Few ideas have escaped the academic literature as thoroughly as attachment theory. Terms that began life describing the behaviour of one-year-olds in a Baltimore laboratory now circulate freely in relationship advice, therapy waiting rooms and arguments between couples at two in the morning. Some of that translation is faithful. A good deal of it is not. It is worth going back to what the original research actually observed, and what it can honestly be said to show.
A doctor with a question about separated children
John Bowlby was a British psychoanalyst who spent the 1930s and 1940s working with children who had been institutionalised, evacuated or otherwise separated from their parents. The prevailing psychoanalytic view held that an infant's bond to its mother was essentially derivative: the baby loves whoever supplies the food. Bowlby did not believe it, partly because the children in front of him grieved in ways that had nothing to do with feeding.
He went looking outside his own discipline for a better account and found it in ethology, the study of animal behaviour. Konrad Lorenz's goslings imprinted on the first moving thing they saw. Harry Harlow's infant rhesus monkeys, given a choice between a wire mother that dispensed milk and a cloth mother that dispensed nothing but contact, spent their time clinging to the cloth. Bowlby drew the conclusion that seeking closeness to a protective adult is an evolved behavioural system in its own right, as basic as hunger, because for most of human history a small child left alone did not survive.
His second idea has aged even better. Children, he proposed, build what he called internal working models: rough expectations, assembled from repeated experience, about whether distress brings comfort and whether other people can generally be relied upon. Those expectations are not memories or beliefs so much as defaults, and they carry forward.
Twenty minutes in a room with toys
Bowlby supplied the theory. Mary Ainsworth supplied the evidence, and she did it by inventing one of the most productive procedures in developmental psychology.
Ainsworth had spent time observing mothers and infants in Uganda before launching a study in Baltimore in 1963, following twenty-six families with monthly home visits through the child's first year. At the end she brought each pair into the laboratory for a short, carefully choreographed sequence known as the Strange Situation. Over about twenty minutes, a toddler of twelve to eighteen months is introduced to an unfamiliar room, an unfamiliar adult arrives, the mother leaves twice and returns twice. The point is not the separation. The point is the reunion, and specifically what the child does with the parent when upset.
Three patterns emerged consistently. Most children protested when the parent left, went to her on her return, were comforted and went back to playing; Ainsworth called this secure. A second group appeared indifferent, showing little distress and ignoring or turning away at reunion, although physiological measures later suggested they were more agitated than they looked; this is avoidant. A third group became extremely distressed and then, on reunion, could not settle, seeking contact while resisting it; this is resistant or ambivalent. Two decades later Mary Main and Judith Solomon added a fourth category, disorganised, for children whose behaviour lacked any coherent strategy at all, often in families with frightening or unresolved histories.
Crucially, Ainsworth could partly predict which pattern a child would show from the home visits: babies whose signals had been read promptly and accurately over the preceding year tended to be the ones who used the parent as a base for exploring and a refuge when frightened.
From the nursery to romance
For twenty years this remained developmental psychology. Then, in 1987, Cindy Hazan and Phillip Shaver published a paper proposing that adult romantic love is itself an attachment process, running on the same machinery. To gather data, they did something delightfully unorthodox: they printed a short questionnaire in a Denver newspaper, offering three paragraph-length descriptions of how a person typically feels in close relationships, and asked readers which one fitted them.
The distribution of answers looked strikingly like the distribution of infant patterns. Roughly three in five picked the description of comfort with closeness and dependence. The rest split between discomfort with intimacy and a persistent worry that partners do not care enough. Their paper set off an entire research field within social psychology, one that now runs largely independently of developmental work.
The four adult styles, plainly
The version that circulates today owes most to Kim Bartholomew and Leonard Horowitz, who in 1991 organised adult attachment around two underlying questions: roughly, do I expect to be worth caring about, and do I expect other people to be available? Crossing those gives four familiar positions.
Secure describes people who are reasonably comfortable depending on a partner and being depended on, who can ask for reassurance without a production and accept it when it arrives. Anxious or preoccupied describes a strong desire for closeness paired with a live worry that it will be withdrawn, which tends to produce vigilance, frequent checking and a sensitivity to small shifts in a partner's warmth. Dismissing-avoidant describes a preference for self-sufficiency, a tendency to downplay the importance of relationships and discomfort when someone gets close enough to depend on. Fearful-avoidant, sometimes called disorganised, describes wanting closeness and fearing it at once, so that approach and withdrawal alternate.
One correction is worth making immediately. Researchers themselves have largely stopped treating these as four types. Since the late 1990s, the standard measures score two continuous dimensions, attachment-related anxiety and attachment-related avoidance, and a style is simply a region on that map. Almost nobody sits squarely in a corner. Being told you are "an anxious attacher" compresses two numbers into a label, and loses most of the information doing it.
What the research does not claim
Three overstatements are worth naming. The first is determinism. The link between how you were parented and how you conduct your relationships at thirty-five is real but modest; longitudinal studies tracking people from infancy into adulthood find correlations that are meaningfully above zero and nowhere near strong enough to predict an individual. Your early years load the dice. They do not throw them.
The second is that a style is a fixed property of a person. It is better understood as relationship-specific. Someone can be reliably secure with an old friend and anxious with a particular partner, and the pattern that shows up depends heavily on who is on the other side of it.
The third is that attachment explains everything about a relationship. It does not touch compatibility, circumstance, money, timing, or the ordinary question of whether two people enjoy each other's company. It describes one system among many: the one that governs what you do when you are frightened and want someone.
Attachment moves
The most encouraging finding is the least publicised. Attachment patterns change. Researchers use the phrase earned security for people who arrive in adulthood with a difficult early history and nonetheless develop the expectations of someone who was well looked after. It happens through long relationships with reliable people, through becoming a parent, through therapy, and sometimes simply through accumulating enough evidence that the old defaults are wrong.
That is the useful takeaway, and it is more modest than the popular version. Attachment vocabulary is a decent instrument for noticing your own reflexes under stress: the impulse to send a fourth message, or to go quiet for two days. Noticing is most of the work. What you do next is not written anywhere in the theory.
