You raise something that has been bothering you. He goes quiet. Not angry quiet, not thinking quiet. Absent. His eyes go to the floor or the phone, his answers shrink to three words, and the more you push for a response, the further away he seems to get. Eventually one of you leaves the room, and the thing you wanted to talk about is still sitting there untouched.

Then it happens again three weeks later, with a different subject and the same choreography.

If that scene is familiar, you are not imagining it. Men withdraw during conflict far more often than women do, and researchers have been documenting it since the 1930s. It has a name, a substantial evidence base, and an explanation considerably more interesting than the one you have probably read.

The short version

• Dropping out of a fight is a real, measurable, usually masculine act. This is one of the most frequently replicated findings in relationship research.
• It is also one of the most damaging habits that a couple can get into, and it is more correlated with unhappiness than defensiveness or criticism.
• Research does not support the popular belief that females are wired to pursue, and males are wired to shut down.
• The same pattern is true for gay and lesbian couples and is just as damaging when the woman pulls away.
• Withdrawal is not predicted by gender. It’s someone who wants to change.• It is flexible, since it is a position rather than a personality trait. Further down this page some things are helpful.

When we’re stonewalling, we’re physiologically flooded

Dr. John Gottman |
The Seven Principles for Making Marriage Work

What the pattern actually looks like

Researchers call it the demand-withdraw pattern. It runs in a loop, and the loop has four moves.

One partner raises an issue and presses for discussion. The other avoids the topic, deflects, or gives minimal answers. The first partner reads the retreat as not caring and pushes harder, often louder and with sharper language. The second experiences the escalation as an attack and retreats further, sometimes into complete silence.

The important thing about that sequence is that every move makes perfect sense from the inside and looks unreasonable from the outside.

She pushes harder because the alternative is the subject disappearing again, and it has disappeared many times before. From her side, persistence is the only thing that has ever produced a response at all.

He retreats further because the conversation has stopped feeling like a conversation and started feeling like a verdict.

He is not trying to end the relationship. He is trying to end the next ninety seconds.

Neither person is behaving irrationally. The pattern is what is irrational, and it survives precisely because each person’s reasonable response is the exact thing that triggers the other’s.

Lewis Terman and colleagues described this in 1938, noting that wives complained their husbands were emotionally or physically absent while husbands complained of feeling pressured and nagged. Nearly a century later, the same loop keeps showing up in the data with remarkable consistency.

How much damage it actually does

It is not a minor irritant. A meta-analysis of 74 studies covering 14,255 people found that demand-withdraw was more strongly associated with reduced relationship satisfaction than any other communication pattern the researchers examined. More than criticism. More than defensiveness. The overall correlation with poorer outcomes was 0.360.

What that number means

Correlations get quoted constantly and explained almost never, so here is the plain version.

A correlation runs from 0 to 1. Zero means two things have no relationship at all. One would mean they move together perfectly, which essentially never happens with human behavior. In psychology, around 0.10 is considered small, 0.30 moderate, and 0.50 large.

So 0.360 is a moderate effect, and moderate is stronger than it sounds. It sits in roughly the same range as the relationship between regular exercise and better mood. It does not mean every couple with this pattern is headed for divorce. It means that across fourteen thousand people, the presence of this pattern reliably tracked with the relationship going worse, and did so more strongly than anything else measured.

If this is happening in your relationship, you are not overreacting by taking it seriously.

Why men withdraw during conflict more often than women

This is worth stating plainly, because it often gets softened into vagueness. The gender skew is real and well documented.

Eldridge and Christensen’s meta-analysis of 14 studies found the wife-demand and husband-withdraw configuration significantly more common than the reverse, with a moderate effect size. The skew is one of the more consistent findings in marital research, visible in study after study going back decades.

It also persists across a lifetime. A longitudinal study of middle-aged and older couples found demand-withdraw occurring at comparable rates throughout the life course, with one particular kind of withdrawal, plain avoidance, increasing markedly as couples aged. Contrary to the theory that men and women grow more similar in later life, the split in who demands and who withdraws actually widened over time.

But buried inside that research is a detail that turns out to matter enormously. The effect was strongest when the topic under discussion was something the wife wanted to change. When husbands raised their own issues, the effect weakened, and in some studies reversed entirely. The husband did the demanding and the wife did the withdrawing.

Hold on to that. It is the thread that unravels the standard explanation.

What happens when men withdraw during conflict

The best-known explanation comes from John Gottman and Robert Levenson, who from the mid-1980s wired hundreds of couples to heart rate monitors and filmed them arguing in an apartment-style lab at the University of Washington. Couples discussed a genuine point of conflict while their physiological responses were recorded continuously.

The threshold

They found that when a partner’s heart rate climbed past roughly 100 beats per minute, the useful part of the conversation was effectively over.

Gottman named this state diffuse physiological arousal, more commonly called flooding. A resting adult heart rate usually sits between 60 and 80. Past 100, adrenaline is up, peripheral blood vessels constrict, and the cognitive functions difficult conversations actually require start losing bandwidth. Listening. Holding two perspectives at once. Choosing words carefully. Remembering what you were about to say.

The person is not choosing to stop engaging. The equipment for engaging has temporarily gone offline.

From the outside this reads as indifference, and it is almost impossible not to read it that way. There is often a small self-soothing movement, picking at a sleeve or straightening something on the table, and the body angles slightly away. From the inside, people describe something closer to drowning.

Why twenty minutes

Recovery is not instant, which is the part most couples get wrong. Gottman’s work suggests roughly twenty minutes for heart rate to return to baseline once someone has flooded.

In the lab’s own count, roughly 85 percent of the people who went silent in heterosexual couples were men. Male partners also tended to show more sustained arousal and took measurably longer to settle, sometimes twenty minutes or more after their wives had already returned to normal.

That difference in recovery time creates a specific and very common trap. She has calmed down and is ready to resolve things. He is still physiologically lit up and cannot. She reads his continued flatness as sulking or punishment. He experiences her readiness to re-engage as the argument starting over.

This is where most articles on this subject stop, and where the pop version takes over: men flood more easily, therefore men withdraw, therefore it is biology and nothing can be done. That conclusion is doing far more work than the evidence can support.

Where the usual explanation falls apart

Three problems, in ascending order of seriousness.

The famous numbers have been challenged

Gottman’s headline findings have attracted real methodological criticism. Richard Heyman’s 2001 paper argued that the well-known divorce prediction figures were closer to postdiction than prediction, because the statistical models were built using data where the outcomes were already known. Fitting a model to results you already have, then reporting how well it sorts those same results, produces better numbers than the model would achieve on genuinely new couples.

Independent replication has also been thinner than the popularity of those numbers suggests. One community-based study of 85 couples that attempted to reproduce Gottman’s affective process models reported that the major findings failed to replicate.

None of this makes flooding fictional. The physiological measurements are real. It does mean the confident claims built on top of them deserve more skepticism than they usually receive.

Men and women are less different than the story requires

The broader claim underneath the pop version, that men and women are fundamentally different psychological creatures, does not hold up well. A large meta-analysis of gender differences by Janet Hyde found that on the large majority of psychological variables, men and women are more alike than different.

Real differences exist. They are mostly smaller than the popular literature suggests, and they mostly shrink further once context and expectation are accounted for.

The test that settles it

Here is the strongest evidence against the wiring explanation, and it comes from an elegantly simple study design.

If the pattern is caused by male and female biology, it should not appear in couples who are both the same sex. There would be no male and female difference available to produce it.

Researchers at Berkeley examined demand-withdraw behavior in 63 heterosexual, gay and lesbian couples. The pattern appeared regardless of couple type. Two men, two women, one of each, it made no difference.

And in every kind of couple, what predicted who demanded and who withdrew was not gender. It was the difference in how much change each partner wanted from the other. The partner who wanted more change did the demanding. The partner who wanted less did the withdrawing. That held just as strongly in same-sex couples as in heterosexual ones.

The researchers concluded that a pattern usually read as a difference between men and women more likely reflects social conventions that hand men more power in relationships and women less.Holley, Sturm & Levenson, Journal of Homosexuality (2010)

And it is just as damaging in reverse

Then there is the cleanest number of the lot. In that 74-study meta-analysis, the effect size for wife-demand and husband-withdraw was 0.380. For husband-demand and wife-withdraw it was 0.392.

Essentially identical. When the woman is the one shutting down, the relationship suffers just as much.

The harm has nothing to do with male psychology. It is the shape of the interaction that corrodes the relationship.

So why is it usually him?

The explanation that fits the evidence best is distinctly unglamorous: in most heterosexual relationships, women want more things changed than men do.

Think about what withdrawal accomplishes. The person who is broadly satisfied with how things are has no reason to open a negotiation, and no reason to keep one open once it has started. Ending the conversation preserves the current arrangement, and preserving the current arrangement serves whoever the current arrangement already suits.

That is not an accusation aimed at any individual man. Most men in this position are not calculating anything. They simply do not experience the same accumulating list of things that need addressing, because fewer of those things land on them.

Socialization sits on top of that. Men are taught earlier and more consistently that visible distress is a liability, and that the correct response to being overwhelmed is to contain it. Withdrawal becomes a practiced habit rather than an instinct.

This is not fully settled, and anyone telling you it is has stopped reading too early. But the weight of the evidence sits with position and power rather than wiring. Which matters practically, not just intellectually. Positions can be changed. Wiring, by definition, cannot.

What actually helps

Five things, roughly in order of how much difference they make.

1. Take a real break, not a token one

If he has genuinely flooded, continuing is not possible in any useful sense. You are talking to someone whose capacity for the conversation has temporarily shut down, and every additional sentence lands as pressure rather than information.

The research points to roughly twenty minutes for the body to return to baseline. A five-minute pause almost never does it.

The catch, and this is the part that gets missed: the break only works if the time is genuinely spent on something else. Sitting in the next room replaying the argument, rehearsing better responses or building the case keeps arousal high and achieves nothing. Twenty minutes of rumination produces someone who returns more agitated than they left.

Walk, shower, do dishes, watch something unrelated. The goal is distraction, not reflection.

2. Agree the rules before you need them

A time-out called in the middle of an argument is almost always heard as abandonment, because in the moment it is indistinguishable from walking out.

The fix is to agree it in advance, on a calm day, when nothing is at stake.

Either of us can call a break. Whoever calls it has to say when we are coming back to it, and we actually come back to it.

The commitment to return is the entire mechanism. Without it, a time-out is just a more polite version of the thing that was already happening. With it, the break stops being a threat, because the subject is not disappearing. It is being postponed by twenty minutes.

That distinction is what lets the demanding partner let go, and the letting go is what lets the withdrawing partner come back.

3. Change how the conversation opens

How a conflict conversation starts predicts a surprising amount about how it ends. A complaint about a specific behavior lands very differently from a criticism of character, even when the underlying grievance is identical.

• “You never help with anything around here” describes who he is. There is no available response except defending himself.
• “I ended up doing the kitchen alone again last night and I was annoyed about it” describes one event and one feeling. There is something to actually respond to.

The first version reliably produces the shutdown. The second frequently does not. This is not about being nicer, and it is not about suppressing the complaint. It is about giving him something answerable instead of something to survive.

4. Notice who is raising what

Since the pattern intensifies around topics the demanding partner wants changed, it is worth honestly auditing who brings things up in your relationship.

If the same person raises every issue, that is information. And if he never brings anything up at all, that is not contentment, whatever it looks like. It usually means he has concluded, on the basis of experience, that raising things does not go well.

Asking directly, at a neutral moment, is more productive than it sounds. Expect the first answer to be nothing. Expect a real answer some time later, probably sideways, probably while you are both doing something else.

5. Come back to it

The part that makes everything above work is returning to the subject. Not the whole argument, just the original issue, once both people are settled.

Couples who do this consistently find the withdrawal reduces on its own over time, because the thing driving the escalation was never the topic. It was the fear that the topic was about to vanish again.

What makes it worse

Three things that feel productive and are not.

Following him. Continuing the conversation into the next room, or through a closed door, converts a temporary shutdown into a longer one. Whatever gets said during that stretch is not being processed.

Requiring him to explain the silence in the moment. “Why won’t you just talk to me” is a reasonable question and an impossible one. Somebody who is flooded usually cannot describe what is happening to them while it is happening.

Treating one calm conversation as the fix. This pattern typically took years to establish, and it re-forms under stress even after real progress. Expect regression. It is not evidence that nothing changed.

If you are the one who withdraws

Most of this article is written for the person on the other side of the silence. If you recognize yourself as the one who goes quiet, a few things are worth knowing.

The physiological response is real and you are not imagining it. You are also not obliged to push through it, and pushing through it rarely produces a conversation anyone benefits from.

But silence without a return is not a neutral act. From the outside it is indistinguishable from not caring, and over enough repetitions your partner stops believing the subject will ever be addressed. That belief is what produces the escalation you are trying to escape. Your withdrawal and their pushing are the same loop, and either of you can interrupt it.

The single most useful thing you can do is say a version of this out loud, in the moment:

I can’t do this right now, but I’m not walking away from it. Give me half an hour.

Then keep the half hour. It is a small sentence that does an enormous amount of work, because it separates the two things your partner cannot otherwise tell apart.

One thing this does not excuse

Flooding is a real physiological state. It is also a convenient thing to hide behind.

There is a meaningful difference between a nervous system that has run out of capacity and a partner who has learned that going silent reliably ends conversations he does not want to have. The two look nearly identical across a kitchen table and they are not remotely the same thing.

The useful test is what happens afterward. Someone who was genuinely flooded comes back to the subject, even clumsily, even a day later. Someone using silence as leverage does not, and the issue quietly disappears until it resurfaces months later in worse condition.

If every difficult topic in your relationship dies this way, and not one of them ever gets picked up again, what you are looking at is not physiology. It is a strategy, and it should be named as one.

Common questions

Is stonewalling a form of emotional abuse?

It depends entirely on what is driving it. Genuine flooding is an involuntary response, and treating it as abuse is both inaccurate and unhelpful. Deliberate silence used to punish a partner, control the terms of a conversation or avoid all accountability is a different thing. The distinction is in the pattern over time, not in any single argument.

Does he actually stop listening, or is he just pretending?

Under genuine flooding, comprehension really does degrade. He is hearing sound and probably tone, but the detailed processing is impaired. This is why repeating a point more forcefully during a shutdown does not work.

How long does it take to change this pattern?

There is no reliable number, and anyone offering one is guessing. What the research does suggest is that the pattern is maintained by the loop rather than by either person’s character, which means it responds to changes in the loop. Couples typically notice the first difference in how quickly they recover, not in how often conflict happens.

What if he refuses to discuss any of this?

Raising it during or immediately after an argument almost never works. A neutral moment, framed around your own experience rather than his behavior, works better. If nothing moves over a long period and the pattern is damaging the relationship, a couples therapist trained in this area is a reasonable next step, and demand-withdraw is one of the most well-mapped problems they deal with.

What to take from all this

He is probably not stonewalling you to win. When a man goes quiet mid-argument, the most likely explanation is a body that has crossed a physiological threshold and a set of habits that never taught him what to do next.

But the reason it is usually him, and not you, has far less to do with how men are built than with who in the relationship currently wants something to be different. That is a question about the relationship, not about his nature.

Nature is fixed. Positions are not.

References

1. Schrodt, P., Witt, P. L., & Shimkowski, J. R. (2014). A meta-analytical review of the demand/withdraw pattern of interaction and its associations with individual, relational, and communicative outcomes. Communication Monographs, 81(1), 28-58.
2. Holley, S. R., Sturm, V. E., & Levenson, R. W. (2010). Exploring the basis for gender differences in the demand-withdraw pattern. Journal of Homosexuality, 57(5), 666-684.
3. Eldridge, K. A., & Christensen, A. (2002). Demand-withdraw communication during couple conflict: A review and analysis.
4. Holley, S. R. The demand-withdraw communication pattern in middle-aged and older couples: A longitudinal study. University of California, Berkeley.
5. Heyman, R. E. (2001). The hazards of predicting divorce without cross-validation.
6. Capaldi, D. M., et al. Generalizability of Gottman et al.’s affective process models. PMC1828692.
7. Hyde, J. S. (2005). The gender similarities hypothesis. American Psychologist, 60(6), 581-592.
8. Gottman, J. M., & Levenson, R. W. Research program on diffuse physiological arousal and marital interaction, University of Washington.
9. Terman, L. M., et al. (1938). Psychological Factors in Marital Happiness.