The Selfie, the Filter, and the Two Bodies

A young woman retouches her ID photo using artificial intelligence. The educator accompanying her tells her: “That’s not you. You’re already very beautiful, but that’s not you.” The young woman herself does not recognise herself in the photo-booth picture; she recognises herself in the retouched image. To my mind, each of us now has two bodies: an embodied body and an image-body, which do not coincide and which perhaps do not need to coincide.

The primal scene of the filter

The scene is ordinary. It plays out every day in educational services, therapeutic consultations, classrooms. A person, often young, often a woman, produces an image of herself that does not look like what the professionals around her see. The image is smoothed, embellished, transformed by filters or by an artificial intelligence application. The nose is finer, the skin smoother, the proportions altered. Faced with this image, adults have an almost unanimous reflex: they grow concerned and name a symptom. Dysmorphia. Narcissism. Loss of contact with reality.

Before issuing a diagnosis, it is worth asking what this person is actually doing. Is she fleeing reality, or is she building something?

The image is never the mirror

The idea that there exists a “true” image of self, of which filters would be a distortion, does not stand up to examination.

The photo-booth picture, which we take as our standard of truth, is not objective. It is an image produced by a machine, in a particular light, at a particular angle, according to conventions inherited from nineteenth-century police anthropometry. The identity card was an invention of surveillance regimes before it was a quest for truth. Alphonse Bertillon, who systematised judicial photography at the end of the nineteenth century, was not seeking to capture the essence of persons: he was seeking to identify them and file them. When we say that a photo-booth picture is “the true image” of someone, we are speaking from within this police lineage, often without knowing it.

Smartphones themselves do not produce “raw” images. They embed automatic artificial intelligence processes, impossible to disable, that smooth skin, optimise lighting, sometimes wait for the smile before triggering the shutter. These are no longer just Snapchat filters: this is built into the machine’s default settings. The “unfiltered” image is already filtered.

As Jean-Luc Godard once put it: it is not a just image, it is just an image. The photo-booth picture, the filtered selfie, the pastel portrait are all constructions. The mirror itself is a construction, since it inverts left and right: the image we see in it is not the image others see of us. No image is “us.” An image is always a proposition, carrying the biases of the machine or the hand that produces it.

The illusion of objectivity, and yet it works

There is a paradox I want to name because it makes me uneasy as well. We believe in the objectivity of images, and that belief operates. In police investigations, photographs are taken of the crime scene, and from these photographs one works as if one were facing the real itself. In Michelangelo Antonioni’s film Blow Up (1966), the photographer character progressively enlarges one of his prints and ends up discovering a body hidden behind a bush: the photograph gives him access to something his direct vision had missed. The image, in that film, seems to touch the real more precisely than the real itself. And we experience this every day. A map, a plan, an X-ray let us act in the world, find our way, make decisions. If the image were pure illusion, it would not work.

And yet it remains a representation. The philosopher Alfred Korzybski formulated in 1933, in Science and Sanity, a sentence that has not aged: the map is not the territory. A map allows us to move through the real world, but it is not the real world. Every map is a projection, that is, a compromise. Flattening the terrestrial sphere requires choosing what to preserve: angles, surfaces, distances. One cannot preserve everything at once. The Mercator projection, which serves as the implicit norm for many Western maps, exaggerates the size of high-latitude land masses and minimises that of equatorial ones: Greenland appears as large as Africa, even though it is fourteen times smaller. The geographer Mark Monmonier devoted an entire book to this question (How to Lie with Maps, 1991), to show that no map can avoid lying, because every map simplifies. Yet we travel by them, we understand the world by them, and we mistake their distortions for self-evident truths.

When I run virtual reality workshops with teenagers, this paradox becomes tangible. We set down a 360° camera, press the button, and the machine records everything around it. One might think that this is the most objective recording possible, since nothing has been framed, nothing chosen. But when we look at the rushes, we see two fisheye images side by side, which look nothing like what our eyes saw during the shoot. The machine captured 360 degrees, true, but its way of seeing is not human. For the image to become legible to us, we need a headset, stitching software, a point of view. The absolute objectivity of the recording produces an image that nobody can look at directly.

The image is therefore not the real, and yet it grants us access to the real through a mediation we forget. This is what the map does: it filters our gaze while giving us the illusion of being in direct contact with the territory. The more precise it is, the more it makes us believe that it is the territory. Our relation to images of self follows the same logic. The filtered selfie, the profile picture, the pastel portrait are maps of the face, not the face. And it is because they are maps that they work in the social spaces in which we use them.

The mask and the face

What remains is to understand what this filtered image does for the person producing it.

Françoise Dolto, in L’Image inconsciente du corps (1984), showed that the body image is not the objective reflection of a biological body but a psychic construction. It is an unconscious representation that the subject forms of their own body, which takes shape in relation to the other, and notably in early exchanges with the mother. This image is relational, and therefore relative to lived experience, to the gazes received, to the words heard. A person can have a healthy body and an altered body image, and the reverse.

The digital filter therefore inscribes itself onto a construction already in motion. It can do so pathologically, when it proceeds from suffering or self-rejection. It can also do so creatively, as a space of experimentation comparable to the mask in psychodrama.

In the therapeutic tradition, the mask is a tool of truth. By wearing a mask, the subject can say things they could not say with a bare face. The mask both protects and frees. It creates a transitional space, in Winnicott’s sense, where the subject can play with identities without being trapped in them. The digital filter can function this way. The young woman who retouches her photo is trying on another face. She plays with the possibilities of her image, she explores a version of herself that does not exist in the mirror but exists in the digital social spaces in which a significant part of her life unfolds.

The two bodies

To think this situation through, I propose to speak of two bodies. Each of us, today, has at least two. An embodied body: the one others see when facing us, with a height, a weight, a face, a voice. And an image-body: the one that exists in the images we produce (profile pictures, selfies, stories, videos, avatars) and that circulate in our place in digital social spaces.

These two bodies do not coincide. They have, in fact, never coincided: before the digital age, the painted portrait did not look exactly like the model, and the photographed person did not always recognise themselves in the photograph. But the non-coincidence is now more explicit and more worked over than before. The filter is its tool. It allows the image-body to be shaped independently of the embodied body.

The historian Ernst Kantorowicz, in The King’s Two Bodies (1957), showed that in medieval political theology, the king possessed two bodies: a natural body, mortal and fallible, and a political body, immortal and sacred. The political body did not disappear with the death of the king: it was transmitted to his successor. The two bodies coexisted in the same person without merging.

What particularly interests me in this story is the role played by images. At royal funerals in France, from the fourteenth century onwards, an effigy of the king was placed on the coffin: a wax or wooden statue clad in the royal insignia. This effigy was not a mere representation. It accomplished something. It maintained the political body in presence during the period of transition between the death of the deceased king and the full recognition of his successor. As long as the effigy was there, the king was not entirely dead. The image kept the political body in existence even as the natural body lay in the coffin. This is a function of images we find again in death masks, in funerary portraits, and more broadly in all the practices where images serve to maintain the presence of an absent body.

This function is not a historical curiosity. It directly concerns us today. When we publish a profile picture on WhatsApp, we are not “representing” our body: we are bringing it into existence in a space where it is not physically present. When a teenager posts a story, their image-body takes its place in the feed of their friends, who interact with it in our absence. The image of self on social networks is not a copy of the embodied body: it is the instance through which we exist in spaces where our biological body cannot be. As with the medieval king whose effigy maintained political presence, our image-body maintains our social presence.

This analogy helps us understand why the filter is such a charged operation. In this logic, modifying one’s image is fashioning the body that will exist where we are not. The medieval effigy was often more perfect, more solemn, more accomplished than the mortal body it replaced, and it was precisely because it was so that it could fulfil its function. The image-body can be smoother, more calibrated, more polished than the embodied body, and that simply makes it another mode of existence.

What the quest for perfection tells us

There remains the question everyone is asking: this quest for the perfect image, this inability in some people to show themselves without retouching, isn’t it, after all, a problem?

It depends. And it is because it depends that a priori judgement is harmful.

For some people, the filter is a game, an experiment, a social costume one puts on and takes off, much as one wears makeup to go out for the evening and removes it on coming home. For others, the filter has become a prison: they can no longer bear to see themselves without it, the gap between embodied body and image-body has become a source of suffering, and they come to want to alter the embodied body to make it resemble the image-body. The development of cosmetic surgery known as “Snapchat dysmorphia,” documented by surgeons, bears witness to this reality.

Between play and prison, there is a whole spectrum. Mediation works in that space. The British psychoanalyst Adam Phillips, in On Flirtation (1994), proposes that flirtation, in the broad sense of an exploration of possibilities without commitment, is a fundamental psychic activity. We need to try out identities without becoming fixed in them, to play with versions of ourselves without having to actualise them. The filter, in its playful dimension, is a tool of this identity-flirtation.

The proper clinical question therefore concerns the quality of the play, not the use of the filter as such. Can the person try out an image and come back, or do they find themselves stuck, the image-body having become the only bearable body? The answer is found in dialogue with them, in listening to what they live, in attentiveness to the suppleness or rigidity of their relation to their images.

The voice and the mirror

A parallel can illuminate this point: the strangeness of one’s own recorded voice. When we hear our voice on a recording, we do not recognise ourselves. It is not us. And yet it is our voice, as others hear it.

The explanation is physiological. When we speak, our voice reaches us by two routes. The airborne route, first: the sound vibrations that travel outside, reach our ears and stimulate our eardrums, as for any sound. The bone-conducted route, second: the vibrations that propagate directly through the bones of our skull, particularly the mandible and the temporal bones, and stimulate our inner ear without passing through the air. This bone conduction modifies the sound spectrum that reaches us: it accentuates lower frequencies and dampens higher ones. Our “inner” voice is deeper and fuller than our “outer” voice. The recording, however, captures only the airborne path. When we listen to our recorded voice, we hear it for the first time as others perceive it, and the dissonance with what we usually believe ourselves to hear is striking.

When I run virtual reality or filmmaking workshops with teenagers, this listening moment recurs as a ritual. They record themselves, listen back, and they say it out loud: “But that’s not my voice.” We adults explain to them that it is exactly the same for us, that we have grown used to it through hearing our voice on telephone answering machines, but that the first time always produces this effect. Yet the disturbance persists for the teenagers. Something of their identity is called into question by this recording that confronts them with a self they do not recognise, and which is nevertheless the self that others know.

This example shifts the question outside the technical field of filters and images. No one says that the voice recording betrays the “true” voice. No one thinks we should ban telephone answering machines, or that singers who listen to themselves in the studio develop auditory dysmorphia. And yet the non-coincidence between voice heard from inside and voice heard from outside is, in its structure, of the same order as the non-coincidence between embodied body and image-body. We have always had two voices; we have only made a drama of it for the face.

In the same way, our image in the mirror does not match the image others see of us, since the mirror inverts left and right. And the photo does not match the mirror, since the photo is flat and frozen whereas the mirror is alive. We constantly navigate between representations of ourselves that do not coincide, and that is normal. The filter is one more element in this constitutive non-coincidence. This explicitation, to my mind, is an opportunity for therapeutic work rather than a problem to be solved.

The felt body, the seen body, the imagined body

Moshe Feldenkrais, in Awareness Through Movement (1972), distinguishes what he calls the self-image from the objective body schema. The self-image, in Feldenkrais, is the representation each of us forms of our body in movement, in space and in time. This representation is strongly biased. Most people, when asked to mime a gesture they perform every day, mime it very differently from what they actually do. We believe we are doing with our body what we imagine ourselves doing with it, and this imagination can be very far from the observable real.

Feldenkrais’s somatic work consists in widening the self-image by exploring movement with attention. The more we move with awareness, the more the self-image becomes precise and the more movement gains in accuracy. There is always a gap between the body as it is and the body as one represents it to oneself, and this gap is itself a working material. It is not a defect to be corrected, it is a constitutive datum of having a body at all.

There thus exist not two but at least three bodies. The biological body, observable from outside. The image-body, which circulates in images. And the felt body, which is the internal representation each of us has of our own body. These three bodies never coincide perfectly. When a young woman retouches herself in a photo, she is not only fashioning an image-body for digital spaces: she is perhaps also expressing something of her felt body, that is, the way she experiences herself. There is no a priori reason for the felt body to coincide with the biological body as an outside photographer would capture it.

Felt identity

The psychotherapist Eugene Gendlin, in Focusing (1978), spoke of felt sense to designate the way we experience our identity bodily, beneath the discourse we hold about ourselves. There is in each of us a sense of self that is irreducible to what we are biologically, nor to what others see of us, and which asks to be heard before being translated into words. The digital filter, in this respect, can be a form of visual translation of this sense. It is not the sense itself, but it is one possible expression of it in the language of images.

Disability Studies, a field of research that developed in the English-speaking world from the 1980s onwards, brought another illumination. Lennard Davis, in Enforcing Normalcy (1995), showed that the very notion of a “normal” body is a historical construction that emerges in the nineteenth century with social statistics. Before that period, the body was thought in terms of an ideal (the perfect, divine, unattainable body), not in terms of a norm. Statistics invented an “average” body that became a “normal” body, that is, a body everyone ought to have. This norm, Davis writes, by construction excludes the majority of real bodies. Rosemarie Garland-Thomson, in Extraordinary Bodies (1997), extended this analysis by showing that bodies described as “extraordinary” (disabled, ill, fat, old, racialised, gender-nonconforming bodies) are not deviations from a normal body that exists in itself: they are produced as deviations by a cultural device that posits a certain body as the standard.

A person who “transforms” with a filter perhaps does not deform themselves: perhaps they draw closer to what they feel themselves to be, by resisting a norm that does not represent them. As one participant put it to me: “there are two images, in fact.” Not a true and a false: two dimensions of an identity.

Sport, which seems remote from questions of mediation through images, brings yet another illumination. In performance psychology, the imagined body plays a central role. High-level athletes visualise their gesture before executing it, and this visualisation does in fact modify performance. Work on mental imagery, studied notably by Allan Paivio from the 1970s onwards and then developed in sport sciences, shows that the imagined body is not an inactive fantasy: it guides motor learning, prepares execution, sustains confidence. The athlete’s imagined body is not their biological body, but it acts on it. The same logic applies, on a smaller scale, to everyday gestures and to representations of self. When we imagine ourselves in a certain way, this image alters our relation to our biological body. This holds for performance, and it holds also for self-confidence, posture, the way one occupies space.

The image of self as working material

If we cease to consider selfies and filters as symptoms to be corrected, what can we make of them in therapeutic mediation? To my mind, they can become rich working material, because they touch on the question of identity itself.

A professional in the social sector was telling me about her team’s idea to launch a research project on avatars, to see what gets projected into an image one can shape. “We were wondering,” she said, “about the fact of having more choice, how the person could put a bit more of the imaginary image into it. Maybe there’s something that comes closer to a projected or felt identity than the image in the mirror.” The filtered image perhaps grants access to a dimension of identity that the image in the mirror does not: that of felt identity.

In a mediation workshop, this duality can become an object of work. One can propose to participants that they make a self-portrait “as I am” and a self-portrait “as I feel myself to be,” show them to the group, listen to what others see in them, and elaborate on the gap between the two, without having to choose which is “true.” The mask is not the reverse of the face. It is an extension of it, sometimes a protection.

Paul Ricœur, in Oneself as Another (1990), showed that narrative identity, that is, the way we tell ourselves to ourselves and to others, is constitutive of personal identity. Our identity is not a fixed biological given; it is a narrative that evolves, that is rewritten, that integrates new experiences. The selfie and the filter are tools of this narrative, in the language of our time.

What we can learn

I have learned something from these exchanges with professionals who work daily with people who use filters and selfies extensively. We adults, the professionals, those “educated by image” as one educator said with involuntary irony, do not understand everything. And our incomprehension is not a diagnosis.

When a young man posts a video of himself shirtless at the gym, the gesture is foreign to me. It is not my culture, not my relation to the body, not what I would do. But the fact that I would not do it does not mean it is a symptom. It is perhaps another relation to the place of the body in social space, a relation I have to learn to understand. The boundaries of intimacy are not the same for everyone, and they shift.

What seems important to me, in our mediation stance, is to hold two things together. There can be suffering in the relation to images of self, and that suffering must be supported. And the recourse to filters and selfies is not for that reason a symptom: it is work, exploration, identity construction with the means of the time. Our role does not consist in bringing people back to a “true” image of themselves, because that image does not exist. It consists in supporting them in their identity work, by opening spaces where they can play with their images, question them, share them, transform them, in a frame of trust that does not judge them.

Some proposals for practical application

I close with a few concrete proposals, because to my mind a reflection on mediation only holds if it takes shape in actual devices. These proposals are not meant to be exhaustive; they indicate working directions I have experimented with or seen experimented with, and which seem to me consistent with what precedes.

The sidestep with the filter. Rather than banning or stigmatising the filter, one can make it an explicit object of work. A possible instruction: each participant produces two self-portraits with their phone, one without filter and one with a filter freely chosen. The two images are projected together in large format in the group, without saying which is which. The other participants say what they see. They do not say whether the image is “true” or “false”: they say what the image does to them, what it evokes, what they see of the person who made it. The explicit instruction is not to rank. This shifts the question of the “true face” toward the question of what each image does, and lets the group discover that both images speak, each in its own way, of the same person.

The gap as material. Another proposal is to ask each participant to choose, from their phone, two images of themselves that seem distant or even contradictory: an image they like and an image they dislike, or two images from different ages, or two images from different contexts. The exercise consists in telling out loud what is at play in the gap between the two. Not in choosing which is true: in inhabiting the gap. This is putting into practice what Ricœur calls narrative identity. We are not an image, we are the narrative that links them together.

The recorded voice workshop. To approach these questions through another door than that of the visual image, one can propose work on the voice. Each participant records themselves telling something simple, for example a daily journey. They listen back to themselves, then play it for the others. The instruction for the others is to say what they hear in the voice. This device shifts the question of non-coincidence between self and image onto the sound terrain, where normative pressure is lower and where it is easier to acknowledge, without drama, that one is not the image one has of oneself. Once that experience has been gone through with the voice, the return to images of the face happens with greater suppleness.

The avatar as exploration of felt identity. In workshops where digital tools allow the creation of avatars (video games, customisation applications, virtual reality software), one can propose an open instruction: create an avatar that does not look like one’s image in the mirror, or that resembles otherwise, or that expresses a dimension of self that one does not usually show. The avatar functions here as a projection more than as a lie. Sharing it in the group, saying what one wanted to put into it, listening to what others see in it: this is exploring felt identity without having to justify it in the biological real.

The collective gaze. All these proposals rest on a single principle, which I have been experimenting with for a long time: the collective gaze on images. An image projected in large format in a group and looked at together changes nature. It steps out of the daily flow. It becomes an object of common thought. And each person sees different things in it, which makes it apparent, by the very fact of the device, that the image does not have a single meaning. This mechanism is powerful for work on the image of self because it says, without having to theorise it, that what others see in me is not what I think I am giving to be seen, and that this discrepancy is precisely the space within which something can be said and worked on.

The ethical frame. All these devices are only worth what the trust they establish allows. A workshop on the image of self can bring up very intimate things, sometimes fragile. The frame must be set out clearly: what is said in the workshop does not leave the workshop, no one is required to share what they do not want to share, and judgements have no place. The mediator is not a judge, much less a therapist: they are someone who proposes a device and accompanies what is at play in it, redirecting if necessary toward competent professionals when something heavier comes up.

The filter is neither the problem nor the solution. It is a material. Everything depends on what one does with it, and on the frame within which this work unfolds.