The Anxious-Avoidant Trap: How to Recognize the Cycle, Break It, and Build Secure Love Together

In Sue Johnson’s Hold Me Tight, the foundational text of Emotionally Focused Therapy, there’s a sentence that should be printed on the wall of every couples therapist’s office: “It is not the differences between partners that cause problems - it is the way those differences are handled when the attachment system is activated” (Johnson, 2008).

The anxious-avoidant trap, sometimes called the pursue-withdraw cycle, sometimes the “demon dialogues,” is what that sentence looks like in real life. One partner activates around closeness; the other activates around distance. The activation feels, to each person, like a response to the other person’s behaviour. It is actually a response to the partner’s response. The cycle runs itself.

Most couples don’t have a communication problem. They have an attachment-cycle problem, and the communication problem is the symptom.

This guide is the relational pillar — the one about what happens when our attachment styles collide. The companion piece is the diagnostic version (the what is my style and how did I get it version) and is the better starting point if you’re still figuring out the labels. This pillar assumes you’ve broadly identified yourself or your partner as anxious or avoidant; the work here is what to do about the dynamic.

A structural note. The most important sections below are written as side-by-side tables: what the anxious partner experiences in a given moment, and what the avoidant partner experiences in the same moment. The tables matter because the single most-missed point about this cycle is that both partners experience it as a response to the other person’s worst behaviour. The dual perspective makes that visible.

Why anxious and avoidant partners find each other

Amir Levine and Rachel Heller, in Attached, lay out an uncomfortable observation: the anxious-avoidant pairing is overrepresented in the population (Levine & Heller, 2010). It isn’t bad luck. It is a magnetic problem.

The anxious partner’s attachment system is calibrated to read intensity as love. A partner who pursues, who is consistently available, who makes their needs known. That partner’s signals are familiar but underwhelming. The anxious partner finds them “boring.” A partner who is unavailable, ambivalent, hard to read, that partner generates the level of activation the anxious system has come to expect. The activation feels like attraction.

The avoidant partner’s attachment system is calibrated in the opposite direction. A partner who is independent, low-need, doesn’t demand much, that partner is comfortable. A partner who is anxious, pursuing, intense — that partner is uncomfortable, but the discomfort is familiar. The avoidant has spent a lifetime regulating against intensity, so an intense partner activates the regulating machinery the avoidant has already built.

Each one’s discomfort is the other one’s known shape.

The magnetics, more carefully

Mikulincer and Shaver’s academic version of Levine and Heller’s framing: the anxious partner runs a hyperactivating strategy (amplifying distress signals, increasing proximity-seeking, sustaining vigilance for rejection cues) while the avoidant partner runs a deactivating strategy (dampening distress signals, increasing self-reliance, suppressing proximity-seeking) (Mikulincer & Shaver, 2007). Each strategy is complete; it can be run alone, but each is also calibrated against an imagined partner. The hyperactivator imagines an absent partner who needs to be summoned. The deactivator imagines an intrusive partner who needs to be held off.

When the two strategies meet, each partner finds in the other the imagined figure their strategy was built for. The anxious partner needs a partner whose availability is uncertain enough to justify the vigilance; an unambiguously available partner gives the hyperactivating system nothing to do. The avoidant partner needs a partner intense enough to justify the deactivation; an unambiguously low-key partner gives the deactivating system nothing to push against. The pairing is, in a strange way, generative; each partner gives the other’s nervous system the input it was built to process.

This is also why each partner’s behaviour reads, to the other, as confirming the worst story. The avoidant withdraws, which confirms the anxious story that love is conditional and exit is imminent. The anxious protests, which confirm the avoidant story that closeness is engulfing and the self must be defended. Each partner’s regulation strategy generates the precise behaviour that the other partner’s regulation strategy is most allergic to. Simpson and Rholes describe this as the stress-amplification property of insecure dyads — under load, the strategies don’t dilute each other; they amplify (Simpson & Rholes, 2017).

The early phase of the relationship masks this. The intensity is read as chemistry; the push-pull is read as passion; the reconciliations after fights are read as evidence of how much both partners care. Tatkin describes this as the “magnetic phase” (Tatkin, 2012). What is actually happening is that two attachment systems are running near the top of their dynamic range, and the resulting arousal is being narrated as love.

This isn’t romantic or destiny. The pairing is not a soulmate signal. But it explains why two adults whose attachment patterns look incompatible from the outside repeatedly find themselves drawn to each other, and why the relationship that results is intense from the start. The intensity is the attachment system, not the love.

The negative cycle, mapped

Sue Johnson’s EFT framework breaks the cycle into three stages: an escalation, a withdrawal, and an isolation. The deep-dive on the EFT loop covers the framework in detail; this section is the introductory map (Johnson, 2004).

The cycle begins with a trigger, usually small. A partner is late to call. A partner doesn’t respond to a text for two hours. A partner uses a tone that the other reads as cold. Whatever the trigger is, it activates one partner’s attachment system. From here, the cycle has a predictable shape:

1. Activation. The anxious partner’s system fires. They feel rejected, abandoned, unseen.
2. Protest. The anxious partner reaches out — sometimes warmly, often with edge. The reach is the protest behaviour: criticism, repeated checking, a sharp text, an accusation.
3. Avoidant activation. The protest activates the avoidant partner’s system in the opposite direction. They feel pursued, suffocated, attacked.
4. Withdrawal. The avoidant partner pulls back — physically (the other room) or emotionally (going flat, becoming functional but distant).
5. Anxious escalation. The withdrawal confirms the anxious partner’s worst fear. They escalate the protest.
6. Avoidant escalation. The escalation confirms the avoidant partner’s worst fear. They withdraw further.

By the end of the cycle, both partners are in physiological flooding (see the deep-dive), neither can hear the other, and the original trigger is forgotten. Both partners are convinced the problem is the other person.

Johnson’s three-stage framing places this six-step sequence inside a larger arc: first, the escalation (the protest-withdraw cycle gathering energy), then the withdrawal (one or both partners disengaging from active conflict but still inwardly activated), then the isolation (the longer-term flattening of the relationship into parallel lives) (Johnson, 2004). Most couples reading this pillar are mid-escalation, escalation-into-withdrawal, or oscillating between escalation and isolation. The exercises later in this article are calibrated for the escalation stage, where the energy is still in the system, and there is something to interrupt. Couples deep in the isolation stage often need professional support, which the later section covers.

What each partner experiences (the dual-perspective core)

This is the section that does most of the work of this pillar. The cycle is hard to interrupt while you’re in it because each partner experiences it as a one-sided assault. Seeing the partner’s experience side-by-side with your own is what makes the cycle workable.

What follows is four common triggering moments, each broken into a parallel table: what each partner thinks, feels in the body, does, and quietly needs. The tables are condensed; the prose underneath fills in the texture. Read both columns, even if you only identify with one. The work of the cycle is reading both.

Same moment, two interpretations

Trigger 1: Partner is 30 minutes late and hasn’t texted.

Anxious partner experience	Avoidant partner experience
Thinks: Something’s wrong. They’ve been in an accident. Or they’re avoiding me. Or they don’t care enough to send one line. The mind narrows to a single question: Am I still chosen?	Thinks: I’m running late, but I’ll be there. Nothing to report. I’ll explain when I get there. The thought of texting en route doesn’t surface; the implicit model is that arrival speaks for itself.
Feels in the body: Heart rate up, chest tight, stomach hollow. Cannot concentrate on anything else. Attention loops on the phone. The body is generating the same response it would generate to genuine danger.	Feels in the body: Slightly hurried, slightly task-focused, mostly neutral. The body is in a low-grade goal-pursuit state. No alarm.
Does: Texts once, neutral. Waits. Texts again, sharper. Composes and deletes a third. May call. May open the location sharing app. May rehearse the confrontation.	Does: Continues the task that delayed them. Phone in pocket. Notifications glanced at and triaged as “deal with on arrival.” Arrives 35 minutes late, walks in expecting the day to resume.
Needs (but rarely asks for): Five seconds of contact, a one-line text that confirms the partner is okay and the connection is intact. The need is not for the partner to be on time. The need is for the gap to be bridged.	Needs (but rarely asks for): To be allowed to handle logistics without performing reassurance during them. The need is not for permission to be late. The need is for the lateness not to become evidence of a deeper failure.

The single message, “Running 30 late, all good, see you soon,” would have cost the avoidant partner three seconds. The avoidant partner doesn’t send it because, from inside the avoidant nervous system, the message feels redundant. They’ll see me when I arrive. The anxious partner reads the absence of the line as the absence of caring. Both partners now walk into a fight neither one understands they’ve already started.

Trigger 2: A pointed text mid-workday.

Anxious partner experience	Avoidant partner experience
Thinks: They’ve been quiet all day. I need to find out where I stand. Sends: “I guess your phone is broken again.” The message is, internally, an attempt at connection. Externally, it carries the edge of the activation.	Thinks: I am in the middle of something. This is a lot. Reads the message as accusation. The mind generates: they’re going to be impossible tonight. I need to brace.
Feels in the body: A burst of relief at sending, now they’ll respond, followed within a minute by a wave of regret and renewed activation when no reply arrives. The cycle of anticipation and waiting is its own small loop.	Feels in the body: A sudden drop in cognitive bandwidth. The meeting that had been absorbing now feels like background noise. A low pulse of dread about the evening.
Does: Watches the phone. Composes a follow-up. Considers a softening text. Considers a sharper one. May open the conversation to re-read what was sent and decide whether it was “really that bad.”	Does: Puts the phone face-down. Returns to the meeting with reduced presence. Tells themselves they will “deal with it later.” Later means after the avoidant nervous system has come back to baseline.
Needs (but rarely asks for): One sentence of contact that says I see you, I’ll be there at 7, we’re fine. The reach was a bid; the need is to have the bid received without retaliation or silence.	Needs (but rarely asks for): Permission to delay the conversation until they can hold it without flooding. The deferral is not abandonment; it is the avoidant version of coming to the conversation as a regulated person.

The anxious partner experiences the delay as confirmation of the worst story. The avoidant partner experiences the delay as self-care. Neither one is wrong about their own experience. The cycle compounds because each partner’s coping move is the other partner’s most-feared input.

Trigger 3: One partner brings up a recurring issue at 11pm.

Anxious partner experience	Avoidant partner experience
Thinks: I’ve been holding this all day. If I don’t say it now, I’ll be up all night. I need this resolved before I can sleep. The urgency is real and embodied; the activation does not tolerate carrying the issue overnight.	Thinks: Not now. Please not now. I have nothing left. Whatever this is, it cannot be productive at this hour. The thought of opening a difficult conversation registers as a demand on a depleted system.
Feels in the body: A churning, restless energy. The nervous system has decided that movement toward resolution is the only path back to regulation. Sitting with the issue feels intolerable.	Feels in the body: Heaviness, a pull toward sleep, an aversion to the brightness of engagement. The body is in a parasympathetic dip and reads the bid as a demand to spike arousal.
Does: Opens the conversation. May start gently. May start with “we need to talk.” If the avoidant partner deflects, escalates. If the avoidant partner goes quiet, escalates differently, louder, more pointed, more questions.	Does: Says “can we talk about this tomorrow.” If the anxious partner accepts, sleeps with low-grade dread. If the anxious partner persists, goes silent. Silence is, from inside the avoidant nervous system, the least-bad option remaining.
Needs (but rarely asks for): A reliable container — a known time tomorrow when the conversation will actually happen so the anxious system can release the issue tonight. The fear of deferred is the fear of never.	Needs (but rarely asks for): A genuine not tonight that is honoured, paired with a genuine let’s do it Saturday morning that is also honoured. The fear of now is the fear of every interaction becoming an ambush.

The deep-dive covers what to do specifically when one partner shuts down. The short version, embedded in the table above, is that both partners are usually correct about their own state, the anxious partner cannot easily defer, the avoidant partner cannot easily engage at 11pm — and the working solution lives in the future-scheduled container, not in winning the present moment.

Trigger 4: Partner doesn’t reach for them during a hard moment.

Anxious partner experience	Avoidant partner experience
Thinks (anxious is the one in distress): I just told them something that mattered. They didn’t move toward me. They didn’t even reach for my hand. They don’t care, or they don’t know how to care, and either way I am alone in this.	Thinks (the avoidant is the one not reaching): They look like they need space. I don’t want to crowd them. I’ll be in the room. They know I’m here. If they need something specific, they’ll say. The non-reach is, from inside the avoidant nervous system, a respectful gesture.
Feels in the body: A sudden cold, a hollowing. The body’s interpretation of the non-reach is abandonment. The activation is full-body and immediate.	Feels in the body: A respectful stillness. The body is tracking the partner with care but is reading the situation as one that calls for presence rather than action.
Does: Withdraws — physically, conversationally, sometimes for hours. The withdrawal is not punitive in intent; it is what happens when the anxious system concludes the partner is unavailable. May later say “I needed you and you weren’t there.”	Does: Stays nearby, reads, makes tea, exists in the same room. Believes they are providing the right kind of support. Is genuinely surprised, hours later, when the partner says they felt abandoned.
Needs (but rarely asks for): A single physical gesture - a hand on the back, a held look, an I’m here that confirms the bond before the words are figured out. The need is for contact, not a solution.	Needs (but rarely asks for): A direct request: please come hold me. The avoidant cannot reliably read the bid; the anxious cannot reliably issue it directly. The result is a near-miss inside the same room.

Marcus (anxious) tells his husband Theo (avoidant) that his father has had a recurrence. Theo, who loves Marcus deeply and is genuinely shaken by the news, sits next to him on the couch and doesn’t speak. Theo believes he is offering steady presence; he had a mother who narrated her own emotions during his crises and he hated it. Marcus reads the silence as Theo not knowing what to feel. The gap between Theo’s intent and Marcus’s experience becomes the second wound on top of the first.

The shape underneath the four triggers

Across the four moments, two things repeat. Both partners are responding to the same situation through nervous systems calibrated by different histories, and both are doing what their history taught them is the safest move. Each partner’s safest move is the other partner’s most-feared input. This is the structural reason the cycle persists.

The four-row logic — thinks, feels in the body, does, needs — is also a diagnostic. Inside a real argument, both partners can usually fill in the does row. Almost no one can fill in the thinks row for their partner accurately. Almost no one can name the body experience their partner is having. And the needs row is the line the cycle most reliably hides because the asking is buried inside the protest or the withdrawal.

What each partner needs (and can’t ask for)

Anxious partner: needs reassurance that the relationship is intact and that the partner is still emotionally engaged, still chooses them. Often cannot ask for this directly because the asking feels weak. So asks indirectly, through the protest behaviour, which is the form the asking takes when direct asking feels too vulnerable.

Avoidant partner: needs space to regulate the nervous system before any further conversation can happen. Often cannot ask for this directly because the asking feels selfish or like abandonment. So withdraws instead, which is the form the asking takes when direct asking feels too charged.

Both partners are asking, in their attachment language, for the same thing: relief from the activation. Neither one’s asking is being heard for what it is.

Exercises to interrupt the cycle

The good news, from forty years of EFT research, is that the cycle is interruptible. Not eliminable — most couples will continue to have versions of it — but interruptible to a point where it no longer dominates the relationship (Wiebe & Johnson, 2016). What follows is six protocols, each with the same structure: name, mechanism (what it changes and why), when to use it, the script (what to actually say), and what to expect the first few times. The protocols are not stacked in order of difficulty; pick the one that maps to where your cycle most often catches.

Exercise 1: The cycle-naming protocol

Mechanism. The single most-leveraged move in the EFT literature is to name the cycle aloud while it is happening. Naming creates a small gap between activation and reaction; into that gap, choice can re-enter. Functionally, it also reframes the problem: the enemy becomes the cycle, not the partner. Sue Johnson’s research found that this single reframe, we have a cycle, and it has us both — predicts relationship recovery more strongly than any specific communication tactic (Johnson, 2008).

When to use it. Early in any escalation, the moment one partner notices the loop starting. Best used before either partner is fully flooded; once you’re flooded, the body cannot reliably execute the protocol.

The script. “I think we’re in our cycle. I’m activated. You’re probably activated too. Can we pause for 20 minutes and come back to this?” The phrasing matters. Our cycle, not your behaviour. I’m activated, not I’m fine, you’re the problem. Pause and come back, not let’s drop it.

What to expect the first few times. The first attempt usually fails. The naming itself becomes an accusation, “you’re in your cycle again” and the cycle continues. Couples need three to five clumsy attempts before the protocol starts to land. The way to make it land faster is to agree on the phrasing in a non-conflict moment, write it down, and treat the words as fixed. The fixedness is the point; in the moment, neither partner can be trusted to invent fresh language.

Exercise 2: Co-regulation by proximity

Mechanism. The body comes back to baseline faster in the presence of a regulated body than alone. Tatkin describes this as the couple operating as a psychobiological unit — two nervous systems that, when in proximity, entrain each other (Tatkin, 2012). The implication is that the first move in an escalation should not be to address the content. It should be to bring at least one nervous system back to baseline first, ideally by proximity to a body that is consciously regulating.

When to use it. After the cycle has been named, before either partner attempts to discuss the content. Also useful as a daily practice, independent of conflict, to build the underlying regulation capacity.

The script. Sit or stand within arm’s reach of each other. Do not yet talk. Both partners breathe slowly — the standard guidance is four-count in, six-count out — for three minutes minimum. One partner may place a hand on the other’s back or chest; the touch is optional but accelerates the effect. The phrase to break the silence, if needed, is the simplest possible: “I’m here.”

What to expect the first few times. The anxious partner often feels the three minutes as agonisingly long; the activation wants to do something. The avoidant partner often feels the proximity as too much; the deactivator wants more space. Both responses are the protocol working — what is being practised is the tolerance of the uncomfortable middle. The deep dive on Co-Regulation Exercises for Anxious-Avoidant Couples walks through five variations.

Exercise 3: The safe-haven protocol

Mechanism. Most cycle moves are protests — indirect bids dressed up as accusations or withdrawals. The safe-haven protocol practices issuing the bid directly, in plain attachment language, without the protest layer. The mechanism is that direct bids are receivable; protest bids are not. Johnson’s Hold Me Tight conversations are the canonical version (Johnson, 2008).

When to use it. In a non-conflict moment, repeatedly, before trying it in a real moment. Most couples need 10–20 low-stakes rehearsals before they can run the protocol during an actual activation.

The script. The anxious partner practises saying, slowly: “I’m scared right now. I’m scared you’re going to leave. I know it isn’t rational. I need to know you’re still here.” The avoidant partner’s practice is to receive the bid, to say, simply, “I’m here. I’m not going anywhere. Take your time.” No defending, no explaining, no well, when you do X I do Y.

The avoidant partner practises a symmetrical move, with their own attachment language: “I’m overwhelmed. I’m not running away from you. I just need to come down before I can talk.” The anxious partner’s practice is to receive that too. “Take your time. I’ll be here when you come back.”

What to expect the first few times. Both partners find their own line genuinely difficult to say. Each one’s protest is, in a sense, easier than the vulnerable version. The exercise is most useful when both partners can name what was hard about saying their line — “I felt weak,” “I felt selfish,” “I couldn’t believe you’d actually stay.” Naming the difficulty is part of the exercise. The deep-dive on How to Self-Soothe Anxious Attachment covers the anxious side; the deep-dive on Dismissive-Avoidant Emotional Needs Checklist covers the avoidant side.

Exercise 4: The time-out done right

Mechanism. Time-outs work when they are bounded, and re-entry is pre-committed. They fail when either partner experiences the time-out as abandonment. The protocol below borrows from Gottman’s flooding research and from Tatkin’s couple-bubble framing: the time-out is a couple move, not a unilateral exit (Tatkin, 2012).

When to use it. When one or both partners are flooded, heart rate elevated, cognition narrowed, and hearing the partner as an enemy. Once flooded, no productive conversation is possible until the body has come back down, which takes a minimum of 20 minutes and often longer.

The script. The partner calling the time-out says, “I’m flooded. I need 30 minutes. I’m not leaving the relationship; I’m leaving the room. I will come back at [specific time], and we will pick this up.” The specificity is what makes this different from withdrawal. The other partner’s job is to honour the time and not pursue during it. The returning partner’s job is to actually come back at the named time, even if they would prefer to keep avoiding.

What to expect the first few times. The anxious partner finds the 30 minutes nearly intolerable and will be tempted to pursue. The avoidant partner finds the return nearly intolerable and will be tempted to extend the time-out indefinitely. Both temptations are the protocol’s stress test. Couples who hold the time and the return consistently for two weeks generally find that the time-to-recovery from a flood shortens by a meaningful margin.

Exercise 5: The re-entry script

Mechanism. Most couples have no agreed script for what to say when they come back together after a rupture. The vacuum is filled by whichever partner speaks first, usually with either a defence or a renewed accusation, and the cycle re-ignites. A pre-agreed re-entry script bypasses this by giving both partners a known opening.

When to use it. After any time-out, after any cycle that ended in silence, at the start of any conversation that follows an unrepaired rupture.

The script. A two-line opening, agreed in advance. Partner A: “I want to come back to what happened. Can I start by saying what I think you were feeling?” Partner B: “Yes. And then I’ll do the same for you.” The structure forces each partner to lead with their best guess at the other’s experience, not their own grievance. The deep-dive on How to Repair After a Fight covers the longer repair arc.

What to expect the first few times. Both partners will get the other’s experience partly wrong on the first try. That is fine. The point is not accuracy; the point is the attempt, the visible effort to enter the conversation through the other person’s door rather than your own. The correction (“close, but it was more like…”) is the conversation. Greenberg and Goldman’s work on emotion-focused couple work suggests that this kind of guided perspective-taking is one of the more reliable predictors of conversational repair (Greenberg & Goldman, 2008).

Exercise 6: Boundary-setting without rupture

Mechanism. For avoidant partners, the deep work is learning to ask for space without disappearing. For anxious partners, the deep work is learning to receive the pause without escalating. Both are skills, both can be practised, both work better when paired.

When to use it. Whenever the avoidant partner needs space.

The script. Avoidant partner: “I need to come down. I’m not going anywhere. Can we pick this up at 8?” Three components: the need (come down), the reassurance (not going anywhere), the specific re-entry (8). Anxious partner: “Okay. I’ll see you at 8. I love you.” The bond statement is what allows the anxious system to release the pause.

What to expect the first few times. Both partners feel the script as wooden. That is the point. Wooden is recoverable; improvised is the cycle. The deep-dives on How to Communicate Boundaries to an Avoidant and Anxious Attachment Triggers Checklist cover the components in more detail.

When the cycle needs professional support

Self-work is meaningful, but there’s a threshold past which a therapist is the appropriate next step. Signs that you’re at that threshold:

• The cycle has been running for years without changing
• The cycle now involves contempt — eye-rolls, mockery, withdrawal of basic respect — not just protest and withdrawal
• One partner has stopped trying to repair entirely
• The cycle is so frequent it’s affecting daily function (sleep, work, parenting)
• Either partner has a history of trauma that is surfacing as the work deepens

The modality with the strongest evidence base for attachment-cycle work is Emotionally Focused Therapy. EFT was developed precisely for this pattern; finding an EFT-trained therapist is the highest-leverage move when self-work is no longer enough (Wiebe & Johnson, 2016). Simpson and Rholes’s review of stress and adult attachment suggests that under chronic relational stress, the regulation strategies of insecurely-attached partners deteriorate rather than improve without external scaffolding — which is the clinical case for not waiting until the cycle has run for another year (Simpson & Rholes, 2017).

The Twogle Check-In is a structured one-session option for couples who want a professional assessment of their cycle before committing to ongoing therapy. The deep-dive pillar on DIY Marriage Counseling covers the spectrum from self-guided to professional work.

For couples experiencing abuse — physical, emotional, or coercive — none of the protocols here apply, and the crisis resources in the disclaimer above this article are the first step. The anxious-avoidant cycle is a pattern between two partners who are both safe and both trying; it is not a framework for safety problems.

Building earned security as a couple

The most misunderstood part of attachment work is the timeline. Couples often hope that once they’ve named the cycle, they’ll stop having it. That’s not how it works. Couples who do the cycle work successfully continue to enter the cycle — they just exit it faster, and the cycle’s intensity reduces over the years (Mikulincer & Shaver, 2007).

The change is measured in months and years, not weeks. What follows is a realistic timeline, drawn from clinical norms in the EFT literature and from what most couples can expect when both partners are actively engaged.

The first three months

The work is mostly naming. Both partners learn to recognise the cycle when it is running and to call it out without the call-out itself becoming a new weapon. The cycle does not yet get shorter; the recognition does. Most couples in months one to three report that they still have the same fights, but they end them with at least one partner saying “that was the cycle again” afterward.

That post-hoc naming is the entry point. Expect setbacks. Expect a fight in week six that feels like all the work has been undone. It hasn’t, single bad weeks are noise, not signal, in the first quarter.

Around six months

If the work is holding, the cycle starts to shorten. Fights that used to last two days resolve in two hours. The flooding episodes still happen, but become recognisable while they are happening. The safe-haven protocol starts to be usable in real moments, not just in rehearsal. Both partners begin to have a felt sense of what the other is experiencing during a cycle, the dual-perspective frame moves from intellectual to embodied.

Around this point, many couples notice a counterintuitive thing: the number of cycles may not have decreased much, but the severity has. The everyday friction is still there. The catastrophic blow-ups are rarer.

Around one year

The cycle becomes one of several patterns in the relationship rather than the defining one. Couples report that they still recognise the anxious-avoidant dynamic but no longer feel governed by it. New, healthier patterns like direct asking, scheduled repair conversations, co-regulation as a daily habit start to occupy the relational space the cycle used to dominate.

This is the point at which Mikulincer and Shaver’s earned security framing begins to apply: not that the underlying attachment style has flipped, but that the partners have developed enough secure-functioning capacity together that their day-to-day operates from security even though their individual attachment histories haven’t been rewritten (Mikulincer & Shaver, 2007).

Beyond a year

The work doesn’t end; it integrates. The cycle, when it appears, becomes a familiar visitor rather than a stranger that destroys the week. Both partners can name when they’re triggered before they act on the trigger. Repair becomes routine, not heroic. The deep-dive on Earned Secure Attachment Case Studies walks through what this looks like over multi-year arcs.

The trade-off, for couples who do the work, is a relationship that genuinely changes shape — not the same relationship with better communication, but a different relationship in which the original wiring no longer runs unimpeded.

Frequently asked questions

What if only one partner recognises we’re in a cycle?

This is the most common situation in which someone reads an article like this one. Usually it is the anxious partner who has done the reading and the avoidant partner who is either resistant to the framework or genuinely unaware of it. There is real work the recognising partner can do alone, but there is a ceiling.

The work you can do alone: regulate your own end of the cycle. The cycle has six steps; you can change steps 1, 2, and 5. Practise self-soothing during activation so the protest does not need to fire as hot. Practise issuing bids in plain language rather than through protest. Practise tolerating the withdrawal without escalating. These changes alone often shift the cycle enough that the avoidant partner notices something has changed and becomes curious. The deep-dive on Self-Soothe Anxious Attachment is the protocol library for this.

The ceiling: a true two-person cycle eventually requires both people. If after three to six months of one-sided work nothing has shifted, and the avoidant partner is unwilling to engage with the framework, that is the question a therapist is for. The Check-In is a low-commitment way to put a third party in the room. The reverse case — the avoidant partner reads first — is symmetrical: the avoidant can do real work on steps 3, 4, and 6, and the same ceiling applies.

How long does it take to break the cycle?

The honest answer: it does not “break.” It loosens. The timeline above gives the realistic markers — three months for naming, six months for shortening, twelve months for de-centring, beyond a year for integration. Anyone who promises a 30-day cycle break is selling something. Anyone who tells you it takes ten years and individual therapy first is also probably wrong; most couples who commit to the work see meaningful change within a year. The middle of those two claims is where the evidence actually sits (Wiebe & Johnson, 2016).

What about fearful-avoidant or disorganised partners?

The pure anxious-avoidant pairing is the cleanest case. Fearful-avoidant partners run both strategies, often in fast alternation, which produces a more chaotic version of the same dynamic. The deep-dive on Fearful-Avoidant Push-Pull Dynamic covers the variant. The exercises here still apply but often need slower pacing and, more often, professional support.

Can the cycle re-emerge after years of work?

Yes — typically under acute stress (a new baby, a death, a job loss, an illness). The re-emergence is not failure. It is the original wiring becoming briefly accessible again under load. Couples who have done the work recognise the re-emergence quickly and apply the protocols faster.

What to try this week

If you’ve never named your cycle out loud to your partner: do that. The script is “I think we’re in our cycle. Can we talk about it not as right-and-wrong but as a pattern we’re both in?” Most couples find that naming alone shifts the next iteration.

If you’ve named it but it keeps running: pick one of the six protocols above. Run it for two weeks. The cycle won’t disappear, but the time-to-recovery should shorten noticeably.

If the cycle has become entrenched: book a Check-In, find an EFT therapist, or commit to one of the longer-arc programs at the deep-dive level. Self-work has a ceiling for this dynamic, and recognising the ceiling is itself part of the work.

Read deeper

• Inside each style: Dismissive-avoidant emotional needs , Anxious attachment triggers , Avoidant deactivation triggers, Fearful-avoidant push-pull dynamic
• In conflict: Pursue-withdraw loop , How avoidant attachers handle conflict , How attachment styles affect communication , Attachment styles and conflict resolution
• The exercises: Co-regulation for anxious-avoidant couples , Sue Johnson EFT exercises , Self-soothe anxious attachment , Help an anxious partner feel secure , How to communicate boundaries to avoidant , Build trust with dismissive-avoidant
• The longer arc: Build attachment security in 90 days , Earn secure attachment , Earned secure case studies
• The connections: Attachment Styles (diagnostic pillar), Communication pillar, Repair pillar, DIY Marriage Counseling

Sources

• Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61(2), 226–244.
• Greenberg, L. S., & Goldman, R. N. (2008). Emotion-Focused Couples Therapy: The Dynamics of Emotion, Love, and Power. APA.
• Johnson, S. M. (2004). The Practice of Emotionally Focused Couple Therapy: Creating Connection (2nd ed.). Brunner-Routledge.
• Johnson, S. M. (2008). Hold Me Tight: Seven Conversations for a Lifetime of Love. Little, Brown and Company.
• Levine, A., & Heller, R. (2010). Attached: The New Science of Adult Attachment and How It Can Help You Find — and Keep — Love. TarcherPerigee.
• Mikulincer, M., & Shaver, P. R. (2007). Attachment in Adulthood: Structure, Dynamics, and Change. Guilford Press.
• Simpson, J. A., & Rholes, W. S. (2017). Adult attachment, stress, and romantic relationships. Current Opinion in Psychology, 13, 19–24. DOI: 10.1016/j.copsyc.2016.04.006
• Tatkin, S. (2012). Wired for Love. New Harbinger Publications.
• Wiebe, S. A., & Johnson, S. M. (2016). A review of the research in emotionally focused therapy for couples. Family Process, 55(3), 390–407. DOI: 10.1111/famp.12229