From: Haddock, G. & Slade, P.D. (Eds.) Cognitive-Behavioural interventions with Psychotic disorders, London: Routledge, ps. 75-85


Chapter 4


Paul Chadwick and Max Birchwood




Auditory hallucinations are traditionally associated with a diagnosis of schizophrenia. In the World Health Organisation’s International Pilot Study of Schizophrenia (WHO, 1973) auditory hallucinations were reported by 73 per cent of people diagnosed as having an acute episode of schizophrenia, yet they can be reported by individuals who have been sexually abused, or suffered a bereavement, as well as by individuals diagnosed as having a manic depressive illness or an affective psychosis. Indeed, because they feature in many different disorders, the diagnostic importance of auditory hallucinations has been doubted (Asaad and Shapiro, 1986).


In addition, it appears that auditory hallucinations are not restricted to clinical groups. Auditory hallucinations can be reported by individuals who, whilst showing signs of a specific clinical disorder, display insufficient evidence for a firm diagnosis to be made (Cochrane, 1983). Again, it appears that under laboratory conditions many ordinary people display a propensity to report hearing sounds which are not there, prompting researchers to speculate that proneness to hallucinate may be a predisposition spread across the general population (Slade and Bentall, 1988). Current opinion in psychology veers towards accepting the possibility that hallucinations lie on a continuum with normality (Strauss, 1969).


The auditory hallucination itself can be a noise, music, single words, a brief phrase, or a whole conversation. The present chapter is concerned only with voices, that is, hallucinations which are experienced as someone talking. The experience of hearing voices is a powerful one that demands a reaction. However, the experience is also very personal. Whilst it is known that a common first reaction to voices is puzzlement (Maher, 1988), individuals evolve different ways of interacting with their voices. Certain people, for example, experience voices as immensely distressing and frightening and will shout and swear at them. In contrast, other individuals might find their voices reassuring and amusing and actually seek contact. Again, in the case of imperative voices, many individuals desperately resist the commands, and comply only at times of great pressure, whilst others comply willingly and fully.


Viewing voices from a cognitive perspective


This diversity in the way in which individuals relate to their voices illustrates the point that voices are not necessarily a problem to the individual concerned; indeed, it is fairly common for individuals to believe their voices to be a solution to a problem. This in turn draws attention to the point that the serious disturbance associated with voices, as with so many other symptoms, tends to be located in the way in which an individual feels and behaves. People who hear voices are typically referred to our service because they are desperate, depressed, angry, suicidal, helpless, harming themselves, isolated, violent, etc. This point is implicit in traditional treatment approaches, which have usually been directed at easing distress and altering behaviour ( e.g. methods of anxiety reduction, punishment procedures) as well as at eliminating the hallucinatory experience (medication, earplugs, headphones). Such treatments were based on the premise that a particular individual's coping behaviour and affect followed necessarily from the nature of his or her hallucination:


“If the voice attacks the patient he or she is depressed and suicidal. If the voice tells the patient to kill others, then, if the patient loses self-control, murderous attacks on others are likely. If the voice tells the patient he or she is wonderful and powerful, grandiose and manic behaviours appear. . . . In other words, the content of the hallucination is directly responsible for the salient and anomalous behaviours associated with schizophrenia.” (Benjamin, 1989, p. 293)


However, this explanation may be too simple. Research has shown how voices with similar contents may evoke different coping behaviour (Tarrier, 1992). Also, an ingenious study by Romme and Escher (1989; also see Chapter 8) has revealed how voices frequently do not evoke a sufficiently strong reaction to bring the individual to the attention of services, even when the content is extremely serious. It would appear that the nature and strength of an individual's response to voices is mediated by psychological processes.


Research on ordinary people indicates the likely extent of this mediation. For example, in Milgram's famous studies, whether subjects could be persuaded to administer what they believed to be a lethal electric shock to other subjects was strongly influenced by their beliefs about the experimenter's authority and power, their own degree of control, and the presumed consequences of disobedience (Milgram, 1974). Likewise, it is possible that the degree of fear, acceptance and compliance shown to voices might be mediated by beliefs about the voices' power and authority, the consequences of disobedience, and so on. For example, an individual who believes his voice to come from a powerful and vengeful spirit may be terrified of the voice and comply with its commands to harm others; if the same voice were believed to be self-generated, however, terror and compliance would be unlikely.


In other words, voices might be viewed from a cognitive perspective. The defining feature of the cognitive model within clinical psychology is the premise that people's feelings and behaviour are mediated by their beliefs, and therefore are not inevitable consequences of antecedent events.


Applicability of the cognitive model to voices


If the cognitive perspective is to be applicable to voices, two hypotheses must be supported. One is that the cognitive model will make sense of why individuals respond in such different ways to their voices; specifically, diverse affective and behavioural responses must be understandable by reference to differences in the beliefs individuals hold about their voices. Second, the cognitive model needs to add to our understanding of voices. That is to say, if differences in voice content accounted fully for people's diverse behaviour and feelings, then from an explanatory point of view the cognitive model would be superfluous ( although it could still have important strategic implications for treatment).


In a recent experiment (Chadwick and Birchwood, 1994), we found support for both hypotheses. We interviewed twenty-six people who had heard voices for at least two years in order to assess their behavioural, cognitive and affective responses to persistent voices. All participants met DSM-III-R criteria for schizophrenia or schizoaffective disorder (APA, 1987). All except one were receiving depot neuroleptic medication at All Saints Hospital, Birmingham; one was in hospital and the remainder were outpatients. Participants volunteered for the study and there were no refusals.


Information was gathered using a semi-structured interview. This covered: formal properties of the voices, including content; beliefs about the voices' identity, power and purpose, and about the consequences of compliance; collateral symptoms that were regarded as supporting the beliefs; other confirmatory evidence; and influence over the voice. Confirmatory evidence referred to actual occurrences that were perceived to support a belief; for example, the belief that voices give good advice would be strengthened if complying with a command led to a desired outcome. Influence concerned whether the individual could determine the onset and offset of the voice, and could direct what it said. Also, the behavioural and affective responses were elicited. It usually took more than one interview to collect all relevant information.


Beliefs about voices: omnipotence, malevolence and benevolence


All voices were believed to be extraordinarily powerful, or omnipotent, and this belief seemed to be supported by four types of evidence. First, nineteen individuals (73 per cent) reported collateral symptoms that contributed to the sense of omnipotence. One man, for example, was commanded by his voice to kill his daughter; he recalled one occasion when she was standing by an open window and he experienced his body being moved towards her. A second man heard a voice telling him that he was the son of Noah, and occasionally when he heard his voice he experienced concurrent visual hallucinations in which he was dressed in a white robe and walked on water. Second, eleven people (42 per cent) gave examples of how they attributed events to their voices, and then cited the events as proof of the voices' great power. Thus, although two individuals cut their wrists under their own volition, both subsequently deduced that the voices had somehow made them do it. Similarly, one man attributed responsibility for having sworn out loud in church to his satanic voices. Third, twenty-one people (81 per cent) were unable to influence either the onset and offset of their voices or what was said, once again suggestive of the voices' power.


Finally, all voices gave the impression of knowing all about people's past histories, their present thoughts, feelings and actions, and what the future held. Frequently voices would refer to behaviour and thoughts of a highly personal and emotive nature, such as a criminal act or personal weakness, which the individual feared others knowing. Perhaps because of this lack of privacy , individuals would often attribute more knowledge to the voice than the content actually displayed; for instance, general statements like "We know all about you" were thought to refer to specific actions. Understandably, this appearance of omniscience left many individuals feeling exposed and vulnerable.


However, because in our sample a belief in omnipotence was ubiquitous, it would not account for differences in behaviour and affect. On the basis of their beliefs about the voices' identity and purpose, people saw individual voices as being either malevolent or benevolent. Thirteen people believed their voice or voices to be malevolent. Beliefs about malevolence took one of two forms: either that the voice was a punishment for a bad deed, or that it was an undeserved persecution. For example, one man believed he was being punished by the Devil for having committed a murder, and another man believed he was being persecuted without good reason by an ex-employer. Six people believed their voices to be benevolent. For example, one woman believed that she heard the voice of a prophet who was helping her become a better mother and wife, and one man believed that the voices were from God and there to help develop a special power. Four people believed they heard a mixture of benevolent and malevolent voices; paradigmatic of this group was a man who was tormented by a group of evil space travellers on the one hand and yet protected and nurtured by a guardian angel on the other.


Three people were uncertain about their voices because of an inconsistency or incongruity in what was said. Uncertainty was defined as having a strong doubt about the voices' identity, meaning or power, where this doubt was the result of the person's deduction. For example, one man was certain that his voices wanted to help, but observed that they had got things wrong: they wanted him to kill himself and move on to the next and better life, yet his religion told him that suicide is a sin and those who commit it go to hell.


Connection between beliefs, coping behaviour and affect


The behavioural responses to voices were organised into three categories:


Engagement was defined as elective listening, willing compliance, and doing things to bring on the voices ( e.g. watching television, being alone, calling up voices).


Resistance was defined as arguing and shouting ( overt and covert), non-compliance or reluctant compliance when pressure is extreme, avoidance of cues that trigger voices, and distraction.


Indifference was defined as not engaging with the voice.


Without fail, those people who believed their voices to be benevolent engaged with them and those people who believed their voices to be malevolent resisted them. Those people who were uncertain about their voices displayed no clear pattern between beliefs and behaviour.


Affective responses to voices corresponded very closely to behavioural responses. All seventeen 'malevolent' voices habitually provoked a combination of negative emotions (anger, fear, depression, anxiety). Ten of the eleven 'benevolent' voices habitually provoked positive emotions (amusement, reassurance, calm, happiness) when they spoke; the one exception was a voice which issued warnings about impending danger and provoked anxiety. All three people who were uncertain about their voices experienced negative affect when these voices spoke.


In order to establish the reliability and validity of these concepts, we have developed a thirty-item Beliefs About Voices Questionnaire (BAVQ; see Appendix 4.1) to measure malevolence (six items), benevolence (six items), resistance (nine items), engagement (eight items) and power (one item). A statistical analysis conducted on a preliminary sample of sixty completed questionnaires has shown the BAVQ to be both reliable and valid (for details see Chadwick and Birchwood, 1995), and has established provisional scoring guidelines to define malevolence (a score of four or more), benevolence (three or more), engagement (five or more) and resistance (six or more). Data from this study strongly supported our proposed connections between, on the one hand, malevolence, distress, positive affect and resistance, and on the other, benevolence and engagement (see Table 4.1).


There is currently much debate about the connection between paranoid thinking and depression. Zigler and Glick (1988) have proposed that paranoia might be a defence against low self-esteem and that certain forms of paranoid schizophrenia might be a defended depression. In this respect we were struck by our finding that voices construed as persecutory (i.e. malevolent) appeared to provoke more depressive reactions than voices construed as benevolent. Therefore, those individuals who completed the BAVQ also completed the Beck Depression Inventory (Beck and Greer, 1987). The results were a striking confirmation of our earlier interview data. Taking a score of fifteen or more to be indicative of clinical levels of depressive symptomatology, and a cut off of three items or more on the subscales of malevolence and benevolence, we found that 'depression' was indeed much more common and severe with malevolent, not benevolent, voices (see Table 4.2). Thus, it might well be that paranoia defends against low self- esteem, but it appears not to defend against depression (Chadwick and Lowe, 1994).


Connection between malevolence, benevolence, and voice content


Having found that differences in coping behaviour and distress were rendered understandable by reference to beliefs about malevolence and benevolence, it remained to be shown that voice content could not account for these differences with equal clarity. In other words, the distinction between malevolence and benevolence needed to say something about the maintenance of voices which could not be said by inspecting voice content alone.


It is clear that there is a link between voice content and the person's associated feelings and behaviour, and therefore that in many cases resistance and engagement might have been predicted on the basis of content. However, the class of belief was not always understandable in light of voice content alone; that is, in eight cases (31 per cent) the beliefs appeared to be at odds with what the voice said. Two voices of benign content were believed to be malevolent; for instance, one of these voices simply urged the individual to 'take care', 'mind his step', and 'watch how he went', yet he believed these words to have been spoken by evil witches intent on driving him mad. The reverse was also true; two voices commanded the hearers to commit suicide, yet both were believed to be benevolent. Three voices commanded the hearers to commit murder (in two instances, of immediate family members), and yet again were believed to be benevolent. Perhaps most strikingly, one woman's voice identified itself as God and yet she disregarded this and believed it to be an evil force.


Weakness of the cognitive model


The explanatory power of the cognitive model was weakest in relation to compliance. In our group the severity of the command, and not beliefs, was the single most important determinant of compliance; there was no compliance with life threatening commands and compliance with mild commands was commonplace. This might be because the relationship with a voice is regulated by wider considerations such as protecting self-esteem and maintaining involvement with others (Strauss, 1989). Certainly, research on ordinary individuals has shown that compliance is affected by the social acceptability of the behaviour in question (Fishbein and Ajzen, 1975). At present we are interviewing people who have acted on serious commands to investigate if such compliance is associated with factors specific to the hallucinatory experience ( e.g. total certainty in the beliefs) or more general predictors of violence (e.g. previous history).



Thus, it appears that individuals' beliefs about voices have an important bearing on how these symptoms are maintained. In our research, the meaning individuals attached to their voices rendered their coping behaviour and affect understandable, when without recourse to the beliefs many responses would have seemed perplexing or incongruous (see also Strauss, 1991). One advantage of the cognitive approach is that it gives individuals time to describe their belief systems, and then uses these descriptions as the starting point for cognitive therapy.




The cognitive approach within clinical psychology ( e.g. Trower et al., 1988) is based on two premises. The major premise states that extreme feelings and behaviour ( e.g. depression and suicide) are consequences of particular beliefs (e.g. '1 am worthless') rather than events (e.g. divorce). The minor premise states that if these beliefs can be weakened using cognitive therapy, then the associated distress and behaviour will diminish.


Cognitive therapy (CT) is now well established as a treatment for a number of non-psychotic disorders (Hawton et al., 1988); for example, using CT to weaken core depressive beliefs seems to be as effective a treatment of depression as prescribing medication (Hollon et al., 1991). More recent evidence has shown its potential in the management of schizophrenia (see Birchwood and Tarrier, 1992 and Chapters 3, 5, 6, and 7). For example, there is growing evidence that some secondary delusions may be weakened using cognitive therapy (see Chadwick and Lowe, 1990).


Traditional treatments for voices have been directly aimed at reducing either the hallucinatory experience (e.g. medication, earplug therapy) or their consequences ( e.g. anxiety reduction methods, punishment procedures). The purpose of using CT for voices is to ease distress and problem behaviour by weakening target beliefs about omnipotence, malevolence or benevolence, and compliance. The possible importance of this new approach is considerable because even the most effective treatment for voices - neuroleptic medication - leaves many voices unchanged (Slade and Bentall, 1988).


The CT we use for people who hear voices draws very heavily on the work of Beck (Beck et al., 1979; Hole et al., 1979; and Ellis, 1962), although we have found it necessary to adapt and develop traditional CT in order to work collaboratively and effectively with individuals who hear voices.


We have found it can be difficult to engage people in CT for voices, because of their powerful beliefs and emotions about their voices. Therefore we have developed a number of strategies to promote engagement and trust. One such strategy is to use our understanding of the connections between malevolence, benevolence, resistance and engagement to anticipate how an individual is likely to feel, think and behave in relation to the voice. This understanding seems to bring individuals a sense of relief. Also, we always inform clients that they may withdraw from therapy at any point without penalty, and this may also reduce anxiety and facilitate engagement. Again, people can meet other hallucinators and watch a video of individuals who have completed therapy successfully discussing their experience; the discovery that others have similar problems, 'universality' (Yalom, 1970), is an important therapeutic process.


The central beliefs are defined early on, together with the evidence used to support them, and we discuss how any distress and disruption attributed to the voices is actually a consequence of the beliefs the individual holds. We emphasise that individuals are free to continue holding their beliefs, and may drop out of therapy at any time; the atmosphere is one of 'collaborative empiricism' (Beck et al., 1979) in which beliefs are considered as possibilities that mayor may not be reasonable.


Disputing a beliefs veracity involves the use of standard cognitive techniques (see Chadwick and Lowe, 1990). At first the evidence for each belief is challenged; this process begins with the piece of evidence the individual rates as least important and progresses to that rated most important. Next the therapist challenges the belief directly. This involves first pointing out examples of inconsistency and irrationality, and second, offering an alternative explanation of events. This alternative is always that the beliefs are an understandable reaction to, and attempt to make sense of, the voices. In our experience this leaves the person searching to understand the meaning of the hallucinations. We conceptualise the voices as self-generated, and try to explore the possible connection, or personal significance, between the voice content and the individual's history.


We use two approaches to test beliefs empirically. On the one hand, we have a set procedure for testing the universal belief '1 cannot control my voices'. First, it is reframed as '1 cannot turn my voices on and off'. The therapist then engineers situations to increase and then decrease the probability of hearing voices. An initial thorough cognitive assessment should identify the cues that provoke voices, and one technique with a high likelihood of eliminating voices for its duration is concurrent verbalisation (Birchwood, 1986). The person rouses and quells the voices several times to provide a complete test.


With all other beliefs the empirical test was negotiated by the client and therapist. It is essential to examine beforehand the implications of the test not bearing out the belief: if the belief will be modified or adapted, or whether the patient has a ready explanation for the outcome that leaves the belief untouched.


A case example


DD was a 41-year-old single unemployed economics graduate with a ten year psychiatric history. For the last three years she had heard voices in half hour bursts, usually in the morning and at bed time. The content was invariably to do with economics, such as 'Infinitely power the rise in inflation', 'Negatively power productivity a million, trillion times'. These and similar statements were usually perceived as commands and occasionally as predictions. DD also held a delusional belief that she could transmit her thoughts using telepathy. Therapy lasted thirteen sessions spaced over six months. Meetings lasted one to two hours.


DD believed the voice to come from the Devil, and that he was using her telepathic power to destroy the British economy. Specifically, the Devil would give a command that in economic terms was disastrous. DD would be compelled to repeat this command and in so doing would unwittingly transmit it telepathically to the Prime Minister, who would act upon it. She believed that if she resisted, then the economy would be saved, but the voice would continue to torment her. In practice, each time the voice began she would resist by saying exactly the opposite of the command, until she finally weakened and repeated the Devil's command, when the voice would stop. She monitored the economy religiously and felt guilt, anger and depression when it dipped.


Two hypothetical contradictions were put to DD. First, she was asked if her beliefs would be altered if she met with the Prime Minister and he assured her that he did not hear her messages. Second, she was asked if her beliefs would be altered if she went out of her way to comply with the commands and the economy was unaffected. DD thought that both these events would weaken her conviction that the beliefs were true.


Whenever the economic news was poor, DD would feel depressed and guilty, inferring 'It is my fault', and she would take the news to be evidence for her beliefs. DD was encouraged to generate and examine an alternative view of events, which was that the main economic indices went up and down regularly and bore no particular relationship to her transmission of messages. This exercise led her not only to question critical evidence for her beliefs, but also to recognise how beliefs influence behaviour and feelings, and are possibly mistaken.


The most obvious inconsistency in DD's beliefs was that on most days she did reluctantly repeat the Devil's commands, and yet the predicted economic disaster had not ensued. Also, there were certain puzzling features in the account; for example, how was the Prime Minister to know that he was to act on the commands, and even should he know this, what does 'infinitely power' require? Again, why should the Devil, an omnipotent being, need to work through DD, as opposed to communicating with the Prime Minister directly?


Following Maher (1988), the alternative explanation offered to DD was that her beliefs arose in response to, and as a way of making sense of, her voice. The beliefs were not labelled as delusions, but were discussed as being a reasonable and reasoned attempt to understand what was a puzzling and alarming experience.


DD first tested the belief that she could not control the voice. She found that by successively reading about economic affairs for five minutes, and then reading aloud material unrelated to economics for five minutes (i.e. concurrent verbalisation), she was able first to increase and then to decrease voice activity. In other words, simple changes in her own behaviour appeared to influence the voice, and this led her to doubt the belief that she could not control it.


In order to test her beliefs about compliance and meaning DD agreed to stop saying the opposite to the voice, and instead, to repeat the commands many times. This process began with comparatively innocuous commands (increasing bus fares and the price of milk) and progressed to the most central (taxation levels and interest rates). Also, with any one command, the principles of systematic desensitisation were applied to reduce DD's anxiety; thus DD would repeat the chosen command several times, rate her degree of distress, relax for a few minutes, and then repeat the procedure again until her distress was negligible. It was agreed that if her beliefs were true then the effects of the tests would appear within two weeks and would be significant, that is, for the belief to be supported bus fares, for instance, would need to double. In all cases the test had no effect and this appeared to weaken DD's beliefs about the consequences of compliance and about the meaning she attached to her voice.


Initially CT worked well for DD. Conviction in all four beliefs about the voice fell significantly and she reported being less fearful, guilty and depressed, and hearing reduced voice activity (see Chadwick and Birchwood, 1994 for full details of this and other cases). By follow-up the extraordinary economic events in Britain in the summer of 1992, including a 5 per cent rise and fall in interest rates in one day, appeared to have undone some of this progress! DD was agitated and her conviction in three beliefs had risen sharply. Since then, earlier gains have been recovered; conviction has once again fallen sharply, DD's voice activity is now far less frequent and, perhaps most importantly, DD is no longer resisting the voice by saying the opposite to the commands.


Personal significance of voices

In our experience, a critical component of CT for voices is often that of drawing out the personal significance of the voices, that is, tentatively making a connection between voice content and beliefs on the one hand, and the individual's history on the other. Indeed, in this chapter we have argued that voices are a powerful experience, one which individuals feel compelled to try and understand. Beliefs about voices are the result of this endeavour and they carry the psychological force of relieving a sense of puzzlement and unease. It has been our experience that whilst many individuals are able to recognise that all or some of their beliefs about voices are mistaken, this can leave them once again struggling to understand the fact that they experience hallucinations.


One possible response open to the therapist is to label the voice a sign of 'illness'. However, there are reasons for not doing this. First, concepts like schizophrenia are of uncertain scientific validity (Bentall et al., 1988) and patients who accept such labels have been found to have a higher incidence of secondary depression (Birchwood et al., 1993). Second, and perhaps of most importance, attributing voices to illness is such an impersonal explanation that it rarely satisfies people. What individuals seem to value is tentatively connecting the content of their own voices to their own histories. For example, in the case of DD, it was suggested that the voice might give a clue as to underlying depressive concerns. The speculation was that DD's childhood experience established a strong need to achieve and earn respect and that as an adult she had experienced a growing sense of failure and inadequacy; indeed, prior to the onset of psychotic symptoms, she had experienced a clinical depression. One effect of her voice was to propel DD into saving the British economy, at great personal cost, and this might be interpreted as both reflecting and offsetting the underlying depressive concerns. As regards this point, DD did experience a drop in mood following the rejection of her beliefs about the voice's identity and purpose; this served to emphasise these depressive beliefs, which were addressed using cognitive therapy.


With other individuals, where there have been grounds to do so, we have made tentative links between voice content and early experience, such as sexual abuse or loss. However, such a link has also been made with more recent experience, such as guilt surrounding abortion. In some instances we have been unable to interpret the significance of the voices.


It is our experience that as beliefs about voices are weakened, underlying themes sometimes emerge spontaneously. We do not interpret this as a reason not to do CT for voices, but rather conclude that certain individuals may require more than symptom-based therapy; however, it is important that such individuals are given the option to address these further issues, with either the same therapist or a new one.




We have argued that viewing voices from a cognitive perspective increases our understanding of the maintenance of voices, and reveals a new treatment approach of considerable promise. However, our research has thrown up puzzling questions to replace those with which we began. For example, if content does not determine whether voices are believed to be malevolent or benevolent, then what does? Also, the cognitive therapy has not been formally evaluated and the limits of its usefulness are unknown. Nonetheless we feel sure that the cognitive model offers clinicians a useful psychological approach to voices, one which will enable them to work collaboratively and effectively to reduce distress and disturbance associated with voices.




American Psychiatric Association (1987) Diagnostic and Statistical Manual of Mental Disorders, revised 3rd edition, Washington, DC: APA.

Asaad, G. and Shapiro, M. D. (1986) 'Hallucinations: Theoretical and clinical overview', American Journal of Psychiatry 143: 1088-97.

Beck, A. T. and Greer, R. (1987) Beck Depression Inventory Scoring Manual, The Psychological Corporation, New York: Harcourt Brace Jovanovich.

Beck, A. T., Rush, A. J., Shaw, B. F. and Emery, G. (1979) Cognitive Therapy of Depression, New York: Guildford Press.

Benjamin, L. S. (1989) 'Is chronicity a function of the relationship between the person and the auditory hallucination?', Schizophrenia Bulletin 15: 291-330.

Bentall, R. P., Jackson, H. F. and Pilgrim, D. (1988) 'Abandoning the concept of schizophrenia: Some implications of validity arguments for psychological research into psychotic phenomena', British Journal of Clinical Psychology 27: 303-24.

Birchwood, M. J. (1986) 'Control of auditory hallucinations through occlusion of monoaural auditory input', British Journal of Psychiatry 149: 104-7.

Birchwood, M. J. and Tarrier, N. (1992) Innovations in the Psychological Management of Schizophrenia, Chichester: Wiley.

Birchwood, M. J., Mason, R., McMillan, F. and Healy, J. (1993) 'Depression, demoralisation and control over illness: A comparison of depressed and non- depressed patients with a chronic psychosis', Psychological Medicine 23: 387-95.

Chadwick, P. D. J. and Birchwood, M. J. (1994) 'Challenging the omnipotence of voices: A cognitive approach to auditory hallucinations', British Journal of Psychiatry 164: 190-201.

Chadwick, P. D. J. and Birchwood, M. J. (1995) 'The omnipotence of voices II: The beliefs about voices questionnaire', British Journal of Psychiatry 165, 773-6.

Chadwick, P. D. J. and Lowe, C. F. (1990) 'Measurement and modification of delusional beliefs', Journal of Consulting and Clinical Psychology, 58: 225-32.

Chadwick, P. D. J. and Lowe, C. F. (1994) , A cognitive approach to measuring and modifying delusions', Behaviour, Research and Therapy 32: 355-67.

Cochrane, R. (1983) The Social Creation of Mental Illness, Essex: Longman.

Ellis, A. (1962) Reason and Emotion in Psychotherapy, New York: Lyle Stuart. Fishbein, M. and Ajzen, I. (1975) Belief, Attitude, Intention and Behaviour: An Introduction to Theory and Research, Massachusetts: Addison-Wesley.

Hawton, K., Salkovskis, P., Kirk, J. and Clark, D. M. (eds) (1988) Cognitive- behavioural Therapy for Psychiatric Problems, Oxford: Oxford University Press.

Hole, R. W., Rush, A. J. and Beck, A. T. (1979) 'A cognitive investigation of schizophrenic delusions', Psychiatry 42: 312-19.

Hollon, S. D., Shelton, R. C. and Loosen, P. T. (1991) 'Cognitive therapy and pharmacotherapy for depression', Journal of Consulting and Clinical Psychology 59: 88-99.

Maher, B. A. (1988) 'Anomalous experience and delusional thinking: The logic of explanation', in T. F. Oltmanns and B. A. Maher (eds) Delusional Beliefs, New York: Wiley.

Milgram, S. (1974) Obedience to Authority, New York: Harper and Row.

Oltmanns, T. F. and Maher, B. A. (eds) (1988) Delusional Beliefs, New York: Wiley.

Romme, M. A. R. and Escher, A.D.M.A.C. (1989) 'Hearing voices', Schizophrenia Bulletin 15: 209-16.

Slade, P. D. and Bentall, R. P. (1988) Sensory Deception: A Scientific Analysis of Hallucination, London: Croom Helm.

Strauss, J. S. (1969) 'Hallucinations and delusions as points on continua function', Archives of General Psychiatry 21: 581-6.

Strauss, J. S. (1989) 'Subjective experience of schizophrenia', Schizophrenia Bulletin 15: 179-85.

Strauss, J. S. (1991) 'The person with delusions', British Journal of Psychiatry 159: 57-62.

Tarrier, N. (1992) 'Management and modification of residual positive psychotic symptoms', in M. Birchwood and N. Tarrier (eds) Innovations in the Psychological Management of Schizophrenia, Chichester: Wiley.

Trower, P., Casey, A. and Dryden, W. (1988) Cognitive-behavioural Counselling in Action, Bristol: Sage.

World Health Organization (1973) International Pilot Study of Schizophrenia, Geneva: WHO.

Yalom, I. (1970) The Theory and Practice of Group Psychotherapy, New York: Basic Books.

Zigler, E. and Glick, M. (1988) 'Is paranoid schizophrenia really camouflaged depression?', American Psychologist 43: 284-90.



1. My voice is punishing me for something I have done.

2. My voice wants to help me.

3. My voice is persecuting me for no good reason.

4. My voice wants to protect me.

5. My voice is evil.

6. My voice is helping to keep me sane.

7. My voice wants to harm me.

8. My voice is helping me to develop my special powers or abilities.

9. My voice wants me to do bad things.

10. My voice is helping me to achieve my goal in life.

11. My voice is trying to corrupt or destroy me.

12. I am grateful for my voice.

13. My voice is very powerful.

14. My voice reassures me.

15. My voice frightens me.

16. My voice makes me happy.

17. My voice makes me feel down.

18. My voice makes me feel angry.

19. My voice makes me feel calm.

20. My voice makes me feel anxious.

21. My voice makes me feel confident.


22. I tell it to leave me alone.

23. I try and take my mind off it. 24. I try and stop it.

25. I do things to prevent it talking.

26. I am reluctant to obey it.

27. I listen to it because I want to.

28. I willingly follow what my voice tells me to do.

29. I have done things to start to get in contact with my voice.

30. I seek the advice of my voice.