Good Enough Was Never a Lower Standard

In the late 1980s Edward Tronick and Jeffrey Cohn videotaped fifty-four mothers and their babies and coded the tape second by second, looking for the moments when the two of them were doing the same thing at the same time. The coordination improved as the infants got older. It never got anywhere near most of the time. Matched states turned out to be the minority condition of ordinary, healthy interaction, and the researchers ended up recommending that these interactions be described by how they move between coordination and miscoordination rather than by how much coordination they contain.

Which means the ordinary case, the one nobody is worried about, is two people missing each other and finding each other again, repeatedly, all day.

That finding arrived roughly thirty years after Winnicott had already said it, and it is the part of him that got lost on the way into general use. The phrase escaped. It now sits in parenting books, in therapy rooms, in the self-talk of adults who have never read a word of him, and it means something close to: relax, you do not have to be perfect. Which is not wrong. It is just so much smaller than what he actually described that it functions as a different idea wearing the same words.

What Winnicott actually described

The good-enough mother, in his account, begins by adapting almost completely to the infant’s needs, and then adapts less and less over time. The reduction is not a lapse. The reduction is the mechanism. He tied its pace to one thing: the infant’s growing ability to deal with her failure.

Read that again with the emphasis where he put it. Not her adequacy. Her failure. The thing the caregiver is doing is failing, on a schedule, at a rate the infant can just about survive.

The reason has to do with what the infant starts out believing. In the earliest weeks, care that arrives the instant it is needed produces something Winnicott thought was necessary and could not be skipped: the infant experiences itself as the cause. It wants, and the world assembles. There is no gap between the need and the arrival, so there is no evidence of anything out there doing the arriving. He called this subjective omnipotence and he did not treat it as a delusion to be corrected early. He treated it as the ground floor.

But it has to come down, and the only thing that brings it down is the world declining to be instantaneous. The delay is the first evidence that something exists which is not you. Every small failure of adaptation, every time the milk is late or the face does not appear, is a data point in the slow discovery that reality has a location outside your wanting.

So the disillusionment is not the price of the arrangement. It is what the arrangement is for.

The documented cost of getting it too right

This is the part that stops being a nice reframe and starts being an actual claim about outcomes, because Winnicott argued that mothering which stays too well adapted past the earliest months does damage, and the observational literature has followed it.

Juliet Hopkins, writing in the Journal of Child Psychotherapy in 1996, traced the development of a baby whose mother was devoted and sensitive and continued to be almost perfectly adapted long past the point where that was useful. The infant did not flourish under it. She was deprived, in Hopkins’ account, of aspects of a sense of self, and of the chance to develop negotiation, concern and reparation. There was nothing to negotiate with. She developed a paranoid orientation, and later rejected her mother outright.

Winnicott’s own prediction was that too-good mothering, sustained too long, produced one of two things: a child who rejects the mother, or a child who stays merged with her in a state of arrested development. Hopkins observed the first. A later observation by Henrietta Otley, published in Infant Observation in 2020, followed the second pattern from the other side, a child whose perfectly attuned care left him unable to make the transition into a reality that would not behave that way, with the arrangement collapsing when he had to be handed to a childminder.

Neither of these is a story about a bad mother. That is what makes them useful. In both cases the care was excellent by every measure the culture uses. The problem was that it did not degrade.

The rate is the entire concept

The popular version does its real damage at this point, and it has nothing to do with lowering standards.

If the failure is the input, then the question is never whether you fail. It is at what rate, and whether what follows the failure is repair. Those two variables are what Winnicott was describing, and they are also what the empirical work has ended up measuring. Mitho Müller and colleagues, looking at how long mother-infant pairs took to get back into a matched state after a mismatch, found that the latency to repair was associated with the infant’s cortisol reactivity. The time it took to close them.

That puts a shape on the whole thing. The infant is not learning that the world is reliable. It is learning that the world is unreliable and comes back. Those produce different people.

And it clarifies what happens at the other end of the range. When the failure is not graduated but overwhelming, when the gesture that comes from inside is not slightly late in being met but is overridden and replaced, the infant does not get the disillusionment on a survivable schedule. It gets an environment that cannot be relied on to meet what comes from inside at all, and the response to that is compliance, and the organisation built on top of the compliance is the false self. Good-enough and the false self are not two separate concepts that happen to share an author. They are the two ends of one axis, and the axis is the rate at which the world fails you.

What the phrase turned into

The migration out of the clinic did something odd to it.

In its original setting, good-enough was a description of a moving process. A schedule. It had a direction, from almost total adaptation towards progressively less, calibrated in real time to a capacity that was itself changing. Nothing about it was a threshold.

By the time it reaches an adult telling themselves their work is good enough, or their parenting is good enough, or they are good enough, it has become a static bar. A line to clear. The permission-slip version, which absorbs the reassurance and drops the schedule.

The reassurance is real and it is doing something useful for people who need it, so this is not an argument for taking it away. But it is worth seeing what got left behind, because the abandoned half is the more interesting one. Winnicott was not saying the standard is lower than you think. He was saying that the failure does the work, that a relationship in which you were never let down would not produce a person capable of being in a relationship, and that the other person only becomes fully real to you at the point where they turn out not to be an extension of your wanting.

That is a claim about how reality gets installed, and it does not stop applying at eighteen months.

The part that does not resolve

The research is definite about the mechanism. Coordinated states are the minority in healthy interaction. Repair, not the absence of rupture, is what tracks the physiology. Care that stays too well adapted has documented costs and Winnicott named both of them before anyone had the tape to check. All of that is settled enough to build on.

What is not settled is what any adult is supposed to do with it.

Because the mechanism as described requires an environment doing something to you, at a rate you do not control, over a long period, without your consent or your understanding. It is not a technique and it was never available as one. The infant does not arrange to be failed well. It either is or it is not, and the shape that leaves is mostly set before there is anyone there to have opinions about it.

So the adult version of the question sits there without an obvious answer. If what makes another person real is their failure to be perfectly available, then the relationships that would do the most for you now are the ones where you are being let down at a survivable rate and are still there afterwards. Nobody can organise that on purpose. You cannot ask to be disappointed correctly.

The good-enough mother is not a permission slip. She is a description of someone failing at exactly the pace a person can grow into, and almost nobody knows they are doing it while it is happening.

FAQ

What does “good enough mother” actually mean?

Not a lower standard. Winnicott described a caregiver who begins by adapting almost completely to the infant and then adapts progressively less, at a pace set by the infant’s growing ability to cope with being failed. The reduction is the mechanism, not a shortfall in it. The failures are what teach the infant that something exists outside its own wanting.

Can a mother be too good?

Winnicott thought so, and predicted two outcomes when near-perfect adaptation continued past the earliest months: the child eventually rejects the mother, or stays merged with her in arrested development. Juliet Hopkins’ 1996 infant observation documented the first, in a child whose devoted and sensitive care left her without the chance to develop negotiation, concern or reparation.

Is it normal to feel out of sync with my baby?

It is the majority condition. Tronick and Cohn coded fifty-four mother-infant pairs second by second and found coordinated states occupy a small proportion of ordinary interaction, improving with age but never approaching most of the time. They concluded these interactions are better described by their movement between coordination and miscoordination than by how coordinated they are.

What matters more, avoiding ruptures or repairing them?

The repair is what the measurement tracks. Müller and colleagues found that the latency to repair after a mismatch was associated with infant cortisol reactivity. The infant is not learning that the world never fails. It is learning that it fails and comes back.