There is a moment in many couple arguments where everything stops working. The partner who was making sense ten seconds ago is now shouting or sneering or going silent. The conversational arc has clearly gone off a cliff. Each new sentence makes it worse. Both partners can feel the conversation is now harmful and neither can stop it. This is the moment John Gottman called flooding. The technical name is Diffuse Physiological Arousal — DPA — and once it has set in, no verbal skill, however well practised, will land. The brain is no longer running the conversation. The body is.
The Gottman 20-minute rule is the protocol for what to do at this moment. Many couples have heard of it. Most couples do it wrong. The version of the rule that works has very specific structure — duration, what to do during the pause, how to come back, what the partner cannot do — and the version most couples instinctively fall into either skips the structure or breaks the rule mid-pause. This piece is the full protocol, with the reasons each piece matters, and the common failure modes that make the same rule work for some couples and not others.
What Flooding Actually Is
Gottman's Love Lab studies, beginning in the 1980s, used heart-rate monitors and skin conductance sensors during recorded couple conflicts. The data showed a consistent threshold: when a partner's heart rate crossed roughly 100 beats per minute (around 80 for highly fit individuals), measurable cognitive changes appeared. Working memory contracted. Attention narrowed. Access to verbal repertoire dropped. Threat-detection circuits dominated. Higher-order capacities — empathy, perspective-taking, complex reasoning — went offline. The technical term DPA captures the diffuse, non-specific nature of this arousal: not the body responding to a specific threat, but the body in a general state of activation that interferes with conversational processing.
Above the flooding threshold, partners cannot make repair attempts that land. They cannot hear repair attempts directed at them. They cannot deliver soft start-ups. They cannot receive specific complaints. The most-rehearsed Gottman techniques and the most-careful NVC vocabulary are equally non-functional. The verbal layer is the second floor of a building whose ground floor is on fire. Putting more furniture on the second floor doesn't help. The fire is the thing. (See emotional flooding in couples.)
How To Recognise Flooding In Yourself
Self-recognition is the first skill. Partners who can name their own flooding before it has fully set in have the best outcomes. The signals are physical first and emotional second. Look for: pounding or noticeably faster heartbeat; tight chest or shallow breath; rising heat in the face; hands clenched or shaking slightly; tightening jaw; the sense that the partner has become an opponent rather than a partner; the impulse to say one more devastating thing or to leave the room entirely; tunnel vision so that the partner's face has stopped being readable.
The "I have to say this now" tell
One of the most reliable single signs of flooding is the urgent felt-need to make your point right now, in this conversation, before any pause. The urgency is the flooding talking. Outside of flooding, the same point can be made calmly in twenty minutes' time. The conviction that it has to be made now is what the body produces when DPA is high. Recognising this conviction as a symptom of flooding rather than as a real time constraint is what makes the pause possible.
How To Recognise Flooding In Your Partner
Your partner's flooding is sometimes harder to read than your own. The signals that work: change in voice quality (higher pitch, louder, or unusually flat); face flushed or pale; rigid posture or fidgeting; shorter, faster sentences; pupils widened; eye contact either too intense or completely avoided; the partner's responses skipping past your words and going straight to their own next point. If you have a partner who tends to stonewall, the signal is different: the partner goes very still, their eyes go down or unfocused, their answers shrink to monosyllables, the room goes quiet. That is flooded stonewalling rather than flooded escalation — same physiology, different output. (See stonewalling in relationships.)
The Protocol — Six Steps That Matter
"I'm flooded. I need twenty minutes."
The naming is critical. A partner who simply leaves the room is doing something that looks like stonewalling, which raises the other partner's threat-response. A partner who names what is happening — even with short, blunt words — signals that this is a regulating move rather than an exit. Use the actual word "flooded" if you have agreed it; use "I need a break" or "I need a pause" if not. The signal is the announcement.
Twenty minutes minimum. Up to an hour. Specify when you'll be back.
The Gottman 20-minute floor is not arbitrary. Heart rate and stress-hormone levels do not return to baseline immediately when the conversation stops. The autonomic recovery curve takes, in most adults, between fifteen and forty-five minutes depending on activation peak. Pauses shorter than twenty minutes mean both partners come back to the conversation still elevated, and the second round flares within sentences. The upper bound is whatever both partners need; the lower bound is non-negotiable.
Different rooms. Or one partner outside.
Sitting on the same sofa during a pause doesn't produce regulation; it produces a quieter version of the same conversation, with neither partner regulating fully. The Gottman protocol specifies physical separation because shared visual field keeps the autonomic system primed. Different rooms work. A walk outside works particularly well for the partner who needs to discharge activation through movement.
Reading, music, a shower, slow breath. Not running the argument in your head.
This is the step couples most often skip. If, during the twenty minutes, both partners spend the time mentally rehearsing what they should have said and what they will say when the conversation resumes, no regulation happens. Heart rate stays elevated. The pause produces no benefit and the second round is just as bad as the first. The pause has to be filled with something genuinely settling. Mindless activity. A novel. A bit of music. A shower. The point is not to think about the conversation. The point is to let the body come down.
No texts. No knocking. No "are we okay?"
The partner who is not flooded — or who is less flooded — has the harder job during the pause: do nothing. Pursuit during a regulating pause raises the flooded partner's activation and restarts the autonomic clock. The pursued partner now needs another twenty minutes to recover what they had just begun to recover. The non-pursuing posture is not coldness. It is faith that the protocol works. The pursued partner will return because the agreement was that they would.
"I'm back. I'm in a better place. Want to come back to it?"
Re-entry should be soft. A small announcement that you're back, an invitation rather than a demand, often an offer of warmth (a hand, a cup of tea) before words. The conversation that resumes after a regulating pause is, in most cases, materially different from the one that produced the pause — both partners are calmer, both have had time to think, both are more accessible. The same content gets handled with much less heat.
Choose a partner who can take a pause without making it a punishment.
LoveCertain matches on values (40%), life stage (25%), attachment (20%) and communication style (15%). Only matches above 70%. £49 once. Full refund if no relationship in 90 days. £99 bonus if there is.
Why Shorter Doesn't Work
Couples who try five-minute or ten-minute pauses almost always report that the protocol "doesn't work for them." It's not them. It is the duration. The autonomic recovery curve for high DPA is measured in tens of minutes, not single digits. A five-minute pause achieves nothing physiologically. A ten-minute pause achieves perhaps a third of what is needed. Twenty minutes is the floor at which most adults are reliably back below the flooding threshold, and for highly activated partners, thirty to forty minutes is more appropriate. The protocol is named for the floor, not the ceiling.
Pauses longer than two hours also lose their function. By that point the autonomic system has been baseline for a while; the additional time becomes avoidance rather than regulation. Two hours is the practical upper bound for most pauses. The exceptions are major rows late at night, where sleeping on it and returning the next morning is more useful than forcing a re-entry at 11pm.
Why Pursuit Doesn't Work
The pursuit-withdraw loop, in flooding form
The most common failure mode of the 20-minute rule is the non-flooded partner pursuing during the pause. The flooded partner names the pause and goes to the kitchen. Ten minutes later, the non-flooded partner appears with a "are you ready to come back? I really need to talk about this." The flooded partner — whose body has not fully recovered — is now back in the conversation prematurely. The second round runs the same arc as the first. Both partners conclude that the pause didn't work. The pause did work; the pursuit broke it. Couples who hold the no-pursuit rule, even when one partner is anxious about the silence, find that the protocol delivers on its promise. The anxious partner's discomfort during the pause is real and is the price of the protocol functioning.
Why The Flooded Partner Must Come Back
The protocol cuts both ways. The non-flooded partner has to not pursue. The flooded partner has to return. The pause is not a permission slip for one partner to indefinitely disengage. The return is a promise: I am stepping back to come back better. Couples in which the flooded partner takes the pause but then does not re-engage at the agreed time turn the 20-minute rule into something that functionally looks like the silent treatment, which destroys trust in the protocol over months. The pause has to be reliable. (See what the silent treatment actually means.)
If the flooded partner needs more time, the move is to communicate that — by text, briefly, no content — rather than to silently extend. "I need another half hour" is a renegotiation of the protocol. Silently failing to return is the violation. The communication is what keeps trust in the protocol intact.
The Pre-Agreement
The single highest-leverage thing a couple can do is agree the protocol when both partners are calm and not in conflict. The middle of a flooded argument is the worst time to negotiate the rules of breaks. Sitting down for ten minutes on a Sunday morning, naming that flooding happens for both of you sometimes, and agreeing the specific shape of pauses you will both use, makes the protocol available in the moment without negotiation. The pre-agreement covers: the trigger word ("flooded" or "pause" or "twenty minutes"), the duration floor (twenty minutes), the no-pursuit rule, the come-back commitment, and what each of you will do during the pause. Couples who have this conversation once find that the protocol works the first time they actually use it. Couples who have not had this conversation often blow up the first attempt because the rules were unclear.
The Sunday-morning pre-agreement, in five minutes
1. "We both flood sometimes. Let's agree what to do when we do." 2. "Trigger word is 'pause' — anyone says it and we stop for twenty minutes minimum." 3. "Different rooms, no texts, no chasing — both ways." 4. "The person who called it comes back at the agreed time. If they need longer, they text just to say so." 5. "When we come back, we start with the smallest version of the conversation, not where we left off." That's the protocol. Most couples haven't actually written it down once. Writing it down — even on a Post-it on the fridge — is the move.
Edge Cases
One partner always calls the pause; the other never does
If only one partner ever uses the pause, check the asymmetry. Either one partner floods much more easily — in which case the protocol is working — or one partner is using "I'm flooded" as a strategic exit rather than a regulating move. The diagnostic: a strategically-used pause produces no real regulation and the same conversation resumes at the same temperature. A regulation pause produces a visibly calmer partner who can engage the conversation differently afterwards. The first is the silent treatment in disguise. The second is the protocol working. (See criticism vs complaint.)
The flooded partner can't come back at all
If, repeatedly, the flooded partner takes a pause and never re-engages — the conversation gets dropped, the issue gets shelved, the next morning carries on as if nothing happened — the protocol is being used as avoidance. The remedy is the meta-conversation, in calm time, about the pattern. Couples who can name this pattern can usually work on it together. Couples in which the pattern is persistent and resistant to naming benefit from therapeutic support. The 20-minute rule is for couples who both want the conversation to continue, just not under flooding.
The pause itself becomes the row
"You're walking away from me again" — said by the anxious partner as the flooded partner steps out — is a familiar pattern in pursue-withdraw couples. The pursuing partner experiences the pause as abandonment. The remedy is two-fold: the flooded partner names the pause more clearly ("I love you, I'm coming back in twenty minutes, this is a pause, not an exit") and the anxious partner does deliberate work during the pause to tolerate the discomfort. Both halves are needed. (See 30 repair attempts that work.)
The Wider Research
The empirical base
Gottman and Levenson's original physiological studies (1985, 1992, 1998) used continuous cardiovascular monitoring during couple-conflict sessions to establish the link between DPA and conversational dysfunction. The 1992 paper in the Journal of Personality and Social Psychology reported the divorce-prediction findings; the underlying physiological work supports the 100 BPM threshold and the recovery-curve data behind the 20-minute rule. The protocol is summarised in Gottman's The Seven Principles for Making Marriage Work (1999, revised 2015) and elaborated in The Relationship Cure (2001). Sue Johnson's EFT framework treats the pursue-withdraw cycle as the most common couple pattern in distressed couples; the 20-minute rule maps onto the withdrawer's regulating need within that framework. Stan Tatkin's secure-functioning work emphasises pre-agreed protocols as a core feature of well-functioning couples.
The Certain Letter
Weekly relationship-science briefings. 4-minute read.
Why This Matters Earlier — In Dating
Two early-relationship implications. First: a new partner's capacity for the announced pause is one of the most useful signals available in the first six months. A new partner who, when overwhelmed, can say "I need fifteen minutes, I'll be back" — and then actually returns — is a partner whose regulating infrastructure is intact. A new partner who, when overwhelmed, either escalates without pause or disappears for hours without word, is showing you the pattern your serious arguments will have in year three. The signal is available in the first three months. (See arguing without destroying the relationship.)
Second: the early conversation about each of you flooding — what it feels like, what it looks like, what you each need from the partner when it happens — is one of the most useful pre-arguments a new couple can have. Most couples have it for the first time during their first major row, which is the worst possible moment. Having it during a calm Sunday in month three, before the relationship has been tested, is the kind of small structural investment that pays off through years. (See when communication breaks down — 7 repair moves.)
For an authoritative external primary source, see the Gottman Institute's article on physiological self-soothing and recovery.
The Encouragement
Flooding is not a moral failure. It is a feature of the human autonomic system that every adult experiences sometimes. The differentiating skill between couples who handle it well and couples who don't is not whether flooding happens but what the couple does with it when it does. The 20-minute rule is the simplest reliable intervention available. It requires no insight, no special vocabulary, no expensive therapy. It requires the discipline of naming the flooding, the discipline of taking the time, the discipline of not pursuing, and the discipline of returning. The discipline is small. The pay-off — the difference between an evening that ends with both partners calm and an evening that ends with both partners damaged — is large. Pre-agree the protocol this weekend. Use it the next time it is needed. Refine it across the months that follow. The whole couple's conflict texture improves within a few months of using it consistently.
