The Attachment You Chose

John Bowlby did not arrive at attachment through the couch. He was a British psychiatrist and psychoanalyst trained under Melanie Klein, and he broke with that tradition to argue from ethology and evolution that the bond between an infant and its caregiver is a hard-wired behavioural system, not a by-product of being fed. The system has a trigger and a goal. It activates under fear, threat or distress. Its goal is proximity to a stronger, wiser figure. Reaching that proximity shuts it down and sends the child back out to explore. Then came the move that carried everything forward. The internal working model. A representation laid down early, holding two questions inside it. Am I worthy of care. Will the other person be there when I need them. Bowlby put the full statement in three volumes with Basic Books. Attachment in 1969. Separation in 1973. Loss in 1980. His own summary of how far the claim reaches sits on page 129. Attachment behaviour characterises humans from the cradle to the grave. I read that line and recognised something I had watched in company after company. The people who held up under pressure were not the toughest in the room. They were the ones who knew, without checking, that somebody was still on the other end of the line.

Mary Ainsworth turned the theory into something measurable. Her Strange Situation runs about twenty minutes. Mother and infant in a room. A stranger enters. The mother leaves. The mother returns. What the child does at the reunion is the data. Three patterns fell out of it. Secure. Anxious-resistant, also called ambivalent. Avoidant. A fourth was added later by her colleague Mary Main, disorganized, for children whose strategy collapses into contradiction and fear. The distribution has held for decades. About 60% of children behave securely. About 20% avoidantly. About 20% anxious-resistant. The canonical write-up is Ainsworth, Blehar, Waters and Wall, Patterns of Attachment, 1978. She also gave us the phrase that does the most work in adult life. The secure base. The caregiver as the platform you push off from to explore the world, and the harbour you run back to when the weather turns. Two functions. Exploration and return. Most of the professional wreckage I have seen up close was a man who was very good at one of them and had never once practised the other.

Cindy Hazan and Phillip Shaver argued in 1987 that romantic love is an attachment process. The paper is Romantic Love Conceptualized as an Attachment Process, Journal of Personality and Social Psychology, volume 52, number 3, pages 511 to 524. The claim was structural, not poetic. The bond between adult partners runs on the same motivational system as the infant and caregiver bond. Proximity seeking. Distress at separation. A safe haven and a secure base. Mutual fascination. They did not run it in a laboratory. They ran it as a newspaper love quiz and got 620 respondents. 56% secure. 25% avoidant. 19% anxious or ambivalent. Set that beside Ainsworth and the finding lands. Infant attachment: about 60% secure, about 20% avoidant, about 20% anxious-resistant. Adult romance: 56, 25, 19. The same shape, decades later, in adults filling in a form in a newspaper. The paper has been cited more than 6,000 times. It is one of the most referenced works in relationship science, for one reason. It closed the gap between a child's first bond and a forty-year-old marriage.

Kim Bartholomew and Leonard Horowitz split avoidance in two in 1991. Cross a model of the self with a model of other people and four styles fall out. Secure. Preoccupied, which is the anxious one. Dismissive-avoidant. Fearful-avoidant. Then the field stopped believing in boxes. Brennan, Clark and Shaver showed in 1998 that the self-report measures hang together along two orthogonal dimensions. Attachment anxiety, which is fear of abandonment and a hair-trigger on rejection. Attachment avoidance, which is discomfort with closeness and compulsive self-reliance. They built the Experiences in Close Relationships scale on those two axes. Fraley, Waller and Brennan refined it in 2000 into the ECR-R, which is still the standard instrument. Here is the practical consequence in one line. There are no attachment types. There are two dials. Everybody sits somewhere on both of them. And almost nobody sits at the same point in a good year and a bad one.

The people who held up under pressure were not the toughest in the room. They were the ones who knew, without checking, that somebody was still on the other end.

John Simpson and William Rholes published the model that explains why this matters more in a bad year than in a good one. The Attachment Diathesis-Stress Process Model, Current Opinion in Psychology, volume 13, pages 19 to 24. Anxious and avoidant orientations are not disorders. They are latent vulnerabilities that decide how a person thinks, feels and behaves when stress actually arrives in the relationship. Two strategies run from there. The highly anxious hyperactivate. They intensify distress, monitor the partner for signs of withdrawal, and seek reassurance in a volume that can produce the withdrawal they were watching for. The highly avoidant deactivate. They suppress awareness of their own distress, withdraw into self-reliance, and do not ask for help even when they need it. Both strategies look composed in a good year. Neither is visible when nothing is at stake. Put the bond under real load and the strategy runs the room. That is the part worth keeping. Calm in a comfortable quarter tells you nothing. The pattern becomes legible only under pressure, and by then most people are reading the behaviour instead of the wiring.

The same review states the counter-case in plain language. Working models can change over time in response to new experiences or events that strongly contradict them. That sentence is the empirical floor under earned security. A relationship consistent enough, and different enough, from the one that wrote the model can rewrite the model. The mechanism has a name. Partner buffering. A responsive partner measurably softens the defences of both the anxious and the avoidant during conflict. Chris Fraley mapped where the field is going in the Annual Review of Psychology in 2019, volume 70, pages 401 to 422. The target of attachment shifts across a life, from parent to peer to romantic partner, while the system itself persists. The live debate is whether insecurity is a defect at all or a context-dependent adaptation. My own read, priced honestly. The probability that a person's attachment strategy is fixed for life is close to zero. The probability that the people around them decide which one they practise is very high. I signed a EUR 20 million personal guarantee. On Christmas Day 2008 the company folded. That is the number in the record, and it is the one everybody asks about. Nobody audits the second number. How available you actually were to the five people who were counting on you while you were busy being necessary somewhere else.