A Reflection on the Glass Onion of the Autism Severity Debate

There is a debate which has pervaded the field of autism research and Autistic and wider autism communities for decades. I takes approximately 400 different forms but, increasingly, I think has a remarkably consistent core.

Sometimes, it’s about autism “levels”. Sometimes, it’s about support needs. Sometimes, it’s “profound autism”. Sometimes, it’s whether late-diagnosed Autistic adults are really Autistic. Sometimes, it’s parents insisting that their child is not represented in Autistic advocates.

The details change, but the basic taxonomy remains remarkably stable: There are Autistic people who have it easy (Autism Lite) and then there are Autistic people who truly suffer (The Real Autism®).

Before I go further, a slight detour.

The Glass Onion

Throughout my life, the lesson I continue to refuse to learn is that most arguments are a glass onion, and yet I insist on peeling back every goddamn layer anyway.

In Glass Onion: A Knives Out Mystery, detective Benoit Blanc says, “I keep returning in my mind to the Glass Onion. Something that seems densely layered... but in fact, the centre is in plain sight.” This resonated with me because I have always had an ability to “cut the shit”, so to speak.

I find it endlessly frustrating when people construct elaborate intellectual justifications around a position rather than saying what they truly feel, fear, value, or want. Humans are remarkably good at dressing emotional, biased, and self-serving beliefs in more socially acceptable clothes. And I am remarkably good at wasting my energy arguing with the clothes, despite my instincts telling me it’s a glass onion.

I present evidence, identify factual inaccuracies, point out logical inconsistencies, carefully interrogate the premise, and ask philosophical and ethical questions. I look for the common goal we share and try to move us both toward it. And then I find myself wondering why we haven’t gotten anywhere. Because sometimes the argument was never really about the argument.

If you disprove Claim A and someone immediately produces Claim B, and then you address Claim B and suddenly we are discussing Claim C, eventually it might be worth considering that the claims are not actually driving the conclusion. The conclusion came first; the arguments are being recruited to protect it. And nothing drives me more bananas than shifting goal posts and a lack of transparency.

Back to Autism

For years, I kept approaching debates about autism “severity” (or whatever form / label it took at the time) as though everyone involved shared the same basic premise: Autistic people are fully human, and the goal is to improve Autistic people’s lives.

Great. Common ground! Now we can examine evidence, interrogate assumptions, and work out how best to achieve that goal. Except, over and over again, that isn’t what happened. From researchers and clinicians, I would encounter arguments riddled with logical fallacies, a refusal to critically evaluate evidence, and a wilful ignorance of the lived experience of #ActuallyAutistic people. From the Autistic and wider autism communities, I would encounter emotionally loaded anecdotes, resource scarcity mindsets, and a wilful ignorance of the lived experience of other #ActuallyAutistic people.

“You don’t understand severe autism.”

“You people should be grateful you can even debate on the internet.”

“I’m the parent of a child with autism!”

That last one is often presented as though loving an Autistic person renders someone incapable of misunderstanding, dehumanising, or unintentionally harming them. Unfortunately, loving another human has never bestowed omniscience upon anyone.

So, I would engage with the stated argument. I would point out factual inaccuracies, identify and question the underlying assumptions, present evidence, try to give the benefit of the doubt, remind myself to remain compassionate in the face of emotion dressed up as logic. I would explain that speech and intelligence are different constructs, that observable “independence” tells us little about the support someone needs, that Autistic people with fluent speech can have substantial support needs, that nonspeaking Autistic people can have high cognitive ability.

Many of the difficulties presented as evidence of “severe autism” are, in fact, conflated co-occurring disabilities and complexities that are not part of autism (though they do, of course, shape an Autistic person’s experience of disability). And that does not make those experiences any less valid or worthy of support; it simply means that, objectively, they are not evidence in an argument about autism severity.

I would highlight that, despite our desire as humans to contain our existential uncertainty with neat, categorical truths, diagnostic categories cannot neatly partition human beings into “mild” and “severe”. Autism is heterogenous.

And so, I thought, if our purported shared goal is improving Autistic people’s lives, surely better information should bring us closer together?

Reader, it did not.

There was always another “but what about…?” Another shifted goalpost. For a long time, I found this absolutely, mind-numbingly frustrating. Because, again, I refused to learn the lesson to trust my instincts telling me it’s a glass onion.

The Centre of the Onion

Here it is, sitting in plain sight: Humans dehumanise Autistic people. Not only non-Autistic strangers; parents can dehumanise Autistic people; professionals can dehumanise Autistic people; researchers can dehumanise Autistic people; Autistic people can dehumanise other Autistic people.

And when an Autistic person’s existence, competence, suffering, joy, communication, autonomy, or perspective threatens the hierarchy someone has constructed, people will perform extraordinary cognitive calisthenics to put that person back where they belong, to maintain a sense of stability.

I don’t think this has ever really been about intelligence, verbal speech, toileting, employment, or living “independently” (the colony froth that one). Those things matter when we are trying to understand an individual person’s needs and make sure those needs are met. But, in this particular debate, I increasingly think they are red herrings, because the boundary keeps moving. What seems to remain consistent, though, is the purported need for a boundary.

A boundary between which Autistic people we perceive as being less than us vs those whom we feel threatened by. And reader, that boundary is spectacularly subjective. And that is why I believe this debate continues. No one wants to reflect on their implicit biases or admit their true motivations, whether that’s to themselves or others.

Take one of the classics: “None of you so-called Autistic advocates speak for my child*. You can type on the internet, and my child can’t even talk.” ^{*Just quickly - of course adults aren’t like a 4-year-old, but I digress.}

Okay. Let’s peel the onion.

The first premise is that someone typing on the internet can therefore communicate using mouth speech. False. Plenty of nonspeaking or minimally speaking Autistic people communicate through typing, AAC, or other forms of communication.

Perhaps the actual distinction is independence? Still no. There are Autistic people of average or “gifted” so-called intelligence who cannot safely live independently, reliably prepare food, manage personal care, navigate the community alone, or complete many activities of daily living without substantial support.

Perhaps it’s cognitive ability? Again, no. If intellectual disability is what we are actually discussing, we already have language for intellectual disability. It isn’t synonymous with autism.

So perhaps the argument is simply that one Autistic person cannot speak for another. Sure. That’s true. No Autistic person speaks for every Autistic person. But neither does a parent, psychologist, teacher, nor researcher. And if the genuine concern were that autism advocacy needs to represent the experiences of Autistic people with the highest support needs, the logical response would be to ask:

How do we make sure those Autistic people are included? How do we support communication? How do we hear from people who use AAC? How do we include people with intellectual disability? How do we gather assent and recognise refusal? How do we make research and consultation accessible? How do we avoid allowing the easiest Autistic people to interview to become proxies for everyone else?

Those are excellent questions. But “[insert biased perception here], therefore your perspective is irrelevant” does precisely none of that. It doesn’t amplify Autistic people with higher support needs, it just silences another Autistic person. And that’s where I think the centre of the onion becomes visible.

Maybe the Threat is the Point

What if the distinction isn’t really between “mild autism” and “severe autism”?

What if, at least some of the time, the meaningful distinction is between Autistic people whose existence confirms someone’s understanding of autism and Autistic people whose existence threatens it?

Because an articulate Autistic adult saying, “I have very substantial support needs” creates a problem if you believe articulate people have Autism Lite™. A nonspeaking Autistic person demonstrating sophisticated thinking creates a problem if you have equated speech with cognition. An Autistic person saying, “That intervention traumatised me” creates a problem if your identity is partly organised around wanting to do good and believing that intervention was necessary.

An Autistic adult recognising their younger self in your child can create a problem if you have spent years believing that people like your child simply do not grow into people like them. And an Autistic person who is happy to be Autistic creates an enormous problem if you understand autism primarily as pathology or tragedy. Suddenly, we aren’t debating diagnostic terminology anymore. We are dealing with threat.

“I don’t want to confront the possibility that I may have misunderstood my child.”

“I don’t want to consider that behaviour I interpreted as meaningless may have been communication.”

“I don’t want to consider that something I did with loving intentions may have caused harm.”

“I am terrified that recognising other Autistics’ needs will mean fewer resources for my child.”

“I need there to be a category separating my child’s suffering from someone’s identity.”

“I don’t want to consider that I might be Autistic too.”

Those are much harder things to say. They are also deeply human. Fear, grief, shame, uncertainty, scarcity, and threat are powerful motivators; I don’t think acknowledging that requires us to imagine everyone involved as consciously malicious. At the same time, an understandable psychological defence can still dehumanise another person. And dressing it up as an argument about “severity” doesn’t make it objective.

Without an Honest Discussion, There Is No Resolution

I used to think we could resolve these debates with better evidence, defining support needs more dimensionally, explaining the limits of psychometrics, talking about adaptive functioning and fluctuating capacity. Maybe we could stop trying to conflate co-occurring disabilities with autism and develop more precise terminology. And those things are still worth doing. But none of that can solve a hierarchy that is fundamentally about who we perceive as fully human.

Give us a “better” classification system and, if we leave that hierarchy untouched, we will simply use the new words to rebuild the old categories. We will find another way to separate people whose internal worlds we are expected to respect from people whose internal worlds we feel entitled to interpret; another way to distinguish the Autistic people allowed to have preferences from those whose preferences become “behaviours”; the people whose distress is communication from those whose distress is a symptom; the people whose refusal is a boundary from those whose refusal is noncompliance.

And perhaps that is why I have found these debates so exhausting. I thought we were disagreeing about the layers; so, I kept peeling, ignoring the core sitting there in plain sight. The question underneath all the others was much simpler:

Why are we so invested in having a “severity” boundary, however we dress it up, and what does that division make it socially or morally permissible to believe about Autistic people—or to do to them?