". . .The fact that things overlap or are hidden
not enter into their definition, and expresses
only my incomprehensible solidarity with one
of them — my body."
Maurice Merleau-Ponty, from "Eye and mind" (1964, p. 173)

have mentioned several times the idea that, when we perceive a picture drawn in perspective from a vantage point other than the center of projection, our perceptual system infers the location of the center of projection and we feel that we are looking at the depicted scene from the vantage point implied by the center of projection. To explain the meaning of such a suggestion, I must first introduce the concept of egocenter. I will then discuss the question of a movable egocenter.

The notion of a spatially localized, visual egocenter that does not coincide with either eye is due to W. C. Wells (1792, cited by Ono, 1981), who was the first to devise a way of locating what came later to be called the "cyclopean eye." One simple method is this: Hold your head still while a friend stands a few feet away and points a stick at you. Have the friend change the position of the stick until you feel it pointing at you perfectly. Record the exact orientation of the stick. Now without moving your head ask your friend to stand somewhere else in the room, to the right or left of where he or she stood before, and adjust the orientation of the stick until it is pointing at you. If you do this several times, and you prolong the lines defined by the various positions of the stick when it is pointed at you, you will find that they all intersect approximately at one point inside your head just behind the midpoint of the line connecting your two eyes. This is the position of what is sometimes called the sighting egocenter (Howard, 1982, pp. 283-91).

In his philosophical essay "Where am I?", which may be the most amusing science-fiction story ever written, Daniel Dennett (1980) proposes a thought experiment: Imagine a surgical procedure that extirpates your brain from your head and connects radio transmitters to the stumps of the nerve cells that carry information from the brain to the rest of the body and radio receivers tuned to the same frequency on the complementary segments of these nerve cells in the body; similarly, this procedure connects radio transmitters to the stumps of the nerve cells that carry information from the body to the brain and appropriately tuned receivers of the sensory nerve cells in the brain. This is no more than, as one of the characters in Dennett's story puts it, ßtretching the nerves"^1. After the operation, as soon as he is strong enough to be taken to see his brain, the hero of the story asks himself why he feels that he is outside the vat looking at his brain, rather than inside his brain being looked at by his eyes. After all, he argues, mental events arc instantiated in the brain, so why does he not feel that he is where his mental events are instantiated? Although we should not take the "results" of Dennett's thought experiment, however plausible, too seriously, it is tempting to infer from them that the reason we feel that we are inside our heads or our bodies is not because all the important bodily or mental processes occur inside our body's skin; the experiment suggests that the spatial location of the machinery that makes mental events possible is irrelevant to our feeling of location. But if the location of the brain does not determine where we feel ourselves to be, what does? Perhaps it is the physical

boundaries of our body that determine v, here w r feel ;v c are. Perhaps we feel that we arc inside ourselves because our skin is where the outside ends. This simple answer can only be part of the truth. For where does our body end and the world begin? If you are walking in the dark feeling your way about with a cane, you are unaware of the pressure of the cane on the palm of your hand; all your attention is focused on the nature of the obstacles revealed by the of the cane. Under these circumstances, if you had to classify the cane as part of the world or part of your body, you would most likely say that it was part of your body. This is true of all tools. It is also true of vehicles. Most of the time when you drive an automobile, you are not aware of your points of contact with the inside of the automobile; it is as if you had grown a shell around you that you now inhabit and that your points of contact with the environment now coincide with the body of the automobile. Thus it is the external boundaries of your auto body and not the spaciousness of the car's interior that determine your feeling of how big a car you are in. In short, the boundary between the world and ourselves is extremely flexible.

Just how flexible this boundary is becomes clear when we consider the readiness with which we adopt virtual viewpoints in a movie theater. When the camera pans (see Figure 10.1), we feel ourselves turning to scan the environment; when the camera tilts, we feel ourselves tilting our heads to look upward or downward; when the camera engages in a tracking or traveling shot (e.g., when the camera is set on wheels or tracks), we feel ourselves traveling forward or backward with the camera. And yet we know all along that we are sitting in a movie theater.

In fact, as Michael Roemer clearly shows, the use of virtual points of view can make the difference between an effective but relatively shallow image and one that endures:

Audiences can be "played" by a skillful movie-maker with a fair amount of predictability, so that even discriminating audiences are easily taken in. At the beginning of Bergman's Wild Strawberries Professor Berg dreams that he is on a deserted street with all its doors and windows shuttered tight. He looks up at a clock that has no hands and pulls out his own watch only to find that its hands are missing also. A man appears on the corner with his head averted; when he turns, he has no face and his body dissolves into a pool on the sidewalk. A glass hearse comes down the street and spills a coffin that opens. Berg approaches and discovers his own body in the coffin. The corpse comes to life and tries to pull him in.

The nightmare quality in this sequence is derivative. The deserted, shuttered street, the clock and watch without hands, the glass hearse, the faceless man are all conventions familiar to surrealist painting and literature. Bergman uses them skillfully and with conviction to produce an effect in the audience, but they are not true film images, derived from life and rendered in concrete, physical terms.

There is a similar nightmare in Dreyer's Vampire. A young man dreams that he has entered a room with an open coffin in it. He approaches and discovers that he himself is the corpse. The camera now assumes the point-of-view of the dead man: we look up at the ceiling. Voices approach and two carpenters appear in our field of vision. They close the coffin with a lid but we continue to look out through a small glass window. Talking indistinctly, they nail down the lid and plane the edges of the wood. The shavings fall onto the window. One of them has put a candle down on the glass and wax drips onto it. Then the coffin is lifted up and we pass close under the ceiling, through the doorway, beneath the sunlit roofs and the church steeple of o small town - out into the open sky.

Here the detail is concrete: an experience is rendered, not cited; she situation is objective and out of it emerges, very powerfully, the feeling that Dreyer is after: a farewell to life, a last confined look at the earth before the coffin is lowered into the grave. (1966, pp. 259-60)

¹ There is, to be sure, considerably more involved in performing such a technological feat, such as ensuring that the blood that flows through the extirpated brain has exactly the same composition as the blood coursing through the brainless body, because could never get drunk, suffer from premenstrual tension, or become sexually aroused.