Why Long-Term Relationships End

— It's the opposite. Fewer of the screamers end up there. Roughly half of divorces come out of marriages that, at the time they were surveyed, showed neither pronounced conflict nor pronounced unhappiness. A flat, even background — and then the filing.

— There's a paper by Amato and Hohmann-Marriott, 2007. They took 509 couples who divorced between two waves of a large American family survey and ran them through a cluster analysis. Two roughly equal piles came out: one of high-distress marriages, one of low-distress. And they checked separately whether the second pile was an artifact — maybe people were just caught on a good day, maybe the instrument was too blunt. No. Not an artifact.

— I'm saying that among the divorced, the two groups are about equal in size. That's not the same claim, but for our conversation it's the more important one: if you're standing in the graveyard of marriages, half the graves are quiet ones.

— That's the interesting part. Nothing was "wrong" with them in the sense you're used to. Picture a house that never caught fire, but where for 20 years running nobody went up to the attic. A fire is an event: the siren, the neighbours, the insurance company, a date on the calendar. A beam quietly going damp doesn't have a single day attached to it.

— Exactly. And everyone comes running to it, the spouses included. Whereas two people slowly turning into flatmates sends up no signal at all — until one day the floor sags. And both of them honestly cannot name the day it started.

— Here's the hardest finding in that paper. People from high-distress marriages reported an increase in happiness after divorcing. People from the quiet ones reported a decrease.

— On average, yes. The authors suggested that in the second case people divorce not out of pain but out of weak commitment to the idea of staying together at all. So there are two different motives: one person is in pain, the other simply no longer believes they're obliged.

— Booth and Amato, 2001, followed the grown children. When a high-conflict marriage broke up, the children did better afterwards — the quality of their own close relationships, support from friends and family, psychological well-being. When a quiet one broke up, they did worse.

— Because the child of a quiet marriage had no reason available that they could understand. Nobody shouted at home, nobody threw dishes, and then suddenly it was over. What are they supposed to conclude from that about how else it could have gone?

— I didn't enjoy it either the first time I read it. It doesn't get lighter. There's a survey of 886 divorcing parents in Hennepin County, Minnesota, 2008. The two most commonly named reasons: "we grew apart" — 55%, and "we were not able to talk together" — 53%.

— Absolutely. And it's not because they're hiding something from the questionnaire. They genuinely have nothing to name. You're asking a person to point a finger at an event, and there were no events.

— And here I'm obliged to hit the brakes. 70% is not a measured share of all divorces. It's a retelling of Amato and Booth's estimate from their own panel: fewer than a third of parental divorces come from marriages with strongly pronounced conflict. The limitation — families with children only. And the direct 2007 measurement gives about half, not 70%. That part was measured — this part got filled in along the way to the reader.

— The whole conversation. This subject is stuffed with beautiful numbers, and most of them grew or rounded in a convenient direction on the way to you. If I don't say every single time where the measured part ends, you'll walk out of here with a confidence I don't have myself.

— Doherty, Willoughby and Peterson, 2011, surveyed 2,484 divorcing parents. About one in four thought the marriage could still be saved. And among the 329 couples where both spouses could be matched up, both of them thought so in roughly one couple out of nine.

— They filled in the questionnaires separately, each under their own number. One writes "I'd be willing to try," the other writes "no," and neither of them is aware. That's probably the loneliest number in the whole field.

— From the inside — almost none. And that's the most unpleasant thing I'll tell you today. Statistics see the difference at large numbers and in hindsight. For one specific couple on one specific Thursday, the same flat, quiet evening can mean either one.

— I do. And I'll tell you in advance: watching for signs will get you nothing. We'll come back to this, more than once.