The most fundamental difference between phantom limbs and other hallucinations is that they can be moved voluntarily, whereas visual and auditory hallucinations proceed autonomously, outside one's control. This was also emphasized by Weir Mitch.e.l.l:

[The majority of amputees] are able to will a movement, and apparently they themselves execute it more or less effectively.... The certainty with which these patients describe their [phantom motions], and their confidence as to the place a.s.sumed by the parts moved, are truly remarkable ... the effect is apt to excite twitching in the stump.... In some cases the muscles which act on the hand are absent altogether; yet in these cases there is fully as clear and definite a consciousness of the movement of the fingers and of their change of positions as in cases [where the muscles of the hand are partially preserved].

Other hallucinations are only sensations or perceptions, albeit of a very special sort, whereas a phantom limb is capable of phantom action. Given a suitable prosthesis, the phantom limb will slip into the prosthesis ("like a hand into a glove," as many patients say)-slip into it and animate it, so that the artificial limb can be used like a real one. Indeed, this must happen if one is to use a prosthesis effectively. The artificial limb becomes part of one's body, of one's body image, as a cane in a blind man's hand becomes an extension of himself. One may say that an artificial leg, for instance, "clothes" the phantom, allows it to be effective, gives it an objective sensory and motor existence, so that it can often "feel" and respond to minute irregularities in the ground almost as well as the original leg.5 (Thus the great climber Geoffrey Winthrop Young, who lost a leg during World War I, was able to climb the Matterhorn using a prosthetic limb of his own design.)6

One might go further and say that a phantom is a portion of body image which is lost or dissociated from its natural, embodying home (the body)-and, as such, as something extraneous, it may be intrusive or deceptive (thus the danger of walking off a curb with a phantom leg). The lost phantom (if one can speak figuratively) longs for a new home, and it will find this in a suitable prosthesis. I have had many patients tell me how they may be disturbed by their phantom at night but relieved in the morning, for the phantom disappears the moment they put on their prosthesis-disappears, that is, into the prosthesis, merging so seamlessly with it that phantom and prosthesis become one.

Knowledge of what one is doing with one's phantom-even without a prosthesis-can be exquisitely refined. As a young student, Erna Otten, a distinguished pianist, was a pupil of the great Paul Wittgenstein, who lost his right arm in the First World War but continued to play with his left hand (and commissioned a number of composers to write music for the left hand). Yet he continued to teach, in a sense, with both hands. In a letter to the New York Review of Books, responding to an article I had written, Otten wrote:

I had many occasions to see how involved his right stump was whenever we went over the fingering for a new composition. He told me many times that I should trust his choice of fingering because he felt every finger of his right hand. At times I had to sit very quietly while he would close his eyes and his stump would move constantly in an agitated manner. This was many years after the loss of his arm.

Unfortunately, not all phantoms are as well formed, as painless, or as mobile as Wittgenstein's. Many show a tendency to shrink or "telescope" with time-a phantom arm may be reduced to a hand seemingly sprouting from the shoulder. This tendency to shrink is minimized by embedding the phantom in a prosthesis and using it as much as possible. A phantom may also become paralyzed or contorted in painful positions, with its "muscles" in spasm. Thus Admiral Lord Nelson, after losing his right arm in battle, developed a phantom limb with the hand permanently clenched, the fingers digging excruciatingly into the palm.7

Such disorders of body image have long seemed inexplicable and untreatable. But over the last few decades, it has become clear that the body image is not as fixed as we once thought; indeed, it is remarkably plastic, and extensive reorganization or remapping can occur with phantom limbs.

If there is interruption of nerve function from injury or disease in the spinal cord or peripheral nerves, cutting off or reducing normal sensory input to the brain, this may cause major disturbance of body image, with strange phantom images superimposed on the real but insentient body parts. This was very striking with a colleague of mine, Jeannette W., who broke her neck in a car accident and became quadriplegic, with a complete absence of sensation below the level of the fracture. She had, in a sense, been "amputated" from the neck down and had little sense of her body below this. But in its place, she had a phantom body, which was unstable and p.r.o.ne to distortions and deformations. She could reverse these, for a while, by seeing that her body still had a normal shape and conformation, and she arranged for mirrors to be set up in her office and in the hospital corridors, so that she could glance up and (in her words) take "visual sips" from them as she bowled past in her wheelchair.

As normal sensation is blocked, body image disturbances can occur very quickly. Most of us have had strange phantom experiences with dental anesthesia, of a grotesquely swollen, deformed, or misplaced cheek or tongue. Looking in a mirror will do little to dispel these illusions, which disappear only with the return of normal sensation. One patient of mine, with the removal of a large brain tumor, had to sacrifice the roots of the sensory nerves on one side of her face. For years following this, she had a persistent sense that the whole right side of her face was "slipping," "caved in," or "missing"; that her tongue and cheek on this side were tremendously swollen and grotesque-looking. She later came to have a leg amputated, and soon after surgery became aware of a phantom leg. Now, she said, "I know what's wrong with my face. It's exactly the same feeling-I have a phantom face."

There can also be extra limbs-supernumerary phantoms-if certain areas of the body are denervated. A striking example of this was described by Richard Mayeux and Frank Benson. Their patient was a young man with multiple sclerosis who developed a numbness on his right side and then experienced, as they wrote,

a tactile illusion that a second right arm was lying across his lower chest and upper abdomen. The extra arm seemed to be attached to the chest wall.... There was only a vague sensation of the duplicate illusory lower forearm, wrist, and palm, but a vivid impression of the fingers lying on the abdominal wall.... The illusion persisted for period of 5 to 30 minutes and was accompanied by a "gripping" sensation of the illusory hand.... The phantom limb sensation was always coincident with feelings of increased stiffness, numbness, and burning [sensations] of the actual right arm.

Nelson's clenched hand exemplifies an unpleasant evolution which phantom limbs may undergo-phantoms which are initially loose, mobile, and obedient to the will may subsequently become paralyzed, contorted, and often intensely painful. Before the 1990s, there was no plausible explanation as to why phantom limbs might get frozen in this way, nor any notion of how to unfreeze them. But in 1993, V. S. Ramachandran suggested a physiological scenario which might explain the progressive loss of voluntary movement so common in phantom limbs. The vivid sense that one could move a phantom limb freely, he thought, went with the brain being able to monitor its own motor commands to the phantom. But with the continuing absence of visual or proprioceptive confirmation of movement, the brain, in effect, might "abandon" the limb. Thus, Ramachandran thought, paralysis was "learned," and he wondered whether it could be unlearned.

Could one, by simulating visual and proprioceptive feedback, dupe the brain into believing that the phantom was once again mobile and capable of voluntary movement? Ramachandran developed a brilliantly simple device-an oblong wooden box with its left and right sides divided by a mirror, so that looking into the box from one side or the other, one would get an illusion of seeing both hands, where in reality one was seeing only one hand and its mirror image. Ramachandran tried this device on a young man who had had a partial amputation of his left arm-his now-rigid phantom hand, Ramachandran wrote, "jutted like a mannequin's resin-case forearm out of the stump. Far worse, it was also subject to painful cramping that his doctors could do nothing about."

After explaining what he had in mind, Ramachandran asked the young man to "insert" his phantom arm to the left of the mirror. Ramachandran described this in his book The Tell-Tale Brain:

He held out his paralyzed phantom on the left side of the mirror, looked into the right side of the box and carefully positioned his right hand so that its image was congruent with (superimposed on) the felt position of the phantom. This immediately gave him the startling visual impression that the phantom had been resurrected. I then asked him to perform mirror-symmetric movements of both arms and hands while he continued looking into the mirror. He cried out, "It's like it's plugged back in!" Now he not only had a vivid impression that the phantom was obeying his commands, but to his amazement, it began to relieve his painful phantom spasms for the first time in years. It was as though the mirror visual feedback (MVF) had allowed his brain to "unlearn" the learned paralysis.

This extremely simple procedure (which was devised only after much careful thinking and a whole, very original theory as to the many interacting factors involved in the production of phantoms and their vicissitudes) can easily be modified for dealing with phantom legs and a variety of other conditions involving distortion of body image.

The appearance of the hand moving, the optical illusion, was sufficient to generate the feeling that it was moving. I described the converse of this in The Mind's Eye, when the existence of a large blind spot in my visual field allowed me, visually, to "amputate" a hand. But if, when I had done this, I opened and closed my fist or moved my now-invisible fingers, a sort of pink protoplasmic extension grew out of my visual "stump" and developed into a (visual) phantom of the hand.

Jonathan Cole and his colleagues have made similar observations, testing a virtual reality system to reduce phantom pain. In their experiments with leg and arm amputees, the amputated stump is connected to a motion capture device, which in turn determines the movements of a virtual arm or leg on a computer screen. Most of their subjects learned to correlate their own movements with those of the on-screen avatar, and developed a sense of agency or ownership, so that they were able to move the virtual limb with surprising delicacy (for instance, to reach for and grasp a virtual apple lying on the surface of a virtual table). Such learning occurred remarkably quickly, within half an hour or so. With this sense of agency and intentionality often came a reduction in phantom pain-and even virtual perception. One man, for example, could "feel" the virtual apple when he picked it up. Cole and his colleagues wrote, "Perception was not only of motion of the limb but also of touch, a virtual-visual cross-modal perception."

In 1864, Weir Mitch.e.l.l and two of his colleagues put out a special circular from the Surgeon General's Office, ent.i.tled Reflex Paralysis. In reflex paralysis, the injured limb is intact, but it cannot be moved; it seems absent or "alien," not part of the body. It is, in a sense, the opposite of a phantom limb-an external limb with no internal image to give it presence and life.

I had such an experience in 1974 during the mountaineering accident in which I ruptured the quadriceps tendon in my left leg. Though the tendon was repaired surgically, there was damage at the neuromuscular junction, and additionally, the leg was hidden from sight and touch, immobilized in a long, opaque cast. Under these circ.u.mstances, where it was impossible to send commands to the injured muscle and there was no sensory or visual feedback, the leg disappeared from my body image, leaving (so it seemed to me) an inanimate, alien thing in its place. This continued to be the case for thirteen days. (Thinking back on this experience, I wonder whether one of Ramachandran's mirror boxes would have helped me to recover movement, and a sense of reality, in this leg sooner. It might have helped, too, had the cast been transparent, so that I could at least see the leg.)

It was an experience so uncanny that I wrote an entire book, A Leg to Stand On, about it. I suggested, only half-jokingly, that readers would more easily imagine such experiences if they read the book under spinal anesthesia, for as the anesthetic blocks activity in the spinal cord, one's lower half becomes not only paralyzed and senseless but, subjectively, nonexistent. One feels that one's body terminates in the middle, and that what lies below-hips and a pair of legs-do not belong to one; they could just as well be a wax model from an anatomy museum. This lack of ownership, this alienation, is bizarre to experience. I found it almost intolerable during the thirteen days in which my left leg seemed alien to me-I wondered, darkly, whether any recovery would occur and whether, if it did not, I would do best to have the useless leg removed.

There may indeed, though very rarely, be a congenital absence of body image in an otherwise normal limb; this is suggested, at least, by the numerous reported cases of what Peter Brugger has termed "body-integrity ident.i.ty disorder." Such people feel, from childhood onward, that one of their limbs, or perhaps a part of a limb, is not theirs, but an alien enc.u.mbrance, and this feeling may engender a pa.s.sionate desire to have the "superfluous" limb amputated.

Prior to 1990, the whole field of phantom limbs and other disturbances of body image could be studied only phenomenologically, from the accounts and behaviors of those afflicted. Such conditions were often ascribed to hysteria or an overactive imagination, but the development of sophisticated brain imaging has changed this by showing the physiological changes in the brain (especially in parts of the parietal lobes) which underlie such strange experiences. This, along with ingenious experiments such as Ramachandran's mirror box, has allowed us to get a clearer view of the neural basis of embodiment, of agency, of self; to bring purely clinical and sometimes purely philosophical ideas into the realm of neuroscience.

Shadows" and "doubles"-hallucinatory distortions of the body and body image-take us into an even stranger realm. If a limb or part of the body is "deanimated" by nerve or spinal cord damage, the deanimated part itself may feel lifeless, inorganic, alien. But if there is damage to the right parietal lobe, a much deeper form of estrangement may occur. The deanimated part of the body-if its existence is acknowledged at all-is felt to belong to someone else, a mysterious "other." Many years ago, as a medical student, I saw a patient who had been admitted to the neurosurgery service for removal of a parietal lobe tumor. One evening, while awaiting surgery, he fell out of bed in a peculiar way-almost, the nurses said, as if he had thrown himself off the bed. When I asked him about this, he said that he had been asleep and awoke to discover a leg-a dead, cold, hairy leg-in his bed. He could not think how someone else's leg had got into his bed, unless-the idea suddenly occurred to him-the nurses had taken a leg from the anatomy labs and slipped it into his bed as a joke. Shocked and repelled, he used his good right leg to kick the alien thing out of his bed, and, of course, he came out after it, and was now aghast because "it" was attached to him. I said, "But it is your leg," and pointed out to him that the size, the shape, the contour, the color were precisely the same in the two legs; but he would have none of it. He was absolutely certain that it was someone else's.8

Over the years I have seen other patients who, in consequence of a right-hemisphere stroke, have lost all feeling and use of the left side. Often they have no awareness that anything has happened, but some people are convinced that their left side belongs to someone else ("my twin brother," "the man next to me," even "It's yours, Doc, who are you kidding?"). Perhaps "my twin brother" is a hieroglyphic way of indicating that while half of the body seems alien, it also seems very akin, almost identical to oneself ... that it is oneself in a strange, disguised way. It needs to be emphasized that such patients may be highly intelligent, lucid, and articulate-and that it is solely in reference to their odd distortions of body image that they make their surreal but irrefragable statements.

The feeling that someone is there, to the left or the right, perhaps just behind us, is known to us all. It is not just a vague feeling; it is a distinct sensation. We may wheel around to catch the lurking figure, but there is no one to be seen. And yet it is impossible to dismiss the sensation, even if we have learned from repeated experience that this sort of sensed presence is a hallucination or an illusion.

The sensation is commoner if one is alone, in darkness, perhaps in unfamiliar surroundings, hyperalert. It is well known to mountaineers and polar explorers, where the vastness and danger of the terrain, the isolation and exhaustion (and, in the mountains, reduced oxygen) contribute to the feeling. The sensed presence, the invisible companion, the "third man," the shadow person-all sorts of terms are used-is well aware of us, and has definite intentions, whether these are benign or malignant. The shadow stalking us has something in mind. And it is this sense of its intentionality or agency which either raises the hair on our neck or produces a sweet, calm feeling of being protected, not alone.

While the sense of "somebody there" is commoner in the hypervigilant states induced by some forms of anxiety, by various drugs and by schizophrenia, it may also occur in neurological conditions. Thus Professor R. and Ed W., who both have advancing Parkinson's disease, have persistent feelings of a presence-something or someone they never actually see; this presence is always on the same side. There may be a transitory sense of "someone there" in attacks of migraine or in seizures-but a very persistent sense of a presence, always to the same side, is suggestive of a brain lesion. (This is also the case with such experiences as deja vu, which we all have occasionally, but which, if very frequent, suggests a seizure disorder or a brain lesion.)

In 2006 Olaf Blanke and his colleagues (Shahar Arzy et al.) described how, with a young woman being evaluated for surgical treatment of epilepsy, they could predictably induce a "shadow-person" by electrical stimulation of the left temporoparietal junction. When the woman was lying down, a mild stimulation of this area gave her the impression that someone was behind her; a stronger stimulation allowed her to define the "someone" as young but of indeterminate s.e.x, lying down in a position identical to her own. When stimulations were repeated with her in a sitting position, embracing her knees with her arms, she sensed a man behind her, sitting in the same position and clasping her with his intangible arms. When she was given a card to read for a language learning test, the sitting "man" moved to her right side, and she understood that he had aggressive intentions ("He wants to take the card.... He doesn't want me to read."). There were thus elements of the "self" here-the mimicking or sharing of her postures by the shadow person-as well as elements of the "other."9

That there may be some connection between body-image disturbances and hallucinatory "presences" was brought out as early as 1930 by Engerth and Hoff, as Blanke and his colleagues wrote in a 2006 paper. Engerth and Hoff described an elderly man who had become hemianopic after a stroke. He saw "silver things" in the blind half of his visual field, then automobiles coming at him from the left, and then people: "countless" people, all identical in appearance and with a clumsy gait, staggering, with the right arm outstretched-precisely the gait the patient himself had when he tried to walk and avoid colliding with people on his left.

But he also had alienation of his left side, and he felt that this side of his body was "filled with something strange."

"Finally," Engerth and Hoff wrote, "the host of hallucinations disappeared, and there then appeared what the patient called 'a constant companion.' Wherever the patient went, he saw someone walking along on his left.... At the moment when the companion appeared, the alien feeling in the left half of the body disappeared.... We would not be in error," they concluded, "if we saw in this 'companion' the left half of the body which had become independent."

It is not clear whether this "constant companion" is to be cla.s.sified as a "sensed presence" or an autoscopic "double"-it has qualities of both. And perhaps some of these seemingly distinct categories of hallucination merge. Blanke and his colleagues, writing in 2003 of body-image, or "somatognosic," disorders, observed that these may take a number of forms: illusions of a missing body part, a transformed (enlarged or shrunk) body part, a dislocated or disconnected body part, a phantom limb, a supernumerary limb, an autoscopic image of one's own body, or a "feeling of a presence." All of these disorders, Blanke stresses, with their hallucinations of vision, touch, and proprioception, are a.s.sociated with parietal or temporal lobe damage.

J. Allan Cheyne has also investigated sensed presences, both in the relatively mild form that may occur when one is fully conscious and in the terrifying form that is often a.s.sociated with sleep paralysis. He speculates that this feeling of "presence"-a universal human (and perhaps animal) sensation-may have a biological origin in "the activation of a distinct and evolutionary functional 'sense of the other' ... deep within the temporal lobe specialized for the detection of cues for agency, especially those potentially a.s.sociated with threat or safety."

Sensed presence not only has its place in the neurological literature; it also forms a chapter in William James's Varieties of Religious Experience. He recounts a number of case histories where the initially horrible feeling of an intrusive and threatening "presence" became a joyful and even blissful one, including that of a friend who told him:

It was about September, 1884, when I had the first experience ... suddenly I FELT something come into the room and stay close to my bed. It remained only a minute or two. I did not recognize it by any ordinary sense and yet there was a horribly unpleasant "sensation" connected with it. It stirred something more at the roots of my being than any ordinary perception.... Something was present with me, and I knew its presence far more surely than I have ever known the presence of any fleshly living creature. I was as conscious of its departure as of its coming: an almost instantaneously swift going through the door, and the "horrible sensation" disappeared....

[On a subsequent occasion], there was not a mere consciousness of something there, but fused in the central happiness of it, a startling awareness of some ineffable good. Not vague either, not like the emotional effect of some poem, or scene, or blossom, or music, but the sure knowledge of the close presence of a sort of mighty person.

"Of course," added James, "such an experience as this does not connect itself with the religious sphere ... [and] my friend ... does not interpret these latter experiences theistically, as signifying the presence of G.o.d."

But one can readily see why others, perhaps of a different disposition, might interpret the "sure knowledge of the close presence of a sort of mighty person" and "a startling awareness of some ineffable good" in mystical, if not religious, terms. Other case histories in James's chapter bear this out, leading him to say that "many persons (how many we cannot tell) possess the objects of their belief not in the form of mere conceptions which the intellect accepts as true, but rather in the form of quasi-sensible realities directly apprehended."

Thus the primal, animal sense of "the other," which may have evolved for the detection of threat, can take on a lofty, even transcendent function in human beings, as a biological basis for religious pa.s.sion and conviction, where the "other," the "presence," becomes the person of G.o.d.

1. It is likely that there was popular or folk knowledge of the phenomenon long before there were any medical descriptions.

Twenty years before Weir Mitch.e.l.l named phantom limbs, Herman Melville included a fascinating scene in Moby-d.i.c.k, where the ship's carpenter is measuring Captain Ahab for a whalebone leg. Ahab addresses the carpenter: