The Scene Before the Subject
A phenomenologist declared there could be no phenomenology of the infant. His wife, an analyst, answered with a single word.
I once heard a philosopher argue that there could be no phenomenology of the infant, and the argument was not foolish. It was nearly airtight. Phenomenology is the disciplined description of experience as it gives itself, and it is done in the first person: I attend to my own experience, I suspend what I assume about it, I vary it in imagination until the invariant structure shows through. The method needs a someone who can perform it—who can reflect, bracket, report. The infant can do none of this. It cannot tell you how the breast appears to it, or what the dark is before there is an inside and an outside to divide it. It furnishes no first-person datum at all. So there can be no phenomenology of it. Of birth, perhaps; of the body, of finitude. But not of the one who has not yet become anyone to have an experience in the first place.
His wife was in the room. She is a psychoanalyst. She called out a single word.
“Bick.”
Esther Bick spent her career producing something very like a phenomenology of the infant, and she did it by the one route the philosopher had ruled out in advance: she gave up the first person. Her method of infant observation, still taught, sends an observer into a family week after week through the baby’s first two years, to watch and do nothing else—to take no notes in the room, to refrain from intervening, to write the visit up only afterward and bring it to a seminar where the temptation to interpret too soon can be held off a little longer. It is description with the intervening hand tied. It is a suspension of the natural impulse—to help, to explain, to make the strangeness familiar—so that the phenomena can show themselves. It is, in other words, an epoché with a stroller, a disciplined bracketing aimed not at the observer’s own consciousness but at a life that cannot narrate itself and must therefore be read from the outside. And it works. It yields the structures: the skin that holds a self together, the terror when the holding fails, the rhythms of gathering and coming apart. The philosopher was right that the infant cannot give the account. He was wrong about what follows. The account can be reconstructed.
That single word is the whole of this essay. Object relations is the phenomenology of the infant that phenomenology said it could not have—and once you see that, you see that phenomenology and psychoanalysis were never the strangers they take themselves for. They are siblings, separated young, who meet again at the cradle.
Begin with the difference Bick’s interruption exposes, because it is sharper than it looks. Freud was not a phenomenologist. Freud was an archaeologist, and the metaphor is his own—the analyst as excavator, clearing the strata, recovering what was buried, reading the present as a ruined surface over deposited layers of the past. Archaeology is a method for recovering contents: the repressed memory, the disavowed wish, the scene laid down years ago and built over. The dig goes downward and backward, after the thing that was hidden, to bring it up into the light where it can be known.
Object relations did something else with the same downward gaze. It stopped digging for buried contents and began describing buried structures—not the particular forgotten scene but the invariant shape of the infant’s affective world, the form that any infant’s experience takes. Klein’s phantasy, the good and bad object split apart. Bick’s skin and the dread beneath it. Bion’s container and contained, the mind built out of another mind’s capacity to hold what the infant cannot yet hold. Winnicott’s holding, his potential space, his mother who is not yet an object because there is not yet a subject across from her. None of this is a recovered content. All of it is a described structure—which is to say, it is phenomenology, the description of how experience is organized—but the structure belongs to someone who cannot do the describing, so it has to be reached the long way, by reconstruction. From the child’s play, where the phantasy stages itself. From the transference, where the earliest configurations come back to be lived again in the room. From the observer’s two years on the sofa in the corner. The infant does not report; the structure recurs, and where it recurs it can be read. Phenomenology by reconstruction is still phenomenology. It simply gives up the privilege of the first person and pays for the loss with patience.
If they are siblings, they had the same parents, and it is worth saying who.
Phenomenology and psychoanalysis are both children of 1900—Husserl’s Logical Investigations and Freud’s Interpretation of Dreams fall in the same two years—and both were born in revolt against the same parent, the naturalism that was then trying to dissolve the mind into physiology and the meaning of a thing into the measurement of it. Both refused. Both insisted that lived experience has a structure of its own that cannot be cashed out in nerves or behavior. Both built a discipline out of suspending the obvious so that the hidden could come forward—and here the family resemblance is almost embarrassing. Husserl’s epoché, the bracketing of the natural attitude, the deliberate refusal to take the surface of the world at its word, has a near-twin in Freud’s prescription for how the analyst should listen: with evenly suspended attention, gleichschwebende Aufmerksamkeit, holding back the urge to select and judge and emphasize, letting the material arrive without grabbing at it. Two disciplines of restraint, each clearing the same throat before the same kind of speech. Both, in the end, are theories of the latent—of the pre-reflective horizon that frames every act of consciousness without being looked at, on the one side, and of the unconscious that speaks through the cracks of what we meant to say, on the other.
So why the estrangement? Three reasons, and they matter, because they tell you where the reunion has to happen. The first is the unconscious itself. Phenomenology is a philosophy of consciousness; even the pre-reflective is, in principle, available to reflection, recoverable by the right turn of attention. The Freudian unconscious is not. It is barred by structure, not by inattention; you can never simply look and find it, only infer it from its effects. To the phenomenologist this is an affront—a region of the psyche declared off-limits to the very self-givenness the whole method is built to honor. The second reason is determinism. Freud’s metapsychology is a machinery of forces, drives and cathexes and economies, and a machinery leaves no room for what we mean by transcendence—for the freedom and the beyond that the existential and ethical strands of phenomenology care about most. The third is access. Phenomenology starts from my own experience, given to me directly. Psychoanalysis works on the other’s psyche, through the transference, by reconstruction, from the outside.
Notice that all three divergences converge on a single scene. The unconscious, the question of freedom, the problem of reaching an experience that is not your own to report—these come to a head exactly where the psyche is pre-verbal and the subject is not yet there: in infancy. And that is the scene at which Bick’s method, the reconstructive observation of a life that cannot narrate itself, is precisely the bridge between the first person and the other. The cradle is not one topic among the things these two traditions might discuss. It is the place the fork rejoins, because there the psychoanalytic object and the phenomenological problem are the same object and the same problem—how a self, and an Other, get made before there is anyone to make them.
Which brings me to the parallel the whole essay was built to reach, and which I owe to a one-line note I have carried for years: that Winnicott and Levinas are describing the same scene from two frames, and that the frames have a common root.
Take Levinas first, because he states the structure in its purest form. Against the whole tradition behind him, Levinas held that ethics is first philosophy—that it precedes ontology, precedes even the question of being—and that it begins with the Other. Not the other as I conceive him, not the other reduced to a theme or an object of my knowledge, because that reduction is what Levinas called totality, and totality is a kind of violence: it takes the Other’s infinity and shrinks it to fit my categories. The Other comes to me as a face, and the face exceeds every concept I could form of it, and it commands me—thou shalt not kill—and it puts me in question, and I find myself responsible for it before I have chosen anything at all. Responsibility precedes freedom. I am, in his hard phrase, hostage to the Other. And the point worth holding for what follows is where Levinas planted this against. He planted it against Husserl, for whom the Other is a problem to be solved by the transcendental ego—the famous strain of the Fifth Cartesian Meditation, the attempt to get the other person constituted out of my own consciousness by analogy, an attempt most readers find never quite succeeds. Levinas’s answer was to refuse the problem entirely. The Other was never mine to constitute. The Other is there first, before me, founding me rather than founded by me.
Now hear Winnicott, in the most quietly radical sentence in the literature: there is no such thing as a baby. If you set out to describe an infant, he said, you will find you are describing an infant and someone—that the baby does not exist on its own but only inside the unit of maternal care, that before there is a self there is a holding, a preoccupation, an environment that adapts itself to a need the infant cannot yet even feel as a need. The mother is not the infant’s first object. She is prior to objects, prior to the subject who would have them. She is there first.
That is the same structure twice. For Levinas and for Winnicott both, the Other precedes the subject and founds it—comes before the self that would otherwise have to constitute the Other and never could. The philosopher reached this as a refutation of the whole egology of modern thought. The pediatrician reached it by watching mothers and babies. And the bridge between the two registers is the figure Levinas himself reached for when he needed an image for responsibility at its most extreme. In his later work the figure of substitution—of being-for-the-other to the point of bearing the other in oneself—is maternity: the body that carries and answers for a life not its own, gestation as the very shape of being-for-another. The Levinasian Other is feminine, he said, by way of being maternal; and though that gendering is contested, and rightly examined, the maternal figure is no decoration in him—it is where his ethics goes when it goes as deep as it can. Levinas reached for the figure of the mother. Winnicott reached for the actual one. Same Other. Two registers—the metaphysician’s and the clinician’s—of a single priority.
Here is the claim that makes the parallel more than an elegant coincidence, and it is the heart of the thing. Levinas describes the structure of the ethical relation. Winnicott describes its formation. The illusory, transitional space that Winnicott found in the nursery is the developmental precondition of the Levinasian openness to the Other—and you can name the precondition in three parts.
The first is illusion. The good-enough mother, by adapting so closely to the infant’s need, lets the infant feel that it created the very breast that was in fact already there—and out of this comes the transitional object, the worn blanket that is neither simply found in the world nor simply made up by the child, but both, held in an intermediate space where the paradox is never to be challenged. You do not ask the child whether it found the rabbit or invented it. The not-asking is the whole point. And the not-asking is exactly the suspension the ethical relation will later require: the capacity to let the Other be at once independently real and met-in-relation, neither possessed as my construction nor crashing in as brute fact, but held in the space between, where I do not demand to settle which. The second part is doubt—the tolerance of not-knowing that Bion, borrowing from Keats, called negative capability, the capacity to remain in uncertainties and mysteries without any irritable reaching after fact. This is what lets the Other stay a mystery rather than collapse into a comprehended theme; the grab for certainty about another person is the small daily form of totality. (I have followed this capacity through both of the essays this one stands beside—there it was the cognitive face of maturity, and there it was the gap that the foreclosed mind cannot reach—and here it does its third work, as the doubt that keeps the Other open.) The third part is play, the activity of the potential space itself, where the infant first risks the not-me and is not annihilated by it, where the relation to what is outside the self is rehearsed in safety. Illusion, doubt, and play: the willingness to let the object be both made and found, the willingness to not-know, and the free rehearsal of the not-me. Together they are the capacity to meet an Other as an Other. Levinas told us what that meeting is. Winnicott showed us where it is built.
And the negative case completes the proof, because it is the same hinge seen from the side of failure. When the environment cannot meet the infant’s spontaneous gesture, the infant builds a false self—a compliant surface organized around managing the world rather than living in it, a caretaker self that hides the true one and, in the worst cases, leaves the person feeling unreal, futile, not quite alive. Now set that beside Levinas, and the false self turns out to be a failure of responsibility in his exact sense. Because responsibility, for Levinas, is response—answer to the face that addresses me. And the false self cannot respond. It can only comply. It meets the Other not as a face to be answered but as a pressure to be handled, a demand to be managed and survived. The compliant self is, structurally, the self that cannot be responsible, because it has substituted accommodation for the spontaneous gesture out of which a real answer would have to come. (And from the other essay: you cannot be responsible for an Other you have split into a good part and a bad one—the Kleinian concern for the whole object, the guilt and the wish to repair, is the affective seed of the very responsibility Levinas describes. Reparation is proto-ethics.)
I have to be honest about an objection, because Levinas would raise it and it is the strongest one available. He would not thank us for giving his ethics a developmental history. For Levinas the responsibility is an-archic—without origin, prior to any beginning, exceeding every story you could tell about how it came to be—and to “explain” it from the nursery looks like exactly the move he spent his life refusing: the reduction of the ethical to something else, to psychology, to a cause, another totality dressed as understanding. The objection is right, and the answer to it is the same discipline that runs through these essays. The developmental account is not a reduction of the ethical relation to its conditions. It does not say the openness to the Other is nothing but illusion and play, any more than the structural reading of autism says a person is a failed overwriting—that was the Bettelheim error, the causalizing of a structure, and it does its damage at every level it is committed on, the ethical no less than the clinical. What the account describes is not the ethical relation itself but the conditions under which it becomes available to a subject—the access, not the ground. The relation stays irreducible. Levinas keeps his infinity. What Winnicott adds is the news that the door to it is built, and can fail to be built, in a room none of us remembers.
Pull back, and the reunion shows its method, which is the last thing I want to name. The move this essay makes—reading Winnicott and Levinas as one scene in two frames—is the move the two essays beside it were already making: reading Klein’s positions and Lacan’s registers as one architecture in two vocabularies. That is not eclecticism. It is a method with a name. It is a descriptive phenomenology of the psychoanalytic theories themselves—the eidetic procedure, free imaginative variation, run not across experiences but across the theories, holding them side by side and varying their vocabulary until the invariant structure they are all circling stands out from the words that differ. Positions, registers, holding, the face: the terms diverge because the schools diverge, and they converge on structure because they are all reconstructing the same scene, the constitution of a self and an Other before there is a subject to report either one. The three essays were a single investigation. This is the panel that turns around and shows you what the other two were doing.
And so back to the lecture, and the one word from the audience. The philosopher was right that the infant cannot speak—cannot reflect, cannot bracket, cannot hand you the first-person account the method seems to require. He drew the wrong conclusion. The infant who cannot give the account is precisely the one in whom the first person is being made, along with the Other it will answer to and the responsibility that will bind them before either has a name; and to reconstruct that scene is not to betray phenomenology but to carry it back to the origin it could never reach from the armchair, because the armchair is already on the far side of everything the cradle decided. The deepest phenomenology is of the one who cannot perform it. That the Other was there first—before the self, before the word, before any report—is not a hole in the description. It is the description. We are each of us constituted by a welcome we cannot remember, and the whole of ethics may be nothing more than the trace it leaves: the dim, binding sense, never quite recoverable and never quite shakeable, that someone was holding us before we were anyone to hold.