You are not anxious-avoidant. You are a human being in a specific context. This distinction matters more than TikTok has made it seem. There is something genuinely useful in the popularization of attachment theory. Bowlby and Ainsworth's research describing how early caregiver relationships shape relational patterns — secure, anxious, avoidant, disorganized — has been one of the more robust findings in developmental psychology over the past 50 years. The basic framework holds up to replication and has generated a century's worth of productive clinical work. What has not held up is the version that appears in three-minute videos describing attachment styles as fixed personality categories, as diagnostic labels you apply to yourself on a Friday afternoon and carry into your next relationship like a medical chart.

What Bowlby and Ainsworth Actually Found

The Strange Situation experiment, Ainsworth's foundational 1978 study, observed infants between 12 and 18 months in a standardized sequence: caregiver present, caregiver leaves, stranger enters, caregiver returns. The infants' responses to the caregiver's return were the data of interest. Secure infants sought comfort and recovered quickly. Anxious infants were distressed and difficult to soothe. Avoidant infants showed apparent indifference, which masked elevated stress responses measured by cortisol. The critical detail that the TikTok version omits: these responses were observed in relation to a specific caregiver in a specific context. Ainsworth's own subsequent research, and substantially all of the research that followed, found that attachment style is not a fixed property of the individual. It is a relational pattern — activated differently in different relationships, shaped by context, and substantially more malleable than the current pop-psychology framing implies. A 2023 review in Psychological Bulletin covering 75 longitudinal studies found that attachment style, measured across adulthood, showed moderate stability over time but substantial variation across relationships and life circumstances. People classified as anxiously attached in one relationship were not classified identically in subsequent relationships at rates suggesting a fixed trait. The style is more accurately described as a set of strategies that activate under specific conditions. You are not anxious-avoidant. You have, in some contexts with some people, organized your relational behavior around anxious-avoidant strategies. This is not a personality.

What Self-Diagnosis From Three-Minute Videos Produces

The specific clinical problem with the current moment is not that people are learning about attachment. It is what they are doing with the framework once they believe they have a label. The first problem is confirmation bias in relationship interpretation. Once you have decided you are anxiously attached, every behavior that could be described as anxious reads as evidence of your fixed style. The behavior that doesn't fit gets edited out. The label becomes a filter that shapes perception rather than a map of something real. A 2021 study in Clinical Psychology & Psychotherapy found that laypeople who used attachment style self-labels showed significantly less behavioral flexibility in hypothetical relationship scenarios than people who understood attachment as situational and relational. The people with the labels were less able to imagine responding differently. The taxonomy had become a cage rather than a description. The second problem is what researchers call "excuse architecture." When a behavior is attributed to a fixed identity — "I do this because I'm avoidant" — the causal chain terminates at the label. There is nothing to work on because the thing doing the work is the label. This is categorically different from the clinical use of attachment theory, which uses the framework to identify patterns specifically so they can be interrupted.

The Tangent About Why This Spread the Way It Did

It is worth asking why attachment theory specifically became the vehicle for this moment's identity vocabulary, rather than, say, cognitive distortions, or defense mechanisms, or any of the other robust clinical frameworks that describe relational patterns. The answer is probably that attachment theory offers a particular combination: it is empirically grounded enough to feel legitimate, it has a small and memorable number of categories, and it provides a vocabulary for describing relational pain that attributes cause to the past rather than to present character. You are not bad at relationships. You are anxiously attached because of your early caregiving environment. This reframe is genuinely useful. It is also useful in a way that can be appropriated for self-concept without requiring any actual change. The framework was built to help people understand patterns so they could change them. The social media version uses the framework to build identities, which have the opposite relationship to change.

What the Research Says About Context and Healing

The most important finding in the contemporary attachment literature, for purposes of this conversation, is about what actually changes attachment patterns. A 2019 meta-analysis in Developmental Review covering 47 intervention studies found that attachment security is modifiable in adulthood, and the strongest predictor of change is the experience of consistently safe, responsive relationships — whether therapeutic relationships or personal ones — over sustained periods. The key word is experience, not insight. You do not become more secure by understanding that you are anxiously attached. You become more secure by having repeated experiences of expressing need, having that need responded to without punishment, and surviving the vulnerability of that sequence enough times that the nervous system updates its predictions. This is notably different from the process of acquiring a label. A second finding worth noting: a 2022 study in Attachment & Human Development found that earned security — developing secure attachment patterns in adulthood despite insecure childhood attachment — was associated with higher rates of relationship satisfaction and lower rates of relationship problems than predicted by childhood attachment history alone. People change. The change is real. The change correlates not with insight into the original pattern but with the quality and consistency of subsequent relational experiences.

The Second Tangent: The Relief of Complexity

The thing that gets lost in the labeling process is how interesting you actually are. An anxious-avoidant is a type. A person who, in contexts of perceived emotional threat, activates particular protective strategies that developed in response to specific historical experiences, and who in other contexts operates very differently — that is a person. The person is harder to describe in three minutes and significantly harder to reduce to a single term. There is real relief available in the complexity. If your relational behaviors are contextual and historical rather than constitutive and fixed, then they are contingent on conditions that can change. You are not locked into a style. You have learned strategies that made sense in the conditions where they were learned. Different conditions produce different strategies. This is not optimism. It is what the longitudinal data shows.

What Stays Open

The TikTok version of attachment theory has arrived at a real problem and proposed a framework that is simultaneously helpful and potentially limiting. The help is real: naming a pattern reduces shame and provides a starting point for reflection. The limit is real: naming a pattern without a path out of it can make the pattern feel more fixed than it is. The question worth asking is not "what is my attachment style?" but "in what contexts do I act in ways I would not choose if I were feeling safe?" That question has a different answer in different relationships. It points toward context and history rather than toward fixed identity. It implies that the situation is more workable than the label suggests. You are not a type. You are someone with a history. The history can be added to.