SOME COMMENTS ON THE NATURE OF POSTHYPNOTIC BEHAVIOR

PETER W. SHEEHAN, Ph.D.1, 2 AND MARTIN T. ORNE, M.D., Ph.D.1

It has long been recognized that a deeply hypnotized subject may be told to carry out some action in response to a specified cue subsequent to the termination of hypnosis. Two characteristics have traditionally been said to identify the phenomenon. First, the hypnotic subject experiences a compulsion to perform the deferred act and second, the hypnotic subject may show a relative lack of awareness of the source of motivation for his behavior (11, 18, 26). A deeply hypnotized subject, given the suggestion in trance to respond to a predetermined cue, will appear to respond compulsively in a quasi-automatic and involuntary fashion, and will usually be subjectively convinced about the genuineness of his response. There is some argument as to how important amnesia is to posthypnotic behavior (14), but the evidence seems to indicate that it has some effect on the consistency with which a hypnotic subject will respond to a given suggestion (12).

A third characteristic is suggested by the work of Fisher (8). Posthypnotic behavior appears to be very much influenced by the subjects' perceptions of the way the experimenter expects them to behave. Fisher maintains that posthypnotic suggestion does not lead to an automatic response which will occur in any setting. Rather, he views posthypnotic behavior as limited to the experimental context. A good hypnotic subject responds only when he thinks it is appropriate, and only when the testing is seen as part of the experimental setting. In his study Fisher suggested to 13 hypnotic subjects that they would touch their ear whenever they heard the word "psychology." After testing subjects on the suggestion following trance, he intimated to them that the experiment was over. Nine of the 13 subjects who had previously responded to the cue failed to respond. Only two subjects continued to respond consistently. Fisher concluded from these findings that posthypnotic behavior is limited to a specific context where the continued performance of a posthypnotic suggestion is a function of the subject's belief that the hypnotist expects him to perform the posthypnotic act.

The implications of Fisher's study are important for the understanding of posthypnotic behavior. They implicitly refute the classic view of the phenomenon as automatic and involuntary and have been used by other workers (1, 2) to support the view that posthypnotic behavior can be explained as compliant behavior. Fisher's findings, however, lend themselves to an alternative explanation. Since the original suggestion was "each time you hear the word 'psychology' you will scratch your right ear," subjects may have appropriately added to it "you will do this as long as the experiment persists," even though the hypnotist had not said so. Any other interpretation of the suggestion would not only be unreasonable, but the hypnotist would also have no conceivable motive for making a request to the subjects to respond indefi-

1 Institute of the Pennsylvania Hospital and University of Pennsylvania. The research upon which this study is based was supported in part by Contract Number Nonr 4731(00) from the Office of Naval Research and by the Institute for Experimental Psychiatry.

The authors wish to thank Frederick J. Evans, Ulric Neisser, Donald N. O'Connell, Emily Carota Orne and Campbell W. Perry for their helpful comments while the paper was in preparation.

Reprint requests should be addressed to the Unit for Experimental Psychiatry, 111 North 49th Street, Philadelphia, Pennsylvania 19139.

2 Now at The City College of the City University of New York.

209

210 SHEEHAN AND ORNE

nitely. If the subjects interpreted the suggestion to mean "as long as the experiment persists," their behavior would be fully appropriate and would offer no implication about the mechanisms of posthypnotic behavior. In this fashion one could readily account for the difference in the subjects' responses in Fisher's experiment. The majority of subjects most likely understood the experimenter as saying they would respond only as long as the experiment lasted, while the two subjects who continued to respond throughout might have interpreted the suggestion literally.

Though Fisher's data would seem to indicate that subjects' expectations affect their posthypnotic behavior, they do not necessarily demonstrate that posthypnotic behavior is limited to a specific setting which is perceived as part of the experiment by the subject; nor do they preclude the possibility that there are different kinds of posthypnotic responders. It becomes important, therefore, to determine whether subjects will respond to a clearly stated suggestion only in accordance with their expectations of what the experimenter wants at the time the suggestion is given, or whether they will respond whenever the response is tested, regardless of context. The latter possibility is tested only by a clear-cut, unequivocal posthypnotic suggestion that subjects are to respond to a predetermined cue. This requires that the hypnotist have a plausible motive in giving the suggestion. The suggestion should be tested in a context which is apparently unrelated to the experimenter, and in one where the experimenter has no way of knowing the nature of the subject's response, whether it occurs or not.

To investigate this issue explicitly, an experiment was devised to test the hypothesis that a posthypnotic suggestion given in a laboratory setting is not limited to the experimental context (19). Seventeen highly hypnotizable subjects and 14 insusceptible subjects were asked to participate in an experiment on 2 succeeding days. The subjects were told that the experiment involved a large number of personality tests, and they knew that they were to see the experimenter on the succeeding day. Insusceptible subjects were instructed to simulate hypnotic performance throughout the entire experiment. The second session legitimized the hypnotist's suggestion that "for the next 48 hours, each time you [the subject] hear the word 'experiment' your right hand will touch your forehead." The subjects were given explicit suggestions of amnesia for their trance experiences. They were tested formally on both days by the experimenter in the experimental room, and were also tested informally on the second day when the experimenter conversed casually with the subjects as he walked down the hall with them toward the experimental room. The crucial tests were made by the receptionist as the subjects left the building and as they returned the next day. On the first day when they were being reimbursed for their participation, the secretary asked them to sign the usual voucher (such as they had typically signed in previous experiences with the laboratory) so that they could be reimbursed for "their participation in the experiment." When they came to the waiting room for the session on the following day, they were asked by the secretary which "experiment" they were in.

The real-simulating model (16) was used to test the adequacy of the secretary's test as one unrelated to the experimenter's concerns. The experimenter was blind to the subject's true identity and simulators were motivated to fake hypnosis. Results showed that no simulator responded over the 2 days when tested by the secretary as they were leaving and entering the building, but 5 out of the 17 hypnotic subjects responded on both days when tested away from the experimenter. The significant differences in performance between the real and simulating subjects when tested by the secretary indicated

211 THE NATURE OF HYPNOTIC BEHAVIOR

that the design was an adequate test of the capacity of some hypnotic subjects to respond posthypnotically in other than an experimental context.

Results strongly reinforced Fisher's finding that the expectations of subjects are important factors to consider in understanding the posthypnotic response. The five subjects who responded outside the experimental setting on the 2 days all stopped responding when the experimenter used the cue word with the clear intent of removing the original suggestion. Later inquiry revealed that subjects understood that the suggestion was about to be lifted and anticipated the removal by ceasing to respond. The finding demonstrates that subtle implications of the experimenter's words are sufficient to produce a major alteration in subjects' behavior, but more than one mechanism is needed to explain the data. Some mechanism is needed to account for the automatic nature of the subjects' response which occurred away from the presence of the experimenter.

Data from the authors' study challenge current understanding of the nature of the posthypnotic response. For this reason, it seems important to examine inferences from the data in some detail. In the light of these inferences, the present paper aims to discuss the nature of posthypnotic behavior and to isolate some of the parameters which seem to affect it. Special attention is paid to the implications of the finding that some hypnotic subjects can respond posthypnotically outside the experimental setting.

AMNESIA AND POSTHYPNOTIC EFFECT

The influence of the subject's conscious knowledge of the content of a posthypnotic suggestion on carrying out the suggestion is an important issue which has obvious relevance to hypnotherapy. Researchers differ in their account of the exact role amnesia plays in influencing the posthypnotic response. For example, contrary to the findings of Erickson and Erickson (7), Barber (1, 2) maintains that posthypnotic amnesia does not aid the efficiency of the posthypnotic response. Results from the authors' study support the general contention (9, 12, 26) that amnesia is not a necessary or sufficient condition for posthypnotic behavior to occur. Three subjects who failed to have amnesia responded appropriately in the presence of the experimenter, and four subjects with complete amnesia failed to respond on all occasions when they were tested formally by the experimenter. The data, however, indicated that those subjects who responded on all occasions, in both the presence and absence of the experimenter, all reported no breakdown of amnesia for the original suggestion. Of eight hypnotic subjects who responded posthypnotically, both inside and outside the experimental setting at some time, six subjects had complete amnesia for their trance experiences, and one of the remaining subjects had amnesia for the suggestion itself. It seems, then, that posthypnotic suggestion can be effective if amnesia is not present, but the best responders (those who consistently respond in a variety of settings) generally are amnesic for their experiences. The data are consistent with previous reports that amnesic and nonamnesic subjects respond to posthypnotic suggestion differently (14, 27).

Many studies on posthypnotic suggestion have investigated the effect of memory of suggestion on the posthypnotic effect by explicitly suggesting to subjects the absence of amnesia in hypnosis (1, 6) ; or by omitting amnesia suggestions in the first instance, thus making it possible for the experimenter to observe the relationship of individual differences in posthypnotic effect to differences in memory of the suggestion (9, 25). For example, a recent study by Edwards (6) showed no change in posthypnotic effect on reaction time when amnesia was terminated in hypnosis prior to retesting the effect. Restored awareness did not influence the efficacy of the post-

212 SHEEHAN AND ORNE

hypnotic suggestion in the hypnotic subjects. The methodology of the authors' study was somewhat different. Subjects were selected who were all capable of amnesia and amnesia was suggested to them explicitly. The consequences of the subject's remembering the events of trance were analyzed by examining the posthypnotic behavior of subjects whose amnesia spontaneously broke down over a 24-hour period when they were tested on a number of different occasions. The differences in results between this and other studies such as Edwards' point to the importance of the fact that, in the present study, the experimenter did not sanction the subject's memory of the suggestion. Any breakdown in amnesia was opposed to the experimenter's request that the subject not remember. The contradiction between the subject's behavior and the experimenter's demand may be an important factor which contributes to the reduced efficiency of the posthypnotic response for subjects whose memory of the content of the suggestion approaches awareness. Conflict is not present in subjects who either experience suggested recall, or fail to forget when no explicit amnesia suggestions are delivered. In the latter instance, subjects simply may not have the expectation that they should forget the events of the previous trance.

When amnesia did break down, data supported clinical findings previously reported by Erickson and Erickson (7). The subjects acted at the level of conscious awareness but frequently showed an element of compulsion in their behavior. One subject's response appeared to be of an essentially voluntary character. Aware of the suggestion, the subject reported that she was faking because she wanted to make the response. She rationalized her behavior as compliance, but the subjective tone accompanying the response was clearly that of compulsion. Curiously enough, this subject responded appropriately more often away from the hypnotist than in his presence. Other subjects adopted idiosyncratic modes of defense, particularly when awareness of the suggestion was partially or completely present. There were fewer instances of defensive reactions when amnesia was present since the posthypnotic response selected for the study was harmonious with ongoing behavior and blended in easily with the subject's ordinary waking activity. Responses which were remote from the original suggestion did not seem disruptive nor at variance with the conscious stream of activity at the time the behavior was elicited, except when the responses had been carried out frequently, or when the subject was aware of the oddity of his behavior. The data point to the need to consider the quality of the posthypnotic act and the frequency of times it is tested. It seems that only the more unusual acts, and frequently tested responses, disrupt the normal stream of consciousness of the waking subject.

THE ROLE OF ATTENTIONAL PROCESSES

Aside from the subjective conviction hypnotic subjects express about the genuineness of their response, one of the important features of their behavior is the nature of the attentional processes brought into play in response to the posthypnotic signal. Some hypnotic subjects failed to respond when the cue word was used with emphasis by the experimenter but responded when the experimenter dropped the word informally in conversation. Response failure was particularly evident when subjects had noticed something peculiar in their behavior. For example, in the postexperimental inquiry one subject reported: "Today during the process of the day I heard the word 'experiment' several times. At first I didn't notice anything, but then I associated smoothing my hair with the one word and I got watching for the word and I tried to keep myself from doing it -- which was a challenge." Under these conditions of partial awareness, the typical response of the subjects

213 THE NATURE OF HYPNOTIC BEHAVIOR

was to resist rather than accept the posthypnotic suggestion. When the subjects knew that their behavior had been influenced by the hypnotist, they reacted to the disruptive character of the posthypnotic act and attempted voluntarily to suppress the behavior. They often succeeded in doing so, even in the presence of the experimenter. It has been observed elsewhere (3, 17) that hypnotic subjects given a posthypnotic suggestion comply only as long as they feel compelled to do so. Any other response on their part would be interpreted by them as cheating.

At some level the cue word must be registered as the relevant signal for response even though the subject reports no memory of the suggestion. Account must also be taken of the intra-individual variation in response to the same cue. One subject asked the experimenter if any posthypnotic suggestion had been given to him on the previous day; he had noticed that his hand moved up to his forehead more than it did normally. When the experimenter asked if the subject thought it was the word "experiment" to which he was responding, the subject answered the question but failed to give the appropriate response to the experimenter's use of the cue word. An inquiry into response failure revealed that the subject was unaware that the signal had just been given. The same subject responded to the cue word used in subsequent sentences during the session. Posthypnotic signals can clearly elicit the posthypnotic response in some instances but not in others. At some level the subject has to be "aware" that the signal to respond has been given.

The role that attentional processes play in eliciting posthypnotic behavior is complex. The data reflect the need to account for the subject's behavior in terms of perceptual sets or expectancies more than in terms of selective attention within the subject's auditory field. Findings are consistent with the operations of sets in everyday life and in experiments on sets functioning in incidental learning (21). In subjects whose amnesia holds, the goal-idea corresponding to the hypnotist's suggestion can be entirely absent from consciousness and the set still operates. There are marked individual differences, however, in the way such a set operates to elicit the hypnotic response in subjects who experience difficulty in keeping the goal-idea from consciousness. If the hypnotist's suggestion can be said to give rise to an expectancy or set which acts as a regulative, directional principle exerting a dynamic, organizing force on the subject's behavior, it seems reasonable to suggest that the motivational state of the subject is one determinant of the set's organizing force. The above subject's report suggests that motivated concern for the oddity of the behavior can override the determining tendency to respond to the cue word. The situation appears somewhat analogous to the influence of motivational states on the recall of incompleted tasks. Just as the familiar Zeigarnik effect fails to appear when subjects are motivated by ego-defense (22), so the tendency to respond posthypnotically can be countered effectively by anxiety and the presence of conflict.

In the study the word "experiment" was tested in isolation and in sentences. Our hypothesis was that subjects would give fewer responses to the cue word when it was embedded into the meaningful context of a sentence. This logic was accepted by simulating subjects in their experimental behavior, but was not illustrative of the performance of the hypnotic subjects. Of seven hypnotic subjects failing to respond at some time to the experimenter's formal test, four failed to respond to the word alone and three failed to respond to it in a sentence (one subject failed to respond to both types of test). Any simulating subject who failed to respond to the cue -- to convince the experimenter of the authenticity of his deception -- chose not to respond to the cue word when it was embedded in a

214 SHEEHAN AND ORNE

sentence. Hypnotic subjects were just as likely to respond to the word when it was embedded in a meaningful context as when it was used alone, except when the cue word had begun to disrupt their normal waking activity.

Cue embeddedness, then, is not a factor which significantly influences posthypnotic performance. The data suggest that hypnotic subjects may respond on some occasions but not on others, and that the sporadic nature of their response depends more on whether the significance of the cue is registered at an appropriate level of awareness. Attentional processes seem to play an important role in this process of registration. When the signal is not registered and subjects are not attending to the cue, they will fail to respond; but when attention is redirected to the signal, they may resume responding in the suggested fashion. Subjects who are partially aware of the nature of the suggestion may resist responding to the cue if it is conspicuous, but continue to respond when it is unobtrusive.

DEPTH OF TRANCE

Within the experimental setting, the ability to respond posthypnotically is related to the subject's depth of trance at the time the suggestion is given (28). Posthypnotic items on standardized tests of hypnosis are regarded typically as difficult items for subjects to pass. Hypnotic subjects who demonstrate considerable trance involvement can execute a posthypnotic response, but the present study shows that not all good hypnotic subjects show an ability to respond outside the experimental setting. Such an ability is shown only by a few subjects who display the best hypnotic performance.

For this study standardized tests of hypnotizability were used initially to select the best subjects from a large group. These subjects were then seen individually and only those were included whom two independent hypnotists had found capable of convineingly demonstrating such difficult phenomena as hallucinations, posthypnotic suggestions and amnesia. They were a highly selected sample who were very susceptible to hypnosis, and all of them responded appropriately to the experimenter's cue at some time throughout the experimental sessions. Although there was a statistically significant association between hypnotizability and extra-experimental response, only 5 out of 14 subjects consistently demonstrated posthypnotic behavior which was independent of the experimental setting. Of these 5 subjects, 4 had been able to respond successfully to every suggestion the hypnotist made over the 2 days. This finding suggests that consistent posthypnotic response both inside and outside the experimental setting is associated with other aspects of hypnotic response. Subjects who failed to respond consistently away from the experimenter also failed on other hypnotic tests which the experimenter had given them throughout the sessions.

Posthypnotic response was also correlated with the experimenter's perception of the subject as susceptible to trance. Of the subjects who responded away from the experimenter, all had been perceived as good hypnotic subjects at the time the suggestion was given. Posthypnotic behavior appears, to be a function of the subject's hypnotic response and the experimenter's conviction as to his trance ability. The data suggest that the conviction of the experimenter is not a sufficient prerequisite for posthypnotic behavior, since five susceptible subjects were perceived as good hypnotic subjects, yet failed to respond at any time outside the experimental setting. It is only for some subjects that the hypnotist's conviction is associated with extra-experimental response. These subjects appear to have a special aptitude for trance. The experimenter's differential treatment of them, however, may be a relevant factor in subjects' utilizing their aptitude for trance.

215 THE NATURE OF HYPNOTIC BEHAVIOR

HYPNOSIS AS "COMPLIANCE"

It is important to consider fully the implications of the view which asserts that hypnotic subjects respond just to please the hypnotist. To test this hypothesis, simulators were used who were instructed to meet all of the experimenter's requests. Results showed that they complied with the experimenter's suggestions more than hypnotized subjects. Simulators (92%) responded significantly more frequently than hypnotic subjects (59%) to the experimenter's informal test as they walked along the corridor casually conversing with him. Some simulators took their role so seriously that they falsely reported to the experimenter in his inquiry that they had responded to the cue word away from the laboratory. The simulators were clearly motivated to accept fully the role of a hypnotized subject as they perceived it, yet they responded differently from hypnotic subjects when tested outside the experimental setting. One may conclude that hypnotic behavior is distinct from compliance. Hypnotic subjects did not respond merely to comply with the requests of the hypnotist.

Posthypnotic behavior is not completely congruent with the role-playing concept of "appropriateness." Hypnotic behavior requires some other explanation. The role-taking theory of hypnosis (23, 24) is based on the assumption that hypnotic behavior is very much dependent on the subject's expectation that he should respond appropriately to comply with the experimenter's suggestions. Although role enactment is characterized by the organismic involvement of the hypnotic subject, the subject performs in an apparently automatic way because he believes that such a change is part of his role as a hypnotized subject (24). Hypnotic interaction is continuous with other social interactions and the units of the social structure are defined by a consensus of expectations. It is clear that expectations play an important role in determining whether a posthypnotic response will occur. But the fact that the posthypnotic response occurs in a situation perceived as inappropriate to the hypnotic context suggests that it is necessary to consider hypnotic responsiveness in other than expectational terms.

Subjects approach the hypnotic session with differing aptitudes and personality traits. It is important to consider whether such personality characteristics account for the individual differences in responsivity to induction. In terms of role-playing theory, one might argue that hypnotic subjects have different "role-playing propensities" from simulating subjects and may be role-playing with greater enthusiasm. It can also be argued that the request to simulators to "behave like hypnotic subjects" is less pressing than the request to the hypnotic group "to be like hypnotic subjects" (5). It is as if both groups of subjects are being asked to act in a play which differs in importance for them. The behavior of the simulators in the present study counters both objections. Simulators role-played with more enthusiasm than hypnotic subjects; they responded more often informally than, hypnotic subjects; and they saw it as part of their role to attempt to deceive the experimenter about their behavior away from the experimental sltuation. The data indicate that simulating subjects played their roles with more enthusiasm than the hypnotic subjects, but the role did not affect behavior beyond what they perceived as the specific experimental situation.

DURATION OF POSTHYPNOTIC SUGGESTION

One variable which is historically viewed as important in the study of posthypnotic behavior is the duration of the posthypnotic effect. Data on the durability of posthypnotic behavior support the notion that the posthypnotic response is considerably less effective than behavior moti-

216 SHEEHAN AND ORNE

vated simply by the need for compliance. The hypnotic response can be distinguished, for example, from the strength of compliance as measured by waking request. Damaser (3) found waking requests a more effective means of causing experimental subjects to perform simple tasks longer and more consistently than posthypnotic suggestions. The finding is similar to the results from the earlier studies of Kellogg (13) and Patten (20). Studies explicitly investigating the duration of the posthypnotic effect have found that the posthypnotic response may last longer than 3 months when retested by the experimenter (4, 15). Erickson and Erickson (7) have reported much longer-lasting effects.

The present study indicates that posthypnotic behavior in some subjects can last at least 24 hours, but the duration of the effect for most hypnotic subjects is very limited. Many who passed the stringent criteria of selection for the study failed to give the required response consistently either in the presence of, or away from, the experimenter. Of 16 hypnotic subjects who responded whenever they were tested on the first day in the presence of the experimenter, only 10 responded again the next day whenever the experimenter tested them; only 5 subjects responded on the second day away from the experimenter. Although the study was not designed to examine the limits of endurance of a posthypnotic response, the situation afforded an unusual variety of tests of the suggestion which demonstrated that the posthypnotic response really endures only in comparatively few instances. Durability of response is not a necessary characteristic of the posthypnotic behavior of hypnotic subjects. The frequency with which long-enduring responses are reported in the literature suggests that other factors may be important. One such factor, which will be discussed later, is the nature of the relationship between hypnotist and subject.

CONSISTENCY OF RESPONSE

The subjects who failed to meet all of the formal hypnotic requests by the experimenter in the study (e.g., the presence of hallucinations, delusions, lasting amnesia, and posthypnotic response in the presence of the experimenter) revealed interesting data on the limitations of the posthypnotic response. The evidence further distinguished hypnosis from compliance where simulators performed differently in the same situation from good hypnotic subjects. Three hypnotic subjects reported on the second day that they remembered the events of the previous session; two of them did not respond posthypnotically when subsequently tested by the experimenter. There was no breakdown of simulated amnesia. The hypnotic subjects' variation in response suggested that amnesia is a difficult condition for hypnotic subjects to maintain and its endurance is a characteristic feature of good hypnotic performance. A more striking feature of the subjects' performance was the inter-session consistency in hypnotic response not shown by those subjects who were asked to comply. No hypnotic subject who failed to respond to a posthypnotic cue on the first day responded to all the formal tests of the suggestion by the experimenter on the second day. Simulators were inconsistent in their response -- failure on the first day often preceded success on the same item on the second day. It seems that once the hypnotic subject ceases to respond to a suggestion, he does not tend to respond again when tested subsequently in the same fashion. This is consistent with the notion of the limited effectiveness of the posthypnotic response. The subject who merely attempts to please the hypnotist demonstrates a wider variety of responses, some of which are inconsistent with previous behavior.

Posthypnotic behavior is limited in many ways. It does not last as long, is more spo-

217 THE NATURE OF HYPNOTIC BEHAVIOR

radic, and is harder to evoke than behavior motivated by compliance. Further studies are needed to examine the conditions under which such behavior increases in effectiveness. These reports must be accompanied by statements about the nature of the subject's rapport with the hypnotist and the degree of personal investment the subject has in the posthypnotic response. Although posthypnotic behavior is limited, the automatic nature of the response, independent of the experimenter, clearly indicates a qualitative change in the organism, a change which cannot be explained by variables pertaining to simulation.

RELEVANCE TO THE THERAPEUTIC CONTEXT

Instances of long-lasting posthypnotic effects are commonly reported in the therapeutic setting. The application to hypnotherapy of the present finding, that posthypnotic suggestion is not as effective as other forms of social control such as waking request, is complicated by the differences between the therapeutic and the experimental setting.

Rapport: In the therapeutic context it is particularly relevant to consider the nature of the rapport between hypnotist and subject. The therapist cares about his patient and patients tell their therapists the results of suggestion. It is conceivable that hypnotic behavior endures outside the therapist's office for a considerable period of time simply because the patient responds in situations which are extensions of the therapeutic context, and which are strongly reinforced by the close personal relationship between two participants in therapeutic interaction. The occurrence of posthypnotic behavior away from the hypnotist raises many implications about the role "rapport" plays in the subsequent elicitation of the posthypnotic response. Data contradict the assertions of Barber (1) and Fisher (8) that somnambulists will obey suggestions from the moment the experimenter gives the command to wake up only until they are convinced that their relationship with the experimenter is no longer one of subject-hypnotist. In the authors' study some hypnotized subjects continued to respond at a time when they had temporarily discontinued their relationship with the experimenter. Posthypnotic behavior cannot be said to occur independently of a formal relationship with the experimenter in the sense that the hypnotist's rapport with the subject may still influence the subject in the absence of the hypnotist. But, once close rapport has been established and the suggestion has been accepted by the hypnotic subject, he can continue to respond quite independently of the presence of the hypnotist in a situation unrelated to the experimenter's interests. It is as yet empirically undetermined whether the behavior would have occurred if the relationship between experimenter and subject had deteriorated since the time the original suggestion was given.

EXPERIMENTAL AND CLINICAL SETTING

Earlier we alluded to the difference between the classical, more clinically oriented view of posthypnotic behavior, and that which emerges from some of the experimental studies. In a therapeutic context one can be impressed with the dramatic effectiveness of a posthypnotic suggestion dealing with a symptom which the patient has been unable to relinquish despite his apparent wish to do so, and the exhortations of others that he alter his behavior. There are important differences between the clinical and experimental settings. These concern the content of posthypnotic suggestion, the expectations hypnotic subjects bring to the setting, and the nature of the relationship between the subject and hypnotist.

The content of posthypnotic suggestion used in an experimental setting differs radically from that found in a therapeutic

218 SHEEHAN AND ORNE

context. Whereas the former typically utilizes behavior which is peripheral to the individual's interest or concern and becomes salient only because it is used for the experiment, therapeutic suggestions deal with behaviors about which the patient has significant feelings. They may be behaviors which the patient normally feels unable to carry out or which evoke anxiety. They may alter basic attitudes in a way that is dynamically relevant. Whatever the case, the posthypnotic suggestion in the clinical setting does not concern neutral material. Not only does the content of the suggestion differ between clinical and experimental contexts, but the expectations of the hypnotized individual are also quite distinct. Whereas experimental participation is implicitly, and often explicitly, defined as a transient "episodic" (see Garfinkel [10]) event, its termination guarantees that the subject will be left unchanged. Therapy promises to alter significant aspects of the individual's functioning, hopefully on a permanent basis. Finally, the relationship between therapist and patent is, as a rule, different from that between experimenter and subject. The therapist often shares with his patients the expectation that hypnosis involves unlimited compliance. In a therapy situation, permanent change is expected by both participants and a close, quasi-dependent, and emotionally charged relationship tends to develop in which the patient reports to the therapist the progress he is making in achieving the shared goals of their joint effort. This is in obvious contrast to the experimental situation where the relationship tends to be relatively impersonal and the common goal is to contribute to human knowledge.

Considering the differences between experimental and therapeutic situations, e.g., purposes, content of suggestions and nature of the relationship between subject and hypnotist, it is not surprising that very different types of observations have been made in the clinical and experimental settings. Evidence suggests that the type of secondary gain currently attached to neurotic symptoms and the relationship between patient and therapist make depth of hypnosis a relatively unimportant factor in determining therapeutic outcome as opposed to dynamic variables. Therapeutic suggestions are known to be effective for subjects in light trance when the situation is defined by doctor and patient as "hypnosis" (18). Neither amnesia nor depth of hypnosis appears to play a major role; the effect of posthypnotic suggestion typically transcends the immediate interaction resulting in posthypnotic behavior in the absence of the hypnotist.

In the experimental setting, the experimenter works with subjects with whom he does not have an intensive relationship, the posthypnotic response is trivial, and the experimental situation is defined as episodic. It is not surprising that the posthypnotic suggestion appears weak and ineffective when compared to a simple request (3). The experimental subject who is given a posthypnotic suggestion recognizes that the investigator is testing the effect of the suggestion and he will carry it out, particularly if there is amnesia for the suggestion. If the amnesia breaks down, he will respond only if he feels compelled to do so. Otherwise his behavior would be interpreted as cheating. In order to prevent the purpose of the experiment from being negated, the subject will often fight his impulse to carry out the suggestion. The fact that in an experimental context a simple request tends to be significantly more effective in eliciting the type of behaviors requested than a posthypnotic suggestion makes the type of finding reported here particularly interesting. If one compares a simple request with posthypnotic suggest on, the request will consistently produce the behavior for a much longer time. However, regardless of the way in which the request was originally formulated, the implicit termination of the experiment will

219 THE NATURE OF HYPNOTIC BEHAVIOR

promptly stop the behavior. The posthypnotic suggestion, while less able to elicit behavioral compliance over a time, is significantly more effective in transcending the immediate experimental situation. It is here that the effect of depth of trance on amnesia and the compulsion to carry out the suggestion becomes evident.

A posthypnotic suggestion given within the experimental setting shares something in common with the nature of a "request" in the therapeutic context: both extend beyond the setting of their origin. The posthypnotic suggestion, however, originates in a setting which has sharply defined limits, whereas the therapeutic context easily carries over to situations away from the therapist (it may be of considerable importance to the effectiveness of therapy that the patient has the opportunity to report results back to the therapist). In the therapeutic setting both a request and a posthypnotic suggestion transcend the immediate face-to-face relationship, but it becomes almost impossible for either the patient or the therapist to be certain as to when the patient responds to what. A request in an experimental setting is typically restricted by the subjects' perception of the experimental situation and does not transcend it; a posthypnotic suggestion, on the other hand, is clearly shown to transcend this context. The differences between the experimental and therapeutic setting with respect to the nature of the posthypnotic request, the intrinsic importance of the behavior to the patient, and the instrumental gains attached to the therapeutic response, all indicate why posthypnotic behavior is much more directly related to the depth of hypnosis in the experimental setting.

SUMMARY AND CONCLUSIONS

Contrary to Fisher's interpretation of his results, the posthypnotic response is not limited to the experimental setting but occurs outside the experimental context in the absence of the experimenter. This finding is of considerable theoretical significance. The expectations hypnotic subjects have about what the experimenter wants are important, but some mechanism is required to explain the quasi-automatic nature of the posthypnotic effect which occurs independently of the interests of the experimenter. Findings are incongruent with the role-playing view of hypnosis. Posthypnotic behavior is behaviorally distinct from compliance and cannot be accounted for except in terms of a qualitative change in the state of the organism. Amnesic and nonamnesic subjects respond to posthypnotic cues differently. The role amnesia plays in establishing an efficient posthypnotic effect in experimental subjects may be related to differences between the experimental and therapeutic contexts. Experimental subjects whose amnesia has broken down have relative difficulty in responding posthypnotically because they have failed to comply with the hypnotist's request to forget. The failure is not offset by any therapeutic gain from the response. The most effective posthypnotic response occurs in those subjects who have an outstanding aptitude for trance and are treated by the hypnotist in the knowledge that they are good hypnotic subjects. Such subjects show a marked consistency in their response, retain their ability to be amnesic, and respond posthypnotically over a considerable period of time. Data draw attention to the complex role attentional processes play in the eliciting of the posthypnotic response. The cue must be registered at a certain level of consciousness for it to elicit the required response, but its eliciting properties may be revoked by contrary motivational states in the hypnotic subject. Concern over the oddity of the response may be one factor which affects the process of registration.

The conclusions in this paper are supported by data gathered during the test of our hypothesis concerning extra-experi-

220 SHEEHAN AND ORNE

mental response. The data suggest the need to explore further the mechanisms responsible for hypnotic response outside the experimental setting. Evidence more tentatively draws attention to the need to investigate systematically the durability and consistency of posthypnotic response with careful recognition of the subject's aptitude for trance, ability to remain amnesic, and the persuasiveness of the hypnotist. Attentional processes in response elicitation require special consideration. Too little is known about what lends hypnotic suggestions their dynamic effectiveness.

REFERENCES

1. Barber, T. X. Hypnosis as perceptual-cognitive restructuring: II. "Post"-hypnotic behavior. J. Clin. Exp. Hypnosis, 6: 10-20, 1958.

2. Barber, T. X. Toward a theory of hypnosis: Posthypnotic behavior. Arch. Gen. Psychiat., (Chicago), 7: 321-342, 1962.

3. Damaser, E. An experimental study of longterm post-hypnotic suggestion. Doctoral dissertation, Harvard University, Cambridge, 1964.

4. Edwards, G. Duration of post-hypnotic effect. Brit. J. Psychiat., 109: 259-266, 1963.

5. Edwards, C. Motivation and post-hypnotic effect. Brit. J. Psychiat., 111: 983-992, 1965.

6. Edwards, G. Post-hypnotic amnesia and posthypnotic effect. Brit. J. Psychiat., 111: 316-325, 1965.

7. Erikson, M. H. and Erickson, E. M. Concerning the nature and character of post-hypnotic behavior. J. Gen. Psyched., 24: 95-133, 1941.

8. Fisher, S. The role of expectancy in the performance of posthypnotic behavior. J. Abnorm. Soc. Psychol., 49: 503-507, 1954.

9. Fisher, S. An investigation of alleged conditioning phenomena under hypnosis. J. Clin. Exp. Hypnosis, 3: 71-103, 1955.

10. Garfinkel, H. Studies in Ethnomethodology. Prentice-Hall, Englewood Cliffs, N. J., 1967.

11. Hammer, A. G. Reflections on the study of hypnosis. Aust. J. Psychol., 13: 3-22, 1961.

12. Hilgard, E. R. Hypnotic Susceptibility. Harcourt, Brace, New York, 1965.

13. Kellogg, E. R. Duration of the effects of posthypnotic suggestion. J. Exp. Psychol., 12: 502-514, 1929.

14. LeCron, L. M. and Bordeaux, J. Hypnotism Today. Grune and Stratton, New York, 1947.

15. Moll, A. The Study of Hypnosis. Julian Press, New York, 1958.

16. Orne, M. T. The nature of hypnosis: Artifact and essence. J. Abnorm. Soc. Psychol., 58: 277-299,1959.

17. Orne, M. T. The nature of the hypnotic phenomenon: Recent empirical studies. Paper presented at the meeting of the Amer. Psychol. Ass., Philadelphia, 1963.

18. Orne, M. T. Hypnosis, motivation and compliance. Amer. J. Psychiat., 122: 721-726, 1966.

19. Orne, M. T., Sheehan, P. W. and Evans, F. J. The occurrence of post-hypnotic behavior outside the experimental setting. J. Pers. Soc. Psychol. In press.

20. Patten, E. F. The duration of post-hypnotic suggestion. J. Abnorm. Soc. Psychol., 25: 319-334,1930.

21. Postman, L. and Senders, V. Incidental learning and generality of set. J. Exp. Psychol., 36: 153-165,1946.

22. Rosenzweig, S. An experimental study of 'repression' with special reference to need-persistive and ego-defensive reactions to frustration. J. Exp. Psychol., 32: 64-74, 1943.

23. Sarbin, T. R. Contributions to role-taking theory: I. Hypnotic behavior. Psychol. Rev., 57: 255-270, 1950.

24. Sarbin, T. R. Attempts to understand hypnotic phenomena. In Postman, L., ed. Psychology in the Making, pp. 745-785. Knopf, New York, 1962.

25. Schilder, P. The Nature of Hypnosis. International Universities Press, New York, 1956.

26. Weitzenhoffer, A. M. Hypnotism: An Objective Study in Suggestibility. Wiley, New York, 1953.

27. Weitzenhoffer, A. M. General Techniques of Hypnotism. Grune and Stratton, New York, 1957.

28. Wells, W. R. Expectancy versus performance in hypnosis. J. Gen. Psychol., 35: 99-119, 1946.