This is the shortest chapter in the guide, and the only one that isn’t built directly on research.
I want to say that up front. Everything else here traces to a study, a sample size, a finding you can check. This chapter comes from a pattern that shows up in enough conversations, from enough men and enough women, that it seemed wrong to leave out just because nobody has run the trial yet.
The pattern is this: one person wants the other held to a fixed set of obligations, while treating their own obligations as conditional.
It doesn’t matter what the obligations are. Traditional, modern, some mix. What matters is whether the terms run in both directions — and whether they still run in both directions when circumstances change.
What it looks like
You’ve seen it. Maybe you’ve done it.
A woman wants a man who provides, protects, leads — and also wants complete freedom in how she spends her time, her money, and her attention. She wants the traditional man. She doesn’t want to be the traditional woman.
A man wants a woman who keeps a home, defers to him, and holds herself to a standard — while treating his own provider role as something he’ll get to once things line up. He wants the traditional woman. He doesn’t want the obligations that came with being the traditional man.
Both of them believe they want a traditional relationship. Neither one does. They want the other person’s half of it.
And notice: this isn’t a values mismatch. Two people can share every value on the list in Chapter Four and still be asking for asymmetric terms. You can both want a traditional home and disagree entirely about who’s bound by what.
What the research does say
There’s no study of reciprocity as such. But three findings sit right next to it, and they tell you this isn’t a small thing.
Employment. In national data, when men were not employed, both husbands and wives were more likely to leave the marriage. When wives were employed, it destabilized the marriage only if she was already unhappy — her job functioned as an exit door from a bad marriage, not as a risk in a good one. The authors’ conclusion was that divorce has to be understood as gendered: the terms aren’t symmetric in how they operate, whatever anyone intends.
Earnings. Using US administrative data, economists found that couples where the wife earns more than the husband report lower satisfaction and face higher divorce risk. The mechanism wasn’t the money. It was gender norms about who’s supposed to earn — the arrangement violated an expectation somebody was still holding, even if nobody had said it out loud.
Labor. Among couples going through the transition to parenthood, a more equal division of labor was one of the few things that buffered the satisfaction drop.
Read together: unspoken expectations about who owes what are a real fault line, they’re gendered, and they surface under exactly the conditions you’re least prepared for — a lost job, a pay change, a baby.
Which is why you find out now.
The rule that makes this a tool instead of an argument
Before anything else: you answer every question in this chapter yourself, in writing, before you ask anyone else.
Not as a formality. As the mechanism.
A reciprocity test that only runs one direction isn’t a test. It’s a list of grievances with a research citation attached. The version of this chapter that helps you is the one that catches your asymmetric asks with the same precision it catches hers.
Chapter Three covered why you need this: you are a poor judge of your own observable behavior and a decent judge of your own internal states. Your expectations are internal. You can find them if you write them down. What you can’t do is see them the way she’ll see them, which is why they go on paper first.
The self-audit
Three questions. Written. Before any conversation.
1. What am I asking for that I’m not offering?
Be specific. Not “loyalty” — what does loyalty require of her, concretely, and what does the equivalent require of you? If you want her to defer on major decisions, what do you defer on? If you want her home life to look a certain way, what does your home life look like?
2. What would I want if the person I’m asking it of asked it of me?
Flip every expectation. If she required of you what you’re requiring of her, would you sign?
3. Which of my expectations are about the relationship working, and which are about how I want to be seen?
This is the one that stings. Some of what you’re asking for is load-bearing. Some of it is about the picture — how it looks to your friends, your family, yourself. The second kind is the kind that gets defended hardest and delivers least.
The arrangement, as lived
Now the questions for her — and the same ones for you, out loud, in the same conversation.
In your last serious relationship, who did what? Walk me through an ordinary week.
Not what you wanted. What happened. Who cooked, who paid, who called whose mother, who made the plans, who handled the crisis.
What did you expect from them that you didn’t get?
Everyone can answer this instantly.
What did they expect from you that you didn’t give?
Almost nobody can answer this at all. Sit with the silence if it comes. A person who cannot name a single thing a partner wanted from them and didn’t get is showing you the edge of their self-awareness — and that edge is exactly where asymmetric expectations live.
The conditionality test
This is the core of the chapter. Asymmetry hides in stable conditions and shows up when things move — because that’s when you learn whose obligations were fixed and whose were contingent.
Describe an arrangement between a couple you know that you’d want. Now describe what each of them does in it.
The first half gets you the picture. The second half gets you whether the picture has two people in it or one.
If we agreed on how things would work, and then your income doubled — what changes?
If mine doubled — what changes?
Compare the two answers. Directly.
If her obligations flex when her circumstances improve but yours stay fixed regardless, that’s the asymmetry. If yours flex when your circumstances improve but hers don’t, that’s the asymmetry running the other way — and you’re the one carrying it.
If one of us couldn’t work for a year, what happens?
If we had a child and one of us had to step back from work, how do we decide which one?
The last question is the sharpest instrument in this chapter. Most people have an answer they’ve never said out loud. Some of those answers are a default — the assumption that it’s obviously her, or obviously him — held so deeply it doesn’t register as a decision. The fact that it’s unspoken is the problem. Two people can walk into parenthood each certain the other one is stepping back.
Reading the answers
What you’re looking for is not a particular arrangement. Traditional is fine. Modern is fine. Anything you both actually agree to is fine.
What you’re looking for is whether the terms are symmetric in their conditionality. Do both people’s obligations hold when circumstances change, or do only one person’s?
Three patterns to name:
Symmetric. Both of you can describe what you owe and what you’re owed, both sets of obligations move together when conditions change, and the parenthood question has an actual answer rather than an assumption. This is what it looks like when it’s working.
One-sided. One person’s obligations are fixed and the other’s are contingent on circumstance. This is a structural mismatch. It will not surface until conditions change, and then it will surface all at once.
Unspoken. Neither of you had answers until you were asked, and the answers you produced don’t match. This is the common case, and it’s fixable — it’s the reason to have the conversation at all.
Why this chapter is here
You could argue this belongs in a book about marriage rather than one about choosing.
It’s here because it’s a selection criterion. A person who is asking you for asymmetric terms is telling you something about the relationship before it starts. And a person who catches themselves asking for asymmetric terms — who reads the self-audit and finds something they didn’t want to find — is doing the only kind of self-work Chapter Three said actually transfers: changing what they’ll do in a specific situation with a specific person.
One more time, because it’s the thing that determines whether this chapter helps or harms: it runs both ways or it doesn’t run. The moment you use it only to audit her, it stops being a tool and becomes the thing it was supposed to detect.
Next: Chapter Six — Told vs. Observed. Why you can’t read people, and the system you use instead.