Advice to walk away from an argument without a word gets passed around as a single piece of wisdom, as though staying quiet and leaving were always the mature move. The research on conflict tells a more divided story. Walking away covers at least two different behaviors, one of which has reasonable support and one of which is associated with relationships going badly. The difference between them turns out to matter more than the walking away itself.
One is the short retreat taken when a person is physically overwhelmed. The other is the habit of going silent and staying gone. They can look identical from across the room.
What flooding is, and why a break can help
The useful concept here comes from the work of John Gottman and Robert Levenson, who spent years wiring couples to physiological monitors and watching them argue. In a 1992 paper in the Journal of Personality and Social Psychology, they recorded 79 newlywed couples having a fifteen-minute discussion about an ongoing disagreement while measuring heart rate, sweat, and blood flow.
Gottman’s term for what happens when arousal climbs steeply during a fight is flooding, or diffuse physiological arousal. Heart rate rises, stress hormones climb, and the body shifts toward a fight-or-flight footing. In that state, people are worse at listening, worse at problem-solving, and more likely to say something they will regret. Gottman’s practical reading of this, developed across his later work on marital dissolution, is that once someone is flooded it takes most people around twenty minutes to settle back down, and that pushing the conversation forward before then tends to make it worse rather than better.
On that logic, stepping away is not avoidance. It is letting a physiological storm pass so that the conversation can resume with both people able to think. This is the version of walking away that has a coherent rationale behind it.
Where the famous prediction claims need caution
Gottman is also known for a much stronger claim, that patterns like these can predict divorce with accuracy figures often quoted above 90 percent. That part deserves more skepticism than it usually gets.
Several of those headline accuracy rates came from models fitted and tested on the same small samples, which tends to flatter a prediction and inflate how well it will work on new couples. The underlying observation, that high physiological arousal and certain conflict behaviors are associated with worse marital outcomes, is real and has held up. The specific promise that a lab can watch you argue and forecast your divorce with near certainty is a different and shakier claim. Both things can be true: the physiology of conflict matters, and the marketing around predicting divorce ran ahead of the evidence.
The other kind of walking away
The second behavior is not a one-off break. It is a durable pattern in which one person pushes to engage and the other pulls back, shuts down, or goes silent. Researchers call it demand-withdraw, and it is one of the better-studied dynamics in relationship science.
In 2014, Paul Schrodt, Paul Witt, and Jenna Shimkowski published a meta-analysis in Communication Monographs pooling 74 studies covering more than 14,000 people. Across that body of work, the demand-withdraw pattern was associated with lower relationship satisfaction, less intimacy, poorer communication, and higher levels of things like anxiety. The overall association was moderate in size, around r equals 0.36 in their combined estimate.
Two cautions are worth stating plainly. This is correlational evidence, so it cannot establish that withdrawing causes the unhappiness rather than the reverse, or that some third factor drives both. And it describes an average across many couples, not a verdict on any single relationship. What it does not support is the idea that habitually going quiet and leaving is a neutral or wise default. On the available evidence, the withdrawing half of that pattern travels with the outcomes most people are trying to avoid.
Why the wordless part is the weak link
Put the two findings side by side and the popular advice starts to look upside down. The part with support is the timing: stepping back when your body is overwhelmed. The part with the worst associations is doing it silently and repeatedly, which is close to the definition of the withdrawing behavior the meta-analysis flagged.
What seems to separate a helpful pause from corrosive withdrawal is not the leaving. It is whether anything is said about it. Gottman’s own guidance is to name the break and commit to returning, something like saying you are too worked up to talk well right now and will come back in half an hour. That single sentence is the difference between a self-soothing break and a door quietly closing on the other person, who is left to interpret the silence however their worst guess suggests.
So the honest answer to when you should walk away without a word is: rarely, and less often than the phrase implies. Walking away to let a flooded body recover is well motivated. Doing it wordlessly is the part the research gives the least reason to recommend.
It is worth keeping the limits of this evidence in view. The physiological work rests on couples arguing for a set number of minutes, on cue, while wired to sensors in a laboratory, which is not quite how a fight unfolds in a kitchen. The recovery figures, including the often-repeated twenty minutes, are averages with wide variation between people, not a timer that applies to everyone. And most of this research has studied romantic couples, so how neatly it transfers to a tense meeting, a sibling, or a friend is more assumption than established fact. The direction of the findings is reasonably clear. The precision suggested by a single number is not.
One scope note. This research is about ordinary conflict between people on roughly equal footing. Situations involving intimidation, coercion, or fear are a different subject with a different literature, and leaving without explanation can be exactly the right call there. Nothing above speaks to those cases.
What the studies cannot tell you is the thing you actually want to know in the moment, which is whether this particular silence is a wise pause or the start of a pattern. That is not a question a heart-rate monitor or a meta-analysis can answer for you.