I. Introduction: The Problem of Perceptual Knowledge

Try the following experiment. Hold a finger in front of your face. Focus your eyes on the finger, but attend to a distant object in the background. If you're doing this right, the background object should appear double and blurry. If you now bring the background object into focus, you will see the finger in your visual field split into two, blurry fingers.

There is, of course, a scientific explanation for why this happens. It has to do with the fact that each of your eyes has a different vantage point on the room; one of the finger-images is produced by your left eye, and the other by your right eye. But the scientific explanation is not our concern here. Our concern is with the philosophical implications that may be drawn from this phenomenon, about the relationship between perception and reality. Obviously, there are not really two fingers in the physical world. Nevertheless, you are "seeing" two of something. Therefore, we ask: What is it that there are two of?

A number of philosophers have put forward the following answer: There are two images of the finger in your mind. It is these images--rather than the actual, physical finger--that you are directly aware of; that is why there appear to be two fingers. Now consider the implications of this position. The finger in your visual field can make a continuous transition from being clearly in focus to being clearly out of focus. Therefore, if the blurry "fingers" that you see are really only images in the mind, it seems that the in-focus "finger" is also an image in the mind. In fact, since the same experiment could be performed with any object you see, all of the objects you are now seeing are images existing in your mind--which you are in the habit of mistaking for real, physical objects.(1)

This idea will probably strike you--it certainly ought to strike you--as quite extraordinary, and perhaps unbelievable. I can see a desk-shaped object in front of me right now; if I reach out with my hand, I seem to feel that same object; it is rectangular, solid; if I sit on it, it will support my weight. Surely this thing is no mere image! Nevertheless, let us pursue the theory to its logical conclusion.

There are similar arguments for the rest of the five senses, to show that what we directly perceive is always an image or "representation" in our minds. Now, if the immediate objects of awareness in perception are always mental images, the question inevitably arises, What reason have we for believing anything other than the mental images exists? Your experience could be qualitatively the same as it is now without there being a physical finger--that is, if only the image existed.

A common way to illustrate this possibility is through the following scenario. Suppose a group of scientists in a technologically advanced future figure out how to remove a human brain from its body and keep it alive in a vat of nutrients. Furthermore, the scientists have figured out exactly how to electrically stimulate the brain's sensory cortex so as to induce sensory experiences qualitatively identical with the experiences of a normal person, say, reading a book. Theoretically, this is possible; all of the brain's "information" about the world (from which it "constructs" our mental images) comes from electrical signals coming into the brain from various nerves. Imagine the scientists duplicate the normal pattern of electrical stimulation. To the brain, everything would appear normal; the brain would have no way of knowing that it was floating in a vat, and so on. In other words, the brain would be experiencing sensory images just like the images you are now experiencing, although there would be no physical objects in the real world corresponding to those images. The brain could "read" a nonexistent book, "walk" down a nonexistent street, and so on. (I use scare quotes because the brain would not really be reading or walking, only seeming to do so.) This sort of scenario is also illustrated in popular movies such as Total Recall and The Matrix, which the reader may wish to view to make the issues more vivid.

All of this is leading up to the question: How do you know that you aren't, right now, a brain in a vat?

Some few philosophers have been led by this kind of consideration to the conclusion that we can never know about the objective world--that is, the world outside of our minds; we can only know, at best, about our mental representations of the world (the "subjective" world). This position is known as "skepticism"--more specifically, it is "skepticism about the external world."

The problem of perceptual knowledge is the problem of how we know that the objects we seem to perceive are real. That is the focus of this book. There are four main positions philosophers have taken on this issue: two forms of realism and two forms of antirealism.

Realists maintain that there is an objective world, existing independent of our perception (a.k.a. "the real world") and that human beings can know about the nature of this world. The two forms of realism are as follows:

1. Direct realists maintain--contrary to the argument given above--that we are directly aware of real, physical objects in perception and that this explains how we know about the nature of those objects.

2. Indirect realists hold, instead, that our awareness of the real world is indirect. They accept arguments like the one given above, which says that what we are immediately aware of in perception is only mental images;(2) however, they say that we can infer the existence of real objects corresponding to our images, because that is the best explanation for why we have the sort of mental images we do.

Realism has two kinds of opponents:

3. Idealists hold that there is no objective world; there is only the mind and the images, thoughts, feelings, and so on in the mind. (This is called "idealism" because the mental images used to be called "ideas.")(3)

4. Skeptics hold that we cannot know that there is an objective world nor, if there is one, what it is like.

In the subsequent chapters, I will defend direct realism against all comers. I will argue that perception makes us directly aware of real, physical things, not mental images. I will develop a theory of epistemic justification to show why we are justified in believing in such objects. And I will show that the arguments philosophers have given in favor of our perceiving mental images--including the above argument from the case of double vision, along with several others--are fallacious, as are the arguments used to show that we cannot know about the objective world.

In a way, I am taking up Descartes's project, or at least a part of his project, but I propose an entirely different way of carrying it out. Descartes reflected that there were many beliefs he had accepted, perhaps foolishly, when he was a child and too young to possess sound judgement. He therefore resolved to make a general examination of his beliefs, to root out the false or unjustified ones and to put his whole belief system on a rational foundation.(4) He began with a consideration of the arguments for philosophical skepticism, which, by and by, he tried to refute, in the course of establishing his knowledge of the external world. Descartes also sought a kind of absolute certainty with which I do not concern myself here, but the essential point I share with him is the desire to verify the rationality of my overall system of beliefs, particularly in the face of the reasons for doubt raised by philosophical skeptics--the sort of reasons I discuss above and that Descartes discussed at the outset of his famous Meditations. I have always regarded Descartes's project, since I first learned of it, as both a natural one and one of the foremost intellectual importance. It is what drew me to the discipline of epistemology at first and, ultimately, motivated this book.

This is probably because I took the arguments for skepticism more seriously than Descartes himself did--though I have since come closer to his view of them. Most of those who read Descartes's Meditations find the skeptical arguments he presents at the outset far more powerful and convincing than his alleged refutation of them later in the book. Yet Descartes appears to have considered them as little more than an academic exercise, for, referring to his arguments for the existence of the external world, he writes:

The great benefit of these arguments is not that they prove what they establish--namely that there really is a world, and that human beings have bodies and so on--since no sane person has ever seriously doubted these things. The point is that in considering these arguments we come to realize that they are not as solid or as transparent as the arguments which lead us to knowledge of our own minds and of God, so that the latter are the most certain and evident of all possible objects of knowledge for the human intellect. Indeed, this is the one thing that I set myself to prove in these Meditations.(5)

Yet generations of philosophy students, myself included, have studied Descartes for his discussion of the arguments for skepticism, leaving his arguments about God by the wayside.

When a person first hears about the brain-in-a-vat scenario, he is apt to have one of three reactions. Reaction #1: "That's stupid. I refuse to talk about that." Reaction #2: "Gosh, maybe I am a brain in a vat. How would I know?" Reaction #3: "What is wrong with this argument? And what can I learn from that about the nature of knowledge?" I would like to encourage you to cultivate reaction #3. Descartes thought one could learn (not from the brain-in-a-vat scenario, but from others like it) that the knowledge of one's own mind and of God was more fundamental and unquestionable than the knowledge of the physical world and that the latter rested on the former. I think, on the contrary, that Descartes's solution to his skeptical problems failed and that one can learn something else from that: that Descartes's theory of perception and perceptual knowledge was fundamentally wrong; that we need to adopt a different theory, a direct realist theory. That is the case I would like to make in this book.


1. Hume, Enquiry, 152; Treatise, 202, 210-11.

2. I do not mean that all indirect realists accept the argument from double vision in particular, but just that they accept arguments with similar conclusions. See chapter VI for survey of indirect realist arguments. Also, not all indirect realists have sought to rest belief in external objects on inference to the best explanation, as suggested in the text--Descartes did not, but he is the only exception I know of.

3. See Berkeley's Principles and Dialogues. I do not discuss idealism at any length in this book, because almost no one believes it, but I think that it rests upon the same mistakes as representationalism.

4. Meditations, opening paragraph.

5. Synopsis of the Meditations, 11.