What holds up over forty years

Studies that have followed couples across decades, and interview research with people in very long marriages, converge on a description that is notably unglamorous.

What they describe

Liking each other. The item that appears most consistently in interview studies, and it is distinguished from love.

People in long, satisfying relationships describe their partner as someone they find interesting and enjoy the company of, which is not the same claim as loving them.

A tolerable set of unresolved differences. Not the absence of conflict but a settled relationship to the disagreements that never resolved.

Getting through something together. Illness, money trouble, bereavement, a difficult period with children.

Shared adversity survived appears repeatedly in these accounts as constitutive rather than incidental — the relationship is partly made of it.

Commitment as a decision rather than a feeling. Many long-married people describe periods of low satisfaction that they did not act on, and describe the not acting as a choice.

This is worth handling carefully. It is a description of what long marriages contain, not a recommendation to remain in unhappy ones.

Practical interdependence. Two people who have built something jointly and who each rely on the other for parts of it.

What does not appear

Notable by its absence.

Sustained passion is not generally described. Nor is the absence of serious difficulty, nor exceptional compatibility, nor communication of an unusual quality.

Most long-term couples describe periods when the relationship was poor, sometimes for years.

Which is worth knowing, because the assumption that a good relationship should not contain such periods causes people to interpret an ordinary phase as a verdict.

The satisfaction curve

Longitudinal research finds an average pattern: a decline over the first years and decades, with some evidence of improvement later in life.

The later improvement is real in several datasets and has an obvious confound — the unhappiest couples have separated, leaving a selected sample.

Studies of older couples also find lower reported conflict and greater expressed affection during observed interactions, which is consistent with something other than pure selection.

The reasonable position: the decline is real, the recovery is partly selection and partly genuine, and the trajectory is not fixed for any individual couple.

The behaviours that recur

Across the research, a short list.

Everyday warmth maintained at a higher volume than conflict.

Complaints made about behaviour rather than character.

The absence of contempt, which is the item with the strongest evidence.

Attention paid to small approaches for connection.

Interest in what the other person is currently thinking about, maintained rather than assumed.

Repair after conflict, accepted as well as offered.

None of it is novel and all of it is difficult to sustain for decades, which is the actual problem.

The honest limitations

This literature is disproportionately from wealthy Western countries, disproportionately heterosexual, and disproportionately drawn from couples willing to be studied.

It also suffers from survivorship: interviews with people in long marriages tell you about long marriages, not about which relationships should have been long.

And it is largely correlational, so the behaviours associated with lasting relationships may be effects rather than causes.

What follows

Less than the volume of advice suggests, and not nothing.

Keep the ordinary part good, because the ordinary part is most of it. Address problems while they are small. Do not express contempt. Stay interested. Expect difficult periods and do not treat them as conclusions.

The absence of anything more sophisticated is the finding rather than a gap in it.

The role of luck

Worth stating, since accounts of long relationships tend to attribute everything to conduct.

Health, money, employment, the behaviour of children and the absence of catastrophe all affect whether a relationship survives, and none of them is fully within anyone's control.

Couples who came through forty years generally had some of this go their way, and describing the outcome purely as the product of good behaviour is not accurate.

It also imposes an unfair reading on relationships that ended under pressures the surviving ones never faced.