The preparation of the chapters contained in this book was supported in part by the United States Air Force under Contract 18(600) 1797 monitored by the Rome Air Development Center of the Air Research and Development Command and under Contract AF 49 (638)-187 monitored by the Air Force Office of Scientific Research of the Air Research and Development Command.
Introduction — manipulations of human behavior
In recent years, concern has been expressed, in both scholarly and popular literature, about the dangers of scientific developments that could be used to control and manipulate human behavior. The fear is frequently voiced that techniques have been developed to an extent which threatens fundamental values of Western civilization.
Anxious alarms and dramatic speculations have overshadowed reports of sober efforts to determine which dangers are real and which imagined.
This book represents a critical examination of some of the conjectures about the application of scientific knowledge to the manipulation of human behavior. The problem is explored within a particular frame of reference: the interrogation of an unwilling subject. A number of scientific areas have figured prominently in speculations regardirig the application of science to the manipulation of behavior in interrogation (69). For this work, scientists who had done research in each of these areas were asked to review the state of relevant knowledge in their fields, to consider whether and how it might be applied by interrogators, and to evaluate the recourse available to highly motivated persons for resisting the attempted influence. Their reports constitute the body of this book.
Attention has been focused on interrogation because of the central position this topic has had in recent public discussions of prisoner-of-war behavior — issues that made scientific methods of manipulating behavior a major public concern. Much of the work in this book was sponsored by the U. S. Air Force because of their interest in the problems which face the prisoner of war. Such aspects of prisoner exploitation as ideological conversion and the elicitation of false con-
*E. L. Freud, Ausgewaehlte Briefe, Berlin: S. Fischer, 1959.
fessions have received relatively more public and academic discussion than the attempts to elicit factual information through interrogation. Nonetheless, the editors believe that there are some major advantages to approaching the broader topic of the manipulation of human behavior by limiting attention initially to the latter type of situation. The background of recent concern with these problems may illuminate some of the considerations leading to the particular emphasis of this work.
The notoriety that Communist exploitation of United Nations prisoners of war has received in the United States gave impetus to professional and lay concern with problems of the manipulation of behavior. Various writers have associated the compliance effected by Communist captors with phenomena observed in the laboratory; e.g., effects reported following experimental work in pharmacology, hypnosis, sleep deprivation, sensory deprivation, semi-starvation, electrical stimulation of the brain, as well as in social-psychological investigations of persuasion and group conformity pressures (3, 10, 13, 14, 16, 20, 23, 24, 28, 32, 39, 67).
The most radical expressions of concern have alleged that techniques for manipulating behavior are now capable, or are at least on the threshold of being capable, of eliminating the determination of the subject as a barrier to successful influence (21, 32, 33, 34, 35, 36, 37, 38, 39, 44, 52, 60). These claims have been challenged: other investigators have been impressed by the strength, stability, and resilience of long-established values and social controls, and by rational regard for self-interest, either as limiting the compliance of those subjected to coercive and persuasive influence attempts, or as enabling the subject to resist completely (2, 4, 7, 11, 15, 18, 50, 53, 54, 55, 65, 66, 68).
Certain other commentators have viewed the successful exploitation of captives in Korea and similar incidents as less indicative of increasing perfection in the arts of influence and coercion. They have attributed the successes of the captor to individual defects of background and stamina in the prisoners who collaborated, or to a general deterioration of the vitality of social values and controls in contemporary society (16, 25, 31, 61, 62, 63).
prisoner-of-war policy. Concepts such as determinism vs. moral responsibility are among the philosophical and value considerations implicit in this debate which have added considerably to the problems already posed by existing biases. There has been dispute regarding the extent to which individual repatriated prisoners of war are legally or morally responsible for deviations from ideal standards of military conduct, especially in the matter of yielding information or making "confessions" to the captor (9, 25, 31, 39, 41, 47, 49, 54, 64).
A central question of fact has stood out in this controversy, quite apart from the moral and philosophical issues raised: whether all individuals, regardless of how strongly motivated to resist, could be made to comply with demands for information, "confessions," or other collaboration by methods employed by the Communists, provided that the intensity, duration, and quality of the pressures were sufficiently great. Popularly, this proposition was phrased: "Every man has his breaking point." It gained acceptance as a premise in policy formulation, although the question remained regarding which, if any, of the situations encountered by United States prisoners of war in Korea approached those extremes of "stress" that no individual could be expected to lesist (40, 48, 64, 68).
Critics of the policy evolved after the Korean War — "The Code of Conduct" — argued that the failure of the policy makers to consider imminent developments in scientific methods of human manipulation has been a more serious error than a lack of understanding of the practices actually encountered by prisoners in Korea. One scientist (39, page 54) has written: ". . . one thing is certain — our national policy concerning the conduct of our prisoners of war has not yet fully faced up to psychological developments which are appearing over the horizon of the future." His own conclusion was that armed forces could protect information from an enemy only by denying it to persons subject to capture or, where this was not feasible, by giving such persons means for destroying either themselves or their memories.
In the study leading to the present book, the U. S. Air Force through their sponsorship has sought an authoritative examination of publicized speculations regarding the possible use of scientific developments in the manipulation of behavior against future prisoners of war. A specific objective has been to examine an enemy's possible use of scientific techniques to elicit guarded information from captives, which is the form of prisoner exploitation having the most immediate and and direct military significance.
very little information on the topic appears in open-source literature. The dearth of sober information on interrogation has had the unfortunate consequence of facilitating the exploitation of United States prisoners of war by Communist captors (64).
Our purpose here has been to bring together in one book authoritative information on methods of behavioral control that have been the subject of considerable speculation in discussions of interrogation. Scientists representing a variety of fields have examined a number of hypothetical means that might occur to an interrogator for eliciting information against the will of his subject. Their attention has been more on what could be done than on what actually may have been done.
Many scholars have observed that science replaces magic and witchcraft as societies secularize. The problems of living in the present age remain much as they have always been, however. They generate many of the same wishes and terrors. The aspirations and anxieties that not so long ago were projected onto conceptions of the wizard and witch are now directed to the scientist.
Two of the most basic of life's problems are linked to the individual's power position vis-à-vis his fellow men: the inability to make others fulfill one's wishes; and the reverse, the fear of being controlled by others, with the consequent loss of the autonomy that is believed to be fundamental to the conception of the self. These opposites are incongruously exaggerated in paranoid thinking, one of the most prevalent mental symptoms of Western man. They doubtless exist in the fantasy of most persons, to extents that differ from paranoia in intensity and pervasiveness.
The profound fascination of the topic under consideration may stem from the primitive, unconscious, and extreme responses to these problems, which gain expression in myth, dreams, drama, and literature. On the one hand, there is the dream-wish for omnipotence; on the other, the wish and fear of the loss of self through its capture by another. The current interest in problems of manipulation of behavior involves basic ambivalences over omnipotence and dependency, which, if projected, find a ready target in the "omniscient" scientist (30).
The "mad scientist" of the horror movie and novel has been with us for many years. More recent fiction has enlisted him in the service of the great nation-state. With the perfection of mass-destruction weapons and the elaboration of totalitarian efforts to control human behavior, the myth has begun to converge with aspects of reality.
Conjectures concerning the prospects of "total annihilation of the human will" appear almost as frequently as those regarding the threat of mankind's total destruction by thermonuclear or similar weapons. Regarding weapons of physical destruction, responsible scientific evidence is offered along with uninformed and ill-informed surmises, both in support of forecasts of doom and in rebuttal. In the case of the threats science poses to human autonomy, however, sensationally speculative expressions, like those of the Brave New World that Aldous Huxley (21) recently revisited, have enjoyed a near monopoly. In professional journals and publications, as well as in statements for popular consumption, scientists have sometimes contributed to uncritical thinking regarding the potential application of scientific developments to the control of human behavior. Some scientists have done so in their zeal to make the public aware of the dangerous tool which the techniques for manipulating behavior could become in the hands of totalitarian and other irresponsible practitioners. A common error has been to assume that some scientific development, or some explicit scientific theory, was being applied by Commanist "brainwashers" and other manipulators (2, 4, 5, 18). Other scientists, and interpreters of science, have also contributed to the identification of the behavioral scientist as a powerful and unscrupulous practitioner of the arts of influence, as in Vance Packard's Hidden Persuaders (45).
It has been pointed out that the ways in which the popular communication media define the problem are akin to those of prescientific times (56). Raymond Bauer (2) has noted the resemblance of the concept "brainwashing" to demonology: the idea of the "brainwashed" does not differ greatly from that of the "possessed" (6). "The battle for the mind" (52), sometimes fought against "the hidden persuader," has many elements of the occult.
Viewing the problem in magical or diabolical terms is not an altogetlier irrational analogy, given the existence of those who simultaneously practice and seek perfection of the means for controlling behavior and conceive of their efforts as directed toward "possessing the will" of their victims. As one of us has pointed out in discussing
Communist interrogation and indoctrination practices (6), the Western conception of "brainwashing," in terms akin to "possession," is matched by the Communists' view of what they seek and how to achieve it as exorcism. Thus, the Chinese Communist leaders not only find nothing to resent in charges that they "brainwash" their opponents (cf. 1), but regard the term as a quite apt and honorable description of what they wish to achieve. To take another example, the symbolic imagery of the medieval Hexenhammer (22) is reflected in the designation of the Soviet World War II inquisitorial apparatus as SMERSH ("Death to Spies") (57).
Scientific sobriety demands that the dry examination of experimental evidence replace the lively books on exorcism of ancient times. As in many contemporary works which are closer to the livelier lore of ancient times, however, the present review deals with human concerns antedating science that are at the root of the central question: "Can man really be made to behave contrary to his profoundest beliefs and his conscious self-interest?"
Symbols of science can be used in a magical way, as much of the "brainwashing" literature illustrates. Various writers have invested the techniques of interrogators with the magic of science by attaching technical labels to what actually have been traditional and pragmatic practices (2). In assuming the attitude of the "hard-headed" scientist toward the problem, there is a danger in falling into an equivalent misuse of science. This would be the case were one, in effect, to attempt to counter those who present a diabolical image of the "brainwasher" by invoking superior scientific deities to frighten this specter away. Thus, magical thinking and projections, as has been indicated, pervade prevalent judgments regarding the significance of the behavioral alterations that interrogators can effect. By substituting impassive scientific names for ordinary language with its intense connotations for human values, the impression may be given of eliminating not only these extravagant judgments but also almost all the human significance of these effects. In this way, for example, "treachery" can become mere "attitude change" or "a shift in the subject's frame of reference."
directed — who is designated following common intelligence usage as "the source" — to be highly motivated to safeguard the information; and that, at least initially, the source regards denying information to his interrogator as "more important than life itself." It apparently has not been a rarity for individuals undergoing interrogation to say "go ahead and shoot" in the face of a convincing threat of death, and yet to reveal the information thus guarded under seemingly mild pressure later (8). Similarly, Western jurisprudence recognizes that lengthy interrogation, even without physical coercion, can produce "unwilling" confessions, true or false, of capital crimes. Divergent interpretations have been placed on reported cases of individuals who have resisted very intensive interrogations without divulging information. Some use it to demonstrate the existence of an unconquerable, inextinguishable human will. Others regard the instances of successful resistance to interrogation as mere illustrations of remediable deficiencies in interrogation technique.
Neither this nor any other scientific volume, in the opinion of the editors, can resolve the differences implicit in these two orientations, or yet other interpretations. On the basis of scientific tests alone, they are difficult to resolve even with a completely deterministic set of assumptions. As the approach of this review illustrates, for any given set of motivations of the source, however powerful, one can at least speculate about possible manipulations to overcome them. On the other hand, it is possible to speculate about methods of heightening motivations and defenses against any conceivable manipulative assault. Exclusively scientific tests probably cannot foreclose either possibility at this time.
Another important qualification to conventional ideas about the ultimate limits of the control of human behavior will become apparent in some of the discussions that follow. The purposes that men have in seeking to control, or to influence, the behavior of others involve the distinctively human capabilities of men and their significance for one another. The major fallacy of the totalitarian interrogator grows out of a poor appreciation of this fact. Some of the chapters here indicate that there are limits to which the ability of a source to reveal information can be separated from his willingness to do so. The analytic divisions we make between such aspects of mental activity as the recall and transmittal of information on the one hand and motivations on the other do not correspond to behaviors that are totally independent of one another in the organism. Furthermore, much of the use one person can make of another resides in the latter's ability to function in a voluntary fashion: in his having
initiative, making choices, preferring, and rejecting. The fallacy of belief in the possibility of total control for any purpose stands out as bizarre in the extreme when acted on by those whose purposes involve the control of self-initiated behavior. An example, simultaneously tragic and ridiculous, is the ideological interrogation.
A system in which mental conformity is sought through coercion and manipulation embodies an ever-present fear on the part of the controllers that conformity will be based on opportunism rather than conviction. In oppressive ideological systems, such as modern Communism, which demand "true sincerity" from their subjects rather than mere outward conformity, the inquisitorial process appears to be a natural development. It is a difficult matter to determine whether thoughts are indeed "true thoughts." The inquisitorial process, being itself highly coercive, reinforces the original suspicion regarding opportunistic conformity. In a vicious circle, coercion is used to produce conformity, generating fears that the conformity produced is insincere, generating in turn further coercion to make it "sincere." The abhorrence of these practices by those subjected to them makes the fears of the controllers well founded and further reinforces the vicious circle.
Under these circumstances, the ultimate test of the loyalty and sincere devotion of the individual to the system is his acceptance of the inquisitorial process itself: the purge, coercion, confession, and the entire paraphernalia of enforced conversion. Malleus Maleficarum (22, page 212) provides an illustration of the manner in which the victim is compelled to adopt the frame of reference of the inquisitor:
At the outset of the study the impression of the editors from their prior investigation of interrogation problems was that the effectiveness of scientific innovations for controlling human behavior may have been exaggerated in most public discussions. Before final evaluations, alarming or otherwise, were ventured from a human standpoint regarding the significance of the control over behavior
which scientific developments will make possible, it was felt that a more sober, systematic, and accurate examination was required from a purely objective perspective. This evaluation relates to the applicability of such developments to a specified type of objective and to the nature and limits of the alterations of objective behavior that these developments will permit a would-be manipulator to induce in a resistant person. Although such an effort cannot settle the philosophical and emotional questions raised about the significance of the control which can be exerted, it can indicate that some are ill-founded and others premature. Yet others relate to very real matters.
This work does not represent an attempt to minimize the problem. The conclusions reached do in fact show that many developments can compound tremendously the already almost insuperable difficulties confronting the individual who seeks to resist an interrogator unrestrained by moral or legal scruples. On the other hand, it can be shown that many of the measures popularly supposed to render an interrogator omnipotent actually have no demonstrable applicability to his purposes. Other measures that appear to have high potential utility for the control or influence of behavior seem to owe their effectiveness to quite different kinds of processes than popularly supposed. Among the latter are "placebo" measures, the success of which depends largely upon the attribution to them of a nonexistent potency by the subject, and at times the manipulator. Knowledge, it appears, is a "two-edged sword" in interrogation.
The latter fact is a source of some comfort. Several scientists have reported on the possible applications of scientific knowledge that might be made by the most callous interrogator or power. The results of their thinking are available here for anyone to use, including the unscrupulous. The alternative is to confer on the would-be manipulator a monopoly of knowledge by default. His success, as the various chapters of this book illustrate, depends heavily on the ignorance of his victims. Skinner (58) has argued that those who are most concerned with restricting the vulnerability of men to control by others have the most to gain from a clear understanding of the techniques employed (pages 320-322). (See also Skinner, 59.)
The question of controlling fundamental attitudes and values may hold greater interest for many than our attention to the eliciting of guarded facts by interrogators. Much concern of recent years regard-
ing behavior control, as has been discussed, has centered on connotations that have come to be conveyed by the term "brainwashing." The source of this concern is the belief that individuals can be "changed" in some fundamental way by devious and mysterious acts of influence. Certainly, Communist practitioners of "thought-reform" visualize the creation of a "new man" as their objective. People of Western nations, frightened and puzzled by these Communist practices, have also felt that the behavior displayed by many victims of such efforts could be explained only in terms of some very basic changes within the individual.
The difficulties confronting attempts to examine such complex issues scientifically argue in favor of dealing first with simpler and more objective forms of behavioral influence. In the "brainwashing" model, we have a basically nonrational attempt to effect nonrational changes of subjective states. Practitioners of "thought-reform" seek "real" changes in beliefs and values. They demand that the victim be "honest, sincere, and full" in his "self-examination, repentance, and change" (27). It is difficult to find objective indicators of the extent to which a "thought-reformer" has achieved "honesty and sincerity," and particularly difficult when given the special ideological meanings such terms have for the practitioners of "thought-reform." As for producing "real" and "fundamental" changes in the person, the superficial and stereotypical concepts about human personality on which Communist " thought-reform" efforts are predicated might lead one to expect that any fundamental changes they produce must be accidental rather than a realization of a deliberate objective.
There is no question that it is possible for men to alter, impair, or even to destroy the effective psychological functioning of others over whom they exercise power. The concepts influence, control, and manipulation denote a certain kind of alteration: the consummation of a purpose of the influencer in the behavior of the influenced. If we wish to examine scientifically questions denoted by the terms influence, control, or manipulation, we must be able to observe objectively and to define in precise terms both the effects sought and those obtained.
A focus on the elicitation of guarded factual information simplifies the analytical problem considerably by posing a model that involves such objectively specifiable purposes and effects. To achieve further simplicity for purposes of this review, the contributors were asked to consider as their primary model interrogations where the interrogator's objectives consisted of obtaining simple, objective information regarding the physical world.
Interrogators in this age of "psychological warfare" increasingly seek "social and psychological intelligence" from their sources. As in most social science interviewing, the content of this type of reporting depends on such factors as the subjective state and the personal and cultural frames of reference of the reporter. Considerable simplification is achieved by avoiding the complex problems of interviewing, which involve influencing persons to report psychological and social information accurately, and the infinitely more complex question of what constitutes accurate information on such topics.
There are various motivations or values which may underlie the resistance of a source to an interrogation attempt. The interest here is in any method through which these bases of resistance may be changed, outweighed, neutralized, or circumvented so that the person comes to behave in a manner he was originally strongly motivated to avoid. The particular form of behavior toward which attention is directed, the imparting of factual information, has various peculiarities. Some of these distinctive features are considered in the reviews. Few experiments, however, have dealt directly with attempts to elicit precisely this form of behavior. The attention of the contributors was broadened of necessity to exploit the relevance of experiments studying interpersonal influence on other forms of behavior.
This book does not pretend to examine the processes by which fundamental and lasting alterations of the value system of a subject come about. Nonetheless, in the review of experimentation on interpersonal influence (Chapter 6), it was imperative to consider knowledge developed through experiments that involved theoretical concepts such as "changes in attitude or belief." From the present perspective, the validity of such observations does not depend upon the degree to which observed changes truly reflect stable and lasting changes in the subject. When a determination is made that later behavior negates some value strongly affirmed earlier in the experiment, or the reverse, the experiment accords sufficiently with the questions being posed here.
Although the kind of influence attempt considered here represents a considerably simpler problem than the attitude changes or even attitude reporting used here for some inferences, it nonetheless involves the production and observation of complex, symbolic, learned human behavior. Thus, evidence regarding the manipulations that are possible of the salivary response or other simple responses of either animals or humans would not provide answers to the questions raised by this review.
We have attempted here to communicate scientific information to scientists, and the work originally undertaken for the U. S. Air Force has been revised and supplemented to this end. Emphasis has been placed on detailing the scientific implications of both the general and the specific subject matters, and their value for theory and research. The number of relevant questions left unanswered by the study points to the need for further investigation of the problem under consideration. The contributors represent a variety of scientific fields, and their material either separately or in the aggregate will undoubtedly hold interest for specialists in still other fields. The writing style here is akin to the broader style of papers designed for presentation at meetings of representatives from several different scientific disciplines.
This work might help the armed forces to offset the lack of knowledge that was in part held responsible for much of the success Communist captors achieved in interrogation of United States prisoners of war in Korea (64). Its value for this purpose is limited in that it assumes an interrogator who pursues his objective of developing information rationally. Past experience indicates that practices encountered by prisoners of war are not determined exclusively by considerations of logic (5). A rational examination of the problem cannot lead to predictions of a nonrational opponent's actions. Historically, there has been frequent resort to coercive practices for eliciting information, despite abundant evidence that such measures are relatively ineffective. Some estimates of what an opponent is likely to do, in addition to those based on considerations of what it will be feasible and advantageous for him to do, are required in devising measures for thwarting enemy exploitation attempts against prisoners of war.
If the present study also receives the attention of interrogators, it may offset their tendency to adopt the sensational stereotypes of interrogation on which many of them appear to have modeled their practice in the past.
Although the generally available literature would probably reflect in an over-all way the achievements of secret research, it is conceivable that some unknown discoveries or applications may have been made. Our contributors have indicated gaps in specially relevant knowledge, many of which would not be pursued intensively in the ordinary course of scientific development. Largely, however, the unanswered questions that are central to the topic of this book also point to critical gaps in present scientific knowledge.
It should be noted that interrogations almost invariably proceed in private. The two major sources of information about them are: practitioners of the "art" and their victims. The former are generally required to guard the details of their craft as secrets; the latter may have a limited perception, understanding, and memory of what they have experienced. It is possible that practice in some respects has advanced beyond the level of the inferences and conjectures presented here. In other respects, experience has proven that some potentialities of interrogation have been overestimated. Free access to the guarded handbooks of interrogators everywhere probably would not lead to any substantial modification in the general conclusions of this review.
Because of its defensive application during interrogation, one aspect of the problem receiving special consideration in this book is the ability of the source to thwart his interrogator by feigning psychological disorder. Malingering is a time-honored tactic. Its discussion here illustrates some of the implications of personality evaluation for manipulative situations.
Personality evaluation historically has been considered a clinical adjunct to manipulation. Its application requires the manipulation to be "tailor-made" to the specific individual differences encountered in the intelligence source. No comprehensive discussion of this topic has been attempted here for several reasons: (a) most means of personality evaluation require the willing cooperation of the subject, which is not likely to be obtainable from reluctant sources; (b) assessments not requiring the cooperation of the source (e.g., observation, graphology, analysis of speech or gestures) yield notoriously poor agreement among independent judges or observers, unless the behavior
is categorized into minutiae that are difficult to interpret meaningfully; (c) knowledge is lacking on how to effect maximal exploitation by differential treatment of sources on the basis of personality information, if it were available in reliable form; and finally (d) consensus on a theory of personality, which is critical to the integration and application of personality data, does not exist.
Published speculations that electrical stimulation of the brain might be employed for purposes of nefarious influence led the editors to believe initially that an examination of this area should also be included in this book. The notion that the action of the brain, and thereby the action of an individual, might be controlled directly is an ancient one. Electrical brain stimulation was one of the methods "Big Brother" used in Orwell's 1984 to control his subjects. Scientists, including Lilly (13, 29), Miller (39), Olds (42, 43) and Sargant (52), have indicated that recent experimental developments give some basis to the fiction-writers' conjectures. Observations, primarily from animal experimentation, led to the following surmises. First, and earliest, were possibilities suggested by observations of Penfield (46) that cortical stimulation might elicit "memory" and some spontaneous verbalization of information. Second, animal experiments raised the possibility that subjective experiences from subcortical stimulation might be so intense as to provide a basis for the administration of reinforcements of unprecedented strength. Also, the possibility was raised that organisms might be made more "teachable" by direct interventions of this kind.
The editors asked Sidney Marvin, M.D., Walter Reed Army Institute of Research, who has been working on subcortical stimulation for pain relief in human subjects, to review current knowledge and techniques in this area from the perspective of this book. Colonel Marvin found that experimentation had not progressed sufficiently to allow for other than conjectural statements regarding the questions raised. The editors believe it sufficient for the purposes of this volume to quote briefly the general conclusions of Colonel Marvin's report:
Also excluded from these pages is a consideration of the role of Pavlovian conditioned reflex theory in interrogation. The notoriety attained by this theory, as explaining the inspiration and effectiveness of Communist techniques of coercive interrogation (20, 32, 36, 52), has prompted studies by other investigators. A number of students of the subject (2, 4, 17, 18, 55, 56) have refuted the contention that Pavlovian theory influenced these practices, whereas Schein (56) and Farber, Harlow, and West (10) indicate the inadequacy of simple conditioning models to account for the kinds of complex behavior patterns produced in the course of interrogation.
Contributors have been free to choose eclectically whatever models and theories appeared most adequate to their respective topics. On the whole, matters of length, level of generality, and organization similarly have been left to the judgment of the individual contributors.
In those cases where the contributors to this book were not themselves highly conversant with interrogation practices, the editors have drawn on their own experience and on research that they have conducted on interrogation in advising the contributors and in editing the chapters. The editors thus actively sought to increase the relevance of the reviews to the realities of interrogation.
13. "Group for the Advancement of Psychiatry". Factors used to increase the susceptibility of individuals to forceful indoctrination: Observations and experiments. New York: GAP Publications Office, December 1956. GAP Symposium No. 3.
17. Hinkle L. E., Jr. "In Group for the Advancement of Psychiatry", Methods of forceful indoctrination: Observations and interviews. New York: GAP Publications Office, July 1957. GAP Symposium No. 4. Pp. 285-292.
18. Hinkle L. E., Jr., and Wolff H. G. Communist interrogation and indoctrination of "Enemies of the State. Analysis of methods used by the Communist state police. (Special Report), A.M.A. Arch. Neurol. Psychiat., 1956, 76, 115-174.
61. "U. S. Congress, Senate". Committee on Government Operations, Permanent Subcommittee on Investigations. Communist interrogation, indoctrination and exploitation of American military and civilian prisoners. 84th Congress, 2nd Session, Senate Report No. 2832, December 31, 1956. Washington, D. C.: U. S. Govt. Print. Off., 1957.
64. "U. S. Dept. of Defense". POW: The fight continues after the battle. The report of the Secretary of Defense's Advisory Committee on Prisoners of War, August 1955. Washington, D. C.: U. S. Govt. Print. Off., 1955.
65. West L. J. "United States Air Force prisoners of the Chinese Communists". In Group for the Advancement of Psychiatry, Methods of forceful indoctrination: Observations and interviews, New York: GAP Publications Office, July 1957. GAP Symposium No. 4, pp. 270-284.
When an interrogation is carried out for the purposes of intelligence, we may assume that it is intended to obtain information and not simply to produce compliant behavior on the part of the man being interrogated. One might describe an interrogator as a man who tries to obtain information from another man who may or may not possess it and who is not necessarily motivated to give the information if he does. The interrogator would like to have this man produce his information rapidly, accurately, completely, and without amendments or additions. In the words of the law, he wants "the truth, the whole truth, and nothing but the truth" — and often he wants this as soon as possible because the information that he seeks has perishable qualities. In the urgency of his need, he may interrogate a man who is injured, fatigued, or in pain. He may use various maneuvers such as prolonged or repetitive interrogation in order to overcome his informant's unwillingness to give information. In doing so he incurs the risk that his efforts may produce compliant behavior without eliciting accurate information.
The information that the interrogator seeks represents what his source still knows about various events, situations, organizations, devices, etc., to which he has been exposed in the past. The most complete and accurate information that he can hope to obtain can be only an approximation of the "true facts of the case" even "under
the best circumstances." It is the purpose of this chapter to inquire into these "circumstances," to consider some of the situations where interrogation may be carried out, how these affect an informant's ability and willingness to give information, and what the ideal circumstances of interrogation may be. The term "circumstances" is taken to mean "the condition of the man being interrogated and the situation in which he finds himself at the time." We shall not consider methods of interrogation or the nature of the interrogation process, but rather the limitations placed upon these by the state of the man who is being interrogated.
The human brain, the repository of the information that the interrogator seeks, functions optimally within the same narrow range of physical and chemical conditions that limit the functions of human organs in general; and it has, in addition, certain special limitations of its own. Any circumstance that impairs the function of the brain potentially affects the ability to give information as well as the ability to withhold it.
Some aspects of the physical and chemical conditions necessary for the normal function of the human brain are rather precisely known. The brain, like other organs of the human body, exists in an "internal milieu" which is maintained in a remarkably steady state by a great number of feedback mechanisms, some quite complex. Any disturbance in the constancy of this milieu brings into play homeostatic responses which may involve the great majority of bodily processes as well as the activities of the man as a whole. When environmental conditions pose a threat, these mechanisms are capable of creating major alterations in the internal economy and in many facets of behavior (59, 129, 130). It is largely in this manner that changes in the condition of the man being interrogated may affect his ability to give accurate information.
The temperature of the human internal environment is maintained very near 37.5° C. A rise in temperature above 44° C (112° F) (10, 33, 121), or a fall below 24° C (75° F) (73, 116) may damage the brain permanently or be fatal. An elevation of body temperature to 41° C (106° F) or above — which may occur during the fever accompanying
disease or during heat stroke — nearly always impairs brain function. Sometimes such impairment appears at much lower temperatures. Similarly, a depression of body temperature to approximately 31° C (88° F) — a level which is sometimes produced artificially during anesthesia or which may occur naturally in men after extreme exposure to cold — also impairs brain function (2, 9, 37, 62, 123). The nature of the impairments of brain function that occur during these and similar disturbances of homeostasis are discussed shortly. These impairments show many points of similarity, regardless of the conditions causing them.
The concentration of the fluid in the internal milieu is maintained remarkably close to 310 miliosmols per liter. An increase in its concentration (as may occur after hemorrhage or after injuries that create shock) may impair the function of the brain. A decrease in its concentration (which can take place if a man is forced to drink excessive amounts of water over a short period of time) also may impair brain funtion (38, 134). 1
The internal milieu contains a number of organic and inorganic substances in solution, and the concentration of each of these is also maintained at a remarkably steady level. Disturbances in the concentration of any of these substances, upward, downward, or in their relative proportions, may impair brain function. This impairment may be produced directly by the effect of these changes on the brain, or indirectly through the impairment of the function of other vital organs, which in turn produce a disturbance of the internal milieu. It would not be profitable to attempt to list the limits of the various elevations, depressions, or relative disproportions of these substances beyond which an impairment of brain function may occur. In practice no single change occurs, but rather a disturbance of the concentration of several. A variety of rather common conditions may produce such disturbances. Among these are excessive sweating, deprivation of water, diets deficient in salt, ingestion of excessive amounts of water or other nonsalty beverages over a short period of time, ingestion of excessive amounts of salt in food when water is restricted, ingestion of sea water in the absence of other water, poison-
1 The awkward term “brain function” is used here because there is no other that denotes all of the complex activities that the higher centers of the brain make possible. (We shall disregard its less complex and vegetative activities.) Terms such as “mentation,” “mental activity,” and “thinking” are inadequate. Consider all the human functions that are absent when the higher centers of a man's brain are inactive, and he is in “coma.” The term “brain function,” as here used, refers to all of these.
ings of various sorts, vomiting or diarrhea from any cause, burns, shock caused by injuries, hemorrhage, and impairment of kidney function (14, 30, 42, 43, 100, 134). Even very rapid breathing, which sometimes occurs in people who are anxious or afraid, may lead to chemical changes in the blood that cause disturbances of brain function (17, 36, 77, 96).
Many of the crude procedures that interrogators have utilized from time to time to make informants "tractable" and to "make them talk" have an adverse effect upon the composition of body fluids: the "hot box" or "sweat box"; the deprivation of water; the salty diet; the "water treatment"; the use of emetics to produce vomiting; and the use of cathartics such as castor oil to produce diarrhea. These procedures have been used by both European and Oriental interrogators in the historical past. They were also in use quite recently in Communist countries, and perhaps still are. American prisoners of war encountered some of them during World War II and the Korean War, and it is likely they will encounter some of them again in the future.
The brain, like other organs, maintains its functions by constant metabolic activity. The basis for this activity is energy obtained by the oxidation of the organic chemicals available from food. Thus, a constant supply of oxygen must be brought to the brain by the blood in the amount of approximately 50 cc per minute (40, 66, 102). The most common way by which the brain becomes deprived of oxygen is by failure of the circulation (65), which may be brought about by loss of blood from hemorrhage, by shock resulting from injury (which has an effect on the circulation quite similar to that of hemorrhage), or by illness. Such failure of the circulation may occur also when a man is forced to stand still in a fixed position for a long time. It is responsible for the common phenomenon of the soldier who faints while standing at attention (20, 22, 89, 90, 107). Transient circulatory failure is also involved in "emotional fainting," which occurs as a result of an acute fall in blood pressure produced by an "emotional" stimulus. Failure of the circulation has other adverse effects on cerebral metabolism in addition to the effects produced by relative anoxia.
Unlike other organs, the brain cannot use proteins and fats as sources of energy, and thus must rely on carbohydrates (65). It is, therefore, peculiarly sensitive to deficiencies in its supply of sugar, a substance normally present in blood. Small increases in sugar concentration, which usually occur after meals, have no discernible effect on brain function, but relatively small decreases in concentration
may have a distinct effect on mood and behavior. Small decreases set in motion homeostatic processes that lead to feelings of nervousness, restlessness, sweating, and inability to concentrate (27, 36, 92). A fall in blood sugar occasionally occurs in people who are anxious or fearful, and seems to contribute to their symptoms. A serious deficiency of blood sugar profoundly impairs brain function (65). This is one of the terminal events of starvation, and it contributes to the final stupor of the starved man (54, 67).
The brain is not dependent on the immediate level in the blood of any foodstuff other than sugar, but it does ultimately suffer if it is deprived of other foods over a long period of time. The prolonged deficiency of protein and fat, which is usual in general starvation, very probably contributes to changes in brain function occurring under these circumstances (1, 18, 54, 67, 82). More immediate and readily recognizable changes in brain function occur when the diet is relatively deficient in one of the accessory foodstuffs, or vitamins, which the body cannot produce itself, but which it requires in minute but definite amounts. Among these the "B" group of vitamins are the most immediately relevant to the brain, probably because they take part in various processes of carbohydrate metabolism. A relative deficiency of thiamin (vitamin B1) causes beriberi; a deficiency of niacin causes pellagra; a deficiency of vitamin B12 causes pernicious anemia; and a deficiency of pyridoxine (vitamin B6) causes nervousness, insomnia, weakness, abdominal pain, and difficulty in walking. All these diseases may be associated with pronounced changes in brain function (95, 99, 120). Beriberi and pellagra have been endemic among prisoners of war from time immemorial (54).
The normal function of the brain is also dependent upon the removal of metabolic end-products from the fluid that surrounds it. Especially when there is impairment of the lungs, liver, or kidneys (the organs primarily involved in maintaining the normal composition of the blood and in clearing it of toxic substances), the internal milieu becomes deranged and impairment of brain function follows. The kidney is the organ most often implicated in this phenomenon because it is highly vulnerable to many common disturbances. Dehydration, surgical shock, hemorrhage (14, 38, 43), or the circulatory impairment produced by very long standing (20, 22, 89, 90, 107), all may impair kidney function to such an extent that the internal milieu may be seriously disturbed.
fever, a profound lowering of body temperature, dehydration, overhydration, disturbances in the composition of the blood, disturbances of respiration, shock, hemorrhage, diarrhea, vomiting, poisonings, starvation (partial or complete), and even static postures that are too long maintained. The disturbance of brain function produced by each of these-and indeed that produced by any homeostatic disturbance, or by any physical or chemical assault upon the brain — is remarkably uniform in many of its features. Even though the symptoms produced by any given homeostatic disturbance (such as overbreathing or dehydration, for example) may exhibit certain idiosyncratic features (such as muscle cramps or thirst), there are fundamental common elements in the disturbances of brain function which follow from all these types of assault.
All severe and uncompensated disturbances of homeostasis produce a syndrome of disturbed brain function which, in civilian life, is most commonly encountered in hospitals. This syndrome (officially, the "brain syndrome") 2 occurs in acute and chronic forms (128), and includes deliria of various sorts, some of which have been given the names of drugs or diseases thought to cause them. The present consensus is that the "brain syndrome," whether acute or chronic, is fundamentally a single syndrome, regardless of its cause (24, 25, 29, 76, 92, 97, 108, 128, 131).
The "brain syndrome" in its fully developed state is an across-theboard impairment of brain function: an impairment of all those aspects of mental activity that are commonly tested when the physician undertakes to assess the "mental status" of the patient. A patient exhibiting this syndrome can no longer carry on his usual complex activities, assume his daily responsibilities, or cope with interpersonal relations. As its symptoms develop, he may become restless, talkative, and delirious; ultimately, he becomes totally confused and lapses into unconsciousness.
The full-blown "brain syndrome" usually occurs in people who are distinctly ill, injured, or depleted. Generally, such people are interrogated only when it is feared that their information may be irretriev-
2 The syndrome has many other names: The “organic reaction syndrome,“ ”symptomatic psychosis,“ “toxic-infectious exhaustive state,” “psychosis with somatic disease,” “dysergastic state,” among others.
ably lost. Their deranged condition is easily recognized, and the information that they give can be evaluated with this in mind. However, under less drastic conditions the syndrome may develop slowly and be difficult to recognize, and its existence may elude an interrogator.
In the earliest stages of the "brain syndrome," the subject experiences the various uncomfortable symptoms associated with his physical state: pain, fatigue, thirst, hunger, drowsiness, or the like. He loses some of his capacity to carry out complex responsibilities accurately, speedily, effectively, and to plan and delay his activities. This is especially noticeable in his impaired ability to meet new situations and his occasional lapses in dealing with familiar situations (24, 25). His interpersonal relations may deteriorate; conflicts arise which he might have avoided under other circumstances. He is likely to become emotionally labile, irritable, depressed, jumpy, and tense, and at other times to be unexpectedly blunted or apathetic in his reaction. His concern for the finer aspects of human behavior-for neatness, accuracy, honesty, veracity, and kindness, as well as patriotism and honor — may fall off to varying degrees, whereas at the same time he shows an increased and at times frantic concern for his more immediate bodily needs such as food, water, sleep, rest, and the alleviation of pain.
In this early stage of the syndrome, the only outward manifestations of disturbed behavior, other than those directly associated with illness, injury, or depletion, are likely to be a slight deterioration of dress, speech, and personal appearance. The subject's performance on short-term tasks, including psychological tests of moderate complexity, may not be outside the normal range (24, 25), especially if he is highly motivated at the time. Yet, despite his ability to rise to a short-term challenge, his performance on tasks generally will be slowed, less accurate, and less effective. If his premorbid level is unknown, a moderate deficit may be undetected. Frank disorientation may be absent. He is more likely to be vague and forgetful about time, place, and person, to have to be reminded, or to make a conscious effort to remember.
People who are injured, ill, or depleted by combat or exposure are often interrogated if they seem to be in good condition and capable of pulling themselves together. Under these circumstances they may be subject to disturbed brain function in this earlier and subtler form. The presence of this condition may not be recognized by either the interrogator or the man being interrogated, even though the source may wish to cooperate with his interrogator and may appear
to be mentally "normal." However, the subject's memory may be especially deceptive. There may be a distinct hiatus in his memory, without its being noticed either by the source or by one who examines him. More often he is vague, uncertain about details, and has temporary blocks of memory, especially for the nuances, or the finer (and sometimes the most important) details.
In this state, the subject may have no frank illusions, hallucinations, or delusions, but he overvalues small events, misinterprets, blames others, and accepts explanations and formulations which he might reject as patently absurd under different circumstances. He does not confabulate, but he may be willing to state that a report is "clearly true," or that an event "actually occurred," when in fact the report merely could be true, or the event might have occurred. His intellectual functions, his judgment, and his insight decline to a similar degree.
As the "brain syndrome" develops, the subject's dress, behavior, and speech deteriorate further. His orientation for time, place, and person becomes increasingly deficient. Initially he may have been quite aware of the impairment of his mental faculties, and his awareness potentiates the apprehension that he may experience. The subject is especially prone to become fearful if his illness is precipitated rather suddenly by acute infection, injury, poisoning, or dehydration. When it comes on more slowly or is due to starvation, his mood may be one of apathy or depression.
The subject is quite likely to have thinking difficulties and sensory experiences, illusions, delusions, hallucinations, and projective or paranoid thinking. Frequently these contain naively transparent elements of wishful thinking. If starved, he may believe that he is about to receive a large meal or he may see it before him. Such experiences may be frightening. If the syndrome develops gradually, he may perseverate, or pointlessly repeat a fragment of thinking, speech, or behavior; or he may confabulate and create figmentary "memories" to cover up actual defects in his memory. Such confabulation may occur even if the subject has a reputation for the utmost adherence to veracity. Since he may be more than usually suggestible (131), the combination of confabulation and suggestibility may make it possible to elicit from him a plausible story that is largely figmentary.
Other mental activities deteriorate also. His intellectual functions fall off. His capacity to calculate, to abstract, to estimate time, to recall items, digits, or stories is impaired. General information that one might expect him to know is not available to him. His judg-
ment is faulty. Although he may at first have had some insight into the fact that he has lost his faculties, later he may have none at all. His memory becomes defective, at first for recent or special events, and later for all sorts of events and topics. The subject's awareness is increasingly clouded, and he becomes more and more drowsy as the process advances to the borderline of the pathological.
The state just described is not uncommon among men who have been through prolonged combat (114) or through prolonged and depleting activities of any sort (4, 39, 83, 124, 135), in men who are injured, who are ill, who have undergone serious exposure to the elements, and who are malnourished or deprived of water. It can be assumed that the U. S. Armed Forces would not deliberately create such a state in prisoners of war, but it is quite likely to occur among them naturally, simply because men often become prisoners of war after strenuous combat, and may be ill or wounded. It can be assumed that future enemies probably will create such a state in American prisoners of war, although they may not do so with any sophisticated intent. Historically, it has been the common practice of captors, police, and inquisitors to isolate their prisoners in places that are cold, damp, hot, unventilated, unsanitary, and uncomfortable, to deprive them of food, fluids, sleep, and rest and medical care, and to beat, torture, harry, overwork and threaten them, as well as to question them interminably with leading questions. Such procedures have been used partly because they make prisoners more "pliable," more "ready to talk," and more "cooperative." They are very likely also to make the information obtained from the prisoner increasingly unreliable.
Some Circumstances under Which Brain Function May Be Disturbed without Demonstrable Disturbance of Other Bodily Functions
The phenomena just considered relate to men who have suffered some disturbance of their homeostasis — some measurable change in the internal environment affecting the body as a whole, other organs as well as the brain. People who experience the effects of isolation, fatigue, or sleep loss may show no measurable disturbance of their general homeostasis. They may nonetheless exhibit impaired brain function, for the brain has special vulnerabilities over and above those that it shares with other organs. It is possible to have disturbed brain function in the absence of any other significant alteration in homeostasis.
The brain of man is an organ that deals with "information," using this term in the technical sense as it is used in communications theory (135). The accumulation and transmission of information in this sense is a characteristic of all living organisms. The nervous system of the higher animals is a specialized apparatus capable of dealing with information in complex ways and thereby greatly increasing the general adaptive capacities of the animal. The exceedingly complex nervous system of man has this fundamental function. It takes in information from the organs of special sense, and from the sensory nerve endings within the body and its surfaces, and transmits this information to the brain. There it is analyzed, organized, evaluated, stored, and used as a basis for organizing the activities of the man as a whole. We might say that "information" arising from the configurations of minute amounts of energy is the substrate for the activities of the brain, in somewhat the same sense that "food" is the substrate for the activities of the gastrointestinal tract.
Deprived of information, the brain does not function "normally." It must have a certain quantity of patterned, meaningful, sensory input from the external environment, and some opportunity to organize its output as behavior (31, 60, 74, 86, 87, 91, 136). Nor can the brain perform one sort of activity continuously and maintain its efficiency. Even though the task undertaken is entirely "mental" (or, as one might say, involves only the carrying out of activity within the brain), and no significant changes in the general physical state of the individual occur as a result of it, the phenomenon of "fatigue" eventually supervenes, and brain function deteriorates (4, 32, 41). In addition to this, the brain requires "sleep" from time to time — a cessation of its "conscious" pattern of activities; otherwise its functions suffer (35, 64, 84, 98, 118). Thus the brain has special vulnerabilities of its own; it cannot function "normally" unless it receives a certain amount of information upon which to operate, and it cannot carry out a single pattern of activities unremittingly and indefinitely.
The experiments of Hebb and others (11, 44, 45, 53, 55, 56, 80, 109, 126), who have concerned themselves with "sensory deprivation," have consisted of putting men into situations where they received no patterned input from their eyes and ears, and as little patterned input as possible from their skin receptors. In some cases there was a
diminution in sensory input itself; but it appears to have been the lack of patterning, the paucity of information, that was important. The subjects were deprived of opportunity for purposeful activity. All of their other bodily needs were taken care of-food, fluids, rest, etc. Yet after a few hours the mental activities of the participants began to go awry. Their capacity to carry out complex tasks and to perform well on psychological tests fell away. They developed illusions, delusions, and hallucinations, a mood of fearfulness, and many of them discontinued the experiment.
Such experimentally contrived situations are by no means the same as those of persons in prolonged prison isolation, yet undoubtedly some aspects of these observations on sensory deprivation are applicable to our understanding of the reaction of the individual to prolonged isolation. It is well known that prisoners, especially if they have not been isolated before, may develop a syndrome similar in most of its features to the "brain syndrome" (57, 58, 91). They cease to care about their utterances, dress, and cleanliness. They become dulled, apathetic, and depressed. In due time they become disoriented and confused; their memories become defective and they experience hallucinations and delusions. In these circumstances their capacity for judgment and discrimination is much impaired, and they readily succumb to their need for talk and companionship; but their ability to impart accurate information may be as much impaired as their capacity to resist an interrogator.
Classically, isolation has been used as a means of "making a man talk," simply because it is so often associated with a deterioration of thinking and behavior and is accompanied by an intense need for companionship and for talk. From the interrogator's viewpoint it has seemed to be the ideal way of "breaking down" a prisoner, because, to the unsophisticated, it seems to create precisely the state that the interrogator desires: malleability and the desire to talk, with the added advantage that one can delude himself that he is using no force or coercion. The prisoner himself may be taken in by this and later stoutly maintain that the interrogator "never laid a hand on me." However, the effect of isolation upon the brain function of the prisoner is much like that which occurs if he is beaten, starved, or deprived of sleep.
The fact that some people, who have been through prison isolation before, or who can create for themselves an active and purposeful inner life of fantasy, can endure isolation for a long time (5, 15, 75) does not vitiate the fact that total isolation effectively disorganizes
many people who are initially obliged to undergo it, even when it is not carried out under circumstances of uncertainty and threat, as it usually is. There appears to be a wider range of variability in the capacity of men to withstand isolation, sleep deprivation, and fatigue than in their ability to withstand dehydration or fever, for example, even though ultimately brain function may be deranged by all these conditions. We shall consider the basis for this.
For reasons not yet known, a brain cannot continue to function without occasional periods of sleep. The amount of sleep men require is quite variable. Some can function effectively for fairly long periods with relatively few hours of sleep obtained at irregular intervals (64, 68). Under experimental conditions men have been known to endure for more than a hundred hours without sleep at all (16, 35, 46, 72, 98, 118, 123), and for more than two hundred hours with only a few brief naps (64). Yet most people deteriorate markedly after about seventy-two hours without sleep, and all deteriorate sooner or later (35, 46, 64, 118, 122). The highest functions suffer first; the capacity to cope with complex and changing situations without making mistakes or errors in judgment is often the first to go. This is followed by a deterioration of dress, speech, and behavior; dullness; emotional lability; defects of recent memory; disorientation; hallucinations, delusions, thinking disturbances; and impaired judgment and intellectual functions, all increasing in severity with the passage of time (35, 46, 64, 72, 98, 118, 122). Even at a fairly late stage of this deterioration, people faced with an acute challenge and highly motivated to meet it may briefly perform complex tasks quite adequately; but ultimately even this capacity is lost (4).
Sleep deprivation affects brain function directly, while producing little or no change in the general internal milieu. Efforts to demonstrate a disturbance of the general homeostasis consistently associated with lack of sleep have been largely futile (84, 118). The constituents of the blood and the function of organs other than the brain may show little or no abnormality in those who have been without sleep for many hours. People whose thinking and behavior are seriously deranged may show "nothing wrong with them" on physical examination or various chemical tests. Their demonstrable defect is a disturbance of brain function.
“Fatigue” is a term which has more than one scientific meaning, as well as more than one popular meaning. We shall use it to denote a group of somewhat similar phenomena which occur in muscles, in reflex arcs, and in the brain. “Muscle fatigue” is precisely defined and denotes the deterioration in the function of a muscle produced by its steady or repeated activity. The phenomenon is readily measured and reproduced. It is associated with measurable changes within the muscular system, and it has its counterpart in the “muscle fatigue” that occurs in the intact man after muscular activity. The fatigue of neural arcs is similarly definable and reproducible. On the other hand, “fatigue” of the man as a whole has been given various definitions (4, 32, 88). It is often seen in people who are depleted or ill, but no measurable bodily change is necessary to it or consistently associated with it. The “fatigue syndrome” may be produced in a man if he is put to performing a given task over and over, without rest and without change. After a while he performs this task less rapidly, less efficiently, less effectively, and with more mistakes. This falling off in his performance on the specific task is usually accompanied by a feeling of “weariness,” or “fatigue.” His impaired performance on this task is not necessarily associated with any other changes in his bodily functions, and his performance on any other task may be unimpaired; indeed, if he is suitably motivated, he may perform even the task that fatigued him quite adequately for a short period of time (4, 32). In addition, the rapidity with which the fatigue syndrome develops is influenced by the attitude of the man to the task that he must perform (4, 32, 81).
“Fatigue” and impaired performance of the degree just described, or of even greater degree, is readily produced by purely “mental” tasks; and if such tasks are continued long enough, performance deteriorates to the extent that the task can scarcely be carried on (4). However, the most extreme degrees of fatigue that have been studied have been associated with combat or with other extremely trying military operations where muscular activity, lack of sleep, and sometimes injury played a part in their production (4, 8, 13, 32, 41, 49, 50, 81, 83, 88, 111, 114, 115, 135). Bodily changes were therefore present. The fatigue that occurs in combat or other military operations is like that occurring during the prolonged and unremitting interrogation carried out by state police in various countries (57). The observed mental phenomena are quite similar in both cases. The profoundly
fatigued man, after combat or military operations, or after prolonged interrogation, shows a deterioration of his speech, dress, and general behavior, emotional lability and blunting, confusion, disorientation, defects of memory, hallucinations, delusions, illusions, paranoid thinking, impairment of intellectual functions, loss of judgment, and very little insight (4, 57, 114). Perseveration and confabulation may occur under these circumstances, as they also may after sleep loss (114). In addition, profound anxiety is often exhibited by those who have been in prolonged combat and who have undergone terrifying experiences (114).
THE MANIFESTATIONS OF DISORDERED BRAIN FUNCTION PRODUCED BY ISOLATION, SLEEP DEPRIVATION, AND FATIGUE
The symptoms of disordered brain function that occur under these conditions differ little, except in detail, from those of the "brain syndrome," as this is produced by disturbances of homeostasis. It is easy to differentiate a man who has been long isolated or who is profoundly sleepy or tired from one who is suffering the effects of pneumonia, gunshot wounds, or starvation; but this differentiation is made upon the basis of symptoms and signs other than manifestations of disturbed brain function. It is not profitable to argue whether or not the symptoms produced by isolation, fatigue, and sleep deprivation should properly be classified under the "brain syndrome." However, if one assumes that the effective performance of complex tasks, alertness, orientation, memory, discrimination, and judgment are dependent upon the function of the brain, then there can be no doubt that isolation, fatigue, and sleep deprivation produce disturbances of brain function. If one accepts that the function of the brain is always associated with electrochemical events occurring within it, then these changes in brain function are, in fact, "organic," as are all brain functions. So far as "organicity" is concerned, the effects of isolation, fatigue, and sleep deprivation upon the brain are different from those produced by pneumonia, starvation, or gunshot wounds primarily in the rapidity of their occurrence and the extent to which they can be reversed. Not all of the phenomena that may occur as a part of the "brain syndrome" have yet been described as occurring during isolation, fatigue, or sleep loss, but this seems to be a function of the limited number of observations that have been made on people who are subject to these conditions in extreme degree.
On the other hand, there is a difference in the predictability of the point at which disturbed brain function will be produced by these various circumstances. One can state within rather narrow limits the
levels of body temperature, blood glucose, or oxygen saturation, beyond which a severe disturbance of brain function can be expected. Isolation, sleep loss, and fatigue, however, present no such narrow limits. It is probably correct to say that if any of these are carried on long enough they will disorganize the brain function of anyone; but the differences in the ability of man to withstand these conditions are very wide. Under experimental conditions, some people have succumbed to sensory deprivation within one-and-a-half hours, whereas others have maintained adequate function for thirty-six hours or more (126). Under prison isolation, as this has been carried out by Russian and Eastern European state police, most prisoners developed symptoms of disorganization within three to six weeks (57); but some have been known to endure this for many months (15, 75), and some have succumbed within days. After forty-eight hours without sleep, some people become disorganized and ineffective, whereas others have been known to go as long as a hundred hours with their functions largely intact (35, 64). A task that will fatigue some men within a short while can be carried on by others for many hours with no evidence of fatigue (4, 32).
It must be conceded that these differences in the ability of men to withstand isolation, fatigue, and sleep deprivation may be based on subtle differences in their genetically inherited characteristics; but if this be true, no evidence has yet been brought forward to substantiate it. On the other hand, there is abundant evidence to indicate that the personality of a man and his attitude toward the experience that he is undergoing affect his ability to withstand it. People who enter prison with attitudes of foreboding, apprehension, and helplessness generally do less well than those who enter with assurance and a conviction that they can deal with anything that they may encounter. Those who readily withdraw into a life of meaningful fantasy and intellectual activity seem to do better than those who are dependent upon activity and interaction with other people. Some people who are afraid of losing sleep, or who do not wish to lose sleep, soon succumb to sleep loss; others, convinced that they can stay awake indefinitely, have done so for well over a hundred hours (64). A great number of reports from industry, and from experimental observations, indicate that the attitude of workers and experimental subjects is as important in determining the rate at which they fatigue as are the tasks they undertake (4, 81). Troops, carried without sleep or rest to the limit of endurance, have been able to straighten up and present a smart appearance for a short time when an appeal was made to their pride (118). In short, the brain, the organ that deals with infor-
mation, also organizes its responses on the basis of information previously fed into it. This information, in the form of a personality developed through the experiences of a lifetime, as well as immediate attitudes and the awareness of the immediate situation, conditions the way that the brain will react to a given situation. There can be no doubt that personality, attitudes, and the perception of the immediate situation seriously influence the ability of a brain to endure the effects of isolation, fatiguing tasks, and loss of sleep.
Some Conditions under Which the Function of the Brain May Be Disturbed by Factors Not Intrinsically Harmful
In a certain sense one might regard isolation, sleep loss, and fatigue as direct assaults upon the brain; for the first can be regarded as depriving the brain of the substrate of its operations, and the last two can be regarded as leading to the depletion of processes in the nervous system which can only be repaired with a period of rest. However, hypotheses such as these cannot easily be applied to the explanation of the human reactions to pain, hunger, and situations interpreted as dangerous or threatening. Here one is dealing with a form of sensory input that may be of great intensity, but which is not in itself disorganizing to brain function even if continued indefinitely. Pain of fairly high intensity may be borne by individuals over a long period of time without necessarily producing impairment of their highest integrative functions (8, 50). In the heat of combat or of physical contests intensely painful conditions may pass quite unnoticed and may not impair performance. When the attention is focused elsewhere by various means, such as hypnosis, the pain of childbirth or of surgical operations may be withstood and not be remarked upon (103, 125). The reaction to pain is therefore quite variable (3, 6, 7, 28, 48, 50, 63, 69, 78, 93, 94, 112, 132, 133). So likewise is the reaction to hunger (21, 34). Hungry men have produced intellectual and artistic output of a high order, and have been responsible for extraordinary military feats. Men have undergone prolonged fasts without significant impairment of their highest faculties. And, so far as dangerous situations are concerned, there are some who are stimulated or even exhilarated by them, and many others who act as if they do not regard them as dangerous at all.
tolerate them indefinitely. The weight of evidence is that it is not the sensory input itself, but the reaction of the individual to this input which may adversely affect his brain function. This is not so with isolation, sleep deprivation, and fatigue, where the effects are intrinsically adverse, and the reaction of the individual is a factor only in determining how long these effects can be withstood. With hunger, pain, and signals of danger, the adverse effects on brain function may be entirely the result of the reaction of the individual. Activities of the brain, initiated in response to incoming information, lead to an impairment of brain function. This special vulnerability of the brain to its own activities, long suggested by clinical observation, has recently received experimental support (24, 25).
The syndrome commonly associated with the reaction to hunger is slow in developing, but it can be expected to occur in the majority of those who are exposed to prolonged hunger from any cause (18, 67). It has been seen among starved populations (1), among inmates of concentration camps (54, 70), and among prisoners of war (47, 61, 67, 104), and has been reproduced experimentally (18, 67, 82). This and other rigors of the prison camp experience were probably responsible for a good deal of the symptomatology that occurred among American prisoners in Korea (79, 105, 113). People deprived of food very soon develop a persistent hunger, which does not leave them until death approaches or nutrition is restored (18, 21, 54, 67). Accompanying this hunger there is a constant preoccupation with food, which may encompass the greater part of waking thoughts and activity (18, 54, 67). As starvation progresses, the niceties of dress and behavior are neglected, and if the lack of food carries with it a threat of death, behavior may cease to be governed by the restraints of "honesty," "unselfishness," "pride," and "honor," which are active under normal circumstances; in short, the very highest integrative functions drop away (18, 54, 67). During the earlier stages of hunger, irritability and emotional lability are the rule, but later profound and continuing apathy occurs (18, 19, 54, 61, 67, 104). In the most advanced stages of inanition, defects of memory, confusion, hallucinations, delusions, and intellectual deficits become evident (1, 54, 67). Advanced inanition is associated with major changes in the physical state of the individual. Although it is quite possible that this state is directly responsible for the derangement in brain function which takes place, the disturbances of behavior and of mood which occur during the
It is not to be taken that the reaction to hunger is always the same. In starved communities, in concentration camps, and in experimental situations, some people endure hunger for a long time, and maintain their highest level functions with very little evidence of disorganization until the effects of illness or lack of food supervene (1, 18, 19, 54, 67, 70). Such people, although they do feel hunger and are aware of it, are able to engage in thought and behavior other than that centering around a preoccupation with food; their symptoms are less outstanding and their behavior is more "normal." So far as is known, this greater ability to withstand hunger has no constitutional or genetic basis; on the other hand, the attitude of the man to the experience that he is undergoing appears to be of great importance in determining his capacity to endure hunger.
The investigation of pain has been especially enlightening in this regard, because many careful laboratory studies have defined the difference between the "sensation of pain" and the "reaction to pain" (3, 50, 52, 110, 132, 133). The sensation of pain seems to be roughly equal in all men, that is to say, all people have approximately the same threshold at which they begin to feel pain, and when carefully graded stimuli are applied to them, their estimates of severity are approximately the same (6, 28, 48, 50, 51, 71, 85). In general, the reaction to pain is in proportion to its severity, and the most intense pains incapacitate men for any sort of complex function during the period of their duration. Yet exception must be taken even to this statement, for when men are very highly motivated, as they may be when their own lives or the lives of others are at stake, they have been known to carry out rather complex tasks while enduring the most intense pain. The variability of human reactions to the moderately severe grades of pain, such as those found in various diseases, is notorious. Some people perform quite effectively over many years while experiencing the pains of chronic headache, peptic ulcer, arthritis, or similar conditions; others with like amounts of pain are severely incapacitated (3, 6, 7, 8, 28, 48, 50, 63, 69, 78, 93, 94, 103, 112, 125, 132, 133).
The adverse reaction to chronic pain of moderate severity is clinically familiar. It is characterized by withdrawal from the more complex and responsible functions of life, a certain amount of irritability
and emotional lability, and concentration upon personal comfort and survival at the expense of the needs of others and of the society. Under experimental circumstances, those who try to "carry on" while experiencing moderate pain show impairment of their performance on complex tasks, impairment of decision making, loss of efficiency, and difficulty in estimating time (8) — symptoms which would be expected to occur in the early stages of the “brain syndrome” and much like those of people who have suffered the destruction of a small segment of their cerebral hemispheres (24, 25).
It is possible that the differences in the way that various people react to pain may be partly determined by their constitutions, for it sometimes appears to the clinical observer that people of “mesomorphic” build, the heavily muscled and big-boned individuals, are those who react to pain with stoicism or with anger and a mobilization for action that temporarily enhances their performance; whereas the lighter and asthenic “ectomorph” often reacts to pain with withdrawal, incapacitation, self-concern, and anxiety. Yet the exceptions to this are many, and the variations in the reaction of the same person from time to time are great. In general, it appears that whatever may be the role of the constitutional endowment in determining the reaction to pain, it is a much less important determinant than is the attitude of the man who experiences the pain (3, 6, 7, 48, 50, 52, 69, 94, 110, 112, 125, 132, 133).
Threats of any sort, direct, implied, or symbolic, are not necessarily derived from sensory input which is intrinsically “unpleasant.” The question of the intrinsically noxious nature of a threat does not even arise. Complex situations, symbols, and small cues arouse potent reactions entirely because of the interpretation put upon them. In different men similar situations produce different reactions. Some men react to ostensibly dangerous situations with continued effective performance. When men react to such situations as threatening, and when their reactions are characterized by anxiety or other intense emotions, these reactions may disorganize their brain function. Intense anxiety, for example, is sometimes associated with defects in every area of performance that is impaired in the “brain syndrome.” Threat is not usually thought of as a “physical” stimulus and the syndrome of anxiety is quite distinct from “brain syndrome.” However, the defect in function that occurs in the anxious man is quite real.
Pain, hunger, and threats are usually thought of as psychological stimuli because the conditions that produce them may not be noxious unless perceived to be so. The features that determine whether or not a man will perceive a given situation as noxious — his personality, his past experiences, his immediate mental set, and the characteristics of the situation — are outside of the scope of this chapter, but we must take due note of their importance.
On the other hand, the psychological reactions to pain, hunger, and threats will be discussed. These reactions are not called "organic reactions," and they are not considered to be part of the "brain syndrome," but this is a sterile distinction.
The same considerations that were applied to the reactions to isolation, fatigue, and sleep loss apply also to those of pain, hunger, and threats. Insofar as mood, thought, and behavior are functions of the brain, the disturbances of mood, thought, and behavior that occur in reaction to pain, hunger, or threat are disturbances of brain function. Insofar as all brain functions are concomitants of electrochemical events in the brain, these disturbances are "organic." Yet the impaired function associated with anxiety is distinguished from the "brain syndrome" because of its reversibility, and because of the relatively greater disturbance of mood and behavior than of intellectual functions, memory, or orientation that usually occur with anxiety. Yet impaired brain function, not entirely distinguishable from the organic reaction pattern, and in effect "permanent," may in some cases be produced by anxiety alone (24, 25).
Quite aside from the question of whether or not the reaction to threats, hunger, and pain may be directly associated with changes in brain function, there is no doubt that it may be associated with notable changes in the function of other organs. When environmental conditions pose a threat, adaptive mechanisms are capable of creating important changes in the internal economy (59, 129, 130). Manifestations of disturbed function of the gastrointestinal and cardiovascular systems are most frequently reported by prisoners (57), but disturbance of any organ system may occur. In the absence of other causes of disease, dysfunctions produced in this manner are not usually fatal, although they may be. When combined with the effects of isolation, loss of sleep, or starvation, they lead to rapid deterioration and sometimes to death. Even if one were to overlook entirely the
direct effect upon the brain of reactions associated with anxiety, fear, or depression, the indirect effect of the homeostatic derangements that often occur at the same time would ultimately be deleterious to brain function.
As the master organ of human adaptation, the brain functions as a whole in enabling man to carry out the exceedingly complex activities of life in the societies that he has erected. Even impairment of the lower level functions of the nervous system, for example, of sight, hearing, or motor function, to some extent impairs his performance of these activities. Yet many of the highest level activities of man remain possible despite such impairment. Milton was blind, Beethoven was deaf, and Winston Churchill was not the last statesman to carry on after he had suffered a cerebro-vascular accident. The part of the brain essential to these highest level activities, without which they cannot be carried on, is the most recent evolutionary development and the part particularly well-developed in man: the cerebral hemispheres, the neopallium. It is this that must be intact for the performance of the creative and responsible tasks that confront a mature man and, in fact, for all those "conscious" activities that are part of being an alert, sentient, and civilized human being.
Within the cerebral hemispheres are many discrete pathways which have to do with lower level functions, including those of sight, hearing, and the motor functions that have been mentioned. These lower level functions are relatively localized. Damage to their pathways can impair them temporarily. But the highest level functions, those necessary for the adequate expression of human needs, appetites, and drives, those which provide the mechanisms for symbolic activity ("memory," "abstraction," "cognition," "integration," "reason," and so on), and those which enable men to tolerate frustration, to deal effectively with threats, and to maintain effective and well-modulated defense reactions, do not appear to be localized within the hemispheres (24, 25).
uring") characteristics, as well as many other characteristics not yet duplicated by man-made apparatus (26). The cerebral hemispheres have no specific pathways associated with "abstraction," "cognition," "integration," "reason," or similar mental activities. Rather, they include a maze of potential pathways, over any of which the complex patterns of activity associated with the highest integrative functions may be set up. Thus, when any part of the hemispheres is damaged, none of the highest integrative activities are entirely lost, but the capacity to perform all is impaired to some extent. It is for this reason, undoubtedly, that anything that impairs the function of the cerebral hemispheres-direct injury, drugs, toxins, diseases, homeostatic disturbance of all sorts, isolation, sleep loss, fatigue, and some reactions to pain, hunger, and threats — ultimately produces a global impairment of the highest integrative activities.
Yet it is also true that some high level functions of the brain are more vulnerable than others. It seems to be a characteristic of the central nervous system that those functions that are "newest" and most complex, those which have appeared most recently in evolutionary development, are most vulnerable and drop out first when the function of the brain is impaired. The cerebral hemispheres provide no exception to this general rule. When they are impaired, the first functions lost are those that are thought to be the most complex and to have been acquired most recently by civilized man: the capacity to carry out the highest creative activities, to meet new, challenging, and complex situations, to deal with trying interpersonal relations, and to cope with repeated frustration (24, 25). Relatively small degrees of homeostatic derangement, fatigue, pain, sleep loss, or anxiety may impair these functions.
As impairment of brain function continues, somewhat less complex activities deteriorate. There is a lessening of the speed and efficiency with which the ordinary tasks of daily life are carried out. Concern about "accuracy," "propriety," "moral rectitude," "honor," and "feelings of other people," and similar "socially oriented" behavior falls away, and an increased concern about sleep, rest, comfort, food, and other bodily needs becomes apparent (24, 25, 54, 67). There is less adherence to niceties in speech, behavior, and dress. Emotional displays lose some of their social orientation. Judgment and insight are less acute. These complex aspects of brain function may be distinctly impaired, whereas orientation, memory, recall, and the capacity to perform well on psychomotor tests are still intact.
With still further impairment awareness becomes clouded. It is at this point that misinterpretations, illusions, delusions, and hallucinations may appear, and "delirium" may occur. Large defects in memory and profound impairment of discrimination and judgment are evident. As they develop, confabulation may take place and perseveration is likely.
It is notorious that threats, pressures, and deprivations, skillfully manipulated by police and interrogators with long practice in their use, will "break" almost any man, "soften him up," "make him cooperate," and "make him talk." They succeed because the most complex, the most "civilized," and the most "socially determined" aspects of human behavior are most affected by these procedures. The "less civilized" behavior patterns, directed at comfort and survival, are brought to the fore in a man whose capacity for judgment and discrimination is diminished. "Honor," "bravery," "security," "loyalty," and "patriotism" then have less weight in determining his behavior; pain, fear, and confusion have more. Rationalizations come easier to him, and points that once seemed important are now unimportant. He becomes more "willing to give information."
"Willingness to give information" is an "attitude," a "mental set," an "increased propensity of the individual to react in a given manner." It is not, in itself, a discrete function of the brain; it is a statement about the likelihood that a given pattern of reaction will occur, provided this reaction pattern can occur. Various degrees of "willingness" exist so long as the brain has any "ability" whatsoever to give information. The "ability to give information" is a statement about the capacity of the brain to furnish information: the possibility that it can do so under any circumstances. "Willingness" and "ability" are not necessarily parallel.
It is easy to see why various police procedures often increase the willingness of men to give information. So far as one can tell, the willingness to give information is not determined by any constitutional factor or by the direct action of any agent from the outside, but by information already within the brain, what might be called its "directions for action." Most of the "directions" which call for a prisoner to withhold information were implanted there by his society. They are the sum total of those learned reactions that have
to do with "loyalty," "honor," "propriety," "security," and so on. As brain function is impaired, information derived from past experience generally becomes less potent as a guide for action, whereas information derived from the immediate experience, pain, thirst, discomfort, and threats to life, becomes more potent. The "attitude" is likely to change, and the man becomes more "willing" to do whatever is necessary to secure his comfort and survival.
The new "attitude," the newly increased propensity, is directed toward doing whatever is necessary to secure comfort and survival. It is directed toward "compliance," toward doing what the situation seems to demand. This new state of the informant may be a trap to an interrogator, especially if he is a vigorous and persistent man with a good hypothesis as to what he might uncover. He is now dealing with a man who is likely to have lost some of his finer capacity for discrimination and judgment, whose insistent physiologic needs impel him toward ready solutions that may serve to relieve him of his discomfort, whose memory for details may be fuzzy and confused, and who is more than usually ready to accept a plausible suggestion. The source is, indeed, more prepared to talk, but he is also more likely to be inaccurate and to give false, misleading, incomplete, or inexact information, of a type like that which his interrogator happens to be seeking. The fact that the giving of this information does not redound to his credit or to his long-term self-interest and the fact that he is prepared to state that it is true, and later to defend his statements, should not be taken as evidence of its accuracy.
Note that these are statements of probability; they are not absolute. "Willingness" is not necessarily enhanced as "ability" deteriorates. Our simple hierarchical outline of the way that brain function falls off is generally true. All the disturbing influences that we have mentioned can be accompanied by the "brain syndrome," and can ultimately cause disorganization and unconsciousness. However, one cannot make a more exact statement, because the precise nature of the symptoms and the facility with which they are produced are dependent upon the personality of the prisoner, what has happened to him before, and how he views the circumstances in which he finds himself at the time (24, 25, 131). These factors have a great deal to do with the form of the "brain syndrome" produced by disturbances in homeostasis. They determine whether a man becomes garrulous or withdrawn, anxious or angry, paranoid or trusting. They likewise determine the form of the "brain syndrome" produced by isolation, sleep loss, and fatigue, and they further have an important influence upon his ability to withstand pain and hunger, and they
Unwillingness to give information is a mental set. If it is strongly imbedded in a man before his capture, it may continue to govern one aspect of his behavior right up to the point of delirium or unconsciousness, no matter what symptoms he may develop. Some people -criminals adhering to the "code of the underworld" (127) as well as prisoners of war adhering to the "finest military traditions" (119) — do not give information although they reach the point of disorganization or death. The evidence suggests that a learned reaction pattern, if sufficiently reinforced, can sometimes govern a specific aspect of behavior as long as a man retains the capacity to carry out that behavior.
From the theoretical point of view it is hard to escape the conclusion that a man is best able to give accurate information when he is in an optimal state of health, rest, comfort, and alertness, and when he is under no threat. This would seem to be the optimal situation for interrogation. Any attempt to produce compliant behavior by procedures which produce tissue damage, disturbances of homeostasis, fatigue, sleep deprivation, isolation, discomfort, or disturbing emotional states carries with it the hazard of producing inaccuracy and unreliability.
However, it is often necessary for the interrogator to question people who are experiencing moderately severe effects of illness, injury, fatigue, discomfort, or anxiety. A body of practical experience indicates that relatively reliable information can be obtained from most such people, if the information sought is neither complex nor extensive. The interrogator faces two special hazards under these circumstances. First, the source may have a fairly serious degree of mental disturbance that is not immediately evident and it may escape the interrogator's attention. Second, any informant in a threatening situation is liable to say whatever will please his captors, even though he may not do so intentionally. These ever-present hazards of interrogation are enhanced under adverse circumstances. It may be assumed, in the absence of evidence to the contrary, that the simpler, the briefer, and the more readily verifiable the information that is sought, the more likely is the evidence of the source to be of value.
On the other hand, granting that various procedures designed to make men more compliant will impair their ability to give accurate information, do these procedures not cause men to give more information than they might otherwise have given? Cannot a man be made to reveal information against his will?
Disordered brain function is indeed easily produced in any man. No amount of "will power" can prevent its occurrence. It can be produced without using physical means, that is, by fatigue or sleep deprivation. Since it may be associated with mental clouding, confusion, lack of discrimination, impaired judgment, and increased suggestibility, it is probably true that most men can be brought to a state where they will agree to statements that are dubious, incomplete, or quite inaccurate. Under these conditions some men will make up entirely fictitious stories incriminating themselves. Therefore, it is usually not difficult to obtain signed "confessions" that are biased, incomplete, inaccurate, or even totally untrue. This is the means by which Communist state police have produced false confessions with great regularity (57), although not with universal success (12, 101, 106, 117, 119).
Most people who are exposed to coercive procedures will talk and usually reveal some information that they might not have revealed otherwise. However, there is no evidence that a man must always reveal a specific item of information that he possesses. Disturbed brain function of the subject does not allow the interrogator to abstract information at will. An interrogator may occasionally trick a disturbed man into revealing bits of information that he had intended to conceal, but information so revealed is likely to be limited and interspersed with unreliable statements. If he elects to do so, a prisoner may endure to death or disorganization without revealing what he knows. Very few men, however, can hold themselves to such rigorous behavior through all the vicissitudes of captivity.
7. Beecher H. K., Keats A. S., Mosteller F., and Lasagna L. Effectiveness of oral analgesics (morphine, codeine, acetylsalicylic acid) and problem of placebo "reactors" and "non-reactors." J. Pharm. exp. Therapeutics, 1953, 109, 393-400.
46. Goodnow, Jacqueline, and Rubinstein I. Effect of sleep loss on problem solving behavior. Washington, D. C.: Walter Reed Army Inst. of Research, Walter Reed Army Medical Center, March 1957, WRAIR-40-57.
57. Hinkle L. E., Jr., and Wolff H. G. Communist interrogation and indoctrination of "Enemies of the State. Analysis of methods used by the Communist state police. (Special Report), A.M.A. Arch. Neurol. Psychiat., 1956, 76, 115-174.
89. Moyer J. H., Mills L. C., Ford R. V., and Spurr C. "The effect of head-up tilted position and ambulation on renal hemodynamic and water and electrolyte excretion in patients with hypertension, with and without renal damage". J. lab. clin. Med., 1955, 45, 179-190.
106. Segal J. Factors related to the collaboration and resistance behavior of U. S. Army PW's in Korea. Washington, D. C.: Human Resources Research Office, George Washington Univer., December 1956. HumRRO Technical Report 33
This chapter is concerned with experimental investigations of the effects upon human behavior of a reduction in either absolute or relative amounts of sensory or perceptual stimulation. Experimental efforts to achieve such a reduction in environmental input to the organism have been referred to in the literature by various terms, of which the most common appear to be "sensory deprivation," "sensory isolation," and "perceptual isolation." Although it is difficult to distinguish between the use of these terms, an effort will be made to examine the similarities and differences found in the growing number of investigations in this problem area. Regardless of the differences in descriptive terminology, these conditions have been observed to produce marked changes in the behavior of subjects exposed to them. These changes have included modifications in thinking, perception, and feeling states, as well as an increase in imagery, often bizarre in content. It is the purpose of this chapter (a) to examine briefly the sources of interest in this problem, (b) to review the extant experimental literature in order to assess the current status of knowledge________________
Partial assistance for the preparation of this report was provided by the Office of Naval Research, Contract Nonr. 1866(29). The author would also like to thank R. R. Holt and P. H. Leiderman for their critical reading of the manuscript and their many helpful suggestions.
Interest in these problems by diverse groups long antedates the current concern. Small (71), in 1900, summarized a great deal of data relevant to the relationship between social life and solitude. He offered the generalization that
Until recently our primary sources of data on this problem have been the autobiographical accounts of prisoners, explorers, and shipwrecked sailors. They have reported rather dramatic, often gripping accounts of the response to isolation in narrow and cramped prison quarters, in endless days at sea, or in stark and desolate polar regions. Reports by Bombard (11), Byrd (15), Ritter (63), and Burney (13), to mention only a few, describe the inexorable monotony of these conditions and report unusual changes in thinking, feeling, and perception. Hallucinationlike experiences seemed quite common and a marked hunger for contact with people and things were manifest. These accounts, as well as a variety of other anecdotal literature, suggested the key role of a varied external environment against which to verify internally arising percepts and ideas. Two recent reviews have summarized these somewhat scattered, largely autobiographical reports (50, 72).
tributed to converging influences coming from three major sources. The first of these sources has been the rapid pace of development in several scientific disciplines. Advances in neurophysiology have led to a gradual revision in our conception of the nervous system and have produced data that provide changing physiological models for psychological events (37). There are now available increasingly sophisticated electrophysiological methods of measuring neural function at various levels of the nervous system. Recent reports (29, 61) have demonstrated electrical changes in the central nervous system followfag reduced sensory input. These changes are believed to have chemical consequences. Neurochemistry has begun to provide techniques for evaluating the nature of these consequences.
Psychoanalysis is another discipline in which interest in these issues is growing. Here, increasing emphasis is placed on the importance of understanding ego functioning and its role in mediating behavior. From this viewpoint, the question may be raised, "If the ego is the executive aspect of personality, enabling the individual to cope with reality, what becomes of ego functions in the absence of an external environment with which to cope?" Although Freud (31) dealt with these issues early in his work, interest in an experimental approach to this problem is relatively recent. The work of Hartmann (36) in elaborating the theoretical basis of "ego psychology" is important in this development.
A second major source of interest in human response to restricted environments has come from the military establishment. Technological developments, as seen in a variety of military applications, have given the pursuit of these questions a new urgency. With the advent of space craft, isolated radar stations, and a generally increased reliance on automated equipment, the problem of efficient functioning in severely restricted, monotonous environments is no longer merely of theoretical or academic interest. The problem of efficient personnel selection and utilization, in a wide variety of these circumstances, has provided marked impetus to the initiation and development of research programs dealing with reactions to limited sensory and social environments.
In this connection, the experience of prisoners of war with Communist "thought-reform" has had similar effects. The revelation that isolation may be one factor in the susceptibility of humans to radical changes in customary behavior and beliefs has heightened interest in the study of isolation. The shocked fascination of the general public, not excepting the scientific community, has served to highlight the need for a systematic understanding of the effects of physical and
social isolation on behavior. Literature on methods of "thoughtreform" or ideological reform has attempted to place these procedures in a context which emphasizes the fact that they are well known and not the result of new discoveries or magical innovations on the part of the Communists (9, 10, 42, 49, 67). In these procedures, solitary confinement and monotonous, barren surroundings play an important role in making the prisoner more receptive and susceptible to the influence of the interrogator. The use of this technique rests not on laboratory science but is part of the empirical know-how of police and military interrogation.
A third major source of interest in these phenomena, although perhaps less dramatic than the foregoing, has come from developments within academic psychology. One such development has taken place in the area of motivation, in which a number of experimenters (14, 34, 58) have attempted to establish the existence and operation of what has been called curiosity or exploratory drive as a primary motive. Attributing a significant role in the determination of behavior to such a drive, we find that this research has arisen in a context which seeks to refute the strongly prevalent view of the organism as a passive receptacle of experience; one which responds only to drive-relevant stimulation. As formulated by Hebb, "Characteristically, stimulus response theory has treated the animal as more or less inactive unless subject to special conditions of arousal." (37, page 244.) In contrast to this approach, studies dealing with the relevance for behavior of a curiosity or exploratory drive indicate that the organism has an active need for experience, and initiates and structures activity in accordance with that need. Studying human response to restricted environments may indicate the mode of operation of the "need for experience."
Another development within academic psychology has contributed to the current concern with the effects of restricted environments on human subjects. Studies of sensory deprivation early in the life of animals, and the effects upon subsequent development and learning, have a relatively long history within psychology. Originally designed to evaluate the relative influence of innate organizational processes (as opposed to learning) on perception, these researches have since been more directly focused on the general effects of early deprivation upon a variety of subsequent behaviors. Although experimental work, because of ethical considerations, has of necessity been confined to animal investigations, clinical and anecdotal evidence such as the reports of Spitz (73, 74, 75) and others (22, 23, 26, 27), and those on "feral man" (70, 71) have supplemented these studies. These reports
have highlighted the importance of a full range of early environmental experience to the development of normal adult functioning. The occurrence of serious and irreversible disruptions of normal development and behavior has been reported. Because this work is beyond the scope of the present chapter, the reader is referred to Beach and Jayne's (5) review of this literature.
Before turning to an examination of the experimental findings, it may be well to consider some of the methodological and conceptual problems raised by research in this area. The diversity of variables involved in a systematic study of response to reduced environmental stimulation makes for considerable complexity. It will be useful to take a brief overview of procedures employed by various investigators.
Three experimental approaches have been identified in the literature (72). In the first of these, efforts were directed toward an absolute reduction of input to the organism from the external world. Lilly (50) immersed two subjects up to three hours in a tank of slowly circulating tepid water, wearing nothing but a head mask that covered eyes and ears. There was minimal visual, auditory, or tactile stimulation. Subjects received an initial set of training exposures to overcome fear of the situation. On the day of the experiment, they were placed in the tank and were instructed to inhibit all movement so far as possible. The use of a soundproof, darkened room (65, 79) as a method for achieving sensory deprivation is similar in intent 'o the foregoing procedure. The subject's arms and hands are enclosed in cardboard cuffs and gloves. Plugs are placed in the ears to reduce further levels of stimulation. Although absolute reduction in sensory input is the goal here, this latter method places less of a restriction on motor activity.
A second approach to reducing sensory stimulation was used by Bexton, Heron and Scott (8). They reduced patterning of sensory inputwhile retaining levels of input at near normal. In this procedure using twenty-two male college students, the subject wore a pair of translucent goggles that permitted the perception of light but not of objects. Auditory input consisted of the masking sound of fan and airconditioner motors, and tactile experience was reduced through the use of cuffs and gloves that permitted no direct exploration of the immediate surroundings.
In a third approach, sensory deprivation consisted of monotony or absence of change in the external environment, e.g., providing a repetitive auditory experience while presenting the subject with an impoverished visual field. In this procedure goggles are not used and the subject is exposed to normally patterned vision of a highly restricted environment. Wexler, Mendelson, Leiderman, and Solomon (80) placed seventeen subjects into polio tank respirators with arms and legs in cardboard cuffs. The repetitive drone of the respirator motor provided an auditory masking sound, whereas the visual environment consisted of the front of the respirator and the blank walls of a screen. Since the ports of the respirator were left open, subjects breathed for themselves. This procedure relies on monotony to achieve its effects and is thus similar to situations in which highly repetitive simple tasks are performed. It is also most similar to the environment of the prisoner in solitary confinement as well as other isolation situations as encountered in real life.
Without attempting a comprehensive survey of methodological problems and issues, some examination of the choices confronting researchers in this problem area may be helpful. What are the limitations and problems of these procedures? Efforts at the absolute reduction of sensory input are limited by the impossibility completely of doing away with sensory experience in a living conscious organism. Even the most sophisticated instrumentation cannot eliminate sensations and perceptions arising from internal body functions. This end point of the stimulation continuum thus must remain theoretical. To the extent to which this goal is relevant to testing a variety of hypotheses, it can only be approximated.
Few if any investigators have attempted a rigorous definition of the terms they have employed. Most have used their experimental methods to provide an empirical basis for their conceptions. Indeed it is understandable that the number of descriptive terms and phrases in the literature is almost as large as the number of investigators. Without becoming too deeply embroiled in the sensation-perception issue, it may be useful to think of attempts at the absolute reduction of intensity of input to the organism as sensory deprivation, whereas reduced patterning and monotony may be more meaningfully seen as perceptual deprivation. The outstanding characteristic of the latter two approaches appears to be the decrease in the structure and variety of input. This inevitably results in a reduction of information. The term "isolation" is one which seems to be relevant to the social dimension rather than to the sensory and perceptual aspects of the various experimental conditions employed. At this stage of
our knowledge, it is unclear as to whether there are different behavioral consequences of sensory as opposed to perceptual deprivation, in the sense used above. It is possible to conceive of this range of stimulus conditions as a complex continuum.
In view of the unique complexities presented by research in this area, it is clear that somewhat arbitrary choices of procedure have been made. These choices must be evaluated in terms of the limitations they impose on the results obtained. Thus the observation of cognitive and perceptual functioning and the descriptions of emotional and affective changes makes simultaneous verbal reports of experience in the experimental situation most desirable. Retrospective reports raise difficult questions about their accuracy and make it impossible to study the concurrence of physiological events and verbal behavior. On the other hand, verbal reports of experiences by the subject during the experiment provide a complex feedback situation. The testing of perceptual and cognitive functions during the experiment constitutes a definite modification of procedure.
One form of control over the subject's sensory experience has been achieved in many experiments by restricting his mobility. This is accomplished by restraining the subject, limiting the space available to him or by instructions to remain still. This limitation of mobility reduces kinesthetic and proprioceptive input. It is difficult to know whether the results obtained are a function of the additional sense of confinement or restriction which goes beyond reduction in sensory stimulation.
Most studies in this field have striven for absolute isolation of the subject from other human contact by avoiding all communication between subject and experimenter. Although social isolation contributes to reduced sensory input, whether this reduction is primarily effective in terms of loss of social contact per se, loss of patterned stimulation from speech, absolute reduction of sensory stimulation, or some combination of these is still to be determined.
Furthermore, the social isolation in these experimental settings is artificial and limited in that the subject knows there is an observer who is interested in his performance. He usually has good reason to suspect that this observer has strong motivation to prevent the occurrence of any long lasting or profoundly debilitating effects. These implicit aspects of the subject-experimenter contract may be major factors in the presumed social isolation seen in experimental studies. These limitations to isolation do not apply to situations such as those of the prisoner or shipwrecked sailor. In the former case there are the additional implications of the status of "enemy" which
undoubtedly also influences the individual's response. The "escape at will" clause present in laboratory studies constitutes a major difference from the motivational conditions of real life isolation situations.
These factors, along with the use of volunteers in experimental studies, constitute serious limitations to the laboratory testing of hypotheses regarding responses to real life isolation and sensory deprivation. We are unable to assess the effects of coercion or the ultimate consequences of prolonged confinement in a deprived environment. These conditions undoubtedly have a profound effect on the motivational aspects of the situation and thus influence response. The inability to replicate these conditions in the laboratory must limit our generalizations from the experimental data. In view of these considerations these data must be interpreted cautiously.
The first experimental work which focused on the response of man exposed to reduced environmental stimulation per se was begun in 1951 in the laboratory of D. O. Hebb at McGill University (7, 8, 28, 38, 39, 40, 41, 69). Although earlier studies had dealt with more limited aspects of this problem, they grew out of an essentially different experimental interest. The McGill studies initially arose out of a concern with the contribution of perceptual isolation to the mechanism of brainwashing and the effects of monotony upon a person with a long sustained watchkeeping task. Previously Mackworth (52) had shown that in a vigilance task requiring prolonged observation, subjects increasingly and strikingly failed to respond to an appropriate stimulus. From this point of departure, the framework of these and other studies was expanded to focus on a wide variety of other variables.
Our approach in reviewing these studies has been influenced by the consideration that in the early stages of acquiring systematic knowledge about a problem, it may be useful to underemphasize considerations of experimental rigor and elegance in favor of developing a richer background of hypotheses and conceptual formulations, even if only at a suggestive level. Because of their exploratory nature, these investigations have often been designed to look for a wide range of possible relationships, rather than to test specific, focused hypotheses. For these reasons this review will not dwell upon limitations of experimental method and procedure. In general, the studies are uneven in quality, and range from carefully designed and
executed procedures to vaguely formulated, poorly controlled observations with small samples. Similarly, measurement in these studies has varied from precise psychophysical calibration to loosely defined clinical judgments unchecked for reliability. The effort has been to provide a comprehensive review of all pertinent studies for whatever light they shed on the problem or support they lend other studies. In reviewing this work we have largely restricted our concern to the psychophysiological aspects of experimental work with isolation and reduced sensory input. No attempt has been made to include the social-psychological aspects of isolation which, while relevant, represent a special subproblem.
For purposes of clarity we shall report the findings in the following categories: perceptual and motor abilities; cognitive and learning abilities; personality findings; feeling states; imagery; and physiology. In addition, we shall consider findings bearing on methodological choices, clinical applications, and a brief survey of theoretical interpretations. Despite some arbitrariness in these classifications and the necessity of considering the same experimental work in several categories, this approach will permit a more coherent view of the evidence within a given experimental domain. In referring to these studies, reduced patterning, imposed structuring, and homogenous stimulation are referred to as perceptual deprivation; absolute reduction in variety and intensity of sensory input will be called sensory deprivation. In a number of experimental procedures elements of both are present.
The problems of vigilance under conditions of perceptual deprivation have been studied by Mackworth (52). Additional literature in this area was reviewed by Holland (44), who summarized these studies as showing a greater over-all percentage of detection when the number of signals per experimental session increases, and a more equivocal finding of an increased probability of detection for longer intersignal times. He interpreted vigilance behavior as a problem of reinforcement scheduling and probability of response. In this context, signal detections serve as reinforcements for observing responses. His own findings confirmed the earlier reports that within a given session, despite individual differences, the use of a larger number of signals increased response rate. He calls attention to the "rather precise control exerted by the environment over the human operator's observing behavior" (page 67).
Hochberg, Triebel, and Seaman (43), working with a different experimental interest, performed a series of studies on the percepts obtained under conditions of spatially homogeneous colored illumination over the entire visual field. These studies were designed to test the hypothesis that a colored Ganzfeld would lose its color under these conditions. Utilizing eyecaps made from halved table-tennis balls, these investigators found that complete disappearance of color was obtained in most cases, despite considerable individual differences in the course of the adaptation process and in the phenomenal content during adaptation.
Similarly, restricting stimulus input to homogeneous visual stimulation, W. Cohen and Cadwallader (20) studied the effects of uniform visual stimulation utilizing a different apparatus. Subjects' exposure to these conditions lasted from three to ten minutes while sitting alone in a room. The findings showed that under both monocular and binocular conditions, subjects reported a temporary cessation of ordinary visual experience after prolonged exposure to a uniform visual field. With increased exposure to these conditions the initial reports of the field as being "foglike" changed to an experience of "blanking out." The "blanking out" or "white-out" phenomenon often persisted for thirty seconds or more and the reappearance of the field occurred only after extensive eye movements and blinking, or with the introduction of an object to the field. Factors that facilitated "white-out" were found to include both extensive prior stimulation and scotopic (rather than photopic) stimulation. A similar finding is reported by Ditchburn, cited by Bruner (12), who showed that if a visual pattern is stabilized on the retina so that it is not even displaced by the natural tremor of the eye, it disappears from view within about six seconds.
Bexton et al. (8), utilizing the procedure of reduced patterning of stimulus input, had twenty-two paid male volunteer college students serve as subjects and compared their performance with an equivalent control group. They were told to lie on a comfortable bed in a lighted cubicle, and they wore translucent goggles, cuffs, and gloves. Auditory input was reduced through the use of a partially soundproof cubicle, a U-shaped foam rubber pillow for the subject's head, and the masking hum of fan and air-conditioner motors, which was fed into earphones in the pillow. Upon leaving, after two or three days in the experimental situation, subjects had difficulty in focusing; objects appeared fuzzy and did not stand out from their backgrounds; the environment seemed two-dimensional; and colors appeared to be more saturated than usual. The experimenters also found deteriora-
tion in visual motor coordination as measured by such tasks as the Wechsler Digit Symbol test, handwriting specimens, and the copying of prose paragraphs. Another study by the same group (69) showed that performance on the Thurstone-Gottschaldt Embedded Figures test declined, whereas no change was manifest in a mirror tracing task. The deterioration of performance on the digit symbol test has since been confirmed by Davis, McCourt, and Solomon (21), who studied ten paid volunteer subjects under different experimental conditions of perceptual deprivation. These investigators failed to find deterioration in the Witkin Embedded Figures test. Vernon and Hoffman (76), after conditions of sensory deprivation lasting twenty-four and forty-eight hours, questioned four subjects about difficulty in focusing, increased saturation of hues, and lack of three-dimensional perception, and reported negative findings for all three phenomena.
Heron, Doane, and Scott (41) extended the duration of their experimental procedure to six days and served as their own subjects. They described the disturbances in visual perception as unexpectedly profound and prolonged, with similar manifestations for all three participants. These effects included apparent movement phenomena (with and without head and eye movements by the observers), distortions of shape, accentuation of afterimages, perceptual lag, and increases in color saturation and contrast. Further work from the same laboratory (28) described the fluctuating curvature of surfaces and lines, and disturbances in size constancy. In addition, these investigators observed that autokinetic effects were harder to abolish, larger figural aftereffects were obtained, and spiral aftereffects were more persistent. A loss of accuracy in tactual perception and spatial orientation was noted.
Freedman, Grunebaum, and Greenblatt (30) studied the effects of isolation and reduced patterning of visual and auditory input upon visual perception. As controls they employed paid male volunteers, who received only social isolation. To induce perceptual deprivation, these investigators used translucent goggles, white noise fed into earphones, cotton gloves, and cardboard cuffs for the subjects' arms. Each of the eight experimental subjects was placed on a bed in a lighted room and was instructed not to move about. The control group of six subjects was similarly treated without the additional restrictions to visual, auditory, and tactile input. Both groups remained in the situation for eight hours and had no contact with the experimenters during this time. They found changes in perception similar to those cited by the McGill group.
in every experimental subject, but none in the control subjects. In some subjects these aberrations persisted for over one hour, and consisted of two-dimensional forms changing shape and size and of straight lines moving and curving. Comparing pre- and postisolation performance, they observed a decrement in size constancy and changes in the Müller-Lyer illusion. In both instances, changes consisted of increased variability of judgment rather than unidirectional effects. Visual-motor coordination, as seen in the copying of Bender-Gestalt figures, was significantly impaired following exposure to the experimental conditions.
An increase in apparent movement phenomena through perceptual deprivation has been demonstrated in a study designed specifically to test this relationship. Ormiston (59) compared thirty minutes of perceptual deprivation, sensory bombardment, and a neutral condition for their effects on the perception of the phi phenomenon with thirty subjects serving in each condition. The deprivation condition was realized through having subjects sit in a bare room wearing translucent goggles, ear plugs, and ear muffs. The sensory bombardment condition exposed subjects to motor tasks, a tape with varied sound effects, taste and smell stimuli, and a variety of colored goggles. The neutral condition consisted of having subjects sit on a couch in a waiting room. A comparison of pre- and posttests showed a statistically significant increase in the perception of phi for the deprived group, whereas the bombardment group showed a trend toward decrease in phi perception. The neutral group showed no change.
Vernon, McGill, Gulick, and Candland (78) studied the effects of sensory deprivation upon a variety of perceptual and motor skills. Eighteen paid volunteer subjects were placed in a small, dark, lightproof, soundproof chamber containing a bed, an icebox with food, and toilet facilities. Subjects wore gauntlet-type gloves to reduce tactile stimulation and inhibit movement as well as the noise of movement. They were instructed to lie quietly and make as little noise as possible. A control group which did not receive sensory deprivation consisted of a similarly motivated group of graduate students. The experimental subjects remained in confinement for one, two, or three days, at the end of which they were required to perform a variety of tasks. The effects of sensory deprivation were assessed by a comparison of differences in pre- and postconfinement scores with those of the control group who were tested at similar intervals. The findings revealed significant deterioration in visual-motor coordination as seen in a rotary pursuit task, a rail-walking task, a mirror tracing problem, and mazes. In perceptual tasks, such as color
perception and delayed auditory feedback, a similarly significant decline in performance was observed. The only task of this series which did not show a decline was a test of depth perception, in which a trend was obscured by the large variability of scores. It should be noted that the mirror-tracing finding in this study contradicts that reported by Scott et al. (69).
Utilizing the shortest periods of exposure to reduced sensory input, Rosenbaum, Dobie, and Cohen (64) studied the effects of 0, 5, 15, and 30 minutes of two conditions of visual deprivation upon tachistoscopic recognition thresholds for numbers. For one group of sixteen subjects, visual deprivation was achieved by blacked-out rubber goggles, while a similar second group received perceptual deprivation through the use of goggles permitting the perception of diffuse light. These investigators found no differences between their two groups; both improved with successive days of practice. The five-minute condition resulted in a lowered threshold for both groups. The thirty-minute condition produced no change from the zero level. Since none of the experiments using longer periods of deprivation measured recognition thresholds, it is difficult to say whether failure to observe changes in this task was a function of an insufficient period of deprivation or whether no relation is to be expected.
In summary, the findings of these studies indicate a generally disorganizing effect of deprivation upon perception. The effects thus far demonstrated have been confined largely to the visual modality. These effects include the following: breakdown in visual-motor coordination, an increase in apparent movement phenomena, increase in color saturation, decline in size and shape constancies, loss of accuracy in tactual perception and spatial orientation, increase in persistence of autokinetic effect, larger figural aftereffects, difficulty in focusing, fluctuating curvature of lines and surfaces, and a general decrease in the efficiency of perceiving relevant stimuli. Although several studies are in disagreement about some of the above details [e.g., Vernon and Hoffman (76)], most reports are remarkably convergent in their findings despite wide differences in experimental conditions. These effects may be best characterized as a general loosening of subjects' ability to perceive reality and the weakening of stable internal norms against which to evaluate perceptual (visual) experieuce. The increase in variability of a number of visual functions and loss of accuracy may be best understood in these terms. The breakdown of internal norms is demonstrated in a variety of other functions and begins to suggest one general parameter which may make isolation and sensory deprivation effective in increasing
A wide variety of studies have referred to subjective reports of difficulty in concentration, attention, and problem solving following isolation and confinement (8, 16, 17, 30, 65, 80). These and other studies have also examined the effects of isolation and deprivation upon a wide range of cognitive functions. Included have been such abilities as those involved in a variety of intelligence test performances, learning and association tasks, logical reasoning, etc. We turn here to a consideration of these aspects of the experimental literature.
The report of Bexton et al. (8) is elaborated in a subsequent study (69). These researchers investigated cognitive performance during isolation and perceptual deprivation. In order to evaluate the duration of the effects, they examined several other functions following four days of isolation. On several occasions during isolation, they had subjects perform tasks such as mental multiplication, arithmetic catch problems, completing number series, anagrams, and wordmaking. Despite the fact that the decline in the twenty-two subjects of the experimental group was not statistically significant for all these tasks, the deterioration due to the experimental conditions was consistent. In a second series they found no change in digit span or analogies during isolation, whereas associative learning tended to decline, but not significantly. In a postisolation series they found significant deterioration in judgment of anomalies and in two block design tasks. Their general findings suggest that performance on intelligence test items grew progressively worse as length of stay in the cubicle increased.
Starting with this observation, Vernon and Hoffman (76) used a procedure of sensory deprivation similar to that described above. They studied the ability of four paid volunteer male college students to learn lists of adjectives after twenty-four and forty-eight hours of confinement. Comparing their experimental subjects to an equivalent control group, they found that the ability at rate-learning improved with continued sensory deprivation. In a follow-up study, nine experimental and nine control subjects, who were all paid volunteer male college students, were compared for ability to learn a longer list of adjectives after twenty-four, forty-eight, and seventy-two hours of sensory deprivation (77). In this instance there were no significant differences between groups in errors or trials to criterion, although
the experimental group made fewer overt errors and showed less variability. Thus, despite failure to confirm their own previous findings, this study did not support the deterioration finding of the McGill group.
Goldberger and Holt (32) studied fourteen paid volunteer male college students under perceptual deprivation conditions similar to those of the McGill experiments. They utilized the halved tabletennis ball procedure of Hochberg et al. (43) to occlude vision and white noise to mask other auditory input. Subjects lay on a bed in a cubicie for eight hours and were encouraged to talk during their time in isolation. The following tests were administered at the end under the experimental conditions: arithmetic reasoning, digit span, and story recall. Subjects were then taken out of the isolation and a test of logical deductions was given. Comparison of the performance of the experimental subjects pre- and postconfinement (without a control group) showed that only the last of these, logical deductions, reflect significant impairment.
Davis, McCourt, and Solomon (21) utilizing a modification of the polio tank-respirator procedure initially described by Wexler et al. (80) kept five pairs of paid volunteer male college student subjects in relative perceptual deprivation for over ten hours. These subjects did not know one another. Although they could talk to each other, they were confined separately and could not see each other. In comparing scores before and after isolation they found no change in performance on a block design task. These authors considered the possibility of procedural variables causing failure to confirm Bexton et al. (8) in finding impairment in block design performance.
B. Cohen, Rosenbaum, et al. (17) studied four normal subjects and six patients from various clinical groups under conditions of brief deprivation. Subjects were seated individually for one hour in an isolation chamber in a comfortable chair. They wore goggles which were either blacked out or else permitted diffuse light perception. Audition was minimized through car plugs, padded earphones, and the masking sound of a fan motor. Their fingers were wrapped in elastic bandages and they wore elbow-length gloves. The instructions were to relax and move as little as possible. Subjects were also told that they would perceive sensations ordinarily below conscious awareness. These experimenters report that there was no "gross cognitive deterioration" under these conditions as measured by the number of word associations produced in two minutes. The small sample size, the brief period of isolation, and the limited measure employed in this study suggests caution in interpreting this result.
S. Cohen, Silverman, Bressler, and Shmavonian (18) reported an exploratory investigation on four subjects exposed singly to four hours of confinement and deprivation while seated in an anechoic chamber, with instructions to keep awake and to estimate the passage of successive thirty-minute intervals. All four subjects showed an increase in performance on digit span, and decrease in arithmetic reasoning, abstraction, and general reasoning. All subjects reported difficulty in problem solving and logical thinking. The small sample size and absence of a control group limit the relevance of these findings.
The few reports available, their currently sketchy detail, and their limited controls make it difficult to arrive at a firm generalization concerning the effects of deprivation and isolation on cognitive skills. Some tentative agreements among the studies may be worth highlighting. It appears that the skill most severely impaired under these conditions is that of general reasoning and problem solving, whether the situation involves verbal-conceptual materials or numbers. On the other hand, in several studies performance on simple recall tasks or rote learning seems either to improve or else does not decline. Tasks that involve analysis and synthesis of visual materials such as block design show equivocal results; in some studies there is deterioration, in others no change is seen. Some of these equivocal results may be a function of differences in procedure or duration of deprivation and confinement. The sequence in which postisolation tests are administered may be a relevant variable here since the duration of the effects, if any, is as yet unknown.
The conceptual analysis of cognitive skills into categories such as reasoning, memory, arithmetic, and manipulation would serve a very useful purpose in these studies. Goldberger and Holt (32) offer a tentative generalization which begins to specify some of the different factors involved in various tasks. They point out that, "Probably any task that can be done satisfactorily in a single brief effort by the use of highly overlearned sets of operations (as in simple arithmetic problems), and any learning or memory performance requiring passive receptivity (cf., digit span, rote learning) rather than reflection and manipulation of ideas (cf., logical reasoning) would be least interfered with by moderate amounts of isolation." (32, page 109.) This formulation appears to account for some of the reported results, but the criteria employed to classify cognitive tasks need further specification and expansion. For example, are there differences in the types of memory or recall involved in digit span as opposed to remote memory? Although the available results are certainly inconclusive, the Gold-
berger and Holt generalization suggests that remote recall would be relatively unimpaired by isolation. This would have implications for one interested in the interrogation problem where, leaving motivational issues aside, it would seem that the information a person might recall when cooperating would be reliable. Furthermore the data suggest a decreased interest in and ability to reason through the complexities of the interrogator-prisoner relationship and thus a decreased ability to cope rationally and effectively with the situation at hand.
Despite the more tenuous nature of the findings in the area of cognitive skills, to the extent to which a generalization is possible, it seems that in addition to the decline in internal norms or standards for perceiving reality, under conditions of deprivation and isolation there is a lessened ability to reason closely and solve complex problems. Should such a generalization be supported by subsequent findings, a relevant question remains about the extent of such impairment in quantitative terms.
An issue related to that of cognitive functioning in isolation and deprivation is that of suggestibility. The reported success of isolated confinement in modifying beliefs and convictions initially directed interest to the question of suggestibility. Despite this concern, systematic data on this problem remain surprisingly sparse.
Heron (39) cited the work of Bexton (7) pertinent to this issue. After twenty hours of perceptual deprivation, twenty-four subjects were exposed to a recorded propaganda message consisting of a ninetyminute talk read in a boring monotone and arguing for belief in various psychical phenomena. To measure attitude change, a series of attitude scales of the Bogardus type were administered before and after confinement. A control group of twenty-seven subjects received both scales before and after a similar interval. Both groups, experimental and control, showed a significant change in attitude after listening to the records. The change, however, was significantly greater for the experimental subjects. Similarly, measures of interest in the topic and assessments of its importance showed a greater increase for the subjects exposed to perceptual deprivation. Although follow-up data were not systematically obtained, incidental evidence indicated that for some subjects at least, these effects persisted for as long as three to four days. Vernon and Hoffman (76) tested subjects for degree of body sway suggestibility following varying periods of
The question of suggestibility has also been approached from a different vantage point. Rather than estimate the effects of isolation upon suggestibility, others have sought to establish a relationship between suggestibility as a personality attribute and response to deprivation. Petrie 1 measured body sway suggestibility in a group of nine subjects who had been exposed to the conditions of confinement and perceptual deprivation described by Wexler et al. (80). Her observations revealed a trend toward a negative relationship between amount of body sway and length of time voluntarily spent in deprivation.
Camberari (16) studied the response to sensory deprivation of twenty male unpaid volunteer psychology graduate students evenly divided into suggestible and nonsuggestible groups. This division was based on the composite scores of these subjects on several tests of suggestibility. Isolation and sensory deprivation were brought about by suspending subjects nude in a tank of water by means of a harness. His over-all findings appear to contradict that of Petrie in that the suggestible group remained in the situation significantly longer (183 minutes) than the nonsuggestible group (111 minutes). Upon closer examination, however, one finds that one of the ten measures in the battery was body sway suggestibility and that for this particular measure there was a negative relationship with length of stay similar to Petrie's.
The Camberari data pose the difficulty of interpreting the notion of suggestibility. The meaning of the term is difficult to assess independent of the operations defining it and the experimental consequences. As such it has limited utility for assessing the effects of deprivation and isolation upon the readiness to modify one's own conviction or belief in favor of those coming from an external authority figure. Once again, leaving aside the complex motivational issues which limit generalization of laboratory studies to real life situations, we are left with the Heron findings that following isolation and confinement, beliefs around a topic such as psychical phenomena change significantly. This observation is also consistent with the hypothesis of a decline in internal perceptual norms and in ability to reason efficiently. It would seem likely that changing the emotional relationship between the authority and the subject would introduce another complex variable which cannot be assessed without data. The tendency to modification of belief in experimental circumstances is
1 Personal communication, 1958.
quite consistent with the reports of the response of prisoners. It should be kept in mind that in the latter situation additional elements of uncertainty, stress, and coercion were brought to bear in inducing these changes (10, 42).
The relationship of personality attributes to tolerance for isolation is one which has significant implications for issues as diverse as personnel selection and personality theory. That the study of response to this situation might be relevant to the study of personality was pointed out: by Hochberg et al. (43).
We referred previously to the work of Camberari (16). In addition to finding differences between suggestible and nonsuggestible subjects in tolerance for sensory deprivation, he observed a number of related personality attributes that seemed to differentiate the two groups. The suggestible subjects appeared to be more productive and more tolerant of regressive behavior, including delusions, hallucinations, and fantasies. The nonsuggesiible subjects, on the other hand, tended to be more threatened by disturbances in body schema, defensive about their intellectual control, and more aware of external factors which reinforce reality.
Wexler et al. (80) studied seventeen paid male volunteer subjects exposed to confinement and perceptual deprivation in a polio respirator up to thirty-six hours. They used a series of personality measures including the Minnesota Multiphasic Personality Inventory and the Edwards Personal Preference Schedule. None of the MMPI scales was related to duration of stay in isolation. For the Edwards test these authors reported a significant negative relationship between need Exhibitionism and length of stay in isolation, as well as near significant positive relationships between the latter variable and need Affiliation, need Succorance, and need Nurturance. They interpreted these findings to mean that subjects with greater tolerance for deprivation relate themselves more genuinely to people and seek more contact and emotional exchange with others. In a second experiment (47), with eleven subjects, under more severe conditions of isolation, these
Unlike Wexler and his associates, S. Cohen et al. (18), who studied four volunteer subjects seated in an anechoic chamber for four hours, found that the two subjects described as "schizoid personalities" on the basis of clinical interviews and psychological tests were comfortable in isolation and were willing to prolong it, whereas the two "fairly well integrated" subjects, who described the experience as unpleasant, were anxious, and felt that they could not tolerate the isolation much longer.
In studying ten subjects consisting of normal, neurotic, schizophrenic, and sociopathic individuals for response to one hour of isolation, B. Cohen et al. (17) reported that normal and neurotic subjects exhibited an increased sensitivity to the residual stimuli in the chamber. The schizophrenic subjects showed no appreciable increase or decrease in their hallucinatory behavior and had a generally positive reaction to the situation, devoid of the anxiety typically exhibited by normal subjects.
A more recent study of the response of schizophrenics to sensory deprivation was performed by Harris (35). Utilizing a procedure similar to that of the McGill group, he placed twelve subjects in isolation for periods up to two hours. He reports that the patients generally tolerated the procedure well. For the most part, hallucinations became less intense and less vivid. Over-all symptomatology either improved or showed no change. These findings appear to be consistent with those of the two earlier studies cited previously.
Working in a different theoretical context, Petrie, Collins, and Solomon (60) attempted to relate pain tolerance, cortical satiation, and perceptual deprivation. Using kinesthetic figural aftereffects, measures of pain threshold, and tolerance for isolation in the polio respirator, their findings tend to support the hypothesis that susceptibility to satiation is associated with tolerance for pain and intolerance for perceptual deprivation. Here satiation is seen as a key factor mediating the perceived intensity of stimulation; the higher the satiability the less intense are succeeding sensations.
Still another approach was taken by Goldberger and Holt (33) in their experiment. They emphasized psychoanalytic concepts, such as resistance to regression and modes of handling primary process material. Fourteen subjects were rated for the maturity with which they handled primary process as manifested in Rorschach test responses. Their verbal behavior during eight hours of isolation and
Two relatively independent reaction patterns to isolation were identified. In the first of these, subjects engaged in a variety of behaviors within the limits set by the situation and instructions. They talked freely, experienced pleasurable affect, little unpleasant affect, thought rationally, and engaged in daydreams, fantasy, and playful thinking without being threatened by the situation. In the second reaction pattern, there was unpleasant affect, anxiety-laden intrusions of the primary process, preoccupation with terminating the experiment, and impaired efficiency in rational or secondary process thinking. They found these two reaction patterns to be significantly correlated in the expected direction with the Rorschach measure of maturity of handling primary process materials. Those who on the Rorschach handled primary process in a mature and effective way were those who reacted in an adaptive way to isolation. Conversely, those who on the Rorschach handled primary process with poor control or avoided it reacted negatively to isolation. This finding is consistent with several others which point to the exaggeration of usual personality defenses under the stress of isolation (18, 56, 65). From this point of view it should be possible, at least theoretically, to predict the dimensions of an individual's response to deprivation and isolation.
An overview of these data emphasizes the truism of marked individuality of response. Whether differences observed among various studies is a systematic function of varying experimental conditions is as yet unclear. Whereas the findings of Wexler et al. (80) and Goldberger and Holt (33) indicate a positive relationship between emotional relatedness and length of stay in isolation, several others have made a different observation. The findings on suggestibility as a personality attribute and those on the relationship to satiation and pain thresholds remain conceptually unrelated to the other work. The Goldberger and Holt demonstration of relationships between preisolation personality attributes and the content of response to isolation is a carefully executed study which has a clear theoretical orientation and makes complex but reliable assessments of verbal and other behavior. Other studies have tended toward utilization of too simplified an index of response such as length of stay which fails to take into account complex behavior during the isolation situation. It may well be that personality variables and their interrelationships are insufficiently reflected in such a simple measure of tolerance for isolation.
In addition it would seem desirable that workers in this area offer a conceptual framework within which to view personality response. Thus specification of terms such as "schizoid" and "withdrawn" may have more meaning, permit replication of procedures, and evaluation of results. Although some of this difficulty in the present studies stems from their preliminary nature, there appears to be some insensitivity to the need for both conceptual and operational specification of measurement and assessment techniques. Progress with the problem of personnel selection and utilization for a variety of tasks, as well as theoretical clarification, awaits such refinement in research programs.
Changes in subjective feeling in response to reduced environmental input has been a common observation in these studies. These observations range over many different experimental conditions, from the uniform visual stimulation presented by W. Cohen (19) in which the pattern of input to one modality is reduced, to sensory deprivation in the water tank suspension procedure used by Lilly (50) and Cambareri (16), where the effort is made at a total reduction of sensory input. We have already mentioned the breakdown in the subjects' ability to concentrate, think clearly, and solve mental problems. Bexton et al. (8) emphasized the progressive increase in irritability during confinement, followed on release by a sense of being dazed and confused. They also noted the presence of headaches, fatigue, and mild nausea, persisting in some cases for twenty-four hours after confinement.
Lilly (50) whose subjects both had a number of trial exposures to the situation in order to get used to it found in the actual situation early feelings of relaxation and enjoyment, followed by tension, restlessness, and an extremely heightened awareness of residual stimulation. This course continued into fantasy and reverie, and finally into the projection of visual imagery. Following the isolation experience, subjects reported a sense of refreshment as though having just awakened from sleep. Camberari (16), on the other hand, utilizing a similar procedure without preliminary exposures, found no such progressive stages, and subjects came out of the immersion feeling fatigued rather than rested. Suggestible subjects felt secure during most of their stay in the tank, although there were some reports of apprehension, fear, and panic. The nonsuggestible subjects generally tended to deny any affective or emotional involvement.
After prolonged visual stimulation, W. Cohen (19) found that his subjects had feelings of drowsiness, excessive yawning, and their voices took on a hesitant, drawling quality. In a study of perceptual deprivation, Hebb, Heath, and Stuart (38) reported that subjects who wore earplugs for three consecutive days while going about their normal activities showed slight to marked irritability, seclusiveness, and personality disturbances not due to discomfort. One subject manifested poor speech coordination while in the experimental conditions. Goldberger and Holt (32) noted that despite the individual differences in response, all of their subjects found the experimental situation frustrating because of lack of things to do, see, and hear, and the physical discomfort attributed to motor restriction. Comparing sensory and perceptual deprivation with sensory bombardment, Ormiston (59) found that the former group reacted with unpleasant affect, whereas the latter group experienced mixed affect of amusement and anxiety.
Davis et al. (21) required their ten subjects to complete an adjective check list describing their feelings following isolation. This was compared with the same check list completed under control conditions. Significantly more somatic complaints, feelings of physical inactivity, and mental clouding were checked following isolation, whereas significantly less satisfaction, elation, friendliness, and impulsivity were reported.
Goldberger and Holt (32) also reported marked variations in postisolation feeling states. Seven subjects described such reactions as feeling dazed, disorganized, groggy, dizzy, and unstable; eight subjects indicated a state of fatigue; four subjects described motivational changes such as losing interest in things; two subjects showed virtually no disttirbance.
In part these responses may be interpreted as a reaction to the novelty and threat of a strange and unfamiliar situation. Undoubtedly these subjective states also reflect the earlier discussed individuality of personality response to these experimental settings. The general response to isolation seems to include boredom, a general state of restlessness related to inactivity, and often anxiety or fear of extreme proportions. Postisolation responses most often seem to reflect fatigue, drowsiness, confusion, loss of time orientation, and a need to reorient one's self to the familiar aspects of reality. These subjective states appear to be consistent correlates of the changes in
perception, cognitive function, and personality function previously described. These feeling states would seem to make one quite vulnerable to new input from a controlled source such as might appear in the prisoner-interrogator relationship. To the extent to which an interrogator becomes associated with the reward of anxiety reduction, providing human contact and directed activity, and thus providing relief for the cumulative discomfort of isolation, he assumes a benevolent role which may be one source of his influence. [See Biderman (9) and Lifton (49).]
Lacking a more adequate descriptive term, we shall in this section use the term "imagery" in dealing with a wide variety of phenomena seen in studies of isolation and deprivation, including what have been called hallucinations, delusions, illusions, fantasies, daydreams, dreams, hypnagogic states, and the like. As pointed out earlier, these phenomena constitute perhaps the most dramatic aspect of this research, and indeed the promise of studying their genesis has undoubtedly been most responsible for exciting the interest of clinically oriented groups. Apart from the autobiographical accounts of isolation, normal persons have rarely reported feelings of depersonalization, breakdowns in body image, and hallucinatory phenomena. Furthermore, the ability to produce such states experimentally brings these concepts much closer to understanding by permitting close observation and perhaps control. With this introductory remark, let us turn to an examination of the pertinent data.
We have already referred to the observation of Hochberg et al. (43) about individual differences in hallucinatory objects. Three subjects in one of his studies reported the appearance of such hallucinatory objects during adaptation, and it was later difficult to convince them that such shapes had not been included as part of the experimental procedure. Bexton et al. (8) reported that these phenomena were largely visual and ranged from simple geometrical forms and patterns, to simple colors, to complexly integrated scenes which were sometimes in color and three-dimensional. The latter often contained dreamlike images. Subjects were able to exercise only minimal conscious control over the content. The images often involved other senses including auditory, kinesthetic, and somesthetic elements. There were also reports of bodily strangeness and peculiar perceptions of body image. Usually these experiences disappeared when the subject began a complex task, such as mental arithmetic.
Time of onset for these phenomena varied from twenty minutes to seventy hours. In the supplementary report (39) it is noted that subjects did not believe in the reality of these experiences and that compared to normal imagery greater vividness is the main difference. Often there was considerable movement of the visual patterns, at times sufficiently disturbing to cause subjects to experience nausea.
To test the notion that residual stimulation was necessary for the production of these phenomena, these same investigators placed opaque goggles on three subjects after several days in isolation with translucent goggles. These subjects had been 'hallucinating' persistently. With this change, they observed an initial increase in vividness, but in one subject this was soon followed by a very marked diminution, and by the total disappearance of these experiences in the other two. With restoration of translucent goggles, the 'hallucinations' reappeared. Two other subjects, run from the beginning with opaque goggles and then shifted to translucent ones, showed an increase in the incidence and vividness of these visual phenomena.
Under the conditions of isolation utilized in Vernon's series of experiments, despite some differences in results, a similar implication emerges (79). Two conditions of sensory deprivation were compared. In one, nine subjects were blindfolded when they periodically had to leave the cubicle. As a result of this procedure they were exposed to a variety of visual stimulation from light leaks, etc. In a second phase, conditions of deprivation were more extreme; the eleven subjects did not leave the lightproof cubicle during their stay. Contrasting the visual imagery reported under these two sets of circumstances, the authors found that the less extreme condition produced many more 'hallucinatory' phenomena. In addition there seemed to be a positive relationship between length of confinement and number of images. The content of the imagery tended toward simple flashes of lights or geometric shapes, rather than meaningful, symbolic, integrated scenes. These authors concluded that confinement permitting the greatest amount of nonpatterned visual stimulation produces the greatest., amount and variety of images. A similarly limited amount of imagery is reported by Ruff et al. (65) who, using both volunteers and nonvolunteer military personnel as subjects observed 'hallucinations' in only two subjects out of more than sixty, run under a variety of experimental conditions of isolation.
Mendelson and Foley (54) observed that a number of polio patients treated in tank-type respirators developed psychoticlike symptoms of disorientation, confusion, 'hallucinations,' and delusions. Physiologi-
cal, psychiatric, and behavioral studies of these patients suggested that these symptoms were not a function of toxic or metabolic factors, but were the result of conditions of life in the respirator. In a follow-up report discussing these phenomena, Mendelson, Solomon, and Lindemann (57) summarized the elements present in these terms: denial of distressing reality; wish-fulfillment in terms of reassuring, pleasurable life situations; rehearsallike anticipation of dreaded possibilities. The authors describe the 'hallucinations' as representing a restitutional attempt on the part of the ego to predict and cope with anticipated stress.
Starting with these observations, Wexler et al. (80) placed seventeen normal adult male volunteers in the tank-type respirator up to thirty-six hours in the perceptual deprivation procedure previously described. These investigators described the incidence of a range of mental experiences in their subjects. These experiences included the occurrence of 'analogies,' 'daydreams,' 'fantasies,' 'pseudosomatic delusions,' 'illusions,' and 'hallucinations.' These phenomena were not related to length of stay in confinement. Under similar, but more severe, conditions of confinement these investigators reported a comparable total incidence of these phenomena, with a higher frequency of occurrence in the categories reflecting more marked deviations from normal imagery (47). This observation held true despite the fact that subjects remained in the experimental situation for a much shorter period.
S. Cohen et al. (18) observed male and female volunteer subjects seated in isolation for two hours in a dark acoustical chamber with no orienting instructions, the only sound being a low hum from a ventilator motor. In this exploratory study the period of deprivation and isolation lasted two hours. Seven of the ten subjects reported unusual visual phenomena ranging from flashing lights to moving objects. The authors suggest classification of the experience on the basis of criteria such as vividness, recognition of subjectivity of experience, conscious control, and emotional accompaniments. Another interesting aspect of this report is the description of changes in the texture and consistency of the chamber walls and floor. Descriptions of an actual metal wall included such adjectives as, "soft," "ruglike," and "spongy, velvetlike." The lack of orienting instructions in this experiment introduces an element of anxiety which may be a factor in the images reported.
Goldberger and Holt (32) also found that subjects in eight hours of isolation reported the spontaneous occurrence of visual and auditory imagery different from that perceived in the normal waking state. These were usually recognized as internal in origin, although several subjects perceived them as being external. These authors feel that these phenomena are most like hypnagogic imagery. In addition they found subjects with body image disturbance, depersonalization, and dreams dealing with the experimental situation. In discussing these data, they argue against the use of the term "hallucination" in that it requires, in addition to an illusory image, a failure to recognize its unreality. On the other hand, it is known that alcoholic patients in delirium tremens, or patients under mescaline intoxication may often recognize the unreality of their percepts and hallucinations. The definition of terms here is complicated by confusion and inconsistency in language usage. Goldberger and Holt suggest consideration of these phenomena in terms of vividness, structure, persistence, realism, and plausibility. From this standpoint they offer the generalization that perceptual deprivation during the period of experimental confinement tends to increase the vividness and structure of imagery without a breakdown in reality testing.
In summary, there are now several studies which point to the importance of some extraneous stimulation for the occurrence of visual and auditory images and hallucinations (39, 65, 79). The role of body movement as a factor in the causation of these phenomena has been cited (30, 32, 65). Freedman et al. (30) attribute the occurrence or nonoccurrence of imagery, at least in part, to differences in motility and kinesthetic feedback. Although this factor is partially confouuded with variations in visual input, they point out that in only two procedures, those of Ruff et al. (65) and Vernon et al. (79), was there free movement and these two report the lowest incidence of imagery. Freedman et al. (30) emphasize the role of motor activity in the general maintenance of spatial and cognitive orientation.
Direct quantitative comparisons of various studies in this area will require agreement about the descriptive parameters of the experiences here subsumed under the rubric of imagery. These dimensions should make possible a more precise evaluation of different experimental conditions for their relationship to the production of these experiences. Similarly it would make possible the detailed comparison of imagery in sensory and perceptual deprivation conditions to
Another problem in the quantitative assessment of the imagery phenomena lies in the fact that most of the data has been obtained by retrospective report. This procedure raises difficulties in the reliable assessment of vividness, frequency, or other suggested dimensions of analysis. It may be that encouraging verbal report during the procedure and making simultaneous recordings of other variables such as physiological responses will make possible independent assessment of these experiences.
Without such independent assessment it would be premature to consider the imagery experienced in deprivation experiments as necessarily indicative of pathology as a number of reports have implied. Such experiences may indeed reflect creative adaptations to a special environment. The evidence that artists, mystics, and religious individuals may utilize isolation for constructive syntheses of experience cannot be ignored. It would seem that the response to the increased awareness of primary process material may be more predictive of adaptation than would the fact of the increase itself or its content. The imagery phenomena may turn out to sustain some individuals confronted by real life situations of isolation and deprivation. On the other hand, the anxiety engendered by the rise of primary process material may serve to increase an individual's susceptibility to external pressures as in the interrogation situation. In general, a broader conceptual framework than that provided by the model of psychopathology would seem to be needed.
Apart from an intrinsic interest in the consequences of reduced stimulation for physiological functioning, such data can also be compared to concurrent verbal reports of the experience. One such index of response thus far studied has been the electroencephalogram. Attention has focused on this instrument in hopes of clarifying the nature of cortical activity in sensory and perceptual deprivation. In addition it has been used to make assessments of the sleep-wakefulness cycle under these conditions.
Heron (39) described the results of periodic EEG tracings on six experimental subjects. The results showed that slower frequencies appear in the parieto-occipital tracings taken at sixteen hours than those taken at the beginning of the isolation period; even after subjects had emerged from isolation for some hours, the records had not returned to their normal state. This finding was confirmed in a
quantitative analysis of wave frequencies within a given time interval. Furthermore, they found that records obtained while a subject was 'hallucinating' showed greatly reduced amplitude and appeared similar to those which might be obtained from a subject in an alerted state. Though these investigators were unable to make accurate determinations of sleep-wakefulness patterns, it was their general impression that subjects slept more during the early part of their stay in isolation and progressively less later in the period of confinement.
Studying the EEG correlates of the "white-out" phenomenon, W. Cohen (19) hypothesized that under conditions of uniform visual stimulation, the "termination" of visual experience should be accompanied by a return of alpha activity. Despite considerable individual differences, he found that in occipital records, strong alpha activity usually followed the onset of "white-out" with a latency of one second. In general, the onset of alpha occurred when visual experience spontaneously became less differentiated. The occurrence of "white-out" seemed to be related to the amount of alpha shown. Thus about half the subjects showed little alpha even when kept in the dark without stimulation.
Vernon et al. (78) in their study of perceptual and motor skills under conditions of twenty-four, forty-eight, and seventy-two hours of sensory deprivation, found that despite the fact that subjects ate well, there was a consistent loss of weight, averaging two and one-half pounds. Measuring strength of grip with a hand dynamometer, they observed gains for control and confined groups except for the seventytwo-hour confined group which showed a slight loss. A third measure was galvanic skin resistance taken before and after isolation. The authors hypothesized a gain in skin resistance because of the quiet and sleeplike conditions for the confined groups. Instead they found that while control group values rose, a statistically significant drop from preisolation conditions appeared in the experimental group. There was, however, no report of control for the humidity in the chamber.
Utilizing a continuous recording of skin resistance throughout isolation on their four subjects, S. Cohen et al. (18) found that resistance rose and remained high for the two subjects comfortable in isolation. For the remaining two subjects, less comfortable with the experience, this autonomic index remained low. Ruff et al. (65) cited observations that confirm the foregoing, and point to skin resistance measures as a useful reflection of arousal, which parallels overt behavioral manifestations during isolation.
ceptual deprivation in the tank respirator procedure (56). General relationships were observed between patterns of EEG, heart rate, and epinephrine-norepinephrine excretions on the one hand, and behavioral measures of activity, verbalization, and emotional responses on the other.
A more detailed analysis of the catechol amine response of ten subjects isolated in the tank respirator is provided in another report by this group of investigators (55). Comparing preisolation, isolation, and postisolation excretion levels, these authors found a generalized increase in both measures followed by a decline after removal from the respirator. They emphasized the wide variability of response, and identified a number of individual patterns which are masked by the group data. Their findings also showed that of the two measures, epinephrine seemed to change more than norepinephrine in response to this stress.
R. C. Davis (24) compared the physiological responses of twenty-two subjects exposed to minimal stimulation with that of twenty-eight subjects who received unpatterned stimulation at near normal levels. The "reduced-stimulus" group lay on a cot in a dark soundproofed room for about forty minutes, whereas the "unpatterned-stimulus" group was given continuous moderate light and sound after five minutes. He recorded circulatory and respiratory variables and muscle potentials from three locations. Comparing the changes in the two groups during isolation, the "reduced-stimulus" group showed a significantly greater increase in muscular and circulatory activity and a decrease in respiratory activity. The author finds the responses of the "reduced-stimulus" group similar to those seen in subjects anticipating a stimulus.
In general, the data on physiological response is sparse, with much of it resting on case observations. Few of the studies utilize precise measurement as well as adequate controls and sample sizes. Nonetheless, a number of indices have by now shown promise of providing useful information about response to sensory deprivation. These include EEG, skin resistance, epinephrine-norepinephrine excretion levels, and muscular, respiratory, and circulatory activity. There has been frequent mention of movement of gross musculature as related particularly to the phenomena discussed in the section on imagery. [The Davis study (24) cited previously measured muscular activity but not imagery.] Such determinations might prove helpful in relating the role of kinesthetic stimulation to behavior generally and to body image in particular. This latter relationship has been a subject of much speculation.
The physiological measures thus far observed have received attention largely due to their presumed relationship to the concept of arousal. The work of the McGill group has been concerned with the use of EEG tracings as an index to the state of arousal (39). Others have used skin resistance measures in this way. The work with catechol amines has focused on the deprivation situation as a stressor agent. Needless to say a variety of other approaches are possible. Thus the measurement of eye movements during the occurrence of imagery and the comparison with eye movements as they appear in dreams [see Dement and Kleitman (25)] may yield important information about the possible similarities of the two processes. Furthermore, the specification of a variety of physiological changes under conditions of sensory deprivation may provide other important clues in the clarification of the entire range of observed effects.
In this section we shall consider several aspects of the literature being reviewed. One issue concerns tolerance for isolation and deprivation as measured by voluntary length of stay in the situation. Findings on this problem have varied considerably, depending on the experimental conditions. Many investigators have not attempted to assess this aspect of the problem. Others have discarded data from subjects who failed to complete a prescribed length of stay. Thus it is difficult to compare the various experimental procedures for degree of stress as it might be reflected in such a measure.
It was suggested earlier that length of stay is perhaps too simplified an index of tolerance for isolation. There is little information available on the relationship between this and other responses in the deprivation situation. Although it is perhaps not independent of problem-solving efficiency, or incidence of imagery, for certain purposes an estimate of tolerance measured in length of stay is important.
The sensory deprivation procedures involving suspension in water (16, 50) appear to be sufficiently extreme as to make a stay of more than three hours quite difficult. Use of a darkened, soundproof cubicle seems to make considerably longer periods of isolation tolerable. Ruff et al. (65) reported that subjects stayed as long as seven days. Both water tank and cubicle procedures presumably attempt an absolute reduction of levels of sensory input. In the latter there is much more mobility and less restriction necessitated by the physical needs of the subject for food and toileting.
tolerated for as long as six days. In the tank respirator procedure, despite the experimenters having set a limit of thirty-six hours, subjects have tended to stay much shorter times. Comparing two sets of conditions of deprivation using this procedure, it was found that under less severe deprivation, subjects stayed an average of 18.8 hours. With more extreme conditions, another group averaged a significantly shorter 8.8 hours (47). Reducing social isolation by allowing subjects to talk to a second subject in an adjacent respirator has been shown to increase the length of stay significantly (21).
Ruff et al. (65) reported on the importance of the subjects' knowledge of the length of confinement. This knowledge added structure to the experience and thus increased the capacity to withstand the situation. They also reported that repeated exposures have a similar effect in reducing the stress of the experimental conditions; an observation also made by Lilly (50). Upon repetition, the situation loses some of its novelty and the subject becomes better able to assume an attitude of passive acceptance. They make the general observation that when subjects are run to maximum tolerance, the decision to leave is made when there is an impending or partial breakdown of personality defenses.
The option of leaving at will, available to all subjects in laboratory research on this problem, complicates extrapolation of these findings to real life situations. Nevertheless, the data on length of stay are useful, insofar as they permit some assessment of the total impact of isolation and deprivation.
Another issue which some investigators have examined is that of orientation in time. In general, these studies have shown a wide range of response, from minimal to gross disorientation in time judgment. Lilly (50) reported a subjective postisolation impression of being out of step with time, as though the day had started all over again following isolation. This was not confirmed by Camberari (16). Wexler et al. (80) reported data for seventeen subjects which showed no consistent pattern of either underestimation or overestimation of time. Comparing these results to those obtained under more severe deprivation, it was found that in the latter conditions average time error was greater (47). This difference did not achieve statistical significance and appears to have been, in part, an artifact of the relative availability of time cues under the two conditions.
In a situation requiring the estimation of successive thirty-minute intervals, S. Cohen et al. (18) found that the two subjects comfortable in isolation underestimated the passage of time, whereas the two disturbed by the experiment overestimated time in isolation. Ruff
et al. (65) cited the importance of awareness of time as an orienting factor during deprivation which adds structure to the experience. Their subjects showed a general tendency to underestimate time. The loss of time awareness often seemed to make the experiment intolerable.
Goldberger and Holt (32) also referred to this lack of time orientation as an important source of frustration in isolation. Time seemed to pass very slowly for their subjects. Despite this, their judgments were surprisingly accurate, with a relatively small but consistent underestimation. This finding suggested to the authors the relative independence of experiencing time from the act of judging time. Ruff et al. (65) have suggested that the underestimation of time is a defensive maneuver where the subject avoids a premature anticipation of his release from confinement. Thus he avoids the frustration of having to remain in the situation at a time when he might otherwise expect release.
The importance of time orientation in influencing response to isolation and confinement is well documented. Burney (13) describes the elaborate procedures he developed for telling time and of his precise knowledge of dates during eighteen months of solitary confinement. Anecdotal reports have cited very complex schemes worked out by subjects to maintain their orientation in time. Just as deprivation and isolation appear to disrupt general cognitive orientations, so too this situation appears to have similar disruptive effects on time perception. As such, resistance to the disintegrative effects of deprivation and isolation might well emphasize the importance of developing orienting anchors in the external environment for both time and space.
Although the implication of most studies thus far discussed has been that deprivation produces "stimulus-hunger," only one study has made a direct attempt at its measurement. The boredom and restlessness mentioned in the section on feeling states may refer to the phenomenon. Lilly (50) has explicitly described "stimulus-hunger" in the following terms.
Bexton (7), using the apparatus previously described, attempted to determine whether perceptual deprivation would lead subjects to elect to listen to materials they might otherwise consider uninteresting or tedious. These subjects, college students, had an opportunity to listen to records of five minutes' duration. These records contained the following types of material: eight repetitions of the 16-bar chorus of "Home on the Range"; two talks for children, taken from a religious primer; radio commercials for soap; and part of a stock market report. Subjects were divided into two groups of four each. One group heard the records before isolation, whereas the second group was told nothing about it until several hours after entering isolation. Once in the experimental situation, subjects were told they could hear any of these materials, whenever and as often as they liked. They found that the four subjects exposed to the material before isolation universally disliked the records and only asked to hear them a total of nine times. The other group asked for the records fifty-three times, and reported that they helped to relieve the boredom. In addition, it was found that the rate of requests for the records was dramatically higher during the second half of the confinement period. Previous exposure to the material seemed to be the principal factor influencing the demand for stimulation.
One major problem that subjects report in the deprivation situation is the lack of things to see, hear, do, or think about. This subjective complaint seems to have clear relevance to the notion of curiosityexploratory drive studied in experimental work with animals. The isolation conditions thus seem to increase receptivity to otherwise dull, uninteresting material. Whether the same is true for "ego-alien" material is not yet known. In the context of other disruptions of the individual's functioning, this effect appears to magnify the affective value of stimulation. Quantification of these phenomena might provide a useful index for comparing the relative severity of deprivation conditions.
We have already referred to the findings of Ruff et al. (65) in highlighting the effects of time structure in increasing the tolerance of subjects for deprivation. Such factors as provision of tasks during isolation, specification of the length of deprivation, and previous exposure to isolation result in making the experimental conditions more tolerable to subjects. The comparison of two conditions of confinement in the tank respirator has also pointed to the increase
in stress and decreased length of stay that accompanies an increase in isolation and reduced contact with experimenters and environment (47). Additional papers discuss other variables operating to influence response to isolation and deprivation (18, 46). These include such factors as experimenter's expectations, subject-experimenter relationship, and the physical setting of the experiment.
Kandel, Myers, and Murphy (45) compared the effects of two sets of instructions on the reporting of visual sensations in ten minutes of darkness. They found that one group, who were told the experiencing of such sensations was to be expected under these conditions, reported significantly more visual sensations than did another group, told that these sensations appeared in psychiatric patients. Prior verbalization of "fantasy material" through exposure to Rorschach cards did not increase the number of sensations reported when compared with a group not given this test.
A number of other procedural issues require further clarification. We have mentioned earlier that all of these studies have employed volunteer subjects, generally paid volunteers, with the exception of that of Ruff et al. (65). When, in one study (80), these volunteers were questioned about their motivation for participating, they offered, in addition to the money, reasons such as aiding science and testing themselves. There is no such data available on possible differential reactions of volunteers and nonvolunteers. Similarly, one must leave open the possibility that solitude and perceptual deprivation sought at the individual's own need or whim may have different effects than when imposed by an experimenter who creates a highly artificial situation so that he may systematically observe the subject. Another issue related to the work of Kandel et al. (45) cited above, is that of the experimenter's purpose in the research. Instructions that stress endurance, or content of thinking, may produce different responses than do those that state the experimenter's interest to be in the problem of rest and relaxation.
These studies highlight the importance of procedural variables and limit the direct comparison of studies utilizing different procedures. They emphasize again the need for specification of experimental purposes and for procedural choices consistent with those purposes. In this connection the importance of evaluating the total context and its implicit motivational and emotional consequences needs to be clearly recognized.
The application and relevance of these findings to diverse areas of interest have proceeded along with the basic exploration of these phenomena. One of the earliest interests in this area was reported by Spitz in a series of articles (73, 74, 75) which dealt with the deleterious developmental and behavioral effects of institutionalization and separation from the mother upon infants. A reading of these reports indicates that perceptual deprivation was a prominent feature of the experience of these children.
A more direct application of isolation and deprivation to clinical procedures has been attempted recently (1, 2, 3, 4). In these procedures, patients with a variety of clinical diagnoses were kept in a darkened hospital room, wore translucent goggles, and had their arms in cardboard cylinders. These periods of isolation ranged from two to six days. Although few of the cognitive changes described earlier were observed, these investigators concluded that deprivation led to a state of disorganization, and in some cases precipitated psychotic reactions. On the other hand, some groups, particularly depressives, seemed to show improvement in the form of increased motivation, socialization, and assertiveness.
Phenomena akin to those reported in isolation and deprivation have also been reported in aviators, especially in high speed, high altitude flying. Citing clinical material on these flyers, Bennett (6) compared their reactions to those seen in isolation studies. Operating in a severely restricted environment with extremely monotonous stimulation, aviators have reported feelings of isolation, unreality, and dreamlike states. Evidence suggests that these feelings, called by some the "breakoff phenomenon," occur at times in approximately one-third of jet pilots.
Earlier reference was made to the response of prisoners who in solitary confinement apparently experience similar reactions. Meltzer (53) reported the occurrence of a range of effects in such prisoners. These included, on the one hand, occasional tense pacing, restlessness, tension, and assaultiveness. On the other hand, some prisoners exhibit a regressed, dissociated, withdrawn, hypnoid, and reverielike state. Hypochondriacal states of a transient type were also seen.
Already mentioned is the work of Mendelson and Foley (54) which showed the importance of isolation and deprivation in polio patients. Two recent papers have appeared that stress the importance of these findings in understanding a number of phenomena seen in medical
practice (48, 81). These reports stress the relevance of this work to the hallucinations of cataract patients, and to the disturbances seen in patients with orthopedic disorders who are suspended in traction.
Perceptual deprivation also has relevance to prolonged and repetitive tasks in man-machine systems, such as long distance driving, flying, assembly line production, continuous monitoring duty at isolated stations, etc. The factor of the external environment and its influence upon behavior is increasingly being recognized for its role in a wide variety of practical situations. The findings discussed in this chapter may account for such things as accidents, loss of efficiency, and emotional alienation often observed in these situations.
Focused research on the responses of human subjects exposed to reduced environmental stimulation has only recently begun. The phenomenal growth of interest in this problem and some reasons for this have been discussed. Many of the investigations referred to in this paper are pilot studies; others leave much to be desired in rigor, elegance, and methodological sophistication. At the same time, these investigations now provide a rich source of new observations and hypotheses, which touch on a wide variety of issues. The findings, tentative though they are, have important implications.
The results of this work, of the research on curiosity or exploratory drive, and of studies on early sensory deprivation converge to provide a revised conception of human motivation. This conception recognizes and emphasizes the "immediate drive value of cognitive experience" as a necessary factor in a theory of motivation (37).
From the viewpoint of theory this work has important implications for several scientific disciplines. Methodologically it makes available a technique for the relatively controlled study of imagery and hallucinations, a problem thus far inaccessible to experimental observation without the use of drugs. Practically, it suggests a whole range of applications from management of medical patients to highway design.
Theoretical accountings of how reduced environmental input produces the various responses described in previous sections have varied widely. Our purpose here will be simply to indicate the range of explanations used and some of the terms of their analyses.
Rapaport (62) discusses these data from the viewpoint of psychoanalytic theory, in the context of the relationship between id and ego functioning. In a detailed discussion of these issues, he states that
in the absence of external stimulation, the ego becomes unable to maintain its autonomy from the id and the "effectiveness of these (ego) structures in controlling id impulses may be impaired" (page 19). The reduced control of these impulses may be manifested in the difficulty of thinking, in the unpleasant affect, emotionality, and content of imagery previously discussed. In this formulation, ego functioning is closely tied to external stimulation. Ruff et al. (65) extend this notion to account for individual differences by using the concept of sharpness of ego boundaries. Orientation in time and space structure the situation and may thus help the individual retain ego autonomy by keeping some ego functions in operation. Since isolation destructures the id-ego relationship, initial variations in the differentiation of ego boundaries may account for the individual differences seen.
The increased awareness of primary process material states in psychoanalytic terminology what others have described as the heightened awareness of internal bodily states. The decline in perceptual stability and reasoning and the increase in anxiety and imagery seem to arise not only as a function of the increased sensitivity to one's own thoughts, feelings, and ideas but also because of the absence of an external reality against which to validate one's inner experiences experientially or consensually. Although periodic "regression in the service of the ego," to use Kris's phrase, can be quite productive and creative, the sustained inability to go through the validation process seems both to increase anxiety and the distance from the social community. Burney's (13) reluctance to leave his solitary confinement after eighteen months, which has been observed in other autobiographical reports as well, may be one manifestation of this process. At the same time, if one accepts Hebb's notion of the "immediate drive value of cognitive experience" (37), the increased receptivity to any cognitive experience as seen in the brainwashing reports becomes more comprehensible.
A recent paper by Bruner (12) places the work on deprivation in a functional context. Perception is seen as instrumental behavior that permits the organism to manage its necessary transactions with the environment. Successful management of these transactions depends on acquiring an adequate internal model of the external world. Early sensory deprivation interferes with the learning of a stable model. Thus the organism becomes limited in acquiring a full range of efficient problem-solving strategies. Deprivation in adult life interferes with the perceptual-cognitive maintenance needs of the organism. Thus, it disrupts the vital evaluation process by which one monitors
and corrects the models and strategies used in dealing with the environment. Freedman et al. (30) utilize a similar notion of "perceptual degradation" to account for their findings. They attribute the observed phenomena to the organism's continuous search for order and meaning in the unstructured perceptual environment provided in their experimental setup.
From a neurophysiological point of view, Lindsley (51) emphasizes the function of the reticular activating system because of its role in attention, perception, and motivation. This system serves the homeostatic function of adjusting input-output relationships. Sensory deprivation is one of a class of conditions which upsets the balance and thus disturbs the regulating function of the ascending reticular activating system. With markedly reduced input, perception is disrupted; attention gives way to distractibility; interest gives way to boredom; and activity is either held in abeyance or becomes highly stereotyped and nonadaptive. Heron (39) points out how sensory and perceptual deprivation may be equivalent. He cites evidence to show that the capacity of a stimulus to evoke and maintain arousal is lost upon repeated exposure of the stimulus. Hebb (37) presents an excellent theoretical discussion of the implications of the concept of arousal and the manner in which these findings bear upon a variety of issues in motivation theory, such as the generality or specificity of drive states, the need for varied stimulation, and the intrinsically rewarding quality of cognitive activity.
These findings are relevant to the problem of interrogation. There are no experimental data available in the studies reviewed bearing directly on the relationship of isolation and deprivation to the amount and accuracy of information which can be obtained when under interrogation. Nonetheless, the findings reported suggest some major parameters which may facilitate or inhibit the disorganizing effects of isolation.
Before considering further the implications of these studies for the interrogation problem, it may be important to point out some limitations. There has been some tendency to equate the effects of sensory and perceptual deprivation studies with those reported under conditions of solitary confinement. One possibly gratuitous assumption in equating the two is that. the nonpatterned stimulation in these studies simply accelerates the debilitating effects observed with social isolation alone. Several studies (21, 30, 32) explicitly control or account for the social isolation variable as contributing little to the effects observed. Schachter (66) studied the reactions of five students to social isolation without interference with ordinary sensory
input. He concludes that for isolation two to eight days seems to produce relatively little of the painful effects seen in the autobiographical reports of sailors and explorers. However, specific investigations of the social factors in the sensory deprivation studies will be necessary in order to make a more precise generalization.
We have earlier elaborated some aspects of the differences in motivation between the experimental situations and the real life conditions. Because of these differences, and of limited data, caution in generalizing the relevance of these experimental studies is necessary. Pending clarification of these issues, some tentative implications may be suggested as relevant.
The loss of internal standards and the absence of opportunity to validate one's ideas against an objective reality would seem to apply in real life as well as in experimental circumstances. The boredom, restlessness, irritability, and other mood changes observed also may well apply. The stimulus-hunger and increased suggestibility which have been observed may make an individual more vulnerable to revealing information he might otherwise withhold, particularly when accompanied by the social uncertainty induced in the interrogation situation. Unprepared for these consequences of isolation and deprivation, like many experimental subjects, an individual may become apprehensive and indeed panicked by his reactions. The appearance of hallucinatory-like phenomena and their emotional accompaniments have often been quite anxiety provoking. On the other hand, previous exposure to these circumstances, familiarity with their consequences, and training individuals in techniques of dealing with them may well increase resistance. Knowledge of the importance of retaining spatial and time orientation, and self-stimulation in concrete tasks, are two examples of techniques for reducing stress by increasing psychological structure. Schachter (66) points out that isolates who are able to keep occupied with distracting activities appear to suffer less and be more prone to the state of apathy. Schonbach (68), in an experimental study, has demonstrated that a state of deprivation is far more bearable under conditions of irrelevant and distractive thought than under conditions where thought is concerned almost wholly with the source of deprivation.
Since direct research on the problem of resistance to interrogation in a realistic setting is difficult, some reliance on the type of study reviewed here is necessary. Further investigation of these problems will undoubtedly continue to shed new light on resistance to the disorganizing consequences of deprivation. However, despite their often dramatic results, these studies have remained within the limi-
tations posed by ethical considerations and have not pushed subjects to their ultimate limits. Indeed, polio patients survive years in respirators without psychosis, whereas prisoners, sailors, and explorers often successfully endure long months of severe deprivation and monotony. Furthermore, the autobiographical evidence, even if selfselected, implies that the long term effects are reversible and in some instances leave the individual with a sense of having achieved a new and better personality synthesis. From this point of view, the findings reviewed must be considered as suggestive, rather than spelling out in final terms the complete and precise parameters of response.
4. Azima H., Vispos R. H., and Azima Fern J. Observations on anaclitic therapy during sensory deprivation, In Solomon P., Kubzansky P. E., Leiderman, P. H. , et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
12. Bruner J. S. The cognitive consequences of early sensory deprivation. In P. Solomon, P. E. Kubzansky, P. H. Leiderman, et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
18. Cohen S. I., Silverman A. J., Bressler G., and Shmavonian B. Problems in isolation studies. In P. Solotrion, P. E. Kubzansky, P. H. Leiderman, et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
19. Cohen W. Some perceptual and physiological aspects of uniform visual stimulation. Washington, D. C.: Research and Development Division, Office of the Surgeon General, Department of the Army. 1958, Progress Report No. 1.
21. Davis J. M., McCourt W. F., and Solomon P. Sensory deprivation: (1) Effects of social contact, (2) Effects of random visual stimulation. Paper read at Amer. Psychiatric Ass., Philadelphia, April 1958.
30. Freedman S. J., Grunebaum H. U., and Greenblatt M. Perceptual and cognitive changes in sensory deprivation. In P. Solomon, P. E. Kubzansky, P. H. Leiderman , et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
33. Goldberger L., and Holt R. R. Experimental interference with reality contact: Individual differences. In P. Solomon, P. E. Kubzansky, P. H. Leiderman, et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
39. Heron W. Cognitive and physiological effects of perceptual isolation. In P. Solomon, P. E. Kubzansky, P. H. Lciderman, et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
42. Hinkle L. E., Jr., and Wolff H. G. Communist interrogation and indoctrination of "Enemies of the State". Analysis of methods used by the Communist State Police. (Special Report), A. M. A. Arch. Neurol. Psychiat., 1956, 76, 115-174.
45. Kandel E. J., Myers T. I., and Murphy D. B. Influence of prior verbalization and instructions on visual sensations reported under conditions of reduced sensory input. Amer. Psychologist. 1958, 13, 334. (Abstract)
51. Lindsley D. Are there common factors in sensory deprivation, sensory distortion, and sensory overload? In P. Solomon, P. E. Kubzansky, P. H. Leiderman, et al. (Fds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
53. Meltzer M. Solitary confinement. In Group for the Advancement of Psychiatry. Factors used to increase the susceptibility of individuals to forceful indoctrinations Observations and experiments. New York: GAP Publications Office, 1956, GAP Symposium No. 3, 96-103.
56. Mendelson J., Kubzansky P. E., Leiderman P. H., et al. Physiological and psychological aspects of sensory deprivation: A case analysis. In P. Solomon, P. E. Kubzansky, P. H. Leiderman, et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
65. Ruff G. E., Levy E. Z., and Thaler V. H. Factors influencing reaction to reduced sensory input. In P. Solomon, P. E. Kubzansky, P. H. Leiderman, et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
78. Vernon J. A., McGill T. E., Gulick W. L., and Candland D. R. The effect of human isolation upon some perceptual and motor skills. In P. Solomon, P. E. Kubzansky, P. H. Leiderman, et al. (Eds.), Sensory Deprivation. Cambridge: Harvard Univer. Press, in press.
The purpose of this chapter is to review available scientific knowledge on the use of pharmacologic agents to influence the communication of information which, for one reason or another, an informant does not wish to reveal. This problem in communication is not an unfamiliar one to the psychiatrist, who often aims to recover unconscious conflicts or memories from the neurotic or psychotic patient in the hope of producing therapeutic benefit. The purpose of this chapter, however, is not so much to review our knowledge on how to bring to a person's awareness, the feelings, impulses, and ideas of which he is not consciously aware; rather, the object is to focus particularly on the problem of getting data from a source of information when the individual is aware of the information but does not want to communicate it, either because the giving might incriminate him or put into possible jeopardy an aggregate of people toward whom he feels strong allegiance, identification, and belonging.
In the physician's customary role as a helping person and as a healer, it is generally contrary to his method of operation to employ any coercion, overt or subtle, to induce a patient to behave in a way that may be detrimental to himself or to his social or national group of origin. Coercion may be used, however, if the patient is considered to be behaving in a manner that is destructive to himself (e.g., a diabetic refusing to take insulin or an alcoholic refusing to stop drinking) or to his social group. Furthermore, the code of ethics, particularly of the psychiatrist, ordinarily binds the physician to keep
confidential the secrets that his patients impart to him, whether or not the patient has been aware or unaware of their nature. In the practice of psychiatry, the code of respecting and keeping the confidences of a patient is considered to be a tool that facilitates the confession or expression of otherwise taboo material from the patient. The psychiatrist's office is, ideally, one place where the patient finds that he has immunity from punitive or disapproving action by the society in which he lives, except for the adverse criticism forthcoming from the patient's own internalized standards of behavior.
Occasionally the psychiatrist's occupation brings him into contact with patients with whom the psychiatrist himself is led into a conflict between his interest in championing the welfare and privileged communications of his individual patient and that of the welfare of the group: family, city, state, nation. In such a quandary, e.g., the question of what to do about a person who has confided his participation in a major crime, the physician's obligation to the individual and to the community may be in opposition to one another. In this position the physician may be forced to disqualify himself as a continuing confidant for the patient until the patient has remedied his social obligation to the state.
Mentioning these situations and the customary attitude of the medical profession has a bearing on the substance of this report. The use of drugs in obtaining a confession from a criminal, or in obtaining information that a source may consciously wish to keep confidential for fear of repercussion to himself or his group, is fraught with ethical conflicts for the physician. This explains in part why there is a relative paucity of systematized published scientific investigation by physiciaits on this matter. The general feeling in western countries regarding the employment of chemical agents to "make people do things against their will" has precluded serious systematic study of the potentialities of drugs for interrogation. It has not, however, precluded considerable speculation on the subject, some of it rather unrealistic.
Much relevant scientific information has been published on the therapeutic employment of drugs. The bulk of the medical articles of the last few years on the effects of drugs on behavior deals with the use and effects of these drugs on the mentally ill population. In fact, a growing avalanche of articles of this type sprang up with the advent of tranquilizing drugs. From this large body of publications, the reviewer aims to extrapolate to the problems of interrogation. Then, there is a notably smaller group of studies that deals principally wiih explorations in methods of assessing the psychopharma-
cologic effects of drugs on relatively normal individuals. From these studies, too, the reviewer aims to transfer what has been learned to the problems of interrogation. Finally, there are the relatively rare published investigations on the use of drugs for purposes of interrogation in police or security procedures; these are reviewed carefully because of their direct relevance.
No published reports have come to the attention of this author detailing the scientific application of drugs by intelligence agencies of any nation as a means of obtaining information. Apparently, what knowledge is available, whether derived from haphazard experience or systematic study, is not accessible in open sources. This reviewer found only two references touching on this topic. Rolin (112) casually claims that the Nazis used mescaline to get information from prisoners at Dachau. In discussing the methods of communist indoctrination of Americans who have fallen into the hands of communists or communist-controlled countries, Hinkle (62) has stated that the methods of Russian interrogation and indoctrination are derived from age-old police methods that have been systematized, and are not dependent on drugs, hypnotism, or any other special procedure designed by scientists. 1
Methodological Problems in Determining the Applicability of Drugs to Interrogation Procedures: Nonspecific Effects of Drugs on Verbal Behavior
One of the crucial questions arising in evaluating the use of a drug for interrogation techniques is what responses are related to the pharmacologic activity of the drug administered and what responses are related to some other aspects of the transactions taking place when a person receives medication from another person. A large variety of nonpharmacologic factors can affect the responses of an individual after getting a dose of medication (see also Masserman and Pechtel, 102). In fact, one of the major problems plaguing investigators of
1 Popular literature contains a number of accounts alleging the use of drugs in interrogations. Recent well-publicized examples include the alleged use of mescaline against Cardinal Mindszenty (S. K. Swift, The Cardinal's Story, New York: Macmillan, 1949); the prison memoir of an American civilian held by the Chinese Communists (R. T. Ryan, I came back from a Red death cell, Saturday Evening Post, January 17, 24, 31, and February 7, 1953); and the account by the Communist editor, Henri Alleg, of an alleged use of sodium pentothal in interrogations he received while held by French forces in Algeria (H. Alleg, The Question, New York: George Braziller, 1958, pp. 91-102).
ps class="continued"ychopharmacologic relationships is how to discriminate the factors that are responsible for observed effects, how to single out factors so their sole effects can be observed, and how to study summation and combined effects of different factors.A series of nonpharmacologic factors within the total transaction of a person giving another person a drug has been found to be more or less capable of contributing to the responses occurring with administration of the drug. These factors may be listed and what is known about each will be taken up separately.
Reactions Due Primarily to the "Placebo Phenomenon," i.e., Reactions to Taking a Medicinelike Substance, Even Though It Is Pharmacologically Inert
There has been an increasing interest in the importance and mechanism of action of the placebo (46, 67, 83, 135, 136). The studies of Beecher and his group (7, 8) indicate that 30 to 50 per cent of individuals are placebo reactors, that is, respond with symptomatic relief to taking an inert substance. If one is interested in the pharmacology of a new drug and tries it out on a group of patients, a third to a half of this group will be relieved of their symptoms by a placebo; they react favorably to the syringe, pills or capsule, regardless of what it contains. Thus they dilute the significant data derived from the half or two-thirds of the group that react only to the active ingredient in the syringe or capsule. In studying a new drug-whether one is interested in applying its pharniacologic effect toward the alleviation of pain, amelioration of emotional distress, or the facilitation of communication of covert information-the scientist is not primarily interested in the subjective and behavioral effects of syringes and pills. Thus the scientist is obliged to take into account the placebo reactors, who must be screened out if one is to get an accurate idea of what the drug itself does.
the investigator can expect that in 30 to 50 per cent of trials pain may be relieved or interrogation may be facilitated.Some additional factors are known which increase the likelihood of a placebo effect:
If one is interested in ascertaining whether a drug produces a given effect to a degree greater than a placebo, it becomes obvious that the effect produced by the drug must exceed the chance variations of the placebo effect to a reliable extent. In experimental investigations exploring the usefulness of drugs for various purposes, the placebo and other nonspecific reactions to medicaments must be separated from the effects specific to the active drug. Devising an experimental study using infrahuman animals to assess the pharmacologic effect of a drug only postpones the assessment of the complicated responses likely to occur when the drug is given to a human being. For the researcher interested in discriminating specific from nonspecific effects of drugs, Beecher (7) has outlined a series of principles and practices on the basis of seventeen drug studies in which he has participated, as follows:
The act of administering a medication usually potentiates its effect since it invokes the status of a professional person and the prestige of social institutions and organizations that are a part of the setting. A general recognition of this fact has made the control of the placebo effect a routine feature of all carefully designed drug studies. Conversely, silent administration has received little or no attention.
Not all psychoactive drugs are equally suited to silent administration. A minimal requirement is the successful masking of the drug by substances otherwise introduced into the body, such as foods, liquids, smoke, or air. From this point of view the ideal drug would be tasteless, odorless, and completely soluble.
Theoretically, the net effect of a silently administered drug should be equal to its effect following routine procedures minus its placebo effect. In practice this effect would be modified by the state of the organism, the general setting in which the subject finds himself, and his typical and persistent modes of reacting, i.e., personality-constitutional factors. One may expect a very different reaction from a subject who is sensitive to his internal, subjective processes than from one who has learned to disregard and reject them in favor of "objective" external cues. Likewise, reactions will vary between subjects who yield to and expand upon their internal experiences and those who
For these reasons it is difficult to specify dosage levels at which a subject is likely to become aware that he is responding to a drug, since so much depends on personality and situational factors and on the subject's previous experience with drugs. In naive subjects moderate doses which noticeably modify their behavior may escape their attention, or be ascribed to other sources, such as fatigue, thirst, apprehension, dyspepsia, etc.
The judicious choice of a drug with minimal side effects, its matching to the subject's personality, careful gauging of dosage, and a sense of timing, makes silent administration a hard-to-equal ally for the hypnotist intent on producing self-fulfilling and inescapable suggestions. Surpassing "magic room" procedures in their efficacy, the drug effects should prove even more compelling to the subject since the perceived sensations originate entirely within himself.
Reactions to Attitudes or Motivations of the Person Administering the Medication and Interacting with the Informant
One of the major problems involved in the assessment of drug effects is distinguishing the psychopharmacologic effect of a drug from that consciously or unconsciously desired by the person administering the drug. Another related problem of consequence is the extent to which a drug effect, noted by one person using the drugs to achieve his special aims, may be expected to occur in the hands of another person using the same drug for an essentially different aim. Although one assumption of this present report is that drug effects are to some extent generalizable from one situation to another, the limitations of such generalizing need to be clarified. The inference exists that the reaction to a specific drug when used by a physician to relieve the symptoms of a patient will produce a similar response when used to extract covert information from a recalcitrant source. In every instance, where such extrapolations are made from one such situation to another, the reviewer does so merely because little or no germane scientific reports are available in connection with the interrogation situation. In every instance where such an extrapolation is made, it is for heuristic purposes, and the generalized ideas and concepts require careful testing and validation.
effect of a drug may be influenced by the incentives given for participating in a drug study. The subjects were former narcotic addicts who volunteered for research. They were accepted after thorough screening by a board of hospital psychiatrists and other professional personnel, with a view to selecting only subjects with histories of repeated relapses to narcotic addiction and very unfavorable prognoses for future abstention from narcotic: drugs. Simple visual-manual reaction times were measured: without administration of drugs; 50 min after subcutaneous injection of morphine; and 50 min after subcutaneous injection of 250 mg of pentobarbital; each was measured under four incentive conditions, defined in terms of the schedule of morphine rewards offered for participation in the experiments. When a fixed reward was given a week in advance of the tests, morphine accelerated and pentobarbital slowed reaction times. When a fixed reward was scheduled for delivery after completion of the tests, neither drug affected reaction times significantly. When the amount of the posttest reward was made contingent upon speed of performance, morphine exerted no significant effect, but pentobarbital accelerated reaction times. When the same group of subjects were retested one to three days later, with posttest rewards again fixed for all subjects regardless of performance, morphine slowed reaction times and pentobarbitat had no significant effect. In other words, depending on the incentive conditions arranged by the investigators, the same dose of either morphine or pentobarbital exerted. either no effect or acted as a "stimulant" or as a "depressant" on simple visual-manual reaction times. Nevertheless, the action of either of these drugs was "specific" with respect to the actions of the other; thus, the action of morphine changed from "stimulant" to "depressant" when conditions changed from "low" incentive (rewards fixed and delivered before testing) to "high" incentive (rewards contingent on performance and scheduled for delivery after testing); whereas the action of pentobarbital changed from "depressant" to "stimulant" when identical changes in incentive were made. Further analysis of the data of Hill et al. revealed that, in comparison with the range of changes in mean reaction times produced by varying the incentive level when no drug was administered, morphine reduced but pentobarbital increased the sensitivity of the subject's performance to changes in incentive level.
A study by Wolf and Ripley (137) illustrates further that the effect produced by a drug depends not only on the particular agent used, the dose and route of administration, but also on the circumstances under which it is given or how its effect is measured. (See also: Rinkel, 110, 111; Sargant, 116.) They observed the effects of amo-
barbital (0.1 to 0.5 gm intravenously) on 700 patients with various complaints, including hypertension and "tension headache." They found that after administration of the drug, the patients were "at ease" if the setting was "secure" and friendly, but the patients were tense and anxious with equal facility if disturbing topics were introduced during interviews. The effects of the amobarbital on headache and blood pressure varied similarly with the nature of the interpersonal milieu. Another illustration is the report of Beecher (7) that a higher percentage of pain relief from various medications was obtained by a sympathetic woman investigator than by a colder, more remote male.
It is now well known that many drugs when taken internally may produce a transient excitant effect where the user becomes euphoric, talkative, and sometimes emotionally responsive. For example, it has been known through the ages that alcohol loosens the tongue during an excitant phase and that a person with enough alcohol may reveal things he would not ordinarily discuss. As is also well known, however, people react differently under the influence of alcohol. Some become depressed and morose. Some become excited and volatile. Some talk freely and others shut up like a clam. As has been previously pointed out, different people may have different reactions to the same drug and similar reactions may occur to different drugs. One cannot always predict what type of reaction may be obtained.
Various factors such as sex, intelligence, and mental and physical condition can influence the speech patterns of an individual. In order to assess the pharmacology of a drug, the predrug differences in verbal communication must be taken into account. Furthermore, there is strong evidence that the pharmacologic effect of a drug interacts with the status of the human organism receiving the drug.
It is obvious and requires no documentation that people with differing intelligence and educational level show large qualitative differences on tests of intelligence. That they speak differently under standardized conditions of eliciting the speech would seem to follow, but this has not heretofore been investigated systematically. It is also
a common observation that the sexes use language differently, if not in a formal, structural way, then in the items of information that they choose to convey in their speech. Gleser, Gottschalk, and John (54) have studied the relationship of word-type usage to gender and intelligence as measured by the Wonderlic test. They obtained five-minute speech samples of a group of ninety occupationally adjusted, medically healthy individuals. These speech samples were elicted by standardized instructions given by a male investigator. The wordtypes were analyzed and scored according to two systems of categories: a grammatical system and a "psychologic" system. The "psychologic" system attempted to classify words according to the emotive, cognitive, and perceptive processes conveyed, and the animate and inanimate objects denoted, regardless of grammatical part of speech. Under these experimental conditions, significant differences were found to occur in the proportion of certain types of words used by men as compared to women. These differences appeared principally among the "psychologic" categories of words. For example, women tended to refer to themselves more frequently than did men. Women expressed more feeling and emotion, and used more negations. They used relatively fewer words referring to place or spatial relations and to denoting destructive action. Significant differences were also found to be associated, step-wise, with level of intelligence. These differences occurred principally among the grammatical categories. For example, the more intelligent adult was found to use significantly more adjectives and prepositions, but fewer adverbs, verbs, and interjections. A multiple correlation of . 65 was obtained between these variables and Wonderlic I.Q. scores. The differences between the sexes in word-type usage tended to decrease at the highest level of intelligence. These investigators have published tables of separate word-frequency norms for males and females and for word categories that vary with intelligence.
In summary, this study illustrates that gender and intelligence influence speech patterns at the microscopic level of word-types. In experimental studies for determining whether or not a drug will facilitate interrogation, the fact that intelligence and gender separately affect speech requires consideration. Suitable controls need to be included in the research design.
Kornetsky and Humphries (77, 78) have shown that there are reactors and nonreactors to drugs and that the reactors are likely to be those who are more depressed (on the D scale of the MMPI), to have more unreasonable fears, and overreact to environmental stimuli (the Pt. scale of the MAIPI). These investigators explored the effects of a placebo, promazine, secobarbital, and meperidine hydrochloride on a series of objective motor, intellectual, and perceptual activities, as well as on subjective responses. The subjective responses were evaluated 30, 90, 150, and 210 min after the drug was taken. Promazine and secobarbital had an adverse effect on the performance of motor tasks but not on simple intellectual and perceptual tasks. Meperidine hydrochloride in 50 to 100-mg doses did not impair performance in any of these same psychologic functions. The more deviant a subject was on four MMPI scales (psychasthenia, depression, hypochondriasis, and hysteria) the greater the effect of the drugs. Also, those who were most affected by one drug were most affected by another. Finally, if large enough doses of a drug were given, all subjects tended to respond in the same manner.
In a well-controlled study, Lasagna et al. (84) concluded: "In addition to dose and route of administration, the nature of the subject and the situation in which a drug is administered are important determinants of drug effects." Using a double-blind technique, these investigators administered, in randomized order, a placebo (1 ml of physiological saline solution subcutaneously), amphetamine (20 mg/ 70 kg body weight subcutaneously), heroin hydrochloride (2 to 4 mg subcutaneously), morphine phosphate (8 to 15 mg subcutaneously), and sodium pentobarbital (50 and 100 mg intravenously) to twenty healthy "normal" volunteers,' thirty patients in a hospital for chronic diseases, and thirty former narcotic addicts serving prison sentences in an institution devoted to the treatment of narcotic addiction. At intervals, before and after administration of the drugs, the subjects completed a questionnaire designed to measure the "subjective" mood changes induced, and this was supplemented by discussing with the subjects their responses to the drugs. In "normal" and, to a lesser degree, in chronically ill patients, amphetamine surpassed morphine, heroin, pentobarbital, and a placebo in producing euphoria. In the narcotic addicts, however, morphine was reported to produce a more pleasant effect than heroin, amphetamine, or a placebo.
Laverty (85) gave randomized injections of sodium amytal and a placebo to forty subjects divided into four groups of ten each: introverted neurotics, introverted normals, extroverted normals, and extroverted neurotics, as assessed by scores on Guilford's R scale for
extroversion. The sedation threshold, as measured by the onset of slurred speech, was highest for neurotic introverts and it decreased step-wise for each of the groups in the order given. In other words, the group of introverted neurotics required the largest amount of intravenous sodium amytal (6.4 mg/kg body weight) before developing slurred speech.
Weinstein et al. in a series of studies (127, 128) have explored the differential effects of intravenous sodium amytal on subjects with preexisting organic brain disease and nonbrain-damaged individuals with illness not involving the nervous system. Some of the brain-damaged patients, who before receiving the drug had expressed awareness of illness and who had good orientation for place and person, with the drug became disoriented for place and grossly misidentified the examiners and explicitly denied illness. Weinstein claims that these changes with sodium amytal occurred only in the brain-damaged individuals; whereas, in nonbrain-damaged individuals receiving sodium amytal, the subjects talked of illness in terms of a third person, used more "concrete" symbols, selectively misinterpreted questions about illness, and misnamed the examiners in "paraphasic" fashion. (Incidentally, Weinstein reports that with the types of patients he studied, sodium amytal did not make them more communicative.) The significance of Weinstein's investigations appears to be that the person with brain-damage gets his sensorium more disorganized with sodium amytal than a nonbrain-damaged individual; an instance, apparently, of the effect of adding insult to injury.
Hoch, Cattell, and Pennes (64, 65) administered sodium amytal, pervitin, and mescaline to each of sixteen patients suffering from the pseudo neurotic form of schizophrenia, twenty-four patients with an overt form of schizophrenia with slight. to moderate deterioration, and nine schizophrenic patients with severe deterioration. With these drugs, especially with mescaline, they found typical physiologic changes occurring, mainly involving the vegetative nervous system. However, with most patients, some aspects of the drug experience seemed to be a direct continuation of previous personality factors. For instance, a patient who showed obsessive compulsive features before the drug experiment would tend to show the same obsessive structure while intoxicated. The same was true about anxiety attitudes, intellectualizations of conflicts, preoccupations with artistic, philosophical, or other matters.
sponded differently. Bensheim (15) thought that the cyclothymic group responded with euphoria and depression to mescaline and the schizothymic group with ecstasy. Lindemann and Malamud (94), experimenting with sodium amytal, cocaine, hashish, and mescaline on the same patients, found that each drug has its specific characteristics, but that the changes produced by a given drug were molded by the patient's personality (see also, 92, 93). Guttman and MacClay (59) and Sarwer-Foner (118) also found correlations between personality and drug reactions. Rubin et al. (114) suggested that the drug reaction be divided into two parts: those responses that are characteristic of a drug regardless of the patient in whom they occur, and those seen in a given patient regardless of the nature of the drug used. The first could be called a collective reaction and the second an individual reaction dependent on the individual's personality.
It has been obvious to those who listen to and study people with personality disorders that the verbal behavior of an individual suffering from an emotional disorder is relatively peculiar, both in form and in content. The depressed person is laconic. The manic person is diffusely wordy. The hysteric and schizophrenic are quite variable in the duration and length of their remarks. There are typical thematic and structural characteristics of the speech habits of patients with these types of psychiatric disorders. (See, for example, Gottschalk et al., 56.)
At present, there are more gaps than facts in our knowledge about the reactions of different personality types to the same and different drugs. Years of intensive research are needed to supply some of this unavailable knowledge.
However, it is already acknowledged that individuals with features of hysterical, conversion, or dissociative reactions are likely to be suggestible and to react strongly to all psychopharmacologic agents, including placebos (83, 85).
Drugs may tend to reinforce the need to give for individuals burdened with feelings of neurotic shame and guilt, especially if such feelings are enhanced by the interrogator. Drugs may also furnish the needed excuse and relief from personal responsibility for sources who violate internalized values and loyalties in revealing information.
The pharmacologic effect of the drug is probably of less decisive influence in facilitating information-getting (although acting as a catalyst) than is the potential readiness of individuals with such per-
sonality features to behave in a typical way under certain circumstances. The consideration of drugs as an aid to interrogation presupposes a thorough understanding of the personality characteristics of the informant and of drugs, to predict what might be expected by their use.
EFFECT OF PHYSIOLOGIC CONDITIONS ON DRUG RESPONSES: BIOLOGIC RHYTHMS, NUTRITIONAL STATES, ISOLATION, AND FATIGUE
There is evidence that the physiologic condition of the individual affects his behavior and responsivity to further incoming stimuli. If so, this propensity is pertinent to our specific interest regarding the use of drugs in affecting the verbal behavior of informants.
It is difficult to ascertain to what extent the behavioral alterations that have been noted under various physiologic conditions are mediated by biochemical changes per se, and to what extent they are secondary psychophysiologic reactions to subtle changes in body chemistry. The answer need not occupy us here, except to note that a chemical alteration within the body is probably one important feature of the varying responsivity of the individual. Under such circumstances, the addition of other chemicals complicates the problem of predicting the behavioral outcome. This is particularly true if the new chemical introduced into the body is mild in its effects, or if it is given in a small dosage.
Biologic Rhythms. Benedek and Rubenstein (12, 13) have studied the relationship of associative material presented by women during psychoanalysis at various phases of their menstrual cycle, as measured by vaginal smears. These two types of data, verbal material and physiologic changes in the vaginal mucosa, were collected and analyzed independently by the two investigators, one a psychoanalyst and the other an endocrinologist. After a long period of collecting such data, the investigators related verbal productions to the phases of the menstrual cycle. A high concordance between the two types of data occurred. This has been validated in clinical studies to some extent. However, because of the importance of the psychophysiologic implications of this classical study, independent validation by other investigators would be desirable. In brief, the investigators found that during the estrogen phase of the menstrual cycle, the women were more extroverted, had more fantasies, dreams, and subjective experiences indicating strivings to be loved and impregnated and had con-
flicts about such strivings. During the progesterone phase, the women were more introverted, were more preoccupied with interests in their own body and self. During the premenstrual phase of the cycle, there were increased references to cleaning out, washing out, evacuating, losing something, and the women were more depressed. Again in these studies, individual variations occurred in relation to the varying hormonal phases of the sexual cycle, depending on the woman's personality type and the kinds of conflicts she had about procreation, childbirth, mothering, etc.
Nutritional States. Studies of the effects of the state of nutrition, especially vitamin deficiency, on human behavior are replete in the medical literature and indicate that neurological and psychiatric disorders may ensue with various vitamin deficiencies, particularly of the B complex. The effects of starvation, voluntary and enforced, in provoking increasing lassitude, apathy, depression, preoccupation with food, flattening of affect, and mood are sufficiently well known and are discussed in another chapter of this study. The more subtle effects of satiation with food and the brief deprivation of food typical of everyday rhythmical eating habits on response patterns to psychologic tests and interviewing procedures have received little careful study, even apart from problems of drug effects.
Clinical psychiatric experience indicates that some individuals become querulous, demanding, restless, even paranoid, and experience hunger contractions if they have not eaten for one to two hours, although they show no demonstrable pathologic, metabolic processes. Other individuals may miss several meals, yet experience no subjective reactions and show no signs of distinctly different behavior. Gottschalk and Gleser (55) did a controlled study of the effect of fasting for twelve hours on the speech patterns of six paid physically healthy and occupationally adjusted volunteers, three males and three females. MMPI's were obtained on all subjects. No homogeneous effect of fasting states on thematic speech variables or on the proportion of various categories of word-types was found under these experimental conditions. In one subject, however, characteristic and repetitive reactions occurred to the stress of the mild fasting, reactions which were principally in the form of significantly increased references to food, home, mainland, mother, and involving attempts to bridge the distances between such objects. Hypoglycemic states were induced by the injection of intravenous insulin in this same subject and the effects of these states were noted. A repetitive, but different, thematic reaction occurred to this experimentally induced hypoglycemia as
compared to voluntary fasting for twelve hours. The investigators concluded that fasting for twelve hours was not enough of a stress to produce consistent effects in the speech patterns of five, out of six, paid volunteers. Susceptibility of reacting to twelve hours of fasting was considered to be due to individual personality features. These investigators did not feel that this initial study gave them enough data to be able to predict accurately which subjects might react and which might not react to this stress.
Isolation and Fatigue. In unnatural biologic states, such as in experimental or enforced isolation from other humans or from ordinary levels of physical stimuli (19, 25, 60, 91) or in loss of sleep (124) emotional disturbances and transient psychotic states have been reported. In states of sleep loss, paradoxical reactions to drugs may occur. In this connection, Wendt (129) is quoted as observing: "There are some interesting things about secobarbital or any of these drugs when ased in individuals under stress, e.g., after 40 hours of enforced wakefulness people become irritable, anxious, apprehensive, and difficult to keep awake. A small dose of secobarbital of 100 mg will wake them up and make them volunteer to go through another night." To this reviewer's knowledge, this phenomenon has not been reported elsewhere and it is important enough to merit. further testing. Such paradoxical effects have been noted clinically with other drugs, notably amphetamine which may have a sedative and quieting effect on restless. anxious, irritable small children with behavior problems presumably associated with psychomotor or petit mal epilepsy (see Forster, 48) and which may reduce the agitation of excited schizophrenics (21).
Methodologic Problems in Determining the Effects of Drugs on Verbal Behavior: Influence of Method of Sampling the Verbal Behavior on the Effect of a Drug
In this brief section, the reviewer, for the sake of completeness, wants to emphasize that the scientist studying this problem must realize that he will discover no more information than his method of evaluation will provide, and that different methods of sampling the verbal behavior of a subject under drugs may give somewhat different information about the psychopharmacologic effect of the drug. Each scientist will tend to use the measuring instrument most familiar to him, and each instrument or technique will have different merits.
The questionnaire of subjective reactions and the directive interview will provide yes and no answers and some qualified answers to structured concepts and hypothesis of the investigator. These methods have the advantage of relative speed and ease of evaluation.
The nondirective interview and the free-associative technique can be applied in a systematic and quantitative way and constitute a valuable means of studying the pharmacodynamics of drugs (see also Kubie, 79 and Wikler, 132), but evaluation of the data is generally slower and more complicated. Although some specific questions in the mind of the investigator may remain unanswered (e.g., is the material fantasy or fact?), these methods may provide considerable information about the drug's effect on symbolic processes, effects, and psychodynamic relationships. To approach definitive answers regarding the potential action of drugs on human behavior, emotion, cognition, and conation, our knowledge needs to be much more complete at the physiologic, biochemical, and psychologic levels of organization. The scientist would best look at his data in as many ways as possible and use a variety of approaches in studying these phenomena (see also Wikler, 131 ; Miller, 103).
In 1921, R. E. House, an obstetrician in Texas, observed in deliveries in which the mother had been given scopolamine that in a certain stage of anesthesia or sedation she might be talkative and reveal things she would not ordinarily discuss. He noted that after childbirth, the mother frequently forgot that she had suffered pain, that she had complained of it, and that she had spoken of personal matters. After the use of scopolamine, often with the addition of chloroform, had proved to have certain advantages in the obstetrical management of a woman delivering a baby, House persuaded himself to extend the use of scopolamine beyond its original purpose to the interrogation of criminal suspects. He gave many enthusiastic demonstrations throughout the United States. As a result, newspapers quickly applied the term "truth serum" to this sedative drug. House's enthusiasm about scopolamine as an adjunct in obstetrics led him to overenthusiastic statements about the value of the drug in interrogation. In 1931, on the basis of two cases, he stated (69) that a person under scopolamine could not lie and that the drug could distinguish the innocent from the guilty. This statement is an example of an
Since House's early experiments with scopolamine (69), which led to the misnomer "truth serom," a great deal more has been learned about drug action. During World War II an interviewing method employing the administration of an intravenous barbiturate was used with disturbed soldiers who were experiencing acute war neuroses in order to allow them, transiently, to relive certain of their battle experiences which were believed to have aroused persistent emotional conflicts (58, 117). In psychotic patients, particularly catatonic schizophrenics, who will not talk and therefore do not participate in psychiatric therapy or reveal any clues to the mental experiences which may underlie their disorder, sodium amytal has been used to facilitate communication with the patient (117). If it works, there is a transient phase that can sometimes be prolonged by injecting the drug slowly, during which the patient will answer some questions and communicate some of his life problems. If the patient passes through this stage into a deeper stage of narcosis there may be a transient period of talkativeness as he recovers from the sedation or anesthesia.
For certain personality types, some drugs lower conscious ego control, thereby facilitating recall of repressed material and increasing the difficulty of withholding available information. The ideal drug for an interrogator would be one which not only accomplishes this feat, but does so without interfering with integrative capacities and intellectual functioning. Because of the uncertainty of the truth or falsity of statements obtained under circumstances of reduced ego control, and because certain drugs may give rise to psychotic manifestations such as hallucinations, illusions, delusions, or disorientation, the verbal material obtained cannot always be considered valid. Such data is not accepted in a court of justice and the information so obtained is not considered wholly accurate by the medical profession.
Jean Rolin (112) has written a book entitled Police Drugs in which he inveighs strongly against the use of drugs for medico-legal purposes. His argument is in part moral, but it is also based on the grounds that there is uncertainty as to the truth of revelations obtained by such means. He concludes that ". . . from a purely medical standpoint, confessions obtained by drugging are valueless and do not give grounds for determining responsibility."
The published experimental studies on the validity of confessions obtained with drugs are few. Much further careful investigation is needed to clarify the problems involved. This reviewer was able to locate only four studies worth reporting here.
The first study is only of borderline relevance and involves the use of intravenous barbiturates as an aid in the differential diagnosis between conversion hysteria and malingering. The author (104) claimed that the use of intravenous sodium amytal was found to be helpful in detecting (and treating) individuals who were suspected of consciously distorting and feigning disability. He found such individuals to be negativistic, sullen, and nonproductive at first under amytal but prone to reveal the fact of and causes for their malingering as the interview proceeded. It was common in his experience to turn up a neurotic or psychotic basis for the malingering. (See Chapter 7.)
The other three studies deserve more detailed review because of their relative superiority, and thus rarity, as research studies in this highly specialized and untouched area.
Redlich, Ravitz, and Dession (109) asked a total of nine university students and professional persons to relate some true shame- or guiltproducing life incident. Then the subjects were asked to invent a "cover story" to be told to another examiner who interrogated them after the intravenous injection of amobarbital, 0.25 to 1.0 gm. In six of the subjects, the "cover story" was given during the amytal interrogation, in one it was mixed with the true story, and in two the true story was given. In nearly all subjects, the "cover story" contained elements of the guilt involved in the true story. However, except in
2 See also MacDonald, 99, 100; Underhill, 125.
the cases of those who confessed the truth, the true story could not be inferred from the story told under amobarbital. An additional finding of interest was that the more normal, well-integrated individuais could lie better than the guilt-ridden, neurotic subjects.
Gerson and Victoroff (53) used amytal interviews on neuropsychiatric patients who had charges against them at Tilton General Hospital, Fort Dix, New Jersey. The patients were told that none of the material from the interviews would be used in the prosecution of charges against them, since it was considered a breach of medical ethics and because the material, derived with the full knowledge and consert of the patient, could not have been presented in court without violating the Twenty-fourth Article of War and the Bill of Rights of the Constitution of the United States.
The researchers first gained the confidence of the patients by discussing their life history. They were not informed that amytal would be used until a few minutes before narcoanalysis was undertaken. It was explained that the drug would make them sleepy and encourage them to discuss things with the doctor that might enable him to gain fuller understanding of the patients' personality and motivations. Under these circumstances, the patients' attitude varied from unquestioning compliance with the procedure to downright refusal to submit to the injection. (One gram of drug in 10 cc of distilled water was injected slowly in the median cubital vein.) Questions relevant to the charges against the patient were not brought up until later, the initial discussion involving relatively innocuous material about the patient's personality and life history which had been discussed previously. A follow-up, waking interview was carried out on the day following narcoanalysis, during which the covert material unearthed during the amytal interview was brought up, and possible psychologic factors in the patient's criminal behavior were discussed. During the follow-up interview, nine patients admitted the validity of their confessions and eight repudiated their confessions. Gerson and Victoroff examined the following factors interfering with the completeness and authenticity of the confessions: (a) inept questioning, (b) tendency of the patient to perseverate on unrelated topics, (c) mumbled, thick, inaudible speech and paralogia, (d) fantasies, (e) contradictory but apparently truthful evidence, and (f) poor rapport between doctor and patient. These experimenters concluded from their study that under sodium amytal subjects could sometimes lie and that their reasoning powers were sometimes present, although much distorted. Although they found amytal narcoanalysis successful for the revelation of deception, they felt that the validity of the information
garnered by this method was not so decisive that it could be admissible in court without further investigation and substantiation. They acknowledged that the doctor could not tell when the patient's recollections turned into fantasy, could not positively state whether he was simulating deep narcosis and actually maintaining his lies, and could not, without social investigation, determine which of contrary stories told under narcoanalysis was true.
Clark and Beecher (30) tested the ability of twenty, paid, volunteer male college students while under drugs to withhold deliberately information during four to eight hours' sessions from a male interviewer, who also administered the drugs intravenously. Thiopental, atropine, amphetamine, methamphetamine, sodium amobarbital, ethyl alcohol, scopolamine hydrobromide, pentobarbital sodium, morphine, caffeine sodium benzoate, and mescaline sulfate were given singly and in combination. The subjects were motivated by their desire for monetary compensation, their perceived importance of the experiment, and pride in their integrity and "will power." The information which the subjects were asked to withhold consisted of (a) two items of personal information (e.g., birthplace, mother's maiden name, etc), (b) an "experimental secret" devised to resemble military intelligence, and (c) a guilt-laden personal experience for which the subject devised a "cover story." All of this information was reduced to writing and deposited with a technician. To check possible forgetting, the subject was asked to produce the withheld information at the end of the experiment, and this was verified against the written version.
In evaluating the considerable ego-integrity maintained by these subjects, it is important to consider that they may have felt relatively secure in a protected experimental situation, in the hands of a responsible experimenter and physician.
In summary, then, clinical experience and experimental studies indicate that, although a person's resistance to communicating consciously withheld information can be broken down with drugs, and particularly sodium amytal, the interrogator can have no easy assurance as to the accuracy and validity of the information he obtains (see also MacDonald's discussion, 100, and that of Inbau and Reid, 71). Furthermore, a lack of crucial information from a subject under a drug does not mean that the subject has no information. An interrogator would have to evaluate many other factors — the personality of the subject, the milieu, other sources of evidence, etc. — to decide how to interpret the outcome of an interview with a drugged informant.
Specific Effects of Drugs on Verbal Behavior, Particularly Drugs Potentially Applicable to Interrogation Procedures
After looking at these efforts to elicit information with a variety of drugs, it may be well to consider each psychopharmacologic agent in turn, for its possible applicability to the interrogation situation.
The major share of studies on the use of drugs in interviewing procedures involves the barbiturates: amobarbital, secobarbital, and pentothal. These drugs have been found useful in treating the acute war neuroses (58, 116, 117), and in civilian practice (23, 32, 106). In psychiatric practice, the purpose of these drugs is to effect a violent emotional response which may have cathartic value for the patient. In the hands of some psychiatrists (117), the emotional reliving enhanced by the drug is not considered necessarily related to a real experience. In order to bring about a high degree of excitement, Sargant (117) has recommended putting the patient back into a past which has been modified by the therapist's invention. In his drugged
Sodium amytal has been found helpful in determining whether or not a subject is feigning ignorance of the English language (96). It is reported that familiarity with a language will show up under the influence of intravenous barbiturates.
Kelley et al. (74) reported that more reliable estimations of intelligence are frequently possible under amytal. They also found that patients under sodium amytal injection gave a greater number of Rorschach responses and fewer rejections of cards. This made diagnoses possible in cases previously considered unreachable. The responses were found to be qualitatively less bizarre and stereotyped, permitting more nuances in personality descriptions.
Brickner (22) has recorded many interviews of patients receiving deep narcosis therapy with barbiturates. The detailed analyses of the verbal productions of these patients have indicated certain typical peculiarities worthy of mention in this review. If they are present, obtained information should be discounted as factual data, although they certainly may reveal in an indirect way some of the gamut of life experiences of the interviewee. Brickner noted the processes of "fractionation" and "recombination" in the verbal productions of patients under deep amytal narcosis. These processes were operative not only at the level of words and word elements, prefixes and suffixes, but also at the level of phrases and clauses, ideas and concepts. The fractionation and recombination manifested itself in the juxtaposition of word fragments, phrases, and concepts which are not ordinarily brought together and in which the connection was often illogical. Brickner believed that this drug process is a caricature of the waking process of comparing new stimuli, percepts, and concepts with others, new and old. He believes that this process of comparison has survival value and is built into human neural structure.
Although the detection and study of such phenomena are of basic research interest to the investigator studying the neuro physiologic correlates of psychologic processes, the decoding of such verbal material by any interrogator seeking factual information is likely to present a very difficult problem. It is probably such phenomena which Gerson and Victoroff (53) observed in their interviews of criminal suspects under barbiturates and which they found to be one of the obstacles to assessing the validity of their informants' verbal productions.
In summarizing the specific effects of barbiturates as facilitants in interviewing, the references already presented and the work of others (32, 101) indicate that, with some exceptions, the following effects on behavior may be expected from the administration of barbiturates to human subjects: (a) decreased attention to stimuli; (b) warmer and more appropriate mood; (c) decreased anxiety; (d) increased contact and communication; (e) reduction of psychotic manifestation.
An increasing variety of nonbarbiturate sedatives have been compounded in recent years. Although they are in wide use, no experimental studies have compared these drugs to the barbiturates as adjuncts to interviewing individuals, either to relieve emotional disturbances or to obtain consciously withheld information.
A list of the chemical and trade name of some of these drugs may be worth including as an illustration of their variety and for heuristic purposes: ethchlorvynol (placidyl), glutethimide (doriden), methyprylon (noludar), methylparafynol (dormison), captodramin (suvren), oxanamide (quiactin).
Administered intravenously to nonpsychotic individuals, researchers have found amphetamine to produce a "push": an outpouring of ideas, emotions, memories, etc. (24, 31, 37, 39, 73, 90). It is of diagnostic help with psychiatric cases by itself (24, 49, 90, 95, 121), or following an intravenous barbiturate (37, 39, 70, 113). It is widely marketed and used in combination with a barbiturate as a mild stimulant drug for patients having neuroses and neurotic character problems.
Brussel et al. (24) claim that methamphetamine hydrochloride is useful in the interrogation of the psychopath who feigns amnesia or withholds vital information which he covers with lies or cautiously alters as he shrewdly weighs his words. These authors claim, perhaps extravagantly, that such a psychopath is powerless under the influence of methamphetamine. Once the drug takes effect, they hold, the tempo of productivity and the insurmountable urge to pour out
It should be noted again that amphetamine and its derivatives are among the main drugs that have been employed in well-designed and controlled studies, showing that the effects of drugs are variable and influenced by personality differences (83, 84, 126).
Pipradrol (meratran) is another drug of the stimulant type which increases not only psychological activity but motor activity as well. In reaction to the "inner push" of ideas, emotions, and speech, some normal subjects report mild euphoria, but others report tension and displeasure. The occurrence of the predominant pharmacologic effects of this drug depends to some extent on the typical personality of the subject (57). Like amphetamine, pipradrol in single, small doses improves the performance of normal subjects in tracking tests (107). It has an advantage over amphetamine in having fewer undesirable side effects, particularly on the cardiovascular system.
This drug has been used in the treatment of patients with "simple depressions" (3, 44, 120). Pharmacologic effects are noticeable in mentally ill patients, but more than a transient therapeutic effect has not been established. No studies are reported on the use of this pharmacologic agent for psychotherapeutic or interview purposes.
Phenidylate (ritalin) is another one of the newer compounds having analeptic effects, such as producing arousal and elevation of mood and increasing the rate of communication. In psychiatric practice this drug has been reported to exert beneficial effects on psychotic patients receiving reserpine (a rauwolfia tranquilizer), which sometimes induces manifest depressive reactions in patients as a side effect (45). But a double-blind, placebo-controlled study has not found ritalin to be of any benefit in chronic schizophrenia (29). The analeptic effects of this drug are well established. The utility for interrogation purposes of the analeptic properties of this drug, as compared to those of other stimulants, such as amphetamine, cannot be evaluated from existing information.
receiving the drug as part of an experimental chemotherapeutic regimen (35). It has been said to improve the performance of normal individuals, enabling them to work more energetically and more effectively, and to need less sleep. Iproniazid has been used in the treatment of mental depression (97) with encouraging results. Definitive well-controlled studies, however, have been reported for neither normal subjects nor mentally ill patients. Several instances of fatal toxic hepatitis have occurred when doses over 150 mg a day were given. Nevertheless, the advent of drugs of such a presumably powerful therapeutic effect in depression, owing either to psychologic conflicts or secondary to chronic somatic illness, opens up further areas of exploration of relevance to the present topic.
It has been common knowledge for centuries that many drugs may modify the behavior of man to the extent of producing psychotic behavior. Hoch et al. (66) have pointed out that acute and chronic psychosis of the toxic type can be produced in susceptible individuals by central nervous system depressants (barbiturates, bromides); central nervous system stimulants (amphetamine, caffeine); analgesics (acetylsalicylic acid, acetanilid); autonomic activators and blockers (atropine, scopolaniine); local anesthetics (cocaine, novacaine); antimalarials (quinine, atabrine); oxytocics (ergot alkaloids); heavy metals (lead, mercury, arsenic); hormones (thyroid, cortisone, ACTH); gases, including low and high oxygen concentration in the inhaled air, and even water in toxic amounts. The interests of psychiatrists in these phenomena have stemmed largely from the supposed resemblance of these psychoses to schizophrenia, because of the time-honored, though as yet unproven, hypotheses (11) that schizophrenia is due to a "toxin." Since Wikler (131) has recently reviewed the literature on this subject critically it will not be done here. Two psychotomimetic drugs are discussed briefly here as examples, from the viewpoint of their psychopharmacologic effects and their possible pertinence to interrogation procedures.
but of no value in the therapy of psychiatric patients. He studied fiftynine patients, seventeen with pseudo-neurotic schizophrenia (Group I), twenty-six with overt schizophrenia but without deterioration (Group II), and sixteen deteriorated schizophrenics (Group III). New psychodynamic material was gained from the first two groups, but relatively little from the last group. Cattell reasoned that the new material obtained in the mescaline state had been condensed and repressed in the drug-free state. Mescaline in general (63, 65, 66, 110) has been found to produce perceptual distortions and hallucinations, accentuation of affective experiences, and increased psychotic manifestations. In some patients contact and communication were increased and in others decreased. Denber and Merlis (40, 41) using 0.5 gm of mescaline in water intravenously on both psychoneurotic and psychotic patients claimed that mescaline induced the production of suppressed and repressed sexual and aggressive conflicts, with a predominance of emotional rather than ideational reactions.
Hoch (63) noted that mescaline in "normals" produced more of an "organic reaction" with some schizophrenic features, as compared to its effect in schizophrenics and latent schizophrenics where mescaline produced more complete schizophrenic disorganization. Hoch also emphasized that mescaline produced a falling off in intellectual functioning.
These articles are typical of the reports on the psychopharmacologic effects of mescaline. When interrogators extrapolate clinical psychiatric observations of this kind to the problems of interrogation, mescaline might serve their purposes in attempting to create an atmosphere of fear or terror in the informant and the illusion of magical overpowering omnipotence about himself. After such a transient state has been created, the susceptible informant might be induced subsequently to reveal information. The perceptual and cognitive disturbances produced by the drug make it unsuitable for obtaining undistorted information while the source is under its influence. From the viewpoint of the informant, the creation of a transient psychotic state by the ingestion of mescaline or lysergic acid might offer him some temporary protection against being successfully interrogated. An interrogator is not likely to consider an individual in a psychotic state a suitable candidate for providing reliable and useful information, at least until the drug effect wears off.
Busch and Johnson (26) gave 30 to 40 gamma of LSD-25 to twentyone chronic psychotic patients (mostly schizophrenics) and later to some psyclioneurotic patients. They noted that the drug transiently increased the mental activity of their patients 30 to 60 rain after ingestion. The effect was a transitory toxic state in which repressed material came forth "sometimes with vivid realism" and emotional expression. Busch and Johnson thought that LSD might be useful as an adjuvant in psychotherapy.
Deshon et al. (42) studied the effect of the ingestion of 1 gamma per kg body weight of LSD on fifteen normal volunteers. Alterations were observed in thinking, speech, emotions, mood, sensation, time perception, ideation, and neurologic signs. The reaction lasted 12 to 16 hr in most cases, but several days in one case. The reaction was typical of an exogenous toxic state, simulating a schizophrenic reaction. The reaction was not specific and the extent to which it was dependent on the basic personality was not determined.
Bercel et al. (16) studied the relation between the type of LSD psychosis produced in normals and of the Rorschach Test findings of the subjects. They could not predict the type of psychotic reactions from the pre-LSD Rorschach, but they could often say from the Rorschach records which normal subjects would show psychotic symptoms.
Abramson et al. (1, 2) found, after administering large doses of LSD-25, that intellectual functioning, as measured by a battery of tests, was disturbed in many spheres. In another study, Levine et al. (89) showed that Rorschach Test scores were altered in the direction of a more psychotic picture.
Davies and Davies (36) treated sixteen mental defectives with LSD-25 in dosages of 20 to 400 gamma in water for as many as twenty-six treatments in three months. Seven patients became more talkative and cooperative. Their "memories were stimulated," but emotional reactions were limited. No lasting benefits were observed. An interesting, unexplained phenomenon was that only two out of the sixteen cases had the expected bizarre hallucinatory experiences.
There are many other reports of experimental and clinical studies employing LSD-25, but those reported here are fairly representative. There is enough given here to suggest the possible applications of LSD to interrogation techniques. It is apparent that this drug impairs perceptual and intellectual functioning. The conclusions reached on mescaline hold equally for the possible applications of this drug to
Recently a large number of new compounds with sedative and anxiety-relieving properties have been introduced, sufficiently different from the classical sedatives, such as paraldehyde, chloral hydrate, barbiturates and bromides, to warrant using new terms to describe them. These compounds are referred to as "tranquilizers" or "ataraxics." They have been investigated clinically on an unprecedented scale in the treatment of psychiatric disorders, particularly the psychoses, and to a lesser extent in the psychoneuroses. The use of these drugs has revolutionized psychiatric therapeutic procedures, especially within the mental hospitals. A thorough review of their effects and mechanism of action in mental illness would be out of place here. Unfortunately, the usefulness of these tranquilizers in exploring psychologic processes and in facilitating communication has not been very extensively tested. Their applicability to interrogation procedures is still speculative. Yet, for the sake of completeness, and to indicate the directions further research might take, the psychologic actions of these drugs deserve mention.
Among the phenothiazine derivatives now in medical use are: chlorpromazine (thorazine), mepazine (pacatal), perhenazine (trilafon), proclorperazine (compazine), promazine (sparine), thiopropazate (dartal), and triflupromazine (vesprin). Of these, chlorpromazine has been most widely used and investigated at this time. Therefore, the psychopharmacologic actions of only this compound are discussed here.
Chlorpromazine (thorazine) . Delay and his associates (38) appear to have been the first to explore chlorpromazine in the treatment of mental illness. They found that the effects of chlorpromazine in patients with manic psychoses were somnolence, decreased responses to external and internal stimuli, pleasant indifference, and decreased spontaneity of speech. Subsequent reports (5, 20, 28, 32, 75, 81, 87, 88, 134) have been in agreement that chlorpromazine is effective in quieting or abolishing severe agitation and psychomotor excitement, whether of manic-depressive, schizophrenic, or toxic origin. Most of these reports agree that the basic disorder in these conditions is not
altered by the drug. In the psychoneuroses, chlorpromazine was reported (52, 80) to decrease anxiety and tension temporarily, but to have no specific ameliorative effects on conversion symptoms, obsession, piaobia, depression, or physical pain. A recent, controlled investigation (123) on the effect of chlorpromazine on the communication processes of psychiatric patients has indicated no specific facilitating effect. In normal subjects, the effect of 10 mg of proclorperazine (a phenothiazine derivative) was compared with that of 10 mg of phenobarbital. No specific differences were found on tests of mental performance, hearing, and pain perception, although a decrement in muscular coordination and efficiency occurred under proclorperazine (14).
Other phenothiazine derivatives are being extensively investigated at this time in psychiatric practice with the hope of finding one with equal or better therapeutic effects and fewer of the side effects of chlorpromazine, such as Parkinsonism, obstructive jaundice, dermatitis, tachycardia, etc.
There are no reports of the use of the phenothiazine derivatives in interrogation and no evidence to support the thesis that these drugs might be of avail to an interrogator's work, except perhaps with excited and agitated informants who might be encouraged to report material in exchange for peace of mind. On the other hand, such drugs might help a harried informant to keep his knowledge to himself
The principal rauwolfia alkaloids being used in psychiatric practice are reserpine (serpasil), deserpidine (harmonyl), and rescinnamine (moderil). Of these, reserpine has been most thoroughly studied. Its actions are discussed here as representative of the group.
Reserpine (serpasil). Reserpine, believed to be the most active of the rauwolfia alkaloids, was identified by Müller et al. (105), and its sedative and antihypertensive effects were noted in animals by Bein (10). Initially the drug was used in the United States for treating high blood pressure (133). From such experiences it was noted that reserpine produced a state of calmness without significant impairment of sensory acuity, muscular coordination, and alertness. The effects of single doses of reserpine (2.5 to 10.0 mg intramuscularly) were found (98) to be more marked in patients exhibiting high levels of psychomotor activity, regardless of the clinical psychiatric diagnosis. Repeated doses of reserpine over long periods of time reduced or
The tranquilizing action of reserpine and chlorpromazine appears to be quite similar. This similarity extends to some of the side effects, such as miosis, lowering of blood pressure and body temperature, increase in appetite, nasal congestion, and Parkinson syndrome. Although certain differences in side effects do occur, they do not require our attention here.
The literature on the effectiveness of reserpine and chlorpromazine as aids to individual (119) and group (34) psychotherapy is extensive. Whereas there is agreement that the drugs lower anxiety and tension, there are such notable differences of opinion about the aims of psychotherapy and about what constitutes psychotherapy that no definite statement can be made regarding this point. If the reviewer were to add his voice to the disharmonious chorus of viewpoints on this subject, he would, on the basis of impressionistic evaluations only, say that the tranquilizers are of some aid in the psychotherapy of those patients who are so agitated, anxious, and hyperactive that they cannot sit still very long for psychotherapy. However, working out the finer nuances of emotional conflicts of a patient in psychotherapy is precluded when the patient is regularly taking a tranquilizer, because the patient appears to become too insulated against emotional reactions to realize or care that he is responding with feeling in some pertinent way.
As with the phenothiazine derivatives, the place of the rauwolfia alkaloids in the potential armamentarium of the interrogation has not been established. There are no relevant articles on this matter. Speculatively, the tranquilizers might be of avail in selected informants who are highly agitated and disturbed, and who might yield information in return for the relaxation they experience with such a sedative. On the other hand, less emotionally disturbed informants might strengthen their resolve to retain information under a tranquilizer. The only way to decide this problem is by experiment. Furthermore, tranquilizers in moderate dosage do not notably impair intellectual and sensory functioning. Therefore, their use probably does not contribute to the distortion of factual information produced.
Such speculations assume that the need of an addict for a drug is so strong as to override many other values, including strong social proscriptions when these conflict with the satisfaction of drug-created needs.
One of the cardinal criteria of addiction is the occurrence of severe reactions when the drug is withdrawn. Withdrawal syndromes occur with opiates, barbiturates, and, recently, an animal study has made this claim for meprobamates (43). Beyond a minimal dosage and time period, the continuous usage of these drugs produces addiction in almost everyone (72). No demonstrable impairment of cognitive or psychomotor functions has been identified in subjects operating under the maintenance dosage to which they are habituated.
Under most of the opiates, the subject is likely to show a keen awareness of a limited segment of reality, a decrease in spontaneity and creativity, a decrease in suggestibility, and an increase in rigidity and compartmentalization of thinking. As with other drugs, the reaction varies widely from individual to individual 50, 131).
Information contributed by an addicted source is naturally suspect, since many addicts have gone to great length, fabrication of information being the least of them, to maintain their drug supplies. If a source became addicted as a sequel to the treatment of injuries, the ability of the interrogator to give or withhold the drug would give him a powerful hold on the source. It appears unlikely that this weapon is so unique as to lead an interrogator to create addiction deliberately. An interrogator who would be willing to produce addiction would not hesitate to employ more reliable and instantly effective means for inducing results as unpleasant as withdrawal symptoms. Since the initial reactions of most subjects to drugs of addiction are unpleasant, these drugs would not appear to have a role as positive motivators, except for subjects experiencing pain.
When the effects of a drug are not consonant with the subject's wishes, the extent to which the subject can successfully counteract these effects without the aid of other drugs becomes an important problem. Inasmuch as this is an unusual condition, little or no empirical information is available. Yet, it may be worthwhile to indulge in some speculation.
If the subject marshals his efforts to fight a drug, one may surmise that he has become alarmed about its effects on him, and that this alarm most likely is in the form of anxiety over losing control. Other instances of this type of anxiety in neuroses, psychoses, and cerebral insult have demonstrated that it feeds on itself. Specifically, the anxiety increases in something like a geometric progression whenever the source of concern is put to the test and adequate control is indeed found to be wanting, with the mounting anxiety itself contributing to further loss of control. Some persons more than others habitually use the mechanism of control and might presumably attempt to do so in this situation.
Small to moderate doses, although affecting the subject, might not alarm him, since the drug effects may be within the range of his ordinary experiences, and since none of the functions which form the basis for his sense of control may have been seriously impaired. Because it is difficult for most persons to succeed in their efforts to relax, the main resultant of any effort presumably is anxiety and arousal. If this is so, the effect of stimulants would be intensified, whereas the effect of sedatives might be counterbalanced to some extent. Here, the question of how massive a dosage of a sedative the subject could counteract would need to be considered. The phenothiazine tranquilizers might be expected to produce a sufficient lack of concern in the subject to prevent his attempting to undo their effect, or, more directly, preclude a state of arousal.
A distinction has been made between interviews carried out for psychotherapy and those to obtain factual information. Although there has been considerable speculation regarding the possible use of drugs for the latter purposes, open publications of serious research dealing directly with such cases are scant. The paucity of reported studies on the matter has obliged the reviewer to include related published material of psychopharmacologic studies. When extrapolations are made from published material of this sort, they are presented as hypotheses, and in every instance require testing and validation.
Apart from any applied research that governments may sponsor for improving interrogation or for aiding their own personnel to resist interrogation, the pursuit of various current scientific and medical interests will doubtless result in developing knowledge of drug action applicable to interrogation. The interest of scientists in employing drugs in research transcends an interest in drug effects, per se. Drugs constitute valuable tools for experimentation directed toward developing basic physiologic and psychologic knowledge, such as the study of neurophysiologic correlates of symbolic and psychodynamic processes. Work by scientists in such areas is also likely to increase knowledge of drugs which may be applicable to interrogation.
A large initial section of this report is devoted to a survey and discussion of the nonspecific effects of drugs and to the difficulties involved in discriminating these effects from the pharmacologic effects of the drugs used. The time spent in describing some of these nonspecific factors is needed to illustrate how the many variables involved complicate the problem of making a judgment regarding the present or potential usefulness of a drug for either therapeutic or intelligence purposes. This section has been included to point out some of the problems which require consideration in designing well-controlled studies in this area. The complexity inherent in psychopharmacologic research requires the integration of all levels of research on drug action: biochemical, neurophysiological and psychological. These problems are multiplied and prediction is lessened when the actions of drugs on living human beings are considered, rather than on isolated nerves, tissues, or animals of simpler neural structure. This reviewer has included only very few bibliographical references to work with animals, and yet a significant portion of excellent experimental, psychologic studies involve animals. This relative omission can be explained by the problem being one unique to human beings: the use of language symbols to communicate and interact with other human beings.
(placebos). Depending on the personality of the subject and the circumstances under which the placebo is administered, 30 to 50 per cent of individuals show or experience a reaction. Well-designed studies can distinguish the pharmacologic effect of a drug from the placebo effect. The possibility is raised that an interrogator might exploit the "placebo phenomenon" with a susceptible subject, instead of employing a pharmacologically active drug.
An examination of the literature demonstrates that the effects of drugs vary with the attitude and motivation of the person administering the medication and the person interviewing the informant. The sex and intelligence of the subject, the presence of mental or physical illness, the occurrence of biologic rhythms (e.g., mensis), state of nutrition (e.g., fasting or nonfasting), degree of fatigue, and experimental or enforced isolation have been found to affect the capacity to react and the reaction of individuals to testing procedures, with or without drugs. The method of sampling the verbal behavior of an individual under the influence of a drug, directive, nondirective, free-associative, etc., also determines the kinds of reactions observed. For these reasons, it is recommended that a variety of sampling methods be used in experimental studies.
When one examines the literature for experimental and clinical studies that bear directly on the use of drugs in interrogation procedures, one finds relatively few studies. Reports dealing with the validity of material extracted from reluctant informants, whether criminal suspects or experimental subjects, indicate that there is no "truth serum" which can force every informant to report all the information he has. Experimental and clinical evidence indicate that not only the inveterate criminal psychopath may lie or distort under the influence of a drug, but the relatively normal individual may, with many drugs, successfully disguise factual data. Less well-adjusted individuals, plagued by guilt and depression, or suggestible individuals, who are compliant and easily swayed, are more likely to make slips revealing withheld information. Even they may, at times, unconsciously distort information and present fantasies as facts. The anesthetic action of the drug, as in narcosis with barbiturates, can interfere with cerebral functioning and promote the presentation of fantasy material as fact, or otherwise alter the form of verbalizations to render them relatively unintelligible. It would be very difficult under these circumstances for an interrogator to tell when the verbal
content was turning from fact to fantasy, when the informant was simulating deep narcosis but actually falsifying, which of contrary stories told under narcosis was true, and when a lack of crucial information coining from a subject under a drug meant the informant had none to offer. To derive useful information from an interrogation in which drugs are employed, an interrogator would have to consider and weigh many important factors: the personality of the subject, the milieu, other sources of evidence, the rapport obtained, and the skill of the questioning. These and other factors affect the validity of information obtained from an informant under sedation. Analogous considerations apply to stimulants.
Advantages and limitations of a number of different types of pharmacologic agents as adjuncts to interrogation can be examined by reviewing clinical and experimental data from the works of psychiatrists, neurologists, psychologists, physiologists, and pharmacologists.
Barbiturates tend to increase contact and communication, decrease attention, decrease anxiety, decrease psychotic manifestations, and make the mood more appropriate and warmer. When combined with interview techniques that aim at arousing emotions, strong emotional reactions may be catalyzed for psychotherapeutic purposes. Barbiturates have been found helpful in detecting whether an individual is feigning knowledge of the English language and in getting mute catatonic schizophrenics and hysterical aphasics to talk. They are of no avail, however, in remedying the speech defects of true aphasics, even transiently. The use of barbiturates has helped to get more reliable estimates of intelligence and personality through psychological tests, particularly in emotionally upset individuals.
The use of various stimulant and antidepressive drugs has been explored, for diagnostic and therapeutic purposes in psychiatric practice, but not to any extent for interrogation. Amphetamine, pipradrol, methylphenidylacetate have in common the capacity to produce an outpouring of ideas, emotions, and memories. An injection of amphetamine following an intravenous barbiturate is said to provoke a striking onrush of talking and activity from psychiatric patients. Without adequately controlling his study, one author claims that methamphetamine produces such a strong urge to talk that the criminal who feigns amnesia or withholds vital information cannot control himself and thus gives himself away. Iproniazid, an antidepressive drug which is relatively slow and sometimes dramatic in its thera-
peutic effect, should be considered for experimentation. This drug, and similar, less toxic analogs which are being developed, might be considered for use in special instances. For example, informants suffering from chronic depression, whether due primarily to emotional factors, situational stress, or physical debilitation, might become very responsive after using a medication of this type. As a class, the stimulants probably present the most obvious exploitative potential for an interrogator.
The psychotomimetic and hallucinogenic drugs, mescaline and LSD-25, have been used largely to study the nature of psychotic conditions and, in a minor way, as an adjuvant in psychotherapy. The use of such drugs by an interrogator would tend to produce a state of anxiety or terror in most subjects, and promote perceptual distortions and psychotic disorientation. Their use could constitute a definite threat to most medically unsophisticated subjects, i.e., the threat of making the subject "crazy." Thus, they emphasize the unrestricted control of the source by the interrogator. When the subject is not under the influence of such drugs, vital information might be extracted as a price for ceasing further medication. An enlightened informant would not have to feel threatened, for the effect of these hallucinogenic agents is transient in normal individuals. The information given during the psychotic drug state would be difficult to assess, for it may be unrealistic and bizarre. On the other hand, from the informant's viewpoint, taking LSD-25, secreted on his person (it is effective in minute dosage), might offer him temporary protection against interrogation, for it is not likely that an interrogator would consider an individual in a psychotic state a reliable source.
The introduction of new drugs like tranquilizers that sedate but do not impair intellectual functioning in moderate dosage (e.g., phenothiazine derivatives and rauwolfia alkaloids) has caused a minor revolution in the psychiatric therapies of agitated psychotic conditions regardless of type or etiology. There is a possibility that these tranquilizers might be of use with selected informants who are highly agitated and disturbed, and who might give information they prefer to withhold in return for the tranquility they experience with such a sedative. Under the influence of this drug, the less emotionally upset informant might find that he can better master his anxieties and keep his resolve to remain silent. These are all speculations which require testing and experimentation.
nance dosage. The motivational effects of obtaining drug supplies, while extreme, are not of a different order for most subjects than those which the interrogator could produce by other more rapid means. The exploitation of addiction probably constitutes a threat to persons previously addicted, or to those who become addicted in the captivity situation as a sequel to other aspects of their treatment, rather than through the deliberate creation of addiction for exploitative purposes.
Another use to which interrogators might put drugs and placebos would involve their ability to absolve the subject of responsibility for his acts. The popular meaning of being "drugged" or "doped" implies that an individual in this state has lost control over his actions and that society will not hold him responsible for them. When the transmittal of information is likely to induce guilt in the source, the interviewer can forestall some of this reaction by the administration of a placebo or drug. In some cases, this will be all that is require4l to remove the barrier to information transmittal. In the avoidance-avoidance conflict between the source's guilt over yielding information and his anxieties over the possible consequences of noncooperation, the "inescapable" power of the drug or placebo serves to justify the source's actions to himself.
What are the over-all conclusions that can be drawn from this review and critical analysis of the use of pharmacologic agents in obtaining information? Are pharmacologic agents of any value to the interrogator in eliciting vital information? The answer is that drugs can operate as positive catalysts to productive interrogation. Combined with the many other stresses in captivity that an individual may be obliged to undergo, drugs can add to the factors aimed at weakening the resistance of the potential informant. However, for many reasons, the use of drugs by an interrogator is not certain to produce valid results. The effects of drugs depend to a large extent on the personality make-up and physical status of the informant and the kind of rapport that the interrogator is able to establish with him. Knowing the pharmacologic actions of a number of drugs, an interrogating team might choose that chemical agent which is most likely to be effective in view of the informant's personality, physical status, and the various stressful experiences he has already undergone. Even under the most favorable circumstances, the information obtained could be contaminated by fantasy, distortion, and untruth, especially when hallucinogenic or sedative drugs are employed.
Means are available to the informant faced with the prospect of being given a drug to loosen his tongue. The informant should know that a drug of itself cannot force him to tell the truth, although it may make him talkative, overemotional, mentally confused, or sleepy. He should also know that the effects of drugs are quite variable from individual to individual, and that those who may use drugs against him cannot predict with certainty what effects will occur in his particular case. To a victim of such attempts the imperfect predictability of many of the direct effects and side effects of any drug offers many opportunities for simulation. It is likely that most nonfatal drugs will have a transient, time-limited action rather than a permanent one. There is no need for the informant to become panicky at any bizarre or uncomfortable reactions he may experience, for these reactions will probably disappear. Instead of passively accepting the administration of a drug, without challenging the interrogator's right to apply such pressure, the informant should effectively delay it, and thus stall a possibly stressful interrogation under a drug. Finally, since the interrogator wants accurate and factual information, the informant can confound the interrogator as to what is fact and fiction by a number of means. He can simulate drowsiness, confusion, and disorientation early during the administration of the drug. He can revel in fantasies; the more lurid the better. He can tell contradictory stories. He can simulate a psychosis; or, if he cares to go so far, he can even induce a transient psychotic state by ingesting a small amount of LSD secreted on his person. By these devices, he can raise serious doubts in the interrogator's mind as to the reliability of the information given by him.
As a final suggestion, this reviewer is inclined to agree with West (130) that the basic training of military personnel can be helpful in developing techniques of resistance to interrogation. A brief course on the limitations of the use of drugs in interrogation and on the kinds of pharmacologic effects to be expected from the different types of drugs would be helpful. Such training could decrease the fear, hypersuggestibility, and other deleterious reactions that evolve from the uncertain, the unpredictable, and the unknown.
4. Azima H., Cramer-Azima, Fern J., and DeVerteuil R. A comparative behavior and psychodynamic study of reserpine and equally potent doses of raudixin in schizophrenics. Morristown, N. J.: The Squibb Institute for Medical Research. Monogr. Ther., 1956 I(2), 15-25.
31. Cleghorn R. A. Drugs that produce deviations in mood, including anxiety presumably without impairing capacities for orientation or at least secondarily to changes in mood. Amer. J. Psychiat., 1952, 108, 568-571.
56. Gottschalk L. A., Gleser G. C., Daniels R. S., and Block S. The speech patterns of schizophrenic patients: A method of assessing relative degree of personal disorganization and social alienation. J. nerv. ment. Dis., 1958, 127, 153-166.
62. Hinlkle L. E., Jr. In Group for the Advancement of Psychiatry, Methods of forceful indoctrination: Observations and interviews. New York: GAP Publications Office, July, 1957, p. 287 f. GAP Symposium No. 4.
84. Lasagna L., von J. M. Felsinger, and Beecher H. K. Drug induced mood changes in man. I. Observations on healthy subjects, chronically ill patients and "post addicts." J. Amer. med. Ass., 1955, 157, 1006-1020.
90. Liddell S. W., and Weil-Malherbe H. The effects of methedrine and of lysergic acid diethylamide on mental processes and on the blood adrenaline level. J. Neurol. Neurosurg. Psychiat., 1953, 16, 7-13.
Rubin M. A., Malamud W., and Hope J. M. Electroencephalogram and psychopathological manifestations in schizophrenia as influenced by drugs. Psychosom. Med., 1942, 4, 355-361.
128. Weinstein E. A., Kahn R. L., Sugarman L. A., and Malitz S. The serial administration of the "amytal test" for brain disease; its diagnostic and prognostic value. A.M.A. Arch. Neurol. Psychiat., 1954, 71, 217-226.
To almost any stimulus from outside the body responds with widespread changes in its physiologic functioning. Many of the changes are invisible and unknown to the person himself. In some situations, these changes have been found to be valuable for indications of the degree of credence which should be given to a person's statements. This chapter reviews the responses available for such interpretations, the validity of the method, and possible improvements and extensions of its use which may occur in the future.
The use of physiologic responses in police interrogation has become commonplace. Such practices, long based on supernatural principles, have in fact been used since ancient times (35). As long ago as the eighteenth century Daniel Defoe proposed a test of this sort with a scientific rationale (31). Actual experiment on physiologic tests of deception seems to have begun with psychologists in Germany early in the century, with Benussi (2), an Italian with German training, offering the most extensive and promising results. Benussi used breathing changes as his principal criterion. A few years later Marston (29) and Larson (23), on the basis of certain experimental work, reported success with systolic blood pressure changes. These two physiologic variables have been the core of commercial "lie detectors" ever since, with the addition of the galvanic skin response following the work of Summers (33, 34) in the 1930's.
also supported by the other services, a group at Indiana University undertook a comparison of variables and combinations of variables that was reported in 1952 (17). Meanwhile the use of the "standard" methods has spread widely as an applied art with a certain body of tradition. (Lee (26) gives an account of current practices.) Controlled experiments, however, on which to base actual practice have too often been lacking. An excellent survey of the current status of the field has recently been provided by Ferracuti (18) in Italian.
The present survey will be organized into the following topics: (a) Evaluation of present practices. (b) Response variables and instrumentation. (c) Interrogation procedures. (d) Data interpretation and operator's decisions. (e) Psychological principles of lie detection.
We may consider "current practices" en masse and ask how effective these have been. This problem is considered by Inbau (20), apparently by comparing "lie detector" results with jury verdicts and confessions. The agreement between detection and the criterion for various sets of data is about 70 per cent, with 20 per cent of the cases discarded as ambiguous.
This figure must be compared, of course, with some percentage of success to be expected by chance. If every case were treated independently, the percentage of success would be 50 per cent. It is, however, common practice to examine a group of suspected persons, of whom it is known that only one is guilty. If the operator then selects one from the group as guilty, his chances of being correct by sheer luck are less than 50 per cent. If, on the other hand, an operator knows, in a particular situation, that most of the examinees sent him are later judged guilty, his "percentage of success by chance" could be much higher.
It would be difficult to analyze field results in greater detail than in Inbau's review (20). The jury decision is an imperfect criterion, of course, and may not be independent of the lie detector results since a prosecutor might be inclined to bring to trial more cases in which the lie detector results were clear. If the test has been used widely for screening, as it is reported to be, many suspects with negative finding on the instrument would not be brought to court. Most of these would be true negatives, and the percentage of success might actually be higher if they were included. The percentage obtained will depend clearly on the group from which it is derived. Further-
more, there is no telling exactly what procedures are used by various examining officers nor just what in the instrument records (or possibly aside from them) influences the judgment as to whether the lie detector test indicates guilt. For determining which methods and conditions give the most valid results, and whether improvement is actually possible, we must turn to experimental comparisons. This experimental knowledge of factors which are likely to influence the outcome could then be used in future attempts to evaluate uses in the practical situation.
The en masse result leads to the conclusion that psychological methods of detection in the criminal interrogation situations do provide information although the amount is uncertain, and the present problem confronting users is to maximize the information.
Laboratory experiments (17, 23, 33) have generally reported greater percentages of successful detection than the figure given by Inbau for field results. He is, however, skeptical of results obtained from these laboratory experiments. The situations are different in many wayssome of the differences tending to favor the laboratory, some the field situation. Although the severity of consequences in the laboratory is much less, the lying is also likely to be of a simpler sort and conditions better controlled. With some of the very high percentages reported for the laboratory studies, there is some question that the criterion may be adjusted to maximize success on one particular set of data and thus cannot be expected to have general application.
At present instruments may be classed into three groups: (a) the traditional ones which have both laboratory and field use; (b) those which have been tried in the laboratory, in some cases incompletely; and (c) those which have possible value but have not been tested for lie detection.
by time of expiration, seems to have dropped out of sight in practice. Benussi's experimental report was that during lying the I/E ratio is increased because of a change in the form of the breathing curve. It is easy to see why the ratio has been neglected in practical work, for it is laborious to compute and the determining points in the breathing cycle are difficult to distinguish, especially in a record taken at the usual slow speed. In the "Indiana study an attempt to evaluate it was abandoned because measurements were so unreliable. Common practice seems to be to regard any marked disturbtnce of breathing as indicative of deception (24).
The technique of recording the breathing rhythm is rather simple. The common pneumatic system is open to criticism because of the nonproportionality introduced by the compressibility of air, the general inconvenience of keeping the system free from leaks, and the awkward readjustment when S throws the recorder off scale by a movement. A simple electrical pickup and recording system can be substituted (17).
The Indiana study considered two aspects of respiration: amplitude and breathing cycle time (the inverse of rate). In amplitude the response in truth telling was an increase, with the maximum 5 to 10 sec after he delivery of a question. In lying the mean response was also an increase, but in a lesser amount. The fact that a smaller increase in amplitude typically indicates deception requires an operator to make a sort of inverted interpretation on this point. Breathing cycle time increased in both conditions, but more in lying. Thus breathing during deception is shallower and slower than in truth telling. These facts are confirmed by the agreement of the two groups. There seems to be much better discrimination between the two conditions when these measures are used in a long series of questions; i.e., they discriminate poorly at first as compared to other measures, but in the long run are among the best discriminators. It may be that breathing in the early part of a series is made irregular by a reaction to the general situation. After some adaptation it becomes possible to compare the responses to questions in purer form. According to some later work (8) the inhibition of breathing seems rather characteristic of anticipation of a stimulus.
One drawback in the use of respiration as an indicator is its susceptibility to voluntary control. If an S wished to produce a confused record he could probably do so by alternating over and under breathing, if he could keep up this or another program in the face of questions. But this drawback is not present in breathing alone. If an examinee knows that changes in breathing will disturb all
physiologic variables under control of the autonomic division of the nervous system, and possibly even some others, a certain amount of cooperation or a certain degree of ignorance is required for lie detection by physiologic methods to work. Respiration, therefore, on balance in the present state of knowledge seems to be one of the better measures.
The systolic blood pressure was first used as a deception indicator by Marston (29) and the method in common use is the contribution of Larson (23). It is evidently the chief reliance of present field practitioners of detection. Inbau (20) and others write that blood pressure is the main channel for the deception reaction in a real situation, although galvanic skin response may have greater power in the laboratory. (It seems unlikely that this is a fundamental difference between the two measures. The manner in which they are used and the length of sitting may be involved. The point would seem to require experimental checking.)
The systolic pressure will typically rise by a few millimeters of mercury in response to a question, whether it is answered truthfully or not. The evidence is that the rise will generally be greater when S is lying than when telling the truth. In using this measure, the operator, consciously or unconsciously, uses some sort of cut-off to separate the two categories. Records are often presented [e.g., (26)] in which a large difference in response to neutral and critical questions makes the decision obvious. However, as might be expected, there are instances which are not so clear. The content of neutral questions will produce variations in the response, and one must then decide whether a response to a critical question is "positive" if it is larger than any other, or if it is larger than average by some amount. It is likely that the criterion is a somewhat variable one in ordinary practice.
The instrument currently in use consists of a pressure cuff similar to that used in medical practice, but equipped with a side branch tube which connects to a tambour through a pressure reducer. The method is to inflate the cuff (on the upper arm) to a point between systolic and diastolic pressure; that is, to about 100 mm of mercury. This may be adjusted from time to time to follow changes in S's blood pressure. Under these circumstances there is a flow of blood to the lower arm only during the upper half of the pulse wave, and there is practically no venous return from the arm since the cuff pressure far exceeds the pressure in the veins, and occludes them. The side branch from the cuff will convey pressure variations to the
The method has long been criticized (4) for its technical incorrectness. It does not give a true measure of systolic (or diastolic) pressure. This criticism has made little impression on those who use the method, since they can exclaim, with some justification, "But it works!" It is a more telling objection that the cuff disturbs the blood supply to the arm a great deal and that it produces many undesirable side effects (7). The practical stoppage of circulation can become, in the course of a sitting, quite painful, and in a long sitting, dangerous. Operators, who are aware of these consequences, release the pressure from time to time to restore circulation. The side effects are such as to produce reactions in the other autonomically controlled variables which one may be measuring, and even in the blood pressure itself.
The Indiana study used a different method, unfortunately also open to these objections to occluding the blood supply. By mechanical means, a steadily increasing pressure was applied to a cuff and the point of complete occlusion determined by means of a pulse detector on the lower arm. The experimental results confirm the opinion that it is one of the better indicators of deception. Again discrimination is poor (almost nil) in the early part of a sitting and improves to a high point later.
Recently the writer (7) investigated the requirements of continuous arterial oressure measurement, and proposed a "closed circuit" method which uses a strain gauge applied to an artery with very little pressure. This device is simple to construct and use and seems well suited to the recording of variations in arterial pressure, although it will not as now developed indicate the base level of pressure. It has been used in a number of tests and experiments to record reaction to stimuli of various sorts (questions, flashes of light, and warning and reaction signals in decision situations). Although it has not been tested in a detection situation, there is good reason to think that it will do at least as well as the occlusion or near occlusion methods.
In the 1930's Summers (33, 34) reported some rather spectacular laboratory results in the detection of deception with the galvanic skin response (GSR). With a certain type of situation he was able to detect lying better than 90 per cent of the time. Since this work, the use of the GSR has increased and apparatus for registering it for detection purposes is made commercially.
The GSR seems to be one of the most easily triggered responseseven slight stimuli will produce the decrease in the electrical resistance of the skin, reflecting activation of the sweat glands (an internal condition, rather than the secretion of sweat). Recovery, however, is typically slow in this variable, and in a routine examination the next question is likely to be introduced before recovery is complete. Partly because of this fact there is an adapting trend in the GSR; with stimuli repeated every few minutes the response gets smaller, other things being equal.
On the other hand, long term changes in skin resistance may have a certain significance. A decrease in resistance which persists for a long period might be more significant of deception than one which has a quick recovery. In any case there is reason to believe that the significance of a change is related to the base level obtaining before it begins (17).
Not all available instruments have a provision for readily determining base level and long persisting trends. The structure of a proper instrument is more complex for the GSR than for the first two physiologic variables, although for modern electronics it is simple. The resistance measuring principle seems most satisfactory; a constant current is passed through S, the I/R drop across him is measured, and its fluctuations recorded.
Such a circuit with a device for automatically setting the recording pen back on scale is described in the Indiana report. For satisfactory recording nonpolarizing electrodes are required, although some commercial suppliers seem to overlook this necessity.
In the Indiana study the GSR was the best of the indicators in a short series though its power of discrimination fell as adaptation progressed. The investigation was concerned, however, only with the short term decreases that follow questions with about a 2-sec latency. There may be still further power in the GSR when it is used differently. The interpretation of the response is certainly made difficult by the confounding adaptation trend, and an interview needs to be planned to allow for such a trend, results being evaluated with regard to it.
The rate, in the form of cycle time, was included in the comparison of the Indiana study. The technique was to use a somewhat faster paper speed and make actual measurements of the time occupied by a certain number of beats. Contrary to the usual expectation the predominant response to questions is a slowing of the rate, reaching a maximum after about 5 sec. In lying the heart rate slows down more than in truth telling. This response is in part also the one produced by loud noises (10), threats of shock (17), and many other types of stimuli not requiring considerable muscular movement. In comparison with the other variables of the comparative study the pulse rate variable discriminates moderately well. Apparently it suffers from the same handicap as GSR — with adaptation, it loses power.
To be interpreted immediately the rate would need to be recorded by a tachometer such as the one described in (13) or that manufactured by the Yellow Springs Instrument Company. These tachometers translate time into a lateral deflection of a recording pen. Since these instruments are operated by the electrocardiogram, they are a bit uncertain if S is not in a shielded room. They are more complex than the GSR instrument and might present problems in the hands of an inexperienced operator. Nevertheless, heart rate may turn out to be a very good adjunct in detection.
The volume pulse has been studied in a number of experiments in the Indiana laboratory in a variety of situations (8, 9 ,10) and it was included in the comparison made in the Indiana lie detection study. The physiologic function is the pulsatile change in the volume of some part such as the finger. The reaction to stimuli is typically a decrease in the amplitude of the pulse wave, which is a manifestation of constriction of the arterioles in the region. This reaction is produced by questions in an interrogation and is greater when S is lying than when telling the truth. Under certain circumstances in a moderately long series of questions the response differentiates well between truth and lying.
The function is measured by some type of plethysmograph. The electrical impedance plethysmograph has the considerable advantage of convenience in attachment. Such an instrument is moderately complex to build but fairly simple to operate.
The pressure pulse is distinguished from the volume pulse in method of recording; a pickup device with a very high resistance to movement is used. The effects recorded are variations of the blood pressure rather than local conditions at the site of pickup. The pulsatile variations are the difference between systolic and diastolic pressure.
A measure of this sort was tested in the Indiana study with rather poor results in detection of deception and more recent studies of other conditions (9) also suggest it is a rather unsatisfactory variable. Probably systolic and diastolic are correlated positively to a substantial degree, but as one or the other has a greater increase, the pressure pulse goes one way or the other.
This variable is in a different class from the previous ones in that it is under the influence of the central nervous system rather than the autonomic division (respiration is affected by both). Although muscular tension can be recorded by other means than electrical, the operating difficulties of the electrical method are less with suitable equipment and the records more interpretable. The EMG can be compared in several places if one desires.
For recording EMG an amplifier is needed, with sensitivity in about the range of the usual electroencephalograph. The output of this needs to be integrated in one way or another to make it easily read (17). The electrodes attached to S are simple metal disks.
The Indiana study included one experiment using EMG in lie detection, but this was carried out independently rather than as part of the battery in which various measures were compared. Results were extremely satisfactory so far as they went. However, the test must as yet be considered exploratory and needs repetition for statistical reasons.
Convenient, portable EMG apparatus is now available, though it is somewhat expensive, and with a little experience an operator could learn to run it. An important technical limitation is the necessity for having S in a shielded room (for example, grounded fly screen), which could be easily provided only in some central laboratory.
tion and thus confuse the record, although the strategy would probably not occur to him. Such a movement would also have the effect of disturbing all the usual measures of lie detection, and evidently is not a common occurrence. The use of the method has some interesting possibilities since conflict between two responses can be recorded (17).
Somewhat similar in operation to the EMG the ocular movements permit the study of a choice of responses. If two or more visual targets are provided the eyes may unconsciously turn toward one or another in response to a question. The method lends itself well to the map or picture exploration method (see the following).
A substudy of the Indiana lie detection research tested the power of the ocular movements as a detector with good results. At least under special experimental conditions good discrimination can be obtained. This method, however, has the handicap of being rather cumbersome and slow. It involves photographing the eye movements with art ophthalmograph and evaluating the record after it has been developed.
Speed of transmission of the pulse wave along an artery has been proposed by the group at Washington State University (5) as a possible indicator of deception. The variable measured is primarily the pressure increase in the artery that follows a heart systole and is propagated through the fluid in the artery in a manner fairly well described by a known equation. (This propagation of pressure wave is quite a different matter from the flow of blood.) If the pressure wave is picked up at two points a known distance apart, its velocity can be calculated. The Washington State group has worked on a device that would do this automatically.
The velocity, according to the equation, depends on the level of blood pressure, and it would seem that in the same individual, variations in pressure would be the principal source of variation in velocity. Consequently, the scheme would be an indirect measure of blood pressure.
a device would require a good deal of monitoring of sensitivity to give suitable readings. The scheme is also rather complex and seems to be an alternative to the simpler method of recording pressure described previously.
Recent instrumental developments (11, 12, 36) make possible the use of the gastrointestinal reactions as a means of detection. According to one method, gastric or intestinal activity is recorded from surface electrodes attached to S, one on the abdomen and one on the arm. Stimulus effects upon GI activity have been demonstrated (11), though much remains to be discovered about them. A probable handicap of the method is the extreme slowness of response and recovery in this sector; it might be necessary to space questions several minutes apart. Nevertheless, the variable might turn out to be highly discriminative.
The EEG is a possible response variable for use in detection although it has never been tested. The problem expected is that under interview conditions the alpha rhythm would probably be blocked most of the time, and there would be very little opportunity for it to exhibit the "arousal" or "alerting" reaction. In other words, the variable may be too sensitive to all sorts of stimuli and reach its maximum response, as it were, too readily, but no one can be certain this would happen without a trial. Equipment for recording the EEG is basically the same as that described for the EMG. An integrator is not generally used in connection with it, though such a transformation might in fact be useful.
A general effect that needs to be taken into account is the rule of adaptation. Almost all the response variables discussed are known to become less responsive with repeated stimulation, some at a greater,
some at a lesser rate (10). Beyond influencing the choice of variable this fact should also regulate procedure. Interrogation would be expected to become progressively less effective as it proceeds, and the diminishing returns would limit length of session: one cannot detect a difference between responses that are practically nonexistent. In other words, instrumental detection would not be expected to combine well with a "wear-them-down" procedure. Profitable sessions would then probably be an hour or less.
Furthermore, questioning or stimulation of other sorts before the instrumental session would undoubtedly also produce a deadening of response. For example, there is considerable transfer of adaptation from one stimulus to another in the GSR. Immediately following a period of highly disturbing events it is possible that an S might be induced to tell the truth, but the situation would be a poor one for instrumental methods, since S is already in such a high state of excitement that increments in the response variables would be small. This can be inferred from experimental results, but it would be well to have direct confirmation.
For similar reasons, possibly because of the same adaptation mechanism, a condition of fatigue or prolonged sleeplessness would be unfavorable in discriminating truth from falsehood. Physiologic reactions are likely to be reduced in such circumstances. This suggestion accords with the common experience of being "too tired to care one way or the other." The stimulus threshold is raised and the person eventually falls asleep — a state of relative indifference to all ordinary stimuli. By the same token one would expect alcohol and barbiturates, and perhaps tranquilizers, to be unfavorable to detection.
Because of these considerations there may actually be a contradiction between trying to secure an admission and detecting lying by instruments. For instrumental detection one needs an S with a lively autonomic (or sometimes central) nervous system, whereas fatigue might favor contradictions and admissions. It would seem that an examiner must determine whether he intends to use instrumental methods as a means of detection or merely as a stage property for intimidatin the subject.
The importance of a state of alertness in S is demonstrated by one of the Indiana studies. A visit to the Chicago Police Laboratories had brought to light the practice of convincing S of the power of the
instrument by "detecting" which card had been selected from a stacked deck. In a situation which also required a certain deception by the experimenter, the Indiana experiments compared the instrumental detections with and without prior demonstration of the effectiveness of the technique. There was a large margin in favor of the "no demonstration" procedure. Apparently when S is convinced that the instrument is infallible, he is resigned and ceases to be excited about the critical questions. A certain amount of contest in the situation seems favorable to detection. The experiment may be taken to show that even for the police officer, honesty is the best policy.
There is a very attractive possibility of weighing in advance the testimony of the instruments. If certain Ss are characteristically "detectable" when telling falsehoods and certain others are not, it should be possible to assign people in advance to the one class or the other. The practical advantage is clear; by deciding at the outset that certain cases are "inoperable" one's percentage of successful detection and confidence in his results for the remaining cases can be enormously increased. (See discussion in the Indiana report (17).) Exclusion of the inoperable cases would be especially advantageous if their number is small. The possibility of this classification depends on the consistency with which individuals respond differentially to critical and neutral questions.
Two experiments in the Indiana series were carried out to test that proposition, one being included in the report, the other being completed too late for inclusion. Both studies, based on the galvanic skin response, found a high degree of consistency among Ss, especially for those on whom detection failed on the first series of questions. The second set of questions also failed to detect deception. The results of the second study, on a larger group, confirm this finding. The implication is that using one or two pretests in which S is lying by instruction on known occasions would serve to distinguish persons susceptible to instrumental methods from those who are not. On the "susceptible" Ss, the certainty of the detection diagnosis will be greatly improved. One point needs further investigation: whether or not a pretest with one set of questions will pick out individuals who are susceptible to detection with a different set of questions. In the experiments mentioned the same questions were used for both pretest and test.
It is possible to query S without demanding replies from him at all, to require yes-no answers to appropriately framed questions, or to ask questions which require explanatory statements from S. Some experimental results (14) lead to the general proposition that if some overt response is required there are greater autonomic and muscular reactions to a stimulus. With larger responses one would expect differentiation between truth and falsehood to be easier. One experiment in the Indiana study confirmed this expectation for lie detection. Subjects who were required to reply "Yes" or "No" to questions gave more differential responses on the instrument (GSR). It also seems probable that requiring a strong verbal or motor response to questions would further increase Ss' recorded reactions and make them easier to differentiate. However, "explanatory" answers should probably be avoided for purposes of instrumental detection. Inasmuch as the overt response required does seem to influence the physiologic records, responses of uncontrolled length would tend to confuse the interpretation.
The possibility of an interviewer and subject having effects upon one another has been brought out in several studies in a psychiatric situation (15, 27). Physiologic reactions apparently occur in each in response to the other. Although no study has been made of such interaction in a police-type interrogation, the occurrence seems quite likely. In any such interview the manner of the operator while asking a question is probably subject to unconscious variation. If, in turn, the interrogator's manner is influenced by the S-if he gets angry or feels sympathetic, for example-the results could be extremely confused. Actually presentation of a set of questions on "flash cards" or in a "memory drum" device may be indicated.
Lie detection experiments have generally dealt with just one plan of questioning: the presentation of a series of supposedly neutral questions with certain critical ones imbedded in it at unannounced places. Responses to the two sorts of questions are evaluated as
though they were independent. Field workers, on the other hand, have developed a variety of ingenious plans (20, 26) which seem sometimes to be more effective. A radically different plan is to let S know when a critical question or group of questions is coming. This procedure is accomplished by going through a question series in the same order often enough so that S knows what to expect. A series of responses is then evaluated as a whole: deception being taken as indicated, on the later repetitions, when responses become progressively larger as the critical questions are approached and die out rapidly thereafter. In some variables there is, further, the possibility of observing a mounting base level as questions near the climax. For a situation other than lie detection, experiments (16, 32) have demonstrated the progressive increase in reaction to stimuli as a noxious question is approached. Conditioned responses are said to be formed to the preceding stimuli on the earlier runs through the series. As a method of detecting deception there is much to be said for this organized schedule of questions. Whereas the unexpected question will produce a brief response, with unidentifiable anticipation probably enlarging responses to neutral stimuli and confusing the issue, the organized schedule permits a detection based on a number of readings, in fact, on the pattern of responses in a whole series. So far, however, an exact method of evaluating the data is not at hand. It would be quite desirable to have an experimental evaluation of this method.
Whatever general scheme of questioning is used, there must be some regard to the adaptative process already described. Size of response to a question in nearly all variables is going to be affected by its position in a series. Current practice evidently recognizes the fact by avoiding the first position for critical questions. A position at the end of a series would be almost as unfavorable, since, other things being equal, the response to a stimulus in that position will be the smallest. Question series of one form or another need to be so planned that the adaptation trend can be discounted in the interpretation.
An operator in field work usually has neither the time nor the means to make a statistical analysis of his results or perhaps even to compute the averages of the several responses to all questions. Apparently the successful operator will learn to decide on the meaning
of his results by some rules of thumb which he has difficulty putting into words even to himself. To exactly what set of cues he is responding it is difficult to say, for apparently no one has made an analysis of the exact difference between records which are judged positive and those which are judged negative or inconclusive.
For the experimenter the problem is different, since he is obliged to say exactly how he has arrived at an answer. His procedure, therefore, will usually be, as in the Indiana studies, to settle upon some rather obvious aspect of response that can be definitely measured, and then find out how well he can do at detection with this information about each response. Naturally, he is discarding a good deal of information while he does this, information about other features of the response which might be supplementary or superior to that which he is using. For example, in most GSR work only the maximum amplitude of the response is considered. It may well be that the duration of the response also has a meaning (beyond its correlation with amplitude). In fact, an unreported portion of the Indiana study indicated that this was true. The field operator might allow for this feature intuitively along with other characteristics such as latency, doubleness or singleness of response, rate of rise, etc. Not that the experimenter could not study these things, but he must take them one by one, test them singly and according to various rules of combination and weighting, a laborious and lengthy process. His hope, of course, is that in the long run he will be able to tell the field operator just what he should take into account to secure maximum reliability of decision. Certain suggestions can be offered a priori and some from experimental evidence.
The logic of the detection task imposes certain requirements. Basically the assignment is one of differentiating two conditions, truth telling and lying, on the basis of readings which are correlated with them ("prediction" in the statistical sense of the word). The operator or experimenter must proceed by finding, for a given S, the mean response to the critical and to the neutral questions. The alternative, of simply comparing critical responses of an S to those of other persons, will not be satisfactory because Ss differ in their level of responsiveness to any stimulus. By representing the responses to critical questions by R o and those to neutral questions by R n, we may say the first quantity to be considered is R c — R n. We must then decide on the basis of data which sign of the result is indicative of lying, if either. We might, under some circumstances, then determine from data how many persons will be correctly classified when the difference is of a certain degree. The number of detections would
Evaluation of this sort would be premature under many circumstances. It may well be that the questions asked are not of equal stimulating power even to those giving true answers. A critical question dealing with a crime or other "sensitive" information would doubtless elicit a larger response than, say, a question about an inconsequential matter. (It is a routine GSR demonstration that words such as "mother," "sex," "hate," etc., produce more response than, for example, the names of common articles of furniture.) Hence it is necessary to complicate the comparison before the decision of truth or falsity can be made. The value Rc — Rn for each S should be compared with that for a group of Ss responding to the same questions. The function to be used is therefore (Rc — Rn)Subject1 — (Rc — Rn)Average Subject. This function the experimenter would then need to evaluate for its detecting power; the field operator needs to arrive at an approximation of this function to reach a proper decision. Having this, the field operator should ideally be provided with a table showing the probability of correct detection and the probability of error for each value of the function. Such tables do not exist at present.
For good discrimination it is essential that measures to be compared be quite reliable. Each recorded response to a question may be considered as partly determined by the question and partly by "accidents" of the environment and in S himself. For example, there may be some influence of a previous question, of the questioner's tone of voice, of even trivial events, noises, etc., in the examining room. Naturally, the first step in securing reliability would be the control of these extraneous factors. A good examining room should be provided where outside events are neither seen nor heard and where S cannot see the examiner nor the operation of the instruments. Despite these precautions there will still be "error" fluctuations beyond the operator's capacity to manipulate. These disturbing effects can be set against one another by the usual technique of repeating the observations and considering the average of the series.
When several physiologic functions are recorded, or several features of one kind of response are measured, there is the further problem of how they should be weighted and combined in making a "prediction." Several variables, rather than one, would be expected to provide a more accurate basis for decision. Aside from the usual advantages derived from multiple measures there is the fact of indi-
vidual specificity of response, demonstrated in a number of experiments (22, 28), which would make multiple measures particularly valuable. Individuals tend to react in one physiologic sector or another. One individual may "specialize" in a heart reaction, for example, and another in the GSR. Recording a number of physiologic variables gives, therefore, a better chance of locating each person's special kind of reaction.
The Indiana studies have imparted some knowledge of the methods of combination. In one investigation a large group of physiologic variables were recorded, more, of course, than would be practical in a field situation with the idea of comparing their effectiveness. "Discriminant analysis" was used in the main analysis of the results. This technique is based upon a computation of optimum weights to be assigned to the measures in order to give maximum discrimination when they are added together in standard score form. The weights derived from one group of Ss must then be tested on another before they can. be considered as firm. This elaborate method gave strange results in the Indiana study. First, by use of it on the original group, truth and lying were differentiated no better than they were by the best single measure. This result is rare with variables that correlate with a criterion and only poorly with each other. Second, the same weights applied to a second group did give a very substantial improvement indeed, whereas by the operation of random fluctuation one would expect the second group always to give worse results than the first. It may be that the second set of data was more reliable than the first, or fit the assumptions of the model better.
It is still possible that a set of weights suitable for transfer to field use could be derived in this way. A particular set of weights, of course, would have to be calculated for the particular grouping of response variables intended for field use. On the other hand, it seems likely that a different method of combination might have greater discriminating power.
In the Indiana group for which discriminant analysis improved detection but little, a simpler method gave very good results. It was found, for example, that merely counting the number of variables showing an indication of lying for each question formed a highly successful lying score. In this method one is avoiding the assumption of linear addition, and the good result of this sort of treatment suggests that it is well to avoid that assumption. There is also a great deal gained in simplicity.
This technique remains to be developed, but it is, in fact, the logical approach in view of response specificity. Suppose, for example, there is an S who responds in a highly favored sector, such as heart rate, and very little in another-GSR, vasoconstriction, and other variables. We may presume that each of the response variables is admixed with a certain amount of random fluctuation ("error"). For such a case the heart rate response could be highly differential of truth and lying, but the other kinds of response would be less than average and, if added to the heart rate response as standard scores, or more simply combined, they would tend to offset the positive heart rate indication. It might come about, therefore, that combining measures in the ordinary fashion would involve greater error proportionate to "true" score than a single measure, properly chosen for each person.
An effective plan for this situation would be to determine by experiment what alternative patterns of response are discriminative of truth and falsehood. A record could be examined for such patterns, or, possibly, the pattern to be expected could be discovered by preliminary trials and a record watched for an exaggeration of this particular S's special pattern.
In field use, it is quite possible that interpretations of this sort are actually made, though rather unsystematically; a large response in one physiologic variable may be counted heavily, and the neutral indications of other variables not allowed to offset it. The multiple record would be used in such cases simply for selecting the most revealing variable for a particular S when he has one.
The effectiveness of lie detection procedures is limited by a lack of knowledge of what psychological principles are involved in successful lie detection. There is little examination in the literature of the basic psychological principles at work. Particularly important is the question: "Just what general properties of a situation provoke the physiologic reactions which make lie detection possible?"
Prima facie it seems improbable that there is a special kind of response peculiar to lying. In the early days Marston (29) recognized that truth and falsity are not psychological categories. Obviously a person can give a false answer merely because he is misinformed. If the answer is true to the best of his knowledge and belief, one would certainly expect no physiologic signs of its falsity. Marston's contention was, therefore, that the physiologic reactions were present
only when there is consciousness of deception. If this is taken to mean that S must know that his statement is untrue, it is probably correct in most cases; that is to say that under some circumstances he would tell the truth. The possibility is not to be dismissed, however, that detection could be accomplished when S had completely suppressed memory of an event or distorted the essential features of it in his own thinking. Evidence of response under these conditions appears in the "subception" experiments (1, 19, 21, 25, 30).
On the other hand, it seems that trivial deception, even when fully known to S, would not provoke much physiologic reaction. A person probably can say he is feeling fine when he is not without the deception coming to light on the instruments.
It is questionable, therefore, whether the idea of "consciousness of deception" makes for an appreciable refinement. If it is true that deception is best with heightened awareness of it, the characteristics of a situation which will heighten that awareness require examination.
The physiologic response in lying, as found in experiments and field trials, is a pattern of changes in the recorded variables. Essentially the same pattern of response occurs when S is telling the truth under interrogation, and detection is possible only because the changes are greater, as a rule, during lying. If the responses are not specific to lying per se, or to the consciousness of lying, then knowledge of just what characteristics of a situation produce them is of first importance.
Three possibilities can be suggested: the conditioned response theory, the conflict theory, and the punishment theory. Each of these implies a somewhat different mode of operation in the detection situation.
According to the conditioned response theory the critical questions play the role of conditioned stimuli and evoke some "emotional" response with which they have been associated in the past. This principle is known to operate in situations other than lie detection. Conditioning of the GSR was demonstrated, for example, in the experiment of Diven (16), and it is a familiar fact that words such as those with sexual or other personal association evoke large responses. It would therefore be expected that questions relating to some fairly traumatic experience would produce especially large reactions. If this is the basis of detection, lies about trivial matters would be nearly impossible to detect. Asserting that a barn is a house, for example, would produce little response from the ordinary individual because neither word is connected with any large reaction on his past life. Denying that he took part in a crime might be
expected to produce a large reaction on this theory, because the crime probably produced a large "emotional" disturbance when it occurred. On the conditioning principle it would further be expected that the bodily reaction would be somewhat different, according to the kind of past experience the question was connected with. [Different stimulus situations do have a degree of specificity in the kind of reaction they produce according to certain evidence (6).] Whether, in fact, the physiologic response in lying differs according to the emotional reaction of S to a question is a problem deserving some investigation.
The simple conditioning theory can, however, hardly be the whole explanation of the lying reaction, for in laboratory experiments, such as some of those in the Indiana study, lying about rather trivial matters according to instruction did lead to enough differential reaction to yield a fairly good detection percentage. In fact, percentages of detection were so high as to suggest that lack of too great general stress is favorable to detection.
The theory of conflict, following the psychoanalytic lead, would presume that a specially large physiologic disturbance would occur when two incompatible reaction tendencies are aroused at the same time. Whether there is a greater disturbance than the sum of the two separate excitations is questionable (3), but at any rate the two would be greater than one. In the process of deception two reaction tendencies may be expected. Long habit would dispose the person to answer a critical question straightforwardly. On the other hand, when he is lying there are circumstances which arouse in him the tendency to denial. In the Indiana studies one experiment was based explicitly on this principle, but with the plan of distinguishing the two response tendencies by different sorts of muscular activity. The experiment gave good results, but not because it was possible to distinguish the two reaction tendencies. A better plan might have been to associate a "yes" answer with one hand and a "no" answer with the other. The purpose may be served, however, if the two response tendencies merely summate in the same place, and this could well be the mechanism by which the usual detection test works.
On the conflict hypothesis, both reaction tendencies would probably need to be strong for good results. Not much can be done about the tendency to reveal information openly. On the other hand, the tendency to deny it may be subject to some manipulation. This suggestion again leads to a paradoxical recommendation: the situation must be so ordered that S makes a strong effort to conceal the infor-
mation. This strategy, opposite to that which might encourage admissions, may in fact be favorable to instrumental detection. The experiment, already described, which showed better detection when S was encouraged to think he might "beat the instrument" lends itself to this interpretation.
If conflict is the basis of the large reactions that signify deception, then there is some danger of confusion with large reactions produced by strictly personal emotional problems. It is an established fact (see the preceding) that words touching on emotionally sensitized areas will produce large reactions, regardless of deception. A question touching on such an area might provoke a reaction greater than that produced by a mild conflict. Errors from this source would be checked by comparing the reactions of several persons to the same questions and, where possible, by comparing a person's reactions to sensitive questions when he may be lying with those he gives to the same subject matter when there is no occasion to lie.
A third possible basis of detection is the punishment, or better, threat-of-punishment principle. According to this idea a person will give a large physiologic response during lying because he anticipates serious consequences if he fails to deceive. In common language it might be that he fails to deceive the machine operator for the very reason that he fears he will fail. The "fear" would be the very reaction detected. More analytically put, S is giving a conditioned physiologic response to the operator's question because the content of the question has been associated with the possibility of unfavorable consequences. Lying is technically, then, an avoidance reaction with considerably less than 100 per cent chance of success, but it is the onl'y one with any chance of success at all. The physiologic reaction would be the consequence of an avoidance reaction which has a low probability of reinforcement, but not too low. If the theory has any validity at all it must be supposed that the physiologic reaction is associated with a state of uncertainty. It does seem that a lie told with a complete certainty of its acceptance would be unlikely to produce much reaction; and on the other hand we have the experimental evidence already mentioned that a lie told with no prospect of success whatever is also poorly detected. For good detection a situation may be necessary where S is willing to gamble on a rather long chance with some hope of success.
To make this punishment theory cover the experimental results one needs to take "punishment" in a broad sense, since in experiments S quite often suffers no serious loss if he is detected. He does, nevertheless, lose the game which he is playing and possibly this is
On this last type of theory successful detection would depend a good deal on S's attitude both toward the instrument and the situation as a whole, for good results would depend on S trying to "beat the game." If the hypothesis is correct, there would be some advantage to finding out what his attitude is and possibly encouraging him to engage in risk-taking behavior. Once again there seems to be all opposition between procedures designed to secure information and those that would lead to the best instrumental detection.
Present knowledge is not sufficient to lead to a decision on which, if any, of these three theories is correct. Since the theories here discussed are not mutually contradictory, it is quite possible that all the conditions referred to are actually operative in some degree in the detection situation. In that event detection would be best when critical questions are associated with somewhat traumatic past events, when S is threatened with possible but not certain punishment as a result of lying, and when critical questions, perhaps by reason of the uncertain consequences, arouse conflicting reactions in S.
Although direct, practical experience is lacking, some general findings of laboratory experiments are applicable. The relevance of many of the experiments for the criminal detection problem suffers from the fact that they involved no "crime." This makes them more pertinent for broader interpretations, however. Experiments concerned, for example, with discovering a person's birth month, when he has been told to say no to every question, are closer to the intelligence interrogation situation than to criminal detection. The concern in these experiments is with lying rather than with crime. From their success, we may conclude that crime is not essential for lie detection.
The intelligence interrogation, however, has certain peculiarities. Studies directed specifically to these distinctive problems would be required for more reliable conclusions regarding the applicability of findings from previous experimentation to practical employments in intelligence interrogations. One may suppose that the person questioned, typically, will have little personal involvement in information sought. The questions frequently will not be about something he has done or for which he feels responsible or guilty. He may or may not know what information is important to his interrogator. Perhaps he is not very deeply motivated to conceal the specific items or information, but loyalties and threatened penalties may dispose him
to do so. If the source regards the matter as unimportant, the motivational aspects of the situation would be rather like those in the common demonstration of detecting which card has been picked from a deck, a trick not difficult to do as a parlor game when a "lie detector" is available. However, if the source is highly motivated toward concealment and anticipates reprisals if he "breaks," the situation is rather like crime detection.
Special considerations also arise in the intelligence interrogation situation because of the kinds of people to be interrogated, their physiologic condition, their emotional state, and their attitudes. They differ from both the suspected criminals and the normal individuals or college students used in most experiments. The effect of factors like these is scarcely known for the groups already studied. It is terra incognita for the sources of factual interrogations.
A special condition for the intelligence situation is that the subject's cooperation in submitting to the instrumental recording might be difficult to obtain. One naturally speculates about the possibility of devising a few recording instruments that would need no attachment to S and might be concealed from him. Considering the complex problems attending overt electrodes and recorders, the information gained from hidden instruments is likely to be quite meager and unreliable. Furthermore, it is not certain that an S who is not aware of the process would actually respond in the same way as one who is. It would seem necessary that interrogators use the ordinary type of instrument and rely on persuasion or coercion to get subjects into it. There is still the possibility that sophisticated subjects would, under coercion, introduce confusion by moving about and controlling breathing. How often this would happen can only be determined through experience.
Nevertheless, on the basis of the facts known from laboratory and field work one might expect that the physiologic methods can be applied to intelligence interrogations with reasonable success. Most of the considerations already discussed would seem to apply.
In spite of the early scientific foundations of lie detection in the work of Benussi, Marston, Larson, and Summers (2, 22, 23, 29, 33, 34) there is at present a rather broad gap between current practice and
scientific knowledge. There is, on the one hand, some information from the laboratory, which could be applied, and there are procedures of questioning, developed in field work, which await experimental testing. Although variation in procedure and in selection of cases makes present field data quite difficult to evaluate, it does seem probable that a significant amount of detection is being secured by physiologic methods. Laboratory experiments generally confirm the success of the technique.
Laboratory science can make some immediate contributions to the improvement of detection methods. Developments have made possible better instrumentation for the recording and analysis of variables which currently figure in criminal detection, and suggest the possibility of recording various others which could increase the accuracy of detection. For some of these additional variables, experimental evidence is already available, others have yet to be tested.
Experiments have also yielded certain results that could be applied to interrogation procedures, of which the following are illustrative. The factor of adaption, differential to particular responses, could be allowed for systematically. The attitude of the examinee influences results considerably; they are better when he does not believe the instrument is infallible. Rather, clearer results are obtained when he believes he has a chance of winning the game. Pretesting of Ss in order to drop a few prospects from consideration would greatly improve the confidence users could have in results on those remaining. Making S take an active part by giving some sort of answer also favors detection. Feedback of the operator's visible reactions has an effect on the source's subsequent physiologic responses. Beyond these facts, experimental evidence bears on a number of other practical matters, such as the order and nature of questions. Detailed instruction in these matters to personnel who might become exposed to such manipulations will go far in frustrating any interrogator who seeks to make use of them.
With respect to the evaluation of results, experiments have done more to set the problem than to answer it. Various possibilities of statistical combinations and evaluations o responses have been tried, but the optimum method is not yet known.
Much could be learned from planned experimental studies of the psychological basis of detection.. Precisely what is it that makes the source's responses different in lying and telling the truth? The difference might depend on a conditioned autonomic response, ona conflict of response tendencies, or on the chance of successful avoidance of punishment, or some combination of these. Knowledge of
3. Bowles J. W. "Electromyographic factors in aircraft control: A muscular action potential study of conflict". Randolph Air Force Base, Texas: U. S. Air Force School of Aviation Medicine, 1956. Rep. no. 55-125.
17. Ellson D. G., Davis R. C., Saltzman I. J., and Burke C. J. A report of research on detection of deception. 1952. (Contract N6onr-18011 with Office of Naval Research.) Distributed by Department of Psychology, Indiana Univer., Bloomington, Indiana.
19. Goldiamond, I. Indication of perception. I. Subliminal perception, subception, unconscious perceptions: An analysis in terms of psycho-physical indicator methodology. Psychol. Bull., 1958, 55, 373-411.
Through the years some lay and professional people have considered hypnosis almost a magical means of influencing others, curative, mystical, bordering on the supernatural. This has been so largely because the phenomenon of hypnosis seems to allow for a high degree of control of the subject's behavior. However, over the years, too, evidence has been accruing to suggest that hypnosis is neither fraudulent as some have maintained nor is it so mysterious as to defy experimental analysis.
Because of the apparent control of behavior during hypnosis it has understandably been proposed as a tool for interrogation. This chapter aims to evaluate these proposals. There is an utter dearth of literature concerning the actual use of hypnosis in interrogation. Either this technique has never been used, or if it has, no one has chosen to discuss it in print. Despite fairly extensive conversations with experts from a variety of countries, the author has found no one who admits to familiarity with its use in interrogation. An approximation to) such usage, however, does exist in isolated instances with criminal suspects.________________
This report will first consider the potential use of hypnosis in the interrogation of captured personnel. Three separate issues are involved here: (a) Can hypnosis be induced under conditions of interrogation? (b) If this is possible, then can a subject be compelled to reveal information? (c) If information is so obtained, how accurate will it be?
The second section will consider proposals advanced for the defensive uses of hypnosis, the problem being the feasibility of protecting personnel from enemy interrogation. Three suggestions will be evaluated: (a) the use of posthypnotic suggestions to prevent subsequent trance-induction; (b) the use of posthypnotic suggestions to induce amnesia on capture for sensitive information; and (c) the use of posthypnotic suggestions to make captured personnel more resistant to stress.
Before discussing the possible use of hypnosis for interrogation, we should like to review briefly what is known about the nature of the state itself. Mere description of the subject's overt behavior is an inadequate definition of hypnosis. He is usually described as passive, apparently asleep, and responsive only to the hypnotist's words. It is true that in the absence of specific suggestions to the contrary the subject seems to be extremely passive and to become unusually dependent upon the hypnotist for direction. However, an individual in hypnosis may also appear to be fully awake. No reliable objective criteria have yet been developed which will unequivocally identify the hypnotic state. This is particularly true in regard to physiological criteria. In the absence of reliable objective criteria, it becomes necessary to describe hypnosis in terms of the subjective events which the hypnotized individual experiences. The cardinal characteristic of the state is that a potentiality exists for the subject's perception of reality to be distorted in accordance with the hypnotist's cues. This distortion may affect any and all modalities of perception in regard to both external and internal events. Although this distortion of reality may be extremely real to the subject and his
behavior appropriate to it, considerable evidence suggests that at some level the individual continues to remain aware of the world as it really exists. Another attribute of the hypnotic state is that the subject experiences it as discontinuous from his normal waking experience. In association with this, amnesia for the experience may spontaneously occur. Finally, the subject generally experiences some compulsion to comply with the hypnotist's requests, along with a striking disinclination even to wish resisting them.
It is inappropriate in this context to review in detail the many theories proposed to account for the clinical observations. We shall briefly consider some of the theoretical views most generally held, since their implications differ markedly regarding the degree to which the state increases the susceptibility of a person to purposeful influence.
Primarily of historical interest are the views of Mesmer (13) and his latex followers, who held that hypnosis, or the Mesmeric trance, results from a flow from the hypnotist to the subject of a force called animal magnetism. This view is important because it is the basis of the lingering lay opinion that hypnosis is in some way an overpowering of a weak mind by a superior intellect. There is no presentday investigator who would defend this position, and in fact it is contradicted by recent evidence.
Since the time of Braid (14) in 1843, the view has been widely held that hypnosis is a state of artificially induced sleep. More recently, Pavlov (56) proposed a similar view when he maintained that cortical inhibition, sleep, and hypnosis are essentially identical. This view is currently held throughout those parts of the world where Pavlovian theory is accepted as a creed. This position implies that hypnosis is a state characterized by a profound neurophysiological alteration and that the subject in trance is somehow passively compelled to respond when appropriate suggestions are given. To the American investigator there appears to be overwhelming experimental evidence against this view. For example, Bass (4) has shown that the patellar reflex, which disappears in sleep, is not diminished in hypnosis. Wells (78) et al. have demonstrated that all hypnotic phenomena can be elicited in a state that in no way resembles sleep, which would lead one to hypothesize that the sleeplike aspect of hypnosis is not intrinsic to the state itself but is rather a result of the suggestion that the subject go to sleep. Barker and Burgwin (3) have shown that the EEG changes characteristic of sleep do not occur in hypnosis, although a true sleep may be induced hypnotically. However, there are two Russian papers (50) which contradict these findings, claiming that the characteristic rhythm of hypnosis resem-
The view proposed by Janet, Prince, Sidis, Corot, etc., which was current at the turn of the century, maintains that hypnosis is a state of temporary dissociation analogous to that which occurs in hysteria. Although this position seems reasonable in view of the similarity of the two conditions, it tells us little about the actual nature of hypnosis. The implicit assumption of this theory-that hypnosis is a sign of pathology — is not generally accepted today.
The Nancy school, especially Bernheim (9), revolutionized thinking about the hypnotic state by introducing the concept of suggestion and suggestibility. This orientation has been supported most notably by Hull (32), who, in a major monograph on hypnosis, concluded that hypnosis is primarily a state of heightened suggestibility. These views focus upon a trait in the subject, suggestibility, which is heightened by hypnotic induction techniques. Hull also relates the phenomenon to a habit, insofar as it becomes increasingly easy for a subject to achieve a state of hypnosis once he has been able to do so. Although the concepts of suggestion and suggestibility provide a bridge between hypnosis and the normal waking state, they do not offer explanations of the causes of the state or of the ongoing processes of hypnosis.
Welch (77) has attempted to explain hypnosis and its induction by an ingenious application of conditioning theory, utilizing the concept of abstract conditioning. He has pointed out that trance induction proceeds from suggestions which are almost certain to take effect to those that are more likely to be resisted. Several suggestions for experimental testing of this theory have never been followed up.
In contrast to the foregoing views, which focus either on the hypnotist or on some trait of the subject, several more recent approaches have been concerned with the interaction between the subject and the hypnotist. Schilder (63), White (83), and Sarbin (61) have all in one way or another emphasized the social relationship which exists in the hypnotic situation and especially the needs of the subject in this context. Also, Kubie and Margolin (40) and Milton Erickson (20) have concentrated on the subject's psychodynamics as being most relevant to the induction of hypnosis. White's view (83) is perhaps the first major formulation of this kind, and it represents a major departure in thinking about the trance state. He emphasizes that hypnosis takes place because the subject wishes to play the role of the hypnotized subject as currently defined by the subject and the hypnotist. It should be noted that the concern is with the subject's
wish to be hypnotized, and this motivation is considered of primary importrace to the induction of hypnosis. All the theories of this group, which might be called the "motivational theories" of hypnosis, emphasize the subject's wish to be in a hypnotic trance. Although other concepts are of necessity evoked to explain various phenomena in hypnosis, the actual occurrence of the trance state is related to the wish of the subject to enter hypnosis. This writer is a proponent of this approach, and the critical comments in this report are undoubtedly colored by this viewpoint.
It is important to recognize that almost no experimental work has been done that would support the validity of these various theoretical views, although there is some evidence already mentioned which tends to refute some of them. The general acceptance of the motivational view is based on the clinical impression of both experimentalists and clinicians that it accounts best for the major portion of the clinical data.
Trance is commonly induced in situations where the subject is motivated a priori to cooperate with the hypnotist, for example, to obtain relief from suffering, to contribute to a scientific study, or (as in a stage performance) to become, temporarily at least, the center of attraction. Almost all the currently available knowledge about hypnosis has been derived from these situations, and it is well to keep in mind the source of these data when one attempts to evaluate the possible utility of hypnosis in situations differing from these.
There is a small body of evidence stemming from the criminal cases in which hypnosis has allegedly played a role, which are radically different from those where hypnosis is normally observed. Because these situations may be more relevant to the questions of hypnosis in interrogation, this body of knowledge deserves particular attention and is discussed subsequently.
The initial problem in utilizing hypnosis for interrogation is to induce trance. It is to be expected that if the subject wishes to withhold information he will not 'wish to enter hypnosis. Therefore, hypnosis must either be induced against the subject's will or without his awareness. A common conception of hypnosis holds that it may be induced without any prior relationship between subject and
hypnotist and regardless of the subject's needs in the situation, with only the hypnotist suddenly gazing at his victim and commanding him to fall asleep. A motivational view of hypnosis would hold that trance induction depends upon the subject's needs of the moment and his expectation that the hypnotic relationship is to fulfill them. In this section we will evaluate trance induction procedures from the viewpoint of their dependence upon a positive relationship between subject and hypnotist and the subject's wish to cooperate.
There are three situations in which hypnosis has been reported to have been induced without the subject's awareness. In the first, hypnosis is induced while the subject is asleep. Another arises when the subject is seeking psychiatric help and hypnosis is induced in the course of a clinical interview with no explicit mention of the process. The third situation involves a trance spontaneously entered by individuals who are observing trance induction in another subject.
1. Sleep. The older literature is replete with statements that hypnosis may readily be induced by giving suggestions to sleeping subjects in a low but insistent voice; the subject becomes gradually more responsive to the suggestions until eventually he enters a somnambulistic state of hypnosis [ Bernheim (9), Braid (14), Binet and Fere (12), etc.]. Unfortunately, there are no cases given to support these statements. As so often the case in hypnosis literature, the statements appear to have been carried over from one textbook to another without any critical evaluation.
In a recent study by Theodore X. Barber (2) sleeping subjects were requested to perform standard hypnotic tasks. He found considerable similarity between compliance to suggestions given during sleep and reactions to customary hypnotic techniques. It should be pointed out that, in his study, Barber requested permission from the subjects to enter their rooms at night and talk to them in their sleep. Several of them remarked that this was hypnosis, and one may reasonably assume that most, if not all, of the subjects perceived that trance induction was the purpose of the study. This study, therefore, tells us little about what would happen if a truly naive sleeping subject were exposed to such a situation. No investigation is available on this point. Casual experimentation by the author failed to demonstrate the feasibility of this technique. The sample consisted of only four subjects, three of whom awakened to ask belligerently what was taking place, whereas the fourth continued to sleep.
Perhaps as in sleep learning (which seems to be effective only in a twilight state), response to suggestion may be obtained only in a receptive subject who has agreed to participate in the study and who is neither deeply asleep nor fully awake. Whether any increase in suggestibility over the normal waking state occurs has never been established.
2. Rhythmic, Repetitive Stimulation. Many trance induction techniques utilize the fixation of the subject's attention on a rhythmic, repetitive stimulus. Thus, metronomes, rotating spirals, mirrors, and swinging pendulums (75) [even the subject's own breathing (39)] have been used to induce hypnosis. In another context, the trance phenomena seen among primitive people frequently occur in ceremonies involving prolonged stimulation by rhythmic drums. Many authors have emphasized the importance of monotonous rhythmic verbal suggestions, especially during the induction stage of hypnosis. Recently, Kroger and Schneider (38) have proposed the use of an electronic aid which gives a repetitive signal approximating the alpha range of ten cycles per second as an adjunct. It is not clear whether these techniques directly facilitate hypnosis or whether they tend to produce a state of drowsiness that is interpreted by the subject as "I'm responding to hypnosis" which, in turn, facilitates further responses to suggestions. Certainly, the use of such techniques or even of monotonous rhythmic speech is by no means necessary in order to induce hypnosis.
All sophisticated discussions of hypnotic trance induction recognize that a successful response to a suggestion will facilitate further successful responses to suggestions. Even early descriptions of eye fixation advise the hypnotist to wait until the subject begins to show signs of fatigue and only then begin to give suggestions to the effect that the subject's eyes are growing heavy. Ideally, the hypnotist times these suggestions to occur immediately preceding the time when the subject begins to experience heaviness. Thus he takes the credit for having induced the state of drowsiness that is an inevitable consequence of eye fixation. Mechanical aids of this type may facilitate induction only to the extent that they bring about an event that is attributed to the suggestive effect of the hypnotist. However, it is also possible, as some of the proponents of these techniques suggest, that a neurophysiological basis exists for the facilitation of hypnosis. In this context it is relevant that road hypnosis and the break-off phenomenon encountered by pilots occurs in individuals subjected to peculiar types of repetitive, rhythmic stimulation despite a high
motivation to retain alertness. An intriguing question on which no evidence exists is the relationship of hypnotizability and susceptibility to road hypnosis or the break-off phenomenon. Furthermore, in the context of this discussion, the utilization of rhythmic stimuli as aids to trance induction is particularly relevant insofar as being subjected to such stimulation does not require the individual's cooperation. Whether an actual relationship exists between the drowsiness which can thus be induced and hypnosis is highly questionable and remains to be investigated. What is a somewhat more likely possibility is that drowsiness may be induced even in the uncooperative subject which may be attributed to some hypnotic influences. This would then tend to make the subject more liable to respond to other suggestions. Clearly, it is an area that might fruitfully be explored. No investigation utilizing such procedures in recalcitrant subjects has been made. In a later section on "magic room" techniques, the implications of using this and related tools are explored.
3. In a Therapeutic Relationship. Studies by Adler and Secunda (1), Sargant and Fraser (62), Schneck (65), and Rosen (59) have used techniques of trance induction which were aimed at preventing the subject from knowing that he was being hypnotized. These techniques all depended upon the subject's desire to obtain help with his problems from a therapist. It is frequently possible to utilize the therapeutic situation in such a manner as to achieve a hypnotic state eventually. For example, the therapist may talk to the patient about relaxing, and the virtues of relaxing, or the virtues of concentrating, thus obtaining his fixation on one particular object. He may suggest that the patient will be more comfortable if he closes his eyes, that in this way the patient can relax more or concentrate better. Thus, in a suitable subject a deep level of hypnotic trance can be achieved in a relatively brief period of time without ever using the term hypnosis and without the subject ever being aware that hypnosis is taking place. Meares (46) uses the neurological examination in this fashion as a test for hypnotizability.
In all the instances cited it must be emphasized that although the subject does not explicitly consent to enter hypnosis, a relationship of trust and confidence exists in which the subject has reason to expect help from the hypnotist. Furthermore, the hypnotist is an individual of high reputation and high prestige and there is some legitimacy in the subject's expectations. Standard medical practice includes many maneuvers by the physician which are essentially meaningless rituals to the average patient, and to which the patient
complies without hesitation because it is assumed by him that this will eventually benefit him. These situations, despite their outward similarity, differ greatly from those where trance induction is attempted by a stranger, without the subject's knowledge or consent.
4. Spontaneous Trance. Subjects who observe hypnosis in a demonstration may spontaneously enter trance. An experience of the author's concerning a psychotherapy patient with whom hypnosis had been used may be cited as an example. The author appeared on an educational television program where he demonstrated various hypnotic phenomena with several subjects. The patient watched the program in a friend's home. She reported that when the author induced trance in the subjects, she went into a trance, coming out of it when the author terminated trance in the television subjects. Spontaneous hypnosis occurred despite the fact that its appearance was a source of embarrassment to the patient since she was in the company of friends.
It is fairly easy to maximize the probability of this occurrence by mentioning the possibility of this phenomenon and conveying one's expectation that this may happen. Here again we are dealing with subjects who are essentially in sympathy with the purposes of trance induction in a situation which is viewed as safe by the individual entering trance. Again, no conclusions can be drawn as to the feasibility of inducing trance empathically in a subject who does not wish to enter trance. It has been rioted clinically that individuals who have negative attitudes about hypnosis do not enter hypnosis under these circumstances. White's (86) study, in which he has demonstrated that subject's attitudes about hypnosis, as shown on the TAT, are predictive of their hypnotizability, is relevant here.
We have been able to uncover only three studies that experimentally test whether a subject can resist the induction of hypnosis. Wells (80) instructed his subject to fight actively against trance induction — the subject was unable to resist. It should be mentioned that this subject had been previously hypnotized by Wells. This study was replicated by Brenman (16) who arrived at the same conclusions.
An even more dramatic experiment is reported by Watkins (74), again dealing with a subject who had previously been hypnotized by the experimenter. A nurse, who was known as a good subject, voiced the opinion to Watkins that under no circumstances could she be hypnotized against her will. He took the challenge and they set up
an experiment. Another nurse and a female psychiatrist were asked to witness the experiment. A dollar bill was placed in front of the subject and she was told that she could keep it if she did not enter trance. However, Watkins is careful to point out that this was a matter of prestige, not of monetary remuneration. Since no restrictions were placed on the subject, she closed her eyes, plugged her ears, talked and shouted. Watkins, speaking close to her ear, suggested that she would feel a pain in her head which would grow stronger and stronger, and that the only relief she would find would be to enter a deep sleep. The subject paused at times, removed her fingers from her ears to hold her head, and said that her head hurt. After six minutes she stopped shouting, tossed the dollar bill at the experimenter, and said, "Here, take it," and went into trance.
In determining the significance of these experiments, we feel that the "demand characteristics" of the situation are relevant. Demand characteristics are defined as those aspects of the experimental situation which implicitly convey the hypothesis of the experimenter to the subject. The author, in another publication (52), has shown that the demand characteristics of an experimental situation may greatly influence a subject's hypnotic behavior. It is clear that at some level a cooperative subject wishes an experiment to "work out," i.e., to help fulfill the experimenter's expectations. If a subject grasps the purpose of the experiment and/or the bias of the experimenter, he is disposed toward producing behavior which will confirm the experimenter's hypothesis. This is particularly true in a hypnotic relationship.
In all three studies, the subject had previous trance experiences with the hypnotist, which, we may assume, initiated a positive relationship between the subject and hypnotist. Although the subject was instructed to resist entering hypnosis, it was in the context of participating in an experiment to test this issue. It seems possible that in all three cases the subject was responding as if the experimenter were implicitly asking the subject to collaborate with him in order to demonstrate that trance could be induced despite the subject's resistance. The subject's motivation in this situation may be conceptualized as: (a) the overt attitude of resistance requested during the experiment and (b) the more fundamental attitude of cooperation to show that trance can be induced against a subject's will. In our view, the latter attitude was more relevant in determining the subject's behavior.
the possibility of inducing trance in a resistant subject who has been previously hypnotized. An experimental situation designed to test this question would have to take two variables into account: (a) the usually positive relationship between subject and hypnotist and (b) the demand characteristics of the situation. These two factors are necessary since in the setting of interrogation the aims of subject and hypnotist are apt to be at variance. One possible experimental design might involve two experimenters: one with whom the subject has a positive relationship, and the hypnotist with whom he does not. It should somehow be conveyed to the subject that the experimenter with whom he has the positive relationship believes (or hypothesizes) that the subject will be able to refrain from entering trance. Under these circumstances, we hypothesize that the hypnotist will be unable to induce trance in the resisting subject. We further assume that if the hypnotist is able to create a positive relationship, he would then be successful. In other words, whether a subject will or will not enter trance depends upon his relationship with the hypnotist rather than upon the technical procedure of trance induction. Admittedly, these predictions are based on extremely subtle phenomena. A test of these hypotheses would necessitate observers trained in evaluating nuances of feelings in order to be able to judge the nature of the relationship between subject and hypnotist. It is imperative that this factor be controlled if we are to draw any valid conclusions about an interrogation situation, since a positive relationship may come into existence only after it has been carefully nurtured.
The same kind of situation could be utilized in studying the induction of trance in resistant subjects who have never before been hypnotized. No data are available on this question. However, the problem is identical to the one discussed above except, perhaps, that additional resistances would be encountered.
In summarizing the evidence we are led to the conclusion that despite many apparent indications that hypnosis can be induced without the subject's knowledge or consent, all these situations seem to depend upon a positive relationship between subject and hypnotist. The most favorable of these situations occur when the subject (a) expects to derive benefit from his association with the hypnotist and (b) has trust and confidence in the hypnotist's ability to help. No reliable evidence exists that hypnosis can be induced directly from sleep in an unaware subject, nor is there good evidence that a subject is unable to resist trance induction if thoroughly motivated
to do so. An unexplored area relevant to this problem is the relationship of rhythmic stimulation and environmentally induced states of fatigue to suggestibility. It is also suggested that the question of whether hypnosis can be induced against the subject's will can be tested only by experiments that control the relationship between subject and hypnotist. In a study that utilized a hypnotist unknown to the subject and where the structure of the total situation was clear to the subject that it was desired and expected that he be able to resist hypnosis, current theory and clinical data lead us to expect negative results. No studies of this kind have been done, however.
Assuming an interrogator were able to circumvent the technical obstacles and induce hypnosis in a subject who wants to withhold information, to what extent would the subject remain master of his fate, even in deep trance? This is an area where wide disagreements prevail among authorities and where experimental evidence is highly contradictory.
Throughout this discussion no differentiation will be made between behavior that results from direct suggestion and that induced posthypnotically. Erickson and Erickson (21) maintain that posthypnotic behavior is performed in a self-limited hypnotic state. All phenomena elicited by means of posthypnotic suggestions may also be seen in trance, although the reverse is not always true. In line with Erickson and Erickson, we feel that the subject carrying out posthypnotic suggestions is in an hypnotic trance state, although at times a less intense one. The difference between the two states, if any, seems to be a difference in degree rather than kind.
Young (84) reports that subjects resist specific hypnotic suggestions if they have decided in advance to do so. Wells (79), on the other hand, reports contrary data. He found that none of his subjects was able to resist the predicted command or, indeed, any other. This contradiction exemplifies the controversial nature of the question of behavioral control in hypnosis. The problem has generally focused on the more specific question of whether a person can be induced through hypnosis to violate major social prohibitions which he has internalized or to commit some self-destructive act. It is the usual practice to use the term "antisocial acts" to refer to such behavior, but in this chapter terms more descriptive of the subjective significance of the act for the person are preferred.
The question is usually phrased in terms of whether an individual will commit antisocial or self-destructive acts in response to direct suggestion. Behavior considered to be antisocial is that which is so defined by the culture in which the individual has been raised. However, the question is complicated by the fact that some behavior is defined as antisocial in one context and as socially required in another, for example, murder vs. the soldier's obligation to fight. One of the major research difficulties is that some behaviors are considered taboo under normal circumstances, whereas they are felt to be legitimized in an experimental setting. The extent to which behavior is legitimized in this manner will depend largely on the subjects orientation both to the behavior in question and toward experimentation. All the material in the following discussion must be viewed in terms of the major difficulty of creating a situation which is contrary to the individual's internalized controls and which cannot be legitimized by the situation in which it is tested. Heron (31), Kline (35), Marcuse (43), Weitzenhoffer (76), etc., have discussed the problem of definition in evaluating the literature on hypnotically induced antisocial behavior.
The early view in this controversy over the elicitation of "antisocial" behavior, which answered the question in the negative, had been generally accepted until recently. Still, such classic authors as Forel (23) and Moll (48) believe that hypnosis is potentially capable of allowing sexual assault.
Supporting the negative view is the classic experiment reported by Janet (34). He asked a deeply hypnotized female before a distinguished group of judges and magistrates to stab people with rubber daggers, to poison them with sugar tablets, and in this fashion to commit several "murders," all of which she did without hesitation. As the company dispersed, the subject was left in charge of some of the younger assistants who, intending to end the experiments on a lighter note, suggested to the subject that she was alone and would undress. This promptly caused her to awaken. It should be noted that the "murders" were committed in such a way as to be play acted, whereas undressing would have certainly been real to the subject. In this classic instance, at least, she had no difficulty in discerning the difference. If, then, hypnotic subjects do not lose contact with the
Wells (80), in an experimental demonstration, induced a subject by means of a posthypnotic suggestion to take a dollar bill from the hypnotist's coat which was hanging on the wall and to accept it as his own money. Thus, in effect, the subject stole a dollar bill. The subject was unaware of this "crime" and denied vehemently that he had committed it. Wells maintains that failures to induce a subject to commit certain acts do not negate this possibility since the subject may not have been hypnotized deeply enough or improper techniques may have been used; whereas even one success demonstrates the possibility of achieving this result.
Brenman (16) conducted a series of experiments involving minor aberrant and self-injurious acts. Thus, in repeating the Wells study, she had a subject remember falsely that she had taken $2 instead of $1. The subject returned the $2 to the experimenter. She also was able to induce one subject to go through other people's pocketbooks, and to have another subject insult an acquaintance.
Schneck and Watkins in two separate reports cite evidence that behavior ordinarily constituting a crime can be produced by hypnosis. Both these reports deal with military situations. Schneck (64) inadvertently caused a soldier to commit a military offense by carrying out a posthypnotic suggestion and thus deserting his duty. It must be remarked, however, that Schneck himself was a medical officer in the army at the time he was conducting this experiment. Although the soldier may have neglected his duty, it was implicitly at the order of the medical officer and Schneck later made certain that no harm came to the soldier because of his military offense.
Watkins' (73) experiments, also conducted in this setting, are of particular relevance to us. Watkins induced a soldier to strike a superior officer by suggesting that the officer was a Japanese soldier and, according to the report, the soldier had to be restrained from inflicting serious injury to his officer. In another more relevant instance, Watkins was able to obtain information from a WAC under hypnosis which she had previously said she would not reveal and which was classified SECRET. The experimental demonstration took place before a professional group. Before the induction of hypnosis the WAC was asked how she would respond to interrogation by the enemy; she replied that she would reveal only her name and serial number. The hypnotist asked the subject to pretend that he was a German military intelligence officer and then proceeded to induct
trance. When the WAC was in hypnosis he represented himself as her First Sergeant and proceeded to question her about classified matters, She answered all of his queries, whereupon an officer stopped the proceedings "in the interest of military security."
Although these demonstrations appear convincing, they are open to the criticism that Watkins was an Army officer and although offenses were apparently committed, no serious damage could possibly result in this setting. At some level, at least, the individuals in question must have been aware of this fact.
A different type of experimental situation was constructed by Rowland (60) and also by Young (86). Two experiments were performed; one required that the subject throw acid at a research assistant, the other that he pick up a rattlesnake. Rowland's original experiments employed an invisible glass which protected the research assistant from the acid, and an invisible wire screen which prevented the subjects from picking up the rattlesnake. Although there were only two subjects in each experiment, all four carried out the hypnotist's commands. No attempt was made to conceal the fact that, in one case, this was a highly corrosive acid, and in the other, that this was a poisonous snake.
Young (86) slightly changed the conditions of the experiment by using a harmless snake which looked almost identical with a water moccasin and replacing the acid with tinted water while the subject was not looking, thus obviating the need for screens or invisible glass which might be perceived by the subject. The similarity of the colored water to the acid was dramatically shown by the fact that in one instance the experimenters themselves became confused and acid was thrown at the research assistant, necessitating the immediate use of first aid. Again the subjects performed both the homicidal and the self-destructive acts in the laboratory. Both experimenters report that normal control subjects in the waking state refused to pick up the rattlesnake or throw the acid when requested to do so.
Although these experiments seem to be extremely convincing, we must take into account the setting in which they were conducted. All the situations were clearly experimental ones, and were perceived as such by the subjects. The hypnotists who request the homicidal or self-destructive behavior are known to the subjects as reputable men. It is highly probable that the subjects, at some level, were convinced that in the experimental situation no serious harm would be permitted to come to anyone. This kind of situation is similar to that of a stage magician who asks a volunteer from the audience to cut off some individual's head with a guillotine which has been
convincingly demonstrated. Under these conditions volunteers from the audience will readily trip the appropriate lever. This could be be construed to be a homicidal act were it not for the fact that the volunteer from the audience knows full well that some kind of trick is operating that will prevent any harm from occurring, even though he cannot see the mechanism of the trick or know how it works. The question may be raised why control subjects in the waking state refused to perform these acts. One wonders whether the expectation that they ought not to do this was somehow communicated to them. Ways in which these objections might be met experimentally are discussed later.
By far the most sophisticated attempt to deal with this problem of the possible recognition of the situation as unreal has been undertaken by Kline (35). Unfortunately, only one subject was involved. He performed an antisocial act, however, which was "not only antisocial but punishable by law." Furthermore, while the subject had agreed to participate in a study to test the legal implications of hypnosis, the act was undertaken in a setting outside of the laboratory which was, to all intents and purposes, "real" (personal communication). The act, which is not detailed in the paper "for reasons of legality and recognizability," was clearly opposed to the internalized inhibitions of the subject. By most reasonable normative criteria, it would be viewed as highly objectionable. Four experimenters, competent hypnotists, failed in their attempts to induce the subject to perform the act. When the subject's perception of the reality situation was altered, however, he was willing to perform the action for three of the four experimenters. The experimenter for whom he refused revealed later that she herself was upset by the nature of the requested act and by the deception. In a further experiment the subject was reassured that the action was all right but no perceptual alteration was used. Under these conditions he was willing to perform the action for only one of the experimenters. It was also possible to induce the subject to perform the act by first requesting him to visualize its performance before directly requesting the action.
This study is particularly interesting in that the subject was willing under some situations to perform an action for the experimenter with whom he had the best rapport but not for the others. He refused to perform this action in the waking state despite the experimenters' attempts at persuasion.
induce the subject to perform the act. The limitations of the study are that only one subject was employed and that the subject was himself interested in investigating the legal implications of hypnosis. In this context, it is interesting both that the subject had amnesia for his action and that after he was finally informed of his behavior he felt that the need to demonstrate the point made the experiment legitimate. Further investigation along this line, especially utilizing subjects less ego-involved in the purpose of the study, would seem necessary in order to draw a more definitive conclusion.
Speaking for the negative in this controversy is an experiment reported by Haupt (30). The subject was a student who was in hypnotherapy with Haupt. The posthypnotic suggestion was given that the student would, upon awakening, pick up Haupt's notebook, leaf through it, and read it. This is an action which the author feels the student would never have dared under normal circumstances. After waking, the student rose, went to the table, looked at the open notebook and asked: "Here you write your notes, don't you?" He made no attempt to pick it up or read it. When memory for the posthypnotic suggestion was restored, the student reported that he had felt a drive to read the notebook but restrained himself. Haupt observes that the subject's behavior was a compromise between the suggestion and what was socially acceptable and that since this minor infraction was not performed, it is not possible to induce more deviant behavior by means of hypnosis.
A fairly elaborate study by Erickson (19), reporting some thirty-six individual experiments, supports the view that violations of social prohibitions cannot be achieved in hypnosis. This study is open to question in view of the reported results in laboratory settings by others. Erickson is known to his subjects as a responsible investigator. The fact that he did not have any positive results would lead one to wonder if he did not implicitly convey his expectations of refusal. In view of the relationship between subject and hypnotist in both the Haupt and Erickson studies, it may be that the subject would act in accordance with the hypnotist's implicit expectations.
In a review of the literature on this subject Weitzenhoffer (75) attempts to reconcile the contradictory evidence on inducing socially prohibited behavior. He points out that attempts which have been successful are those in which the subject was given a hallucinated pseudo-situation which redefined the behavior as socially acceptable. An instance of this would be the Wells' (80) demonstration. He induced the subject to "steal" a dollar bill by being told it was his own money. Thus, from the subject's viewpoint he was no longer
committing a transgression. Weitzenhoffer attributes failure to induce subjects to perform "antisocial" acts to those situations in which the subject perceives the transgressive nature of his behavior. This explanation, although seductive at first glance, does not appear to do justice to the literature. Erickson attempted in some instances to create this type of situation and obtained negative results. On the other hand, Schneck was unaware of the normative implications of his posthypnotic suggestion at the time it was given. Nor was there any attempt to disguise the dangerous nature of the situations in the Rowland or Young experiments.
It seems appropriate, in this context, to note that frequently subjects in hypnosis appear to show an increase of super-ego-type inhibitions. This has been pointed out by Gindes (27) and has been observed by the author. Bramwell (15) reports a case that clearly illustrates this point. A patient suffering from pulmonary disease was treated by hypnotic suggestion by her physician in the presence of a nurse. Before trance was terminated, the physician remembered that he had not examined the patient that week, and asked her to bare her chest so that he could examine her. Much to his amazement, the patient refused to do so despite the fact that this was a routine procedure to which she had never objected in the past. After the patient was awake, the physician again asked her and she permitted him to proceed with the examination without any objection. The nurse asked the patient sometime later why she had refused in hypnosis, and the patient expressed disbelief that she had done so. Under some circumstances, at least, behavior normally prohibited but appropriate to the situation will not be carried out in hypnosis. Apparently, under hypnosis the subject may interpret interpersonal motives and intentions differently from when they occur in the waking state.
To satisfy the requirements of an adequately controlled investigation of violations of internalized prohibitions in hypnosis, a number of conditions would have to be met. These have not been dealt with in any experimental study to date.
As has been pointed out previously, the experimental situation legitimizes much behavior which the subject, in other contexts, views as contrary to his internalized prohibitions. It is desirable to determine whether the behavior is also legitimized in the experimental setting by subjects who are not hypnotized. One way in which this can be determined economically 'is to utilize a control group of
subjects who are highly motivated to simulate hypnosis in order to deceive the experimenter. If the experimenter is not aware that the subjects are simulating, he will treat them as he does real subjects. If these controls perform the antisocial act, we may assume that the experimental situation itself has legitimized behavior that appears to be antisocial. A refusal of the control subjects to perform the given action would lend support to the hypothesis that the behavior cannot be legitimized solely by the experimental situation.
An additional possibility must be considered. Abundant evidence exists that under some circumstances of social legitimization, individuals indulge in behavior that is ordinarily viewed as antisocial; for example, lynching behavior, or extreme exhibitionism and sexual license in association with drinking or marijuana. In some instances, hypnosis may provide the legitimization for behavior which the person wishes to perform but which he feels he cannot do under normal circumstances. It is not clear whether it is hypnosis per se or the hypnotic situation which is instrumental in the production of these acts. Clinical evaluation of each experimental subject thus becomes necessary for an understanding of the motivations involved.