Every chapter so far has been about choosing. This one is about what you’re choosing.
Selection research quietly assumes people know what they’re signing up for. Mostly they don’t. Ask a man what marriage requires and he’ll describe a feeling — loving her, wanting to be there, knowing she’s the one. Almost nobody describes the actual demands, because almost nobody has been told what they are.
This chapter tells you. Some of it is hard reading. It belongs at the end of a guide about selection because the thing you’re selecting for should be visible before you select.
The line goes down
Start with the finding that reframes everything else.
You’ve probably heard marriage follows a U-curve: happy at the start, dips through the child-raising years, recovers once the kids are gone. It’s a comforting shape. Get through the hard middle and you come out the other side.
It isn’t real.
When researchers followed the same couples across five waves over seventeen years — rather than comparing different couples at different stages — the U disappeared. What they found was decline at all durations and no late-life recovery.
The U in the older studies was survivor bias. Unhappy marriages end. Survey couples married thirty years and you’re surveying the ones that lasted, which makes the thirty-year group look happier than the twenty-year group. The miserable twenty-year couples are still in your sample. The miserable thirty-year couples are divorced.
So the honest baseline: average marital happiness declines, gradually, and does not bend back up on its own.
That’s not a reason to avoid marriage. It’s a reason to stop expecting the relationship to sustain itself. The couples who do well aren’t riding a natural arc back up. They’re doing something.
The predictable shocks
Four things are close to certain over a long marriage. None of them is a surprise, and all of them surprise people anyway.
A child
The transition to parenthood is the best-documented hit to relationship quality in the literature.
A meta-analysis found parents report lower satisfaction than non-parents, with the effect most pronounced for mothers of infants — 38% of mothers of infants reported high marital satisfaction, versus 62% of childless women. Satisfaction declines further with each additional child. Interestingly, the effect is worse for higher-income couples and for more recent generations.
An eight-year prospective study of 218 couples found something sharper. Couples who had a baby showed a sudden deterioration in relationship functioning right after the birth — visible both in what they reported and in what observers rated — and it persisted for years. Childless couples in the same study declined too, but gradually, without the drop. The birth itself does something.
Mothers and fathers changed by similar amounts.
What buffers it: relationship quality going in, a planned and wanted pregnancy, and a more equal division of labor. That last one connects directly to Chapter Five — the couples who had the conversation about who steps back are the ones who handle it better than the couples who each assumed it was the other.
Money
Job loss is the shock, and the research on it is unusually precise about why.
Analyzing decades of national panel data, economists found divorce risk rises after a spouse’s layoff — but not after a plant closing, even though the financial hit is comparable. And not after disability, despite similar long-run income loss.
The difference is information. A layoff says something about the worker. A plant closing says something about the plant. The marriage destabilizes when the loss reads as a verdict on the person.
The effect is concentrated in the first one to three years. And here’s the part worth holding onto: for job losses more than five years back, layoffs were associated with lower divorce — the authors’ interpretation being marriages that survived a crisis and came out stronger.
There’s a gendered pattern too. When men are not employed, both spouses become more likely to leave. When women are employed, it destabilizes the marriage only if she was already unhappy — the job is an exit resource from a bad marriage, not a risk in a good one.
Which is a warning for any man whose entire sense of his own worth is loaded onto his earning. Chapter Three flagged this: the provider-status framing sets you up badly for the one shock most likely to arrive.
Illness
This is where a famous claim needs correcting, because you’ve almost certainly absorbed it.
In 2015, headlines went around the world reporting that divorce risk rose when wives got sick but not when husbands did. Men leave sick wives.
The study was retracted. Researchers at another university couldn’t replicate it. The original author found the error — one line of code had counted marriages lost to attrition as divorces. The reported 32% divorce rate was actually about 6%.
The corrected analysis found no general gender effect. One narrow exception survived, for heart problems specifically. That’s it.
Most marriages survive serious illness. What predicts staying is the relationship’s quality beforehand, the practical capacity to provide care, and resources. Caregiving is genuinely hard on the caregiver’s own health — that literature is large and real. But the story about men walking out on sick wives came from a coding error, and it should stop being repeated.
Sex changes
Over decades: aging, illness, menopause, erectile dysfunction, medication, exhaustion, and the ordinary erosion of novelty.
This is normative, not pathological. What predicts long-term sexual satisfaction isn’t preserved frequency — it’s the same thing Chapter Four flagged. Couples who believe sexual compatibility is built handle the changes; couples who believe it’s found read every change as evidence the relationship was wrong.
That belief is measurable, it’s assessable before you commit, and it predicts better than anything about your current sex life does.
What actually holds it together
Given a declining baseline and four predictable shocks, what keeps a marriage?
The research points to three things, and the first one is not a feeling.
Dedication, which is a decision
Relationship researchers distinguish two kinds of commitment.
Dedication is wanting to be with this person: a couple identity, a long view, prioritizing the relationship, willingness to sacrifice for it. It’s a stance you take.
Constraint is everything that makes leaving costly: shared money, a house, social pressure, the sheer difficulty of unwinding a life. Chapter Eight covered how these accumulate.
Dedication is what predicts a marriage that’s actually good. Constraint is what holds a couple together through the low stretches — long enough for dedication to recover, in a marriage worth recovering.
That’s a genuine benefit and a genuine danger, depending on which marriage you’re in. Constraint in a good marriage is ballast. Constraint without dedication is a cage, and the research is clear-eyed about that.
The practical form: commitment is a decision that becomes a structure that supports the feeling through the years when the feeling is thin. Not the other way around. A marriage running on feeling alone has nothing to draw on when the line goes down — and the line goes down.
The belief that it’s built
The sexual growth-versus-destiny finding generalizes.
People who believe a relationship is something you build handle conflict constructively, report higher satisfaction, and stay more committed. People who believe it’s something you find make fast compatibility judgments, invest heavily once they’ve decided, and become brittle when normal difficulty arrives — because to them, difficulty means they got it wrong.
There’s a darker edge to destiny belief the research names directly: it can produce resistance to all threats to the relationship, including conflict, including abuse, including infidelity. “Meant to be” is load-bearing in a way that keeps people in places they should leave.
Growth belief has its own failure mode — it can keep someone working on a relationship that can’t be fixed. Which is why the next section exists.
Repair, and the limit of forgiveness
Repair is the skill. Not avoiding conflict — coming back afterward, closing it, and having it stick.
But forgiveness has a documented dark side that almost nobody hears about, and a father reading this should hear it.
In a diary study of newlyweds, a spouse whose bad behavior was forgiven was nearly twice as likely to repeat it the next day. Over four years, more-forgiving spouses saw psychological and physical aggression stay flat, while less-forgiving spouses saw it decline.
A follow-up found what determines which way it goes: the offending partner’s agreeableness. Forgiveness reduced repeat offenses among agreeable partners, who felt obligated not to exploit it. It increased them among disagreeable partners, who learned they could.
So forgiveness is not unconditionally good. Forgiveness paired with accountability and changed behavior is repair. Forgiveness that substitutes for accountability teaches the other person what they can do to you.
The floor
Everything above — dedication, growth belief, repair, forgiveness, constraint as ballast — is advice for staying. All of it applies above a floor.
Below the floor, the evidence says something different, and this guide is going to say it plainly.
Where there is aggression, coercive control, or a partner who exploits forgiveness rather than responding to it, the “work harder” stance is not virtuous. It’s the mechanism by which people stay too long. The forgiveness research says it directly: with a partner who takes advantage, forgiving more produces more of what you’re forgiving.
Chapter Six listed the control screen. It’s exempt from idealization and it’s exempt from this chapter too. Growth beliefs, dedication, and repair are for relationships that are fundamentally sound and difficult. They are not for relationships that are unsafe.
Knowing the difference is the point of the whole system.
For remarrying fathers: the legal floor
The rest of this chapter is about what a marriage demands emotionally. This part is about what it demands administratively, and it’s the part men with children from a prior relationship most often get wrong.
This is general information, not legal advice, and state variation is substantial. Every item here needs a lawyer in your state.
The elective share can override your will. Nearly every state guarantees a surviving spouse a portion of the estate — commonly about a third — regardless of what your will says. So leaving everything to your child can be defeated by your new spouse’s claim. And leaving everything to your new spouse can disinherit your child. Both of the obvious moves fail.
The standard tool is a QTIP trust: your spouse receives income for life, and the remainder passes to your children. She can’t redirect it. But the elective share can still be claimed against it — unless it’s waived in a prenuptial or postnuptial agreement. That waiver is the only complete fix.
Prenups fail for predictable reasons. They’re recognized in every state, and what invalidates them is consistent: inadequate financial disclosure, signing under time pressure right before the wedding, no independent counsel for both parties, grossly one-sided terms. What makes one survive: negotiate early, disclose completely with documentation attached, each of you with your own attorney, adequate time. Some states add specific requirements — California, for example, requires independent counsel or a separate written waiver plus a seven-day review period.
Child support can’t be waived in a prenup. That’s public policy everywhere.
Beneficiary designations don’t follow your will, and divorce doesn’t reset them. Retirement accounts and life insurance pass by designation, outside the will entirely. About 26 states have statutes automatically removing an ex-spouse — but those statutes do not apply to employer-provided ERISA policies, where the named beneficiary controls no matter what. There’s case law where an ex-wife lost a claim because an ERISA policy named someone else, and case law running the other direction. Update every designation yourself.
Remarriage doesn’t change your child support. A new spouse’s income is not counted in the calculation. Indirect effects exist — reduced living expenses could free disposable income a court considers, and new children may create a hardship deduction in some states — but remarriage alone isn’t grounds for modification.
Separate property becomes marital property quietly. Depositing an inheritance into a joint account, using premarital savings for the marital home, retitling your house into both names — any of these can convert what was yours into what’s both of yours. Protection is boring: separate titling, no commingling into joint accounts, records that let you trace it, and language in the prenup.
Check your decree. Many divorce decrees require the paying parent to maintain life insurance naming the ex-spouse or children. That’s enforceable, and forgetting it after remarriage causes real problems.
On raising it: practitioners are unanimous that the prenup conversation happens early — both for enforceability and for the relationship. There’s no rigorous research on how it affects a couple, but the practitioner consensus is consistent: frame it as mutual protection and shared financial planning, lead with your own reasons, both people with their own counsel, financial values discussed before specific numbers.
And notice this is the same disclosure exchange from Chapter Seven. Doing it once serves both purposes.
What you’re actually signing
Here’s the whole thing, said plainly.
You’re signing up for a relationship whose average satisfaction declines and doesn’t naturally recover. That will take a sharp, lasting hit if you have a child. That will face a money shock, an illness, and decades of sexual change. Where forgiveness can teach the wrong lesson, where your own commitment will have to be a decision rather than a feeling for stretches of it, and where the thing that carries you is a stance you take repeatedly rather than a person you found.
That’s the real offer. Not the one in the vows.
And people take it anyway, and a lot of them are glad they did — because a marriage where both people know what it demands and choose it anyway is different from one where two people expected it to carry itself.
The point of this guide was never to talk you into it or out of it. It was to make sure that if you say yes, you’re saying yes to the actual thing.
Next: the instrument. Everything in this guide, in a form you can actually run.