Hearing multiple inner voices was once treated as a sign of psychosis – but could we learn deeper truths by listening to what they say? In an edited extract from his new book, psychologist Charles Fernyhough explores a radical new approach to mental illness.

“I don’t know. It sounds daft, but sometimes he says things that are really funny.”

The speaker is Margaret, a woman in her seventies with a bright, open face and a gentle smile. She has come along today with her daughter, who is desperate for advice on helping her mother to cope with her frequent voices. There are around 20 of us in the room. Apart from a couple of clinical psychologists and a few of us from the academic team, everyone gathered here in a conference suite at Durham University on a cold, sparkling spring afternoon, is a voice-hearer. We are hosting the event with our special guest Jacqui Dillon, chair of the UK Hearing Voices Network and an old friend of our project.

“Yeah, mine say funny things, too.”

Before Margaret came into this room, she had never met another voice-hearer. Now she is surrounded by them. I see her deep in conversation with Julia, a writer of a similar age who has come to talk to us several times about her experiences. Two elderly ladies, sipping tea and chatting about the voices in their heads. Julia is an old hand, but Margaret is in entirely new territory. She looks radiant, ­ transformed. I have the sense that a life could be changing in front of my eyes.

Groups like this come together around a starting assumption that voices are meaningful, and that they convey valuable emotional messages. The idea that voices can have profound human significance has deep roots, featuring, for example, in the psychoanalyst Carl Jung’s argument that hallucinations contain a “germ of meaning” which, if identified accurately, can mark the beginning of a process of healing. The idea is antithetical to the traditional biomedical view from psychiatry, which has tended to see voices as neural junk, meaningless glitches in the brain – yet many people are now finding solace in this approach.

The first voice that Longden heard was benign, commenting on her actions in the third person: ‘She is leaving the building.’ ‘She is opening the door.’

In Eleanor Longden’s gripping and heart-rending book, Learning from the Voices in My Head, based on her popular TED talk, she describes how her psychotic breakdown as a student led to a diagnosis of schizophrenia, from which she was told there would be no recovery. The first voice that Longden heard was benign, commenting on her actions in the third person: “She is leaving the building.” “She is opening the door.” When Longden mentioned her voice to a friend, the beginnings of a positive relationship with the voice began to unravel. Longden was urged to seek medical help, and her college doctor referred her to a psychiatrist – the beginning of a journey from a straight-A student to cowering, degraded psychiatric patient. One consultant psychiatrist told Longden that she would have been better off with cancer than schizophrenia, “because cancer is easier to cure”.

Like countless others, Longden has benefited from the alternative framework of understanding that the Hearing Voices Movement has given her. After her grim days in hospital, she was seen by an enlightened psychiatrist, Pat Bracken, who helped her to understand her voices not as symptoms of disease but as survival strategies. She had been brutalised by her experiences, and her psyche was struggling to adapt. Longden began to understand her voices as resulting from the horrific and organised sexual torture she had suffered as a small child. “It was a blasphemy,” she wrote of her abuse, “a desecration beyond expression; and it left behind a tiny child whose mind broke and shattered into a million tiny pieces.” Longden grew into an academically successful, organised young woman who became adept at presenting a serene face to the world. Behind the façade, though, was a mind torn apart by a “psychic civil war”.

With Bracken’s help, Longden resolved to make sense of her voices. She learned of the Dutch psychiatrist Marius Romme’s claim that voices are messengers communicating important information about unresolved emotional problems. In Romme’s metaphor, it makes no sense to shoot the messenger just because the content of the message is unpleasant. Instead, the Hearing Voices Movement approach is to encourage voice-hearers to try to understand the events that led to the emotional distress that the voices are expressing. “The question we should be asking,” Longden says, “is not ‘What’s your problem?’ It’s ‘What’s your story?’”

Hage’s epiphany in understanding her alien voices was a simple realisation: “I’m not a schizophrenic, I’m an ancient Greek!”

The movement has its origins in a particular therapeutic partnership between Romme and one of his patients, a young Dutch woman named Patsy Hage. As a conventionally trained physician, Romme’s instinct was to see Hage’s destructive, troubling voices as meaningless symptoms of a biomedical illness. But Hage insisted that her voices were as real and significant as the deities to whom those around her prayed. She was influenced by Julian Jaynes’s theory of how our ancestors in the era of the Iliad experienced the gods speaking to them, and she was reassured by the closeness of her experience to what Jaynes claimed was once the default mode of thought. As she put it to Romme, her epiphany in understanding her alien voices was a simple realisation: “I’m not a schizophrenic, I’m an ancient Greek!”

Romme’s view of his patient started to change, and he began to take her testimony more seriously. When they appeared together on Dutch television to discuss the work they had been doing and to appeal for voice-hearers to contact them, there was an overwhelming response. Around 150 of the people who contacted them had found ways of living contentedly with their voices. As the movement grew, its basic tenets began to cohere: that hearing voices is a common aspect of human experience, which can be distressing but is not inherently a symptom of illness; and that voices bear messages about emotional truths and problems which can, through taking the voices seriously and with appropriate support, be resolved.

At first glance, this looks utterly different to the idea that voices result from atypical processing of inner speech. Rather than pointing us to the processing of language and the speech perception networks in the brain, it suggests that we should look for a link with traumatic memories. Clinically, and from the evidence of personal testimony, this makes perfect sense. What scientific support, though, is there for the idea that voices are about memories of the past?

One way of addressing this question is to ask whether those who hear voices show any differences in how they process memories. Flavie Waters and her colleagues at the University of Western Australia have proposed that auditory hallucinations result from a failure to inhibit memories that are not relevant to what the person is currently doing. The idea, which has gained support from Waters’ experimental studies, is that people who have such experiences are particularly bad at keeping irrelevant information out of consciousness. Coupled with a problem with context memory – which refers to the ability to recall the details of the context in which an event happened – this can lead to memories intruding into consciousness shorn of the contextual anchors that would usually allow us to recognise them as a memory rather than a hallucination. Another line of evidence comes from associations with trauma. There is now very strong evidence for a link between hearing voices and early adversity, particularly childhood sexual abuse. In one recent study led by Richard Bentall, childhood rape was strongly and specifically associated with hallucinations later in life. To give an indication of the strength of the relationship, Bentall likened it to that between smoking and lung cancer.

The memory account faces serious problems, though. For one thing, it has to explain how memories of horrible events can lie dormant for so many years before resurfacing in early adulthood – the peak time for the diagnosis of disorders like schizophrenia. Another problem is that memory simply doesn’t work like that. Creating a memory is a reconstructive process that involves bringing together lots of different sources of information, some of which are erroneously included, having not figured in the original event at all. We are particularly bad, even over a short period, at recalling the exact words that people say to us: we tend to recall the gist of messages rather than their verbatim information. That would be problematic if we wanted to propose that voices are literal reproductions of earlier conversations. In general, the idea that a traumatic memory can faithfully reproduce the details of the event and then be reactivated decades later doesn’t fit with what is known about how we remember.

The missing link between trauma and voice-hearing might be a psychological phenomenon known as dissociation. First described by the French psychiatrist Pierre Janet towards the end of the 19th Century, dissociation refers to the phenomenon whereby thoughts, feelings and experiences are not integrated into consciousness in the usual way. The connection with hearing voices comes through the finding that people who live through horrific events often describe themselves dissociating during the trauma. It is as though there is some drastic attempt by the psyche to remove itself from the horror that is unfolding: drastic because it effectively involves the psyche cleaving itself into pieces. This state of dissociation may then create a pathway for the hallucinations, by causing the mind to “disown” the fragments of voices emerging in their consciousness.

Dissociation perhaps just describes an extreme version of what is normal in the rest of us

Dissociation perhaps just describes an extreme version of what is normal in the rest of us. Eleanor Longden is now a postdoctoral researcher specialising in how the psychological processes of dissociation might help to explain the occurrence of trauma-related voices. “I think we all have multiple aspects of self,” Longden told me recently, “and that this is an experience most people will relate to: a part that’s very critical, a part that wants to appease everyone, a part that is playful and irresponsible, and so on. Voices feel more disowned and externalised in most cases, but essentially I think they represent a similar process.”

Voices, then, might give us important clues about the fragmentary constitution of an ordinary human self. This view of a dissociated self is already paying dividends in research on voice-hearing, although much remains to be done in explaining exactly what these fragments of self look like, and how they operate and behave. A dissociation account is still some way from explaining how traumatic memories become transmuted into hallucinations – and in particular why those experiences are so often verbal – but it is a promising avenue for future research. In particular, it might provide a partial answer to the puzzle of how a voice can be both “me” and “not me”.

Perhaps the most valuable thing about thinking about voices in this way is its implications for managing the experience. If voices are at least partly about things that happened to you, then they give you something you can work with. They hold out the prospect of recovery.

This is an idea that the Hearing Voices Movement has used to powerful effect. When Eleanor Longden offered forgiveness to her most unpleasant voice, the tone of its communications with her changed. “Essentially,” she told me, “it was a process of making peace with myself, because the negative voices embodied so many painful memories and unresolved emotion. A turning point was coming to the realisation that the voices weren’t the actual abusers; rather, that they represented my feelings and beliefs about the abuse. So while they appeared incredibly negative and malicious, they actually embodied the aspects of me that had been hurt the most profoundly, and as such needed the greatest amounts of compassion and care.”

Although voices may not be memories in the literal sense, we can learn much about methods for dealing with them by looking at how people learn to live with memories for horrible events. Therapy for traumatic memories, such as those that characterise post-traumatic stress disorder, involves encouraging the sufferer not so much to try to forget the event as to remember it more accurately. That means fitting it into a network of memories in such a way that it can become less intrusive, distorted and autonomous.

As far as voices are concerned, that means recognising that one’s voices are strange, disruptive, twisted facets of the self – but they are still parts of the self. If they’re in some sense about memory, you can come to terms with them in the same way as you can come to terms with unpleasant memories: by integrating them back into the psyche from which they have become split away.

One way of easing this process of integration is to give the voice some concrete external reality. In a technique favoured by many in the Hearing Voices Movement, the voice is brought directly into the therapy session. Known as Voice Dialogue or Talking with Voices, the method involves the facilitator asking the voice-hearer whether she can speak directly to the voice. In one of Eleanor Longden’s case studies, the therapist (in this case the Dutch psychotherapist Dirk Corstens) asks to talk to one of the participant Nelson’s dominant voices, who goes by the name of Judas.

After establishing contact with Judas, Nelson adopts a military stance and begins pacing around the room (he had previously served in the army), while speaking in Judas’s voice, with its characteristic abrupt sentences. At the end of the session Nelson comes back into himself and reports that he was aware of everything that went on. The therapy had positive and lasting effects, with Nelson finding a new rapport with his mental visitors and, crucially, effecting a rapprochement between his distinct voices – and thus parts of his self – which had previously been at war.

Another technique that involves externalising the voice bears several parallels with Voice Dialogue, although it emerged from a rather different quarter. The English psychiatrist Julian Leff recently had the idea of setting up a therapeutic situation in which patients who heard voices could interact with them through a computer-based avatar. Using face-generation software, patients are first asked to create a face that matches the voice they want to work with, and to use voice-synthesis software to construct a voice to go with it. The therapist, sitting in another room, then speaks to the patient in the voice of the hallucination, with the face of the avatar animated in sync, although at the flick of a switch he can flip back to being the comforting, guiding therapist. Typically this will involve encouraging the patient to stand up to the voice and challenge what it is saying. A pilot study showed impressive results, with reductions in the frequency of voices and also beliefs in their malevolence and omnipotence. Leff believes that the visualisation of the hallucinatory experience allows patients to gain control over their voices, particularly when they are frightened about how the voice might react if it was confronted or challenged. 

Jacqui Dillon points out that there is nothing particularly new in this approach. When I meet her for lunch on a sunny November day at the Barbican Centre in London, she tells me that the Avatar project is thoroughly well meaning but appears to lack theoretical foundations. “It does seem to be making the whole thing much more convoluted and technologically heavy than it needs to be,” she explains. “Because essentially what seemed to be working was that you’re taking these things as being real and meaningful to me, and you’re kind of having a conversation with me about them. The fact that you’ve got to go in another room and do that, through an avatar that we have to create, seems to me a bit of a palaver, really. Why can’t we just sit in the same room and have that conversation?”

I’ve known Dillon for several years. She is a highly influential figure in the mental health world, travelling frequently to give talks about her own life and experiences and her role in the growing international movement. As our project has developed and she has seen that we are genuinely interested in understanding voice-hearing, we have become friends, although there are still a few things that we disagree about. Like many others in the Hearing Voices Movement, Dillon is sceptical about the inner speech model of voice-hearing – the view that hearing voices happens when you don’t recognise your own ordinary inner speech – because it seems not to do enough justice to the meaning of the experience. Catching up with her today is partly about persuading her that my being interested in how voices work – wanting to understand their mechanisms in the mind and brain – doesn’t entail trying to explain them away. Or even wanting to make them go away, in the sense of stopping them altogether.

The voices in our heads contain a diverse variety of forms, dialogue, representations of remembered events, interactions with visual imagery and other sensory experiences

I suspect that some of the antipathy to the inner speech model stems from overlooking the complexity of the phenomenon. “How can it just be inner speech?” people ask me. I respond that there is nothing “just” about inner speech. The voices in our heads contain multitudes: not just a considerable variety of forms, but also diverse voices in dialogue, representations of remembered events, interactions with visual imagery and other sensory experiences, and so on. Saying that inner speech cannot explain why voices are meaningful only makes sense if you hold that inner speech is not meaningful. As the research shows, nothing could be further from the truth.

Dillon hears more than a hundred different voices, and she is far better placed than I am to say whether the scientific theories bear any relation to the felt experience. She has described the hearing of a voice as being like receiving “a telephone call from your unconscious”. It’s a message that, however horrible and disruptive, you feel compelled to listen to. “It’s an aspect of self, isn’t it?” Dillon tells me. “Even if it’s an unpleasant aspect. I find this with a lot of people who hear voices, when you really ask them, they don’t really want them to go away. They don’t want them to be as distressing, but there’s this sense that they’re part of me in some way, even if they’re saying these really troubling, upsetting things.”

Exactly the same question comes up when we ask about how people relate to traumatic memories. Even when memories are terrible, we seem to want to hang on to them. They are parts of our selves, even if they are aspects that we can only look upon with horror. But therapy can help people cope with them in a more constructive way, reprocessing the events so that they become less intrusive, disruptive and distorted. When I was researching my book on the topic, I talked to a lorry driver, Colin, who had had his horrific and distorted memories of a road traffic accident reshaped by a method called EMDR. During the worst phase of his post-traumatic stress disorder, he would have given anything to make the memories go away, but now, after therapy, he accepts them as part of himself. I would say the same thing about a couple of highly distressing memories that I myself have. They’re horrible, but they are mine. Isn’t this precisely what some people say about their voices?

Dillon agrees that the experience of distressing voices is not something that can be best dealt with by banishing them. If we can reshape our memories into a more palatable form, perhaps voice hearers can also mould their own pieces of remembered experience into a story that is not quite as shattering to their sense of who they are. This chimes with Dillon’s intuitions as a voice-hearer. “I suppose I would see those as aspects of a person’s self, and the analogy I would often use is family therapy… it’s a dysfunctional family. I would see them as fragments of a self, and those aspects make up a person.”

And we are all fragmented. There is no unitary self. We are all in pieces, struggling to create the illusion of a coherent ‘me’ from moment to moment. We are all more or less dissociated. Our selves are constantly constructed and reconstructed in ways that often work well, but often break down. Stuff happens, and the centre cannot hold. Some of us have more fragmentation going on, because of those things that have happened; those people face a tougher challenge of pulling it all together. But no one ever slots in the last piece and makes it whole. As human beings we seem to want that illusion of a completed, unitary self, but getting there is hard work. And anyway, we never get there.

In the meantime, people like Dillon are left with the voices. Joking with her at the cafe at the Barbican, I can tell she’s happy with the company. “Most of the time they make me laugh,” she said later in an interview. “They’re insightful, they’re loving, they’re comforting, and they help me to feel less alone. They know me better than anybody else and they’re always there for me.”


Charles Fernyhough is an author and psychologist at Durham University in the UK. This is an edited extract from his new book The Voices Within. Photography by Olivia Howitt (@oliviahowitt); the images are purely conceptual, and are not portraits of any of the people interviewed for the article.

Join 600,000+ Future fans by liking us on Facebook, or follow us on Twitter, Google+, LinkedIn and Instagram.

If you liked this story, sign up for the weekly bbc.com features newsletter, called “If You Only Read 6 Things This Week”. A handpicked selection of stories from BBC Future, Earth, Culture, Capital, Travel and Autos, delivered to your inbox every Friday.