Assumptions, and how to check them

Couples who know each other well develop models of each other that are efficient, mostly accurate, and progressively out of date.

The overconfidence finding

Research on how well close partners understand each other has produced a consistent and slightly deflating result.

People are more accurate about a close partner than about a stranger, and considerably less accurate than they believe.

Studies comparing predicted preferences against actual ones have found that the gap between confidence and accuracy is larger for long-term partners than for people who have known each other briefly.

The proposed explanation is that familiarity produces confidence faster than it produces information, and that after a certain point people stop checking.

Where the models go stale

Preferences change. Food, activities, how someone wants to spend a weekend, what they find stressful.

The version formed in year two is applied in year twelve, and it is applied confidently.

Circumstances change what someone needs. Someone who wanted company after a difficult day at twenty-five may want silence at forty, and will not necessarily announce it.

Positions soften. Views held firmly a decade ago are frequently held loosely now, and partners continue to avoid subjects on the basis of an old reading.

The other person has stopped saying. The most consequential version — where a preference was expressed, not accommodated, and eventually dropped.

Mind reading in conflict

The specific version that causes damage.

Statements attributing motive — you did that because, you always think that, I know what you're going to say — are common in couple arguments and they are corrosive for a structural reason.

They are unfalsifiable claims about another person's inner state, and they remove the other person's standing to describe their own experience.

The discipline is to describe the observable and ask about the rest. "You didn't reply and I assumed you were annoyed — were you?"

Which converts an accusation into a question, and the question can be answered.

Checking deliberately

A few practices that surface stale assumptions.

Ask questions with unknown answers. Discussed elsewhere on this site and it belongs here too. Conversation in long relationships defaults to exchanging events, which updates nothing.

Ask what they would do differently. About the week, the arrangement, the holiday, the division of tasks. Standing arrangements are rarely revisited and frequently no longer suit either person.

Notice when you are predicting. The thought "they'll say no" is a prediction, and predictions in long relationships are frequently used to avoid asking.

A useful exercise is to test a few of them: ask the thing you assumed the answer to. The rate of surprise is generally higher than expected.

Ask what you have wrong. Directly. "Is there anything I keep assuming about you that isn't true?"

An odd question that produces unusually specific answers.

The efficiency argument

The counterweight, because constant checking would be exhausting and irritating.

Models of each other are useful. A couple who had to verify every preference would not function, and the shorthand is one of the genuine benefits of a long relationship.

What is worth checking is the small set where the stakes are high or where the model was formed a long time ago — how the other person wants support, what they want from a holiday, what they find difficult, what they want more of.

Those are worth a conversation every few years, and almost nobody has it.

Being asked

The other side, and it requires something.

When a partner checks an assumption, the useful response is an accurate answer rather than the one that maintains the existing arrangement.

People frequently confirm an outdated preference because correcting it feels like a complaint about the years it went unaddressed.

It is not one. Saying that something changed is information, and it is a good deal cheaper to give than to have discovered later.

The shared calendar of preferences

One small practice that some couples adopt.

A short written note of the things each person currently wants more of and less of, revised occasionally.

It sounds excessive and it addresses the specific problem that these preferences are stated once, absorbed imperfectly, and then never updated.

Having them written removes the argument about whether something was ever mentioned, which is otherwise a recurring and unresolvable dispute.