You've probably had the experience of being in an argument and sensing that the intensity has grown beyond the thing you are arguing about. The dishes, the comment at dinner, the thing they forgot, the content matters, but it no longer seems to explain the heat between you.
In moments like these, communication is only part of what is happening. Conflict can activate expectations about closeness, rejection, autonomy, and safety that were shaped long before this particular conversation. This piece looks at what happens when those attachment responses become active: why one person may pursue while another withdraws, why it becomes harder to think flexibly when emotions rise, and why repair can fail even when both people genuinely want it.
This piece maps the mechanisms rather than the felt experience. If you want to understand what these patterns feel like from the inside: the shame, the confusion and the experience of knowing your pattern and still being unable to stop it, the companion piece speaks directly to that: Why Knowing Your Attachment Style Doesn't Change How You Feel
What Happens When Conflict Starts to Feel Threatening
Conflict can produce real physiological changes. Heart rate may increase, muscles tense, attention narrows, and emotions become more intense. For some people, activation leads toward confrontation, urgency, or pursuit. For others, overwhelm brings withdrawal, numbness, or difficulty finding words.
These changes do not make thoughtful communication impossible, but they can make perspective-taking, flexibility, and repair considerably harder. Our capacity to communicate tends to shift when we feel threatened or overwhelmed; the parts of thinking that allow us to hold complexity, consider the other person's perspective, and generate nuanced responses become less easily accessible.
Polyvagal theory is one framework sometimes used to describe these shifts, distinguishing between states of social engagement, mobilisation, and shutdown. It can be clinically useful as a map. But the important point here is simpler: what feels manageable in a calm moment can feel genuinely overwhelming in the middle of conflict, and understanding that helps explain why two people who care about each other can find themselves unable to reach each other.
How Attachment Patterns Shape Conflict
Attachment patterns are, at their core, learned ways of managing the anxiety that arises when closeness, separation, or relational threat is activated. They become most visible during conflict, because conflict is one of the primary contexts in which fears about abandonment, rejection, engulfment, and not mattering come to the surface.
Anxious attachment in conflict
Someone with a more anxious attachment pattern may detect distance, real or perceived, as the primary threat during conflict. Silence, a flat tone, a partner's withdrawal, an unresolved tension: these can register as signals of abandonment, producing an urgent pull toward connection.
This can show up as pursuit, repeated attempts to restore closeness, often escalating in intensity as the perceived distance increases. Over-explaining and over-reasoning may follow, as an attempt to close the gap through words, though the underlying need is connection rather than information. Cues that might otherwise be minor such as a delayed response, a shift in tone can become outsized when the attachment system is highly active.
This is not drama for its own sake. It reflects something learned when connection felt unpredictable: that staying alert and reaching toward people was the most reliable way to maintain closeness.
Avoidant attachment in conflict
Someone who tends toward avoidance may detect emotional intensity or proximity as the primary threat during conflict. A partner's activation, urgency, or demands for engagement can register as overwhelming, producing a pull toward withdrawal or distance.
This can show up as going quiet, leaving the room, focusing on practical details rather than emotional content, or becoming monosyllabic. For someone who learned early to manage difficult feelings alone, a partner's emotional intensity may simply exceed what they can process without needing to step back.
When someone is highly overwhelmed, engagement may become genuinely difficult rather than simply unwanted. This is worth understanding, not as an excuse for the impact of withdrawal, but as a starting point for working with it differently.
When these patterns meet
When an anxious attachment pattern and an avoidant one meet in conflict, the result can become self-reinforcing. The more one partner pursues, the more the other withdraws. The more the other withdraws, the more urgent the pursuit becomes. The argument may be carrying attachment fears that neither person has yet named.
This dynamic is explored in more depth in the companion piece on anxious and avoidant patterns colliding.
When closeness itself activates both fear and longing
People with more fearful or disorganised attachment patterns may move between pursuit and withdrawal, sometimes quite quickly. Closeness can bring both relief and alarm, which can make conflict particularly confusing, reaching toward repair and then pulling away once vulnerability increases.
Some people may also experience blankness, detachment, or difficulty staying emotionally present when conflict becomes overwhelming. This isn't a fixed feature of a particular kind of nervous system, it's a response that can emerge when the situation exceeds someone's current capacity to remain present.
What looks like a communication breakdown is often two nervous systems doing exactly what they learned.
When This Framework Doesn't Apply
Attachment and nervous system frameworks can help explain many repetitive conflict patterns. They should not be used to flatten important differences in power and responsibility.
A relationship in which both people become activated during conflict is not the same as a relationship in which one person uses intimidation, threats, degradation, coercion, or control. Activation is something both partners may experience. Abuse is a pattern of behaviour that one person does to another.
Communication and regulation strategies cannot make an abusive dynamic safe. Understanding that a partner's controlling behaviour may have relational roots does not make it acceptable or make the relationship safe enough for the vulnerable repair work this piece describes.
Where there is fear, coercion, or a consistent pattern of one person having to organise themselves around the other's reactions, the issue may be significantly more than an attachment cycle. Specialist domestic abuse support is more appropriate than couples-focused communication work in those circumstances.
Why Repair Fails
Repair, the act of reconnecting after a rupture, is one of the most clinically significant moments in any relationship. John Gottman’s work on couple relationships has long emphasised the importance of repair, how couples reconnect after things go wrong, rather than expecting healthy relationships to be free of conflict.
And yet repair fails repeatedly in relationships where significant attachment activation is present. Understanding why matters.
Attempting repair before sufficient regulation
The most common failure of repair is timing. One or both partners attempt to reconnect while still highly activated, still flooded, still in a threat-response mode, before sufficient regulation has occurred.
Meaningful repair tends to require the capacity to tolerate vulnerability, to hear the partner's experience without becoming immediately defensive, to acknowledge one's own contribution to the rupture, and to extend genuine care to the partner's distress. These capacities can become considerably harder to access when either person is highly activated.
When repair is attempted prematurely, a few things tend to happen. Someone with an anxious attachment pattern may apologise quickly — not from genuine understanding, but to restore connection and end the intolerable distance. Someone with an avoidant pattern may offer repair as a way to end the emotional intensity and return to equilibrium. Both produce repair that is brittle: the rupture has been quieted rather than genuinely addressed. The unprocessed relational material accumulates and re-emerges in the next conflict.
Repair is usually more productive once both people have enough regulation to remain present and reflective. This is not avoidance; it is sequencing.
Repair attempts misread as continued threat
A second common failure: a genuine, well-intentioned repair attempt is misread by a still-activated partner as a continued threat.
For someone with an anxious attachment pattern, a partner's quiet, calm offer to talk may register as emotional distance; the calm tone, which is a sign of regulation, is interpreted as further disengagement. For someone with an avoidant pattern, an emotionally activated repair attempt, however loving in content, may register as escalation.
Both partners end up responding to what they perceive in the other's repair attempt rather than to the repair itself.
The absence of a repair template
For people whose early relationships didn't include consistent experiences of rupture followed by reconnection, repair may not simply be uncomfortable, it may be genuinely unfamiliar. If caregivers didn't model the process of acknowledging harm, taking responsibility, and reconnecting, there may be no established sense of what repair looks or feels like, or how to tolerate the vulnerability it requires.
Repair asks people to hold their own distress while attending to a partner's, to acknowledge their own contribution without collapsing into shame or deflecting, and to trust that vulnerability won't be used against them. For people without that relational history, each repair attempt can feel like stepping into territory they don't know how to navigate, which is itself activating.
Attachment Activation and Behaviour
Attachment activation can strongly influence behaviour during conflict, but it does not remove choice, impact, or responsibility.
What appears on the surface, pursuit, defensiveness, withdrawal, appeasing, or shutting down, may partly reflect a protective response that has become active. Understanding that response can make the behaviour easier to work with. It does not automatically make the behaviour harmless, and it does not transfer responsibility for impact to the person who experienced it.
Several things follow from understanding activation during conflict.
Reasoning alone is often ineffective when someone is highly activated. Attempting to reason with a flooded partner tends to increase their activation. What is needed first is not explanation but sufficient settling for both people to be able to think again.
Regulation is not a pause from the real work; it is the prerequisite for it. The capacity to engage with challenging relational material requires enough space to remain present and reflective. This is not avoidance. It is sequencing.
One partner's regulated presence can affect the other's capacity to settle. A calm, non-pressuring presence changes the relational field in ways that often operate beneath conscious awareness through voice tone, facial expression, and the absence of escalating demands. This is not a guarantee, and it doesn't remove the other person's responsibility for their own behaviour. But it matters.
Sometimes the presenting issue and the attachment meaning underneath it are both important. A disagreement about housework may also evoke a deeper sense of not mattering. A request for space may also activate fears of abandonment. Good repair attends to both rather than treating the practical issue as irrelevant.
What Regulation Can Look Like
A named pause with a return commitment is different from withdrawal. I'm overwhelmed right now. I need twenty minutes. I'm coming back to this. This names what is happening and commits to re-engagement, giving one partner an anchor rather than open-ended silence, while giving the other the space they need to settle.
Physiological grounding before attempting to talk tends to be more effective than attempting to talk through high activation. Slowing the breath, reducing physical tension, noticing the ground beneath your feet can shift something before conversation resumes.
Self-disclosure tends to be less activating than accusation. I feel frightened when you go quiet is different from You always shut down. The first describes what's happening internally. The second reads as a threat, which tends to escalate the partner's activation regardless of whether its content is accurate.
Titration of intensity matters. If the full emotional content of a rupture is too activating to address in one conversation, it doesn't need to be. Smaller, more frequent moments of connection and repair accumulate over time.
People differ in how long they need before they can return to a difficult conversation productively. The important distinction is between taking enough time to regain perspective and using distance to avoid ever returning to the issue. A specific agreed return time tends to meet both partners' needs more effectively than open-ended distance.
What becomes possible once both people have sufficient regulation: genuine acknowledgment of impact, accountability for one's own contribution, curiosity about the partner's experience rather than defence against it. These are not natural by-products of conflict — they are capacities that require enough settling to access.
The Slow Work
What changes the pattern is not understanding it; it is experiencing something different, repeatedly, in a context where there is enough safety to do so. Gradually, the expectation shifts: rupture does not have to mean abandonment; closeness does not have to mean losing oneself; the relationship can hold both people's distress without collapsing.
This is slow work. It is not achieved through insight alone.
Individual therapy can provide a consistent relationship in which needs, vulnerability, conflict, and repair can be explored differently from the patterns a person already knows. Over time, that experience can begin to update what someone expects from closeness and conflict more broadly.
Couples therapy can help both partners recognise the cycle, slow it down, and practise different ways of responding to one another — with a third presence able to observe what's happening in real time and help both people do something different with it.
Neither is quick. Both are possible.
Understanding the mechanics of what happens during conflict is one layer of this work. But attachment doesn't only operate in the domain of conflict and repair — it shapes how you experience yourself, how you understand your own contradictions, and why knowing your pattern doesn't stop it from running.
If that dimension of attachment is where you are right now, the companion piece speaks directly to it: Why Knowing Your Attachment Style Doesn't Change How You Feel
If you recognise your relationship in the patterns described here, the cycle that keeps returning, the repair that doesn't quite reach, trauma-informed therapy can help. This work begins by understanding what each partner has learned about closeness and conflict, mapping what runs between them, and building the capacity that makes genuine repair more possible.
📧 kat@safespacecounsellingservices.com.au
0452 285 526
Frequently Asked Questions
Can I do this work individually, or does my partner need to be involved?
Both have value, and they work differently. Individual therapy can help you understand your own attachment responses, develop greater capacity to regulate during conflict, and begin to change the pattern from your side, which does affect the relational dynamic, because your partner's nervous system is responding to yours as much as to the content of what you say. Couples therapy works with both nervous systems simultaneously, mapping the pursue-withdraw cycle and interrupting it in real time. For many people, individual work is the right starting point, particularly if the relationship isn't yet safe enough for joint work, or if your partner isn't willing to engage.
Why do we keep having the same argument even after we've resolved it?
Because resolution at the content level doesn't always address the underlying relational activation. If both nervous systems haven't returned to social engagement before repair happens, what gets produced is a brittle close — the rupture is quieted rather than genuinely repaired. The unprocessed relational material accumulates and re-emerges in the next conflict, often attached to new content. The argument looks different on the surface; the activation underneath is the same. What tends to produce more durable repair is both people returning to regulation before attempting to reconnect, and the repair conversation addressing the relational need underneath the content — the fear of abandonment, the sense of not mattering, the feeling of being controlled — rather than only the presenting issue.
My partner says they need space after an argument, but I need to resolve things immediately. How do we manage that?
This is one of the most common expressions of the pursue-withdraw dynamic, and the nervous system basis for it is real: different people genuinely need different amounts of time to return to a window of tolerance in which repair is possible. For some people, particularly those with anxious attachment patterns, unresolved rupture is itself activating — the uncertainty of an unresolved conflict can feel as threatening as the original conflict. For others, particularly those with avoidant patterns, attempting repair before sufficient regulation has occurred produces worse outcomes than waiting. What tends to help is agreeing in advance — outside of a conflict moment — on a pause-and-return protocol that names a time commitment rather than leaving the return open-ended. The anxious partner needs an anchor. The avoidant partner needs space. A specific agreed return time can meet both needs simultaneously.
I understand what's happening when we argue, but I still can't stop escalating. Why?
Because understanding and regulation are not the same thing, and they don't operate through the same mechanisms. You can have clear intellectual insight into a cycle while your nervous system is still running the older protective response — particularly under stress or when something in the current conflict activates an older relational wound. What changes the felt-sense response over time is not further insight but accumulated relational experience of something different: enough repeated experiences of conflict being navigated safely, repair happening and holding, vulnerability not being used against you. This is slow work. It is not a failure of willingness or intelligence that understanding doesn't immediately stop the reaction.
Does this mean couples therapy is always necessary?
Not always. Some couples develop sufficient understanding of their own and each other's patterns, and sufficient communication and repair capacity, through individual therapy, reading, or their own reflection and practice. What tends to distinguish relationships that can shift the pattern without formal couples work is both partners' genuine willingness to understand their own activation and take responsibility for their own contribution — not just to understand the cycle intellectually, but to do the harder work of changing their behaviour in the middle of it. Where the cycle is deeply entrenched, where repair has repeatedly failed, or where one or both partners are carrying significant attachment wounds, couples therapy provides something that self-directed work usually can't: a regulated third presence that can observe the cycle in real time, interrupt it, and help both nervous systems practice something different together.
What does it actually feel like when the pattern starts to change?
Usually quieter than people expect. Change in entrenched relational patterns rarely arrives as a dramatic shift — a conflict that resolves cleanly, a repair that lands perfectly. It tends to show up in smaller moments: noticing mid-argument that a familiar pattern is running, rather than only recognising it afterward. Recovering more quickly after activation. A repair conversation that reaches something real rather than producing only a temporary close. A moment of genuine curiosity about the partner's experience in the middle of a difficult exchange, where previously there was only defensiveness. These moments don't announce themselves as progress. But they accumulate — and over time they build a different relational history, one the nervous system gradually begins to use as its reference point for what conflict and repair mean.
