[up] [home]

A Theory of Vigilance and Dreaming

Montague Ullman

Department of Psychiatry, Maimonides Medical Centre, Brooklyn, New York, U.S.A.

The Oculomotor System and Brain Functions, Proceedings of the International Colloquium held at Smolenice 19 - 22 October, 1970. Edited by Vladislav Zikmund. Butterworths London, Publishing House of the Slovak Academy of Sciences, Bratislava 1973

Theories abound when the need arises to do something drastic about an intolerable and highly resistive state of ignorance. In the case of dreams, before the advent of the REM era one monumental theory overshadowed all others for at least five decades although dissenting views never quite disappeared. With the new influx of factual data about sleep and dreams came a reawakening of interest in clinical theory and a general openness to experimentation, testing and discovery.

One difficulty is that it is hardly likely that theorizing at any one level can do justice to the complexity of the problem, yet very few theory builders seem humble enough to admit this. Theory building by multidisciplinary consensus appears to be what is needed and what unfortunately is lacking. In the absence of this there is apt to be an excess of dysfunctional partisanship. If could set up such a multidisciplinary arrangement for theory building, assuming, and I think rightly so, that much of the necessary information is already at hand, if would include the following experts along with the questions I would address to them:

(1) The neurophysiologist: What are the underlying brain mechanisms for REM activity and dreaming? How do such mechanisms influence the cycle of altered organismic states that occurs with sleep?

(2) The psychologist: How is the unique life experience of the individual reflected in the dream? What are the dynamic considerations pertaining to dreaming that transcend any narrow theoretical orientation and would have to be explained satisfactorily by any dream theory?

(3) The psychiatric clinician: How are disturbances in human relationships reflected in dreams? What, if any, specific dream patterns emerge in connection with the major behavioural disorders?

(4) The clinical sociologist: How do social arrangements and institutions influence dream style? In what sense do dream elements have social points of reference as well as personal ones? Conversely, to what extent can the assignment of social salience to dreams influence the social milieu?

There are a number of others who could offer more specialized help:

(5) The psychopharmacologist: What is the relationship of artificially induced, altered states of consciousness to dreaming? How do drug-induced visions and hallucinatory phenomena fit into the general scheme of things and; more specifically, how do they relate to the natural alternation of thought modalities characteristic of waking and dreaming?

(8) The computer expert: What analogies can be developed to explain two basic life processes, i.e. the conservation of identity and the necessity for growth? How is sameness maintained in the face of the necessity for growth? It may be that only the first part of this question is really suitable for computer experts, but several intriguing theories involving both aspects have already appeared (Gaarder, 1966; Shapiro, 1967; Dewan, 1969).

(7) The ethologist: What is there in the comparative study of animal behaviour that can help clarify question concerning the phylogeny and perhaps the ontogeny of REM?

At the risk of introducing a controversial note I would include one other category:

(8) The parapsychologist: The current tempo of inquiry can no longer permit the insistent reports of parapsychologists to be swept under the rug. The parapsychologist would have to be included as a bona fide member of the team and would be asked to struggle with the question of how paranormal events are related to normal and altered state of consciousness.

With the possible exception of the fourth category all of the above experts are on the scene but working separately. There have been sporadic beginnings with regard to the social dimension in the past and a very occasional glimmer in the present (Bastide, 1966).

It would take a group such as the above or one similar to it to arrive at a unitary theory of dreaming. In its absence any theoretical model or theory, such as the one to be presented, has a one-sided bias built into it.

My own area of competence is in the clinical arena, with some spillage into the last, the parapsychological. My interest in dreams has drawn me to each of the other spheres but with no pretensions other than that of interested observer. A delight in conceptual schemes and a dissatisfaction with classical dream theory led me, just prior to the work on REM sleep, to formulate a theory of dreaming based on the notion of vigilance (Ullman, 1958, 1961, 1966). The theory appeared to be consistent with the emerging discoveries concerning REM sleep (Ullman; 1958): With one major modification[1], I believe it remains compatible with current experimental findings, at least as interpreted by one leading investigator (Hernandez-Péon, 1966, 1967).

The essential feature of the theory is that while asleep there is, under ordinary circumstances and except for occasional motor automatisms such as sleep-walking, only one major behavioural act that occurs, namely, the change in the internal state of the organism leading to awakening. Because we do it every day of our lives the drastic nature of this change from sleeping to waking tends to be overlooked. It is certainly the most marked qualitative change in state that the organism experiences in the course of a day. Compared to it variations in levels of arousal that characterize changing interest patterns while awake are relatively trivial.

It follows that there must be a mechanism available to the organism to bring about such a change in state should a need arise of sufficient urgency to interrupt the otherwise physiologically controlled cyclic variations that occur during sleep. Such a need might be envisaged if something either so painful or so intense were to arise during the night that its continuation was incompatible with the continuation of the sleep cycle.

A second feature of the theory involves the general statement that consciousness, in whatever form it is experienced, is not epiphenomenal but exerts a unifying, integrative, executive function. Furthermore, this function cannot be separated either from the afferent input occurring at the time or the efferent output, be it immediate or delayed. Waking consciousness is organized in relation to a series of selectively attended-to events,. each one having afferent components as well as immediate, delayed or potential efferent components. We have a certain say in relation to the stimuli we attend to and the actions that we take. As a consequence of this, waking consciousness has a distinctly 'voluntary' quality.

When we turn our attention to dreaming the first thing we note is that we are conscious while we are dreaming but that the way that consciousness is experienced, as well as its specific form and content, is qualitatively different from waking consciousness. It is involuntary in quality, sensory in form, and creatively organized in content. All theories of dreaming must account for these characteristics. Despite these rather dramatic differences, and operating within whatever limits they impose, dream consciousness is integrative and just as contingent as waking consciousness upon the capacity to selectively attend to an afferent input as well as to effect a behavioural response. The difference lies not in the presence or absence of  any of these components. It is to be found in the nature of the component parts themselves. In comparison to the waking state the dreamer is selectively attending to a different environment, scanning it in a different manner and responding to it differently. These differences stem from the fact that there is an enforced attention to a self-generated afferent input and that the possibilities for a behavioural output are narrowed to a single act, namely, the act of awakening. This can occur reflexively in response to the overwhelming nature of the dream itself, more gradually as a consequence of the dreamer's struggle to terminate the dream, or simply as the passive endpoint of a dream sequence. Under most circumstances no overt behavioural change occurs and the only change to be noted is the shift to another level of sleep. In any event, consciousness experienced under these circumstances has an 'involuntary' quality. It reflects the fact that the entire occurrence, from beginning to end, is largely outside the dreamer's control. When one examines the form and content of dreaming in relation to this kind of unitary behavioural endpoint the notion of vigilance suggests itself. In fact, one might analyse the sequence of events that occur in dreaming by drawing an analogy to orienting behaviour that occurs in the waking state. A line of reasoning will now be presented in support of the thesis that dream consciousness is an elaborate form of orienting activity designed to attend, to process and respond to certain aspects of residual experience, with an endpoint being reached in either the continuation of the sleeping state or its interruption and consequent transformation to awakening. This is a summary statement of the vigilance hypothesis.

In pursuing the analogy to orienting behaviour we will concern ourselves with (a) the characteristics of the environmental field (b) the characteristics of the scanning apparatus, and (c) the characteristics of the response system.

The environmental field. What is the nature of the field in which novel stimuli arise compelling attention? In the waking state the field is external to the individual and is spatially and temporally organized. In the dreaming state the field is temporally organized and is internal to the individual. While awake stimuli appear as aspects of an external environment and summon the attention of the organism by virtue of certain properties which are in part intrinsic and in part related to the state of the organism as well as its prior experience with it or with similar stimuli. The dreamer, on the other hand, stands (or lies) in relation to a subjective field encompassing his entire past personal history. The attention-compelling stimuli, while they have the same general characteristics of similar stimuli in the waking state (surprise, ambiguity, unfamiliarity) assume a different form. They are internally generated and consist of affective residues connected with events of the recent past. The so-called day's residue, seemingly trivial in itself, assumes importance because of its link to conflict or unaltered life content. Its recency accounts for its early reappearance in the field of consciousness and its disturbing quality for its signal or orienting effect.

The scanning apparatus. What we have designated as the affective residue operates reflexively or automatically as a scanning mechanism. Ranging over the entire longitudinal history of the person it exerts a polarizing influence, drawing to itself and mobilizing aspects of past experiences that are related to it in emotionally meaningful ways. More recently, and in a different context, Dewan (1969) has called  this 'emotional tagging' and has identified it as a device facilitating memory storage and consolidation. Here it is viewed as an energizing or mobilizing effect necessary to help the sleeping organism to fully assess the meaning and implications of the novel or disturbing stimulus and through the participation of a conscious monitoring process either to allow the sleep cycle to remain intact or to engage in an arousal process leading to awakening.

The parameters of the past history brought under scrutiny are therefore determined by their emotional contiguity with the initiating stimulus, as well as by the nature and intensity of that stimulus itself. Any such stimulus, which had little or no orienting properties during the waking state  and in fact which was overlooked at a conscious level, assumes such remarkable properties in the sleeping state for three reasons. By virtue of its recency it reappears on the scene when the necessary level of cortical arousal obtains. It commands attention because it emerges on what presumably was a pre-existing quiet conscious field or one characterized by random, familiar stimuli. It maintains attention because of its resonating and reverberating connections with unsettled, incomplete or unmastered fragments of past experience.

By engaging in the above operations the organism is responding to three important questions that could be expected to arise when cortical activation reaches the level necessary to usher in the peremptory form of consciousness known as dreaming. Once the latter occurs, a quiescent, absent or familiar conscious field is replaced by a novel, unfamiliar, and strange one that has the added uncomfortable feature of being enforced and involuntary. In the face of all of this the organism makes an attempt to regain a sense of mastery by seeking the answer to three questions. What is happening to me? What are the implications of what is happening to me? What can I do about it?

To explain the unique way, in which he comes to grips with these questions we now turn to an examination of the very special response system at work during periods of dreaming.

The response system. When it comes to explaining 'dream-like' features of dreaming most theorists seem content to accept prevailing ideas concerning primary and secondary processes as originally set forth by Freud feeling they are too well established to challenge. Explicitly or implicitly, and with the use of correct-sounding but nevertheless scientifically pejorative terms, dream consciousness is dealt with as release phenomena linked to the instinctual, the primitive and the animal sides of our nature. I submit that this is an enduring myth, the continued operation of which blinds us to the self-serving quiet or open violence of our waking lives. If this facet of our existence becomes more visible to us while dreaming it is not as a consequence of any release phenomenon. Quite the contrary - it comes about as a consequence of the most exquisite honesty we are capable of as we allow ourselves to be confronted with what it is we actually feel at the moment that we feel it, particularly when the consequences may be disturbing or painful and are not known in advance. This is the human dimension to dreaming which has always been easy and convenient to overlook and which, indeed, has always been overlooked except by poets, writers and a handful of dream theorists.

What do I mean by honesty used in this context and how can it be related to the illogical, bizarre presentations we encounter in our dreams? I think the best way to start is to go back to the conceptual scheme first proposed by Pavlov when he described the primary and secondary signalling system. He indicated that over and above the realm of unconditional responses there were in man conditional sensory (primary signalling) and verbal response (secondary signalling) systems. He further indicated that such basic alterations in consciousness as waking and dreaming could be understood in terms of the relationship between these two systems at the time. While awake, the secondary signalling system is relatively dominant over the primary system, with the converse holding true for dreaming. Both systems remain in operation in both states, with the more dominant system determining the quality of consciousness experienced at the time. The verbal system, relating to a selected input and a willed output evokes a voluntary quality to consciousness. The sensory system, when dominant, responds to an unconsciously selected sensory input and results in a largely passively arrived at output; hence, the involuntary quality of our feeling, or sense of consciousness while dreaming.

It is almost impossible to study the dream response system without distilling enduring truths out of the work of Pavlov and Freud. At the same time it is most important to eschew the mischief that can be brought about by blind adherence, rigid discipleship and defensive posturing. With regard to Pavlov I have already alluded to what I consider useful and enduring, namely, the concept of the two signalling systems and the relative dominance of the primary signalling system over the secondary during dreaming and similar altered states of consciousness. The enduring contributions of Freud concerning dreams have to be stripped away from some of their theoretical underpinnings and, in my opinion, can be simply stated:

(1) Dreams can be used to therapeutic advantage.

(2) The basis for this advantage lies in (a) the discovery that a day residue incorporated in a dream touches on conflictual material, (b) the conflict is explored in connection with the development of the individual and (c) in the process of dealing with the conflict healthy and defensive operations can be identified.

(3) Dreams thus incorporate a wider range of information bearing on the current conflict than that is immediately available to the dreamer in the waking state.

You will note nothing about wish fulfillment, instinctual gratification or symbolic disguise, t do not think any of these are necessary. Having said this we can return to our theme of the nature of the response system and examine it with regard to both form and content.

Formal aspects of dream consciousness. At the time dreaming occurs the EEG generally indicates a state of increased cortical excitation, similar to the waking state but even more intense. It has been likened to a state of hyperarousal. At the same time consciousness is experienced in the sensory mode, generally but not exclusively visually. Whereas the state of global excitation might be explained on the basis of a disinhibiting action, the resulting sensory effect cannot be explained by any simple release mechanism. All we can say with any degree of certainty is that in the absence of a supply of external afferents and without access to somatic efferents and under the circumstances of cortical excitation that obtain while dreaming there is a radical shift or transformation in the organizational activities in the brain itself. This change can be characterized in Pavlovian terms as a shift toward the relative dominance of the primary signalling system over the secondary. It must, however, be understood that the primary signalling system is no longer a virginal system operating exclusively at a sensory level. It has undergone modification by a lifelong exposure to a human milieu in which verbal conditioning has occupied a prominent place. Primary signalling system dominance at the human level does not imply that sensory experiences are released from higher nervous control but rather that life experience, including the most subtle abstract elaboration of it are experienced in a sensory mode. Far from being understandable in primitive, atavistic, or release terms, it represents an enormously interesting application of creative energies dedicated to the task of designing visual metaphors that express both what we feel and think at a given moment. As dreaming proceeds there is a succession of visual images (metaphors in motion) which reflect both the experiential panorama unfolding under the polarizing influence of the affective residue and the coping mechanisms brought into play, both healthy and defensive.

The adaptive significance of this can be further understood as a system of internal monitoring of the affective intensity of the particular sample of experience under scrutiny. In line with the vigilance hypothesis, awakening would occur if the affective level cannot be contained and rises above a critical threshold. The neurophysiological substrate for this kind of interaction would logically involve the connections between the cortex and the reticular activating system. The latter, responsive to external sensory input while awake, modulates variations in level of arousal. While asleep it is likewise sensitive to sensory input but his time endogenously generated. Instead of modulating levels of arousal, since all dreams seem to maintain our unflagging interest, it stands ready to influence the arousal process itself should awakening be the adaptive measure of choice in response to the dream presentation.

The content of dreams. I have consistently stressed the all or none possible behavioural endpoint and the radical nature of the change of  state that may ensue in the course of dreaming. The reasons for this emphasis will become apparent when we next consider the psychological requirements of a vigilance hypothesis.

These requirements are essentially twofold. The nature of the impinging threat (novel or disturbing stimulus) must be identified for what it is, rather than for what one may wish it to be, that is, it must be identified and portrayed with a rigorous honesty. Since feelings are our most accurate guide to the way the world impinges upon us, it is the felt residue of a waking event which is transformed into a sensory event in the dream.

The initial presentations are then developed through their linkage to related past experience into a series of presentations which also reflect the available range of coping mechanisms, including, as a last resort, the mechanism of awakening. Why this thorough-going, rigorously honest exploration of a current issue? The answer is that the full implications of a presenting issue have to be explored in depth in order to assess whether or not the sleep cycle can continue uninterruptedly or whether they are, in fact, serious enough to warrant a radical change of state.


In developing a vigilance hypothesis I have drawn an analogy to waking exploratory activity occurring in response to novel stimuli. I have indicated that dream consciousness is more than the passive consequence of failure to maintain secondary process thinking. It represents the active assumption of an alternate mode of consciousness suitable to the maintenance of vigilance operations in the deeper stages of sleep. The immediate transformation of a disturbing felt residue of recent experience into a visual metaphor represents the dreamer's effort to identify and assess an attention-compelling stimulus. The metaphorical quality highlights and dramatizes the felt aspects of the intruding stimulus. The sensory aspects highlight the potential arousal effect this stimulus may have by virtue of its sensory impact upon the reticular activating system.

The further evolution of the metaphorical system occurs dialectically as the initial metaphor leads, by the process of emotional contiguity, to linkage to past relevant experience. By virtue of the continuing impingement of these sensory presentations upon the activating system there is a continuing assessment of safety or threat in connection with the developing metaphors. Safety is compatible with the continuing biological regulation of the sleep-dream cycle. A sense of threat, arising when the effect mobilized is too intense, invokes the only major behavioural response possible for the dreamer, namely, awakening.

Vigilance theory and the current experimental scene. The theory offered above has been clinically derived. Therein lie bath its strengths and its limitations. Any dream theory has to be compatible with clinical data. In my opinion vigilance theory succeeds in doing this with fewer metapsychological assumptions than classical dream theory. Its limitations arise from the fact that in being based on spontaneously recalled dreams it may not go far enough in explaining all dreams or explicitly deal with other possible tasks subserved by dream consciousness. This comes most clearly into focus when vigilance theory is compared with dream models based on computer analogies (Gaarder,1966; Shapiro,1967; Dewan,1969). Here the emphasis shifts to the data processing aspects and the assessment of recent experience in relation to the problem of memory storage. Dream consciousness is seen as managing the spillage from the day's excess input with the dreamer involved in a kind of play back mechanism. Data are examined in their fit with past experience and anticipated future needs as they are processed for long term retention.

 The chief difficulty with the computer model as I see it is that it regards dream consciousness from the point of view of the waking state. Insufficient account is taken of the executive function of consciousness in meeting the needs of the sleeping organism at the time dreaming is occurring.

There is a close fit between vigilance theory and the neurophysiological model of sleep and dreams developed by Hernandez-Péon (1966, 1967). The main tenets of his model are that slow sleep and rapid sleep unfold in a continuum as a unitary process, with rapid sleep, at least in cats, representing the deeper stage. He then postulates hypnogenic and vigilance systems in reciprocal antagonistic relationship. As sleep develops the vigilance system is, more and more, inhibited by the hypnogenic system. While awake the vigilance system maintains selective attention through its cortical inhibitory effect exerted by reticulofugal inhibition. In the deeper stages of sleep, with maximal inhibition of the vigilance system, there results cortical and limbic disinhibition which is subjectively experienced as dreaming. This release brings into play recent, remote memory and motivational systems, all of which enter into dream consciousness. While Hernandez-Péon does not link dream consciousness specifically to vigilance operations, the neurophysiological substrate is compatible with the theory as it has been outlined. Although vigilance theory is essentially derived from, and related to psychological facts, it articulates with neurophysiological theory around three important issues: (a) the existence of functional interrelationships between the cortex and the reticular activating system governing arousal levels and capable of inducing arousal, (b) the existence of a unique state of cortical excitability making possible a more global orientation to present experience, and (c) the theory is predicated on the physiological control and monitoring of the REM state and its vicissitudes but leans upon the possibility that, once the cortex is sufficiently activated for dreaming to occur, psychological factors can play a role in determining, the final behavioural outcome.


BASTIDE, R. (1969); The sociology of the dream. In: G. E. Grunebaum, R. Callois (Eds.), The Dream and Human Societies. Univ. of California Press, Los Angeles, pp. 199-211

DEWAN; E. M. (1969): The programming (P) Hypothesis for REMs. Air Force, Cambridge Research Laboratories, Bedford, Mass.

GAARDER, K. (1969): A conceptual model of sleep. Archs gen. Psychiat. 14, 253-260.

HERNANDEZ-PÉON. R. (1966): Attention, sleep; motivation and behavior. In: P. D. Bakan (Ed.), Attention. Van Nostrand Co., New York, pp. 181-204.

HERNANDEZ-PÉON, R. (1967): Neurophysiology, phylogeny, and functional significance of dreaming. Expl Neurol. 4, 106-126.

SHAPIRO, A. (1967): Dreaming and physiology of sleep. Expl Neurol. 4, 56-81.

ULLMAN, M. (1958): The dream process. Am. J. Psychother. 12, 671-690.

ULLMAN, M. (1961): Dreaming, altered states of consciousness and the problem of vigilance. J. nerv. ment. Dis. 133, 529-535.

ULLMAN, M. (1966): A nocturnal approach to Psi. Proceedings of the Parapsychological Association 3, 1966. Presidential address, IXth Annual Convention of the Parapsychological Association, New York City, Sept. 9, 1966.


Evans: In 1964 Newman and I first put forward the idea of the dream as being a process comparable to the rewriting, up-dating and clearing of old program ‑ operations that must be performed with modern computers. The brain is obviously a computer of a special kind and as with other computers must be controlled by sets of programs. We feel that it is exceedingly likely that the brain would need to set time aside with the system 'off-line' when existing programs are updated and rewritten in the light of information fed in the course of the day. As the system is off-line (a condition which human beings have come to call sleep) these programming operations are carried out at an unconscious level. Interruption of the sleep state will cause conscious mind to 'catch the system at work' and a 'dream' will be reported. Incidentally, Dr. Ullman while I am very dubious about telepathy and kindred phenomena, I don't see in principle that these ideas of ours conflict in any way with your own notions about sleep.

Ullman: I did not mean to imply that computer analogies were incompatible with the point of view expressed, but that they were limited in that they seemed to be exclusively concerned with either a storage function or the processing of information for use in the subsequent waking state. The emphasis then, is on a point of view about dreaming taken from the waking state. I believe the emphasis should be a more phenomenological one which would shift the focus to a concern with the needs of the sleeping human organism. These needs involve decision making during states of consciousness that differ qualitatively from waking consciousness. The processing of information at this time has to be oriented to such immediate needs as the question of relative safety in remaining asleep as well as the longer range needs of the waking state.

[1] Earlier, when dreaming was thought to occur in the lighter stages of sleep, vigilance theory was related to the problems of transitional consciousness.