Long before couples arrive in a therapy room describing themselves as “anxious” or “avoidant,” they are usually describing something much simpler: a pattern that repeats. One partner reaches, the other retreats. One needs reassurance nightly; the other needs space to think. Neither is wrong — but neither quite understands why the same argument keeps finding its way back to the table. Attachment theory, first developed by the British psychiatrist and psychoanalyst John Bowlby, offers one of the clearest maps we have for understanding why.
Bowlby’s Starting Point
Working in London in the years following the Second World War, Bowlby was troubled by what he observed in children separated from their caregivers — in hospitals, in wartime evacuations, in institutional care. He concluded that the bond between infant and caregiver was not simply a matter of feeding and comfort, as the psychoanalytic orthodoxy of the time suggested, but a biologically rooted survival system in its own right. A child is wired to seek proximity to a trusted adult, and the quality of the caregiver’s response — consistent and attuned, or unpredictable and unavailable — becomes the template, or what Bowlby called the internal working model, through which that child will later read every close relationship it enters.
This is the idea that makes attachment theory so useful in adult relationship work: the internal working model does not stay behind in childhood. It becomes the lens through which we interpret a partner’s silence, a delayed reply to a text, a raised voice, or a hand withdrawn mid-sentence.
The Four Attachment Styles
Bowlby’s theory was later tested and extended by his colleague Mary Ainsworth, whose “Strange Situation” studies identified distinct patterns of infant response to separation and reunion. These patterns map closely onto what we now describe, in adult relational language, as four broad attachment styles.
Secure
Roughly the outcome, Ainsworth found, of consistently responsive caregiving — a caregiver who was reliably available, attuned, and able to soothe distress without being frightened or overwhelmed by it themselves. The child learns two things simultaneously: that others can be trusted to respond, and that their own needs are legitimate enough to voice. Neither lesson has to be relearned in adulthood; both are simply assumed.
A securely attached adult tends to trust that others will be available and responsive, and that they themselves are worthy of care. In relationships, this often looks unremarkable — which is precisely the point. Conflict is navigated without catastrophising; closeness and independence are both tolerated without one threatening the other; a partner’s bad mood is read as a bad mood, not as evidence of abandonment. Securely attached people are generally comfortable asking for support and equally comfortable being alone, can voice a complaint without it becoming a referendum on the relationship, and tend to assume good faith when a partner’s behaviour is ambiguous rather than defaulting to the worst-case reading. It is worth saying plainly: this is not the same as an easy or conflict-free relationship. It is the capacity to move through conflict without the bond itself feeling at risk.
Anxious-Preoccupied
Here, closeness is longed for but rarely felt as secure enough. This pattern often develops where caregiving was inconsistent — warm and attentive at times, unavailable or preoccupied at others — teaching the child that connection cannot be relied upon passively; it has to be actively monitored and pursued, or it may disappear. Ainsworth’s infants with this pattern were noted to protest intensely at separation and yet remain difficult to settle on reunion, unable to fully trust that the caregiver’s return meant safety.
In adulthood this can surface as heightened sensitivity to a partner’s tone, facial expression, or availability; a need for frequent reassurance that can feel, to the anxious partner, like simple honesty about what they need, and to the other partner, like it is never quite enough. Ordinary distance — a quiet evening, a slow reply, a partner absorbed in their own thoughts — is often interpreted not as neutral but as a sign of withdrawal or rejection, which can prompt protest behaviour: repeated calls or texts, rehearsing the relationship’s problems, or picking a fight simply to re-establish contact. Underneath this is rarely a wish to control a partner; it is a nervous system that has learned, correctly at the time, that connection had to be fought for.
Avoidant-Dismissive
Where a child’s bids for comfort were consistently rebuffed, discouraged, or met with mild disapproval — caregiving that may have been perfectly adequate practically, but emotionally under-responsive — self-reliance becomes the safer strategy. Rather than protesting separation, as the anxious infant does, Ainsworth observed that these children appeared outwardly unbothered by a caregiver’s absence and return; internally, however, physiological measures showed their stress response was very much activated. The strategy is not an absence of need, but a learned suppression of its outward expression, because expressing it previously brought no useful response.
The adult version often prizes independence, feels genuinely crowded by demands for emotional closeness, and may minimise, deflect, or intellectualise feelings rather than voice them directly — sometimes appearing entirely calm in moments that would distress most people, or professing not to have strong feelings about a situation that clearly affects them. Partners often experience this style as distant, self-contained, or unreachable, and can mistake the composure for indifference. It is more accurately understood as a deep-seated protection against disappointment: if closeness was never reliably met with care, the safest position is not to need it too visibly.
Disorganised / Fearful-Avoidant
Identified by Ainsworth’s successor Mary Main, this pattern typically arises where the caregiver was, at different times, both the source of comfort and the source of fear — often in the context of frightening, chaotic, unpredictable, or frightened caregiving, and frequently associated with unresolved trauma or loss in the caregiver’s own history. The child faces an unsolvable dilemma: the person they must turn to for safety is also the person they need protection from, and the infants in Main’s studies showed exactly that contradiction — approaching the caregiver on reunion and then freezing, turning away, or moving in strange, incomplete gestures rather than settling.
The adult expression can look contradictory from the outside: a simultaneous longing for closeness and dread of it, oscillating between pursuing a partner and pushing them away, sometimes within the same conversation. This is often the pattern behind relationships that feel intense and unstable — pulling a partner in, then feeling engulfed and creating distance, then fearing the distance has caused real damage and pulling back in again. Because it combines features of both the anxious and avoidant strategies without the organising logic of either, it is usually the most distressing pattern to live inside, for both partners, and the one most likely to benefit from working with a therapist rather than being worked through alone.
How This Translates Into Our Relational Patterns
What makes attachment theory so clinically valuable is not the labels themselves but what they explain about the choreography couples fall into. A common configuration in the consulting room is the anxious partner pursuing and the avoidant partner withdrawing — each response intensifying the other. The pursuer’s protest (“why won’t you talk to me?”) is, at root, a bid for reassurance; the withdrawer’s silence is, at root, a bid for safety. Read only at face value, the exchange looks like incompatibility. Read through an attachment lens, it looks like two nervous systems doing exactly what they learned to do to survive closeness.
A few patterns I see regularly in couple work:
- Escalating conflict that seems disproportionate to its trigger — often a sign that an old attachment wound, not the current disagreement, has been activated.
- One partner experienced as “too much” and the other as “too distant,” when both are, in fact, regulating the same underlying fear of disconnection in opposite directions.
- Difficulty repairing after rupture — not because the couple doesn’t care, but because neither partner’s working model included a reliable experience of repair.
- A partner who appears entirely self-sufficient in daily life becoming disproportionately dysregulated around a specific relational trigger, such as lateness or a delayed reply.
- Discomfort naming feelings aloud, and a marked hesitancy around commitment — moving in together, engagement, a proposal of marriage — even within a relationship the partner genuinely values.
This last pattern is one of the most common presentations of avoidant attachment in couple work, and often the most misread. From the outside, hesitancy about marriage or reluctance to talk about feelings can look like ambivalence about the relationship itself, and the more anxious partner will frequently experience it that way — as evidence they are not loved enough, or not enough. But for the avoidant partner, the discomfort is rarely about the relationship’s worth. Committing formally, or naming an emotion out loud, converts something that has felt safe precisely because it was undefined into something fixed, visible, and — in the internal working model — something that can therefore be lost. The instinct to deflect, delay, or intellectualise (“we don’t need a piece of paper,” “I’m just not a talker”) is usually a protective strategy against that exposure, not a verdict on the partner or the relationship. Naming this distinction explicitly in the room — separating the fear of commitment from the depth of the commitment actually felt — is often where couples begin to find their way out of the pursue-withdraw cycle.
None of these patterns are fixed diagnoses. Bowlby himself, and researchers since, have been clear that attachment style is not a life sentence — it is a working model, and working models can be revised. This is sometimes described as earned security: through a stable relationship, through therapy, or through a corrective relational experience, an anxious or avoidant pattern can shift meaningfully towards security.
Why This Matters in the Therapy Room
In my own clinical work, I find attachment theory most useful not as a way of categorising a couple, but as a way of slowing a couple down — helping each partner see their own reaction as a strategy with a history, rather than as an indictment of the relationship itself. This sits comfortably alongside an existential-phenomenological approach: attachment theory describes the pattern, but it is the lived, present-moment experience of the couple in the room — how it actually feels to reach and not be met, or to be reached for and feel unable to respond — that allows the pattern to be understood, and eventually, to change.
Understanding one’s own attachment style, and one’s partner’s, rarely resolves a relationship on its own. But it very often does something just as important: it replaces “you always” and “you never” with a more compassionate and more accurate question — what is this pattern trying to protect, and what would it take for both of us to feel safe enough to put it down?
