Quick answer
After someone goes quiet, the next conversation isn't a fresh startup script. A 1990 study of 31 couples found that people press harder for a change they want and pull back when the change is their partner's.
In 1990, Andrew Christensen and Christopher L. Heavey reported on 31 couples and two conflict situations: one in which the husband wanted a change in the wife, and one in which the wife wanted a change in the husband. This piece starts after that kind of talk has already gone bad. Someone pushed. Someone went quiet. You aren't looking for a startup script for a brand-new argument. You're looking at the next conversation, the one that has to live with the shutdown that just happened.
What the two roles mean
In that paper, demand means pressure through emotional requests, criticism, and complaints. Withdraw means retreat through defensiveness and passive inaction. Those definitions matter once the room has gone cold, because the next hour often repeats them in a lower voice. The person who keeps listing hurts is still in demand. The person who answers with a shrug, a joke that ends the topic, or a long silence is still in withdraw. Passive inaction counts. You don't have to slam a door to retreat.
Wife-demand/husband-withdraw was more likely than the reverse. That gender difference showed up especially when the topic was a change the wife wanted. Read the second sentence before you turn the first into a slogan. The pattern wasn't a flat claim that women always pursue and men always flee, in every topic, at every hour. It was stronger when her change was the subject.
Separate analyses in the same paper cut across that gender result. Both partners were more demanding when they discussed a change they wanted. Both were more withdrawing when they discussed a change their partner wanted. Men were overall more withdrawn than women. Women were not overall more demanding than men. If you only remember the famous half of the finding, you'll mishear the fight you just had.
You might be thinking the gender result means the person who wants the change should stop asking. The paper doesn't say that. A change can be fair and still trigger demand from the person who wants it and withdraw from the person who's being asked. Wanting a change isn't a character flaw. Going quiet isn't proof that the change was unreasonable. The finding is about the structure of the moment: whose request is on the table.
After the shutdown
Picture the kitchen twenty minutes later. The words have stopped. One of you is rinsing a cup that's already clean. The other is rehearsing a better speech that will probably come out as the same speech. Before you reopen anything, name whose change was under discussion. If it was yours, the study suggests you were more likely to be the demanding one. If it was theirs, you were more likely to withdraw. Both can be true across a month, in different talks. In this talk, one direction was active.
That naming is a way to use the finding. It isn't a line the authors handed couples, and it isn't a clinical protocol. It does give you a question that's more specific than "why are you like this?" You can say, as a suggestion for your own mouth and not as a quotation from the study, "The change was mine, and I piled on complaints. I want another try that doesn't do that." Or, "The change was yours, and I left the room without leaving the house. I'm willing to come back if we can start from the request instead of from my defense."
Watch the risk on the second try. If you open with a longer list of complaints, you may be repeating demand. If you agree to talk and then offer only yes-no answers, you may be extending withdraw. Defensiveness, in the paper's definition, is part of retreat, not a separate hobby. Criticism and complaints are part of pressure. You can care about the issue and still choose a sentence that's a request rather than a prosecution.
The next conversation is also a bad time to pretend the last one didn't happen. A startup script imagines a clean beginning. Aftermath doesn't give you one. Your partner heard the shutdown, or did the shutting down. If you skip that fact, the new talk inherits the old ending. You can mark it in a single sentence and then move to the request that started the conflict. You don't need a speech about your childhood to do that. The 1990 paper is about the structure of demand and withdraw during conflict, not a full biography of either spouse.
When the issue will come back anyway
On the Gottman.com conflict page, last updated May 13, 2026, perpetual problems account for 69% of relationship conflict. Every couple has some. They grow out of personality, or out of the kind of life each person needs, and you meet them again. In-laws, sex, how children are disciplined, and housecleaning are the page's examples of topics one pair can settle and another pair will keep reopening. The fight you just had may be one of those returning topics. If it is, the next conversation isn't a final hearing. It's another pass.
The page says the aim with a perpetual problem isn't to solve it once and for all. The aim is dialogue that communicates acceptance, and that dialogue can include humor, affection, and even amusement. That's a different job from winning. Acceptance doesn't mean you drop the change you wanted. It means the talk can hold the difference without treating your partner as an obstacle to be worn down. Humor belongs if it keeps you on the same side of the problem. It doesn't belong if it punishes the person who went quiet.
If you can't establish that dialogue, the conflict becomes gridlocked. The page's picture of gridlock is wheels that spin: effort without travel. Hidden agendas underlie the issue. Discussions in that state lead to painful exchanges or icy silence, and they almost always involve criticism, contempt, stonewalling, and defensiveness. Shutdown sits next to stonewalling for a reason. Icy silence is one of the endings the page warns about. If your "next conversation" is only a quieter silence, you have not left gridlock. You have furnished it.
This essay won't give you seven techniques. A list like that pretends the hour after a shutdown is a class you can ace with steps. The research cited here is narrower. One study shows that people press harder for the change they want and pull back when the change is the partner's. A conflict page says you can keep talking about a problem that returns, with more warmth than a gridlocked fight allows, and that warmth is the point when a solution won't stay put. You can use those two ideas in the next talk. You can't honestly stretch them into a guarantee that the talk will end in agreement.
If you're deciding whether to speak tonight or tomorrow morning, pick the moment when both of you can stay in the room. Then ask one question before you restate the charge.
Whose change was on the table when the talking stopped, and can you name it without adding a new complaint?
Studies named above
- Andrew Christensen and Christopher L. Heavey, "Gender and social structure in the demand/withdraw pattern of marital conflict," Journal of Personality and Social Psychology 59, 73–81, 1990.
- Gottman.com, "Managing Conflict: Solvable vs. Perpetual Problems," page last updated May 13, 2026. https://www.gottman.com/blog/managing-conflict-solvable-vs-perpetual-problems/
Frequently Asked Questions
- Does this piece teach a fresh opening for a fight that hasn't started?
- No. It assumes the clash already ended in retreat, and it won't hand you a seven-part class.
- Do the findings say women are always the ones applying pressure?
- No. In that sample the wife-pursues pattern was more common, especially for a change she wanted, but women as a group weren't the heavier demanders, and men were more withdrawn overall.
- Who is more likely to go quiet?
- Both partners pulled back more when the requested change belonged to the other person. Knowing whose request was on the table tells you more than a gender slogan.
- Can humor and warmth belong in a problem that will not stay solved?
- Yes, on the conflict page. The point of that dialogue is acceptance, including amusement, rather than wheels that spin without travel.
- What should I name before restating the complaint?
- Name which person wanted the change when the room went quiet. Then try a request instead of another stack of criticisms.