Witness

Before

There is a condition that precedes every human life as it is actually lived. Ordinary language has no name for it, because ordinary language was built by people who had already left that condition behind. It cannot be remembered, either. The self is the apparatus that would do the remembering — the autobiographical machinery, the capacity to look back on a life — and that apparatus did not exist yet, inside that condition. So it is not an experience in the usual sense. It is the condition that comes before experience has anyone to have it.

In the earliest weeks of life, an infant does not yet distinguish between itself and the world around it. The developmental psychologist Margaret Mahler called this the symbiotic phase — a period when infant and mother function as one, with a single barrier between that unit and everything outside it. Mahler’s account is built on years of direct observation of infants over their first three years, and it remains the foundational map of this territory. In the symbiotic phase, the infant does not experience its mother as a separate other. There is no other yet. There is no self to register the distinction. When a need arises and something meets it, that isn’t two events happening in a relationship — need and satisfaction are a single continuous condition. The infant, in Mahler’s account, is the world it inhabits.

Mahler’s larger argument is that biological birth and psychological birth are two different events, happening on two different timelines. Biological birth is dramatic, observable, and over quickly. Psychological birth is not. It is a slow, internal process — the infant gradually separating from the symbiotic unit, a self gradually emerging that knows itself as distinct from its mother. Mahler’s framework maps this process carefully: the subphases, the developmental milestones, the specific behaviors that mark each transition. But her framework’s job is to map the process, not to name what the process costs. That’s the piece she leaves unstated.

What it costs is the original condition.

Sigmund Freud came close to naming this condition, though in a different context. In the opening pages of Civilization and Its Discontents, he took up something the poet Romain Rolland had described: a felt sense of indissoluble bond with the external world as a whole, an “oceanic feeling” that Rolland believed was the true source of religious experience. Freud’s reading of it was developmental. He argued the oceanic feeling is a preserved residue — what’s left over from the infant’s primitive ego-feeling, from before the ego existed at all. In Freud’s account, the ego emerges at the moment the infant first discovers that its sources of satisfaction are external, are not-me. The clearest example is the absence of the breast: the infant learns there is a difference between itself and the thing it needs. The self is the product of that discovery. Before the discovery: the oceanic condition. After it: a subject who will spend the rest of its life trying to recover something it has no name for.

Freud treated the oceanic feeling as an interesting artifact — a psychological curiosity, useful mainly as evidence for a developmental account of religion. This piece reads the same structure differently. The oceanic state is not just a backdrop that development happens against. It’s what development ends. And its ending isn’t just one milestone among others. It is the event that produces the self in the first place — the necessary rupture without which there would be no person left to feel anything at all.

So, before moving on: there is a condition, prior to selfhood, that has no name and cannot be remembered, because the self that would remember it is the very thing produced by that condition’s ending. That ending is where the piece begins.

The Threshold

Biological birth and psychological birth, then, are not the same event — the previous section already established that. Between them lies a period that ordinary developmental language has no good name for. Pediatricians have recently started calling it the fourth trimester: the first three months of life outside the womb.

But the mother’s knowledge of the infant doesn’t begin at birth. It begins months earlier, in a register that has no counterpart in any relationship that comes after. The infant makes itself known somatically — through pressure, through movement, through the way it progressively transforms the body that contains it, through the increasing discomfort of a container approaching its limit. The mother knows this specific presence through her own body, before she has ever seen it, heard it, or held it. She knows it from the inside. All the knowing flows in a single direction. It is the mother who knows; the infant cannot yet know anything back.

Birth, then, does not introduce the infant to the mother — introductions are for strangers, and the infant is not a stranger to her. What birth does is translate what the mother already knows into every one of her remaining senses at once. Labor itself is one of the most physically extreme experiences a human body undergoes: pain at a level that would signal catastrophic damage in any other context, a body opening in ways that will permanently change it. That ordeal ends with the infant arriving into the full force of a gaze that has been accumulating for nine months. For the first time, the mother meets — with sight, sound, touch, all at once — something she already knows more completely than it will ever be known again. The infant is not a stranger to her. It is the source of everything that just happened, finally made visible.

What the infant arrives into is not care in the ordinary sense. It is the concentrated attention of a being who has been in continuous bodily relationship with it for its entire existence — who knew it before it had a self, before it could signal anything, before it knew itself. This piece will call that condition coverage, and will define the word precisely later on; for now, take it to mean something close to total, uninterrupted knowing, with no gap yet between knower and known. The attention the infant receives in the days and weeks after birth is historically saturated in a way no relationship that comes after can replicate. Nobody will ever know this self the way the mother knew it before it was a self.

The biology extends this pattern. For the first several days after birth, before mature milk arrives, the breast produces colostrum — a substance so different in composition from milk that it belongs to its own biological category. It is not food in the ordinary sense. It is a transition medium. It is dense with antibodies, calibrated to the marble-sized stomach of a newborn, built to coat the infant’s gut, transfer the mother’s immunity directly, and seed the microbiome with what the infant’s own system cannot yet produce. In other words: the mother’s body keeps performing, now from the outside, the same functions the womb was performing from the inside. The umbilical cord has been cut. The physiological continuity has not. Colostrum even changes its composition daily as the infant develops, adjusting in real time to what the transition requires — until mature milk comes in, and the bridge has finished its work.

This is the threshold, then: the infant is outside the container that held it, and still inside the original condition. Both are true at the same time. The self is pending — not yet produced, not yet lost, suspended in the one moment when the original condition and the separated world briefly overlap. Coverage is still as close to total as it will ever be again: the mother’s entire orientation is toward the infant, and the infant’s entire world is the mother’s presence. But a gap has opened that did not exist before. The infant is outside now. It is being held from the other side of a boundary that is new. And here, for the first time, coverage begins to become something else — what this piece will call witness.

At six to eight weeks, something new happens: the infant smiles. This isn’t the reflexive grimace newborns show in the first days — a neurological firing with no social content behind it. This is different. The psychoanalyst René Spitz studied its emergence in the 1940s and called it the smiling response. It’s the first evidence that the infant is beginning to register the other as other — the first signal that one specific presence, this face, is distinct from the rest of the world and produces something in the infant when it appears. It is the infant’s first relational act. The infant isn’t responding to stimulation in general. It’s responding to a presence, specifically. And — this is the crucial part — it is acting. The smile is the infant’s first successful attempt to produce a specific response in a specific other person. The mother’s face changes, warms, turns toward the infant more fully. The loop closes. For the first time, the infant has discovered that it can do something to the world.

The mother receives the smile as arrival. Developmentally, though, it’s also the first evidence of departure. To smile at another, the infant must have begun to register that the other is other — which requires, at some level, that the boundary between self and world has already started to form. So the same discovery cuts two ways at once: the infant who discovers it can act on the world has also discovered, in the same moment, that the world is not itself. The smile is the threshold’s first visible crack. It is the moment coverage shows its own edge, from the inside. The original condition does not end here, not yet. But this is where it begins to know that it will.

So: birth doesn’t create the gap between infant and mother — it holds both conditions at once for a moment, coverage and separation overlapping. The first true crack in that overlap is the six-week smile, the infant’s first relational act, and its first evidence that coverage is starting to give way to something the piece will call witness.

The Rupture

Psychological birth — the first real experience of separation — is not a developmental achievement in the way walking is a developmental achievement. Walking is something a self learns to do. Separation is different in kind: it’s what produces the self that will go on to learn everything else.

This is the structural claim underneath everything else in the piece, and it’s worth stating precisely. The self does not exist first and then undergo separation as something that happens to it. The self is the separation. It is the residue of the original rupture — the record of what ended. To become a self is to become something that is no longer everything.

This claim can’t be confirmed by a controlled study, and it’s worth being honest about why. To test it at the level of precision that would settle the question, someone would have to induce in an infant the exact conditions that reveal the structure of this loss — and no ethical framework permits that experiment. But the control condition isn’t just practically unavailable. It’s unavailable because reaching it would require an instrument the structure itself rules out. The self cannot observe the event that produced it, because the observer is the product of the very thing being observed.

What can be studied is the consequence of the original condition’s absence.

The Signal

In the autumn of 1989, Western journalists and aid workers entered Romania after the collapse of the Ceaușescu regime, and they encountered something they had not anticipated. The state’s orphanages held more than one hundred thousand children — many abandoned by families who couldn’t afford to raise children under the natalist policies Ceaușescu had mandated — and those orphanages were eerily quiet. The children were not crying. They had cried, earlier, as infants do. But in an environment where crying produces no response — no face turning toward them, no warmth, no adjustment of anything at all — the cry stops.

Not because the need stops — the need does not stop. What stops is the signal, because signaling into silence isn’t really signaling at all. The cry is an instrument, calibrated to a social environment that’s supposed to respond to it. When that environment returns nothing, the instrument goes quiet. What changes is the architecture built to express the need — not the need itself.

The journalists and aid workers who first looked at these children located the cause in the room — assumed, in other words, that something was wrong with the caregivers. But the research record shows otherwise. The caregivers were not indifferent. They brought soap and food from their own homes. They dug the graves of the children who died. One carer, interviewed decades later, described watching the children rock back and forth — the self-soothing that had replaced the human contact that never came — and said it was the hardest part of the work. What she could not change was the ratio: one caregiver for every fifteen infants, one for every twenty-five toddlers. A single person cannot be the receiver for twenty-five signals at once, no matter how much she wants to be. The witness function — the piece will define this term precisely soon — was not withheld by anyone in that room. It was arithmetically foreclosed before anyone in that room had a chance to withhold or give it.

The architecture that produced that ratio wasn’t in the room at all. It sat three levels up, in the policy apparatus: the natalist decree that had manufactured these children in the first place, the state ideology that treated institutional care as superior to family care, and the systematic underfunding that made the ratio what it was. The caregivers were near the very end of a resource chain that had been deliberately starved from the top. In other words: the cause was not where the observers’ instruments were calibrated to look for it. They were looking in the room. The cause was in the policy above it.

The mechanism is not particular to orphanages. It runs wherever power has organized itself around the systematic refusal to witness.

René Spitz had documented this exact sequence decades earlier, in the 1940s. He studied infants in foundling homes in South America — institutions that provided adequate physical care but almost no maternal contact — and observed three phases in how infants responded to the absence of an attachment figure. First, protest: the infant cries, reaches, demands. Then despair: the infant withdraws, slows down, goes quiet. Finally, detachment: the response to the attachment figure gets actively suppressed, built into the infant’s functioning as a kind of standing architecture. Spitz called this anaclitic depression — a term that has been partly absorbed into clinical language even as the mechanism behind it has been partly forgotten. What Spitz was watching was not a psychological disorder grafted onto an otherwise normal infant. It was the infant’s nervous system learning the actual shape of the world it found itself in.

John Bowlby built on Spitz’s work, along with his own observations of children separated from their mothers during wartime hospitalizations, and formalized this sequence as a biological system rather than a psychological curiosity. In Bowlby’s framework, attachment is not learned behavior. It is an innate behavioral system, shaped by evolution, operating from the first hours of life. An infant is biologically built to seek closeness to a caregiver — to signal, reach, cry — because for most of human evolutionary history, closeness to a caregiver was survival. The sequence Spitz observed — protest, despair, detachment — is Bowlby’s sequence too, and in Bowlby’s own words, each stage does more serious harm to the infant than the one before it. Detachment is not recovery. It is the defensive exclusion of the attachment function: the signal apparatus going quiet, not because the need is gone, but because the need has learned there is no one there to receive it.

Decades later, Romanian data confirmed all of this — at scale, and down at the neurological level. Charles Nelson, Charles Zeanah, and Nathan Fox ran the Bucharest Early Intervention Project starting in 2000: the only randomized controlled trial ever conducted of foster care as an alternative to institutional care, following 136 institutionalized children across two decades. What they found was that psychosocial deprivation didn’t just produce behavioral problems. It produced measurable changes in brain development — reduced gray matter, disrupted cortisol regulation, high rates of attachment disturbance and developmental deficit. Children placed into foster care in roughly the first two years of life showed markedly better outcomes than children placed later, across every brain and attachment measure the study tracked. And in most domains, even the children who received foster care never fully caught up to children who had never been institutionalized at all. The architecture inside them had changed. What had changed was not some minor or incidental thing.

The children moved from institutions into families recovered — partially. And the partiality is the finding, not an asterisk on it. Something was addressed by restoring the witness function. Something else was not fully recovered, precisely because it had never been present to begin with.

So: the cry that stops is not a resolved problem — it’s an adapted-to problem. Spitz named the sequence, Bowlby explained why it’s built into biology, and the Romanian data show the cost is measurable in the brain itself, not just in behavior. What was missing in those orphanages is the thing this piece has been calling coverage, and is about to name properly: witness.

The Structure of the Loss

The Romanian case is usually presented simply as evidence for the importance of early attachment — and it is that. But it’s also something more specific: it demonstrates the structure of the original condition, by showing exactly what happens when that condition is missing from the very start.

In ordinary development, coverage recedes gradually, and the recession is scaffolded — supported at every step. The symbiotic phase (recall: the early period when infant and mother function as one) doesn’t end abruptly. The mother doesn’t withdraw. She stays present while the infant differentiates, available as the fixed point against which separation gets measured. Mahler has a term for this: checking back. A crawling infant who has moved a few feet away from its mother will turn around, again and again, to confirm she is still there. The self that is forming, forms in the presence of the very thing it is forming away from. The loss is real. But it happens inside a continued relationship with the source of what’s being lost.

What the Romanian orphanage removes is exactly that scaffold. There, the witness function is absent not at the end of the symbiotic phase, but from the very beginning. An infant cannot differentiate its way out of a relationship that was never there to begin with. Instead of gradual recession, what it encounters is structural silence: a world that does not register its signals, does not respond to its need, does not confirm that anything is present at all.

The infant who stops crying has not resolved the problem. It has adapted to a world in which the problem cannot be resolved. Those are not the same thing — and mistaking one for the other, reading the silence as stability, reading the stillness as adequacy, is the single most visible instance of the mechanism this piece is examining.

This loss precedes language. It precedes memory. It precedes even the self that would eventually need to report it. Adults who grew up in those institutions could not say what was missing, because they had no words for it and no memory of anything different. The flatness, the disconnection, the inability to reach others or be reached by them — these were simply the furniture of the only world they had ever known. There was no prior condition to compare it against. There was no memory of coverage to register as absent. There was no instrument inside them capable of measuring the distance between what was and what might have been.

So: ordinary loss happens gradually, inside a relationship that stays present while the self forms — Mahler’s checking-back. Institutional loss removes the relationship itself, from the start, which is why it can’t even be named by the adult who lived through it. That difference is why the next section has to define, precisely, what was actually missing.

What Witness Is

This is the word the piece has been circling since the very first section, gesturing toward the definition each time without quite giving it. It needs precision now, before it can carry the piece’s weight.

Witness, as this piece uses the word, is not attention. A camera provides attention. Witness is not care, either — an institution can provide care, as the Romanian orphanages show, without providing witness. Witness is not even love in the ordinary sense, since ordinary love is a feeling held privately by one subject. Witness is something prior to all three. It is the presence of a contingent, responsive other: a subject who not only registers that the self exists, but responds to its states reliably enough that the self learns three things at once, through that response.

That it exists.

That what it feels is real.

That what it does matters.

These are not three separate confirmations arriving through three separate mechanisms. They are one confirmation, delivered through a single structure — the other who receives the infant’s signals, reflects its states back, and responds to its actions. Go back to the six-week smile: the infant who smiles and sees the mother’s face change in response has not merely been seen. In that same moment, it has discovered three things at once. That it is a self. That its interior states have a shape the world can read. That it can act on the world, and the world will respond. Existence, emotion, and agency all arrive together, from the same source, through the same single event.

The developmental research backing this is precise. When a caregiver reflects an infant’s emotional displays back to it — meeting distress with concern, meeting delight with delight — that mirroring isn’t simply reassuring. It is the actual mechanism through which the infant comes to know its own internal states. An infant does not know what it feels until it sees what it feels reflected on the face of a responsive other. So the self that forms after psychological birth doesn’t arrive already equipped with self-knowledge. It acquires self-knowledge through witness, afterward. To be seen, at this level, means learning that one is a self that can be seen. To have one’s states reflected means learning that one has states at all. To have one’s actions met with a response means learning that one is the kind of thing whose actions produce consequences in the world.

But witness — even total witness — is not what the original condition provided. This distinction is the hinge on which the piece turns, and it must be held precisely.

Witness requires a gap. It requires a self on one side to be witnessed, and another self on the other side to do the witnessing. It requires a signal to cross the distance between them, and a response to travel back. The infant in the symbiotic phase cannot be witnessed in this sense, because there is not yet a self there to witness — no subject on one side of a gap, reaching toward a subject on the other. What the original condition provided was not witness. It was coverage: the complete, continuous, historically saturated attention of a mother whose knowledge of the infant came before the infant’s own knowledge of itself — attention directed at a presence that could not yet register that it was even being known. Coverage requires no gap, because there is no gap yet. Coverage requires no signal, because nothing has separated far enough to need to send one.

When psychological birth happens, the gap opens. The self forms. And what the self inherits — what it is actually made of — is the shape of what just ended. It arrives capable of witness at the exact same moment coverage stops being possible. It can now be seen, reflected, confirmed. But it can no longer be covered. The mother who knew it from inside her own body — who carried nine months of somatic knowledge before the first full-sensory encounter, who adjusted colostrum daily to whatever the transition required — that mother is now on the other side of a gap. Her knowledge of the infant is still more complete than anyone else’s will ever be. But now it has to cross a distance to reach the infant. Everything, from here forward, has to cross a distance.

This is why nothing ever fully satisfies. The self spends its life seeking, through witness, the thing only coverage could ever provide. No matter how complete, how sustained, how loving witness becomes, it can only confirm the self across the gap. Coverage was the condition before the gap existed at all. The self cannot recover coverage through witness, because witness is the very structure that replaced coverage the moment the gap formed. None of this means seeking more or better witness is wrong — it addresses a real need, and addressing it matters. What it does not do is address the original condition. The original condition was never a state of being witnessed. It was a state in which being witnessed wasn’t even a category yet — in which the question of whether the self was known couldn’t arise, because the self and its knower hadn’t yet become two separate things.

The child who says look at me has already understood all of this, without knowing that it understands it. The request isn’t really for observation. It’s for confirmation that crossing the gap is still possible — that the world, on the far side of that distance, is still oriented toward the self that’s reaching across it. The climb up the play structure, the drawing held up for approval, the small accomplishment: none of it feels complete until it has been received by someone else. The doing and the being-seen-to-do turn out not to be separable at all. This is the six-week smile, grown up into language. The mechanism is identical — it’s just that now the self has grown large enough to name what, as an eight-week-old, it could only demonstrate.

The need for witness, then, is not one human need sitting alongside a list of others. It is the structure of selfhood itself: the shape of the gap the self was born into, the outline of what the original condition was, traced backward from the self that formed when it ended.

Every relationship that comes afterward tends to get calibrated — below the level of conscious awareness — against this original structure. Not against a remembered experience; remember, the original condition can’t be remembered, because the apparatus that would do the remembering didn’t exist inside it yet. It’s calibrated against a structural expectation instead: the sense that total coverage, total response, total adequacy is the baseline, and that every actual relationship falls short of that baseline by definition. This calibration happens before language, before cognition, before there is even a self capable of explaining why its relationships feel insufficient. It gets installed in the architecture before the architecture has a name for itself.

The standard account of loneliness locates its cause in failed relationships. That’s a category error. It mistakes variation within a structure for the source of the structure itself. And the standard account can’t see past this, for a specific reason: the instrument doing the observing — an ordinary self, using ordinary introspection — is built inside the very condition it’s trying to describe. It can’t see that loneliness isn’t produced by inadequate connection. It’s produced by the recession of something that connection, however adequate, was never designed to replace in the first place.

None of this dissolves the reality of ordinary interpersonal experience. Relationship failure is real. Deprivation, neglect, and rupture later in life are real, and they act on the structural condition this piece is describing — they intensify it, distort it, sometimes partly mask it. What the piece is identifying is the baseline underneath all of that variation: a structural condition every self inhabits, regardless of whether its later relationships turn out to be good or poor, loving or withholding, adequate or catastrophically insufficient. The structural condition is what those variable relationships operate on top of. It is not what they explain.

The claim can be stated simply: the self is what remains when coverage ends, and what remains spends its existence seeking through witness what witness was never designed to provide.

So, to gather the definition in one place: coverage is what preceded the self — total, gapless, requiring no signal. Witness is what replaced it once the gap opened — real, valuable, and structurally incapable of closing that gap back up. Loneliness, on this account, is not a failure of witness. It’s the ordinary condition of being a self at all, reaching for something witness was never built to restore.

The Long Negotiation

The self that emerges from psychological birth doesn’t emerge into emptiness, though. It emerges into a world full of relationships — all of which partially approximate the original condition, and none of which fully restore it.

The mother remains, for one. She is now an object within the infant’s world rather than the whole of that world, but she is present, responsive, attuned. And beyond her, a tribe forms around the growing child — family, peers, community — with each relationship extending the network of witness a little further, distributing across many people what was once concentrated in just one. The architecture of human belonging, across cultures and across history, is really an architecture of witness distribution: the ongoing attempt to maintain, through collective structure, something that approaches the coverage that was total at the very beginning.

The psychiatrist Irvin Yalom, cataloguing what he called the ultimate concerns of human existence, named one of them existential isolation: the unbridgeable gap between one person’s interior world and every other person’s interior world. Unlike ordinary interpersonal loneliness — which more or better connection can actually address — existential isolation, in Yalom’s account, cannot be closed. It’s built into the structure of being a self. As far as it goes, Yalom’s framing is accurate. What it doesn’t do is locate where the gap came from. For Yalom, existential isolation is simply a given of existence — the condition a person finds themselves in, inexplicably, once they examine their situation carefully enough. This piece proposes that it is not inexplicable at all. The gap is the gap left behind by the original merger’s end. The self did not arrive into isolation as a separate fact about it. The self is the isolation — produced by the rupture, carrying its shape forward.

Once this structure is visible, the institutions humans have built around it become legible too. Religious traditions across very different cultures organize themselves around experiences that dissolve or suspend the boundary between self and world: the unio mystica of Christian contemplatives, the fana of Sufi practice, the dissolution of ego reported in certain meditative states. These are not simply diverse phenomena that happen to feel similar to each other. They are recognizably related attempts — running through different channels, in different traditions — to approach the condition that preceded the self. The attempts are serious. Many people report genuine transformation through them. But there’s something none of them can produce, no matter how disciplined the practice: the original condition itself. That’s because the original condition is, by definition, the condition before there was a self to practice anything at all.

Romantic love reaches for the same territory. The experience of falling in love — the dissolving of the boundary between self and other, the feeling that the other isn’t separate anymore, that the world is finally complete — is not a distortion of reality. It’s an approximation of the original structure, produced by a nervous system that was built during the symbiotic phase and still carries that architecture forward. The approximation is close enough to produce genuine recognition, and that recognition is why falling in love feels like coming home. But the approximation is not close enough to hold. The self remains. The separation remains. And the moment it reasserts itself — when the other becomes distinctly other again, when the merger fades — may be the loss felt most acutely of all. The approach was close enough to remind the self of what it’s carrying. It just wasn’t close enough to let the self set that weight down.

So far, three of the negotiation’s attempts: the tribe distributing witness across many people, religious practice reaching for the pre-self condition directly, romantic love approximating it through the body. None of them close the gap. All of them are worth doing anyway.

The agency piece of witness — the confirmation that what the self does actually matters — turns out to be just as fundamental as the rest, and removing it turns out to be just as consequential. In the late 1960s, the psychologist Martin Seligman ran experiments, initially with dogs subjected to inescapable electric shock, and documented something he called learned helplessness. Repeated exposure to a world that doesn’t respond to action rewires the brain’s circuits to default to passivity — even later, once action becomes possible again. Dogs who had learned that nothing they did made any difference did not attempt to escape once escape became available. They lay down and accepted the shock. What Seligman was observing was not a belief the dogs had formed and could later be argued out of. It was a structural adaptation — the brain’s agency architecture concluding its experiment and setting the result as a fixed premise. The symptom profile he documented — motivational deficit, cognitive impairment, social withdrawal, apparent despair — is the clinical profile of depression. And the mechanism that produces depression in adults is the same mechanism the Romanian orphanage installs at the level of an infant’s very first calibration, before its architecture is even set.

The evidence runs the other direction too, showing what happens when agency is restored rather than removed. In 1976, the psychologists Ellen Langer and Judith Rodin ran a study at a Connecticut nursing home. Residents on one floor were told they were responsible for their own lives at the facility — including caring for a plant they got to choose themselves. Residents on another floor were told the staff would take care of everything, including watering a plant that had been assigned to each of them. The care given to both groups was attentive and genuine in both cases. What differed was contingency: one group’s actions produced consequences in the world, and the other group’s did not. At three weeks, the responsibility group showed measurably higher alertness, activity, and wellbeing. At eighteen months, the study reported a mortality rate in the responsibility group that was half that of the comparison group. A plant. The choice of which night to see a film. Permission to arrange one’s own furniture. None of these are profound restorations of the original condition — they’re the thinnest available approximation of the confirmation that what one does matters. And even that thin approximation was enough, set against a comparison condition of genuine, attentive, but non-contingent care, to produce a difference that shows up in the mortality data.

The communal deathbed worked the same territory from the opposite end of life. Picture the dying person inside a tribal container — the medieval English household, where neighbors gathered at a deathbed as a structural requirement, not a courtesy; or the Sora community, conducting its dialogues between the living, the dying, and the dead. In both cases, the dying person was held inside a sustained, role-bearing presence that their membership in the community demanded. The witness function, which had been distributed across a whole lifetime of relationships, concentrated itself again at the moment of departure. This is not a coincidence. The approach of death reduces the self back toward the condition it came from: its boundaries attenuate, the ordinary structures of selfhood grow less stable, and what remains is something closer to the original permeability it started with. The tribe’s presence at the deathbed was the witness function meeting that returning permeability — not restoring the original condition exactly, but holding the self as it approached its own dissolution with something close to the coverage that preceded its formation in the first place.

When the tribe left the deathbed — when institutional medicine moved dying into the hospital, when the community’s presence was replaced by professional care on a schedule — what was lost wasn’t merely comfort or ceremony. What was lost was the witness function, right at the moment of maximum need for it. In 1965, the psychiatrist Elisabeth Kübler-Ross pulled a curtain around a bed in a Chicago hospital, sat down, and started asking dying patients what the experience was actually like. The stages she documented — denial, anger, bargaining, depression — are usually read as a universal, invariant interior experience of dying. This piece reads them differently: they were shaped by the specific conditions Kübler-Ross was observing, the conditions of institutional, often solitary dying — the self navigating its own dissolution without the witness structure that dissolution had historically required.

The anthropologist Piers Vitebsky documented something that makes this especially clear among the Sora of eastern India. When their children converted to Christianity and withdrew from the community’s shamanic dialogues, Sora parents grew afraid to die. Vitebsky’s account is precise about what that fear actually was: not a fear of death. A fear of dying unwatched. That distinction matters. The fear was not of ending. It was of ending without witness — of the self’s dissolution happening outside the structure that had held it across its entire lifetime.

Put the two cases next to each other and the shape becomes visible.

At the beginning, the infant who stops crying in the Romanian orphanage.

At the end, the dying person in the hospital room.

The same structure. The same absence. The coverage that was total before the self existed — and that became witness the moment the gap opened — has been receding ever since.

That’s the negotiation, gathered up: tribe, religion, love, agency, and finally the deathbed — five different channels, all reaching for the same thing, none of them able to close the original gap, all of them still worth having. The final section explains why closing that gap was never going to be possible, structurally, no matter how good any of these got.

What Cannot Be Seen

The instrument of observation, in all of this, is the self. And the one event the self cannot observe is the event that produced it. This is not a contingent limitation — not a gap that better introspection, or more careful psychology, could eventually close. It is structural. The self cannot perceive the original condition, because the self is what emerged when that condition ended. There is no vantage point from which to observe the merger, because the observer is constituted by the merger’s rupture. The fish cannot see the water. More precisely: the fish is what happened when the water separated from itself.

This is why the loneliness never resolves. Every attempted solution — every relationship, every practice, every community, tradition, or ritual covered in the last section — operates from inside the very condition it’s trying to address. The self that seeks reunion with the original condition is itself the entity produced by the original rupture. It cannot find what it’s looking for, because what it’s looking for is the condition before it existed. The entity doing the seeking is, itself, the evidence that the return it’s seeking is impossible.

It’s worth addressing directly: this claim can’t be falsified, and that’s not a weakness in the argument. It’s a description of the exact epistemic situation the piece has been tracking all along. The control condition — the original merger, the state before psychological birth — is inaccessible. Not because it’s been lost to the historical record. Not because no one has run the relevant study. It’s inaccessible because the only instrument that could confirm or refute the claim is built inside the very event it would need to observe from the outside. The Romanian data comes closer than anything else to empirical demonstration here — not proof of the original condition’s structure, but a demonstration of what happens in its absence. The damage is the evidence of what was there.

This framework has limits, and they’re worth stating plainly. Where neurological disruption is severe, where development follows an atypical path, where later structures impose competing constraints of their own, the mechanism’s explanatory power is reduced — not absent, but no longer the primary thing doing the explaining. The piece is not claiming that this structural condition explains every instance of loneliness, depression, or failed relationship, or that it operates with equal force in every self. The claim is more specific than that: beneath the variable overlays of particular lives and particular relationships, this condition is present as the structural baseline of selfhood. And that baseline is what the variable overlays operate on top of — it’s not what they explain.

The loneliness itself is real. Its source sits prior to memory, prior to language, prior to the self that experiences it. Naming that source doesn’t relieve the loneliness. What naming it does do is dissolve one particular misattribution — the habit of blaming the loneliness on failed relationships, inadequate connection, the specific insufficiencies of the people and institutions that have populated a life. Those failures are real. They are simply not the cause of the deepest version of what’s being felt. That cause is structural, prior, and inaccessible to the self that carries it around.

The self is not what separated from the original condition. The self is the separation.

What was there before has no name. The self arrived too late to learn it.

Loneliness is not a wound that healing will close. It is the shape of what was whole before the self existed to know it.

References

• Bowlby, John, Attachment and Loss (Basic Books, 1969)
• Freud, Sigmund, Civilization and Its Discontents (1929)
• Gergely, György and Watson, John S., “The social biofeedback theory of parental affect-mirroring,” International Journal of Psycho-Analysis 77 (1996)
• Langer, Ellen J. and Rodin, Judith, “The effects of choice and enhanced personal responsibility for the aged,” Journal of Personality and Social Psychology 34 (1976)
• Mahler, Margaret S., Pine, Fred, and Bergman, Anni, The Psychological Birth of the Human Infant (Basic Books, 1975)
• Nelson, Charles A., Zeanah, Charles H., and Fox, Nathan A., Romania’s Abandoned Children (Harvard University Press, 2014)
• Rodin, Judith and Langer, Ellen J., “Long-term effects of a control-relevant intervention with the institutionalized aged,” Journal of Personality and Social Psychology 35 (1977)
• Seligman, Martin E.P. and Maier, Steven F., “Failure to escape traumatic shock,” Journal of Experimental Psychology 74 (1967)
• Spitz, René A., “Hospitalism,” Psychoanalytic Study of the Child 1 (1945); “Anaclitic Depression,” Psychoanalytic Study of the Child 2 (1946)
• Vitebsky, Piers, Dialogues with the Dead: The Discussion of Mortality Among the Sora of Eastern India (Cambridge University Press, 1993)
• Watson, John S., “The development and generalization of contingency awareness in early infancy,” Merrill-Palmer Quarterly 12 (1966); “Smiling, cooing, and ’the game,’” Merrill-Palmer Quarterly 18 (1972)
• Yalom, Irvin D., Existential Psychotherapy (Basic Books, 1980)

Copyright © 2026 Lloyd W. Taylor