Letting Go Was Never the Goal

Freud's essay Mourning and Melancholia gave twentieth-century grief theory its central image: mourning as labour. The bereaved person has libidinal attachment invested in somebody who no longer exists, and the work of grief is to withdraw it, piece by piece, until it can be invested elsewhere. Successful mourning ends in detachment.

Call this grief work. It dominated clinical practice for most of a century, and it defined a healthy outcome as letting go.

Then, in the 1990s, a group of researchers published work on what bereaved people actually do, and the answer was not that.

They talk to the dead person. They ask what he would have thought. They keep objects, not as relics but in use. They notice themselves acquiring his opinions. They tell the children what she would have said. They continue, in a changed form, to have a relationship with somebody who has died.

This was given the name continuing bonds, and the crucial finding was not that it happens, which anybody bereaved could have reported. The finding was that it is normal: common, cross-cultural, and compatible with functioning well. It is not a stage people pass through on the way to detachment, and people who do it are not stuck.

One observation that made the old model look local rather than universal is Japanese practice around the butsudan, the household Buddhist altar. The dead are addressed daily, offered food and tea, told news, included. This is ordinary, expected and socially supported. Under the detachment model it is unresolved grief on a national scale, which should have been a clue that the model had generalised from one culture's expectations about emotional independence.

The clinical consequence of getting this wrong was not neutral.

A model that classifies a coping strategy as a symptom does active harm. People were encouraged to put things away, to stop the conversations, to break the habit of consulting somebody who was gone, on the grounds that these were failures to accept the death. For many of them, those were the practices holding things together.

Be precise about what the newer view claims, because it is easy to overstate. It does not say grief is static. The relationship persists and it changes: the person at ten years is not having the conversation they were having at ten days, and the frequency, the intensity and the content all shift. It also does not say that every form of continued attachment is healthy. Some are not, and distinguishing them is clinical work rather than a rule.

Finally, stand back and look at this book as a whole, because the four lessons are one lesson.

Predicted breakup distress: a model built on a couple of dozen students. The divorce rate: an arithmetic error between two populations. The five stages: observations of dying patients applied to the bereaved. Grief work: a theoretical construct from one essay, generalised to every culture.

In each case a confident, memorable claim was built on a small or unrepresentative base, spread because it was useful to repeat, and then used to tell people how they ought to be feeling. And in each case the corrective is the same three questions: who was actually observed, how many of them were there, and was this ever tested on anybody else.