The Four Communication Styles That Harm Relationships: Gottman's Communication Patterns and How to Change Them
By Mark Ryan, BACP Accredited Psychotherapist and NCPS Accredited Relationship Therapist
When couples come to a first session, they often describe the same argument in different words. The trigger is different each time. The intensity is different. But there is a sequence underneath it that feels almost choreographed. One partner says something that lands like an attack. The other defends. The first one escalates. The second one shuts down. Both walk away feeling unheard, again. Almost every couple who has been together for any length of time has a version of this. And almost every version contains some combination of four specific patterns that have been studied more rigorously than any other set of behaviours in the field of relationship research. They are called the Four Horsemen.
Why these four patterns matter
In the 1970s and 80s, the American psychologist John Gottman set up what became known as the Love Lab. Couples were invited to spend extended periods in an apartment-style research setting while their interactions were filmed, coded second by second, and tracked alongside physiological measures like heart rate and stress hormones. Over decades and across multiple studies, Gottman and his collaborator Robert Levenson identified a small cluster of communication patterns that, when they appeared regularly during conflict, predicted relationship breakdown with striking accuracy. In the most cited of these studies, conducted in 1992, the researchers were able to predict which couples would divorce within six years with around 93% accuracy, based on observing a fifteen-minute conflict conversation (Gottman and Levenson, 1992).
The four patterns Gottman called the Four Horsemen of the Apocalypse, borrowing the biblical image of inevitable destruction. They are criticism, contempt, defensiveness, and stonewalling. The name is theatrical, but the finding is sober. These are not the only things that go wrong in relationships, but they are the patterns most consistently linked to breakdown. They also tend to appear in a sequence. Criticism opens the door. Contempt follows when criticism goes unresolved. Defensiveness arrives in response to both. Stonewalling sets in when defensiveness has nothing left to defend against. Once all four have become the default style of conflict, the relationship is in genuine difficulty.
The good news, and it is genuinely good news, is that each of the four has a recognised antidote. None of them are character defects. All of them are learned patterns, usually inherited from the families we grew up in. They can be unlearned. What follows is an honest description of each, the form it takes in everyday relationship life, and what really helps.
Criticism: the one that feels like honesty
Most couples do not realise they are criticising. They believe they are being honest. They believe they are naming a real problem. In a sense they are. The issue is not whether there is a problem worth raising. The issue is how it gets raised.
Gottman drew a careful distinction between a complaint and a criticism. A complaint is about a specific behaviour. A criticism is about a person's character. "I felt hurt when you forgot our anniversary" is a complaint. "You never think about me, you are so selfish" is a criticism. The first names a behaviour and a feeling. The second names a global flaw in the other person. The first invites a response. The second invites a defence.
In the therapy room, criticism is usually the easiest of the four to recognise once it is pointed out, and the hardest to give up. It feels reasonable. It often is reasonable in the sense that the underlying frustration is real. The partner on the receiving end, though, hears something quite different. They hear that they are the problem, not their behaviour. Over months and years of this, something hardens. They stop hearing the underlying need at all, and start hearing only the verdict on who they are.
The antidote Gottman called the gentle start-up. It is a way of beginning difficult conversations that names the behaviour, names the feeling, and names what you need, without dressing any of it up as a verdict on your partner. The structure is simple. I feel X about specific behaviour Y, and I need Z. "I felt lonely on Sunday when you were on your phone for most of the afternoon. I would love to plan something together next weekend." The same underlying message, delivered in a way that the other person can receive.
Couples often resist this at first. They feel it is artificial, that they should not have to script how they raise things in their own relationship. The honest answer is that most of us were never shown how to raise things well, and learning a structure for it is no more artificial than learning any other skill that does not come naturally. After a while, the structure becomes invisible. What remains is a different tone of voice, a different opening, and a partner who can hear what is being said.
Contempt: the most dangerous one
Contempt is the one that quietens the room in a couples session. The therapist sees it, both partners see it, and the temperature drops. It is the eye-roll. The sarcastic laugh. The mocking imitation of the other person's voice. The dismissive sigh. The "Oh, here we go again." It communicates one thing very clearly, and it does not need words to do it. The message is: I am above you. You are beneath me.
Of all the four horsemen, contempt is the one most consistently linked to relationship breakdown. Across multiple studies, Gottman identified it as the single strongest predictor of divorce, more powerful than any of the other three (Gottman, 1994). It is also linked to physical health. Couples who show high levels of contempt towards each other have been found to suffer more frequent infectious illnesses, with researchers suggesting a connection to suppressed immune function under sustained relational stress.
Contempt rarely arrives suddenly. It is usually criticism that has fermented. When complaints go unaddressed for long enough, something turns. The frustration is no longer about a specific behaviour, or even about a pattern of behaviour. It has hardened into a general view of the other person as inferior, foolish, contemptible. Once that view has settled in, every interaction is filtered through it. The partner stops being a person and starts being an example of why the other one is right to feel superior.
The antidote is not, on the face of it, complicated. It is appreciation. Specifically, it is the deliberate, daily practice of noticing and naming what you value in your partner. Gottman's research found that stable, happy couples maintain a ratio of around five positive interactions to every one negative interaction during conflict (Gottman and Silver, 1999). Couples heading for separation come in closer to one to one, or worse. The five-to-one ratio is not about avoiding conflict. It is about the broader emotional climate within which conflict happens. A relationship with a strong base of warmth, appreciation, and small daily kindnesses can absorb a great deal of difficulty. A relationship without it cannot absorb very much at all.
In therapy, I often ask couples to make appreciation a practice rather than a feeling. Not because the feeling does not matter, but because the practice tends to bring the feeling back. Naming, out loud, one specific thing your partner did that you valued, every day, even when you do not feel like it, especially when you do not feel like it. It is one of the most boring-sounding interventions in couples therapy. It is also one of the most powerful.
When couples argue, they often think they’re fighting about money, sex, time, or chores. But what they’re really fighting about is emotional safety.
Defensiveness: the most automatic one
Defensiveness is the one almost everyone does. It feels like self-protection. When you are being criticised, or when you think you are being criticised, defensiveness is the immediate, almost involuntary move to protect yourself by deflecting the blame. Most of us learned it in childhood, in homes where being wrong felt unsafe. We carry it into adult relationships without noticing.
Defensiveness takes a few common shapes. Counter-attack: "Well, you did the same thing last week." Innocent victim: "I cannot believe you would say that to me, after everything I have done." Excuse-making: "I was going to do it, but then I had to deal with something at work." Yes-but: "Yes, I did forget, but you have to admit that you have been distracted too." All of these have one thing in common. They communicate, in different words, that the problem is not you.
From the inside, defensiveness feels reasonable. You really were going to do the thing. You really have been having a hard week. The criticism really was unfair in its framing. All of that may be true. The problem is that defensiveness, even when justified, almost never resolves a conflict. It escalates it. Your partner, who came to you with something they wanted heard, is now arguing about whether they are allowed to have raised it at all. Nothing has been received. Nothing has shifted. The original concern is still sitting there, and now there is a fresh layer of disconnect on top of it.
The antidote is taking responsibility, even partial responsibility, for a piece of what your partner is naming. Not all of it. You do not have to agree with the whole account. You have to find the bit that is true and acknowledge it. "You are right, I did forget. I am sorry. Let me think about how to make sure I remember next time." The defensiveness has been set down. The conversation can move forward.
This is one of the hardest moves in couples therapy. Partners often experience it as backing down, as losing, as conceding territory they should be defending. It is none of those things. It is the move that makes resolution possible. Couples who learn it find that the underlying conflicts shrink, often quite quickly, because the conflicts are no longer doubled by the defensiveness layered on top of them.
Stonewalling: the one that looks like silence but is not
Stonewalling is the fourth horseman and often the last to arrive. It is what happens when a partner has reached the limit of what their nervous system can absorb in a conflict and shuts down. The conversation continues around them. They do not respond. They look away. They leave the room. They go quiet. From the outside it can look like calm, or indifference, or stubbornness. It is almost never any of those things. It is overwhelm.
Gottman's physiological research found that stonewalling tends to be accompanied by a specific bodily state he called flooding. The heart rate climbs above 100 beats per minute, sometimes considerably higher. Stress hormones spike. The capacity for rational conversation effectively goes offline. Whatever the partner who is stonewalling looks like on the outside, on the inside they are in a state of acute nervous system arousal, and no useful conversation is going to happen until they come back down.
Research has consistently found that in heterosexual couples, men stonewall more often than women, by a substantial margin. This is not because men care less. It appears to be linked to differences in how male nervous systems regulate under conflict stress, and the pattern starts early. Boys are often taught from a young age that emotional expression is unwelcome, that anger is dangerous, that quiet is safer than tears. By the time they reach adult relationships, the shutdown response is well practised. None of this is a moral failing. It is a learned, embodied pattern.
The antidote is self-soothing, which is to say physiological self-regulation, combined with a negotiated time-out. The shape of it looks like this. When you notice you are flooded, you name it. "I am too overwhelmed to keep talking about this right now. I need twenty minutes. I will come back at half past, and we can carry on." Then you take the twenty minutes, away from the conversation, and you do something that brings your nervous system back down. A walk. A breathing practice. A shower. Not scrolling your phone or rehearsing the argument in your head. Something that settles the body. Then you come back, on time, and you continue.
This is straightforward to describe and surprisingly difficult to do. The partner who is stonewalling has to learn to name what is happening, rather than just disappearing. The other partner has to learn that the time-out is not abandonment, but a precondition for the conversation continuing usefully at all. Both are skills. Both can be learned. There is a longer piece on this specifically, on what to do when one partner withdraws and shuts down, for couples where stonewalling has become the central pattern.
The cascade: how the four arrive together
It is rarely just one. The four horsemen tend to appear together, and they tend to follow each other in a recognisable sequence. Criticism comes first. Contempt arrives when criticism has gone unaddressed long enough to harden. Defensiveness shows up in response to both. Stonewalling sets in when defensiveness has run out of road. Each one invites the next, and each one makes the next more likely.
In therapy, recognising the cascade in your own relationship is often the first real turning point. Couples come in saying "we have a communication problem." They leave the first session knowing they have a specific, named, well-studied pattern, and that it is shared by most couples in difficulty. That alone tends to shift something. The vague sense of doom becomes a concrete thing that can be worked with. The piece on how to stop arguing with your partner covers what to do once you can see the pattern.
What the research predicts, and what it does not
The 93% accuracy figure is striking, and it is worth handling honestly. It refers to specific studies of conflict conversations, conducted under controlled conditions, where trained coders watched fifteen minutes of a couple discussing a difficult topic. It does not mean that any couple displaying any of these four behaviours is heading for divorce. Every couple displays all four at times. The question the research asks is about pattern and proportion, not presence and absence.
The other crucial finding, often missed when the four horsemen are discussed, is about repair. Gottman found that 84% of newlyweds who showed high levels of the four horsemen, but who were able to repair effectively after conflict, were in stable, happy marriages six years later. The presence of the four horsemen is not, by itself, the danger sign. The absence of repair is. Couples who can interrupt the cascade, name what is happening, soften, apologise, return to warmth, can absorb a great deal of difficult conflict and remain close. Couples who cannot repair, even where the conflict itself is less frequent, often cannot. There is more on this in the piece on repair after a fight.
What changes in couples therapy
Most couples can name at least one horseman in themselves and one in their partner within the first session of therapy. That recognition is the easy part. The harder, slower part is changing the patterns, and that is what the working phase of couples therapy is largely about.
The work tends to involve a few specific things. Slowing down the cascade so that each partner can see, in real time, when they are sliding into one of the four. Practising the antidotes, often clumsily at first, in the room with the therapist before trying them at home. Building the underlying ratio of positive to negative interactions, deliberately and daily, so that the relationship has a base of warmth to draw on when conflict arrives. And making sense of where each partner's particular horseman came from, because most of them are inherited rather than chosen. The defensive partner often grew up in a home where being wrong felt unsafe. The contemptuous partner often grew up in a home where contempt was modelled as a normal way to express frustration. Understanding the origin does not excuse the pattern, but it makes the pattern less personal, which makes it easier to change. There is a separate article on what to expect from your first couples therapy session in London, which covers how the work begins.
A final word
The four horsemen are descriptive, not prescriptive. Naming them is not a diagnosis. Most couples see themselves in all four, at least sometimes. The point is not to feel doomed, or to feel that your relationship is uniquely broken. The point is that these patterns are recognisable, well understood, and changeable. The couples Gottman studied who were able to interrupt the cascade, learn the antidotes, and build the underlying warmth back up, were the couples who stayed together and were happy six years later. The patterns are not the destination. They are the signal that something needs attention, and that something is workable.
If you recognise any of the four horsemen in your relationship and would like to talk about how couples therapy might help, you can book a free thirty-minute consultation. Consultations are available in person at the Kensington, Pimlico, or Angel locations, or online by video. If you are weighing up the decision, there is also a guide on how to choose a couples therapist in London.
About the Author
Mark Ryan is a BACP Accredited Psychotherapist and NCPS Accredited Relationship Therapist based in London. He runs Rise and Grow Therapy, a private practice specialising in couples and relationship therapy, with consulting rooms in Kensington, Pimlico, and Angel, Islington. He works with couples of all backgrounds and configurations, with particular experience supporting LGBTQ+ relationships and sex-positive practice. His primary modalities include integrative relational psychotherapy and Imago Relationship Therapy.
To explore whether couples therapy might be right for your relationship, you can book a free thirty-minute consultation, in person or online.
References
Gottman, J. M. (1994). What predicts divorce? The relationship between marital processes and marital outcomes. Hillsdale, NJ: Lawrence Erlbaum Associates.
Gottman, J. M., and Levenson, R. W. (1992). Marital processes predictive of later dissolution: Behavior, physiology, and health. Journal of Personality and Social Psychology, 63(2), 221-233.
Gottman, J. M., and Levenson, R. W. (2002). A two-factor model for predicting when a couple will divorce: Exploratory analyses using 14-year longitudinal data. Family Process, 41(1), 83-96.
Gottman, J. M., and Silver, N. (1999). The seven principles for making marriage work. New York: Three Rivers Press.
Hewison, D., Casey, P., and Mwamba, N. (2016). The effectiveness of couple therapy: Clinical outcomes in a naturalistic United Kingdom setting. Psychotherapy, 53(4), 377-387.