Take our Relationship Attachment Test

Find a therapist to strengthen relationships

Attachment style may be less an emotional habit than a predictive model a developing mind settles into.

Two axes matter: the content of predictions about a caregiver and how precise—how reliable—they are.

The intrauterine environment is a first statistical world, shaping expectation before a child draws a breath.

It is tempting to describe attachment styles as emotional residue—the good-enough or not-good-enough caregiving of the early years, laid down and carried forward. That's not wrong, only incomplete.

A new open-access paper in Frontiers in Psychology by Erica Santaguida, Giuseppe Pagnoni, and Massimo Bergamasco (2026) adds a layer of useful specificity. Attachment style, the authors argue, is what happens when a developing mind approximates a predictive model of its world—and the deciding variable is not only what the child comes to predict about a caregiver, but how reliable those predictions turn out to be. With responsive caregivers, the child—in my view—learns she can also shape the world, reading it in how the caregiver bends in response to her efforts to predict and correct. In one version, the world can adjust to her. In another, it is unyielding.

A Hall of Smart Mirrors

Content and precision are distinct axes, and separating them brings the developmental picture into focus.

Secure attachment, here, is a model with high precision and positive content: I can confidently predict that reaching out will be met, and that confidence is exactly what frees a child to stop monitoring the relationship and turn toward the world. Avoidant attachment is the surprising case—high precision again, but negative content: I confidently predict I will not be met, so I stop asking and rely on myself. Anxious, or ambivalent, attachment is the low-precision case, where care was inconsistent enough that no reliable prediction could form, and the system stays chronically on, burning resources it cannot spare, never quite able to triangulate on a safe-enough model.

These are not either-or scenarios but a matter of degree—of how often a bad surprise arrives, and whether it........

© Psychology Today