The Anxious–Avoidant Trap: Why You Keep Dating Each Other (And What Actually Works)

It's 11:47 pm. You've sent two messages. He's read the first one and is "typing…" but nothing arrives. You can feel the conversation in your chest. He, on his side of the city, is staring at the screen and feeling a small, unmistakable pressure he can't quite explain — and so he puts the phone face-down and goes to make tea.

That, in one snapshot, is the anxious–avoidant trap. Two people who probably love each other, doing things that feel reasonable from the inside, that read to each other as exactly the worst thing the other person could be doing right now. Neither is the villain. Both are running an old script.

The pattern has had a thousand names. Couples therapists call it pursue-withdraw. Attachment researchers call it anxious–dismissive coupling. TikTok calls it the anxious-avoidant trap or the situationship spiral. The internet has spent the last few years discovering it as if it were new. It is not new. It has, in fact, been one of the most-studied dynamics in relationship science for nearly forty years, and the research has both bad news and good news. The bad news is the pairing is structurally over-represented in the dating pool, so you're going to keep meeting each other. The good news is the script is changeable — but the standard advice ("just communicate better") is roughly useless, and the moves that actually work are not intuitive.

This is the long version of what's happening, why, and what to actually do about it.

The Four Attachment Styles, in One Honest Paragraph Each

Most explanations of attachment theory start with a babies-in-a-room experiment from 1978 and end with a quiz. We're going to do the babies-in-a-room thing because the science actually matters here, but we'll keep it short.

The framework began with British psychiatrist John Bowlby, who in his 1969 book Attachment and Loss argued that humans are born wired to form attachments to caregivers, and that the quality of that early attachment shapes how the person handles closeness for the rest of their life. His collaborator Mary Ainsworth, a Canadian-American psychologist working at Johns Hopkins, then ran a famous lab procedure called the Strange Situation (published 1978), in which one-year-olds were briefly separated from and reunited with their mothers. Three patterns leapt out: secure, anxious-resistant, and avoidant. A fourth — disorganised — was added later by Main and Solomon in 1986.

In 1987, two psychologists at the University of Denver, Cindy Hazan and Phillip Shaver, published a now-classic paper in the Journal of Personality and Social Psychology showing that the same patterns described adult romantic relationships, not just toddler-caregiver ones. That's the bridge that turned attachment theory into the relationship lens it is today. Hazan and Shaver's 1987 paper also gave us the now-standard adult breakdown, with later refinement by Bartholomew and Horowitz (1991) splitting avoidance into two flavours.

The contemporary four-style picture, with rough population estimates summarised in Mikulincer and Shaver's Attachment in Adulthood (2007), looks like this:

Note the three things that aren't in those descriptions: a moral judgement, a fixed life sentence, and an excuse. Attachment styles are tendencies, not personalities. They're shaped by early experience and modified by every important relationship since. They predict probability of certain moves under stress; they don't determine the moves you actually make.

The anxious–avoidant trap is what happens when categories two and three meet on a dating app.

Why Anxious and Avoidant Keep Finding Each Other

If you've been single for any meaningful stretch of the last decade, you've probably had the experience of looking around and feeling like the entire dating pool was either anxious women and avoidant men, or anxious men and avoidant women, depending on who you were dating. The pattern feels like a conspiracy. It's actually structural.

In their 2010 book Attached, psychiatrist Amir Levine and writer Rachel Heller identify three reasons the pairing is over-represented:

1. Securely attached people pair off and stay paired. They handle conflict well, recover from setbacks, and don't generate the on-again-off-again drama that keeps a relationship in the pool. So the dating pool is, on any given Tuesday, disproportionately composed of insecurely attached singles. The remaining math means anxious-avoidant matchups happen more often than chance would predict.
2. Anxious people misread avoidant signals as desirable. The avoidant partner's emotional unavailability doesn't read as a warning to an anxious nervous system; it reads as mystery, challenge, not too easy. A securely attached prospect who texts back promptly and is unambiguously interested can feel "boring" or "too keen" by comparison. The contrast effect is real, and it's a bad guide.
3. Avoidant people misread anxious intensity as proof of their own value. A pursuing partner doesn't trigger an avoidant person's "this is exhausting" alarm at first; it triggers a "this person really wants me" feeling that, in someone with deactivated attachment, scratches a specific psychological itch. A secure prospect's calm interest, by contrast, doesn't satisfy that need at all, so the avoidant person tends to drift toward the anxious one.

The result is a dating market that selects, again and again, for exactly the pairing that's hardest to make work. It's not that you're broken. It's that the matching algorithm of romantic attraction is, for some people, broken.

The Four Phases of How It Plays Out

The trap doesn't show up on date one. It shows up in a predictable sequence — predictable enough that once you've named the phases, the relationship feels less like fate and more like a script you can edit.

The Gottman Institute's longitudinal research, summarised in The Seven Principles for Making Marriage Work (1999), identified pursue-withdraw as one of the strongest behavioural predictors of divorce. The closely related concept of emotional flooding — the physiological shutdown that happens during high-conflict moments — frequently rides on top of the same dynamic. The cycle isn't just unpleasant. Left untreated, it does measurable damage.

Why "Just Communicate Better" Doesn't Work

Generic relationship advice tends to reduce the trap to a communication problem. Tell each other what you need! Use "I" statements! Schedule time to talk! If you've ever been inside the cycle, you already know why this advice tends to fail: the trap isn't a misunderstanding. The trap is two attachment systems running their own incompatible defensive scripts, faster than language can catch them.

Specifically:

• Anxious pursuit isn't "asking better." When the anxious partner is activated, what comes out is rarely a clean request. It's protest behaviour — a term Levine and Heller popularised in Attached. Protest behaviour includes texting more frequently than feels natural, "testing" the partner with silences, escalating the stakes ("if you cared about me, you'd…"), withholding affection to provoke a response, and, in its more extreme form, threatening the relationship to force a reaction. Each of those moves is the attachment system trying to re-establish closeness, and each of them lands on the avoidant partner as a reason to withdraw further.
• "I need space" doesn't sound like what it is. When the avoidant partner says this, they mean "my nervous system is overwhelmed and I need to deactivate." What it sounds like to the anxious partner is "I am ending this relationship in slow motion." The same words trigger opposite interpretations in the two attachment systems. Telling the avoidant partner to use a different phrase doesn't help, because the underlying need is real and it has to be expressed somehow.
• The two systems don't share a goal in the moment. When activated, the anxious partner's nervous system is asking "are you with me?" The avoidant partner's nervous system is asking "can I have a moment to not be merged with someone?" These aren't reconcilable in the same five minutes. Standard couples advice that says "find a compromise" is asking both nervous systems to do something neither can do while activated.

This is exactly why Sue Johnson's Emotionally Focused Therapy (EFT), the most-validated couples therapy approach for attachment-driven conflict, doesn't start with communication skills. It starts with helping each partner identify the underlying attachment fear behind the surface move. Johnson's 2008 book Hold Me Tight describes the core therapeutic insight as recognising that attachment conflict isn't a logic problem; it's a felt-safety problem. You can't reason your way out. You have to renegotiate the safety first.

Four Things That Actually Move the Needle

Here, finally, is the part that's worth the read. None of these is fancy. They're the moves that, in our experience and in the research, actually interrupt the cycle.

1. The "concrete return time" rule.

When the avoidant partner needs space, the most effective thing they can do isn't shorten or eliminate the request — it's bookend it. Replace "I need some space" with "I need an hour, and I'll come back at 9." The anxious partner's nervous system isn't soothed by reassurance about feelings; it's soothed by predictability. A specific return time turns "you might be leaving" into "you'll be back at 9," and the activation drops sharply. The avoidant partner gets the deactivation they actually need; the anxious partner gets the certainty they actually need. This single change, used consistently, reduces protest behaviour faster than any conversation about it.

2. The anxious partner's pre-pursuit pause.

When the anxious nervous system is activated, the impulse is to move toward the partner: text, call, drive over, escalate. The intervention is a pause before pursuing. Not "don't pursue ever" — that suppresses a real need. The pause is between "feeling activated" and "doing the move," and during the pause the anxious partner does something else with the activation: a walk, three minutes of slow breathing, writing the unsent text in a notes app, calling a friend (specifically not the partner). The activation usually drops by half within ten minutes. The thing the anxious partner then does with the partner is dramatically more in proportion to what's actually happening, rather than to what their attachment system fears is happening.

3. The avoidant partner's small unprompted check-ins.

Avoidant partners often think their job is to respond to the anxious partner's bids. The intervention is the opposite: the avoidant partner sends small, low-effort, unprompted signals. "Thinking of you." "How's your morning." A photo of something funny. The investment is tiny — twenty seconds, no profundity required. The reason this works is that anxious nervous systems run on signal frequency, not signal magnitude. Three two-line messages over a day calm the system more than one long love letter. Sending these before the anxious partner reaches for them prevents the activation entirely; sending them only in response is too late.

4. A shared vocabulary for the moment.

Couples that successfully exit the cycle almost always develop a private code for the cycle. "I'm in protest mode." "I'm deactivating." "I think we're doing the thing." The phrase doesn't need to be elegant; it needs to be shared, and it needs to be available to both of you in the moment without explanation. The shorthand short-circuits the cycle by collapsing minutes of escalation into a single labelled state. You can disagree later about whose move started it. In the moment, you only need to name it together. EFT therapists often call this developing a map of the dance, and the map alone, even before behavioural change, frequently reduces the frequency of full-blown cycles by half.

Earned Secure: Attachment Style Is Not a Sentence

One of the more demoralising assumptions floating around the internet's attachment discourse is that you're stuck with whatever style you have. The research disagrees. Mikulincer and Shaver's Attachment in Adulthood, the standard academic text on adult attachment, surveys longitudinal studies and concludes that adult attachment is moderately stable, with somewhere around 25–30% of people showing meaningful style change over multi-year periods. Stability is the default; change is real and not rare.

The clinical literature has a name for the move from insecure toward secure: earned secure attachment, a concept that emerged from Mary Main's work on the Adult Attachment Interview at UC Berkeley starting in the 1980s. Earned-secure adults grew up insecurely attached but, through some combination of stable adult relationships, deep reflection on their early caregiving history, and (often) therapy, developed adult functioning that's indistinguishable from people who were securely attached as children.

The three known routes to earned secure are:

1. A long-term relationship with a securely attached partner, in which the secure partner doesn't get pulled into the cycle and consistently models a different response. This is the route the dating-pool math makes hardest, but not impossible.
2. Therapy, especially attachment-focused therapy or EFT, which gives you a structured way to identify your style's defensive moves and run experiments in safer settings.
3. Sustained reflection on the early relationships that shaped your style — typically with a therapist, sometimes alone, often via writing. The pattern Main observed was not "people who had perfect childhoods become secure"; it was "people who can tell a coherent story about their imperfect childhoods become secure." Reflection, not retroactive perfection, is the mechanism.

The last point is worth sitting with. Earned secure isn't a fix-your-past project. It's a make-sense-of-your-past project. The childhoods don't change. The narrative around them does, and that turns out to be most of what's needed.

What This Looks Like for Your Specific Relationship

The pattern is general. Your relationship is specific. The most useful thing two people in an anxious-avoidant pairing can do, after reading something like this, is slow down and have an explicit conversation about what each of you sees in the trap, without using it as ammunition.

This is, unsurprisingly, what we built Heart to Heart for. The mode hands you turn-by-turn questions designed to surface exactly the kind of attachment-level information the cycle prevents you from sharing. Questions like "What's a small thing I do that lands harder than I realise?" or "What do you wish I'd notice that you don't say out loud?" are, structurally, what an EFT therapist would prompt for in session. The advantage of the structured-prompt format is that neither partner has to be the one who brought it up, neither has to feel exposed for asking, and the cadence of turn-taking interrupts the activated escalation that normally derails this conversation. We've also written about why structured questions work better than "let's talk" for couples who find the unstructured version disastrous.

If you want a lighter way in — fewer attachment-level questions, more "I had no idea you thought that" — our list of 80 "how well do you know me" questions covers similar ground at a softer entrance angle. And the 36 Questions That Lead to Love sequence, originally validated in Aron et al.'s 1997 study, works well as a one-evening exercise that bypasses the cycle by being too structured for either partner to derail.

The Honest Closing Note

The anxious-avoidant trap is real, well-documented, and probably more common in your social circle than you realise. It's also not a verdict on either of the people stuck in it. Two anxious-avoidant partners with shared awareness of the pattern, mutual willingness to interrupt their own scripts, and a few low-effort changes (concrete return times, pre-pursuit pauses, unprompted check-ins, shared vocabulary) can absolutely build a stable, warm relationship. Two people running the cycle without naming it can stay together for decades and never escape it.

The single biggest predictor of which of those happens is whether both partners can talk about the cycle as a cycle — a third thing, separate from each of them, that they're collaborating to interrupt — rather than as a list of grievances against each other. That conversation is hard to start cold. It's much easier with a structure. That's most of why we build the structures we do.

Frequently Asked

Can two anxious people or two avoidant people work?

Two anxious partners often function well early on (lots of mutual reassurance, high contact frequency) but can struggle when both nervous systems activate simultaneously and there's no one in the room to be the calm one. Two avoidant partners often coexist comfortably for years with low conflict, but the relationship can drift toward roommate-mode without either partner noticing because neither is generating the bids that would surface the drift. Both pairings work; both have specific failure modes. Awareness of those failure modes, more than the pairing itself, predicts long-term outcome.

How do I know my own attachment style?

The most-validated free tool is the Experiences in Close Relationships – Revised (ECR-R) questionnaire developed by Fraley, Waller and Brennan (2000), used in academic research. The popular online quizzes from Levine and Heller's Attached are decent first-pass orienting tools but are not psychometrically validated. The single most useful self-test is non-quantitative: when conflict starts, do you tend to move toward the partner (anxious) or away from the partner (avoidant)? Most people know within five seconds. (If you've been using MBTI to make sense of your relationship, our companion piece MBTI vs Attachment Style walks through why one of these frameworks predicts couples and the other doesn't — and how to use both honestly.)

Should I tell my partner their attachment style?

Telling another adult what their attachment style is, unprompted, almost never lands well. It frames you as the diagnostician and them as the patient. The better move is to share your own first ("I think I tend to be anxious in relationships, and here's what that looks like for me") and let them volunteer where they recognise themselves. The asymmetric version — diagnosing them but not yourself — is both annoying to be on the receiving end of and frequently inaccurate, because the partner's behaviour you observe is the dance of both attachment systems, not theirs in isolation.

Is "anxious-avoidant" the same as toxic?

No, and conflating them is a TikTok-era misuse. "Toxic" describes a relationship in which one or both partners behave in ways that consistently violate the other's wellbeing, often with intent or pattern. "Anxious-avoidant" describes a structural mismatch in attachment systems that can produce painful dynamics even when both partners are good people doing their best. Plenty of toxic relationships have nothing to do with attachment style. Plenty of anxious-avoidant relationships are warm, decent, and worth saving. The categories are different and shouldn't be flattened.

How long does it take to break the cycle?

Realistic ranges from couples therapy outcome research are roughly 12–20 weeks for noticeable change with weekly EFT sessions, and longer for stable transformation. Without therapy, with two motivated and aware partners using the kinds of interventions described above, change is real but slower — usually months for the cycle frequency to halve, longer for the underlying activation pattern to shift. The fastest results come from the avoidant partner unilaterally implementing concrete return times and small unprompted check-ins, which often surprises both partners by how much it shifts the relationship's overall climate within weeks.

Two attachment systems, running incompatible scripts, can still build something that lasts. The script is older than either of you. It's not stronger than the two of you choosing, on a Tuesday, to interrupt it.