← Back to Kaleidoscope Systems
The Cognitive Ecologist
Six literatures that have almost nothing to do with each other, all arriving at the same place — and what that place implies about whether autism is intrinsically disabling.
Long read · with a synthesis diagram · content note: burnout, alcohol use disorder, suicidality mentioned in passing
Someone got angry with me online. That’s the occasion for this, and I want to be honest that it is the occasion, because the argument came first and the anger only made me finally lay it out.
What they were angry about, underneath the specifics, was this: I don’t experience autism as intrinsically disabling, and I say so. To some autistic people that sentence sounds like a betrayal — like I’m denying our reality, or distancing myself from people with higher support needs, or handing ammunition to a system that would love an excuse to cut services. I understand why it lands that way. I’ll come back to it.
But I didn’t arrive at that position by vibes, and I didn’t arrive at it by reading one thing. I arrived at it because six separate bodies of literature — gestalt psychology, ecological cognition, minority-stress epidemiology, executive-function measurement, disability theory, and epistemology — kept handing me the same shape, and eventually I couldn’t un-see it. This is that shape.
Start with the thing that made all of this necessary. My life has produced a set of verdicts that cannot all be true of one person.
Gifted, as a child in a stable home. Failing, a few years later, once the scaffolding came away. Brilliant at the kitchen table and hollowed out on a restaurant floor, the same year, sometimes the same day. Capable of sustained, complex thought at home and, in an assessment room, no impairment found — denied. Received completely, without translation, walking in a forest with my friend. Read as contempt and hubris in a comment thread, the same week.
One clarification, because the first verdict is the one most likely to be misread as advantage. I was tested as gifted. I was never given gifted education. That gap was made of class and circumstance: I moved through several elementary schools, and my parents — correctly, in my view — would not agree to separate me from my sisters in order to place me. So the label arrived and nothing came with it. What came instead was a sentence I heard in various forms for years: she doesn’t need as much as the others. A high score used as a reason to withhold support rather than provide it. The verdict itself became the justification for the absence of help.
Nothing in the middle of that changed. My processing didn’t change, my intelligence didn’t change, my intent didn’t change. So either I am a wildly unstable person, or the verdicts were never measuring me. That was the question I went looking for an answer to, and I found six of them.
Ten environments, ten verdicts
The centre never changed. Only the ring did. Read the verdicts against each other and they cannot all be describing the same object.
That is the anomaly. Not a hard life — a set of readings that contradict each other while the thing being read stays still. Any explanation located in me has to account for why the same person keeps producing incompatible results. None of them can.
Positionality, verdict by verdict
An analysis without positionality is unfinished. So: what kind of system each of those rooms was, and who I was inside it.
Environment
The system
Who I was in it
Stable childhood home
Family. Private. No institutional gaze.
A white child in a poor but safe home. Not materially secure — the privilege here was safety and legibility, not money. I was tested as gifted and never given gifted education: multiple elementary schools, a family that would not separate me from my sisters, and a label that arrived with nothing attached.
College at 18, self-funded
Private higher education. Scholarship, student and parent loans, working throughout.
The safe environment was gone and nothing replaced it. Scored high once, now failing — while working and carrying debt. Whiteness bought me the benefit of the doubt for a while. It never bought me support.
24 years of work
Restaurants, nonprofit, government. Slow wage progression, no accommodation, high sensory load.
Working class in those rooms regardless of my degree. I worked my way up, but the wage progression was slow enough that what it cost me exceeded the material safety it bought. In the nonprofit and government roles it was my justice sensitivity that ended things — and there is no accommodation category for that. This is where class stopped protecting me.
Assessment room
Clinical diagnostics. Impairment-location.
A verbal, articulate, insured adult who could reach and pay for a four-session assessment. That access is privilege. It produced extreme scores — AQ 43 against a threshold of 26, RBQ-2A 48 against 26, CAT-Q masking 155 against 100, ASRS-5 19 of 24 — and a finding of no impairment, on the grounds that my needs were “carefully and well accommodated.” They were. I had bought the accommodations myself. privilege. It produced a finding of no impairment.
Two years in the system
State bureaucracy. Eligibility-gated.
Someone with enough institutional literacy to file, appeal and persist for two years — and it still failed. Most people cannot afford two years.
Comment thread
Social media. Context-stripped, no peer support.
An unmasked autistic woman in public. Here whiteness is not shelter; visibility is the exposure.
The home I live and work in
Self-designed environment. Chosen relationships.
Materially privileged enough to have one. This is the single largest privilege on this list and I will not pretend otherwise.
Forest, with a friend
Chosen relationship. Diverse, multicultural, humanistic. No legibility demand.
Received whole. Note what kind of room this is. My safest environments are the diverse ones.
My university
Diverse, multicultural, humanistic. No legibility demand.
Not disabled here. Nobody requires me to be linear to be taken seriously. And I paid for this environment with my future — the debt is beyond belief. Access I bought, not access I was given.
My chosen community
Heavily diverse. Access first, not ideology first.
Not disabled here either. The organising principle is what makes it work — nobody has to pass a purity test before their access needs are met. Difference is the starting assumption, not the thing being adjudicated.
Read the right-hand column and the pattern is not the one people expect. My privilege shows up as access and safety — a safe childhood home, an assessment I could eventually reach, two years I could spend appealing, a home I could redesign. Not money in childhood; there wasn’t any. It does not show up as whiteness protecting me from harm. The rooms that damaged me were normatively white, professional, legibility-demanding institutions. The rooms where I am received whole are the diverse, multicultural, humanistic ones. Whiteness is the environment that harms me. It is not the one that shelters me.
Six sources on the left. The mechanism each one handed me in the middle. Where they converge at the bottom. None of these fields were talking to each other. That’s the part that convinced me.
What I was reading
The mechanism it handed me
Gestalt psychology & gestalt language processing
Koffka 1935 · Prizant 1983 · Blanc 2012 · Hutchins et al. 2024 · Hoerricks 2026
Meaning arrives as a whole; the parts resolve afterward. So context isn’t decoration on the message — it is part of the message. Strip it and the meaning doesn’t degrade, it inverts.
Ecological & situated cognition
Bateson 1972 · Gibson 1979 · Varela et al. 1991 · Barab & Plucker 2002 · Clark 2008
Ability is not a quantity sitting inside a person waiting to be read off. It emerges between a mind and its conditions. Barab and Plucker asked it outright: smart people, or smart contexts?
Minority stress & weathering
Meyer 2003 · Geronimus et al. 2006 · Botha & Frost 2020 · Raymaker et al. 2020
Chronic structural stress becomes measurable biology. Geronimus proved it at population scale with allostatic load. So the environment doesn’t just inconvenience a body. It writes on it.
Executive-function measurement
Hill 2004 · Geurts et al. 2009 · Garfinkel et al. 2015 · Guazzo 2024 · Groba et al. 2026
Fractionate a construct and the deficit dissolves. Geurts’s review found no consistent evidence for cognitive flexibility deficits. Groba: same accuracy, slower clock. The scores were tracking fit.
Disability theory & disability justice
UPIAS 1976 · Oliver 1990 · Sins Invalid 2016 · Mingus 2017 · Chapman 2023
Disability is the disadvantage imposed by how things are organised, not the difference in the body. That’s the founding distinction, and it means “intrinsically disabling” is a contradiction in terms.
Epistemic injustice
Haraway 1988 · Fricker 2007 · Milton 2012 · Botha 2021 · Chapman & Carel 2022
My report on my own interior is pre-judged inadmissible, and the shared language needed to describe it is missing. Which is precisely why the contradiction stayed invisible for forty-odd years.
All six arrive here
Convergence
Observable capacity is a property of a relationship, not of a person.
So every verdict measures the room as much as the mind. And the difficulty I have spent my life being blamed for does not sit where everyone kept looking for it.
If one field had told me this, I’d have suspected the field. What I couldn’t explain away is that a Gestalt psychologist writing in 1935, an ecological perceptual theorist in 1979, a public-health epidemiologist measuring allostatic load in Black Americans, a group of autism researchers who noticed their own flexibility literature wouldn’t replicate, the founders of the British social model, and a philosopher of testimony all handed me the same structural claim. None of them were trying to say anything about autistic adults on social media.
Convergence across unrelated methods is the strongest evidence a synthesis can have. It’s also, I’ll admit, exactly the kind of pattern a whole-first mind is built to find and most at risk of over-finding. I hold it as a strong working position, not a proof.
And there is a pattern in that table I want named, because it cuts against how this argument usually gets dismissed. My safest environments have been the diverse ones. My family. The forest. A diverse, multicultural, humanistic university where nobody requires me to be linear in order to be taken seriously. A chosen community organised around access first rather than ideology first. The environments that disabled me were the normative ones — the mass-processing school, the restaurant floor, the eligibility office, the comment thread policing who counts. Being white did not make those rooms safe for me. What made rooms safe was whether difference was expected there.
I want to be exact about this, because it matters for how the argument should be judged. If this were only a reading list, it would be a literature review with a personal anecdote attached. That is not what happened, and pretending otherwise would misrepresent the method.
Three things were running at once, and none of them would have been sufficient by itself.
First
Lived experience
Forty-seven years inside the centre of that diagram. This is what produced the anomaly — the contradiction that needed explaining. No amount of reading would have handed me the question, because from outside the ring there is nothing obviously wrong.
Second
Praxis
Twenty-five public workshop sessions. Rooms every week. Teaching it, being corrected in it, watching it fail and get revised against other people’s lives. This is what tested the pattern — and what kept it from being a story that only fits one biography.
Third
Scholarship
Six literatures, a library of nine hundred sources. This gave me the vocabulary and the mechanisms — the ability to say “allostatic load,” “situated ability,” “testimonial injustice” instead of “something is wrong and it isn’t me.”
Experience gave me the question. Praxis gave me the check. Scholarship gave me the words.
And the real-time data backs this up. Scheeren and colleagues (2025) used ecological momentary assessment — smartphone prompts through the day rather than a retrospective survey — and found that masking varies with who is present. It is not a stable trait someone carries. It moves with the room.
Take away the lived experience and I would never have noticed there was anything to explain. Take away the praxis and I would have a theory that fit exactly one person, which is not a theory. Take away the scholarship and I would have a strong intuition and no way to say it, no way to test it, and no way to be argued with — which is the condition I lived in for most of my life.
This is also why I keep insisting the disagreement be about the ideas. The lived experience is not the evidence base; it is the thing that made the evidence base necessary. Those are different claims, and only one of them is mine to be corrected on.
I need to say this plainly, because it sits underneath every argument I’ve had this year.
This is not rhetoric. It is not a debating position I adopted. It is not internalised ableism, it is not denial, it is not privilege talking, and it is not whatever else has been assigned to it. It is my life. And that is exactly why it hurts so much when someone leaps at me with all of that — because the thing being diagnosed as a bad argument is just the shape of forty-seven years.
I have lived some of it in material privilege and some of it not at all, and I have not hidden either — it is in the table above, verdict by verdict. What I’m asking is that the privilege be located accurately, because when “that’s a white privileged argument” gets thrown at this, it lands somewhere the evidence isn’t.
An analysis without positionality is unfinished. That’s mine. Now argue with the argument.
Here is the part people get angry about, said as precisely as I can.
I believe the world disables autistic people, comprehensively, whether or not any instrument can find a reason to. What I cannot make work is the claim that the disability is inside me. And my objection isn’t only theoretical — it’s that the intrinsic framing fails on contact with the actual system.
Get the chronology right, because it is the whole argument. The system has been slowly disabling me since I left the one environment that was built for me — my family. That is not a metaphor. Inside my family I was legible, translated, received whole. Every environment after it has cost me something to occupy, and the cost accumulated for decades before anyone put a name to any of it.
So I did not arrive at the disability system healthy and get broken by it. I went looking for a name in my forties because the environment had already become unsurvivable — not because something in me had finally revealed itself. The disabling was decades old by then. The diagnosis was me trying to find out what to call it.
The assessment itself was thorough and genuinely life-changing, and it was done by someone who works from the neurodiversity paradigm. So my diagnosis does not contain the word impairment. I was fine with that. I didn’t see myself as impaired either. I knew perfectly well what was happening: the system was the thing I was being forced to accommodate, and I needed accommodations in order to keep accommodating it. Deficit framing wasn’t necessary for that. Or so I thought.
This is the part that is more complicated than my own summary of it, and I want to correct myself. My assessment did not find nothing. It documented a great deal: an AQ of 43 against a threshold of 26; repetitive behaviours (RBQ-2A) at 48 against 26; an ADHD screen (ASRS-5) of 19 out of 24; a CAT-Q score of 155 against a threshold of 100, which is a very high masking presentation; interoceptive differences, including difficulty naming emotions and recognising hunger; and sensory needs that have measurably affected my quality of life. What it declined to do was translate any of that into the specific word the funding apparatus requires. It described an Autistic brain style, with strengths and differences, and it was right to.
And then, on the final page, there is a medical-necessity checklist. Treatment of a mental disorder? Yes. Services realistically expected to improve condition or alleviate impairment? Yes. Ticked, signed, dated — because otherwise it isn’t billable.
So the document contains its own contradiction, and it is not the assessor’s fault. She practises in a state that is genuinely neuroaffirming — which is good, and rare, and ahead of its own policy. That is the actual shape of the problem: the clinical culture has moved and the funding architecture has not. Nine pages of accurate, strengths-based, non-pathologising description, wrapped in a form that only pays out for a disorder and an impairment. The affirming report and the billing code cannot both be true, and the system pays the billing code.
Then I took an accurate, non-pathologising diagnosis to an apparatus that only funds pathology. The disability system doesn’t run on the social model; it runs on impairment-location. It looks inside a person for a finding. My assessment hadn’t recorded one, because there wasn’t one to record in those terms — so the disability system didn’t find one either. No one can see my impairment. Not because it is hidden. Because it was never in me to begin with.
And I want to be exact about what those two years were. That was not a leisurely appeal. I was in burnout. I was, in that period, genuinely disabled — losing function in real time, unable to advocate for myself precisely because of what I was trying to get help for. It was a desperate attempt to get support while disabled, and it cost me two years, significant debt, and function I have not fully recovered.
Which is where the demand comes in, and it is the practical heart of all of this. Autistic burnout is a disability caused by something outside us. If we will not name it that way, it stays unfundable — not a condition, not an impairment, just a person mysteriously failing. And people in it cannot advocate, by definition, because the capacity to advocate is one of the things it takes. We deserve support when this happens to us. That requires a category that locates the cause where it actually is.
So when I say I don’t experience autism as intrinsically disabling, I am not making a smaller claim than my critics. I’m making a claim about where the harm lives — and therefore about where you have to intervene, and about which framing gets used against us in an eligibility determination.
Some autistic people do have findable impairment. Nonspeaking autistic people, people with co-occurring intellectual disability, people whose support needs are continuous and will remain continuous under any environmental design anyone could build. Nothing in this framework denies that, and nothing in it asks them to reinterpret their own lives. The world disables them too, and the assessment finds something. Both are true.
I’m aware the sentence I’m defending has a history. “Autism isn’t inherently disabling” has been used for decades to split us into deserving and undeserving, to deny services, to insist that high-support-needs autistic people don’t need what they plainly need. Mine isn’t that argument. It sounds like it in the first three seconds, and I don’t get to be surprised when people respond to the first three seconds.
I also know disability identity was liberation for a lot of people — the thing that finally stopped the self-blame. I am not devaluing that. I’m saying it wasn’t available to me, because the institution that confers it looked at me and said no, and I am not willing to organise my self-understanding around a category I was refused entry to.
Everything above is ten readings of a white mind. That is a real limit on what the ring can show, and I want to state it as a limit rather than a disclaimer. If the model is right — if verdicts track environments rather than people — then the pattern I have described should be more visible at harder intersections, not less. Mine is the mildest version of it.
I am not going to imagine my way into that. But the model makes specific predictions, and other people have already tested them.
Predictions, and who documented them
If verdicts track environments, the same profile should be identified later, and misdiagnosed first, for Black children. Mandell and colleagues found exactly that — later age at diagnosis among Medicaid-eligible Black children, and a documented pattern of other diagnoses arriving before autism. Mandell et al. (2002; 2007; 2009)
If access is what converts difference into disability, service use should diverge by race even at equal eligibility. It does. Bilaver et al. (2021); Burkett et al. (2015)
If chronic structural stress becomes measurable biology, the load should compound where the hostility compounds. Weathering, at population scale. Geronimus et al. (2006)
If legibility is the currency, autistic trans people should face compounded illegibility — and be asked to prove themselves twice. Documented as intersectional stigma. Maroney & Horne (2022); Strang et al. (2018)
If identification pipelines produce verdicts, the same underrepresentation should appear in gifted programmes. It does. Ford (2010)
Read together, that is not my argument being weakened by intersectionality. That is my argument being confirmed by it, by researchers who were better placed to test it than I am. And it is why the framework has to be intersectional at the level of design rather than as an appendix — Crenshaw (1989, 1991) and Collins (2000) for the analytic, Schalk (2022) and Brown, Ashkenazy and Giwa Onaiwu (2017) for the accounts from inside, Botha and Gillespie-Lynch (2022) for the recovery of neurodiversity as a category of intersectionality in the first place.
What I can tell you is where my ten verdicts came from. What happens to a Black autistic woman in that eligibility office, or a trans autistic person asked to be legible twice over, I can only imagine — and I am not going to. I would rather point at the people who know, and be argued with about whether the mechanism holds.
Which is the whole ask, and I will keep making it until it lands. Argue with the argument. Tell me the convergence is illusory. Tell me environments cannot account for that much variance. Tell me I have mistaken a privilege for a mechanism. All of those are real objections and I want them. What is not an argument is my character, my whiteness deployed as a conclusion rather than a variable, or a screenshot of one sentence offered as evidence of who I am. That is the same inferential error the entire essay is about — reading an interior off an output — and it does not stop being an error when it is aimed at me.
Which brings me back to why any of this got written this week.
Someone took one sentence of mine, produced in the environment with the least available context and no peer support, and read my character off it. Screenshots offered as evidence — evidence of an interior, gathered from output. That is the exact inferential move all six of those literatures say you cannot make. It is the same move that scored me as failing at nineteen, unreliable at thirty, and not-impaired-enough at forty-five.
I don’t think it was fair. I think it was the pattern, happening again — and this time I had the diagram for it.
I don’t need anyone to adopt my position. I’d like the disagreement to be about the ideas, which is what I keep asking for and rarely get. Tell me the convergence is illusory. Tell me impairment effects survive every environment and my forest day was luck rather than fit.
But if you want to run that objection, you have to run it against more than one afternoon. Eighteen years inside my autistic family, where no impairment appeared. The home I live and work in now, where none appears. Other people’s homes I walk into, where none appears. My forest day was not a lucky outlier — it is what happens every single time the conditions are right, and it has happened across decades and across households. The impairment shows up when the environment demands it and disappears when the environment doesn’t. That is not a trait behaving. That is a mismatch behaving.
Those are still real objections and I’d rather have them than agreement.
None of this is only about me, and it isn’t only theoretical. I sit in rooms every week with people who are spending their energy convincing each other that they are not the problem. That is what the analysis is for. But reassurance is not redress, and eventually somebody has to be able to file something.
I tried to get justice for what happened to me at work. There was no help and no support for that process, and the process itself was ableist from end to end. There is no functioning route for environmental harm — no mechanism that recognises “this workplace produced this collapse” as a claim rather than a complaint. Nothing is working for those of us who have to work.
So this is where Kaleidoscope Systems is pointed. We are interested in policy that recognises harm caused by the environment — that treats autistic burnout as an externally-caused disability, that funds support at the point of collapse rather than demanding a finding inside the person first, and that gives people an actual route to justice when this happens instead of a two-year search for an impairment that was never there. Not a framework to sell you. Policy, and a mechanism.
What I won’t accept is being told what I meant. I have lived in the middle of that diagram for forty-seven years. Everyone else has only ever seen the ring.
Not a framework to sell you. Policy, and a mechanism.
The ask
Policy that recognises harm caused by the environment. That treats autistic burnout as an externally-caused disability. That funds support at the point of collapse rather than demanding a finding inside the person first.
And a real route to justice when this happens — instead of a two-year search for an impairment that was never there.
Because if we keep moving ahead of the policy, only people with institutional and class privilege will survive a burnout. Which is how it has always worked. It is why so many of us are not here.
And I have to be honest about class, because our current policy protects people with money and working-class people already know it. We are not stupid. We know what the two-year search costs, and we know who can afford to conduct it. We cannot stop working. We have families. We have children. We have aging parents. The system is built as though disability arrives to someone with savings, flexible time, and an advocate — and it quietly excludes everyone for whom stopping means losing the house.
I have twenty-three-year-olds. And I have to ask myself the question directly, because it is the one that actually keeps me at this: do I want to send my children into hostile environments with no recourse? Into workplaces that will extract exactly what those restaurants extracted from me, with no policy that recognises the harm, no mechanism when it happens, and a diagnostic gate they will have to fail their way through before anyone helps them?
Because that is the current arrangement. Not a hypothetical. That is what we are handing them.
I almost didn’t make it. Twice. The reason I am able to write this at all is that I was gifted the conditions to heal — time, safety, people, a room I could redesign around my own nervous system.
That should not be a privilege. It should not be luck. It should be policy.
If we cannot stop, burnout takes us out. You may be one lost job from a lost home. One collapse from couch-surfing. One from a marriage ending. So why would we not try to prevent that for autistic people?
Instead we gatekeep. We tell people they must have been disabled their entire lives in order to count. That is not advocacy. That is a private club with hidden rules — and it recreates the exact neuronormativity we say we oppose. Almost nobody wants to look at that.
I have done a great deal of work on the addiction and on the suicidality, and here is what I keep finding: so many of our autistic siblings do not know that they are not failing because they are insufficient. They are failing because the environment is harming them. Nobody has ever told them there is a difference.
Every claim above rests on somebody’s work. Each of these has been checked against the published record — so you can go and disagree with them directly rather than with me.
Argue with me
Not agreement. Objections. Tell me the convergence is illusory. Tell me impairment effects survive every environment. Tell me what my positionality blinds me to. Tell me the one I most need answered: is fighting for this necessary? And if you would rather not argue at all — tell me your story instead, or whether any of this matches your life, or what you would want policy to actually do. All of it is welcome.
I read everything. I will not always agree, and I will say so. But a real objection is worth more to me than agreement.
← Back to Kaleidoscope Systems
A note on how this was made: the synthesis, the argument, and the voice are mine. I use AI as an access tool — to hold six literatures in view at once and to help translate a whole I grasp all at once into ordered, sequential prose. Given that the essay is about the cost of that translation, the irony is doing its own work.
The Cognitive Ecologist · Companion to “How I Connected the Dots”
One mind held constant. Ten rooms. Ten incompatible readings — and the literature that explains each one.
Companion essays · each node of the ring diagram, cited
The ring diagram makes a claim without arguing it: that the verdicts on my life contradict each other because they were measuring rooms rather than a person. These ten short essays are the argument. Each takes one environment, one verdict, and the research that accounts for it.
They are deliberately short. The point is not to exhaust any one of them but to show that every node has a literature behind it — that the pattern is not a story I told about myself after the fact.
01Stable childhood home
I was tested as gifted as a child, in a house that made almost no unmanaged demands of me. The sensory world was known. Recovery was available. The scaffolding I did not yet know I depended on was simply there, and nobody had to name it because nobody had to fight for it.
What a test measured under those conditions was a whole-first mind operating under light load with the door open. Barab and Plucker (2002) asked the question that reframes this entirely: smart people, or smart contexts? Their argument is that intelligence, expertise, and talent have been wrongly construed as traits sitting inside individuals, when they emerge in person–environment transactions. A gifted score obtained in an unusually well-fitted environment is not a reading of a child. It is a reading of a fit.
The label also flattened something. I was hyperlexic — decoding print years ahead of my age — and simultaneously dyslexic and dyspraxic. Grigorenko, Klin and Volkmar (2003) note that hyperlexia has resisted tidy categorisation precisely because it presents as disability and superability at once. The single word gifted took an uneven, spiky profile and produced one reassuring number from it.
And it is worth saying what the label did and did not come with. I was tested as gifted. I was never given gifted education. That gap was made of class and circumstance — I moved through several elementary schools, and my parents, rightly, would not agree to separate me from my sisters in order to place me. So a high score arrived and nothing came attached to it except a sentence I heard in various forms for years: she doesn’t need as much as the others. The verdict became the justification for the absence of help. That is the same logic I would meet again decades later in an eligibility office, and I did not recognise it at the time.
And identification systems are not neutral. Ford (2010) documents the persistent underrepresentation of culturally different students in gifted education — referral bias, testing practice, deficit thinking. Which door you arrive through determines the valence assigned to overlapping phenomena. I arrived through the gifted door, so my intensity was called overexcitability rather than dysregulation (Mendaglio, 2008).
Barab & Plucker (2002), Educational Psychologist, 37(3) · Grigorenko, Klin & Volkmar (2003), JCPP, 44(8) · Ford (2010), Gifted Child Today, 33(3) · Mendaglio (2008), Dabrowski’s Theory of Positive Disintegration
02School, once the supports came away
Then the world withdrew the scaffold, in the ordinary way that childhood ends, and the same mind began producing worse output. I failed classes I had been told I was too bright to fail.
Nobody read this as the field fragmenting faster than I could hold it. We read it as the earlier verdict revised downward. This is the cruellest form of the measurement error, because you internalise it: you conclude that you peaked at ten.
The developmental data supports the environmental reading rather than the revealed-deficit one. Rosenthal and colleagues (2020) found that impairments in real-world executive function increase from childhood to adolescence in autistic people — not because the mind degrades, but because environmental demand rises steeply while support falls away. The gap widens as the room gets harder.
The scores did not reveal what had always been wrong with me. They tracked the withdrawal of what had always been holding me.
This is also the point at which girls and late-identified children disappear from the record. Bargiela, Steward and Mandy (2016) documented women diagnosed in late adolescence and adulthood describing exactly this sequence — “pretending to be normal,” professionals missing them because of their gender, and the accumulating cost of going unrecognised. Jadav and Bal (2022), in a sample of over 4,600 autistic adults, found later recognition associated with more co-occurring psychiatric conditions. Failure at this stage is not a diagnostic finding. It is a support failure that gets recorded as one.
Rosenthal et al. (2020), Autism & Developmental Language Impairments, 5 · Bargiela, Steward & Mandy (2016), JADD, 46(10) · Jadav & Bal (2022), Autism Research, 15(11)
03The restaurant floor
After I had my children, the work I could get was restaurants. A restaurant does something specific and ruinous to a gestalt processor: it shatters the field into a hundred competing, unrelated parts — tickets, timers, voices, heat, motion — with no coherent whole to hold them. Brilliant at the kitchen table and hollowed out on the floor, the same year, sometimes the same day.
It is worth being precise about the class story here, because it is not simply that the work was low-paid. I worked my way up over twenty-four years. There was progression. But the progression was slow enough, and the cost of holding those rooms high enough, that what it took out of me exceeded the material security it bought. That is a specific kind of trap: the job is not bad enough to be obviously unsustainable, and not good enough to fund what surviving it requires. You stay, because leaving costs more, until staying costs everything.
It was not only restaurants. I tried other rooms, and they failed in a more revealing way. In nonprofit work and later in a government role, the sensory load dropped — and something else took over. I saw the systemic patterns immediately, the whole shape of what was wrong, and I did not keep my mouth shut. I did not know then that this was justice sensitivity, that it was unaccommodated, or that it was a documented disposition rather than a personality defect. I only knew I was somehow the problem in every room I could see clearly.
And there is a chronology inside that I need to name, because it is the hinge of the whole story. The nonprofit years are when the mask came off — not because I chose to unmask, but because the thing holding it on had been alcohol, and I got sober. The mask was never a skill I had. It was a chemical achievement. Alcohol lowered the sensory gain, muted the social monitoring, and loosened the executive grip I was otherwise holding by force, and the self-medication pattern is documented rather than mysterious (Khantzian, 1997) — substances get selected for the specific load they relieve. Take the substance away and the surface it was maintaining goes with it. So the room where I could finally see the systemic patterns clearly is the same room where I no longer had any way to make my seeing palatable.
It was sobriety that made me open my mouth.
Justice sensitivity is a measured construct with four distinct facets, and the other-oriented ones — observer and perpetrator sensitivity — track empathy, role-taking, and social responsibility rather than grievance (Schmitt et al., 2005). Autistic adults have been shown to be more prosocial toward socially distant others in a study specifically designed to test and reject the “it is just rigidity” explanation (Forbes et al., 2026). So what read as difficulty was a disposition doing exactly what dispositions do, in institutions with no route for it.
That is the part the word unreliable conceals. In the restaurant the load broke me. In the nonprofit and the agency, my accurate reading of the system broke my standing. Different mechanism, same verdict. Which is precisely the point: the verdict tracked the room, not me.
Thelen and Smith (1994) give the mechanism. In a dynamic-systems account, higher-order behaviour emerges from the real-time interaction of lower-order components rather than from a dedicated module. So when one channel is obstructed, the cost does not stay local. It propagates. Load spent upstream is load unavailable downstream.
Some of my channels were permanently costlier to run. Motor-coordination differences are among the most consistently documented features of autistic samples (Fournier et al., 2010), and dyspraxia is a concrete, named load on motor planning — one of the very subsystems the model lists. Everaert and colleagues (2024) show the principle even in ostensibly helpful settings: adding a bodily task can backfire when it competes with the cognitive one for the same budget.
And much of what those environments extracted was not the work. It was the camouflaging — the effortful production of a fluent, unbothered surface, which carries its own documented cost (Hull et al., 2017). Run at a chronic deficit for years and the result has a name: burnout, described by autistic people as having all of your internal resources exhausted beyond measure (Raymaker et al., 2020). The reasoning had not left me. The pathway was congested, night after night, until the congestion became the condition.
Khantzian (1997), Harvard Review of Psychiatry, 4(5) · Schmitt et al. (2005), EJPA, 21(3) · Forbes et al. (2026), Autism, 30(2) · Thelen & Smith (1994), A Dynamic Systems Approach · Fournier et al. (2010), JADD, 40(10) · Everaert et al. (2024), European Journal of Special Needs Education · Hull et al. (2017), JADD, 47(8) · Raymaker et al. (2020), Autism in Adulthood, 2(2)
04The assessment room
Then the room built specifically to find out what is wrong with me looked, and found nothing it recognised.
This is the node where the whole diagram turns, because the assessment literature undermines its own instruments. Geurts, Corbett and Solomon (2009) reviewed the cognitive-flexibility research across multiple sites, methods, ages, and autism subtypes and concluded that no consistent evidence for cognitive flexibility deficits was found — while noting that clinicians assume inflexible everyday behaviour maps directly onto lab flexibility deficits. They called the mismatch a paradox.
It is not a paradox. It is a part-sampling instrument reporting on a whole it never measured. Groba and colleagues (2026) separate the components and find late-diagnosed autistic adults matching non-autistic peers on accuracy while differing mainly in reaction time. Same destination, different clock — scored as a deficit by any measure that prices speed. Guazzo (2024) reaches for the corrective from the working-memory side: attend to the strategies and resources a mind deploys, not to raw performance scores that silently penalise a differently-routed cognition.
Garfinkel and colleagues (2015) show what happens when you do this properly. Fractionate interoception into accuracy, sensibility, and awareness and the three turn out dissociable — most people’s belief about their body-sense does not match their measured body-sense. One word had been doing three jobs, and the contradictions in the literature resolved once it was split.
Feuerstein, Feuerstein and Falik (2010) draw the methodological conclusion the psychometric tradition resists: what a mind can do with appropriate mediation is a truer index of capacity than what it produces alone under standardised deprivation. Dynamic assessment unlocks the door and reports on the house. Static assessment measures the doorway and writes up the rooms.
Geurts, Corbett & Solomon (2009), Trends in Cognitive Sciences, 13(2) · Groba et al. (2026), Research in Autism, 131 · Guazzo (2024), IJESSR, 7(2) · Garfinkel et al. (2015), Biological Psychology, 104 · Feuerstein, Feuerstein & Falik (2010), Beyond Smarter
05The comment thread
The environment with the least available context, no peer support, and no channel for repair. One sentence, read as a statement about my character.
Gestalt language processing explains why this room is uniquely hostile to me. Prizant (1983) argued that autistic language had to be understood as a whole before it could be understood at all; the framework has been elaborated in autistic and clinical communities (Blanc, 2012) and examined empirically (Hutchins, Knox & Fletcher, 2024). The units I speak in carry the emotional weather of the rooms they were learned in. Strip the context and the meaning does not degrade gently. It inverts.
Milton (2012) supplies the structural point: breakdowns in mutual understanding between autistic and non-autistic people are a bidirectional mismatch, not a unilateral autistic failure. Cheang and colleagues (2025) finally tested the direction almost nobody tests — whether non-autistic people can accurately read autistic emotion — and the assumption of one-sided deficit stopped looking obvious. The misreading is a property of the coupling.
Hoerricks (2026) gives the sharpest available formulation of what is actually being demanded: an extractive demand for clarity, in which legibility becomes the condition of being heard. And Fricker (2007) names the injustice precisely — testimonial, when the knower is pre-judged unreliable; hermeneutical, when the shared language needed to describe the experience does not exist.
Screenshots offered as evidence. Evidence of an interior, gathered from output. That is the exact inference the whole literature says you cannot make.
Prizant (1983), JSHD, 48(3) · Blanc (2012) · Hutchins, Knox & Fletcher (2024), ADLI, 9 · Milton (2012), Disability & Society, 27(6) · Cheang et al. (2025), Autism, 29(9) · Hoerricks (2026), Neurodiversity, 4 · Fricker (2007), Epistemic Injustice
06Two years inside the disability system
I entered that process to get accommodations so I could keep accommodating an environment that was already disabling me. Two years later I had burnout, debt, and lost function I did not have to lose — and no support.
Autistic burnout is the mechanism, and it is now well-described. Raymaker and colleagues (2020) defined it from autistic accounts as the exhaustion of internal resources beyond measure; Higgins and colleagues (2021) refined the definition through experts by lived experience; Summerill and Summers (2025) synthesise the evidence base. What all of it shows is a condition caused by sustained environmental demand — and one that removes, by definition, the capacity you would need to advocate for help with it.
The access literature explains why the door stayed shut. Mason and colleagues (2019) systematically reviewed barriers to healthcare access for autistic adults; Nicolaidis and colleagues (2015) titled their findings with a participant’s own words — respect the way I need to communicate with you. A system that requires sustained self-advocacy in a communication style I cannot reliably produce under load is not neutral machinery that happened to fail. It is a filter.
And the harm is not metaphorical. Botha and Frost (2020) show minority stress predicting distress in autistic adults; Geronimus and colleagues (2006) demonstrated at population scale that chronic structural stress becomes measurable biology. I had nothing findable when I went in. I had significant functional impairment when I came out. The apparatus built to recognise disability manufactured it, then declined to fund what it had made.
Raymaker et al. (2020), Autism in Adulthood, 2(2) · Higgins et al. (2021), Autism, 25(8) · Summerill & Summers (2025), Research in Autism · Mason et al. (2019), JADD, 49(8) · Nicolaidis et al. (2015), Autism, 19(7) · Botha & Frost (2020), Society and Mental Health, 10(1) · Geronimus et al. (2006), AJPH, 96(5)
07The home I live and work in
Three years of relative safety, most of it in a room I designed around my own nervous system. By most measures I am organised, inventive, and capable of sustained complex thought. Nothing in me changed to produce that.
Klein and Macoun (2024) give the construct: person–environment fit. Their work reframes camouflaging as a function of fit rather than a trait of the person, which moves the question from why do you mask to what is this environment demanding. Masking, on this account, is contextual rather than dispositional — a response to what a room asks of you.
The flourishing literature is newer and better than its reputation. Hedlund and colleagues (2025) built an autistic-led conceptualisation of what a good autistic life actually consists of, from autistic women’s lived experience. Dantas and colleagues (2025) review autism and thriving critically. Lampinen and colleagues (2026) document self-reported strengths and talents of autistic adults — a rare thing to find measured at all.
And the tools are not cheating. Clark and Chalmers (1998) argue that when an external resource plays the right functional role in a cognitive process, it is part of that process. Hutchins (1995) shows mature cognitive work routinely running across people and artefacts rather than inside a single head. Dictation, scaffolds, structured environments and language models carry the serial labour of translating a whole into ordered parts. They do not think the thought. They render it.
Klein & Macoun (2024), New Ideas in Psychology, 75 · Hedlund et al. (2025), Frontiers in Sociology, 10 · Dantas et al. (2025), Autism in Adulthood · Lampinen et al. (2026), Autism, 30(1) · Clark & Chalmers (1998), Analysis, 58(1) · Hutchins (1995), Cognition in the Wild
08A forest, with a friend
A whole day, no translation in either direction. When the exact word would not arrive on cue, nothing broke — she waited, or she already knew, or the word turned out not to matter because the whole thing had already landed.
This is the node people assume is sentimental, and it is among the best-evidenced of the ten. Crompton and colleagues (2020) tested information transfer along chains of autistic, non-autistic, and mixed pairs. Autistic-to-autistic transfer was as effective as non-autistic-to-non-autistic. The mixed chains were where information degraded. The difficulty is in the crossing, not in us.
Which is Milton’s (2012) double empathy problem demonstrated experimentally, and it is why my forest day is data rather than luck. Change the interlocutor and my communicative competence changes, with nothing inside me having moved.
De Jaegher (2013) supplies the theoretical frame: sense-making as participatory, something enacted between people rather than performed by one and received by the other. Belek (2019), doing ethnography with autistic adults, shows how community concepts like overload and shutdown get learned and then filled with meaning — a shared vocabulary that makes experience sayable. And Mingus (2017) names the thing I actually had that day: access intimacy, the felt sense of someone getting your access needs without a negotiation.
A forest receives you whole or not at all, which happens to be the only way I arrive anywhere.
Crompton et al. (2020), Autism, 24(7) · Milton (2012), Disability & Society, 27(6) · De Jaegher (2013), Frontiers in Integrative Neuroscience, 7 · Belek (2019), Medical Anthropology, 38(1) · Mingus (2017), Leaving Evidence
Ninth environment · My university
A diverse, multicultural, humanistic institution where nobody requires me to be linear in order to be taken seriously. No verdict was issued, because no deficit presented. The same mind that could not produce an assessable impairment in an eligibility office produces graduate-level work here without accommodation, which the intrinsic-impairment account cannot explain.
Barab and Plucker (2002) asked the question this node answers: smart people, or smart contexts? Their argument is that ability is not a quantity inside a person but something that emerges in person–environment transactions — so an institution can constitute competence or preclude it. Feuerstein et al. (2010) make the methodological version: what a mind can do with appropriate mediation is a truer index of capacity than what it produces alone under standardised deprivation.
Botha and Gillespie-Lynch (2022) supply the reason diversity is the operative variable rather than kindness. Where difference is expected, the interpretive resources for it already exist — which is Fricker’s (2007) hermeneutical injustice solved by ordinary institutional design rather than by anyone’s good intentions.
I want the cost named, because it is the whole class argument. I did not receive this environment. I bought it, and the debt is beyond belief. Access I purchased is not access I was given, and a model that only works for people who can finance their own accommodation is not a model. It is a market.
Nobody here has ever asked me to be legible before being taken seriously. That is not a therapeutic achievement. It is an institutional design choice.
Barab & Plucker (2002), Educational Psychologist, 37(3) · Feuerstein, Feuerstein & Falik (2010), Beyond Smarter · Botha & Gillespie-Lynch (2022), Human Development, 66(2) · Fricker (2007), Epistemic Injustice
Tenth environment · My chosen community
Heavily diverse, and organised around a single principle: access comes before agreement. Nobody has to pass a purity test before their needs are met. Difference is the starting assumption rather than the thing being adjudicated. No verdict here either — and this is the node that separates my argument from the one it gets mistaken for.
Mingus (2017) gives the mechanism: access intimacy, the felt sense of someone understanding your access needs without a negotiation. Sins Invalid (2016) gives the organising principle in its founding form — interdependence and collective access, built by disabled queer and trans people of colour precisely because rights-based movements had centred one kind of disabled person.
Freeman (1972) explains why the structure has to be explicit rather than assumed: there is no such thing as a structureless group, only groups whose structures are unnamed and therefore unaccountable. And Táíwò (2022) names the failure mode this community avoids by design — elite capture, where a collective liberation practice gets narrowed to the priorities of whoever in the room is most comfortable.
The contrast with the comment thread is the finding. Both rooms are autistic. One is organised around access, the other around adjudicating who qualifies. I am not disabled in the first and I am disabled in the second, which means the variable is not autism and it is not other autistic people. It is what the room is for.
Two autistic rooms, opposite outcomes. The variable was never the people in them.
Mingus (2017), Leaving Evidence · Sins Invalid (2016), Skin, Tooth, and Bone · Freeman (1972), Berkeley Journal of Sociology, 17 · Táíwò (2022), Elite Capture
Companion to “How I Connected the Dots.” Every source cited here is in my research library and can be checked. Where a claim is mine rather than the literature’s, I have tried to say so in the sentence rather than in a footnote. A note on method: the synthesis and the argument are mine; I use AI as an access tool to hold many sources in view at once and to help render a whole I grasp all at once into sequential prose.