HYPNOTIC BEHAVIOR: - CiteSeerX

Journal oj Abnormal Psychology1966, Vol. 71, No. 1, 42-51

HYPNOTIC BEHAVIOR:

THE DIFFERENTIATION OF TRANCE AND DEMANDCHARACTERISTIC VARIABLES a

KENNETH BOWERS 2

University oj Illinois

Hypnotic Ss received posthypnotic suggestions (a) to begin all sentencesto the ostensible experimental (Taffel) task with "he" and "they," and (b)to be unaware of and amnesic for this fact. Waking simulator Ss receivedidentical suggestions preceded by instructions to behave later as though theyhad been hypnotized when they received the suggestions. During a post-experimental inquiry with a different E, 8 of 14 hypnotic Ss were amnesic fortheir experimental behavior; none of the 13 simulating Ss were amnesic (p <.0005). All simulating Ss testified that their use of "he" and "they" was vol-untary; the reverse was true for 12 of 14 hypnotic Ss (p < .00005). The resultsof this experiment demonstrate that hypnotic behavior is not wholly reducibleto acting in accordance with demand characteristics, and suggest that there isa "state" of hypnosis within which suggestions have a peculiarly potent effect.

Most investigators interested in hypnosisbelieve that there is an hypnotic state whichfundamentally differs from the waking state(As, 1963; Erickson, 1944, 1952; Marcuse,1959; Orne, 1959; Shor, 1960, Weitzenhoffer,1953, 1957, 1963; White, 1941). If there is astate of hypnosis, it would seem that thereought to be manifestations unique to it; yetnearly all of its purported behavioral indiceshave turned out to be spurious and/or evinci-ble by procedures other than those tradition-ally termed hypnotic (Orne, 1959; Barber &Calverley, 1962, 1964a; Barber & Hahn,1962; London & Fuhrer, 1961).

It is the contention of the present investi-gator that the search for behavior unique tothe hypnotic state is understandable but some-what misguided, and furthermore, that un-equivocal demonstration of the existence ofan hypnotic state is in no way contingent upona demonstration that certain behaviors areunique to it. It may well be that almost anybehavior, however bizarre, that is elicited froma somnambulist can be duplicated by a wak-ing person providing only that (a) he is privy

paper is based on a doctoral dissertationsubmitted to the Department of Psychology at theUniversity of Illinois. The author wishes to thankthe thesis chairman, William Gilbert, for his con-tinued support and guidance on this project, andJohn Williamsen for his untiring aid and helpfulsuggestions.

2 Now at the University of Waterloo, Ontario.

to information concerning how and in whatmanner he is supposed to behave, and (b)the stakes for a successful performance arehigh enough.

Nevertheless, it has been suggested thatequivalent behavior obtained from simulatorand hypnotic subjects is a sufficient basis forinferring an equivalence in their state, thatthis inferred state-equivalence casts doubtupon the reality of the hypnotic as distinctfrom the waking, motivated state, and thattherefore "the concepts of 'trance' and 'hyp-nosis' may no longer be useful . . ." [Barber& Calverley, 1962, p. 388]. It is here claimedthat any inference to equivalent states fromequivalent behavior must ultimately devolveupon the conditions under which similaritiesof behavior are obtained (Weitzenhoffer,1963). It is moreover claimed that the con-trol conditions of many previous hypnotic ex-periments demonstrating no difference be-tween hypnotic and simulating subjects dis-allow any such inference.

The effect of demand characteristics andexpectancies (by whatever name) on hyp-notic behavior seems to be widely acknowl-edged (Erickson, 1944; Orne, 1959; Sidis,1906), but their effect on the behavior ofsimulating controls has not been equally ap-preciated. As a result, simulators are oftendifferentially exposed to procedures whichinform them of what constitutes "hypnotic"

42

HYPNOTIC BEHAVIOR 43

behavior and/or which communicates theinvestigator's expectancy that the simulatorshould behave in a prescribed, hypnoticlikefashion in order to make the experiment asuccess (e.g., Barber & Calverley, 1963, Ex-periment I; Orne, 19S9). These kinds of in-structions are generally termed motivatinginstructions—and no doubt they do enhancemotivation; but they do so in part by alteringthe experimental demand characteristics andthe subject's expectancies of how he is tobehave. Most subjects in an experiment willordinarily behave in a way commensurate withwhat they feel is expected of them (Orne,19S9, 1962); to explain motivated controlsubjects' behavior with reference to motiva-tion only thus seems to be misleading andinaccurate. By contrast, hypnotic subjectsare often told just to let happen whateverhappens without attempting to help or hinderthe operator. Similar behavior elicited fromsubjects in each of these two groups couldquite obviously occur for two different rea-sons. The present investigator submits thatselectively providing only the simulatingsubjects with expectancies eventuating in be-havior comparable to that evinced by hyp-notic subjects confounds the effects of trancevariables on one hand, and demand character-istics and expectancy variables on the other.Such a confounding disallows any inferencethat task-motivated and hypnotic conditionsare equivalent in essence as well as in effect.

Inasmuch as the influence of demand char-acteristics has proven to be ubiquitous forexperimental subjects in general (Orne, 1962;Orne & Scheibe, 1964), the fact that controlsimulators will respond to them is not at allsurprising. The strategy of the present experi-ment takes advantage of this fact and focuseson a slightly different question, namely: willhypnotic subjects respond to an alteration indemand characteristics (in the same way thatsimulators do) when so responding is opposedby previously administered, posthypnoticsuggestions?

Through the use of posthypnotic amnesia,the present experiment will attempt to dem-onstrate that the effects of demand character-istics and trance variables are distinguishable.The following two working assumptions aremade: (a) that in general, a simulator sub-

ject's failure to verbalize presumably amnesicmaterial or to justify previous behavior is adirect effect of a group of variables subsumedunder the rubric "demand characteristics,"and that subsequent and appropriate manip-ulations of the demand characteristics candispose such a subject to verbalize what heremembers; (b) that an hypnotic subject's"amnesic" and "nonjustifying" behavior is adirect effect of some variable other than de-mand characteristics and that subsequentmanipulations of demand characteristics de-signed to elicit recollections or justificationsfrom such a subject will fail in their aim. This"other" variable responsible for an inabilityto remember or to justify behavior is assumedto be somehow related to an altered state ofthe subject like that ostensibly induced byhypnosis. Let this be called the trance varia-ble.

The above two assumptions have focused ontwo aspects of a subject's behavior: a failureto verbalize, which is referable to some pre-sumably amnesic state of the organism, andthe presence or absence of justifying behavior.The latter notion requires some further elab-oration. For the purposes of this study, a per-son who justifies his previous behavior is onewho claims that the behavior in question wasperformed voluntarily; a person who does notjustify prior behavior either denies or doesnot admit the occurrence of the behavior, orclaims that it was performed involuntarily. Itseems reasonable to suppose that the presenceor absence of the volitional quality of experi-ence is related to the presence or absence ofthe altered state ostensibly induced by hyp-nosis (Austin, Perry, Sutcliffe, & Yeomens,1963; Erickson & Erickson, 1941; Miller,Galanter, & Pribram, 1960).

The dependent variable of this experimentwas subjects' verbal reports collected in anostensibly postexperimental inquiry consist-ing of standardized questions about the ap-parent experimental (Taffels) task. Use of

"The Taffel task (Taffel, 1955) is a task orig-inally utilized in verbal conditioning experiments.The only equipment consists of a series of 3 X 5index cards on each of which appears (in differentorder on each card) the personal pronouns I, he,she, you, we, and they. A different past tense verb islocated above the pronouns on each card. The sub-

44 KENNETH BOWERS

subjects' postexperlmental reports as the de-pendent variable has been suggested previ-ously by Orne (19S9) and Barber and Cal-verley (1962), and it has been utilized in amodified form by Fisher (19S4) and Austinet al. (1963).

The hypotheses to be tested are as follows:I. There will be a positive relation between

the hypnotic condition and a failure to verb-alize the use of "he" and "they" in a post-experimental inquiry.

II. Fewer hypnotic subjects than simulat-ing subjects will justify their use of "he" and"they" on the Taffel task during a postexperi-mental inquiry.

METHODSubjects

The subjects were selected from a volunteer groupof approximately 450 undergraduate women at theUniversity of Illinois, most of whom were enrolledin an introductory course in psychology. Some ofthe subjects were enrolled in more advanced psy-chology courses, and a very few of them heard aboutthe experiment from friends and volunteered theirservices. Subjects from this group were selected forthe initial phases of the experiment on the basis ofa score earned on the Harvard Group Scale of Hyp-notic Susceptibility (Shor & Orne, 1962). Subjectswho passed 8 items (including the posthypnotic am-nesia item) out of 12, and who expressed a willing-ness to do further individual work in hypnosis wereseen for 1 or 2 more sessions during which trainingand practice in attaining deep hypnosis were given.During these training sessions, no specific suggestionsof posthypnotic amnesia were administered. Thepurpose of this deletion was to minimize later effectsof subjects' expectancies concerning this phenomenonon the eventual dependent variable.

After the training sessions 10* out of the 12 itemsof the Stanford Hypnotic Susceptibility Scale FormC depth scale (Weitzenhoffer & Hilgard, 1962) wereadministered to each subject. A score of 7 or morewas considered criterial. After both screening pro-cedures were completed, 37 subjects remained. Eightsubjects served in pilot work; 29 were run in thefinal study. Subjects were randomly distributed intoan experimental hypnotic group (hypnotic subjects),and a control, simulating group (simulating sub-jects). One of the simulating subjects was highlydeviant in her performance to the Taffel task, so her

ject is asked to make up sentences to the indexcards, always beginning each sentence with one ofthe pronouns, and using the verb as the predicate.

4 The posthypnotic amnesia item was deleted forreasons outlined in the text. The arm rigidity itemwas deleted because it was consistently utilized asan initial warm-up suggestion for all subjects intheir individual training sessions.

responses to the inquiry procedure were not ana-lyzed. The data of another subject were lost due toa breakdown of the recording equipment.

ProcedureThe general strategy of the experiment was to

manipulate the experimental situation so as to elicita "confession" from most simulating subjects con-cerning their voluntary use of "he" and "they," andthen to see how identical procedures affected thehypnotic subjects. It was not so easy as it mightseem to elicit remembering and justifying behaviorfrom a highly susceptible, hypnotically trained sub-ject who was consciously committed to playing therole of a posthypnotic subject. Despite the diffi-culties, however, a four-stage procedure was evolvedthat consistently resulted in the desired "confessions"by the pilot simulating subjects. This procedurewas subsequently applied to the hypnotic subjectsand simulating subjects in the main study as out-lined below:

I. At the beginning of the final (third or fourthindividual) meeting, the hypnotist (Experimenter 1)turned on the portable tape recorder which was onhis desk in full view of the subject. (This was stand-ard procedure for all of the individual practice ses-sions as well.) He then hypnotized the subject bymeans of the arm levitation technique with the sug-gestion that the subject would be more deeplyasleep than she had ever been before by the timeher hand touched her face. Experimenter 1 thenread suggestions to the sleeping hypnotic subjectstating in effect that (a) she would always beginsentences to a task, soon to be presented, withwhat was being called the correct response, that is,with either "they" or "he"; (ft) that she would beunaware of this fact "just as you are unconsciousof your rate of breathing or the fact your eyesblink," and (c) that upon awakening, she wouldneither remember nor be reminded of any of thesesuggestions. The hypnotic subject was then awak-ened and introduced to Experimenter 2, who pre-sented the ostensible, experimental task. Experimenter1 was absent from the room at this time. The simu-lating subjects were treated exactly the same wayexcept that prior to the reading of the above hyp-notic suggestions, they were awakened and readthe following instructions:

In a moment . . . I am going to read you ex-actly the same instructions hypnotized subjectsare read . . . and later on I want you to pretendas though you had been hypnotized when youwere read the instructions, and act just like youthink the subject who had been hypnotized wouldact under the circumstances. Do you understand?Fine. Here are the instructions.

II. Experimenter 2, who was ignorant of thegroup to which any particular subject belonged, firstadministered standard instructions for the Taffeltask. Twenty-nine instead of the usual 80 cardswere presented. After the subject completed theTaffel task, Experimenter 2 presented 2 MAPS

HYPNOTIC BEHAVIOR 45

pictures. The subject was asked to make up S sen-tences to each of the 2 pictures. After the subjectcompleted the 10 sentences, Experimenter 2 leftthe office and Experimenter 1 re-entered.

III. Experimenter 1 sat down behind the desk andswitched off the tape recorder on his desk. Simul-taneously, through the use of a foot-pedal control,he switched on another tape recorder hidden fromthe subject's view in a large desk drawer. This ma-chine thus provided verbatim records of the sub-sequent (Part IV) inquiry procedure. He then an-nounced to the subject, "Well, the experiment isover. Do you have any questions about it?" Nomatter what the subject said, Experimenter 1 even-tually worked the conversation around to the pointwhere he could say without apparent irrelevance:

The purpose of the experiment was to see if theway subjects make up sentences to the index[Taffel] cards has an influence on the way theymake them up to the [MAPS] pictures. And thereason I had a third person administer the taskis because theoretically at least I could have un-wittingly biased the responses in the way I wantedthem to come out.

The reason for turning off the mock tape re-corder, announcing that the experiment was over,and providing the subject with a raison d'etre forthe experiment was to alter the situation in such away that it would no longer be perceived as experi-mental in nature. It was assumed that such analteration would reduce, if not eliminate altogether,a subject's further need to perform successfully(Barber, 1962).

The subject was allowed to make any final com-ments, and just before she stood up, Experimenter I,in an apparently spontaneous but actually stand-ardized fashion, asked the subject if she could talkto another person (Experimenter 3) for a moment.This person was purportedly doing pilot research inan experiment concerning verbal behavior and hadexpressed interest in talking to Experimenter 1'ssubjects. After Experimenter 1 had made appropriateexplanations to the subject and had extracted herpermission to be interviewed by Experimenter 3,Experimenter 1 said, "And listen, since we're allfinished here, I want you to be completely honestwith him, regardless of what I've said before." Thepurpose of the italicized portion of Experimenter 1's"closing speech" was to convince the subject thatExperimenter 1 wanted her to tell everything toExperimenter 3 she could remember, despite anysuspicions she might have had about whether theexperiment was in fact over. Thus, two factors wereoperating to promote the subject's "confession"concerning her use of "he" and "they." First, theexperiment was declared to be over, thereby remov-ing an experimental subject's normal tendency toperform successfully from that point on. Secondly, anew and different demand was established to theeffect that the subject was to be honest despite anyprevious instructions.

IV. Experimenter 3, on any given occasion, didnot know in advance of his interview with a subjectwhether she was an hypnotic subject or a simulat-ing subject. Experimenter 3 was introduced to thesubject, thanked her for giving him a few extramoments, and acknowledged his good fortune inbeing able to use Experimenter 1's subjects as partof his pilot data. In a conversational manner, Ex-perimenter 3 then asked the subject a series ofstandardized questions embedded in a matrix ofirrelevant questions and comments designed to mini-mize the subject's suspicions. There was no attemptto adhere to a specific schedule, and the conversa-tion proceeded in what appeared to be a spon-taneous fashion. Experimenter 3 did not discouragesubjects from spontaneously elaborating answers orasking questions. When Experimenter 3 did ask thepertinent questions, however, he asked them in orderand in their standarized form insofar as the circum-stances allowed. These interview sessions were taperecorded, and the kind of responses that the subjectsgave, judged in terms of the criteria presented be-low, served as the dependent variable of this study.The standardized questions follow:

(a) Tell me, how did you go about making upsentences to the pictures? (pause) For example,was there any regularity or consistency about thekind of sentences you made up to either of thetwo pictures?

(6) Was there any similarity in the way youmade up sentences to the pictures and the wayyou made them up to words on the index cards?

(c) Was there any regularity or consistency inthe way you made up sentences to the index cards ?

(d) From the start to the finish of each sen-tence, tell me step by step how you went aboutmaking up sentences to the words on the indexcards.

(e) [If there has not yet been any acknowledg-ment of the use of "he" and "they":] Well, oneof the things you haven't mentioned, perhapsbecause you didn't think it was important, is thefact that you started your sentences with "he"and "they," (pause) Hmm? (pause) Do you re-call now? [Regardless of answer:] I'm wonder-ing why you didn't mention the use of "he" and"they." Several possibilities occur to me. One isthat you couldn't remember the words, another isthat you could remember the words but youdidn't think you were supposed to mention them,and still a third is that you just didn't think itwas important. Which of these three alternativessounds the closest to the way you felt?

The questions up to this point were concernedwith eliciting information about posthypnotic am-nesia. In order to be counted as amnesic, a subjecthad to say nothing about the use of "he" and"they" through Question d, and in response to Ques-tion e, say that she could not remember using thewords.

46 KENNETH BOWERS

The following questions were concerned with theability of a subject to justify her behavior and wereasked only if the subject immediately acknowledgedor ultimately recognized her prior use of "he" and"they."

(/) At the time you were actually doing thetask did you realize that you were starting allyour sentences with "he" and "they"? [If "yes,"Experimenter 3 went on to Question g; if "no,"then /i was asked.]

(/i) When you say "no" to this question, it'spossible to mean one of two things. One possi-bility is that even though you knew you wereusing "he" and "they" it was really very naturaland easy to do this, and therefore you simplydidn't have to think much about it as you wentalong. The second possibility is that you reallydidn't have the slightest idea that you were start-ing your sentences with "he" and "they." If thiswere the case, you were probably very surprisedto discover that this was what you had done.Which of these possibilities seems closer to thetruth, the first one or the second one? [If thesubject replied with the former, it counted thesame as if she had said "yes" to question /. Ifthe subject replied with the latter, no furtherquestions were asked.]

(g) If you were aware of using "he" and"they," did you realize at the time why you wereusing the words?

(h) At the time did you use the words involun-tarily or voluntarily? And by voluntarily I meandid you use them on purpose because you knewyou were supposed to, that is, because they hadbeen designated as correct for this experiment?

The subjects were said to justify their behavior onlyif they claimed that their use of "he" and "they"was voluntary in response to question h. The sub-jects were said not to justify their behavior who re-plied to question h that they had responded in-voluntarily, or who were never asked this questionbecause of its irrelevancy in light of previous answers.It was assumed that only simulating subjects wouldjustify their behavior and the validity of this as-sumption was tested by seeing whether hypnoticsubjects and simulating subjects could be accuratelyidentified on the basis of this criterion.

After Experimenter 3 completed his interview, heleft the office and Experimenter 1 returned. If thesubject was a simulating subject, Experimenter 1thanked her for helping Experimenter 3, and ad-monished her to say nothing about what had hap-pened during the final session. If the subject was anhypnotic subject, Experimenter 1 rehypnotized her,removed the suggestion to begin sentences with "he"and "they," and awakened her. Neither hypnoticsubjects nor simulating subjects were ever told theactual purpose of the experiment at this time. Thisprecaution was taken to minimize the importanceof any information leaks that might occur despiteExperimenter 1's plea for silence.

RESULTS

There was no significant difference in thehypnotic subjects' and the simulating sub-jects' tendency to use "they" and "he" on theTaffel task nor was there any significantdifference in the frequency with which hyp-notic subjects and simulating subjects usedthe "correct" pronouns to begin sentences tothe MAPS pictures.

Although the effect of hypnosis did notproduce significant differences on a task ex-plicitly defined as experimental, there werequite striking differences in the "post-experimental" reports of hypnotic subjectsand simulating subjects. Before reporting theactual findings, however, a word should beinterjected about how the subjects' reportswere classified. It might seem that the explic-itness of the criteria for amnesia and justi-fication would disallow any equivocation injudging their occurrence for any given sub-ject. Indeed, this was the case for the vastmajority of the subjects. However, a smallminority of subjects responded in such away that made it difficult to ascertain un-equivocally whether or not they fell into oneresponse category or another. This occasionalambiguity occurred for both the dimensionsof amnesia and justification. In order to arbi-trate this matter, two independent judgesassessed all of the tape-recorded responses.The data of the judge least favorable to theexperimental hypotheses are reported below.In confirmation of Hypothesis I, there was ahigh relationship ( <j> — .62; p < .0005) be-tween the hypnotic condition and the lack ofverbalization of "he" and "they" during the"postexperimental" inquiry. All of the 13simulating subjects verbalized the use of"he" and "they" at this time whereas 8 outof 14 hypnotic subjects did not. Subjects'use of "he" and "they" was agreed upon 89%of the time by the two judges. In confirma-tion of Hypothesis II, the use of justificationas a criterion in discriminating hypnotic sub-jects succeeded 25 out of 27 times. All 13simulating subjects justified their behavior;12 out of 14 hypnotic subjects did not (one-tailed x2 = 20.05; p<.00005). Two inde-pendent raters agreed in their classification of

HYPNOTIC BEHAVIOR 47

justifiers versus nonjustifiers 92 % of thetime.

Three possible objections might be raisedabout the present findings. First, it might beargued that Experimenter 1 was not naiveabout the condition of the subject and thatthis knowledge could have subtly and selec-tively affected the way he vocalized his clos-ing speech (Rosenthal, 1964). In an attemptto circumvent this criticism entirely, Experi-menter 1 did not give the closing speechduring early pilot work. Instead, Experi-menter 3 announced the end of the experi-ment after ostensibly collecting data forExperimenter 1. Following this announce-ment, Experimenter 3 plied the subject withsome "informal" questions of his own, theanswers to which were utilized in makingjudgments concerning amnesia and justifica-tion. Now, the research program consisted ofseeking and finding a procedure that wouldconsistently elicit "remembering" and "justi-fying" behavior from simulating subjects andthen to see how the identical procedures af-fected the hypnotic subjects. Unfortunately,under these initial pilot conditions, simulatingsubjects did not "confess" even though theywere convinced the experiment was over.Their rationale for continued withholding wasa belief that only the hypnotist could releasereally hypnotized subjects from previoussuggestions by announcing the end of theexperiment. The effect of these confoundingexpectancies obviously had to be eliminatedfor both hypnotic subjects and simulatingsubjects, so Experimenter 1 thenceforthannounced the experiment's finale.

In lieu of obviating completely the possibil-ity that Experimenter 1 would convey dif-ferential treatment effects through subtle andconfounding vocal cues, it was possible totest for their existence. The 27 repetitions ofExperimenter 1's closing speech were tape-recorded and four independent judges5 wereasked to judge from these recordings whetherExperimenter 1 was talking to an hypnoticsubject or a simulating subject on any givenoccasion. Three judges met with totallychance success in their judgments, but one

"One senior psychologist, two advanced graduatestudents in clinical psychology, and one PhD inelectrical engineering served as judges.

judge did discriminate successfully (p < .05).When this judge was asked what his ration-ale for categorizing subjects was, he repliedthat Experimenter 1 seemed more convincingand persuasive with the hypnotic subjects!If this were in fact true, that is, if his successwere not simply improbable, it seems plausi-ble that hypnotic subjects would have beenmore prone to "confess" than simulating sub-jects all other things being equal. In any event,it seems unreasonable to suppose that naiveexperimental subjects could have been con-sistently influenced by vocal cues the verypresence of which were so unreliably assessedby sophisticated judges set to discriminate.

A second possible objection to the presentfindings concerns the possibility that the in-quiry procedure might have aroused subjects'suspicions that the experiment was not over,and that this suspicion might in some way beused to explain the present results better thanthe independent variable. To check out thispossibility, Experimenter 1 telephoned eachsubject after all the subjects had been seenfor the last time. He asked, "By the timeExperimenter 3 was actually asking you hisquestions, did you have any notion that infact my experiment was not over?" Ninehypnotic subjects and six simulating subjectssaid "No," three hypnotic subjects and fivesimulating subjects said "Yes."8 (Fisher ex-act probability test insignificant at .05 level).At least by their own account, the hypnoticsubjects and simulating subjects did not differsignificantly in their perception that Experi-menter 3's questions constituted a postexperi-mental inquiry. It follows that insignificantdifferences in how the inquiry procedure wasperceived could not account for the highlysignificant differences of hypnotic subjects'and simulating subjects' verbal reports to theinquiry.

A third objection that might possibly belevelled at the present experiment is this.Actually, there were two differences in theway the hypnotic subjects and the simulatingsubjects were treated: hypnotic subjects butnot the simulating subjects received theirsuggestions while hypnotized; simulating sub-

6 Two hypnotic subjects and two simulating sub-jects were not included in this computation becausetheir answers disallowed objective classification.

48 KENNETH BOWERS

jects but not hypnotic subjects received spe-cial instructions to behave later as if theyhad been hypnotized when they received thesubsequent suggestions. Since there were twodifferences in the way the groups weretreated, any significant results could as easilybe attributed to instructional differences asto hypnosis.

This is a logical argument, but not anentirely reasonable one when the purpose ofthe simulating instructions is clarified. Ifthese instructions had been deleted, thensome of the simulating subjects would havebeen apt to make wrong assumptions abouthow they were to behave. For example, itwould be quite plausible for a simulatingsubject, unprepared by simulating instruc-tions, to assume that subsequent hypnoticsuggestions were not supposed to "take"simply because both he and the hypnotistknew he was not hypnotized upon receipt ofthe suggestions. Such an assumption by asimulating subject could be expected to leadto obviously nonhypnotic behavior. The ex-plicitness of simulating instructions thusserved to maximize the likelihood of all simu-lating subjects behaving hypnotically insofaras this behavior is determined by the experi-mental demand characteristics. The purposeof maximizing this source of hypnotic behav-ior in simulating subjects was to minimizethe differences between the two groups insofaras this difference might be a function of thehypnotic subjects' sensitivity to the demandcharacteristics (Fisher, 1954). In short, thefunction of the simulating instructions wasto enhance the likelihood of accepting thenull hypothesis if in fact hypnotic behaviorwere nothing but responsiveness to demandcharacteristics. Therefore, any significant dif-ferences between groups that did appear seemto be the result of something other than simu-lating instructions, namely, the presence ofthe trance variable in hypnotic subjects only.

It must be admitted that the above line ofthinking is only reasonable, and that themerely reasonable can occasionally be as mis-leading as the merely logical. For example, itmight be true that demand characteristicsfor hypnotic subjects were totally differentfrom those of the simulating subjects simplybecause the hypnotic subjects knew or be-

lieved themselves to be hypnotized and thesimulating subjects knew that they were not.This is a somewhat vexing possibility noteasily accessible to experimental control.Barber and Calverley (1964b) suggest thata slight variation of this possibility is thecase, but their conclusions in this and similarexperiments are suspect due to faulty meth-ods described below. Anecdotal evidence sug-gests, however, that such belief or knowledgeis not a crucial demand variable in hypnosis.Occasionally, one hypnotizes a highly sus-ceptible subject who afterwards claims (of-tentimes with some vehemence) that he knewat the time and knows now that he in factwas not hypnotized. This claim typically doesnot square with the clinical facts, for ex-ample, unacknowledged and spontaneousposthypnotic amnesia. Evidently in thesecases it is not the demand characteristicssupplied by the knowledge or belief ofbeing hypnotized that eventuates in hypnoticbehavior.

DISCUSSION

Inasmuch as differences due to experi-mental demand characteristics were mini-mized, the results of the present investiga-tion seem to constitute one of the more con-vincing experimental demonstrations of the"altered state" theory of hypnosis. The suc-cess in this regard was made possible in partby avoiding a methodological pitfall commonto investigations which have concluded orimplied that the concept of an hypnotic stateis expendable. The flaw that has promptedthis conclusion is simply stated as follows:typically, the experimenter provided too muchinformation in the initial (hypnotic, simu-lating, task motivating, etc.) suggestionsabout what kind of (dependent variable) be-havior was expected. If, for example, bothhypnotized subjects and task-motivated(simulating) control subjects are told theyare to see an (hallucinatory) cat on theirlaps or hear Jingle Bells played on an imagi-nary phonograph (Barber & Calverley, 1964a)then both groups of subjects have a goodidea of what is expected of them: they aresupposed to say that they saw a cat andheard the song. Given the admission of hal-lucinations by both hypnotic and task-

HYPNOTIC BEHAVIOR 49

motivated subjects, it is impossible to knowif the hypnotic subjects are simply respond-ing to the demand characteristics as thecontrol subjects are, or if they so respond be-cause they "see" and "hear" the suggestedphenomenon as a function of having receivedthese suggestions in a hypnotic trance.

Barber and his co-workers are particularlyvulnerable to this criticism. He has presenteddata from many experiments all of whichlead him to conclude that simply giving directsuggestions to waking, task-motivated sub-jects is sufficient to elicit behavior that can-not be differentiated from "hypnotic" behav-ior. If this were all he concluded, there wouldbe little to disagree with. However, he furtherimplies that since hypnotic subjects do notbehave differently from task-motivated con-trols, the basis of their behavior must besimilar. This does not seem to be a legitimateconclusion.

It is very important to be clear about this.A person can voluntarily perform a tic thatto all appearances is indistinguishable froman involuntary, nervous tic. The behavior isthe same; the basis of the behavior quiteobviously is not. The actual differences under-lying these two kinds of tic behavior are ofparamount theoretical and practical impor-tance, for example, the unconscious as op-posed to conscious origins of the nervous tic,the implications this has for treatment, andso on. Equally important issues are at stakein the experimental study of hypnosis.Consequently, insofar as one is grappling withquestions about the nature of hypnosis, thereis a point of diminishing returns implied bythe exclusive use of paradigm that tends, byits very nature, to emphasize the similari-ties between real and simulated (or task-motivated) hypnotic behavior rather than toclarify the nature of the differences under-lying them. That hypnotic and control sub-jects do behave similarly is certainly avaluable finding, particularly when it is notan expectable result on strictly a priorigrounds. The question remains, however,whether or not the similarity in behaviorreflects a similarity in the basis of behavior.

There seem to be at least three competinganswers to the question of why hypnotic sub-jects and their controls do in fact behave

similarly: (a) the control subjects are some-how hypnotized by virtue of having receivedtheir suggestions in a motivated state (apossibility that seems almost too remote toconsider); (b) hypnotic subjects are merelytask motivated and/or responsive to the ex-perimental demand characteristics (a pos-sibility that Barber seems to favor); (c)control subjects simply respond in a moreor less purposeful manner to the task de-mands, whereas the hypnotic subjects' behav-ior is determined both by demand variablesand by state or trance variables.7 The pres-ent investigation seems rather unequivo-cally to confirm this last alternative. Previousinvestigations, however, utilizing the usualhypnotic-simulator paradigm have not reallylegitimized a choice among these options.Their strategy has generally consisted solelyof proving the null hypothesis that hypnoticand simulator subjects can behave similarly.But no matter how many times this sameparadigm is used to "prove" the null hypothe-sis of behavior equivalence, it will never suf-fice to accept or reject the null hypothesisconcerning the underlying bases of similarbehavior. To maintain contrariwise, that is,to conclude that because behavioral manifes-tations of hypnotic and simulating subjectsare similar, the behavioral substrate must be,is a little like saying that the person whovoluntarily performs a tic must be as neuroticas the genuine tiqueur.

The present investigation suggests thatearlier studies showing no difference betweenhypnotic subjects and their controls did sofor a very good reason: they compared groupsin terms of behavior that was entirely

7 These alternatives are not necessarily exhaustive.Indeed, much of the literature on social persuasionand conformity behavior seems pertinent to Barber'smethods. Oftentimes, for example, Barber presentswhat amounts to false norms to his task-motivatedcontrols, telling them, for instance, that so fareveryone who has really tried to see a cat on theirlap has been able to do so. A subject's subsequentadmission that he indeed did see a cat bears a cer-tain resemblance to the behavior of subjects in anAsch type experiment whose responses on a variety ofjudgment tasks are altered greatly by the influenceof false norms (Crutchfield, I9SS). It has beenshown by factor analysis that this general kind ofresponsiveness to social norms bears little or norelation to hypnotic susceptibility (Moore, 1964).

50 KENNETH BOWERS

explicable on the basis of demand charac-teristics alone, thereby minimizing the pos-sibility of isolating any special effects thehypnotic state may have had. Whether hyp-notic subjects behaved the way they didsimply as a response to demand variables orwhether for uniquely hypnotic reasons couldnot, therefore, be determined.

The present study obviated this interpre-tive dilemma by making critical alterations inthe hypnotic-simulator paradigm usually em-ployed. In previous studies, the influence oftrance and task demand variables has been inthe same direction. This merger confoundedtheir effects in a way that obscured anyinfluence the trance variable may have pos-sessed. In this study, therefore, the directionsof influence of trance variables and subse-quent demand variables were purposely an-tagonized, thereby allowing a legitimate in-ference from the dependent variable behaviorconcerning which of the variables was pre-potent.

The following is a summary account of therevamped hypnotic-simulator paradigm as itwas employed in this experiment. Neither thehypnotic subjects nor the simulating subjectsreceived any specific information about howto behave after the announced conclusion ofthe experiment. Thus, demand characteristicsinherent in the initial suggestions, and spe-cific to the dependent variable behavior wereminimized. The subsequent termination ofthe experiment, together with a directive tobe honest despite previous suggestions' alteredthe situation for the simulating subjects sothat it became all right for them to "confess.1-The fact that an identical alteration in thedemand characteristics did not similarly af-fect the hypnotic subjects is, of course,the critical finding of this investigation. Theeffects of receiving suggestions under hypnosisevidently take precedence over subsequent,countermanding alterations in the demandcharacteristics. It is reasonable to concludethat hypnotic behavior is not wholly re-ducible to acting in accordance with demandcharacteristics, and that hypnosis seems inpart to be an altered state within which sug-gestions have a peculiarly potent effect.

REFERENCES

As, A. Hypnotizability as a function of nonhypnoticexperiences. Journal of Abnormal and Social Psy-chology, 1963, 66, 142-150.

AUSTIN, MARGARET, PERRY, C., SUTCLIPFE, J. P., &YEOMENS, N. Can somnambulists successfullysimulate hypnotic behavior without becoming en-tranced? International Journal of Clinical andExperimental Hypnosis, 1963, 11, 175-186.

BARBER, T. X. Towards a theory of hypnosis: Post-hypnotic behavior. Archives of General Psychiatry,1962, 7, 321-342.

BARBER, T. X., & CALVERLEY, D. S. "Hypnotic be-havior" as a function of task motivation. Journalof Psychology, 1962, 54, 363-389.

BARBER, T. X., & CALVERLEY, D. S. An experimentalstudy of "hypnotic" (auditory and visual) hal-lucinations. Journal of Abnormal and Social Psy-chology, 1964, 68, 13-20. (a)

BARBER, T. X., & CALVERLEY, D. S. Toward atheory of hypnotic behavior: Effects on suggesti-bility of defining the situation as hypnosis anddefining response to suggestions as easy. Journalof Abnormal and Social Psychology, 1964, 68,585-592. (b)

BARBER, T. X., & HAHN, K. W. Physiological andsubjective responses to pain producing stimulationunder hypnotically-suggested and waking-imagined"analgesia." Journal of Abnormal and Social Psy-chology, 1962, 65, 411-418.

CRUTCHPIELD, R. S. Conformity and character.American Psychologist, 1955, 10, 191-198.

ERICKSON, M. H. An experimental investigation ofthe hypnotic subject's apparent ability to becomeunaware of stimuli. Journal of General Psychol-ogy, 1944, 31, 191-212.

ERICKSON, M. H. Deep hypnosis and its induction.In Le Cron (Ed.), Experimental hypnosis. NewYork: Macmillan, 1952.

ERICKSON, M. G., & ERICKSON, E. M. Concerningthe nature and character of post-hypnotic be-havior. Journal of General Psychology, 1941, 24,95-133.

FISHER, S. The role of expectancy in the perform-ance of post-hypnotic behavior. Journal of Ab-normal and Social Psychology, 1954, 49, 503-507.

LONDON, P., & FUHRER, M. Hypnosis, motivation,and performance. Journal of Personality, 1961,29, 321-333.

MARCUSE, F. L. Hypnosis: Fact and fiction. Balti-more: Pelican, 1959.

MILLER, G. A., GALANTER, E., & PRIBRAM, K. H.Plans and the structure of behavior. New York:Holt, 1960.

MOORE, ROSEMARIE K. Susceptibility to hypnosisand susceptibility to social influence. Journal ofAbnormal and Social Psychology, 1964, 68, 282-294.

ORNE, M. T. The nature of hypnosis: Artifact andessence. Journal of Abnormal and Social Psy-chology, 1959, 58, 277-299.

HYPNOTIC BEHAVIOR 51

ORNE, M. T. On the social psychology of the psycho-logical experiment: With particular reference todemand characteristics and their implications.American Psychologist, 1962, 17, 776-783.

ORNE, M. T., & SCHEJBE, K. E. The contribution ofnondeprivation factors in the production ofsensory deprivation effects: The psychology of the"panic button." Journal of Abnormal and SocialPsychology, 1964, 68, 3-12.

ROSENTHAL, R. Experimenter outcome-orientationand the results of the psychological experiment.Psychological Bulletin, 1964, 61, 405-412.

SHOR, R. The frequency of naturally occurring"hypnotic-like" experiences in the normal collegepopulation. International Journal of Clinical andExperimental Hypnosis, 1960, 8, 151-163.

SHOR, R. E., & ORNE, E. C. Harvard Group Scaleof Hypnotic Susceptibility. Palo Alto, Calif.:Consulting Psychologists Press, 1962.

Srois, B. Are there hypnotic hallucinations? Psycho-logical Review, 1906, 13, 239-259.

TAFFEL, C. Anxiety and the conditioning of verbalbehavior. Journal of Abnormal and Social Psy-chology, 1955, 51, 496-501.

WHITE, R. W. A preface to the theory of hyp-notism. Journal of Abnormal and Social Psychol-ogy, 1941, 36, 477-505.

WEITZENHOFFER, A. M. Hypnotism: An objectivestudy in suggestibility. New York: Wiley, 1953.

WEITZENHOFFER, A. M. General techniques of hyp-notism. New York: Grune & Stratton, 1957.

WEITZENHOFFER, A. M. The nature of hypnosis,Part II. American Journal of Clinical Hypnosis,1963, 6, 40-72.

WEITZENHOFFER, A. M., & HILGARD, E. H. StanfordHypnotic Susceptibility Scale, Form C. Palo Alto:Consulting Psychologists Press, 1962.

(Received February IS, 1965)