As this is not a paper for publication and is largely based on talks I have given, I have been economical with references.
There are several contexts in which hypnosis, in the formal sense, can be said to be taking place. Four important contexts are the clinical setting, the research laboratory, places of entertainment, and training courses. In each of these, and in other contexts, the aims of hypnosis and the underlying assumptions made about it tend to differ to the extent that, from the standpoint of psychology, we might be talking about different phenomena.
Let's start by considering what is traditionally termed 'academic' or 'experimental hypnosis' - hypnosis that is undertaken with normal volunteer participants for the purposes of academic research. It is mainly the outcome of such research that informs the scientific theories of hypnosis that have been developed over the last few decades.
The defining features of experimental hypnosis may be simply expressed as follows:
The hypnotist repeats suggestions, ideas and imagined scenarios to the subjects who, if they are sufficiently suggestible, are able to respond in a manner they experience as realistic and effortless.
So the first property of hypnotic responding is that it is a very real experience for the subject, often very vivid and intense. In response to even a simple suggestion such as arm levitation (the idea that their arm is becoming so light that it will float upwards), some subjects are quite surprised by their response. Indeed, with suggestions that call for quite unusual experiences, such as positive and negative hallucinations or changes in personal identity, the effect on the subject can be quite profound.
The second property of true hypnotic responding related to the above is that it seems to the subject that the experience is happening to them; it does not seem that they are using some deliberate strategy to produce the experience. For example, in the case of a suggestion of amnesia, it does not seem to the responsive person that they are deliberately trying to block the material from coming into consciousness; the amnesia appears to happen automatically. Similarly, when the person is responding successfully to a suggestion of anaesthesia in the hand, they do not report, for example, that they are deliberately distracting themselves when the experimenter is testing their awareness of a tactile stimulus. When the person responds to a suggestion of arm levitation, it does indeed seem to them that their arm is rising without any effort on their part.
So here is the good news. No serious researchers in the field of hypnosis now doubt that responsive hypnotic subjects are having very realistic experiences. Therefore theories of hypnosis have to explain the realistic and effortless nature of the experiences of subjects when responding to hypnotic suggestions.
Now there is one further property of hypnotic suggestibility that may help in the construction of such theories. I shall not describe this just yet, but let us see if we can deduce what it is from the following experimental studies.
In a series of investigations Martin Orne (1,2) gave hypnotic subjects the suggestion that when they opened their eyes they would not be aware of the chair in front of them. Some responsive subjects did not report seeing the chair, yet when they were asked to stand up and walk forward, they avoided the chair. On the other hand, subjects who were asked to pretend that they were deeply hypnotized were more likely to collide with the chair. In another experiment, subjects who were able to 'hallucinate' a person sitting opposite them in a real chair were also able to describe parts of the chair which would normally have been obscured by its occupant. Simulators were more inclined to deny being able to do this. Finally Orne (3) described how he age-regressed a subject to a period in his life when he was only able to speak and understand German; when asked several times in English whether he understood English (using a different phraseology each time) he replied, 'Nein'. (More common is the observation that age-regressed subjects respond without difficulty to questions and instructions that, in reality, they would not have understood at the age to which they are regressed.) Orne called this phenomenon 'trance logic'.
Trance logic and the associated experimental demonstrations have drawn criticism both on methodological and theoretical; it is not typical of all hypnotically suggestible subjects and may be evident in non-hypnotic contexts. One reason for this may be that responding to hypnotic suggestions involves a complex interplay of a range of psychological processes, so people differ widely and idiosyncratically in how they respond and how they report their experiences. Nevertheless the above demonstrations exemplify a theme that has emerged from investigations of suggested experiences over a range of modalities.
Consider, for example, what happens when it is suggested to subjects that they cannot hear their own voice. Some very susceptible subjects do report that they are unable to hear their voice. But what do you suppose happens under the condition of delayed auditory feedback, when their speech is played back to them over earphones with a very slight delay? Usually people find it very difficult to speak normally if they hear their speech output delayed in this way. What about hypnotic subjects who can no longer hear their voice? Is their speech should also affected? Yes! (4).
Something equivalent to this has been demonstrated in the visual modality. For example some researchers first suggested to subjects that they were blind and then showed them a list of words with unusual spellings. Some subjects reported that they could not see the words. When they were later given a spelling test their performance was clearly influenced by the words they had been shown earlier but which they denied being able to see (5). This may remind you of Orne's demonstrations in the case of positive and negative visual hallucinations and we shall shortly see that the two phenomena may be related.
By now you may be detecting a common theme or process underlying these findings. Is this theme present in the case of suggested post-hypnotic amnesia? Indeed it is. For example, subjects are first shown or listen to a list of words to remember and are then given suggestions for post-hypnotic amnesia for the word list. Some now deny being able to recall the words, yet appropriate electroencephalographic (EEG) changes are still detected when they are presented with items from the word list for which they deny any conscious recall; these EEG changes do not occur with new material (6,7). Rather more convincingly, the supposedly 'forgotten' material still interferes with the recall of a word list that was presented earlier that was not included in the amnesia suggestion (8). In other words you can still have proactive and retroactive interference from words that the subject was supposed to have 'forgotten'. Indeed, the 'forgotten' words influence performance on other tasks such as word associations and the completion of word fragments. This is now a well-established finding (9).
So what is the common theme underlying this experimental research. Clearly in the case of post-hypnotic amnesia explicit memory (i.e. conscious recall of the material) is impaired but not implicit memory.
More generally we can say that when responding to suggestions, responsive subjects may be able to inhibit from consciousness perceptual and cognitive experiences that they normally would be aware of but they still respond to these unconsciously or implicitly. So, for example, they cannot consciously see the chair in front of them, or hear their own voice, or see the word list, or recall something they have just heard but they automatically respond as though they have done.
This disconnection between explicit and implicit awareness is not unheard of in psychology. For example, in the case of the psychiatric condition hysterical blindness, patients appear blind to their surroundings yet may successfully negotiate their way around furniture whose position in the room they have no way of knowing, likewise patients with certain lesions in the visual context.
Hence it may be that subjects who respond successfully to the hypnotist's suggestion are able to employ some kind of dissociative or inhibitory mechanism that operates within the perceptual-cognitive-behavioural process. This inhibits from conscious expression some experience that the subject would normally be aware of. I shall illustrate this in a moment.
This idea forms the basis of several highly technical neurocognitive theories of hypnosis. One of the earliest and most influential of these is due to Ernest Hilgard who in 1976 wrote an account of this in a book titled Divided Consciousness. He called this 'neodissociation theory' because he construed the mechanism of dissociation as a normal function, manifested when, for instance, one assigns attentional priority to particular activities or experiences while unconsciously engaging in other activities or being unaware of other experiences potentially available to consciousness. Examples are driving a car while holding a conversation with a passenger (at times concentrating on one while automatically and successfully executing the other) and not being aware of pain while attending to an emergency. It may be that responsive subjects are able to deploy this kind of mechanism in accordance with the requirements of the hypnotist's suggestions. Thus when responding, for example, to an arm levitation suggestion the subject is consciously aware that the arm is rising but is unaware of the effort of lifting it; the arm thus appears to them to be moving 'on its own'. Or when responding positively to a suggestion of glove anaesthesia, the subject withdraws attention from the experience of their hand being touched but the cognitive effort required is, unlike in other circumstances, outside of awareness. More recent theories are summarised by Heap (10).
Of central importance to experimental hypnosis is the fact that people vary in their hypnotic suggestibility and this characteristic can be reliably measured using standardised scales. There are several such scales each consisting of a number of different types of suggestion that the investigator reads out, and the subject's response to each is assessed subjectively and/or objectively. The scores are summed to yield a total score that is the person's suggestibility as measured by that scale. Suggestibility scores in the general population bear a rough approximation to a bell-shaped distribution.
Theories of hypnosis must account for several characteristics of hypnotic suggestibility, such as:
At this point we should remind ourselves that the origins of hypnosis lie not in the experimental psychology laboratory but in the medical or healing context. Modern hypnosis may trace its lineage to the ideas and practices of Franz Anton Mesmer in late 18th century France. This is true of both experimental and clinical hypnosis, which I am going to discuss now. However, clinicians prefer a more remote ancestry in the sleep temples of ancient Egypt and Greece.
Let us first remind ourselves that hypnosis as construed by the experimentalists and theorists is a normal psychological process in which the hypnotist provides suggestions, ideas and imagined scenarios to which susceptible subjects are able to respond in a manner that seems to them to be realistic and effortless or involuntary.
Now, this may also describe a typical session of hypnosis in the clinic. The clinician or therapist may administer all the hypnotic suggestions that the experimentalists investigate and obtain the same results. However, just doing that is not treatment or therapy and is not what is meant by clinical hypnosis or hypnotherapy. The assumptions and the goals of clinical hypnosis are very different from those of experimental hypnosis.
The central concept in what we are now calling clinical hypnosis is that during hypnosis the patient is in some kind of psychological and physiological state termed a 'trance' that is advantageous for therapy and healing. In the clinic, the trance state is nearly always produced by a traditional hypnotic induction. This consists of series of suggestions and imagery intended to encourage the subject to physically and mentally relax and to maintain an inner focus of attention while remaining alert to the hypnotist's communications. Traditionally it has been assumed that by this means the subject enters an altered state of consciousness or trance. He or she remains in this trance state until alerted at the end of the hypnosis session.
While in this trance state that the subject is very responsive to suggestion (we might even say that true hypnotic responding requires the subject to be in a trance state). The more profound the suggested alteration in experience, the deeper the trance has to be. So, for example, a response to the suggestion that the extended arm is becoming too heavy to be kept in the horizontal position only requires the subject to be in a light trance; a suggestion of a negative visual hallucination requires the subject first to enter a deep trance. And so on. People vary in the trance capacity and hence in their responsiveness to various suggestions.
Hyper-suggestibility is not the only property that has been traditionally claimed for the hypnotic trance. Other alleged properties have included:
It is generally acknowledged that many of the claims made about the hypnotic trance state have not withstood experimental scrutiny. More could be said about spontaneous post-hypnotic amnesia and pain but not here. Note that 'hyper-obedience or automatism' is not the same as effortlessness of responding mentioned earlier. 'Communication with the unconscious mind' is a property still claimed by hypnotherapists.
So, hypnosis as understood by clinicians could be described as:
An altered state of consciousness usually induced by suggestions of mental and physical relaxation and inner absorption, during which the patient is more receptive to the clinician's therapeutic communications and which facilitates access to the patient's unconscious.
This is clearly a much broader and far-reaching conceptualisation than that of experimental hypnosis
Let us study the three assumed properties of hypnosis as outlined above.
From now on you are going to feel more confident and more able to stand up for yourself.
As soon as you become aware of these feelings of anxiety, you will take in a deep breath and as you breath out you will experience a sense of calmness and relaxation…
Each and every time you think of having a cigarette, the image of this diseased lung will immediately come to mind and you will immediately put down the cigarette
These suggestions are repeated several times and may be rehearsed in imagination.
Now you could simply make these statements in a convincing manner to the patient without doing any hypnosis. But it is assumed that with hypnosis, the patient is in a special state of mind that is particularly receptive to these therapeutic communications.
What is the situation regarding experimental hypnosis?
Following a typical relaxation induction in the laboratory, on average there is indeed an increase in suggestibility as measured by suggestibility scales. However, it seems that the increase is on the whole rather modest and does not occur for everyone, some individuals even being less suggestible following an induction (perhaps because they are averse to the idea of hypnosis). For example, Kirsch & Braffman (11) found that following a hypnotic induction:
Hence the gains in suggestibility scores are modest. But why do these gains occur at all? Is it because the person has been put into a trance by the induction?
Well suppose, prior to administering a suggestibility scale we substitute the traditional relaxation induction with 'motivating suggestions'. This is a series of instructions to the effect that we are shortly going to ask subjects to bring to mind certain ideas and situations and it is most important that they concentrate and use the full powers of their imagination to create the suggested images, the feelings, and the experiences as vividly as possible. This procedure usually leads to increases in suggestibility scores equivalent to those when a classic induction procedure is used (12).
Now let's do 'an induction' that is the complete opposite of the traditional procedure. Instead of telling the subjects they are feeling more relaxed, tell them they are feeling more energetic! Say that, instead of narrowing their attention on their inner experiences, they are becoming more alert and aware of everything that is happening around them, as though their mind is 'expanding'. Under such conditions, suggestibility scores increase overall by the same margin as when a traditional induction procedure is used with just subtle differences amongst some individual items. The same effects are even observed when, in place of a comfortable chair or couch, participants are provided with an exercise bicycle and told to pedal continuously as you administer your 'alert induction' as it is called (13).
Laboratory experiments have even obtained the same results using 'sham' inductions such as inhalation of a gas from a cylinder (actually air) 'that has a powerful effect on suggestibility' or a pill with 'hypnosis' stamped on it (14).
What does all this mean? It seems that for experimental hypnosis the effect of the induction on suggestibility is to prepare or prime subjects to respond to the suggestions to follow - that is, to enhance their involvement and expectation. You can do this with a traditional sleep or relaxation procedure, but you could choose any of the others described above, or even invent your own.
Despite this, as in the clinic, laboratory investigations of hypnosis still use traditional induction methods. But modern theories of hypnosis do not require the concept of trance although one theory (the theory of dissociated control) retains the hypnotic state as an explanatory concept, probably needlessly (10).
As a rejoinder to this, it may be asserted that:
The fundamental issue here is that while responding to suggestions does appear to be a key idea in both experimental and clinical hypnosis, is it the case that suggestion and suggestibility as studied in experimental hypnosis is the same as in clinical hypnosis? Or do they involve different psychological processes?
As we know, suggestions, and in particular posthypnotic suggestions, play a significant role in clinical hypnosis. They are intended to bring about appropriate changes in the way patients behave, think and feel in those situations they find problematical.
However, unlike in the laboratory, the aim of hypnotic suggestions is to bring about desirable of an enduring nature, not changes that last only for the duration of the therapy session (unless the intervention is, say a single-session medical or dental intervention); hence the extensive use of post-hypnotic suggestion in the clinic.
Now, in the laboratory it has been found that quite dramatic changes in a person's behaviour, thinking, emotional state and apparent personality can be effected by suggestion and post-hypnotic suggestion and the response feels very compulsive and automatic to the highly suggestible subject. However there are limitations on their influence once the hypnotic context is terminated. There is not a great deal of direct evidence that, in the experimental context, hypnotic suggestion itself can bring about the kinds of significant and enduring changes in a person's life that are witnessed when a person responds successfully to therapy.
One rejoinder to this is that laboratory studies of suggestion normally consist of one session whereas a course of therapy consists of several treatment sessions. Also the patient is encouraged to practise self-hypnosis and the therapeutic suggestions may be put on a tape or CD for home use. Perhaps the potency of the hypnotic suggestions is thereby strengthened.
So what we need to do is to look for other evidence that hypnotic suggestion in the clinical context is the same as that in the experimental context - i.e. they involve the same psychological mechanisms.
In order to proceed with this question, let's widen the discussion. The term 'suggestion' is used to describe and explain human behaviour in quite a wide range of situations, not just experimental and clinical hypnosis. For each of these contexts we can identify what the key outcome is of the use of suggestion.
In the case of experimental hypnosis, as we have seen the key outcomes are realistic and effortless responses to the suggestions administered.
What about everyday examples of persuasion, the outcome of which, as a result of verbal suggestions, is that a person changes their beliefs or agrees to do something that initially they resisted (e.g. making a purchase or committing an antisocial act)?. Is the persuader's use of suggestion related to hypnotic suggestion as studied by experimental hypnotists?
Now consider interrogative suggestibility. In this case the context is the coercive police interview of a suspect. The key outcome is a false confession to the offence under investigation (or false incriminating statements) elicited from the suspect. Does suggestion in this context involve the same psychological mechanisms as those for experimental hypnosis?
How about placebo, the therapeutic benefit that accrues from simply providing the patient with some treatment, regardless of whether this of itself has any effect on the patient's condition? The key outcome here is some improvement in the problem being treated (improved mood, reduced anxiety, better sleep, fewer headaches, and so on. The placebo effect is often ascribed to 'suggestion'; does this involve the same psychological processes as suggestion in the hypnosis laboratory?
Now think about the charismatic, evangelistic preacher who, after a powerful, riveting speech to his or her congregation summons them to 'Come forward! Come forward! If you wish to devote your life to Jesus, come forward!' Clearly the key outcome is for people to rise from their seats and come to the front. Is the preacher using hypnotic suggestion in the same way as the experimental hypnotist?
Then we have stage hypnosis. Clearly the intended outcome of the stage hypnotist's suggestions is that the volunteers behave in a dramatic and hilarious manner for the entertainment of the audience.
And finally we return to hypnotic suggestions in the clinic, where the key outcome is (as with placebo) beneficial changes in the condition being treated. Are hypnotic suggestions in the clinical equivalent to those studied in the laboratory?
How can we address the above questions about the relationship of suggestion in each of these contexts to hypnotic suggestion in the laboratory? One way we can do this is to see whether the outcomes identified in each case are related to the hypnotic suggestibility of the individuals concerned as measured in the laboratory.
In the case of experimental hypnosis we can be confident the measured outcome of a suggestion (a realistic and seemingly effortless experience of the suggested effect) will be significantly related to the participants' scores on a hypnotic suggestibility scale that does not include that suggestion.
Likewise if persuasion were the same as or similar to hypnotic suggestion as defined by experimental hypnosis then we would expect the outcome of such a persuasive influence, namely the targeted change in belief, opinion or behaviour of the person concerned, to be related to their hypnotic suggestibility score. In fact the evidence suggests that no such relationship exists.
Concerning interrogative suggestibility, there have been no studies directly investigating the hypnotic suggestibility of known false confessors. Indirect evidence from measures of interrogative suggestibility presents a mixed picture, but the bulk of the research again indicates no strong relationship (17).
Investigation of the relationship between hypnotic suggestibility and positive response to placebo is very limited. However, evidence suggests that responding to placebo analgesia involves different mechanisms to responding to hypno-analgesia.
There has been very little systematic research into stage hypnosis. Participants are initially selected by the entertainer for their positive responding to simple suggestions such as being unable to separate their clenched hands. This means that the participants who find their way to the stage should be somewhat higher than average in their hypnotic suggestibility (and not necessarily extremely suggestible) and this was the finding of one study of just one stage hypnosis show (16). Some deselection may take place as the performance progresses: recall that the desired outcome is for a hilarious, over-the-top performance; the personality traits and other characteristics required for this outcome may have nothing to do with hypnotic suggestibility.
Last, and most importantly, there is clinical hypnosis. Are the outcomes of hypnotherapeutic interventions related to the hypnotic suggestibility of the patients and clients as is measured by experimental hypnotists? Although there are clinical studies of hypnotic treatment of different problems that do report a significant correlation, this tends to be weak and many other studies have found no relationship. There is some evidence that the relationship is slightly stronger for the treatment of children and for acute pain.
A meta-analysis of some of the better conducted clinical trials of hypnosis for a small range of medical and mental health problems (18) concluded that overall, hypnotic suggestibility accounted for only 6 per cent of the variation in clinical outcome. (Surprisingly some experimentalists still admonish their clinical colleagues for not administering suggestibility scales to their patients prior to embarking on treatment.) This evidence suggests that outcome of clinical hypnosis is determined by factors unrelated to those that are associated with experimental hypnosis.
There is a second piece of evidence that supports this conclusion. Other than is the case with experimental hypnosis, in most if not all of the contexts mentioned earlier in which suggestion is deemed to be at work, the perceived attributes of the individual using suggestion and his or her relationship with the other person have an important effect on outcome. This is not so with hypnotic suggestions administered in the laboratory.
Over the years, it's become clear that hypnosis as investigated in the laboratory and hypnosis as practised in the clinic as a therapeutic medium have grown so far apart that now there remains only a small area of overlap. This overlap may be, for example, where clinical hypnosis is used in a single session to assist a patient cope with the pain and anxiety of some medical or dental procedure, rather than in multi-sessional applications of hypnotherapy. It is clear that the model used to inform the practice of clinical hypnosis in which the establishment of a trance state is central to positive responding to the therapists communications, does not apply to experimental hypnosis.
Experimental hypnosis has come on strongly, particularly in the last 20 years. Clinical hypnosis has not done so. Demonstrating the effectiveness of clinical hypnosis is one thing, but it is vital that we acknowledge that it is different from experimental hypnosis because equating the two versions of hypnosis is unworkable and it leads to so much confusion. If the clinicians and therapists can liberate themselves from having to answer to their academic colleagues they can then develop their own coherent theory of clinical hypnosis. The assumptions of clinical hypnosis that I summarised earlier indicate where much of the research should be directed: e.g. the nature of the therapeutic trance and the claim of facilitated access to 'the unconscious'. This will help establish clinical hypnosis on a sounder theoretical basis. Of special interest is assessing individuals for their likely responsiveness to hypnotherapeutic procedures given the failure of standard hypnotic suggestibility scales that have proved so effective in the investigation of experimental hypnosis.