35

Introspection, Explanation and Perceptual Experience:

Resisting Metaphysical Disjunctivism

Aaron Zimmerman

1. Seeing, Experiencing and Reflecting on Experience

My target will be Martin’s form of disjunctivism: the metaphysical claim that veridical visual

experience and hallucination are not species of a single genus. That is, according to Martin,

there is no non-disjunctive, fundamental kind of thing—visual experience—such that

hallucination and veridical experience are both types of things of this kind. I think there are

good reasons to reject such a view.

Let’s begin our assessment by supposing that a subject, Max, does indeed see a sphere in

front of him. And let’s suppose that I establish this fact by locating the sphere, locating Max,

and discerning a certain relation between them. When I verify that the latter sees the former,

what fact do I therein come to know? If common thinking is any guide, I have established the

existence of a complex event—an event that involves, as constituents, both Max and a sphere

wholly distinct from him.

Now we might ask, as metaphysicians of mind, whether the existence of this seen object

or any other is an “essential” aspect of the complex event that is Max’s seeing this sphere. If this

sphere were not in front of Max, and there were nothing of the kind in view, mightn’t Max still

be seeing something sphere-like in nature?

Suppose that Max is hallucinating extremely hard, and while it looks like there is a sphere

in front of him, he is really just staring into space. Does Max see something in this case? To

suppose not, is to do no more than reject a particularly aggressive version of the sense data

35

theory, as most of us now do. If there is nothing distinct from your mind before you, you aren’t

seeing anything at all. So while it looks to the hallucinating subject as though there is a sphere in

front of him, and he really does think that he is seeing a sphere, he is simply mistaken in this

belief.

Note, though, that there is a natural way of thinking about disjunctivist metaphysics of

the sort embraced by Martin on which these relatively common sense reflections on what it is to

see something already result in a mild form disjunctivism, albeit a form of disjunctivism limited

to the seeing of objects.1 Often when one hallucinates one does not see anything, as there is

nothing there to be seen. (And a subject seriously impaired by hallucination often cannot see

those things that are in fact there to be seen.) Seeing something is therefore an entirely different

kind of thing than is hallucinating something. Seeing involves both an observer and a seen

object. In contrast, hallucination is nothing more than an event in the observer’s mind.

Of course, the disjunctivist view of perception advocated by Martin is much more

controversial than this, as Martin’s thesis concerns visual experiences rather than instances of

seeing things.

The naïve realist claims that some sensory experiences are relations to mind-independent objects…taking

experiences to be episodes or events, the naïve realist supposes that some such episodes have as

constituents mind-independent objects. In turn, the disjunctivist claims that in a case of veridical

perception like this very kind of experience that you now have, the experiential episode that you enjoy is of

a kind which could not be occurring were you having a hallucination (2006, 354).2

1 Timothy Williamson’s (2000) view of “seeing that p” is a comparably mild form of disjunctivism. Cf. John Hyman (1999) and (2001). 2 See too, “The Naïve Realist . . . claims that our sense experience of the world is, at least in part, non-representational. Some of the objects of perception—the concrete individuals, their properties, the events these partake in—are constituents of the experience . . . The motivation for disjunctivism, I suggest, is a desire to hold on to Naïve Realism” (Martin, 2004, 273-4).

35

Now Martin claims that his “naïve realist” account of visual experience, “best articulates

how sensory experience seems to us to be just through reflection” (ibid.). Indeed, his stated goal

in “On Being Alienated” is to save this wholly “reflective” understanding of experience from a

skeptical train of thought: the kind of thinking that drives some philosophers to countenance

sense data as the objects of hallucinatory experience, and lands many more into thinking that

even veridical visual experience is representational in nature.3 So our first question is whether

Martin is right about this. If our accurate visual experiences really are “as [they] seem to us to be

just through reflection” must they have mind-independent objects as essential constituents?

Indeed, this first question might turn out to be our last. For if we don’t ordinarily take our

accurate visual experiences to essentially involve extra-mental objects of perception, we won’t

have to worry with Martin about whether our visual experiences really have this feature.

Of course, as we’ve already noted, our common sense conception of someone’s seeing

something is indeed a disjunctivist one. So I take it that what seeing something seems to us to be

through “reflection” alone is something that (essentially) involves a mind-independent object as

one of its constituents. But surely, whatever they are, visual experiences are supposed to be

located in our heads. And if the common folk, when they reflect on their experiences, picture

these episodes as unfolding in their heads, they are unlikely to think that a certain range of

them—namely, the veridical ones—essentially involve things that are located outside of their

heads. To be fair, this does not mean that Martin’s disjunctivist account of visual experience

cannot be right. But it would suggest that few of us are going to be convinced of its truth in the

3 The modern sense data theorists against whom Martin argues include Jackson (1977), Robinson (1994), and Foster (2000); the representationalists are too numerous to name, though see Byrne (2001) for extensive citations.

35

absence of a compelling argument.4 It would suggest that Martin must do more than merely

defend the disjunctivist account from skeptical assault.5

Indeed, one might wonder whether the folk even have the concept that philosophers

associate with “veridical visual experience.” So one might doubt whether there is a pre-

theoretical conception of veridical visual experience to rescue from skeptical critique.

Admittedly, if our concept of visual experience is theoretical in nature, the theory is an

extraordinarily old one, as philosophers since at least Aristotle have discussed cases in which the

appearances of objects vary without change in the objects themselves.

Why is it that to those who are very drunk everything seems to revolve in a circle, and as soon as the wine

takes hold of them they cannot count objects at a distance . . .? Is it because the vision is continually

disturbed by the heat of the wine? The same thing then happens to those who are drunk as when an object

appears double if one puts it close to the eye. (Problems 874a; 1984, 1346)6

In his famous discussion of “Skepticism with Regard to the Senses,” the young Hume lumps this

observation together with others made in the intervening millennia.

4 I have in mind here an argument like Hilary Putnam’s (1975) famous case for the conclusion that the meanings of our words are not in our heads. Chalmers and Clark (1998) argue for an externalist account of the memories, thoughts and other semantic “vehicles” that we commonly suppose are in our heads, an externalism that resembles Martin’s in greatly exceeding Putnam’s conclusions about meaning or semantic content. 5 Might Martin just mean by “having a veridical visual experience of x” what the folk mean by “seeing x”? It is telling that Susanna Siegel, before critiquing Martin’s view, feels she must inform the reader of a departure from ordinary thought and talk. “Throughout this essay, I’ll use ‘veridical experience’ to pick out instances of perception [e.g. seeing] that are not hallucinations or illusions. So there will be no such thing as a veridical hallucination in this sense of ‘veridical’” (2008, 205). Once the concept associated with “veridical experience” is distorted in this way—once it becomes analytic that hallucinations are inaccurate—“veridical experience” may well be cognitively equivalent to “seeing” and the events to which both terms apply may well have objects as their essential constituents and so differ (in at least this way) from even perfect hallucinations. (The question I am asking is whether Siegel could have achieved the same effect by saying, “‘Having a veridical visual experience of x’ is herein to be read ‘seeing x’.”) At any rate, I choose to think that Martin is not pushing for linguistic or conceptual reform. On the reading of the texts I find most natural, Martin is merely arguing (albeit on wholly a priori grounds) for naïve realism, where this implies no more than that having a veridical visual experience of something is a great deal like seeing it in that both events essentially involve an experienced/seen object as an extra-mental constituent. (That is, on my reading, Martin treats naïve realism as a synthetic—albeit an a priori knowable—truth.) I suspect that Martin also embraces the associated property identity— that having a veridical visual experience of something just is seeing something—but I am unsure of this, as there are passages—e.g. (2004, 294)—in which Martin seems to countenance veridical hallucinations. 6 While of unquestionable antiquity, the authorship of Problems is disputed.

35

When we press one eye with a finger, we immediately perceive all the objects become double, and one half

of them to be remov’d from their common and natural position. But as we do not attribute a continu’d

existence to both these perceptions, and as they are both of the same nature, we clearly perceive that all our

perceptions are dependent on our organs, and the disposition of our nerves and animal spirits. This opinion

is confirm’d by the seeming encrease and diminution of objects, according to their distance; by the apparent

alterations in their figure; by the changes in their colour and other qualities from our sickness and

distempers; and by an infinite number of experiments of the same kind; from all which we learn, that our

sensible perceptions are not possest of any distinct or independent existence. (T 1.4.2.45) 7

Let’s focus on the first of these experiments, the double vision of which both Aristotle

and Hume speak. When I press my eyeball while looking at a sphere, something will of course

change. (Hume is right about that.) But there is no reason to infer that what I am seeing changes

in the case. (This is where Hume goes off the rails.) I am still just seeing a sphere, and I know

that it has not changed. (There remains a single sphere before me, and I know this without the

aid of philosophy.) Instead, Hume’s experiment puts me in a position to infer that the change

I’m aware of—the doubling of my vision—must be caused by the depression of my eyeball.

And since the depression of my eyeball is unlikely to cause a change in the sphere before me, it

must instead change something in my head: a visual image, impression, sensation or experience.

Moreover, if common sense is any guide, the visual experience I become aware of when I

conduct Hume’s experiment can be evaluated in something like that manner in which we

evaluate maps, portraits, and other iconic representations. My experience is accurate prior to the

depression, as there is then a single sphere before me, and that’s how things then look. But my

visual experience is then distorted when I depress my eyeball; it then becomes inaccurate. For

after the eyeball’s depression there is a sense in which it now looks to me as though there is a

7 As to the sources of these experiments, David and Mary Norton cite in addition to Aristotle: Lucretius, Sextus Empiricus, Hobbes, Montaigne, Rohault and Collier. See their edition of the Treatise referenced below at 477, n. 45.

35

second sphere, a sphere removed from the “common and natural” position occupied by its

original, even though I know that there is no sphere at this second location. After the depression,

things are not entirely as they look.

It seems, then, that the conception of visual experience we get from a simple experiment

known to philosophers for thousands of years is of something in the head that represents things

outside the head in either an accurate or inaccurate manner. It is not, therefore, the naïve realist’s

conception of visual experience.

Now, what does all this show? It shows that when Martin writes of how “sensory

experience seems to us to be just through reflection” he must not have in mind the experiments

of Aristotle and Hume and the lessons common sense draws from them. It would seem, that is,

that Martin is not using “reflection” to denote the hybrid of observation, introspection, and

inference described above. For that method of thought leads us away from naïve realism towards

a distinctively representational view of experience.

Let’s call the form of reflection driving Aristotle and Hume “a-introspection” and

characterize it as follows.

A-Introspection: Forming beliefs about one’s feelings, thoughts, and experiences making

full use of any experiments, observations and inferences that seem relevant.

Clearly, a-introspection involves introspection properly so called, as Aristotle and Hume rely on

introspective judgments about the “doubling” that marks their visual experience when their

eyeballs are depressed or the object before them is too close to clearly see. But a-introspection

clearly involves much else besides introspection. For instance, both theorists must infer their

35

representationalist conclusions from their introspective observations and the further (extra-

introspective) assumptions articulated above.

Now I have been arguing that this augmented form of “introspection” yields a conception

of visual experience at odds with naïve realism. The question for Martin, then, is whether there

might be another more reliable guide to the nature of visual experience—a method or form of

introspection that might lead us to discount the results of a-introspection and embrace Martin’s

naïve realism instead.8

2. A Purely Introspective Conception of Experience

When Aristotle and Hume reflected on the nature of their visual experiences, they felt free to

conduct experiments and reason from their observations. And, as we have seen, this process of

reflection speaks against naïve realism. So if a-introspection does not support naïve realism,

what does?

To answer this question, Martin turns away from observation and inference to the wholly

introspective methods employed by phenomenologists. While looking at the sphere (with wholly

un-depressed eyeballs) I am supposed to simultaneously see the sphere and introspect the largely

accurate experiential episode I am then enjoying. And naïve realism is supposed to be the

conception of my visual experience I find most congenial when I am introspecting in this

manner. We can call this form of reflection “b-introspection” to distinguish it from its more

inclusive rival.

8 Note that it would be wrong to suggest, as Martin sometimes does, that theorizing about the nature of hallucinations is the only thing driving theorists away from naïve realism. As the quotes from Aristotle and Hume make clear, hallucinations are only one among a variety of disturbances, disruptions, and patent inaccuracies in our typical visual experience that would seem to have proximate (or internal) causes and are consequently cited as evidence for the view that experiences are head-bound representations. Scott Sturgeon (2006, 2008) makes a related observation when he describes the spectrum of cases that has entirely accurate experience and pure hallucination as its “good” and “bad” endpoints.

35

B-Introspection: Forming beliefs about one’s feelings, thoughts, and experiences in an

“immediate” way—i.e. without relying on experiments, extra-introspective observations

or any inferences from them.9

Of course, even if we were convinced that the results of b-introspection support naïve

realism, we would not yet be warranted in accepting that view. For we also know that the results

of a-introspection run contrary to naïve realism. Thus, even if Martin convinces us that b-

introspection really does support a naïvely realistic view of visual experience, we must still

weigh our competing sources of evidence to determine which source, if either, should be trusted

in the case on hand. Which guide to experience is more reliable: the a-introspective hybrid of

introspection, observation and inference that leads us to infer a representational view of

experience from double vision and the like, or the purely a priori b-introspective method that is

supposed to lead us into naïve realism?

Now I do not think we can answer this difficult methodological question in a fully

general way. For there are, admittedly, b-introspective judgments that are properly assigned

enormous epistemic authority: my current judgment that I am in pain and my current judgment

that I am thinking about disjunctivism are two clear examples. And it is indeed hard to see how

the judgments in this privileged class could be overturned by experimentation, observation and

inference. Nevertheless, one might reasonably doubt whether Martin’s b-introspective judgment

that his accurate visual experience involves an extra-mental object as an essential constituent

9 Eric Schwitzgebel equates introspection with what I’ve been calling “a-introspection” (this volume, ch.1). But I can see no objection to isolating the non-inferential aspects of this process and introducing a label for them alone. It is of course a further claim that when we use “introspection” we more often intend to pick out the less inclusive belief-forming process (b-introspection) than the more inclusive one (a-introspection). It may be that philosophers typically use “introspection” to pick out b-introspection while the usage of lay people is less restrictive.

35

falls into this same highly authoritative category. For surely, highly abstract, highly theoretical

judgments about one’s experience are at least more amenable to third person refutation than are

those core cases of introspective report on which discussions of first person authority rightfully

focus.

But the case against metaphysical disjunctivism is even stronger than these reflections

suggest. We do not have to content ourselves with arguing that a-introspective arguments for

representationalism overwhelm the support naïve realism gains from b-introspection. For Martin

is just wrong in thinking that the results of b-introspection support naïve realism. Instead, as I

will now explain, it seems to me that both a-introspection and b-introspection support the

representationalist view against its rival.

Note that there are many different thoughts I can have about my visual relation to the

sphere before me. Crucially, I can think or reflect on the fact that I am seeing the sphere and I

can also reflect on the fact that I am having (or “enjoying”) a visual experience of it. And while

it is relatively uncontroversial that the event that is my seeing the sphere has an extra-mental

object as one of its (essential) constituents, if my attention is instead turned to my experience of

the sphere—if I am thinking of what is now happening as an accurate experience of something

extra-mental rather than as the seeing of that thing—I am not thinking of what is now happening

in a naïvely realistic manner. Or, at any rate, Martin fails to argue that naïve realism supplies us

with the best description of what I take my experience to be when I reflect in the manner he

recommends.

In sum, experiments known to the ancients lead most of us to posit visual experiences

that are representational in nature. If I eschew these experiments and try to get a more direct fix

on the kind of visual experience I am enjoying while seeing the sphere before me, I initiate a

35

process of b-introspection. But to form beliefs about the experience I am having I must decide

how to best conceptualize or characterize that experience. If I think of what is going on as my

seeing the sphere, I am led to a naïvely realistic view of what is going on. But if I think of what

is going on as my having an accurate visual experience of a sphere, I am not led into thinking of

the episode as the naïve realist does. Instead, since this episode is nothing more than an

experience (albeit an accurate one) it would seem to be something in my mind, something that

could exist in the absence of that wholly external object of which it is an experience.

Disjunctivism does not emerge from either a-introspection or b-introspection on our experiences,

but must instead be motivated with an independent argument or source of evidence.

3. Non-Conceptual Purely Introspective Awareness of Experience

Though Martin does not explicitly address the line of objection advanced above, there are crucial

moments in his defense of disjunctivism in which I find him positing a non-conceptual form of

introspective awareness of visual experience. (Whether Martin really believes in non-conceptual

awareness of experience is something I will address below.) And if we are indeed “acquainted”

with our visual experiences when we introspect, Martin can invoke this awareness in reply to the

case I have presented against him. Perhaps, as I’ve suggested, it is up to Max to decide whether

to think about what is going on in the scenario I have described as his seeing a sphere or his

having a veridical experience of such, where in the first case he is apt to think of what is going

on in disjunctivist terms, and in the second case he is not. But Martin might insist that Max is

also introspectively acquainted with what is going on in the case. He has, that is, an

introspective awareness of what is going on that does not await his decision to think of it as an

35

episode of seeing or a veridical visual experience. We can call the method we deploy to secure

this kind of acquaintance with states of our own minds “c-introspection.”

C-Introspection: Focusing one’s attention on one’s feelings, thoughts, or experiences

without necessarily categorizing them (as, say, visual experiences or events of seeing) or

drawing conclusions about what they really are.

Now if there is such a thing as c-introspection, Martin might insist on its providing

evidence for naïve realism. When Max c-introspects his visual experience of the sphere before

him, Martin might claim, that experience will seem to him to essentially involve a sphere.

Martin’s only remaining task, then, would be to argue, as he acknowledges he must, that visual

experience really is as it is presented to us when we c-introspect.

Our question, then, is whether there is such a thing as c-introspection, and if there is,

whether this primitive take on our visual experiences presents them to us as essentially external-

object-involving.

Now there are those who doubt the very coherence of c-introspection. For example,

Sydney Shoemaker rejects a “narrow perceptual model” of introspection by arguing that

introspective knowledge is “fact-awareness” not grounded in “thing-awareness” (1996, ch. 10;

cf. Tye, 2009b, ch. 5). According to Shoemaker, Max can know via introspection that he is

having an experience of a sphere, and he can know (in some way or other) that he is seeing a

sphere. But Max cannot be introspectively aware of the event in question if this is supposed to

be something distinct from his awareness that it is an event of a certain type. According to

Shoemaker, that is, Max cannot simply focus his attention on his visual experience without

35

therein thinking of it as a visual experience, an event of seeing something, or an event of some

other kind. In consequence, neither Max nor we can c-introspect. On Shoemaker’s lights,

introspection is nothing more than the formation of introspective beliefs or judgments.

Introspective judgment has no extra-conceptual (or pre-conceptual) preliminary.

I have argued elsewhere that Shoemaker’s critique of the narrow perceptual model of

introspection is too ambitious (Zimmerman, 2008). Though our introspective knowledge of our

beliefs and thoughts is indeed wholly non-perceptual in nature, our knowledge of our sensations

and experiences resembles paradigmatic perceptual knowledge in various respects. For instance,

in stark contrast with our thoughts and beliefs, we can quite easily describe mistaken judgments

about our own experiences that don’t have their source in irrationality or conceptual confusion.

And a careful examination of these mistaken judgments unearths at least some evidence of c-

introspective awareness of visual experience.

The phenomenon is perhaps most clearly exemplified by a scenario Gareth Evans

discussed:

Consider a case in which a subject sees ten points of light arranged in a circle, but reports that there are

eleven points of light arranged in a circle, because he has made a mistake in counting, forgetting where he

began. Such a mistake can clearly occur again when the subject reuses the procedure in order to gain

knowledge of his internal state: his report ‘I seem to see eleven points of light arranged in a circle’ is just

wrong. (1982, 228–9)

Evans’s subject does not have an irregularly malleable visual experience that morphs to match

his miscalculations. Instead, he has a somewhat inaccurate view on the character and content of

a largely stable stream of experience. He seems to see ten points of light in front of him, but he

mistakenly thinks that he seems to see eleven.

35

There are at least three features of the case worth noting. First, the events it describes are

fairly unexceptional. Relevantly similar cases can be described that don’t involve numerical

concepts, as when one is asked by a doctor to characterize bodily discomfort and mistakenly

describes gas pain as nausea, or is asked by a cook to identify the flavors of a complex broth and

mistakes the taste of lime for that of vinegar. Second, Evans’s case lacks even the faintest whiff

of irrationality. The subject he describes is guilty of nothing more exotic than a failure to keep

track when counting. Third, and most importantly, the example demonstrates that cognitive or

conceptual acts like counting often intervene between the enjoyment of a visual experience and

the formation of an introspective belief or judgment affirming the existence and character of that

experience.

Admittedly, it would be odd to posit a (second-order) seeming-experience that intervenes

between one’s (first-order) visual experience and the introspective belief or judgment that one is

enjoying that experience. And since visual experiences do intervene between the extra-mental

objects we see and our knowledge of the existence of those objects, introspective knowledge of

experience and perceptual knowledge do differ in this important respect. But we have seen that

introspection of experience resembles perception in an equally important way. Both acts require

categorization, counting, and other conceptual preliminaries. And when these preliminary acts

of conceptualization are unsuccessful—as when Evans’ subject miscounts—mistaken

introspective judgments can result. In consequence, there are introspective errors similar in kind

to errors of perceptual judgment.10

10 These reflections rebut Martin’s (2006, 407) claim that one must either: (a) accept (second-order) introspective appearances to which our (first-order) experiences needn’t conform; (b) posit an infallible mechanism of introspection; or (c) deny, as Martin does, that we have “a distinct perspective on our inner lives from that we take when experiencing the [extra-mental] world” (2006, 407). I have argued that we have a “distinct perspective” on our “inner lives.” It is constituted by our (second-order) introspective beliefs about—and, perhaps, our c-introspective awareness of—various first-order states of mind (experiences of spheres and the like). On this view,

35

At the very least, then, the relation between our introspective judgments and our

experiences can be seen to be different in kind from that which obtains between our first order

beliefs and thoughts and our (second-order) introspective knowledge of them. Descartes’ belief

that he exists is self-verifying—the judgment is made true by its execution. And the same is

probably true of Tyler Burge’s (1988) class of cogito-like judgments: judgments of the form “I

am thinking p.”11 But our introspective knowledge of our experiences is not like this. Indeed,

the difference remains in view when we turn from our introspective knowledge of those thoughts

we merely entertain to our knowledge of our own beliefs. Evans’ subject does not figure out

whether he believes that there are eleven points of light before him by turning his attention to his

belief and counting. Instead, his knowledge of his belief is drawn directly from it (Zimmerman,

2006b). In contrast, Evans’ subject does turn his attention to his experience when determining

how many lights he seems to see. He first becomes c-introspectively aware of that experience.

And only then, after (incorrectly) counting the lights he seems to see, does he form a (mistaken)

judgment as to the number of lights he is experiencing. C-introspection is not only possible; it is

a typical preliminary to b-introspective judgment.

There are, admittedly, other possible descriptions of the case that do not invoke c-

introspective awareness of experience. One might insist, for instance, that Evans’ subject can

only exercise perceptual attention, and then either replace or augment its more natural product—

the judgment that there are x number of lights in front of him—with the (wholly conceptual)

judgment that he is having an experience of x number of lights. According to this line of this

thought—which may indeed be Evans’ take on the scenario—advocates of the perceptual model

introspection is fallible, but not because our (first-order) experiences aren’t always as they introspectively seem. (Again, there are no seeming experiences intervening between our experiences and our knowledge of them.) Introspection is fallible because we can incorrectly conceptualize or categorize our experiences when formulating our introspective judgments. 11 See Zimmerman (2006a) for caveats.

35

of introspection mistake cases in which we more carefully focus our perceptual attention on

external objects for cases in which we c-introspect upon the visual experiences that represent (or

contain) those objects.

But this reply is not plausible in the case of hallucination. If Evans’ subject is

hallucinating ten lights, focuses his attention on what he takes to be the circle of lights before

him, and then counts as best he can, he will of course think that he has focused his attention on

the lights he can see. (He will think he is more carefully examining those points of light, not his

experience of them.) But we will know, looking in, that his attempts to engage the mechanisms

of perceptual attention were for naught. Since there were no lights there for him to see, he only

succeeded in making himself more fully aware of features of his experience. And surely, it is

introspection that makes one aware of these features of one’s experience, not vision. Our

subject’s visual experience represented ten points of light arrayed in a circle before him. And,

unbeknownst to him, when he tried to focus on the lights before him, he wound up dwelling

upon those features of his experience that help make it a representation of said lights. He c-

introspected.

Now, though I think these considerations make the existence of c-introspection a

reasonable conjecture, I will not pretend that they force its acceptance. And this is clearly not

the place to evaluate the wealth of arguments that have been presented both for and against the

perceptual model. So let us just suppose, if only for the moment, that reflection on Evans’

example (or its hallucinatory variant) has convinced us that Shoemaker is mistaken, and that we

can indeed get an introspective fix on our visual experiences without conceptualizing or

describing them as events of experiencing or events of seeing. With this assumption in place, we

can allow our initial subject Max “thing-awareness” of his veridical visual experience of the

35

sphere before him—a kind of awareness that cannot be equated with either his knowledge that he

is having an experience of a sphere or his knowledge that he is seeing a sphere. So when Max c-

introspects in this manner, will his experience seem to him to involve an extra-mental

constituent?

To answer this question we must each do our best to c-introspect and report what we find.

So I encourage you to turn your attention to your visual experience and to verify whether the

claims I am about to make jibe with those to which you are inclined.

Now I find it difficult to focus my attention on my current veridical visual experience

itself; it is much easier to focus perceptual attention on the sphere I can now see. But it is not

nearly as hard for me to focus my attention on the veridical visual experience I enjoy when my

vision is blurred, doubled or distorted in some other manner.12 But when, on these occasions, I

manage to focus on my experience itself, I do so by shifting my attention away from the extra-

mental objects I am seeing. Indeed, if I don’t do this, how can I be sure that I am not just more

acutely seeing those extra-mental objects that I can see, or more carefully dwelling on the fact

that I am seeing them? What distinguishes such acts from c-introspecting the veridical visual

experience itself?

Surely, if there is such a thing as c-introspecting an experience of a sphere, there must be

something that distinguishes this act from attentively focusing on a seen sphere, as attentively

watching something isn’t itself an introspective act. (Only primates can introspect, but many

“lower” species can look, watch, and examine what they see.) But it seems that the only feature

that could distinguish the two acts from one another is a difference in the focus of attention.

When I attentively focus on the sphere, I more carefully examine its shape, color and the like. In 12 Note that these are still instances of veridical experience. (They are neither hallucinations nor illusions.) I am still seeing the blurry object in view; as we representationalists would say, my experience still correctly represents the existence of this object despite mischaracterizing its color, shape and exact location.

35

contrast, when I c-introspect my visual experience of the sphere, I focus my attention on its

blurriness and other clearly perspectival properties of the event. But if c-introspection essentially

involves such an “inward” shift of one’s attention, the object of c-introspection won’t present

itself as an extra-mental object, nor will it seem to contain such an object as an essential

constituent. Instead, I will have used c-introspection to focus my attention on precisely those

features of my visual experience that do not reflect the properties of the sphere it enables me to

see. And though it is an open question whether my current visual experience really could exist in

the absence of any seen object, it does not seem to me that c-introspection provides any evidence

that this experience—or an experience of the same type—could not be wholly hallucinatory.

Surely, my vision’s blurriness is not an indication of the sphere’s true shape. Isn’t it then

possible that none of the features on which I am focused have any extra-mental reality? As far as

I can tell, c-introspection can do nothing to answer this question in the negative.

In the end, then, I suspect that Martin rejects c-introspection altogether. (I will say more

about this below.) For if Martin did countenance c-introspection, he would stand guilty of

conflating it with attentive seeing. He would stand accused, that is, of running together the act of

focusing introspective attention on experience with the act of focusing perceptual attention on its

objects. Why else would he say that when you turn your attention to your visual experience, it is

presented to you as essentially external-object-involving?

At any rate, whatever explains the divergence between Martin’s introspective reports and

my own, I am certain that at least my case does not answer to Martin’s description of how

attention to experience is supposed to pan out. When I do my best to c-introspect my current

visual experience it does not seem to me to essentially involve the external object of which it is

an experience.

35

4. Explaining Introspective Indiscernibility

I have argued that a-introspection, b-introspection and c-introspection all speak in favor of

representationalism and against a naively realistic view of visual experience. Still, I do not want

to hang too much on what might prove to be peculiarities in my own case.13 So I want to admit,

if only for the sake of argument, that when we c-introspect on our veridical experiences we take

them to essentially involve external objects in the way in which events of seeing essentially

involve the extra-mental thing seen.

And yet, even if the disjunctivist is granted this, the view is still open to dispute. For, as

Martin admits, there will be cases in which a subject cannot tell via introspection alone whether

she is hallucinating or instead enjoying a veridical visual experience. And when we consider

cases in which a subject cannot tell via introspection whether she is, say, hallucinating a sphere

or actually seeing one, it is natural to suppose that there is something common to both the

hallucinatory and veridical experience beyond the mere fact that they cannot be told apart by the

subject in question, where this further commonality explains why the subject cannot tell these

experiences apart.14

There are actually several forms such explanation might take. First, we must ask whether

b-introspection is preceded by c-introspection in the case. Does the subject merely judge that her

visual experience has not changed despite the move from veridical experience to hallucination?

Or does she ground this judgment in some more primitive (non-conceptual, quasi-perceptual) act

of c-introspection? Next, we must decide which features of the subject’s experience cause and 13 My worries may be well founded. On the basis of a (rather limited) survey of those philosophers who publish their introspective reports, Hellie (2007, 266-9) argues that “phenomenological study” of our experiences provides support for something like naïve realism. 14 This is a common criticism of disjunctivism ably mounted by Siegel (2004, 218-23); Hawthorne and Kovakovich (2006, 179); Byrne and Logue (2008, 89-90); and Tye (2009a, 560).

35

rationalize her judgment that nothing has changed. Does she judge that her experience remains

the same because her pre- and post-hallucinatory experiences are qualitatively identical? Or

might distinct representational features play this role? Does she judge that her experience has

not changed because what things are like for her when she is hallucinating is exactly the same as

what things are like for her when she accurately perceives what is before her? Or does she judge

that nothing has changed in her experience because her hallucinatory experience represents

precisely what her veridical experience represents, where what an experience represents cannot

be equated with what things are like for the subject enjoying that experience?

Different answers to these questions yield different substantive theories of introspection.

But what is striking about disjunctivism, is its incompatibility with any such theory. For the

disjunctivist’s commitments prevent him from accepting any substantive explanation of

introspective indiscernibility.

We can make this criticism more precise by concisely stating the explanatory thesis

metaphysical disjunctivists must reject.

Substantive Explanation of Introspective Indiscernibility (SEII): If a mature, reflective

subject judges on the basis of careful introspection that her (relatively simple) visual

experience before t is the same as her experience after t, where one of these experiences

is hallucinatory and the other veridical, her experience before t must be similar to her

experience after t along some dimension, where the similarity in question explains (i.e.

causes and rationalizes) her judgment that nothing has changed.

35

The disjunctivist must reject SEII. For if there were a positive (representational or qualitative)

characteristic common to both hallucinatory and veridical visual experience, we could use this

characteristic to pick out the genus of which both kinds of experience are species (Martin, 2006,

367). And this would impugn the disjunctivist’s claim that hallucination and veridical

experience of a sphere are not members of a natural or non-disjunctive kind. They would instead

be experiential representations of a sphere, or experiences of the qualitative kind: sphere*. Thus,

it seems that even if we grant the disjunctivist the phenomenological support he claims for

himself, we can rationally reject the position on offer as explanatorily impoverished.

Now Martin tries to reply to this critique by arguing that no explanation of the

introspective indiscernibility of hallucinatory and veridical experience is necessary (2006, 393-

6). We can, he thinks, fully characterize the agent’s “perspective” when she is hallucinating by

saying that she cannot discriminate her situation from one of veridical visual experience (ibid.).

So we needn’t say that she is then aware of some positive (representational or qualitative)

features of her experience: features that would equally qualify her experience were she actually

seeing something external to her mind. Of course, it is not uncommon to say that on both

occasions it looks to the observer “as though” there is a sphere before her; and it is not

uncommon to say that “what it is like” to enjoy a perfect hallucination of a sphere is no way

differs from what it is like to really see one. But, in Martin’s view, what we’re trying to convey

here is just the fact that we cannot tell the two episodes apart. “It looks to the observer in both

cases as though there is a sphere in front of her” and “What its like to see a sphere in no way

differs from what it is like to vividly hallucinate one” do not state facts of similarity from which

sameness in experiential kind can be inferred.

35

Might this be true? The first thing to note is that the view is unattractive as stated.

Suppose we focus on a case in which a subject, Sam, sees a sphere before him right up until the

moment at which the sphere is removed and he begins to enjoy a perfect hallucination of it.15

When we ask Sam whether anything has changed (in either the sphere or his visual experience of

it), he will honestly answer in the negative, as it looks to Sam as though nothing has changed.

And those of us who reject Martin’s disjunctivism will explain Sam’s inability to discriminate

the hallucination from the veridical perception in terms of what we at least take to be distinct

facts about Sam’s experience. Despite the disappearance of the sphere, Sam’s experience of it

has remained qualitatively the same, and his experience still represents to be the case precisely

what it did before. But Martin must insist that these quite natural explanations are defective.

According to the disjunctivist, our descriptions are either false, or they fail to assert anything

beyond the fact that Sam cannot tell his two different experiences apart. But why should we

abandon our explanatory practices as Martin requests? Because veridical experience

(purportedly) seems to us to involve objects when we c-introspect upon it? But why should the

latter consideration weigh more heavily in our thinking than the former? Why not save SEII by

rejecting the (purported) introspective appearances on which Martin relies?

5. An Internal Tension in Martin’s Reply

We have seen that neither a-introspection nor b-introspection provides support for naïve realism,

and that Martin must therefore accept the existence of c-introspection if he is to provide the least

shred of evidence for naïve realism and the disjunctivist conception of hallucination and

veridical experience he bases upon it. Moreover, as we have also seen, Martin maintains that no 15 William Fish might reject the very possibility of Sam’s case as I’ve described it, as he insists the possibility of an unnoticeable shift form veridical perception to perfect hallucination cannot be safely assumed (2009, 87). But I cannot see why Fish demands defense of a possibility so easy to imagine and describe.

35

explanation of the introspective indiscernibility of veridical visual experience and perfect

hallucination should be provided. Since all such explanations depict hallucination and veridical

experience as species of a common genus, the disjunctivist must reject them all. But there is a

tension here. For in arguing against SEII—and so against substantive explanations of

introspective indiscernibility—Martin would seem to reject the very existence of substantive

introspective knowledge. And in so doing, he robs himself of the only potential source of

support for his belief in naïve realism.

Martin’s argument for why we need not explain why subjects cannot tell their perfect

hallucinations apart from their veridical experiences hangs on the idea that introspection is

importantly different from perception. Though the sphere that Sam can see is, for example,

distinct from Sam’s initial visual experience of it—as the sphere is an extra-mental object—

Martin claims that Sam’s visual experience of the sphere is not distinct from his introspective

awareness of that experience. On Martin’s lights, we don’t need to explain why Sam judges that

nothing has changed when his experience shifts from veridical perception to perfect hallucination

because the fact that his experience remains qualitatively or phenomenologically the same is

nothing beyond his introspective sense that nothing has changed in it.

However things seem from the subject’s perspective with respect to her phenomenal consciousness is how

phenomenal consciousness must be…

The subject’s perspective on her own sense experience constitutes sense experience being that way for her .

. . If she really is in a situation in which from her perspective it is as if she is having an experience as of a

white picket fence, then that constitutes her being in the situation of having an experience as of a white

picket fence. (2006, 391-3)

Now I think is natural to read these statements as claiming a kind of infallibility for our

introspective judgments. And if this reading is correct, Martin must deny the very cogency of

35

Evans’ example of mistaken introspection judgment discussed above. (One cannot be led by

introspection to the false conclusion that one seems to see eleven points of light if false

introspective judgments are simply impossible.) Moreover, whatever one thinks about the

plausibility of positing infallible powers to introspect, the dialectical power of this maneuver

cannot be denied. For if our judgments about our visual experiences really were infallible,

Martin would be right to reject the folk psychologist’s attempt to explain Sam’s judgment that

nothing has changed in his experience by appealing to qualitative and representational

constancies in that experience. Why? Well, Martin is right to reject as incoherent infallible

mechanisms of introspective discrimination (2006, §7). And if we cannot be credited with

infallible powers of introspective discernment, our (purportedly) infallible introspective

judgments about what we are experiencing must be assigned some other source. (Perhaps those

theorists driven to accept the invariable truth of our introspective judgments would feel pressed

to invoke conceptual considerations. Like Descartes’ judgment that he exists, and the clever

man’s belief that he believes at least one thing, our introspective judgments about what we are

experiencing would have to be construed as making themselves true.) So Martin is right about at

least this: If Sam’s introspective judgments about his experience were truly infallible, his

judgment that nothing changes in his experience when his veridical experience gives way to

hallucination really could not be grounded in an appreciation of some respect in which his

veridical experience and his hallucination are independently similar to one another.

There is a problem, however, with endorsing this line of reasoning and following Martin

along a train of thought that would seem to lead him from the infallibility of our introspective

judgments to a denial of SEII. And this is just the patent implausibility of the claim that

introspection is infallible in the sense we have described. Surely, Evans’ subject is just wrong

35

that he seems to see eleven points of light. Surely, our introspective judgments are sometimes

mistaken. Of course, as we have already noted, there are indeed important differences between

perception and introspection. But fallibility is something the two sources of knowledge would

seem to have in common.

It is perhaps because of these worries that Martin swears off the natural reading of the

text that we have been considering. The disavowal comes when Martin confronts the

Wittgensteinian objection that infallible knowledge isn’t really knowledge at all. If my opinion

on some matter can’t be wrong, the Wittgensteinian protests, its truth isn’t much of an

achievement; and, the critic continues, knowledge must be won, not awarded by default. So if

our introspective judgments really can’t be wrong, they can’t really constitute knowledge.

In reply to this sort of objection, Martin says that “a subject’s perspective on her own

mind” is not meant to denote her beliefs and judgments about how things seem to her.

The constitutive connection is between the subject’s perspective on his or her own mind, how it seems to

be, and how his or her mind then is. This need not be identified with the judgments he or she actually

makes. (2006, 392, fn.42)

Presumably Martin is not just saying that the subject’s perspective on her experience need not be

identified with her introspective judgments. Instead, the reader is being told that she should not

so identify the introspective perspective when interpreting Martin’s claim that this perspective is

infallible. For if the reader did so interpret the phrase, she would conclude that Martin thinks our

introspective beliefs about what we are experiencing are invariably true. And Martin is assuring

the reader that he does not endorse this patently false doctrine.

But if the introspective perspective is not constituted by introspective judgments, and it is

not constituted by introspective experiences of our experiences—because, as Martin rightly

asserts, there are no introspective appearances—what exactly is the introspective perspective

35

supposed to be? I don’t know whether Martin has arrived at a firm answer to this question. The

remarks he does make suggest that he endorses something like Shoemaker’s (1994) view

according to which a first-order mental state will “double” as an introspective awareness of itself

so long as it is embedded in a mind with the requisite concepts and powers of reasoning. And if

this is indeed Martin’s view, he would have Sam’s visual experience of the sphere itself

constituting Sam’s introspective perspective on that experience (given its embedding in Sam’s

sufficiently sophisticated and well-functioning mind).

But even if we suppose that Sam’s initial veridical visual experience of the sphere

somehow “doubles” as his awareness of such, Martin must tell some story about how this

minimal form of introspective awareness might somehow morph into the full-blown judgment

that the sphere is an essential constituent of that experience. For recall that Martin does not

support his assertion of naïve realism by citing experiments, observations and inferences.

Instead, the view is supposed to be motivated by introspection or “reflection” alone. But how

exactly is a belief in naïve realism supposed to emerge from introspection when a subject’s

introspective awareness is nothing more than the (suitably embedded) experience of which she is

aware?

We can bring the problem this question poses more clearly into focus by shifting our

discussion from our imagined subject Sam to Martin, the theorist himself. To do this, let us

suppose that Martin is enjoying a veridical visual experience of a lavender bush in bloom before

him. How does Martin transition from his minimal introspective awareness of this visual

experience—an awareness that is supposed to consist in his suitably embedded visual experience

of the blooming bush—to his confident judgment that this experience involves the lavender bush

as one of its essential constituents? Surely a veridical visual experience of a bush is not itself a

35

belief in naïve realism. But Martin says nothing about how a subject might transition from the

one to the other.

In sum, Martin argues against SEII by claiming that introspective awareness of a visual

experience of a mind-independent object is nothing more than that very experience embedded in

a suitably sophisticated mind. And it is clear why he argues for this extraordinarily minimalistic

conception of introspection. For if introspective awareness of experience were more substantial

than this, it would be hard to see how it could be infallible. And if introspection were fallible,

folk psychologists would be right to explain its successes in cases like Sam’s by appealing to

those positive representational features his hallucination and veridical experience of the sphere

have in common; features that justify and render true his belief that his experience has not

changed; features that are not present when, as sometimes happens, a subject mistakenly judges

that nothing has changed in her experience. But Martin cannot consistently characterize

introspection in this minimalistic way. For naïve realism is supposed (by Martin) to derive its

support from introspection alone. And it is clear that Martin must characterize introspection in a

more substantial light if he is to convince us that it really does provide evidence for a naïvely

realistic conception of veridical visual experience. Indeed, it is hard to see how Martin’s varying

descriptions of introspection—invoked, as they are, in response to two very different challenges

to his view—might be unified in such a manner as to leave his case for disjunctivism intact.

6. Disjunctivism and Psychological Explanation

Visual experience doesn’t just interact with our beliefs about it. Crucially, psychologists of both

the folk and academic varieties invoke experience to explain beliefs about external objects—

beliefs that are not themselves preceded by introspection upon experience. And it can be seen

35

that many of these beliefs respond to hallucination and veridical perception in exactly the same

way. Thus, reflection on cases reveals that we often treat a subject’s hallucinatory and veridical

experience as equivalent in kind when explaining her efforts to render correct perceptual

judgments. And Martin has not successfully argued for the impropriety of this practice.

To illustrate the kinds of folk psychological explanation I have in mind, let’s suppose that

Sam initially sees not one large sphere before him, but ten small ones, and that at some point he

begins to hallucinate the spheres without being able to tell that a change has taken place. There

are two variations to consider.

In one case, Sam is asked how many spheres are before him prior to the onset of his

hallucination, and after correctly counting them up, he concludes there are ten. Now a third

party asked to explain what has occurred will no doubt point out that both Sam’s eyesight and his

ability to count were involved. For in the absence of either he would not have correctly

ascertained the number of spheres. But to arrive at something more determinate than this our

folk psychologist must run afoul of one of the two differing philosophical theories on hand, as

each offers a different account of just how looking and counting operated to affect Sam’s

response. On Martin’s account Sam has a veridical visual experience of the spheres, and

correctly counts the spheres that partly constitute that experience. According to the

representationalist, Sam has a veridical visual experience of the spheres and he correctly counts

the spheres his experience represents.

But now suppose that Sam isn’t asked for the number of spheres until the hallucination

has already begun, and that he again does his best to count and arrives at the conclusion that

there are ten spheres before him. Now in this case, the two theories will diverge. Despite the

change in distal stimulation, the representationalist’s explanation of how Sam arrives at his

35

answer will not change. Again Sam correctly counts the spheres his experience represents. But

the disjunctivist must now provide a radically different explanation. Sam cannot now count the

ten spheres that partially constitute his experience, as there are now no spheres to be counted. So

the disjunctivist must in this case appeal to a representation of the spheres, or, at the very least,

invoke something other than Sam’s counting spheres.

And yet, though the disjunctivist might appeal to an experiential representation to explain

Sam’s numerical judgment in the hallucinatory case, he cannot give the same explanation of the

conclusion Sam reaches when veridically perceiving. For to do so would be to treat

hallucination and veridical visual experience as instances of a common kind. It would turn out

that both of Sam’s experiences are representations of ten spheres: experiences that represent ten

spheres in such a way as to inform and render intelligible Sam’s efforts to count what is before

him.16

So we must ask: When our ordinary subject explains the two cases under discussion will

she embrace the disjunctivist’s conclusions? Again, I have only my own case to go on, but I

predict that you will join me in a negative answer. When explaining Sam’s answer in the

veridical case our folk psychologist will feel pressed to cite the very same relation between

looking and counting that obtains when Sam hallucinates. Again, Martin’s disjunctivism runs

into implausibility. We are left looking for a good reason to revise the common sense

psychological practice that we have identified.

To be clear, I am not denying that various facts about Sam’s thinking, reasoning, and

acting are best explained by adverting to the (necessarily spheres-involving) fact that he sees ten

spheres. His successfully sorting the spheres into two groups of five is surely a fact of this 16 Note that an operation of conceptualization can be shaped and informed by (or based on) an experience without its being about (or of) that experience. (Sam’s counting the spheres can be based on his experience of them without his therein counting his experience or elements of it.) See Evans (1982, 227) for discussion.

35

kind.17 But we have limited our view to cases in which “visual experience” is used to explain

someone’s judgments or behavior, so explanations in terms of “seeing spheres” or “seeing that

there are spheres” are beside the point. And it is instructive to note that the fact that Sam might

correctly sort the spheres in the veridical case couldn’t have a corollary in the hallucinatory case,

for when he is hallucinating, Sam can’t find any spheres to sort. Indeed, it is because Sam

counts those spheres he seems to see in precisely the same way in both of our initial scenarios

that we feel compelled to provide a uniform explanation of what has happened. The

commonality in effect calls out for a commonality in cause. We posit a single experiential origin

to explain a remarkable similarity in cognitive result.

7. Explaining the Nature of Experience

I have argued that Martin’s disjunctivist approach is incompatible with folk psychological

practice. But critics have also argued against Martin’s attempt to explain the nature of

experience in terms of veridical experience and events or states indiscriminable from it. My aim

in this section is to explain this criticism of Martin’s account, to describe Martin’s response to it,

and to explain why Martin’s response cannot be applied to the case we have already presented

against the disjunctivist view.

Consider, as an instance of the kind of criticism I have described, Susanna Siegel’s

(2004) argument against Martin’s equating visually experiencing a sausage with having an

experience introspectively indiscriminable from a veridical visual experience of a sausage, an

17 See Williamson (2000, 60-4) for cases in which we quite naturally cite an agent’s factive (or external-object-entailing) states of mind to explain her actions. Cf. Pettit (1986) and Child (1994, 204-16). Martin claims, without argument, that explanations of judgment and behavior phrased in terms of “perceptual states” (e.g. states of seeing something) are in “competition” with explanations in terms of “sensory experience as something common across the [veridical and hallucinatory] cases” (2006, 371-2). But he never considers the view, defended in the text above, that different explananda (counting v. sorting) demand differing explanans (experiencing v. seeing).

35

equation that is supposed (by Martin) to capture the idea that both hallucinations of sausages and

veridical visual experiences of sausages are indeed visual experiences of sausages even though

they fail to constitute a common (non-disjunctive) kind. Siegel argues that when we turn to a

classification of the hallucinations of non-human animals, the equation we have described will

come to grief.

She argues with a dilemma, the first horn of which has Martin characterizing

introspective indiscriminability in a positive manner, so that two experiences are introspectively

indiscriminable if the subject enjoying them is (positively) disposed to judge them the same

(2004, 96). Since dogs cannot introspect, they are not positively disposed to judge hallucinations

of sausage to be the same as veridical experiences of such. And yet, despite this failing, the

dog’s hallucination of a sausage supper is surely an experience of sausage. Thus, Martin must

reject the positive characterization of introspective indiscriminability on offer as too restrictive.

But this just brings Martin to rest on the dilemma’s second horn. For suppose that he

instead adopts a negative characterization of introspective indiscriminability by allowing that

two experiences are indiscriminable so long as the subject enjoying them is not disposed to judge

them different. The dog is not disposed to judge that his hallucination of sausage is different

form his veridical experience of such because he cannot make—and so isn’t disposed to make—

introspective judgments of any kind. So the negative characterization of introspective

indiscriminability correctly classifies both events as experiences of sausage. But then the dog

also lacks a disposition to judge that his veridical experiences of his ball and his bowl of water

differ from his veridical experiences of sausage. In consequence, the negative characterization

of introspective indiscriminability incorrectly classifies as experiences of sausage the dog’s

35

veridical visual experience of his ball and water bowl (2004, 97). In sum, the positive

characterization that Siegel considers is too strong, and the negative characterization too weak.

Martin (2006, 379-83) replies to this objection by invoking an “impersonal” notion of

introspective indiscriminability—a strategy explicated and endorsed by Scott Sturgeon (2006,

2008). To say that the dog’s perfect hallucinations of sausage are introspectively indiscriminable

from veridical visual experiences of such is not to say that the dog would judge them the same.

It is instead to say—speaking impersonally now—that these events just cannot be told apart after

even the most careful efforts to introspect. Introspecting a difference between them is, to use

Martin’s phrase, “impossible simpliciter,” and not just impossible for this or that subject or

species (2006, 381). Correlatively, to say that the dog’s veridicsl experiences of sausage are

introspectively discriminable from his veridical experiences of balls and bowls of water is not to

say that he can tell these experiences apart. It is just to say—without relativizing to the dog or

any other subject—that they can be told apart on the basis of introspection. When we are trying

to give an account of what sausage experience is, we should feel free to idealize in this manner,

Sturgeon suggests, just as we should feel to idealize when giving an account of doxastic

rationality in terms of logical omniscience and full probabilistic coherence (2006, 196-7).

Let us suppose that Martin and Sturgeon are right about this. And let us suppose, in

consequence, that being a visual experience of ten spheres in a circle can be equated with—or

shown to be necessarily co-extensive with—being introspectively indiscriminable from a

veridical visual experience of such. Still, what does this have to do with our efforts to explain

why Sam, when hallucinating the spheres before him, came to judge that there were ten spheres

before him? Perceptual concepts and perceptual judgments develop before introspective

concepts and judgments. So Sam might hallucinate his spheres and judge there to be ten spheres

35

before him without judging that he is having an experience of ten spheres and without actually

trying and failing to discriminate his experience from a veridical experience of ten spheres. So

what does the fact of the hallucination’s impersonal indiscriminability from a veridical visual

experience of ten spheres have to do with that hallucination’s motivating and rendering

intelligible the judgment he actually makes in the case on hand?

When doxastic rationality is supposed to consist in logical omniscience and full

coherence it is not at the same time supposed to account for our real-world, real-time judgments

and decisions. We invoke these idealizations to characterize the working of an ideal mind and

the norms to which many ordinary thinkers aspire, but we must instead cite the non-idealized

properties of the reasoning processes of (non-ideal) people to provide adequate explanations of

their real-world, real-time beliefs and actions. Indeed, modern economic theory is learning

precisely this lesson from behavioral economists. You can use idealized conceptions of

rationality when constructing a model of ideal economic activity. But if you can replace these

idealizations with a more realistic description of the processes that actually generate our beliefs

and choices you will invariably arrive at better explanations and predictions of our actions and

the economic system they embody.

So which event drove Sam to judge there to be ten spheres before him? His non-ideal,

all-too-real hallucination of ten spheres. And which features of this hallucination best explain

why it had this effect on his judgment? Its non-ideal, all-too-real representational and qualitative

features. Sam’s perfect hallucination is indeed impersonally introspectively indiscriminable

from a veridical visual experience of ten spheres. But when he doesn’t in fact introspect upon

that experience, and he doesn’t in fact mistake it for a veridical experience of ten spheres, the

observation that Sam’s hallucination is (in some impersonal sense) indiscriminable from a

35

veridical perception does not enter into (much less exhaust) the best folk psychological

explanation of the judgments to which this experience actually gives rise.

8. Scientific Explanation

What then of scientific psychology? I suspect that investigation of Sam’s nervous system would

find exactly the same interactions between the neural correlates (or realizations) of counting and

visual processing occurring both before and after Sam begins to hallucinate. That is, though it is

an empirical conjecture—a conjecture that could only be conclusively verified by impracticably

(and immorally) producing a series of perfect hallucinations through surreptitious cortical and

(perhaps) retinal stimulation—I suspect that there is no neurological difference between veridical

perception and perfect hallucination, nor any difference in the impacts these events have on the

neural realization of counting at a distance and the judgments of number in which such counting

issues. And since the neural realizations of perfect hallucination and veridical visual experience

will turn out to be identical, and the two will turn out to have identical impacts on many of the

forms of cognition to which vision is essential, neuroscientists will be tempted to infer that the

realizers of hallucination and veridical visual experience are species of a single neurological

genus. Won’t they then infer that this neurological genus is the correlate or realizer of visual

experience: the folk psychological genus of which veridical visual experience and hallucinatory

experience are both species? I suspect that they will, and that Martin’s claims about the intuitive

nature of naïve realism will not move them from this stance. Only an explanatory or predictive

35

advantage could tempt a neuroscientist to alter her interpretation of the results I think likely.

And Martin’s theory of experience offers neither.18

Of course, Martin might reject my predictions as unlikely or overly speculative (as would

Fish, 2009, ch. 5). Or he might just ignore them altogether. But even if he focuses exclusively

on folk explanation, the case against disjunctivism will not go away. For, as we have seen, we

folk do introduce a common experiential kind to explain commonalities in the perceptual

judgments of those who are seeing and those who are merely hallucinating. Neither folk nor

scientific psychology recommends Martin’s disjunctivist view.

9. Conclusion

I have argued at length that Martin is mistaken when he says that our reflective conception of

visual experience is only accurate if naïve realism is true. Introspection does not support but

actually undermines naïve realism. This is especially true of conceptual forms of introspection.

When a subject reflects on the fact that he is seeing something he thinks of what is happening in

a naively realistic light, but when he reflects on the distinct fact that he is having a veridical

visual experience of something, he thinks of what is happening in representational terms. Martin

thus needs to invoke non-conceptual introspective awareness of experience (c-introspection) to

motivate the naively realistic picture of veridical visual experience on which his disjunctivism is

based. But it is hard to see how Martin can countenance c-introspection while at the same time

18 Based on a careful examination of contemporary vision science, Tyler Burge (2005, 22) argues for a “Proximity Principle” on which, holding fixed background psychology, the same proximal stimulation of the visual system will yield tokens of the same perceptual state type regardless of differences in distal cause. (Burge’s phrase “proximal stimulation” is meant to include light striking the retina, though because this light needn’t have been reflected off the surface of an object, Burge countenances analogs of perfect hallucination: illusory perceptions—entertained in the absence of external objects—that have the same proximal origins as—and so are type identical to—the veridical experiences from which they are introspectively indiscernible.) Burge is surely right about the contemporary science, so the predictions I’ve made in the text are conservative.

35

rejecting substantive explanations of the introspective indiscernibility of veridical visual

experience and hallucination. Moreover, even if Martin were right that introspection supports

naïve realism, our explanatory practices would provide us with a strong reason to reject the

disjunctivist conception of visual experience to which naïve realism gives rise. If common

thought is any guide, our hallucinatory and veridical experiences share a positive

representational and qualitative character.

35

References

Aristotle. 384-322BC/1984. “Problems.” In The Complete Works of Aristotle, Vol. 2, edited by

J. Barnes, 1319-1527. Princeton, NJ: Princeton University Press.

Burge, Tyler. 1988. “Individualism and Self-Knowledge.” Journal of Philosophy 85.11: 649-63.

Burge, Tyler. 2005. “Disjunctivism and Perceptual Psychology.” Philosophical Topics 33.1: 1-

78.

Byrne, Alex. 2001. “Intentionalism Defended.” The Philosophical Review 110.2: 199-234.

Byrne, Alex and Heather Logue. 2008. “Either/Or.” In Disjunctivism: Perception, Action and

Knowledge, edited by A. Haddock and F. Macpherson, 57-94. Oxford: Oxford University

Press.

Chalmers, David and Andy Clark. 1998. “The Extended Mind.” Analysis 58: 10-23.

Child, William. 1994. Causality, Interpretation and the Mind. Oxford: Clarendon.

Evans, Gareth. 1982. The Varieties of Reference, edited by J. McDowell. Oxford: Clarendon.

Fish, William. 2009. Perception, Hallucination, and Illusion. New York: Oxford University

Press.

Foster, John. 2000. The Nature of Perception. Oxford: Clarendon.

Hawthorne, John and Karson Kovakovich. 2006. “Disjunctivism.” Proceedings of the

Aristotelian Society, Supplementary Volume 80: 145-83.

Hellie, Benj. 2007. “Factive Phenomenal Characters.” Philosophical Perspectives: Philosophy of

Mind 21: 259-306.

Hinton, J. M. 1967. “Visual Experiences.” Mind 76: 217-27.

Hume, David (1739-40/2000) A Treatise of Human Nature, edited by D. F. Norton and M. J.

Norton. Oxford: Oxford University Press.

Hyman, John. 1999. “How Knowledge Works.” The Philosophical Quarterly 49: 433-51.

Hyman, John. 2001. “Knowledge and Self-Knowledge.” In Wittgenstein and Contemporary

Philosophy of Mind, edited by S. Schroeder, xx-xx. Basingstoke: Palgrave.

Jackson, Frank. 1977. Perception: A Representative Theory. Cambridge: Cambridge University

Press.

Martin, M. G. F. 2004. “The Limits of Self-Awareness.” Philosophical Studies 120: 37-89.

Reprinted in Disjunctivism: Contemporary Readings (2009), edited by A. Byrne and H.

Logue, 271-317. Cambridge, MA: MIT Press.

35

Martin, M. G. F. 2006. “On Being Alienated.” In Perceptual Experience, edited by T. S. Gendler

and J. Hawthorne, 354-410. Oxford: Clarendon.

Pettit, Philip. 1986. “Broad Minded Explanation and Psychology.” In Subject, Thought and

Context, edited by P. Pettit and J. McDowell, 17-58. Oxford: Clarendon.

Putnam, Hilary. 1975. “The Meaning of ‘Meaning’” in Minnesota Studies in Philosophy of

Science, Vol. VII: Language, Mind and Knowledge, edited by K. Gunderson, xx-xx.

Minneapolis: University of Minnesota Press. Reprinted in Putnam’s Mind, Language, and

Reality: Philosophical Papers Vol. 2 (1975), Cambridge: Cambridge University Press, 215-

71.

Robinson, Howard. 1994. Perception. London: Routledge.

Shoemaker, Sydney. 1996. The First Person Perspective and Other Essays. Cambridge:

Cambridge University Press.

Siegel, Susanna. 2008. “The Epistemic Conception of Hallucination.” In Disjunctivism:

Perception, Action and Knowledge, edited by A. Haddock and F. Macpherson, 205-24.

Oxford: Oxford University Press.

Sosa, Ernest. 2007. A Virtue Epistemology: Apt Belief and Reflective Knowledge, Vol. 1.

Oxford: Oxford University Press.

Sturgeon, Scott. 2006. “Reflective Disjunctivism.” Proceedings of the Aristotelian Society,

Supplementary Volume 80: 185-216.

Sturgeon, Scott. 2008. “Disjunctivism about Visual Experience.” In Disjunctivism: Perception,

Action and Knowledge, edited by A. Haddock and F. Macpherson, 112-43. Oxford: Oxford

University Press.

Tye, Michael. 2000. Consciousness, Color, and Content. Cambridge, MA: MIT Press.

Tye, Michael. 2009a. “The Admissible Contents of Visual Perception.” The Philosophical

Quarterly 59.236: 541-62.

Tye, Michael. 2009b. Consciousness Revisited. Cambridge, MA: MIT Press.

Williamson, Timothy. 2000. Knowledge and Its Limits. Oxford: Oxford University Press.

Zimmerman, Aaron. 2006a. “Self-Verification and the Content of Thought.” Synthese: 149: 59-

75.

Zimmerman, Aaron. 2006b. “Basic Self-Knowledge: Answering Peacocke’s Criticisms of

Constitutivism.” Philosophical Studies 128: 337-379.

35

Zimmerman, Aaron. 2008. “Self-Knowledge: Rationalism v. Empiricism.” Philosophy Compass

3/2, 325-352.

35