According to the Shorter Oxford English Dictionary the primary meaning of the word “nightmare” is a “female monster supposed to settle upon people and animals in their sleep producing a feeling of suffocation,” and it only secondarily refers to “a bad dream producing these or similar sensations.” It is, in fact, a word that harks back to the largely lost world of evil spirits, demons, incubi, and succubi, when dreams were not, as they are now, products of our own mental processes but were portents, threats, visions, and visitations from another world.

Something of its former horror and glamour still, however, surrounds the word “nightmare.” Many people still believe that there is some specific nocturnal experience, an encounter with a real nightmare, which they may one day have, or, more soberly, that there is one specific kind of dream, distinct from all others, which alone merits the designation “nightmare.”

Ernest Jones seems to have been one of the last people to believe in nightmares, and in his On the Nightmare (1951), which is still in print, he asserted that there are such things as true nightmares and that they have three cardinal features: agonizing dread, a sense of oppression and weight upon the chest leading to difficulty in breathing, and a conviction of helpless paralysis; features that would indeed be present if one were being overlaid by a female monster. However, Jones, being a psychoanalyst, did not, of course, believe in the real existence of visiting monsters; he asserted that nightmares are “morbid phenomena” with a specific pathology, being “an expression of a violent conflict between a certain unconscious sexual desire and intense fear,” the certain unconscious desire being in both sexes “the feminine or masochistic component of the sexual instinct.” In other words, Jones maintained that there are such things as true nightmares, which closely resemble the visions that medieval man blamed on intrusive monsters, incubi, and devils, and that in men they are the result of repressed homosexual wishes and in women the result of repressed heterosexual ones.

There is, however, something very peculiar about Jones’s thesis. He claims that it is based on his own clinical experience as a psychoanalyst, which would lead one to expect that he would be able to cite twentieth-century examples of true nightmares, but in fact all the examples he gives date from the eighteenth century or even earlier.

True or classical nightmares seem in fact to be something of a myth or at least a rarity, and twentieth-century Europeans and Americans do not, it seems, often have the kind of dream that led their medieval ancestors to believe in intrusive monsters, succubi, and incubi. In an earlier work, Sleep and Dreaming (Little, Brown, 1970), Ernest Hartmann reported that true incubus nightmares had only very rarely been reported by subjects sleeping in dream laboratories, while “very frightening” dreams were “not quite so rare.” In his present work, The Nightmare: The Psychology and Biology of Terrifying Dreams, he in effect undermines the very category that he, or his publisher, has chosen for his title.

He does this in two ways. First, he argues that the classical nightmare or incubus attack is not a dream at all; it is, he says, not remembered, though the awakening from it in terror may be. It lasts too short a time for a dream to occur during it, it occurs during the wrong stage of sleep. It is really a form of night terror, a night terror being a condition best known to pediatricians in which a sleeping child sits or stands up, screams in terror, remains poised between sleep and wakefulness for a few minutes, and then either awakens or falls asleep again, remembering little or nothing of the episode later.

Secondly, Hartmann defines a nightmare as a “long, frightening dream awakening the sleeper out of dream-sleep,” so that a nightmare is no longer, as the ancients and Jones believed, a particular type of dream which might perhaps have a specific meaning, but one end of a spectrum. It is only definable quantitatively according to the intensity of the anxiety experienced during it. If I understand Hartmann rightly, he holds that all dreams from which the dreamer awakens spontaneously in a state of anxiety can appropriately be called nightmares. He also calls such dreams “dream anxiety attacks.” “I hope it is clear,” Hartmann writes,

that the person in the classical descriptions of “nightmares” who sits upright in bed, or gets out of bed, screaming, with a glazed expression on his face, is a person having a night terror, not a nightmare. Nightmares and night terrors are indeed quite different phenomena psychologically, physiologically, and occur in quite different persons, as we shall see.

Hartmann, then, takes the position that classical nightmares are not nightmares at all, or even dreams, and uses the word to describe intense anxiety dreams which awaken the dreamer. His reasons for doing so derive from his interest in sleep and dream laboratory research and from his faith in the reliability of psychological tests and surveys. The evidence derived from such sources does indeed appear to be that the subjects who arrive in dream laboratories with a history of having frequent “nightmares” divide into two groups: there are those who have their “nightmare” early in sleep when their electroencephalograph records nondream sleep (called Stage 4 sleep) and do not remember any dream even though the laboratory workers can observe their terror. And there are those who have their “nightmare” late in sleep when their electroencephalograph records dream sleep (D-sleep) and can report in vivid detail the dream that has awakened them.


This laboratory finding sounds clearcut and straightforward, and one can appreciate how Hartmann arrives at the idea that the former seemingly contentless terrors are not dreams at all, but disorders of arousal. But as Hartmann himself says, “Life is never quite this simple,” and there are, it seems, cases in which it is not clear whether “the patient or subject” has had a night terror or a nightmare (dream anxiety attack).

There are cases in which the type of experience is not entirely clear, even when one asks all the right questions. This can be due to poor memory, or poor descriptive abilities on the part of the patient or subject. Or it can be the fact that the person actually has both nightmares and night terrors concurrently—I have seen six or seven such cases—that makes description of and delineation between them difficult.

These “six or seven such cases” make one think. If it is possible for someone to have concurrently night terrors that are not dreams, and nightmares (dream anxiety attacks) that are dreams, one cannot help wondering whether night terrors and nightmares in Hartmann’s sense can really be such different types of sleeping experience as his definitions of them suggest, a doubt which is increased by his reporting that victims of night terrors do sometimes recall a single frightening image: “Something is sitting on me,” “I am choking,” “Something is closing in on me.” These images all sound suspiciously like the tail ends of classical nightmares that were too terrifying to be remembered in detail. On the only other occasion when I have myself written on nightmares and anxious dreams, I came to the conclusion that in nightmares the dreamer is imagining some situation that threatens his identity, his very sense of himself as agent, while in anxious dreams he is being confronted by some aspect of himself that he has as yet failed to integrate into his personality.

These “six or seven such cases” also raise doubts about the validity of Hartmann’s hypothesis that sufferers from nightmares (dream anxiety attacks) and night terrors are quite different kinds of people. This hypothesis, which derives from interviews and psychological tests performed on some 120 people, asserts that the sufferers from nightmares (dream anxiety attacks) “had some features of schizophrenia or a vulnerability to schizophrenia; they had artistic tendencies and a kind of openness and sensitivity.” The sufferers from night terrors

included all kinds of people with no specific psychopathology and no particular artistic tendencies; their psychological tests did not show the unusual characteristics found in the tests of the nightmare sufferers. Many were psychologically quite average; some had psychopathology, but not of any single type; a subgroup seemed unusually tightly controlled, but the relationship of this characteristic to night terrors (were they “holding in” or “holding down” angry impulses excessively?) was not clear.

The trouble with distinctions of this kind is that they are too vague, and too dependent on psychological tests performed on rather few people. The truth of the matter appears to be that fifty lifelong sufferers from dream anxiety attacks, who still as adults had more than three nightmares a week, scored significantly higher on the paranoia, psychasthenia, and schizophrenia scales of the Minnesota Multiphasic Personality Inventory than did twenty-four vivid and ordinary dreamers who were used as controls, and that most of the fifty who suffered anxiety attacks impressed the psychiatrist who interviewed them as being open, gifted, sensitive people. One would have to have greater faith in psychological tests and psychiatric interviews than I do to take seriously formulations based on such shaky foundations.

These fifty nightmare sufferers, who must, it seems, have all been Bostonians, were recruited by newspaper advertisements asking for persons who had frequent nightmares to volunteer to take part in a medical psychological study. It would have been illuminating had we been told something of the social and economic forces that led them to converge on the dream laboratory. They all claimed to remember having had nightmares as children, and to still be having “totally real” nightmares more than three times a week. Three-quarters of them claimed to have been very sensitive children, two-thirds of them claimed to have “precognitive dreams, out-of-the-body experiences, or other such experiences,” and to have had traumatic experiences (muggings, rapes, violent deaths of persons close to them) after they had started having nightmares. More than two-thirds of them had had either long or short-term psychotherapy before offering themselves as subjects at the dream laboratory. Almost all of them had experimented with alcohol and drugs, but were not drug users at the time of the study.


They were all between the ages of twenty and thirty-five, and

one striking feature was that none of those who were employed held ordinary blue- or white-collar jobs. A number were musicians (both composers and performers), painters, poets, writers, craftspersons, teachers, and therapists…. Many saw themselves as rebels or “outsiders” to a greater or lesser extent. Some of them actively rejected society—materialism, mediocrity, bourgeois values, and so on—while others saw themselves more as rejected by society. Most, as we shall see, felt from childhood on that they were a little different from other people.

Dr. Hartmann seems in fact to have collected a group of people who must, I would surmise, be far from representative of Americans as a whole, and much of what he says about them, about their vulnerability to schizophrenia, their openness, their defenselessness, their sensitivity, and their thin and permeable ego boundaries must, it seems to me, be an account of the psychology of a particular dropout cultural group and not an account of the psychology of the typical American sufferer from dream anxiety attacks. His thesis would have been more convincing if his sample of subjects had included even one factory worker, lumberjack, or longshoreman.

I find it impossible to believe that susceptibility to nightmares in Hartmann’s sense of the word is confined to the single social group from which he drew his sample. Clinical experience convinces me that patients in all diagnostic categories may have dreams that would satisfy Hartmann’s criteria for a nightmare, although rarely if ever as frequently as his Boston group claimed to have them. I must also confess to being skeptical whether his subjects really did have as many terrifying dreams as they claimed: a tendency to self-dramatization could well have been one of the group’s attributes.

The same criticism cannot be leveled at Dr. Hartmann’s discussion of traumatic nightmares—i.e., of long, terrifying dreams which reproduce exactly or almost exactly some past event that has actually happened to the dreamer—which is largely based on studies of Vietnam veterans. Although, again, the number of subjects studied was small, the picture that emerges seems clear enough. Those soldiers who already suffered from nightmares before they went to Vietnam suffered more of them after they returned, and included frightening wartime scenes in them even though they had never been in action. Those soldiers who had been in “reasonably good mental health” before their military service but developed traumatic nightmares after being in action had still been in their teens when they had the combat experience that they reproduced in their nightmare; and all but one of them had lost a close buddy in action. Those soldiers who had been in combat but had not developed traumatic nightmares had been in their twenties when they were in combat; they had, they said, “quickly learned not to get too close to anyone in Vietnam.” In his discussion of the dreams of war veterans something of the reality of human suffering and tragedy breaks through the jargon and the statistics.

Such glimpses of the painful realities that must, one presumes, be part of the experience and life history of those who frequently have terrifying dreams occur only rarely in Hartmann’s book. Although the jacket describes him as a “famed analyst,” Hartmann’s approach is basically experimental and statistical. As a result he has little to say about the meaning, symbolism, and interpretation of dreams, though he is clearly impressed by the extent to which terrifying dreams reproduce infantile feelings of helplessness. He also thinks that people who have frequent terrifying dreams do not use such relatively mature defense mechanisms as repression and isolation, as hysterical and obsessional patients do, but instead either “accept their idimpulses” or use “primitive coping mechanisms, such as projection.” People subject to nightmares have, he says, “characteristics opposite to those of the obsessional character.” They lack what Wilhelm Reich called character armor.

However, in stressing that his nightmare subjects have thin ego boundaries, he is not, I think, primarily concerned to report an observation that could be of clinical importance, but rather to introduce a concept that could, he believes, be used to bridge the conceptual gap between the psychology of mind and the physiology of the brain. In Chapter 10 he discusses

in detail the biology and especially the chemistry of the nightmare. If nightmares are an indication of thin boundaries, and an indication of a certain kind of vulnerability, as we have suggested, perhaps the chemistry of the nightmare also forms part of the chemistry of thin boundaries in the mind and the chemistry of these vulnerabilities. I propose that there is a basic structural concept of thin or thick boundaries underlying our many uses of the word and that this structure must have a brain biology…. For instance, it is not impossible that thickness of boundary could refer to “insulation” or resistance to spread of excitation from one part of the forebrain to another, though the situation will probably turn out not to be quite so simple.

In this passage, which is hardly notable for its clarity, and in Chapter 10 itself, Dr. Hartmann is, I think, suggesting, first, that differences in sensitivity, vulnerability, and the capacity to construct ego defenses may have constitutional roots, and, secondly, that these constitutional roots may be reflected in the chemistry of the brain. This idea remains, however, purely speculative, as Dr. Hartmann himself admits:

Persons who have frequent nightmares would be persons who have a tendency to greater or more rapid activation of these dopamine systems or who are more sensitive to such activation. It is possible that this same characteristic if present from birth might make these people more vulnerable to psychosis.

In other words, the biology producing a nightmare may involve an altered balance of the four neurotransmitters discussed.

This passage contains too many qualifying woulds, mights, possibles, and mays to carry much conviction; the fact appears to be that certain drugs concerned with neurotransmission do indeed either increase or decrease the frequency with which people dream or have nightmares.

Similarly, Hartmann’s insistence that night terrors are not dreams while nightmares are does not arise from the observation that they have a different psychological meaning, but from the fact that the electroencephalographs that accompany them display different patterns. This must indeed signify something, but Hartmann himself quotes a sufficient number of anomalous findings—e.g., dreams occurring in the “wrong” stage of sleep—to make one doubt whether the difference between them can be as absolute as he would, it seems, like to believe.

The jacket of The Nightmare claims that it is “rich in scientific data yet written in layman’s language,” but I think that laymen will in fact find it heavy going. The details of the various psychological tests performed on the various groups of subjects are most confusingly presented, while the last chapter, “The Biology of the Nightmare,” presupposes a familiarity with brain anatomy and pharmacology that I suspect few laymen possess.

This Issue

March 28, 1985