Mirror, Voice and Reciprocal Role

Elia, I., 2004. Mirror, Voice and Reciprocal Role. Reformulation, Spring, pp.6-8.

Have any of you ever used the mirror to help clients recognize their self-to-self reciprocal roles (RRs)? Using it briefly in the session or at home, after written and diagrammatic reformulations have been presented, can supply a concrete experience of key RRs. I will describe how I became interested in the signals from and to the face, and then how the mirror might be used in CAT. Any comments, suggestions, or warnings would be most welcome from readers who have worked clinically or experimentally with mirrors and/or facial expression.

My curiosity about facial expression began when I looked at the habitual ‘faces’ that clients brought to therapy. They were so different from each other but so reliably the same within each person. That is to say, the habitual set of the features was the same or nearly the same at the start and sometimes throughout the session from week to week. A 30-year-old beauty culture trainee always arranged her carefully madeup face in what she thought was an attractive immobility and simply waited. A 25-year-old medical student came in with a beautiful smile and then began to cry within five to ten minutes, for all but two of the 16 sessions. A 35-year-old father, husband, and teacher had the thin-lipped severity of an embittered 70-year-old. I wondered if they experienced their faces as I did: respectively, hostile and resentful; lovable and hopelessly needy; judgmental and angry. How exactly did they look at themselves, that is, what face did they make in the mirror? How did they look to themselves, that is, what face did they see?

In some way, what was usually ‘on the face’ seemed to be connected to what was usually ‘on the mind’, but I did not know how. As therapy progressed, the events that had shaped the facial expression and the inner voice were revealed—respectively, a competition with a sister plus critical, dismissive experiences at home and school; a brutal father whom the client resembled facially and a pretty but cold mother from whom love had to be “squeezed” by an extremity of need; a ‘perfect’, religious upbringing, the ‘holier than thou’ superiority of which had been dented by tragic loss. But did they play the same roles toward themselves? And if so, did they notice the script they were muttering, keening, or shouting inside their own heads?

The other kindling for my interest in faces came from something I read, while doing research for a physical anthropology project on the cultural beauty requirements that permanently alter the bodies of girls and women, often without their consent or knowledge of the consequences. In one of my sources, Raymond Bull and Nichola Rumsey’s ‘The Social Psychology of Facial Appearance’1, a study2 was reviewed in which researchers found a significant relationship between the physical abuse of children (2-7 years) and the length of the child’s face (i.e., they had a low ratio of forehead to face height for their age). The researchers wondered if these children, who looked older than they were because of their low cranial-facial ratio, were being expected to act older and were abused if they did not. Bull and Rumsey, commenting on the study, said that for this to be true “one would need to presume that the abusive “caregivers” (who would normally be the child’s parents) were unaware of the child’s true age.” And the reason this is so pertinent to a discussion of facial expression and role learning is that in spite of it seeming to be a very plausible statement, later research3 and personal experience made me realize that it was probably incorrect. The mind may know the age of the child, but the face is a more potent signaler than is knowledge. So too, the client may know s/he wants to be happy, but perhaps the potent signaler of her/his own face, along with the problematic inner voice, make her/him behave in undermining or abusive ways towards the self.

The cross-cultural research of Jones3 and others4, which has gone some way toward establishing the physical parameters common to faces considered attractive (by high consensus), shows that perception of beauty is probably experienced deeply, quickly, and automatically. Physical features of age, gender, attractiveness, and mood seem to be long-evolved signalers, which evoke reflex responses of rapprochement or avoidance. Naturally, this is a generalisation, and conscious choice and training can alter these responses. But alteration of these fast responses to the face may not be as easy as our cultural belief in knowledge and free will makes us expect.

The physical features that evoke deep responses do not have to be consistent with underlying fact. They only need to be convincing. When, as an adult, I saw an old photo of me standing in my crib (‘cot’) and asked my mother why she had kept me in a crib when I was five, she replied that I had only just turned two. My instant assessment of my age had been based on my appearance, not on thinking. Back when I was two, my mother’s knowledge of my true age also seems to have been overridden by the signals of my height, long hair, and face. According to eyewitness report, she would smack and shout, if I sang or vocalized when adults were asleep in the morning. I realize that this does not prove that she saw me as five rather than two and therefore expected more considerate behaviour, but similar experiences with my daughter made me suspect it was true. When she was three, a paediatrician friend said that she looked five. I needed to remind myself and others how old she really was in order to halt or correct mistaken expectations and scoldings, e.g., if she could not tie her shoelaces or zip her jacket.

So physical features, which influence how we feel and behave, need not be consistent with the facts, only convincing. That is why makeup and costumes work in theater and film. A role is a powerful thing, and it is not complete without an appearance, a face. It is that face and overall appearance that makes us feel strong emotions, while in the audience. And in some sense, we are all in the audience to our own faces in the mirror and to the role-bound, habitual faces of others.

I thought the mirror might help clients and me discover how our faces were created in early, dialogic interactions. It did not seem fair or safe to ask them to look at themselves in a mirror for me, until I knew more about it. So I looked in the mirror myself. I wondered if I would find any of the RRs outlined in the letters of reformulation that I had received in my therapies. It was not difficult to see in my reflection the dull, ‘tin’ expression of ‘controlled’. This view was corroborated by old class photos, which showed my tepid, sad, but dutiful smile, so much less alive than other smiles in the group. Sometimes there was another aspect to the reflection, full of cold hate and threat. I had known I could turn on a ‘deadly’ face, when needed in the NY subway, but not that it was how I might look at myself.

Why hadn’t this other face been spotted and named in therapy? And what good would it do to recognise a face consistent with a particular role?

I asked others, who saw me all the time, to describe my habitual expression: alert, intent, preoccupied, tired, cross. So I looked in the mirror again and noticed behaviour, as well as expression: not unlike my mother, I honed in on flaws, worried about them, tried to correct, made them worse, became obsessed, got cross, and worried some more. All of this simmered away as my usual mode of operation. I wasn’t aware that I was playing any sort of role toward myself. There was the feeling that it was stable, unchangeable; this was just me.

Clients have said the same thing to me about their roles and procedures: ‘It’s just me.’ Perhaps the good that might come of using the mirror would be that adopting a new expression could reduce the background tension and permit a new role to be experienced.

Had I not been telling a therapist friend, who was learning some CAT, about the idea of ‘voice’ in RRs, and had she not then given me a little book, entitled ‘What to say when you Talk to Yourself’5, I would probably never have thought of changing my inner voice along with my facial expression. The face I altered by adding a small, tentative smile, which did make me feel different. I had recently seen ‘Vanilla Sky’ in which Tom Cruise playfully winked and smiled at his mirror image. While I suspected he had a lot more to smile about than I did, I forged on. It did seem artificial at first, but when I added a couple of novel words, such as ‘Well done!’, something really shifted. It seemed unbelievable that I could feel better in myself because I had looked at and spoken appreciatively to myself. It was a choice.

It was odd and often hard to remember to do it, but gradually I added new verbal and visual pieces to the repertoire, on the theme of self-liking, appreciation, and care. Looking at and wondering about the face I usually showed myself helped me hear, for the first time, the inner articulations that were just part of my everyday life, virtually every minute: ‘Got to… then, got to…. and I’ve got to…’ Also: ‘Should have…ought to…’ These were connected to feelings of pressure, worry, and hopelessness. For the first time, I knew how the feelings got there, and I experienced being able to change them, at least a bit, and sometimes dramatically.

Hearing the inner voice, which was ‘mine’ but of which I was usually unaware, also helped me understand how a client might suddenly assert that they had heard me say something about them, during session, that had never even crossed my mind no less my lips: ‘You are an idiot.’ ‘It’s all your fault.’ ‘You are not caring.’ If I were right about reflections, perhaps such transference ‘hearing’ was provoked by my reminding them of a parent’s expression or stance. Perhaps what they thought they heard was also sourced by the covert parental articulations that they ‘heard’ (felt), each time they looked in the mirror and saw an aspect of the critical or resentful parent. The statements that they were convinced I had said were plainly related to statements they had reported their parents making to them.

It was easiest to see how the reflection may have been helping to produce self-critical inner statements in the case of the student who facially resembled her father. I never asked her to look in the mirror in my presence. She had often told me of her dissatisfaction with her appearance, her belief that she looked like a man (which she most certainly did not), and her feeling of being unlovable. As this was around the time when I had tried saying ‘Well done!’ to myself, I suggested she try the same, both mentally and in front of the mirror at home, with a smile, if possible. This helped empower her, on at least one occasion, to decline the invitation of a man who had previously forced unprotected sex. It was in the session after this experience that she managed not to cry at all. The mirror-speech technique was strong but very difficult for her to continue to use, lifelong negative scripts and the parental mirror image being so powerful.

Then I asked her what her father’s favourite colours were, and she said she did not know, didn’t know if he had any, well, maybe blue and black because he mostly wore that. When I pointed out that she wore black and blue to all but one session, her curiosity was sparked. Could it be that we not only feel about and speak to ourselves as our parents did, but that we unconsciously wear the costume and the facial expression to go with the role they played towards us and that we now play toward ourselves?

I wonder if mirror work plus positive self-statements could be part of the ‘compassionate retraining’, which Paul Gilbert discussed in ‘Working with Shame’6. He mentions a study in which self-critical students could easily generate critical mental images, but they found it difficult to ‘see’ vivid, compassionate self images. So relearning internal images and associated feelings might be made easier with conscious alteration of the mirror image.
I have now asked four clients if it would be all right if I had a look at how they looked at themselves in the mirror. Each agreed to hold and look in an old-fashioned hand mirror, for a few seconds. Afterwards, if they wished, they said what they saw and I said what I saw. The first thing that I noticed about all of them was that not one smiled at their reflection, even if they perpetually smiled at me or began every session with a smile.

They would frown for an instant, look blank, bemused, or uncomfortable and pass the mirror back to me. If it could be noticed and admitted that their reflection appeared somewhat angry, blank, or worried, etc it was possible to think about what the response to such a face would be. To the angry: guilt, not good enough; to the blank: no feeling, distance; to the worried: tense, striving, hopeless. Each one seemed to be the child’s response to the parent’s look. But the client who frowned and admitted that he often felt dissatisfied/angry with his appearance showed me that it could work both ways. With respect to self-to-other roles, he said that he could be contemptuous and placating at the same time. Maybe it is the same with enactment of the RRs within or toward the self; the image in the mirror flashes between inner parent and child very quickly and subtly.

The same client commented about connecting positive inner speech with mirroring: “It was poignant. The feeling of being liked, approved of. It was sad to me that I knew I had not had it as I grew up.”

The student who always cried said, “I had to avert my eyes. I could look at my face to put on makeup, but I had never looked in my eyes like that. I was afraid to look up. It seemed wrong to smile at me. I had done so many things wrong, made so many mistakes. It felt as if something bad would happen. It was difficult to smile at me.” And when had something bad happened in the past? When she was with her father, looking at him! And who else had made so many mistakes and done so many wrong things? Her father, whom, I suspected she saw in her own reflection.

Roz Caroll’s ‘Love in a scientific climate’7 alludes to studies confirming “the depth of the impact of facial expression… often overriding the experience of verbal communications”. It does seem that early facial mirroring, which is a spontaneous, back-and-forth copying of facial expressions between mother (or other) and baby, is the beginning of a deep imprinting. In addition, metaphorical ‘mirroring’, which occurs through attunement to the young child’s actions and feelings through parallel modalities (e.g., mother copying a child’s movement with coordinated vocalizations), builds upon the visual copying.

Perhaps the durability of this system, for individuals and species, shows that copying equips us with our parents’ way of surviving. If, later on, we find that what was so deeply imprinted is not as survival-adaptive as we might wish, it is extremely difficult to change. If it is true that what we feel and do comes to be linked with the signal of a particular face, then it would make sense to try to change feelings and acts by altering the facial expression.

A therapist may have to notice the habitual face of the client, and be the one for whom its expression does not override reason. The tight-lipped, judgmental, cold face, for example, perhaps ought not to provoke in the therapist its usual reciprocal of over-striving and placating or of cold, competitive condescension. Being aware of what a client’s face communicates about RRs enables the therapist to try for something different, with her own face, body language, and verbal feedback. This can be part of the revision process, whether or not a mirror is used.

Recognising and revising their habitual expression and inner voice allows clients to befriend the self. This new friendship or compassion can affect self-to-other interactions too. One client, who had been reworking his self-to-self procedures, noticed that he was overly challenging toward his colleagues and held them in contempt, feeling as he did when he had first looked thoughtfully at his own image. He re, respectively alised that the anxiety generated from the ‘image and voice’ made him provocative and angry at work. When he could see the face that he usually carried to work, he drafted some positive, self-respecting statements and gradually released himself and others from ‘overly critical—criticised and shamed’.

The shift from self-to-self work to self-other behaviour happened to me too. After about three months of noticing and trying to intervene with my goading inner voice, which had always made me walk very fast, I was charging home from the bus station, one beautiful spring evening, when the possibility of walking more slowly sprang into my mind, as something utterly novel. With the same tone of reassurance that I had been rehearsing over the previous weeks, the words, ‘Don’t worry, you’ll still get everything done.’ were just there. I felt relieved and entitled to enjoy a leisurely walk. Later, the ‘Don’t worry, you’ll get done what needs to be done.’ phrase carried over to my work. I realised that sometimes I had been making clients go too fast, just as the child who looked older than she was had to do more and go faster. So I settled down calmly and confidently to wait and repeat and reassure in therapy. It helped so much.

Some of the questions that have come to mind about mirror work are: ‘How much is enough? Where will it all end? Will I be talking to myself all the time, winking and smirking at myself in the mirror for hours?’ While discussing this with one insomniac client, she remembered telling me about holding her nephew in her arms until he quieted down and fell asleep. Maybe enough work is just the amount needed to quiet us down sufficiently to be able to sleep, eat, work, and relate without ‘pain’. I would be grateful to know if anyone else is using the mirror and rescripting the inner voice to assist clients in recognising and revising their roles and procedures.

Irene Elia


Bull, R. and Rumsey, N. (1988) The Social Psychology of Facial Appearance. London: Springer Verlag, p. 157.

McCabe, V. (1984) Abstract perceptual information for age level: a risk factor for maltreatment? Child Development. 55, 267-276.

Jones, D. (1995) Sexual selection, physical attractiveness, and facial neoteny. Current Anthropology. 36, 5, 723-748.

Jones, D. and Hill K. (1993) Criteria of facial attractiveness in five populations. Human Nature. 4, 3, 271-296.

Helmstetter, S. (1991) What to Say when you Talk to Yourself. London: Thorson’s.

Gilbert, P (2003) Working with Shame. Reformulation. Summer Issue, 13-15.

Carroll, R. (2003) Love in a scientific climate. The Psychotherapist. Spring/Summer Issue, 11-12.

Petition to NHS England - The Case for Funding Training in the NHS 2021 Alert!
ACAT's online payment system has been updated - click for more information

Full Reference

Elia, I., 2004. Mirror, Voice and Reciprocal Role. Reformulation, Spring, pp.6-8.

Search the Bibliography

Type in your search terms. If you want to search for results that match ALL of your keywords you can list them with commas between them; e.g., "borderline,adolescent", which will bring back results that have BOTH keywords mentioned in the title or author data.

Related Articles

Mirror Workshop Summary Held On The 9th July 2010
Elia, I., 2010. Mirror Workshop Summary Held On The 9th July 2010. Reformulation, Winter, p.47.

Workshop Presentation Synopses from the 2005 ACAT Annual Conference
Carroll, R., Elia, I., Compton Dickinson, S. and Webster, M., 2005. Workshop Presentation Synopses from the 2005 ACAT Annual Conference. Reformulation, Spring, pp.7-9.

Three linked things: Bodies
Maggie Gray, 2013. Three linked things: Bodies. Reformulation, Summer, p.40.

Working With Shame
Gilbert, P., 2003. Working With Shame. Reformulation, Summer, pp.13-15.

Comment on James Turner’s article on Verbal and Pictorial Metaphor in CAT
Hughes, R., 2011. Comment on James Turner’s article on Verbal and Pictorial Metaphor in CAT. Reformulation, Winter, pp.24-25.

Other Articles in the Same Issue

A Cognitive Analytic Multicomponent Psychotherapy Program, for the Treatment of Severe Personality Disorders in an Intensive Outpatient Unit.
Mirapeix, C., Landin, S. and Alvarez, V., 2004. A Cognitive Analytic Multicomponent Psychotherapy Program, for the Treatment of Severe Personality Disorders in an Intensive Outpatient Unit.. Reformulation, Spring, pp.10-13.

Book Review: Managing Intense Emotions and Overcoming Self Destructive Habits: Lorraine Bell
Hobson, J., 2004. Book Review: Managing Intense Emotions and Overcoming Self Destructive Habits: Lorraine Bell. Reformulation, Spring, p.32.

Continuing Professional Development
Buckley, M., 2004. Continuing Professional Development. Reformulation, Spring, p.2.

Mirror, Voice and Reciprocal Role
Elia, I., 2004. Mirror, Voice and Reciprocal Role. Reformulation, Spring, pp.6-8.

Naming or Shaming?
Nuttall, S. and Scott Stewart, M., 2004. Naming or Shaming?. Reformulation, Spring, p.3.

The Big Issue - Report from UKCP
Knight, M., 2004. The Big Issue - Report from UKCP. Reformulation, Spring, p.31.

The Use of the Dyad Grid in CAT - An Example
Bristow, J., 2004. The Use of the Dyad Grid in CAT - An Example. Reformulation, Spring, pp.22-30.

Untying the knots: relational states of mind in Cognitive Analytic Therapy?
Potter, S., 2004. Untying the knots: relational states of mind in Cognitive Analytic Therapy?. Reformulation, Spring, pp.14-21.


This site has recently been updated to be Mobile Friendly. We are working through the pages to check everything is working properly. If you spot a problem please email support@acat.me.uk and we'll look into it. Thank you.