The Uncertainty That Isn’t the Autistic Problem

I say this in my head roughly every time I open social media and see another autistic person claiming that autistic people are cognitively rigid.

I am, by any honest measure, cognitively flexible.

So why do we keep volunteering for the one deficit nobody actually made us take?

It’s like watching someone apologize for a crime committed by their evil twin.

Here’s the receipt.

We are not inherently cognitively inflexible. And what gets called “intolerance of uncertainty” may be doing a spectacularly bad job of describing what is actually happening.

I knew this long before I found the paper, because I’ve been reading myself for years, and the data never matched the diagnosis.

A 2025 grounded theory study did something almost suspiciously simple: it asked autistic adults what uncertainty actually feels like, instead of handing them a scale, adding up the numbers, and calling the score an explanation.

What came back was much more interesting than “autistic people can’t handle not knowing.”

Known uncertainty, the kind you can anticipate, prepare for, research, rehearse, or Google at 2am, wasn’t necessarily the problem. The problem was uncertainty you cannot plan your way out of no matter how hard you try.

And that distinction matters enormously, because now we are no longer describing a fixed deficit inside an autistic person. We are describing a relationship: how much unpredictable, unplannable uncertainty does an environment generate, and how much meaningful control does the person have to reduce it?

That’s not a personality trait. That’s math.

And suddenly a whole lot of my life makes more sense.

I didn’t burn out because my brain was constitutionally incapable of uncertainty. I burned out in workplaces practically engineered out of it.

Vague instructions. Invisible expectations. Same-day changes. Shifting priorities. Contradictory demands. “We’ll figure it out as we go.”

Give me complexity and I can play with it all day. Give me ambiguity and I will probably find twelve interpretations you hadn’t considered. Give me chaos I can map and I’ll build you a model.

But put me inside a system producing consequential uncertainty that I have no power to resolve, then measure my distress and call the distress evidence of my cognitive inflexibility?

Oh, come on.

I was halfway through writing this when I got on a call with a peer support group I’m part of — real people supporting one another, not a clinic. The facilitator, in her own check-in, brought up the exact same thing before I’d said a word about the study.

She was talking about planning a dinner with her husband. He’ll look up a place the day of and be fine. She needs to know where they’re parking, what’s on the menu, whether it’s outside, well before they walk in the door. Not because she can’t handle not knowing where they’re eating. Because she’d already done the exact sorting the study describes without a name for it — some not-knowing she can work with, some she can’t, and the difference is whether she can do anything with the gap in advance.

Another member of the group talked about a chaotic meeting the week before — people talking over each other, no order, impossible to follow who was saying what. The fix wasn’t “everyone needs to be less excitable.” It was structure: raise your hand, and the room quietly sorts itself into an order everyone can track. Known unknowns instead of unknown unknowns. Same mechanism, completely different context.

I read the study out loud to both of them, right there on the call. Neither of them had read it. It didn’t matter. They’d already lived the finding.

That’s the part that gets me. This wasn’t a case of research handing autistic people language for something we hadn’t noticed. It was research finally catching up to something a room full of us had already worked out for ourselves, informally, without a grant or a coder or a publication behind it. The theory didn’t teach us the mechanism. It confirmed one we’d already built out of necessity, one dinner reservation and one hand-raise at a time.

Here’s the part that gets me. The Intolerance of Uncertainty Scale has been in use for years. It asks you to rate statements like “uncertainty keeps me from living a full life” on a five-point scale, tallies the numbers, and hands you a score. It was never built to ask a follow-up question. It has no mechanism for “wait, which kind of uncertainty, specifically?” It just assumes uncertainty is one undifferentiated substance, and that some people simply have less tolerance for it, the way some people have less tolerance for spicy food.

Autistic people scored high on that instrument for years, and the field treated the number as the finding. It took someone actually sitting across from fifteen autistic adults and asking “okay, but what does that feel like” to discover the instrument had been quietly averaging together two completely different experiences: uncertainty you can work with, and uncertainty that exists specifically to make working impossible.

That’s not a minor methodological footnote. That’s the difference between measuring a person and measuring a mismatch, and mislabeling the second one as the first for the better part of two decades.

Maybe the deficit was never “can’t tolerate uncertainty.” Maybe we kept measuring what happens when uncertainty exceeds agency, and mistaking the resulting stress for a property of the person.

That is a very different research question. And, inconveniently, a much more political one — because “the person can’t tolerate ambiguity” asks nothing of the environment, while “the environment manufactures more unresolvable ambiguity than any nervous system could reasonably absorb” asks something of the workplace, the school, the clinic, the system that built it that way and never once had to answer for it.

This is the same move I keep finding, every single time someone actually bothers to ask instead of assume. A blunt scale hands you a number. The number gets narrated as a trait. Somebody finally asks a better question, and the trait dissolves into a mechanism, and the mechanism runs straight through the environment. Every time. I’m starting to think “we just haven’t asked properly yet” should be the default hypothesis for most things said about autistic cognition, not the last resort.

So no, I am not intolerant of uncertainty. I am intolerant of systems that generate uncertainty they refuse to plan for, and then hand the bill to the person standing closest to it.

Where I keep landing

This is also, unglamorously, the entire argument behind Kaleidoscope Systems. Not “here is a deficit, here is a coping strategy for it.” The five-layer model I teach — self, relationship, setting, institution, culture — exists because the layer that is usually blamed is the person, and the layer that is usually doing the damage is everything surrounding them. This study is a small, precise instance of exactly that move, done properly, with a method instead of a hunch.

I did not build a company around a feeling. I built it around watching this same pattern reappear, in domain after domain, for years — and getting tired of it being treated as news each time.

— Sher

Source

Lennuyeux-Comnene, L., Yates, J., & Gaigg, S. B. (2025). What is uncertainty? A grounded theory of the role of uncertainty in anxiety in autism. Autism in Adulthood, 7(6), 739–751. https://doi.org/10.1089/aut.2022.0085