What Actually Predicts Lasting

The research on what makes relationships last is considerably less exciting than the research on falling in love, and considerably more useful.

Several findings replicate reasonably well.

Contempt is the most damaging thing that happens in a relationship, more damaging than anger, more damaging than frequent disagreement. The reason is structural. Anger is about an act and can be answered by addressing the act. Contempt communicates that the person is beneath consideration, and there is no repair aimed at that, because the message is not about anything that could be fixed.

How a disagreement opens predicts a surprising amount about how it ends. The first few minutes carry much of the outcome, which is one of the few findings in this entire course with a direct practical implication that a person can act on tonight.

Small bids for attention matter. A remark about something outside the window, a comment on an article, a hand on a shoulder: minor attempts at connection, easy to miss and easy to ignore without any cruelty being involved. Whether they are noticed and answered accumulates, and the accumulation is eventually what the relationship is.

And the most counterintuitive one. How a partner responds to good news predicts more than how they respond to bad news.

That looks wrong until you see why. Bad news summons support from almost anybody, including people who do not much like you; there is a social script and most people can perform it. Responding well to somebody's success is different. It requires genuinely wanting their life to go well independently of you, including in ways that might take them further from you, and it has no script. Enthusiasm that is merely polite is detected immediately.

Now the caveat, which is not optional.

Effect sizes in this literature are modest. These are averages over groups, and prediction at the level of an individual couple is poor. A great deal of confident advice has been sold on the back of these studies, including precise claims about predicting divorce, and the confidence is not supported by the underlying work. Take the findings as things worth attending to, not as a diagnostic instrument.

Which closes the course, because five books have been arguing with each other and the argument resolves here.

The Greeks said love has parts, and Sternberg measured three of them and found they run at different speeds. The troubadours and García Márquez built a love out of longing alone, which is passion with nothing attached, and made it last fifty-one years by never letting it arrive. Austen and Aristotle both insisted that loving somebody requires perceiving them accurately, which is intimacy, and which is work. Jeong describes commitment accumulating quietly without anybody deciding anything. The animals suggest the machinery is old, shared, and not ours alone.

And the last finding is that the part you can actually influence is the least glamorous part of it. Not the longing, which arrives unbidden and leaves the same way. Not the chemistry, which is doing what it has done for a very long time. What responds to effort is how you open an argument at eleven at night, and whether you look up when somebody tells you something good happened.

Nobody has ever written a poem about that, which is the most interesting fact in this course.