Anxious Attachment, or Autistic Nervous System? Why the Question Matters

A client's partner describes it with real concern: every schedule change, every last-minute plan shift, every transition from one activity to the next produces visible distress — sometimes tears, sometimes a kind of frozen shutdown, occasionally irritability that looks, on its surface, like anger at the partner personally. The partner has read about attachment styles and has started to wonder, out loud, whether this is anxious attachment: a fear that plans changing means the relationship itself is unstable, a nervous system bracing for abandonment every time something shifts.

The client, who was diagnosed autistic in adulthood, has a different account of the same moment. It was never about whether the partner would still be there. It was about not knowing, until the last possible second, what was actually going to happen — a completely different kind of threat than relational abandonment, arriving through a completely different mechanism, that happens to produce some of the same visible symptoms.

Both explanations are plausible from the outside. Only one of them is usually correct, and figuring out which one requires taking a closer, more skeptical look at a psychological framework that gets applied to autistic people constantly, often without anyone asking whether it was built with them in mind in the first place.

Attachment theory, developed by John Bowlby and operationalized by Mary Ainsworth's Strange Situation Procedure, classifies how infants and children respond to brief separations from and reunions with a caregiver — secure, anxious-preoccupied, avoidant-dismissive, or disorganized — based on observed behavior: proximity-seeking, distress at separation, and specifically how comfort is sought and received upon reunion. The underlying claim is that these patterns reflect an internal working model of relationships, shaped by the consistency and sensitivity of early caregiving, that goes on to influence how a person relates to others across their whole life.

This is a genuinely useful framework for a huge number of clients. It was also built, tested, and normed almost entirely on neurotypical behavioral expression — proximity-seeking that looks like moving toward a caregiver, comfort-seeking that looks like eye contact and physical closeness, distress that reads through facial expression and vocal tone in expected ways. None of that is a given for an autistic child or adult, whose actual relational security may be expressed through entirely different, equally valid channels: parallel play instead of physical closeness, an object or routine used as a genuine attachment figure alongside a person, comfort-seeking through shared interest or information rather than eye contact and touch.

This isn't a fringe concern. When researchers have applied the Strange Situation Procedure directly to autistic children, the results have been messy in a specific, telling way. Several studies find that a meaningful minority to just under half of autistic children classify as securely attached using standard coding — lower than typical population rates — while other studies have found that using unmodified coding schemes, nearly all autistic children in a given sample end up coded as "disorganized" until researchers manually separate autism-specific behaviors, like stereotyped movements, from behavior that actually indicates relational distress. Researchers studying this population have explicitly raised the concern that their own measurement tools may lack what's called discriminant validity here — meaning the tool may simply be capturing level of autistic functioning rather than actually measuring attachment security at all.

In plainer terms: a framework built to read neurotypical behavioral signals, applied without modification to autistic behavioral signals, risks systematically misclassifying autistic security as insecurity, and autistic distress-about-predictability as distress-about-relationship, simply because the coding manual was never built with this population's actual expression patterns in mind. This is the same category error covered in Not Narcissism — a framework built by and validated on one neurotype, applied to another, producing a mismatch that gets blamed on the person being observed rather than on the limits of the observing tool.

Here's the differential that actually matters clinically, and it doesn't require throwing out attachment theory — it requires not defaulting to it as the only available explanation. Insistence on sameness — a documented, core autistic trait, distinct from attachment style — describes a nervous system that regulates through predictability and destabilizes through unexpected change, largely independent of who is present or how secure the relationship is. This isn't a metaphor for anxiety. It's a different mechanism entirely: reducing uncertainty in the environment functions as its own form of regulation, in the same way proximity to a trusted caregiver functions as regulation in classic attachment theory — except the object of the regulation is the predictability of the situation itself, not reassurance about a specific relationship.

This produces the exact ambiguity in the vignette above. A meltdown at a sudden schedule change (see Autistic Meltdowns vs. Shutdowns for the fuller mechanism) can look identical whether the underlying mechanism is I'm afraid this means you're leaving me or I had no way to prepare for this and my nervous system is now overloaded. The visible behavior — tears, shutdown, irritability, difficulty being soothed in the expected way — can be nearly the same. The actual internal experience, and therefore the actual clinical need, is completely different.

A few concrete questions tend to separate the two mechanisms more reliably than surface behavior alone. Does the distress track with who is present, or with what is predictable, regardless of who's there? Does advance warning and a clear structure for the change reduce the distress significantly, even without any reassurance about the relationship itself? Does the person seek comfort in ways that look nonstandard — an object, a routine, a return to a special interest — rather than proximity to a specific person, and does that comfort-seeking actually work, on its own terms, once it's recognized as comfort-seeking rather than avoidance? Is the distress specific to transitions and changes generally, including ones with zero relational content at all — a canceled train, a rearranged room, a menu item unexpectedly unavailable — rather than concentrated specifically around moments that could plausibly threaten the relationship?

None of these questions are diagnostic on their own. Together, they start to separate a nervous system organized around relational security from one organized around environmental predictability — which are not mutually exclusive, and can absolutely coexist in the same person, but which call for genuinely different clinical responses.

Misreading insistence on sameness as attachment insecurity has a specific, damaging downstream effect: it implicates the caregiver or partner. Attachment theory's whole causal story runs through caregiver sensitivity and consistency, which means labeling autistic distress as "anxious attachment" quietly suggests that a parent or partner failed to provide something they were, in fact, providing all along. Research on this population has explicitly flagged this risk — caregiver sensitivity does correlate with attachment classification in autistic samples too, but researchers have been careful to note that correlation doesn't establish that caregiver insensitivity caused the pattern, given how entangled caregiving behavior and a child's autistic traits already are with each other. A parent or partner blamed for "creating" insecurity in someone whose distress was actually about disrupted routine is being handed an injury that was never theirs to carry.

The reverse error causes its own harm. Real attachment wounds — genuine inconsistency, genuine relational injury, a real fear of abandonment rooted in real early experience — can get waved away as "that's just the autism" in exactly the pattern covered in Ableism as a Weapon as diagnostic overshadowing. An autistic person absolutely can have both insistence on sameness and unresolved attachment trauma at the same time, and defaulting to the neurological explanation because it's more available doesn't serve them any better than defaulting to the relational one.

The practical implication isn't choosing one framework over the other. It's building a genuine differential before assuming either one, and being willing to hold both simultaneously when both are actually present. For distress rooted in insistence on sameness, the intervention that actually helps looks like environmental scaffolding — visual schedules, advance notice, structured transition warnings, predictability built into the environment itself. For distress rooted in genuine attachment insecurity, the intervention looks like consistency-building and relational trust work. These are not interchangeable, and offering one when the other is needed doesn't just fail to help — it can actively confirm to the client that they weren't actually understood, which, for someone already navigating a lifetime of being read through the wrong interpretive framework, is its own familiar and specific kind of injury.

The client from the opening of this piece didn't need his partner to work on relational reassurance, though reassurance is never unwelcome. He needed a shared calendar, visible in advance, with changes flagged the moment they were known — and once that scaffolding existed, the tears at sudden transitions mostly stopped, not because the relationship had become more secure, but because the actual problem had never been the relationship's security in the first place.

This article is for educational purposes only and is not a diagnostic tool. Attachment patterns and autistic traits can only be accurately assessed by a qualified clinician working directly with the individuals involved.

Related reading in this series:

• Not Narcissism — the same surface-behavior/underlying-mechanism differential, applied to narcissism instead of attachment
• Autistic Meltdowns vs. Shutdowns — the fuller mechanics of the overload response this piece keeps returning to
• Ableism as a Weapon — diagnostic overshadowing, and how a real diagnosis gets used to dismiss a real grievance
• The Family's Sin Eater — the family-systems framework this whole series keeps returning to
• RSD or Anxious Attachment? The ADHD Differential — the parallel differential for ADHD and rejection sensitivity

© 2026 Love Psychotherapy, LLC. All rights reserved. Love Psychotherapy® and the Love Psychotherapy heart logo are registered trademarks of Love Psychotherapy, LLC.

Well wishes. 🙏

Mx. Love C. Dialogos, LMFT Licensed Marriage and Family Therapist | Buddhist Chaplain Pronouns: They/Them