing is necessary. Otherwise, it would be difficult to explain why phenomenal visual experiences are not continually like double imagessince one has separate left-eye-caused and right-eye-caused early visual states. It would also be difficult to explain why one seems not to see arrays of incoming brightnesses, i.e., why phenomenal experiences do not represent distant objects as being "behind" two splotches of light near ones eyes.
Many rival theories of phenomenal consciousness can accept the Nonphenomenology of Perception. Theories diverge over what needs to be added to nonphenomenal sensing to reach phenomenal experience. Inner-sense theories hold that experience arises only when states are innerly sensed, states that otherwise ("on their own") would be nonphenomenal. Call the states directly produced by inner sense "I-states," since they are typically inner-directed (or about ones own mental states). Similarly, call the states directly produced by outer sense "O-states," since they are typically outer-directed (or about things other than ones own mental states, such as environmental or bodily entities). Typically, I-states are sensitive to neural or functional features of O-states, while O-states are sensitive to environmental features of environmental entities (I illustrate some such features under Accusation 14, below). On an inner-sense view a perceptual or imaginative experience as of some entity E (an object, event, etc.) essentially involves both O-states and I-states of O-states, as illustrated here (where the dashed arrows signify causation that is typically present but not strictly required):
The important point is that on a properly formulated inner-sense theory experience is compositional. An experience as a whole need not be innerly sensed. Rather, a component of an experience (an I-state) represents another component of the experience (an O-state). In a derivative way, the experience itself may be said to be innerly sensed in virtue of having a part that is innerly sensed, just as a house may be sensed in virtue of having a part that is sensed. And since the inner sensing is itself part of the experience, the experience may be said to represent itself, "reflexively," in addition to representing features of E.
I will now sketch four styles of argument I use elsewhere for this overall view. The first is analysis of the concept of phenomenal consciousness (under review-ahereafter, "Stopgap"). From wholly nonphenomenal premises about inner sense we can deduce, as a matter of conceptual necessity, that there is something it is like to have mental states. Roughly, if a mental state essentially appears to its bearer to have some feature F, then it is "F-like" for its bearer, so it is "like something" for its bearer. Strictly speaking, for a state innerly to appear to its bearer as a whole, rather than to some small part of its bearers brain, the I-state must be poised for wide and systematic influence on cognition and conationit must be available to "central" systems of belief and desire (see Fodor, 1983). So the diagram of experience must be complicated as follows (where the solid arrow signifies causation that is required for full-blown experience):
In addition, O-states may (and normally, do) have their own outputs to central systems. Neither inner sensing of a state without central availability, nor central availability of the state without inner sensing, suffices clearly for phenomenality of the state. A state that is innerly sensed, but not in a centrally available way, is perhaps like something but not like something for its bearer. A state that has impact on central systems, but that is not innerly sensed, is perhaps for its bearer but not like something for its bearer. Only as a unit does centrally available inner sense insure that the state is like something for its bearer.
The second style of argument is inference to the best explanation of allegedly troublesome features of qualia (1994hereafter, "Qualia!"; see also section 7 of "Stopgap"). If the presence of (centrally available) inner sensing explains there being something a state is like, the specific content of the inner sensing should explain what the state is like, its "phenomenal properties" (or "qualia"). Roughly, if a mental state essentially innerly appears to be Q, then the state has phenomenal property (or quale) Q. The adjective "phenomenal" in "phenomenal property" functions similarly to the adjective "apparent." This means that an experience can have a phenomenal property (or quale) Q without having (property) Q. The appearance of Q may be false; an I-state may misrepresent an O-state as having Q, just as an O-state may misrepresent the features of environmental objects. But even if the appearance is false, the experience has quale Q. The property of having quale Q is literally a property of experiences, namely, an experiences property of representing itself in certain (e.g., inner-sensory) ways as Q. To be sure, a quale (e.g., Q) of an experience is itself a real property, with real (intentional) relations to the experience, even if it happens not to be a property of the experience. This captures what is true in qualia "realism"the property Q can exist and bear the "is a quale of" relation to an experience, and can even be had by the experience. It is also the key to explaining features of qualia such as:
The theory explains all of this while preserving insights from philosophical frameworks that have often been thought incompatible with these claims, such as:
One advantage of the inner-sense theory is that it makes sense of so many apparently plausible though apparently conflicting views of qualia.
Third, there is inference to the best explanation of phenomenal illusions (under review-bhereafter, "Illusions"; see also 1996bhereafter, "Nonphenomenal"). Inner sense provides an independently plausible unified explanation of two widespread illusions built into phenomenal experience. The "image illusion" is an appearance of phenomenal objects subjectively having normal environmental properties: one "sees mental images" as having shape, "feels pains" as throbbing, "hears inner speech" as having variable pitch, etc. The "transparency illusion" is an appearance of normal environmental objects objectively having phenomenal properties: from various perspectives one perceives constant-shaped objects as having varying modality-specific shape-looks and shape-feels, etc. The key to explaining the illusions is that there are "binding confusions" between O-states and I-states. O-states and I-states represent features that are not in fact had by the same things (at best, O-states represent features of environmental entities and I-states represent features of O-states). The confusion is that these states mistakenly attribute these features to the same objects, or, in other words, that the states operate as if they are about the same objects. O-states and corresponding I-states represent the bearer of these features in a unified way, as if by using the same variable, name, or dummy constant (say, x):
The transparency illusion results when the inner (neural or functional) features Q represented by I-states are "projected" onto the environmental or bodily objects represented by O-states. (This does not mean that perceptual experience is wholly illusorythe O-states may be perfectly accurate about the features of the objects.) Conversely, the image illusion results when the outer (environmental or bodily) features P represented by O-states are misattributed to the inner objects represented by I-states (namely, O-states). (Attention determines which illusion is presentwhen attention is directed outwardly, O-states have priority over I-states and x is identified as an outer object, but when attention is directed inwardly, I-states have priority over O-states and x is identified as an inner object.)
The final style of argument is inference to the best explanation of correlations. The illusions are present in all clear cases of phenomenal consciousness, despite the otherwise wild heterogeneity of these cases (see sections 1 and 3 of "Illusions"):
Call these the "Qualitative Quartet." The transparency illusionthe appearance as of objective looks and feelsis rife in normal perception, misperception, dreams, and diffuse bodily sensation. The image illusionthe appearance as of subjective likenessesis most at home in normal imagination, degraded perception, thought, and nondiffuse bodily sensation. By contrast, inner sense and the illusions are absent for attitudes and moods (see "Illusions"). On the inner-sense theory this explains why phenomenality is absent for attitudes and moodseven when they are conscious in some sense (I argue for this absence on independent grounds in "Nonphenomenal"). While there is often something it is like when we have a conscious mood or attitude, this is not due to the mood or attitude itself but due to coexisting Quartet experiences. Conscious moods and attitudes are not themselves phenomenalthey may be for a creature but they are not like something for the creature.
I have said that I-states count as sensory due to their use, i.e., due to their functioning as outer-sensory states, as O-states. The binding confusions between O-states and I-states are one illustration of this. The unification of O-states and I-states into phenomenal experience is another. Once I-states are produced, in whatever way, they are processed like O-states, as further states in particular sense modalities. For example, I-states in vision help to produce visual beliefs, help to control visuomotor skills, and are not introspectively distinguished in kind from visual O-states. I-states should differ from inner judgments (or beliefs) in several ways that O-states differ from judgments, on plausible psychological assumptions (cf. Fodor, 1983). Compared with judgments, O-states (i) are largely under stimulus control, (ii) are in control of reflexes and motor skills, independently of countervailing desires, (iii) are resistant to rational inference and will, (iv) are restricted in content, and (v) are not easily matched in content by desires, memories, communication-intentions, etc. Similarly, we should expect I-states (i) to be largely under control of O-states, (ii) to guide reflexes and skills, (iii) to be difficult to form by will or reason, (iv) to involve a limited range of contents, and (v) to be difficult to remember and express in public language. The account I offer makes many commitments about the function of inner-sensory states, but makes no commitments about the internal structure of the process of inner sense.
It would be difficult to complete a list of the prominent philosophers who have brought charges against inner sense. Plaintiffs range from staunch eliminativists about qualia (Gilbert Harman, Daniel Dennett) to staunch realist/antireductionists about qualia (Colin McGinn, John Searle, David Chalmers). They are aided by defenders of rival reductionist theories of phenomenality (Robert Kirk, Michael Tye) and of introspection (Gilbert Ryle, Sydney Shoemaker, William Lyons), and also by some doing pro bono work, not obviously in the service of their own positive views (Dan Lloyd, Christopher Hill). The higher-order-thought theory of consciousness is perhaps the closest kin to inner-sense theories, but its staunchest proponent (David Rosenthal) and staunchest opponents (Fred Dretske, Güven Güzeldere) harmonize that inner sense is a nonstarter. Some complaints are even due to erstwhile defenders of inner sense (Brentano, Armstrong)! In addition, although I do not discuss them, there are accusations due to a long line of phenomenologists (including Edmund Husserl and Jean-Paul Sartre) and, at the other extreme, accusations due to a long line of opponents of the myth of the given (including Ludwig Wittgenstein and Wilfrid Sellars).
First, I discuss arguments against the very existence of inner sense, arguments based largely on introspection and phenomenology. Scientifically-based arguments against the existence of inner sense follow. Then I turn to an assortment of arguments to the effect that, even if inner sense does exist, it is unnecessary for phenomenal consciousness or insufficient for phenomenal consciousness. In a few cases I violate this classification in order to juxtapose accusations that may be answered similarly. Nevertheless, the discussions are for the most part independent, with some explicit cross-references, so readers are encouraged to follow their own priorities. For convenience, in the following table I list the accusations I address.
Phenomenology v. Existence of Inner Sense
Shoemaker presses the accusation as follows:
I agree that when we introspect experiences, we don't get new experiences of the experiences. On my inner-sense account an experience represents itself, in virtue of having component I-states that represent component O-states. The two states (O- and I-) combine to form a single experience.
The accusation would be correct if every sensing did "involve" a sense-experience, as Shoemaker alleges. Then each distinct sensingthe O-state and the I-statewould involve a distinct experience. But this cannot be assumed in objecting to inner-sense theories; it violates the Nonphenomenology of Perception. Without the I-state, the O-state yields no experience as of an object. Without the O-state, the I-state yields no experience as of an object (see Accusation 8). The position is that outer sense and inner sense together generate experience in the first place.
This accusation is similar to the preceding one except that it concerns duplicate qualities rather than duplicate states. Even if it is granted that inner sense involves only a single experience, a plaintiff might expect that this single experience would involve distinctive outer-sensory and inner-sensory qualities. As Lyons argues:
According to this accusation there is a fundamental disanalogy between inner and outer sense: inner sense doesnt have "its own" phenomenal properties. I accept the point that inner sense does not have its own phenomenology. However, I wish to deny the premise that it should, if it is to be fully analogous to outer sense. The premise is based on the view that outer perception has its own phenomenology. This is plausible so long as we focus only on full-blown conscious outer-sensory experience. But many outer-sensory states are much humbler, given the Nonphenomenology of Perception. According to a properly formulated inner-sense theory there is nothing it is like simply to have an O-state, nothing it is like simply to sense outerly. Ordinary outer sense is not sufficient for phenomenology; unadorned O-states are not phenomenally conscious and involve no phenomenal qualities. So it is not a disanalogy that I-states dont have their own phenomenology, distinct from that of ordinary O-states.
The accusation alleges that "introspecting a patch of blue" borrows its phenomenal qualities from "perceiving a patch of blue." But the qualities of "introspecting a patch of blue"at least for inner-sensory introspection, i.e., innerly sensing an O-state as of a patch of bluearent "borrowed from" but bestowed upon "perceiving the patch of blue," a state that would otherwise be nonphenomenal. My position is that outer sense and inner sense together generate phenomenal properties (qualia) in the first place. A quale belongs to them jointly, i.e., to the experience they jointly compose. Güzeldere considers a response that is similar in effect:
But the response is not ad hoc. It is the essence of any inner-sense theory of phenomenality. It appeals to a principle that is fundamental to inner-sense theories from the outset, and plausible on independent grounds of subliminal perception, blindsight, early-visual states, etc.the Nonphenomenology of Perception.
Several authors have remarked on what G. E. Moore (1905, 450) calls the "diaphanous" nature of visual experience, the apparent difficulty or impossibility of noticing qualitative features of experience rather than features of the external world. When one tries to attend to features of normal experiences, one normally "sees through" the experiences to nonmentalenvironmental or bodilyobjects. Given diaphanousness, there seems to be no role in normal visual experience for inner sensing of mental phenomena, distinct from outer sensing. Gilbert Harman presses this accusation as follows:
Defenders of introspectible intrinsic qualia may claim not to satisfy Harmans prediction (see Block, 1990, 72-74); this response gains some plausibility in cases of degraded vision (e.g., blurred or doubled vision), discussed in Accusation 18 below. It is also not clear how, if at all, Harmans claims are supposed to generalize from visual perception to visual imagination, where features are often introspected as if they belong to alleged mental objects within experience. Also, Harman runs together "attending" to something and "experiencing" it, thereby forgetting about inattentive experiencea common mistake I highlight in Accusations 15-18, below. Nevertheless, Harmans prediction is plausible in the main.
In whatever cases Harmans prediction stands, it shows that seeings do not seem introspectively to have intrinsic properties. Harman moves directly to the "eliminativist" conclusion that seeings have no introspectible intrinsic properties. This is a leap. For all Harman has shown, it may be that one (innerly) senses an O-states having some intrinsic property Q, but senses it as an environmental objects having Q. Just as one can perceive an object as a different object, so one can perceive an objects having a property as a different objects having the property. I hear the ventriloquists speaking, but I hear it as the dummys speaking. Likewise, in watching a big-screen, fast-paced, high-visual-fidelity film consisting of life-size images, it is surprisingly hard to see colors and shapes as being on the screen rather than in the scene. If one blocks out the surrounding room (perhaps by looking through a tube), it does not seem that one is looking at a screen at all. Nevertheless, what one is seeing is a screen, and the colors and shapes one sees are on it. By contrast, when the film becomes degraded or blurred, it is easy to see the colors and shapes as being on the screen. One does not see subtitles or credits as being in the scene rather than on the screen; perhaps these are similar to apparent afterimages that can seem "superimposed" over normal perceptual experience.
In a slightly different context, Harman himself concedes the basic point: "one might be aware of intrinsic features of experience without being aware of them as intrinsic features of experience" (1990, 42). He argues that such a mistakenly-perceived property would not be "psychologically significant" in any sense that would amount to an "objection to functionalism" (1990, 42-43). I am not concerned to dispute this; the primary aim of Harmans discussion is defending functionalism, and in that I wish him well. But his concession undermines his secondary aim, of attacking inner sense. The latter aim is unnecessary to the former, since a properly-formulated inner-sense theory need not be in any way hostile to functionalism.
In order to resist Harmans eliminativism about introspectible intrinsic properties of experience, one does incur a serious burden of explaining what these properties are, and why they seem to belong objectively to trees and other environmental objects rather than to experiences. The inner-sense theory delivers, in its explanation of the transparency illusion: as I argue in section 2 of "Illusions," I-states and O-states use the same "terms," and so when attention is directed outwardly and objects are identified as outer, the properties represented in inner sense are mistakenly attributed to the outer objects. Harman might object that we should only as a last resort suppose that normal perceptual experience is subject to an illusion. However, he has no leverage to make such an objection. In "Illusions" (section 1) I argue that we would be subject to a widespread illusion even on Harmans own view, according to which we see "relational features of the tree from here."
The illusion here consists in the fact that we experience many of these allegedly "relational" features of trees as being stuck on the trees, without relation to us. As we move around a tree, we in fact bring about and participate in varying spatial relations to the treeand thereby bring about and participate in what Harman wishes to call the trees "size-from-here" and "shape-from-here." But we do not seem to create or participate in any size-ish or shape-ish properties of the treeall such features seem intrinsic to the tree. It does not seem to one as if by moving one brings about new shape-ish features in the way one brings about new spatial relations to the object. Rather, it seems that, from here, one can discover that the tree has this objective lookstuck on one side of it, saywhile from there, one can discover that the tree has that objective lookstuck on another side of it. Harman is incorrect in suggesting that shape-looks and size-looks are "presented as relations between the viewer and the tree." So if Harmans faith in diaphanousness were aptif appearances of diaphanousness were not deceivingthis would license the elimination not only of introspectible intrinsic properties of experience, but also of introspectible nonintrinsic relations, including relations that are suited to explaining the difference in sensitivities to the varying looks of objectslooks of their shape, location, reflectance, etc. The argument would prove too much.
Lyons considers a view very much like my position on inner sense, namely, the view that we introspect "the products of an act of perception" (1986, 105). He presents the following argument against this position:
Thus, Lyons suggests that to embrace inner sense of sensory states, one must be a "veil-of-perception" theorist. And since we have reason to reject the latter view (see the next accusation), we have reason to reject the former.
I agree that there are no "phenomenal sense data," "interposed" between environmental objects and ones sensings of environmental objects. But a properly formulated inner-sense model of experience is not committed to sense data. Inner sense as I have described it requires two things: O-states of environmental objects, and I-states of the O-states. O-states arent "interposed" between environmental objects and our perceptions of themthey are our perceptions of environmental objects. And I-states certainly arent "interposed" between objects and our perceptions (O-states) of them; they come "after" O-states in the cognitive flowchart. Inner sensings are directed not at entities interposed between objects and ones perceptions of them, but at ones perceptions themselves. The causal chain in perceiving an apple neednt proceed from the apple to an inner sensing and then to a sensing of the apple. Rather, on a more natural view, the causal chain goes directly from the apple to a sensing of the apple (an O-state), and then (in cases in which the apple-sensing forms part of a phenomenal experience) to an inner sensing (an I-state) of the sensing of the apple. Both outer sense and inner sense are "direct" in that neither requires mediation by further sensings. Nor are inner sensings required for outer sensings to be centrally available to beliefs and desires. Each kind of sensing can and does have its own independent impact on central systems.
There are two traditional ways to press this accusation. One arises as an objection to the view that there are sense data interposed between environmental objects and sensings of them. As Ryle argues:
Whatever the merits of this as an argument against meddling sense data, it does not affect the present view, which is not committed to sense data (see the previous accusation).
The second traditional argument arises from the claim that all mental states are conscious, in the sense that one must be conscious of all ones mental states. Brentano considers the following argument:
Again, there is no commitment on the present view to the idea that all mental states are conscious, in any sense or disjunction of senses of "conscious." (For example, the states that provide arguments for the Nonphenomenology of Perception are good candidates for wholly unconscious mental states.) So inner sense cannot be shown to involve a regress in either of these traditional ways.<5>
Brentano defends both inner perception and the claim that all mental states are conscious. He avoids infinite regress by maintaining that inner perception is reflexive, that an inner perceiving represents itself (in addition to other things):
This view is partially preserved on the present account, according to which experiences represent themselves. But experiences represent themselves only in a derivative sense, by containing parts (I-states) that represent other parts (O-states). I think nonderivatively reflexive representation would be difficult to explain (see section 6 of "Stopgap"). In any event, it would severely weaken the analogy between inner sense and outer senseO-states are certainly distinct from their objects, so I-states should be distinct from their objects, O-states. Phenomenologists after Brentano have accordingly tended to see (nonderivative) reflexivity as a rival to inner sense.
Puzzlingly, Lloyd has pressed a distinct regress argument against reflexivity views; if it is successful it would also work against the present view that phenomenally conscious experiences are derivatively reflexive:
I do not see how to make sense of this accusation. Talk of "distinct states, C = C˘ = C˘ ˘ = " would be an "absurdity"how could states be "distinct" yet "="? But this is Lloyds talk, not Brentanos or mine. At most, given Lloyds terminological use of primes, double-primes, etc., reflexivity views are committed to the following innocent claim: every state of (phenomenal) consciousness has (not "is") an infinitude of distinct names (not "states"): "C," "C˘ ," "C˘ ˘ ," . There is also no reason to think that "distinct contents" entail "distinct states." A state can have a single conjunctive content combining two distinct contents(for Brentano) that the glass is empty AND I am (hereby) aware that the glass is empty, or (for me) that x is empty AND x is Q. Or a state can be ambiguous or multiple in meaningmeaning both "the glass is empty" and "I am (hereby) aware that the glass is empty," or both "x is empty" and "x is Q." Such maneuvers are not applied ad hoc to conscious experience; I argue elsewhere (1996a), on independent grounds, that individual, token, mental states are normally multiple in meaning.
This is another very puzzling accusation against reflexive representation, which would hardly seem worth discussing were it not put forward by two philosophers who have thought most deeply about inner-directed representation. Rosenthal argues:
A mental state can represent itself without its being conscious (or aware) of itself, just as a state can represent a banana without its being conscious (or aware) of the banana. Just as the sentence "This sentence contains five words" has itself as a referent, so a mental state can have itself as a referent. The reflexivity of experiencewhether Brentanos nonderivative kind or my derivative kind (see the previous accusation)certainly does not require experiences to be experiencers. I have no good guess about why it would be thought otherwise.
Shoemaker argues that "[n]o one thinks that one is aware of beliefs and thoughts by having sensations or quasi-sense-experiences of them" (1994, 255).<7> I agree, wholeheartedly. If attitudes were phenomenal experiences, this would undermine an inner-sense theory of phenomenal consciousness. But on independent grounds it is plausible that attitudes are not phenomenal (see "Nonphenomenal"). There is nothing conscious attitudes themselves are like, although there is often something accompanying states are like, states from the Qualitative Quartetperceptual experiences, bodily-sensational experiences, imaginative experiences, and experiences in the stream of thought (I give some examples in the next paragraph). The absence of inner sense for attitudes is not a weakness of inner-sense as a theory of phenomenal consciousness, but one of its greatest strengths. A serious difficulty for theories of phenomenal consciousness is to explain why attitudes are not phenomenal, even when they are conscious in some other sense. Of course, attitudes can be excluded by fiatfor example, I think that Tye does this by a sheer stipulation that phenomenality requires "nonconceptual" content (1995, 137-139). But to my knowledge only inner-sense theories provide an explanation of the nonphenomenality of (conscious or unconscious) attitudes.
If we do not innerly perceive our attitudes, how do we know about them? Is inner sense completely irrelevant to attitude self-knowledge? Shoemaker turns the screws:
On the contrary, I think it is utterly normal, in becoming conscious of ones attitudes, to have accompanying phenomenally conscious "symptoms" of the attitudes (see "Nonphenomenal"). For example, ones standing, unconscious belief that snow is white may cause one to form an auditory image of quickly saying the words "snow is white" (or "I believe snow is white" or "Mon Dieu! La neige! Blanche!" or ). There are normally more aspects to this verbal imagery, which help one to determine which kind and strength of attitude (belief, desire, suspicion, etc.) is revealed via the thought. In cases of belief, for instance, one may imagine saying the words in an assertive tone of voice, and without any concomitant proprioceptive sensations of suppressed giggling, or auditory images as of appending "NOT!," etc. In addition to such verbal imaginings, there are often nonverbal imaginings, e.g., of a white expanse of snow, and perhaps visual imaginings of words. And consider William James interesting introspective report:
It is no part of my proposal that each belief or desire has a canonical phenomenal symptom, and I do not here assert (or deny) that every bit of attitude self-knowledge proceeds via phenomenal symptoms. But to say the least, we should be very suspicious of any view of introspection that marginalizes such an important and nearly ubiquitous entryway into knowledge of attitudes (see note <9>, below).
Searle argues as follows against the analogy between introspection and perception:
To answer this accusation, it is necessary to describe how the inner sense theory is supposed to apply to "conscious thought." We must be careful about what counts as a "thought." In one sense, "thought" refers to propositional attitudes generally, and in a related sense it refers to indicative propositional attitudes specificallybeliefs, hypotheses, etc. These senses are irrelevant to the present theory, which is not intended to apply directly to attitudes (see the previous accusation). In the phenomenally relevant sense of "thought," a thought is an element of the "stream of thought." First and foremost, this includes the fairly slow, roughly serial, and typically deliberate phenomena we try to describe as "talking to oneself" or "thinking in words." (There is also what we are tempted to describe as "thinking in images"say, visual images of objects or words.) Such "inner speech" is simply conscious auditory imagining (or hallucinating) of speech. On the present view, it involves O-states of linguistic items, and I-states of such O-states.
We must also be careful about adopting Searles suggested procedure of "taking away" elements of a sensory encounter. The task should not be one of voluntarily controlling the existence of objects and states. This cannot always be done in the case of outer senseI can close my eyes to prevent myself from seeing a bicycle, but when I am awake and a loud noise erupts, I cant voluntarily prevent myself from hearing it. The (alleged) causal connections between O-states and I-states, in normal subjects, are much more difficult to control voluntarily. Instead, the task is to conceive of the absence of elements of a sensory encounter, and determine what remains. But how is this determination to be accomplished? Certainly introspective reports in Searles style have a role, but they are not the only legitimate source of evidence.
The inner-sense theory applied to conscious inner speech has the following consequence: if either O-states of words or I-states of these O-states are absent, then there is no conscious experience as of words remaining. But is Searle correct that there is nothing left at all? It is plausible that there can be O-states of words without accompanying I-statesperhaps these underwrite unconscious or subliminal linguistic processing. This would mirror the perceptual casevisual O-states without visual I-states underwrite unconscious or subliminal visual processing. Searles introspective method is wholly silent on whether such nonintrospectible states are possible.
Could there be I-states as of O-states, in the absence of accompanying O-states? Perhaps not for normal subjects, in the absence of breakdown. (Outer hallucinations do not require mechanical breakdown within the subjectan abnormal environment sufficesbut inner hallucinations would.) Still, what might occur in such a case? There is some leeway on my account, which is appropriate for such an unclear case of breakdown. I think it most likely that a state consisting only of I-states would result in no experience at all. But given that qualia are primarily determined by the contents of I-states, perhaps I-states alone would be sufficient for an unusual kind of experience. The closest familiar comparison would be to experiences as of mental imagery with environmental properties (as contrasted with experiences in normal perception and nonlucid dreaming). These experiences occur when there are both I-states and O-states, but when the I-states have more salience and strength than the O-states. So perhaps if the O-states are wholly absent, the experience would be as of something inner, but without environmental propertiesin the case of visual imagery, an apparent seeing of something mental without apparent shape, size, color-reflectance, etc.; in the case of auditory verbal imagery, an apparent hearing of something mental without apparent pitch and volume. These might be comparable to cases of perceptual agnosia, in which patients can perceive environmental objects without being able to recognize them: one might be able innerly to "see an image" or "hear voices in ones head" by qualia without being able to "recognize" any normally correlative apparent shape, size, color-reflectance, pitch, volume, etc. To the extent that we have words and concepts ready for environmental properties but not for qualia, such mental imagery might be singularly ineffableperhaps even a step toward the "pure experiences" countenanced in Buddhist texts: alleged experiences as of no ordinary perceptible properties, whether instantiated outerly or innerly. Again, Searles examination of his own experience leaves entirely open whether such psychologically unusual experiencesexperiences that are not clearly as of experienced outer or inner objectsare possible.
Shoemaker is willing to accept that introspections and their objects are distinct (see the previous accusation), but he argues against inner sense by appeal to a related feature of normal perception:
According to Shoemaker, if introspection is inner sense, then innerly sensed mental states should similarly be independent of inner sense. However:
His strategy is to show that "there is a conceptual, constitutive, connection between the existence of certain sorts of mental entities and their introspective accessibility," at least for creatures that have "the ability to conceive" of those entities (1994, 272-273). The disanalogy with outer sense is that one cannot be entirely introspectively blind to these mental entities, while one can certainly be perceptually blind to trees, mountains, etc.
His argument for this conclusion is long, complex, and worth pursuing, but not here. Even in advance of the details I think it can be shown that the overall argument is fallacious. It would prove too much. Just as there is an "independence condition" for normal perception, so there is an one for normal belief. Trees, mountains, houses, automobiles, and human bodies, etc. can all in principle exist without being believed in, and even without there being creatures capable of believing in them. If the argument yields any evidence that we do not have inner-directed sensings, it yields equally good evidence that we do not have inner-directed beliefsinner sense would be "disanalogous" to outer sense only in the same sense that inner belief would be disanalogous to outer belief. If Shoemaker is right about the conceptual relation between introspection and introspected, then this would be a "fundamental difference" between belief and introspection! But of course that should not incline us to doubt the existence of innerly-directed beliefs.<9> Nor should Shoemakers argument, whatever the details, incline us to doubt the existence of innerly-directed sensings.
Shoemakers proceeds to argue, case by case, that introspectibility is necessary for the existence of pain feelings, perceptual experience, agency, and propositional attitudes. This provides another simple way to see that the argument presents no threat to the inner-sense theory I propose: the (fundamental) objects of inner sense are nonphenomenal O-states, rather than phenomenal pain feelings or perceptual experiences (see Accusations 1 and 2). I agree that pain feelings and perceptual experiences are dependent on introspection, and in addition my view explains this: the inner sensings of parts of these experiences are themselves parts of the experiences.
Rosenthal offers inner-sense theorists a gift:
As it turns out, the gift is a Trojan horse:
If Rosenthals idea is simply to disarm the "differential-response" point as a source of support for inner sense, then I wouldnt mind. Nor would the citizens of Troy have minded if the Greek soldiers had simply jumped out of the horse, dismantled it, and taken it away peacefully. But Rosenthal seems to think his argument works as an objection to inner senseindeed, as a bit of scrutiny that inner sense cannot withstand.
I agree that mental states are a heterogeneous class.<10> This heterogeneity is not a threat to my inner-sense theory of phenomenal consciousness, for two main reasons. First, inner sense is restricted to sensory O-states, rather than attitudes and other mental states. Second, my view is not committed to there being a single faculty of inner-sense. I-states are not states of a special inner-sense modality, but states of the separate outer-sense modalities themselves. There are inner-visual I-states of outer-visual O-states, inner-auditory I-states of outer-auditory O-states, etc. This is necessary (at least given our cognitive architecture) for the I-states to be bound with the O-states, generating image illusions and transparency illusions. Rosenthal is correct that qualia in various modalities are heterogeneous, but the various inner sense faculties are equally heterogeneous.
Science v. Existence of Inner Sense
Although this accusation is often mentioned in passing, it is rarely taken to be seriously damaging to inner sense theories. In part this is because it is unclear just what is plausible or implausible, neurophysiologically. To my knowledge, the boldest statement of the accusation is due to Rosenthal:
Leave aside the contrast between "internal to conscious states" and "due to inner perception"if inner sense is part of a conscious experience, the two are compatible. Rosenthal provides no positive reason to believe in "the absence of any such organ or mechanism," so the matter stands where inner-sense theorists have left it: with some support for a philosophical bet on inner sense, together with a wait-and-see attitude towards brain science.<11>
Of course, inner perceptions are not generated by literal inner eyes and ears, that are affected by light-waves or air-waves inside the head, transducing them into neural signals. The relevant innerly sensible features of O-states are already neural (or functional, but realized neurally), so there is no need for inner transduction. Although there are outer sense organs that produce O-states, there need not be inner sense organs that produce I-states. The mechanisms (if any) by which O-states cause I-states are not what make the latter sensory, in my view. Whether these mechanisms involve literal inner sense organs, or (hardwired or flexible) neural connections, or some kind of inference, or the cooperation of Martians or angelsall this is irrelevant to what it is like to have O-states and I-states, so long as the process (if any) is itself hidden from introspection.
Against inner sense and other "Theater" theories of consciousness, Dennett objects:
My inner sense theory requires no phenomenal gland. As I described in response to the previous accusation, I-states live among O-states, in the various sensory modalities themselves. (These are the same neighborhoods where Dennetts "multiple drafts" live.) Counting by gross modalities, I would say, there are several observers distributed around the brain. Counting by processes that determine the specific contents of individual I-states, there are perhaps thousands upon thousands of microobservers. I mean for these claims to be empirically testable eventually, of course, given neuropsychological advances aided by introspection and by an appreciation of the various philosophical theories of consciousness. Since on my view I-states accompany O-states, it should not be surprising that current, unaided neuroscientific techniques are unable to detect them as I-states rather than as further O-states.
In the absence of direct neuroscientific evidence for or against inner sense, Dennett turns to indirect "engineering" considerations. He argues that inner sense would waste precious cognitive resources, if it involves anything distinct from mere judgments about experienced events. For example, he considers the "phi phenomenon" of vision, in which a flash of light is followed very rapidly by a second, nearby, flash, and one seems to see, not two separate points of light, but a single light moving from the first spot to the second. On any view, ones visual or cognitive system must somehow (consciously or unconsciously) register the second point of light, and only then jump to the mistaken conclusion that there was light at the intervening points (without the registration of the second flash, we would have no good explanation of the direction of the concluded motion). A natural idea for defenders of inner sense is that one doesnt merely conclude that there was motion, because in addition representations of the intervening spots are "filled in" so that the entire path appears lit. Dennett objects as follows:
There are three resources that Dennett says would be wasted by inner sense: discriminatory effort, time, and representational media (a defender of inner sense need not posit mental "paint"neurons will do nicely, so long as they realize states that are innerly sensed). I respond to these charges more fully in "Qualia!," but I will summarize here:
Though I take comfort in the absence of negative engineering arguments against inner sense, I concede that inner-sense theories would profit from a richer positive account of the engineering value of inner sense.<12>
Scientific investigations of introspection have revealed widespread "confabulation" in self-access. In trying to identify ones beliefs and motivations, one systematically but sincerely reports attitudes one thinks rational or statistically normal in the circumstances, even if one doesnt have them (see Nisbett and Wilson, 1977). Lyons argues on this basis that introspection is more a matter of theoretical inferences than a matter of inner sense:
There is little room to doubt that introspection is often theory-laden in this sense. But the only question for the present inner-sense theory is whether introspection of phenomenal properties is theory-laden. The model of rationalizing or statistical guesswork does not extend easily to accessing ones own current qualia. For example, untutored subjects offer consistent and apparently reliable reports of stinging (rather than throbbing) pain "feelings" when a limb has restricted blood flow. It is implausible to suppose that they are inferring these feelings in anything like the way one might form beliefs about anothers pain feelings, since there are no commonsense principles of rationality according to which one should feel stinging rather than throbbing, nor need the subject know any relevant statistical information about how people do feel in these circumstances.
Rosenthal describes a candidate case of phenomenal confabulation:
On the contrary, confabulation is a very extreme diagnosis. There is a much simpler explanation. As Rosenthal says, over time "we retain the visual information" from a large number of fixations between eye movements. This means that at each moment, while we newly represent details only from a small region, this new information is fit into a structure of old (well, moments-old) details about surrounding regions. At each moment we do have detailed visual representations of (nearly) the entire visual fieldmore detailed, anyway, than could be explained by the new information alone. The surrounding detail is not confabulated, but, in a sense, visually remembered. The only rub, apparently, is that visual memory does not flag the information as old, so we see surrounding features as present. In a normal environmentas opposed to a psychologists labthis leads to little error, and to virtually no error that matters practically. Rosenthal misleads when he says that we are aware of little visual detail. We are aware of little new visual detail, but we represent (as current) much more detail.
Dennett argues that one can easily be wrong about the "changes and constancies" in ones experience over even brief intervals of time (1988, 59). This is evidence against especially reliable memory access to what it was like to have past experiences, though it leaves open the possibility of especially reliable (or even infallible) access to ongoing experiences (for a fuller response to Dennetts argument, see "Qualia!").
An inner-sense theory needs an account determining which properties of mental states are innerly sensed. Plaintiffs against inner sense have emphasized the implausibility of innerly sensing the contents of mental states. As Shoemaker remarks:
This presents a difficulty for inner-sense accounts of introspection. Shoemaker quotes Paul Boghossians (1989) query about how anyone could be "in a position to know his thoughts merely by observing them, if facts about their content are determined by their relational properties."<14> Shoemaker emphasizes that this point is not restricted to propositional attitudes:
I agree that inner-sense cannot (simply and directly) account for introspective knowledge of content properties. I advance inner sense not as a general theory of introspection but as a theory of one specific kind of introspection, the kind (I claim) that explains phenomenal consciousness. I-states represent features of O-states, but they do not represent the representational relations of O-states. These are not plausible candidates for being innerly sensed, since they are relations between O-states and other things that are not innerly sensede.g., environmental objects.
What other features of O-states might I-states represent? Well, what other features do O-states have? Instead of O-state content, I-states might represent formalintrinsic, physicalproperties of O-states. These could include abstract "syntactic" features of O-states, and even features of O-states specific to their "hardware" realization, such as the rough number of neurons that realize an O-state, or their rough average rates of firing. (I do not assume that I-states involve complex descriptive concepts of neural structure or of psychological function; another possibility is that they respond to specific neural or psychological features in more primitive ways, just as in outer sense one may be sensitive to wavelengths or molecular motion without having concepts of waves or molecules.) As an illustration, consider Christopher Peacockes example of seeing two same-sized trees, at varying distances from one:
As I described in Accusation 3, Harman and Tye say that we represent the trees as having different "sizes from here." I am skeptical of this kind of view (see section 1 of "Illusions"), and I think the inner-sense account presents a viable alternative (one compatible with functionalism, representationalism, and other views dear to Harman and Tye). Peacocke argues that "you simply enjoy an experience which has the feature" of different sizes-in-the-visual-field. However, it is not simply that the experience "has" this visual-field feature, in the way it might "have" the feature of being realized in, say, molecules. In addition, one is normally sensitive to an experiences visual-field features (in a way one is not normally sensitive to its being realized in molecules). An inner-perception account can explain the visual-field differences in Peacockes two experiences, as well as ones sensitivity to these features, as follows. Compared with O-states about the distant tree, O-states about the nearer tree are realized by (or causally connected to) many more O-states in retinotopic maps in the early visual system. This is just the sort of property to which inner sense may make one sensitive.
I-states may also represent interrelations among O-states, relations that cannot be reduced to their intrinsic (formal) monadic properties. These interrelations are better candidates for inner sense than O-state content, because their relata are all innerly sensed. As illustrations, consider cases of double vision and blurred vision. Typically we are sensitive to the doubleness or blurriness of such experiences. In double vision, we may innerly sense of two O-states (say, two matching perceptions of an edge) that they are two in number. In blurred vision, we may innerly sense of a certain O-state (say, a perception of an edge) that it is in a causal interrelation with an unusual set of other O-states (say, states that involve neurons unusually and poorly "lined up" in the retinotopic maps in primary visual cortex). In normal focal vision, by contrast, we are typically sensitive to the nondoubleness and nonblurriness of our experience. This could be explained by our innerly sensing related structural features of O-states. We detect of an O-state of an edge that it has no distinct matching O-state, and that it is in a causal interrelation to a standardly "lined-up" set of other O-states.
In addition to content features, there is another kind of feature that makes something an O-state: the functional relations that characterize modalities such as vision and hearing. But for the same reason that I-states do not represent the contents of O-states, it may not be plausible that an I-state represents the modality of an O-statewhether it is visual, auditory, etc. At least in one "wide" sense, modality features are also in part relations to entities that are not innerly sensede.g., sense organs. Instead of the wide modality of O-states, I-states might represent the force of O-states, which I think is or includes their degree of attentiveness. Perhaps this is reflected in O-state form, somehow. Attentive states may be realized in more rapidly firing neurons, for example. (Also, commitment to wide modality properties need not preclude commitment to "narrow" modality propertiesperhaps in another sense an O-states being visual or being auditory can be explained in terms of its form, force, or relations to other O-states. If so, inner sense might be sensitive to narrow modality.)
People v. Phenomenal Necessity of Inner Sense
Lyons cites the following result as a consensus among developmental psychologists:
The worry is this: if children don't introspect, and if a kind of introspection helps constitute phenomenal experience, doesn't this mean that children lack phenomenal experiences?<15> For this and many other reasons it is crucial to distinguish between attentive inner sense and inattentive inner sense. Inner sense is required for phenomenal consciousness, but this inner sense need not be attentive. Inner sense of any sort produces I-states about specific O-states. But attentive inner senseand much else besidesmay normally be required for the development of general introspectively-based concepts of mental states. The age of inner sense predates the age of introspectionconstrued as involving inner-directed judgment and/or reports.
McGinn raises a similar objection with respect to many animals:
It is perhaps true that animals have their attention glued to the external world, but this does not show that their awareness is glued thereperhaps they have inattentive awareness of the internal world. Furthermore, on the inner sense view this awareness need not be a matter of having (luxurious) "thoughts" about mental states, but instead can be a matter of having more primitive sensings of mental states.
Sometimes arguments for this conclusion rest on the same conflations that undermine the previous accusationequating awareness in general with attentive awareness (thereby forgetting about inattentive awareness), and with judgment (thereby forgetting about sensing). Chalmers, for example, argues:
An inner-sense account of experience is not committed to continual noticing of (or attention to) ones experiences, and is not committed to continual judgments about ones experiences. Strictly speaking, it is not even committed to continual inner sensing of ones experiences. The view is that one has continual inner sensings, inattentive or attentive, in ones experiences (see Accusations 1 and 2).
Hills argument in support of the present accusation runs as follows:
The "clarity" of these examples depends, I think, on an unclarity about the term "sensation." The crucial question is whether sensations are necessarily phenomenalon whether there is necessarily something it is like to have a sensation. Hill begins his book by characterizing sensations as "concrete sensory events" (1991, 3); by the Nonphenomenology of Perception, this indicates that sensations need not be phenomenal. But he also sometimes uses the term "sensations" in such a way as to suggest that sensations are phenomenal by definition. Let us call nonphenomenal sensations "N-sensations" and phenomenal sensations "P-sensations." It is clear that there can be N-sensations of which one is not aware in any wayI have been calling these "O-states" all along. But this does not mean that there can be P-sensationsphenomenally conscious sensory statesof which one is not aware in any way. Again, I grant that one need not form beliefs about ones P-sensations, and that one need not attend to them. But Hills examples do not provide evidence that inattentive inner sense is unnecessary for phenomenality.
Dretske has produced a novel argument to the effect that there are (phenomenally) conscious features of experience of which the creature is unaware in any way. The argument begins with a distinction and two definitions:
(1) All awareness comes in two kinds: being "thing-aware" and being "fact-aware."
(2) Perceiving an object or event (or state) x = being thing-aware of x.
(3) Perceiving or judging that p = being fact-aware that p.
(Dretske does not discuss whether the perceiving must be experiential, or whether wholly subliminal, nonphenomenal perceiving is sufficient for thing-awareness or fact-awareness.) He then diagrams two side-by-side "constellations" of black, irregularly shaped spots, labeled "Alpha" and "Beta." His instructions are for the reader to "[g]lance long enough to assure yourself that you have seen all the elements" of the two constellations, encouraging the reader to "change fixation points in order to foveate (focus on the sensitive part of the retina) all parts" (1993, 272-273). The two constellations differ in that Alpha contains a single spot that Beta lacks, an easily visible spot of roughly the same size and shape as the rest. Dretske names this "Spot":
(4) Spot = the difference between Alpha and Beta.
Now consider a reader, Reed, who follows Dretskes instructionstaking care to perceive, focally, every part of Alpha, including Spotbut fails to become aware that Alpha and Beta differ. (This seems to be the normal case.) So, by stipulation:
(5) Reed perceives Spot.
From this Dretske concludes:
(6) Reed is thing-aware of the difference between Alpha and Beta.
And also by stipulation:
(7) Reed is not aware that Alpha and Beta differ.
From this Dretskes concludes:
(8) Reed is not fact-aware of the difference between Alpha and Beta.
According to the next step in Dretskes argument, Reeds experiences of Alpha and BetaE(Alpha) and E(Beta), respectivelyare themselves "qualitatively different," in much the way that Alpha and Beta are different. The basic idea is that Spot is itself experienced, and does not merely lie in the direction of the gaze. If so, E(Alpha) must contain an experiential component, caused by Spot, that E(Beta) lacks. Dretske calls this component "E(Spot)":
(9) E(Spot) = the difference between E(Alpha) and E(Beta).
Given (7), it is also plausible that:
(10) Reed is not aware that E(Alpha) and E(Beta) differ.
From this Dretske concludes:
(11) Reed is not fact-aware of the difference between E(Alpha) and E(Beta).
Dretske marshals further arguments (see below) for the claim that:
(12) Reed is not thing-aware of the difference between E(Alpha) and E(Beta).
Given (1) and (9), Dretske concludes from (11) and (12) that:
(13) Reed is not aware of E(Spot), in any way.
This means one can have a conscious experience (or conscious element of experience) of which one is wholly unaware.
I think Dretskes argument rests on a fallacy of ambiguity, ambiguity in the locution "fact-aware of the difference between ." I will try to explain the problem informally, by giving examples of fact-awareness that Dretske does not block. Then I will try to root out the ambiguity more formally.
I agree with Dretske that Reed is at no time fact-aware, about Spot, that it is a difference between Alpha and Beta. But this is only one particular fact about Spot. There are many other facts one could represent. Consider the moment when, during his scan of Alpha, Reed (by hypothesis) foveates Spot. He plausibly has several bits of fact-awareness about Spotthat it is there, that it is black, that it is a spot, etc. Otherwise Reed would not be so confident that Alpha is entirely composed of black spots. It need not be that Reed judges (or believes), about Spot, that it is there (is black, is a spot, ). Dretske sometimes writes as if fact-awareness requires belief-like representation, and as if sense-like representation is restricted to thing-awareness. But this identification is incorrect. For example, just as one can believe that something is Fusing a "concept" of being Fone can sense that something is Fperhaps using a "percept" or "sensory mode of presentation" of being F. And, again, this sensing need not be attentive. Nothing Dretske says counts against this kind of inattentive outer-sensory fact-awareness about Spot. Further, if Reed has fact awareness about Spot, then by (4)since Spot is the difference between Alpha and BetaReed has fact awareness about the difference between Alpha and Beta. He is fact-aware that a certain object is there (is black, is a spot); and this object is the difference between Alpha and Beta. Strictly speaking, (4) it is not quite correctSpot is not "the" difference but "a" difference. Spots color (black) is also a difference. Even if all Reed sees is that blackness is theresomehow without representing Spot as any kind of objectthis is enough to have fact-awareness about a difference between Alpha and Betaabout Spots color. On almost any theory of visual representation (at least, any theory compatible with wholly unconscious sensings), such a fact is sensed when Reed foveates Spotsensing of such facts is present (or at least begun) in the retina, long before the visual individuation of objects.
The same points apply to fact-awareness about E(Spot). Reed of course lacks the particular bit of fact-awareness, about E(Spot), that it is a difference between E(Alpha) and E(Beta). But he may have much other fact-awareness about E(Spot)for example, that it is there, that it is like this, that it is an experience, etc. As with outer fact-awareness about Spot, this inner fact-awareness about E(Spot) does not require belief-like representation. It is compatible with inattentive inner sense about E(Spot). The facts about E(Spot) that are innerly sensed may include facts about E(Spot)s form, relations, or force (see Accusation 14)that Q-ness is present, for some such feature. Dretskes argument fails to show that an experience can have phenomenal features of which there is no awareness at all. Inattentive inner-sensory fact-awareness is perfectly safe from Dretskes criticisms.
More formally, the central fallacy affects three transitions in Dretskes argument: from (7) to (8), from (10) to (11), and from (11) and (12) to (13):
(7) Reed is not aware that Alpha and Beta differ.
(8) Reed is not fact-aware of the difference between Alpha and Beta.
(10) Reed is not aware that E(Alpha) and E(Beta) differ.
(11) Reed is not fact-aware of the difference between E(Alpha) and E(Beta).
(11) Reed is not fact-aware of the difference between E(Alpha) and E(Beta).
(12) Reed is not thing-aware of the difference between E(Alpha) and E(Beta).
(13) Reed is not aware of E(Spot), in any way.
These inferences trade on an ambiguity in the following schema used in (8) and (11)where r, a, and b are schematic terms:
(F) r is not fact-aware of the difference between a and b.
This sort of locution is not formally introduced by (3), in terms of "r is fact-aware that p." For (8) to be supported by (7), and for (11) to be supported by (10), (F) must mean the same as:
(F˘ ) r is not aware that there is a difference between a and b.
(7) and (10) do, respectively, trivially support the following:
(8˘ ) Reed is not aware that there is a difference between Alpha and Beta.
(11˘ ) Reed is not aware that there is a difference between E(Alpha) and E(Beta).
The trouble is, (11˘ ) and (12) do not support (13). (11˘ ) does not rule out all kinds of fact-awareness about E(Spot); it merely rules out awareness of the specific fact, about E(Spot), that it is a difference between E(Alpha) and E(Beta). For (11) and (12) to cover all relevant ways of being aware of E(Spot)so that together with (1) they support (13), (F) must mean something broader than (F˘ ). To derive (13), (F) must mean the same as:
(F˘ ˘ ) r is not aware that p, for any p that is about any x that is a difference between a and b.
(13) is supported by (12) and the following:
The trouble is that (11˘ ˘ ) is not supported by (10). The upshot: if (F) is used as (F˘ ) then, we may grant, the argument proceeds fine through (11). But even if (12) is also granted, (13) would not follow. On the other hand, if (F) is used as (F˘ ˘ ), then (8) doesnt follow from (7), and (11) doesnt follow from (10).
I could end here, but for completeness it is worth considering Dretskes argument for (12), his argument against inner-sensory thing-awareness of E(Spot). He begins by crediting Rosenthal with "a convincing case against inner sense," describing the argument I discussed above as Accusation 2, and citing pages in which Rosenthal advances many arguments I address elsewhere in this paper (see Accusations 6, 8, 10, 11, and 13). Dretske adds only the following considerations:
This argument rests squarely on the prior argument against fact-awareness of Spot and E(Spot)the arguments for (8), (11), and (13)and collapses if, as I have argued, the prior argument does. Dretske has no good argument that the viewer lacks all fact-awareness of Spot and E(Spot)about which intrinsic properties they have, for example; all he has is an argument that the viewer lacks the fact-awareness that Spot and E(Spot) have certain complex comparative relations: that they differentiate Alpha and E(Alpha) from Beta and E(Beta).
Hill considers the following natural formulation of an inner-sense model:
Hill argues against this view by providing examples of ways that attending to ones sensations can change them. He calls the first "volume adjustment," in which, by increasing ones attention to a sensation, one increases either its "intensity" or its "internal articulation." He illustrates these as follows:
The second effect of attention is what Hill calls "activation," in which one seems to bring a sensation into existence, rather than merely modifying one:
A theory of phenomenal consciousness should explain how it is possible for attention to have these effects on its object. Hills challenge is this: how can inner sense explain this, if it is to be analogous to noninterfering outer sense?
Some outer sensing does change (or interfering with) its object. Feeling an object with ones fingertip changes the object as well as the finger. The sense in which feeling is more intimate than seeing may help explain why we more naturally speak of "feels" than of "looks" in discussing phenomenal consciousnesswe speak of "raw feels" rather than "raw looks" (even in the visual case!), and we substitute "what it feels like" for "what it is like" more naturally than we substitute "what it looks like." But it would be a missed opportunity for the inner-sense theorist to resort to outer feeling to preserve the analogy with outer sense. I think Hills cases show no disanalogy with outer vision, and, more importantly, add crucial support to an inner-sense theory.
As I described in response to the previous accusation, Hills term "sensation" can be taken in either a nonphenomenal or a phenomenal sense. N-sensations are "concrete sensory states"O-statessuch that, by the Nonphenomenology of Perception, there is nothing it is like to have them. P-sensations are full-blown sensory experiences, phenomenal by definition. Care with respect to this ambiguity enables inner sense to address the current accusation. Inner sense may be noninterfering, like outer sense, in that it does not change (the internal properties of) N-sensations. At the same time, inner sense may explain volume adjustment and activation, because it does change P-sensations. Even if I-states do not change their objects, O-states, they may change states of which they are a part, experiences.
Consider first how inner sense explains activationthe "dawning of a new sensation" "when one turns ones attention to a new region of phenomenal space" (1991, 126). Hill claims that activation is "flatly incompatible" with the inner eye hypothesis (1991, 126). However, I-states are tailor-made to bring new P-sensations into existencewithout I-states, there are no experiences, but instead only O-states. Just as shifting outer-sensory attention to some location can generate O-states of formerly unperceived objects at that location, so shifting inner-sensory attention can generate I-states of formerly unperceived O-states. This "activates" a phenomenally conscious experience (a P-sensation) where there was none before: in keeping with the Nonphenomenology of Perception, O-states alone are not phenomenally conscious.
Can Hill argue that there is activation of N-sensations as well as of P-sensations? He does not do so, and in fact what he does say weighs against the claim that N-sensations are brought into existence by introspection:
With a few amendments to make this unconscious processing seem less intellectual, I would map this theory-sketch onto my framework as follows. Hills "packets of information" are my "O-states" (or "N-sensations"). Hills "sensations" (in this passage) are my "experiences" (or "P-sensations"). Hills "appropriate information-processing mechanisms" are the processes that form my "I-states." Given Hills points, then, I-states operate on preexisting O-states to form experiences. There is no attentional "activation" of N-sensations (O-states); and although there is attentional activation of P-sensations (experiences), this is explained on the inner-sense theory. All this is compatible with the claim that I-states do not interfere with their objects, O-states. Activation does not display a disanalogy between inner and outer sense. On the contrary, it provides data that support my inner sense account.
Next, consider how inner sense can explain volume adjustmentincreases in the "intensity" and "internal articulation" of a sensation. Unlike activation, volume adjustment concerns changes to a preexisting experienceto a preexisting O-state/I-state pair. By attending to the O-state, one changes the content of the corresponding I-state (i.e., one "notices" something different about the O-state). Just as increasing outer-sensory attention to an environmental object can increase the detail or strength of ones O-states of the object, so increasing inner-sensory attention to an O-state can increase the detail or strength of ones I-states of the O-state. This "adjusts the volume" of the qualia, changing the "internal articulation" or "intensity" of the experience. On the view I have proposed, changes in the content of I-states change the "qualia" of experienceswhat it is like to have an experience is determined by the contents of its component I-states. (There need be no change to the "internal" properties of O-states, so this does not compromise the idea that inner sense is noninterfering.) Furthermore, on independently plausible assumptions about the contents of I-states (see Accusation 11), the inner sense account explains the detailed changes in intensity and internal articulation of experience. Changed I-states representing the form and interrelations of O-states may explain changes in Hills "internal articulation"level and clarity of detail, level and salience of gestalt organization, and so on. Changed I-states representing O-state force may explain changes in Hills "intensity"sound-volume, taste-vividness, pain-severity, itch-importunity, pressure-strength, and so on. Again, evidence that initially seems to threaten the inner sense view in fact confirms it.
Hills volume-adjustment cases are similar to the often-remarked changes in experience that come from concentrated exposure in a perceptual domain such as wine-tasting or music-listening. Inner sense provides a plausible explanation of these changes as well. Such changes can occur quickly, as Dennett illustrates:
We would like an account of what constitutes such changes in experience: in what way does the experience (or the qualia) get more complex? And how do these changes result from long-term or short-term "training"? Although training may build more complex concepts (of the sort that figure in judgments), these do not seem sufficient for changing experience. One can acquire new complex concepts about wine and music simply by reading books, for example. On the other hand, if it is suggested that some concepts can be acquired only by tasting wine (or wine-like stimuli) or listening to music (or music-like stimuli), why shouldnt we think the tasting or listening changes experiences directly, rather than by detouring through concepts? It seems implausible to suppose that the listener in Dennetts example needs to acquire a new concept, rather than some kind of temporary auditory modification. But although training may build more complex outer-sensory discriminatory abilitiesbigger and better O-statesthis does not seem to account for all such changes (e.g., in Dennetts example, the sounds "complexity" is discriminated "all along"the O-states do not become more complex in content). If the perceptual modifications are not (always, only) due to changed concepts and O-states, what other kind of modification could there be?
The inner-sense account provides a modifiable level of representation besides concepts and O-states: namely, I-states. Training in wine-tasting or music-listening may change experiences, not merely by developing ones theoretical concepts, and not necessarily by changing ones outer-sensory discriminatory abilities, but by changing ones inner-sensory sensitivities to ones outer-sensory states. I agree with Dennett that "the complexity apprehended in the third plucking was there all along (being responded to or discriminated)," but this is only to say that the O-statesnot the I-stateson the first and third plucking are constant and complex. On my account, the change in qualia is due to a change in I-states. The I-states on the first plucking may represent the complex O-states as simple, while the I-states on the third plucking represent these same complex O-states as complex.
In my responses to Accusations 15 and 16, I described ways in which philosophers have failed to apply the distinction between inner-directed sense (whether attentive or inattentive) and inner-directed attention. Drawing this distinction is compatible, however, with maintaining that inner-directed attention is a kind of inner-directed sensing, as illustrated in the discussion of the previous accusation. Perhaps one way of attending to perceptual states is by having attentive inner sensings of these states. Lyons rejects a related view:
This argument is a bit too sweeping. The best way to view something attentively is to focus ones gaze on it. But this cannot handle all cases of attending to vision. Normally one is unaware of the fuzziness of ones visual periphery (see Accusation 13), but one can willfully "attend to" the periphery and thereby notice the fuzziness with which one sees things there. This must be different from shifting ones gaze to the periphery and so removing the fuzziness. Similarly, in degraded cases of vision, such as crossed-eye vision, it does not seem difficult to attend to the degradationbut it is the experience, not the environment, that seems degraded. Nevertheless, in general Lyons is right that it is at least very difficult to redirect attention from environmental objects to experiences of them.
A more general defense of my inner-sense view emphasizes that I-states are directed at O-states, not (except derivatively) at full-blown experiences. Accordingly, it is not a commitment of the present view that one can attend to full-blown experiences, much less that one can do so regularly, easily, or on command. I think a process can be "sensory" even if it is incapable of supporting attention, but others may disagree. If they are right, then my view does have a commitment to the claim that one can attend to O-states as well as to environmental objects. Whether there is such attention is not a matter to be settled purely introspectively, but by a combination of introspection and psychological theory. The cases described in the previous accusationHills activation and volume-adjustment, wine-training and music-training, Dennetts guitar caseprovide some evidence that there can be attentive inner sense of O-states, distinct from attentive outer sense of environmental objects.<18>
A similar response applies to the following informal experiment proposed by Shoemaker:
The present view is not committed to attentive inner sense of full-blown experience. Even if it were, it would not be committed to the idea that attention can be redirected at will. Outer attention is not always at our command. Raise both of your hands before you, and have a friend set your left one on fire. Now shift your attention to your right hand. Can you do so? Probably not, because your representation of your left hand captures your attention at the expense of your right hand, for excellent natural-selective reasons. Similarly, for excellent natural-selective reasons, environmental objects capture attention, at the expense of mental states. This does not show that mental states are (inattentively) unsensed, any more than the burning-hand case shows that ones right hand is (inattentively) unsensed.
Lyons considers and rejects a distantly-related response to his initial argument:
Although I think there is attentive inner sense that is distinct from attentive outer sense, it is no part of my view that in a single case one can have attentive inner sense while also having attentive outer sense. (On the contrary, it is crucial to my overall view that inner sense and outer sense are rivals for attentionsee Accusation 21, below.) Being a "spectator" does not require being an attentive spectator. Nothing Lyons says even begins to militate against the inner-sense view.
Shoemaker presses another alleged disanalogy between introspection and outer sense:
It may be that the force of this accusation turns on an ambiguity of the term "perspective." In a broad sense, a "perspective" is a way of representing something, and in a narrow sense, a "perspective" is a "spatial perspective," a perceptually relevant spatial relation between perceiver and perceived. I think one can introspect and innerly sense one and the same mental state in different ways, with different broad perspectives. For example, Dennetts guitar-plucking case, discussed in response to Accusation 17, is arguably a case in which one-and-the-same O-state can give rise to different I-states, and consequently to different experiences. It is more plausible that inner sense lacks varying spatial perspectives. This is unsurprising, because inner sense does not represent spatial properties at all.<19>
Strictly speaking, spatial perspectives are not required for outer sensing, especially in outer-sense modalities that do not represent spatial properties. Tasteas distinguished from tactile sensation via the tongueis a good example. We cannot taste where something is, and equally we cannot adopt different spatial perspectives in taste. (I leave aside the fact that different spatial regions of the tongue respond selectively to different flavors. For simplicity, imagine a tongue with varying taste buds intermingled uniformly, or imagine a tongue containing only sweetness receptors.) The difference Shoemaker identifies between inner sense and (some kinds of) outer sense is, at root, a difference in subject matter. But of course, inner-sense theorists are interested in pursuing an analogy with outer sense only modulo differences in subject matter. We would turn a deaf ear to a plaintiff who argued that inner sense has an inner subject matter and sosurprise, surpriselacks a crucial feature of outer sense.
People v. Phenomenal Sufficiency of Inner Sense
20 Why wouldnt inner sense simply be inner blindsight?<20>
Blindsight subjects have portions of their visual field"blindsight regions"in which they (plausibly) deny that they have conscious experiences of specific stimuli. However, in some sense they have sensory states sensitive to these stimuli"blindsight states." The key diagnostic feature of their condition is that these blindsight states lack substantial influence on central processes of belief and desirethe subjects dont believe they are having experiences or sensory states of any sort, they are unable to report on what they perceive, they must be coaxed into responding, and when they do respond, they say they are only "guessing." It is difficult to see how these features alone could utterly "turn off" qualia. Perhaps they explain why blindsight states are like nothing for the person as a whole, but they do not seem to explain why blindsight states are like nothing, period. Something else seems to be missing; on the present view, the absence of inner sense, whether centrally available or not, explains the lack of qualia.
In short, there can be inner blindsight, in the sense that there can be centrally unavailable inner sensing. But this does not present a difficulty for the inner-sense theory of phenomenal consciousness. Innerly blindseen states are (or would be) phenomenal (i.e., like something), but not phenomenal for the person as a whole. This proposes to elevate, above the realm of clashing intuition, the controversy about whether or not "normal" (outer) blindsight is phenomenal. Blindsight states are not innerly sensed in a centrally available way. If they are not innerly sensed at all, they are nonphenomenal. If they are innerly sensed in a centrally unavailable wayif they are innerly blindseenthey are phenomenal but not phenomenal for the person as a whole. In principle we can discover the relevant ins and outs of blindsight.
Dennett provides a complicated argument that the distinction between conscious and unconscious states is vaguethat some psychological states are neither determinately conscious nor determinately unconscious. He takes this result to weigh against what he calls the "Cartesian Theater," including any view according to which inner sense helps explain the conscious/unconscious distinction. As I argue more fully in "Qualia!," vagueness is not an objection to an inner-sense theory of consciousness, because it is independently plausible that inner sense is vague. After all, outer sense is vaguethere is vagueness in the distinction between mere reflexive reactions to stimuli and (even unconscious) sensory representings of stimuli. In "Stopgap" I also discuss briefly the vagueness in the notion of "central availability"how central is central, and how available is available?
Perhaps this is to jump out of the frying pan and into the fire, however, if the following argument of Chalmers is correct:
We need to distinguish two respects in which conscious experience can be "faint": what is had can be faint, or the having of it can be faint. Chalmers claim is plausible on the first readingeven if there is something-faint it is like to be a mouse, there is something it is like to be a mouse. But perhaps instead there faintly-is (i.e., indeterminately is) something it is like to be a mouse, in which case it would neither be true nor false that there is something it is like to be a mouse. The possibility of indeterminately-having qualia is opened up by the vagueness of "central availability," and by the possibility that a state may neither be determinately an inner-sensory state nor determinately a mere inner reflex reaction, but somewhere in the vague between.<21>
This accusation arises in many disguises.<22> Kirk issues the following explanatory challenge:
Dretske objects similarly to the idea that sensing a statebeing "conscious of" the state, in one senseis sufficient for a state itself to be conscious (phenomenally or otherwise). Unlike Kirk, he focuses on states closer to home:
To further tighten the screws Dretske appeals to unusual perceptions of brain states:
Dretskes talk of "higher-order experiences" and "lower-order experiences" seems to embody his commitment to Accusation 1about experiences of experiences. Given the response to that accusation we may reconstrue his terms as "I-states" and "O-states," respectively.
There is a single key to answering each of these complaints: a state (or event) is "like something" in the strongest, phenomenal sense only if it is necessarily like something, in order to exist at all. There is something it is like for one to have a certain state only if there is something it is like simply for one to have that state. (Compare: there is normally something it is like to fall off a cliff, but there is not something it is like simply to fall off a cliffwhile in a coma, say. For this reason, falling off a cliff is not itself a phenomenal state.) Given a sensory construal of "being like something" as "appearing some way"a construal I motivate at length in "Stopgap"this means that a state is like something only if it necessarily appears some way, in order to exist at all. In short, phenomenal states are subjective states in the sense that for them to be is for (components of) them to be sensed. This obtains for experiences, on the inner-sense view, because they essentially involve ones sensings of (components of) them. The inner sensings are necessary in order for the experiences to exist. By contrast, in Kirks case, outer sensings are not necessary for "external events" or "external objects" to exist. (All bets are off if idealism is truemaybe then "external" objects such as bananas do have phenomenal properties and "external" events do involve conscious experience.) Similarly, in Dretskes "internal state" cases, heartbeats and the rest exist independently of being sensed. In Dretskes final cases, the subjects brain states also exist independently of someone elses sensing them via brain scans. The kind of inner sense required for phenomenal consciousness is built into the experience innerly sensed. (So Kirks wondermentthat "the" perception of internal events necessarily involves experienceis misdirected from the start. Only a very specific kind of perception of internal events is necessary for experience.)
A fuller answer to the accusation stems from the conditions necessary for generating the phenomenal illusions"image illusions" of phenomenal objects in ones mind, and "transparency illusions" of phenomenal properties in ones environment. The illusions are generated when O-states and I-states undergo binding confusionswhen they are mistakenly applied to the same object. I-states and O-states interact to form the illusions (with or without external entities). But no such illusions are generated when O-states (without I-states) "interact" with external entitiesbananas, heartbeats, others brain states, etc. Bananas and heartbeats are not representational entities at all, and so are not co-applied with O-states about them. Submitting to cross-examination:
Books, bananas, and heartbeats are not sensory states at all, and so do not function in a unified way with O-states about themfor instance, they are easily distinguished introspectively. The same is true of sensory states in others brains, states of subjects different than the one who has O-states about them.
Again, these brain states would exist independently of his sensing them, and would not be bound with his sensings of them or function in a unified way with his sensings of them.
Such a case is difficult to describe coherently, but even if it is possible, it would not generate a binding confusion. For states to be confused, they must be distinctas O-states and I-states are.
For two states to produce the illusions, they must exist at the same time (although they can have different onsets or lifespans)as O-states and I-states do.
Okay, thats a harder case, requiring discussion of further conditions that crucially shape the illusions. So far I have described only constraints on the confusedly bound states themselves; there are also constraints on the features represented by the confusedly bound states (I discuss all of these constraints in more detail in "Illusions," section 2).
First, the image and transparency illusions are normally complementary, varying according to whether attention is focused inwardly (by giving priority to I-states) or outwardly (by giving priority to O-states). For attention to be the determining factor, there must be a deep attentive rivalry among features represented by I-states and O-statesit must be difficult simultaneously to attend inwardly and outwardly. It is difficult to attend to an imagined banana while also attending to ones (alleged) mental image of the banana. It is difficult to attend to the shape of an object while also attending to the (alleged) look of the shape. By contrast, the features represented by O-states are rarely rivals for attentionit is normally easy simultaneously to attend to the shape and reflectance, or shape and sound, or flavor and texture, etc., of an object.
Second, there are normal epistemological differences between "outer" and "inner" properties that are crucial to explaining what seems strangely ghostly or private about mental images and the varying looks and feels of a constant shape. A subject need not (and normally does not) perceive or think of them as mental, and he need not (and normally does not) perceive or think of them as being in some strange space. Rather, what is ghostly or private about looks, feels, and mental images is that they are normally only accessible monomodally, by one sense modality in one subject. The objects and properties represented by O-states are, by contrast, normally accessible polymodally. We are disposed to test our visual O-states of shape against tactile states, and to test our visual O-states about color-reflectance against (reports of) the visual states of others. But we have no natural dispositions to test I-states against the verdicts of other senses or other perceiversunderstandably so, since in the circumstances that have shaped our natural dispositions, the neural or functional features represented by I-states are only accessible monomodally. In the image illusion, then, we have no independent dispositions to test the identifying properties of alleged images, the properties innerly sensed. In normal perception, objects do not seem monomodal like images, because we do have independent dispositions to test their identifying properties, the properties outerly sensed. Instead, some of their alleged properties seem monomodal, the properties innerly sensed but projected outward.
Now consider how these points apply to the residual putative counterexample, involving outerly perceiving ones own immediately preceding, but persisting, outer perception. In this case, there are two distinct O-states, one that begins first and persists, and another that results from outer perception of it. Call these "O1" and "O2," respectively, and suppose they represent properties P1 and P2, respectively. Further, suppose they each use the same variable or term, "x." Then O1 represents that x is P1, and O2 represents that x is P2. Since these are temporally overlapping states, unless the case requires total breakdown in processingrendering it too unclear to carry much weight as a counterexamplethe system should treat these features P1 and P2 as compatible with one another, as color and shape are mutually compatible. (Otherwise, if P1 and P2 were mutually incompatible, as squareness and roundness are, O1 and O2 would have to compete with one another for credence.) Either P1 and P2 are the same feature, or they are not, depending on how the case is elaborated. But if they are one-and-the-same, then "they" cannot be rivals for attention and "they" cannot differ in monomodality and polymodality. Even if they are distinct features, they are all (apparently) compatible, outer-perceptual features, and so they should not be rivals for attention any more than shape and size are, and should not be treated as monomodal any more than shape and size are.
It matters to the phenomenal illusions that inner sense is inner rather than outer, in the sense that its subject matter is represented in a manner quirkily unlike normal, outerly sensed, subject matterin a monomodal manner (so that the innerly-sensed subject matter seems private or ghostly), and in an attentively rival manner (making it difficult to attend to the private or ghostly subject matter in addition to the normal subject matter). On my account this is why inner sense can be sufficient for phenomenality even though outer sensewhether of bananas, heart beats, books, or brain statesis not.
Perhaps the most important feature of my inner-sense account, for addressing a wide variety of the accusations, is:
The keys to answering most phenomenologically-based accusations against the existence of inner sense are (a) and the following:
Two further features, together with (a), block most scientifically-based accusations against the existence of inner sense:
Accusations to the effect that there is too little inner sense to explain all phenomenal experience are answerable by (a) and the following features:
Finally, these two features help to deflect arguments to the effect that inner sense is insufficient to explain phenomenal experience:
Of course, other considerations are needed to make a positive case that inner sense exists, and that it is necessary and sufficient for phenomenal consciousness. But the prosecution has not shown this to be impossible, or even unlikely.
Ladies and gentlemen of the jury, the facts are clear, and the decision rests on your ability to do what is right. Search your conscience. Is it to be filled with guilt, or with inner sense?
<2> See also: McGinn, 1982, 50-51; Rosenthal, 1990, 34-35; Dretske, 1993, 17; Güzeldere, 1995, 343. Dretske (1995, 176) cites similar views from Dennett, Harman, Shoemaker, Tye, and the psychologist Donald Hebb.
<3> See also Tye, 1995, 30-31 and 135-136. Shoemaker argues that outer sensing "standardly involves" sensing of intrinsic features of environmental objects (1994, 253); in a complicated discussion he flirts conflictedly with using Harmans argument to show that inner sense is disanalogous to outer sense (1994, 264, 301, 310-311). Similar points are often made by those who press Accusation 2.
<4> Other critics have indirectly suggested a connection between inner sense and sense data. Harman attributes the initial plausibility of inner sense to a failure to "distinguish between the properties of a represented object and the properties of a representation of that object," and continues:
Shoemaker argues that "reasons for rejecting the sense-datum theory are at the same time reasons for rejecting the act-object conception [of perception] as applied to pains and after-images" (1994, 262). By the "act-object conception" he seems to mean the reification of pains and afterimages, with all their perceived properties. It should be clear that on the present view there is no commitment to throbbing pain-objects and purple afterimage-objectsthese result from image illusions. But there is no reason to deny the "act-object" conception of inner sense as applied to O-states, or the neural structures realizing them. As I argue below, this carries no commitment to sense data. Shoemakers point shows no disanalogy between inner sense (of neural objects) and outer sense (of environmental objects).
<5> For a contemporary presentation of Ryles argument in the context of criticizing inner sense theories, see Kirk, 1994, 9-10. For a presentation of what appears to be Brentanos argument, see Searle, 1992, 171.
<7> He emphasizes that "the perceptual model of introspective self knowledge has its greatest plausibility" for sensory events, as opposed to propositional attitudes (1994, 261). Rosenthal also argues that there is a special implausibility to inner perception of attitudes, as follows:
Of course, in the end Rosenthal and Shoemaker reject an inner-sense model for sensory states as well as for attitudes.
<8> See also Searle, 1992, 97. This seems to be a standard objection to inner sense in the phenomenological literature (being pressed, instead, in favor of Brentanos reflexivitysee Accusation 5).
I have no disagreement with the idea that introspection (at least often) produces beliefs. I do claim that such beliefs are often caused via inner-directed sensings (see Accusation 7), as opposed to Shoemakers bare "wired" connections. I do not see how Shoemakers view can explain lawful and systematic patterns among introspectible and nonintrospectible states, without an ad hoc assumption, for each pattern, that there happens to be the right "wiring" to higher-order beliefs. For example, why does it seem easier to introspect ones fleeting thoughts than ones deeply held beliefs? After all, deeply held beliefs would have more time to "grow" wired connections to beliefs about them. On my inner-sense account, fleeting thoughts, but not deeply held beliefs, are realized in O-states (typically auditory imaginations of speech; see the previous two accusations). These are subject to inner sense, and thus are easier to introspect than beliefs. Likewise, why on Shoemakers account arent brain states governing autonomic bodily functions (easily) introspectible? There could have been wired connections from these to beliefs, so why arent there? I claim that we lack (easily attained) beliefs about these states because we lack inner sensings of these statesthey are motor or glandular "commands" rather than sensory O-states.
<10> Indeed, the class is far more heterogeneous than a division into "intentional" and "sensory" states would reflect: leftovers include being depressed (but not depressed that p), being forgetful, being funny, being bold, etc.
<12> Questions about the driving forces of and constraints upon evolutionary design are usually very difficult to answer. If asked why we should have inner sense, I would offer a speculation along the following lines: by allowing one to detect certain qualities of outer sense, inner sense allows one to detect the quality of outer sense: whether and how the representing is degraded, whether it is imaginative, whether it is obscured, and so on. This may facilitate behavioral or inferential "corrections," including behavior aimed at improving the quality of the outer sense (shifting position, squinting, etc.). What I would urge now is that the same mystery arises about phenomenality and qualia: why do we have them? Worry about the "functions" of inner sense could be a selling point for it as an account of phenomenality and qualia! Given how difficult it is to understand what functions phenomenality plays, if any, it would be surprising if a philosophical theory of phenomenality appealed to a phenomenon with obvious functions.
<14> Boghossian does not use this point against the existence of inner sense; his concern lies elsewhere. For use of the point against inner sense, see Dretske, 1995, 108-109; Güzeldere, 1995, 349-352.
<15> See also Dretske, 1995, 18, who presses the same accusation as Lyons but lowers the hurdle to "3 or 4 years." With the stakes so high, it matters little whether children become introspectors at eight years old or four years old or eight months old.
<16> Similarly, Lyons says that "[b]ecause our ability to attend to what is pouring into our sensory receivers is severely limited, we can concentrate on very little of it" (1986, 100), and Kirk says that "[m]uch of the time when we are enjoying ordinary conscious perceptual experience, what interests us is the things out there that we can see, hear, smell, and so on" (1994, 153). This emphasis on attention, concentration, and interest betrays the conflation underlying the accusation.
Hill illustrates in a surprising way how easy it is to miss the possibility of inattentive inner sense. Hill never mentions inattentive inner sense, despite considering inattentive inner belief (which he calls "basic awareness") and attentive inner sense (which he calls the "inner eye"). He says that basic awareness is "an essentially passive phenomenon" rather than a matter of "[a]ttending concentrating, focusing, and scrutinizing," and he says it forms "beliefs" rather than sensings about mental states (1991, 117-118). About the inner eye he says that "one attends" to a mental state "by adjusting ones internal scanning device," forming sensings rather than beliefs about mental states (1991, 119).
Hill comments that in Pillsburys case of activation "one notices a phenomenal change; one experiences the birth of an itch." By contrast, in the volume-adjustment case of the itching leg, "[a]t most there is an increase in prominence."
<18> Contrary to Lyons claim, attentive inner sensation would be one kind of "attention to" outer sensing. This is compatible with Lyons proposal that "attending to sensing" is, in another sense, merely an "adverbial modification" of sensing. To my ear, Lyons sense is better expressed as "tending to" sensingdoing the outer sensing carefully. Attentive outer sensingtending to outer sensingdoes not require a second-order representation of the outer sensing. Likewise, attentive inner sensingtending to inner sensing, doing the inner sensing carefullydoes not require a third-order representation of the inner sensing.
<19> Mental images do seem to have locations, but I maintain (in "Illusions") that this is due to the spatial contents of O-states, mistakenly generating image illusions, and not due to any spatial contents of I-states.
<20> I have not seen the accusation in print, but I wish I had a dime for every time it has been posed to me in conversation. (I would make a fortune off of certain individuals.) I elaborate the following response to the accusation in section 5 of "Stopgap."
<21> Chalmers provides no argument that consciousness has sharp boundaries, and he does not discuss Dennetts arguments that consciousness has vague boundaries. Independently of Dennetts claims, I believe it is plausible that the distinction between nonphenomenal and phenomenal sensing is vague, commonsensically. There seem to be fuzzy boundaries at the edges of ones visual field, for example, and intuitively it is hard to tell exactly when there becomes something it is like to see a peripherally moving object. Similarly, when ones attention is drawn to various faint pressure sensations or itches around ones body, it can be hard to answer whether there was something they were like a moment earlier. And exactly when does one lose all phenomenality as one falls asleep? Exactly when does one lose all phenomenality as an anesthetic dose is slowly increased? Reflective common sense is not committed to a sharp conscious/unconscious distinction. But if we have states that are neither clearly conscious nor clearly unconscious, then there should be possible creatures (perhaps, mice) that have only such states. Chalmers conclusion is as implausible as Dennetts is unrevolutionary.
I do not understand the difficulty Searle envisions. The last sentence seems to suggest that if one generates a phenomenally conscious experience by innerly sensing an O-state, then that O-state must "in itself" be a phenomenally conscious experience (of something external). I agree, if all this means is that the O-state must "itself" be an essential part of a phenomenally conscious experience. Otherwise, if it means that the O-state must be sufficient for a phenomenally conscious experience distinct from the experience generated by inner sensean experience that would persist if the O-state were "by itself," innerly unsensedthen Searles accusation collapses into Accusation 1.