Attachment styles: the four patterns we learn closeness by, and how to spot your own

“I don’t understand myself. When everything is fine, I want him closer. When he actually is close, I want to leave the room.”

A woman told me this convinced there was something badly wrong with her. There wasn’t. She had described something that has a name and a history, something researchers have been measuring for sixty years: that all of us carry a default strategy for handling closeness when we feel uncertain. That strategy once paid off. Today it often doesn’t, yet it still switches itself on before we manage to think anything through.

This piece is an overview: where the idea of attachment came from, which four patterns the literature describes, how they show up in adult life and — the part I find most important — why four boxes are a useful shortcut and a poor truth. I have written elsewhere about how a pattern repeats across relationships, in the post on why we keep ending up in the same place in relationships; I won’t go over that ground again here.

Where the idea of attachment came from

In the 1950s John Bowlby worked with children who had been separated from their parents for long stretches — in hospitals, institutions, after the war. The explanation of the day was that a child attaches to whoever feeds it, and that too much comforting spoils it. Bowlby argued something else: that seeking closeness when afraid is a separate, innate system, as basic as hunger. A child does not cry to manipulate. It cries because its nervous system has registered that it is alone, and because across evolution a child alone was a child dead.

The key idea that came out of this is the secure base. An available adult is not there to keep the child near, but so that the child can move away. When there is somewhere to come back to, the child goes further. It is a paradox that repeats almost unchanged in adulthood: people with reliable closeness are as a rule more independent, not less.

A pattern is not character, it is strategy

Before the four types, I have to say the thing that popular articles most often leave out. An attachment pattern is not a personality trait and not a label for a person. It is a strategy — the best available answer to the environment in which someone learned what closeness brings.

If the adult’s response was reliable, it paid to show distress directly. If the response was unpredictable — sometimes warm, sometimes absent — it paid to show it louder and for longer, because only then was it noticed at all. If closeness meant rejection or discomfort, it paid to quiet the distress and manage alone. None of these three routes is wrong. All three are intelligent. The trouble starts only when the strategy stays and the circumstances change.

So in my room I never say, “You’re the anxious type.” I say, “What happens inside you when he doesn’t reply for three hours?” That question goes further than any label.

The four patterns, briefly

In the 1970s Mary Ainsworth designed a procedure known as the strange situation: in a furnished room, a one-year-old is briefly left by the mother, who then returns. The interesting part was not the leaving but the return. Children who let themselves be comforted and then went back to play were described as securely attached. Children who settled with difficulty and both sought contact and fought it off were called ambivalent, or anxious. Children who turned to the toys on the reunion and looked entirely unbothered — though their heart rate said otherwise — were called avoidant.

The fourth pattern came later. In the 1980s Mary Main and Judith Solomon reviewed recordings that coders could not place anywhere: children stopped mid-approach, rocked, approached backwards, froze for a moment. They named it the disorganised pattern, because what was missing in it was precisely a strategy — the child had no single plan for what to do with its own fear.

Secure attachment: what it actually means

In the literature security is not measured by how calm someone feels, but by what they do when they are not calm. A securely attached person shows distress early enough, says it directly enough, and can also let it go once an answer arrives. They are not naturally more placid. They simply have less work to do translating a need into something more acceptable.

In a couple this sounds disarmingly simple: “Today was hard, I need you to be around for a while.” For many people that sentence is unsayable. Not because it is difficult, but because in some earlier life such a sentence carried a price. When someone learns it again, what often happens is not that conflicts become fewer, but that they become shorter.

I also want to break the notion that securely attached people are the ones who had a lovely childhood. The samples contain people from very different families. What recurs more often is not the absence of trouble, but that there was at least one person around the child who was responsive often enough and predictable enough to be counted on. Not perfect. Reliable enough that the child could work out what would happen when they came for help.

The anxious pattern in outline

I will be deliberately brief here, because this pattern deserves a post of its own. In short: the attachment system is set sensitively and fires quickly. Silence, a change in tone, a shorter message than usual — all of it goes through a filter that reads threat into ambiguity. Then comes protest: calling, explaining, checking, sometimes reproach. The purpose of the behaviour is not control, however it looks from outside. The purpose is contact.

What these people most often hear is advice to be less needy. That almost never works, because it only adds shame. More useful than suppression is understanding what anxiety does in the body, and practising how to move from rumination to a clear request.

The avoidant pattern in outline

Equally brief. This person is not cold. Physiological measures show their body does react during conflict, sometimes more strongly than that of the partner who is doing the talking — they simply do not show it. They learned that expressing distress does not lead to relief but to discomfort, so the shortest route is through withdrawal: silence, work, errands, the appearance of calm.

The most common misreading is that this person does not care. As a rule they do care, the route to words is just longer. So in therapy I never start by getting them to talk more; I start with making closeness physically bearable.

The fourth pattern, which is not just the first three mixed

The disorganised pattern — in adults usually described as fearful-avoidant — is not simply “anxious and avoidant at once”. What makes it distinct is that the system meant to carry a frightened person towards another, and the system that pulls them away from danger, both fire at the same address. That is not a comfortable blend of two strategies but the absence of one.

Because this pattern is often bound up with what the body carries forward, many people recognise it through physical rather than mental signs — I wrote about this in the post on trauma that stays in the body. It deserves its own treatment and its own therapist, not an online quiz.

Why the boxes mislead

Since the late 1990s adult attachment has mostly been measured not as four types but along two continuous axes: how much anxiety there is about being abandoned, and how much avoidance of closeness. Each of us has a value on both. The four types are really the four corners of that space, and most people sit somewhere in the middle, which is why everyone recognises a bit of themselves in every description.

This is not pedantry. It has a practical consequence: if you understand yourself as “the anxious type”, you have drawn a line the data does not contain. If you understand that you score high on the anxiety axis and that the score drops in some relationships, you have asked a question that can be investigated. And a pattern belongs to a particular relationship: the same person can be secure with one friend and not with a partner.

Where the pattern shows in daily life

In my office I see it in small things. Who writes first after an argument. What someone does when a partner says “we’ll talk about it later”. Whether a person says they are disappointed or decides it isn’t worth mentioning. How they take a compliment. How long it takes, after an unanswered call, for a story to form.

It also shows where we don’t expect it: in setting boundaries, where the anxious side sets none for fear of loss, while the avoidant side sets them so high that nobody gets close. And in shame, which often stands just behind both.

Where this framework does not hold

Now the part that popular accounts almost never include. Attachment theory is a good description, but not an explanation for everything.

First: a pattern does not predict fate. Meta-analyses of transmission across generations show an association that is moderate, not determined — a large share of the variance remains unexplained. If someone forecasts your relationship from your childhood, they are going further than the data allows.

Second: much of what we attribute to attachment may be something else. Chronic illness, exhaustion, depression, shift work, poverty and insecure housing do to a relationship exactly what we like to call avoidance. Before I explain a pattern by childhood, I always ask how much a person sleeps and how many hours they work.

Third: where there is violence or coercive control in a relationship, the language of attachment becomes dangerous. “We both have our pattern” sounds balanced, and that balance obscures the fact that one of the two is not safe. That is not a cycle, it is a matter of safety, and it is a different conversation.

Fourth: the categories arose in particular cultures and particular family arrangements. A childhood in which the evenings belong to a grandmother rather than the mother is not thereby insecure; a procedure developed elsewhere simply measures it differently.

What I do with this in therapy

Nothing spectacular and nothing fast. First we describe what happens in the body in the first two seconds after a trigger, because everything else is decided there. Then we look for when the pattern does not fire — there are always more such moments than a person notices. Then we practise sentences that state a need plainly. And we work as we go with the resistance this raises, because the strategy was once protection and will not step aside out of politeness.

If you are considering exploring this with someone, the post on how to choose a therapist may help. With attachment patterns, the working relationship itself is half the work.

What to remember

The four patterns are not four kinds of people but four strategies for handling closeness when it feels uncertain. Each arose as a sensible answer to some environment. They are measured along two continuous axes rather than as boxes, and they differ across relationships. What you bring from childhood explains part of the story, not all of it. And the most useful question is not “which type am I”, but “what happens in me when I feel alone, and what do I do in that moment”.