User:Oolong/injustice

This page in a nutshell: Wikipedia's coverage of autism-related topics is worse than it should be, partly because we aren't doing enough to correct epistemic injustice.

This essay has been published by Thinking Person's Guide to Autism under the title How Wikipedia Systematically Misleads People About Autism.

Wikipedia, the free online encyclopedia that anyone can edit*, aims to present “a neutrally written summary of existing mainstream knowledge in a fair and accurate manner”.

In many ways, it is phenomenally successful in its pursuit of this goal. Despite what its detractors say about it, it usually does a pretty good job of summarising the state of human knowledge about totally uncontroversial topics.

Its coverage of subjects like chemistry and maths, for example, is only occasionally wrong. A much bigger problem is its being edited by people who don’t know how to present complex information in an accessible way — which is not really anybody’s fault. The whole site (in common with this article) is mainly driven by the twin urges to infodump, and to correct people who are wrong on the internet. In many ways, that’s a beautiful thing! It’s powered the creation of an extraordinarily complete and passably reliable encyclopaedia. If it sometimes means readability comes as an afterthought, well, there are worse things.

Where Wikipedia’s model breaks down is chiefly where there is no general consensus on facts — especially where a topic is also relatively niche. Its editing is nominally consensus-based, after all; when a subject is objectively contentious, that gets tricky. The problem is worse when it is a subject that people are passionate about, and worse still in rapidly evolving fields.

As anyone familiar with autism research, practice or discourse should be aware, all of the above apply to autism.

Historically, psychiatrists and psychologists have been seen as the main authority on the nature of autism, even though most have no direct personal experience of it. Autistic activists, advocates and researchers^[1] have been pointing out for decades that non-autistic experts have been making grave and damaging errors in their descriptions of autism.

A large proportion of autistic people and those working with and around us have come to see traditional deficit-based, medical frameworks for understanding autistic people as limited and often unhelpful.

The neurodiversity paradigm^[2] situates differences like autism as aspects of human diversity to be accommodated and understood, rather than treated or eliminated. On this view, the problems faced by autistic people are not best understood in biomedical terms, but social ones, similar to those of other marginalised minorities.

Originating in autistic communities in the 1990s^[3], this perspective has gained more and more acceptance in recent years: among wider neurodivergent populations, notably ADHD communities, as well as among researchers and practitioners.

As well as reframing how neurodivergences (or ‘neurodevelopmental conditions’) are seen, neurodiversity implies a power shift: a change in who gets to say what autism is. This is a question of epistemic injustice — specifically testimonial injustice. Historically, autistic people have not been seen as authorities on our own experiences, with non-autistic psychologists and psychiatrists overwhelmingly setting the narrative. In other words, autistic people’s self-knowledge was largely dismissed, if not ignored outright. Many scientists have now come to see this as a scientific mistake, and an ethical failure.

These failures have had many practical and theoretical consequences. One glaring example is the Double Empathy Problem: as autistic scholar Damian Milton pointed out in 2012^[4], empathy always involves two or more people. It makes no sense to claim that autistic people have impaired empathy, once you realise that we have no more trouble empathising with non-autistic people than they have empathising with us. A series of very persuasive studies^{[5][6][7][8][9][10]} have backed up this point, starting about five years after Milton published. Empathising with people who have very different experiences from you is just difficult in general, especially if you haven’t heard very much about those experiences. Acknowledging this means recognising that vast swathes of autism research, and many popular descriptions, are built on extremely shaky foundations.

Again, this is just one example of the power of listening to autistic people, when it comes to the interpretation of our own experiences. It makes for much better science^[11][12], but paradigm shifts are always messy — the ‘normal science’ that researchers usually engage in is thrown into disarray, with extensive re-evaluation following.

In recent years there have been various academic neurodiversity conferences, such as the Stanford Neurodiversity Summit happening yearly since 2019, and ITAKOM 2023 (It Takes All Kinds of Minds) in the UK. We have seen best-sellers like NeuroTribes, as well as scholarly books like The Neurodiversity Reader^[13] and Critical Neurodiversity Studies^[14], among others^[15][16]. In 2023, Wiley announced a “multi-journal special issue” on neurodiversity across 38 journals, while Sage launched a scientific journal called Neurodiversity. Autism charities and government bodies talk about it more and more, and there is ever more discussion of how other developmental differences, like ADHD and dyslexia, fit into this picture.

In short, neurodiversity has gone mainstream. What was for a long time a slow-building movement has —at least within English-language academia and autism communities — become almost impossible to ignore.

Even so, it is fair to say that many researchers and medical professionals still view autism as some kind of pathology. So do some charities, although even Autism Speaks talks about neurodiversity now, and removed the word ‘cure’ from its mission statement in 2016 — although some might still doubt their sincerity.

Meanwhile (again, talking here about the English-speaking world) support for the neurodiversity paradigm seems to be somewhere close to universal among autistic-led organisations, and has been very much the norm in online autistic communities for many years.

In some important ways, the situation mirrors where we were a few decades back with sexuality: when homosexuality was still something that psychiatrists would diagnose, although most gay people did not see their sexual orientation as a mental illness.

It may seem obvious, looking back, that the justifications for treating same-sex attraction as a mental illness were always dubious, but specialists at the time would point to things like high rates of depression and anxiety in queer communities, difficulties fitting into structures like traditional marriage, and the supposed ‘naturalness’ or ‘normality’ of heterosexuality.

It took sustained social pressure, as well as eminent psychiatrists putting their careers and reputations on the line, for homosexuality (and other minority sexualities) to be removed from the diagnostic manuals and start to be accepted as an ordinary part of life. Wikipedia has an excellent entry on this, as it happens, but I have to wonder how it would have dealt with this shift as it was happening.

Another useful point of historical reference here is anthropology, which has been grappling with its colonialist legacy for about a century now. Historically, anthropologists — positioning themselves as ‘objective’ observers — would visit societies they had no first-hand knowledge of, and impose their own interpretations on customs and social structures. The people they were trying to describe typically had little or no control over how they were represented in published work about their lives.

Many modern anthropologists are quite embarrassed by the arrogance of their forebears. For some decades now, it has been common for them to recognise the importance of respecting different standpoints: that different perspectives allow for legitimately different understandings of things; that assuming a detached observer can arrive at objective truths about peoples, while their subjects are unreliable by default, is unjust and unsound.

The shift in anthropology partly reflects a recognition of the limits of the scientific method, when it comes to making sense of human beings. We should not (and in most contexts, do not) rely on scientific methodology when we seek to understand people or cultures. Careful observation, experimentation, and theorising can all be valuable tools for understanding humans; but so can conversations, stories and direct experience. Objectivity is not always possible in this context,^[17] and neither is it necessarily desirable.^[18]

Anthropology is by no means perfect as a field, but it has come a long way from its brazenly colonialist past. Psychology and medicine, likewise, have slowly started to recognise the importance of listening to what their subjects have to say about their own experiences. Medical practitioners are increasingly advised to approach patients with cultural humility; more and more autism researchers are pursuing participatory research, and citing autistic community sources in their work.

So where does all this leave Wikipedia, with its many substantial autism-related pages, collaboratively written over at least two decades? The short answer is that it doesn’t look good. Anyone looking up autism or any associated topic there will be presented with perspectives on autism and related matters which are widely considered outdated, and often harmful.^[19] Anyone seeking to bring its coverage up to date is likely to face deeply-engrained resistance, often from volunteer editors with many years of experience maintaining and policing the site.

Where disagreements exist, Wikipedia tries to aim for articles “with an impartial tone that document and explain major points of view, giving due weight for their prominence.” But what does ‘impartiality’ mean in light of irreconcilably different perspectives on a topic? How do you weigh up major points of view? Who decides the amount of weight to afford to different sources, when historically one group has wielded unchallenged power, but recent years have seen a profound shift in that?

As a first step, Wikipedia says it aims to “fairly represent all significant viewpoints that have been published by reliable sources, in proportion to the prominence of each viewpoint in those sources.”

The next question, of course, is what counts as “reliable sources”. Wikipedia’s general answer is “independent, published sources with a reputation for fact-checking and accuracy” in a relevant context. Newspapers are typically regarded as relatively reliable sources on Wikipedia, but in the context of autism, a series of studies^[20][21][22] have found that “newspapers focus on adversities associated with autism and adopt an ableist perspective, and tend to use negative and stereotypical language to refer to autism and autistic people.”^[22] Is that a problem for their reliability? I would say so, but if someone else disagrees, who decides who’s right?

With something as controversial as autism, different people have very different ideas about which sources are reliable. Should we be looking for sources that have a reputation for fact-checking and accuracy among autistic people? Among autism researchers? Practitioners? Family members? All of the above? What about random members of the public?

There is no consensus among autism researchers on what constitutes a reliable source, let alone in the wider autism community. Aside from the methodological and ethical implications of the neurodiversity paradigm, there are also autism researchers working in many different fields, who do not always share the same assumptions: psychology and neuroscience of course, but also disability studies, philosophy,^[23] anthropology,^[24] education, sociology and more.

Many autistic people, and increasingly many researchers, distrust autism research that hasn’t at least been fact-checked and reviewed by some of the people it’s about — and preferably conducted with autistic people from the start.^[25] It is very significant that Autism in Adulthood, with autistic people on its editorial team and at least one reviewing every paper, currently has the highest Impact Factor of any autism journal,^[26] despite only launching in 2019.

As Pellicano & den Houting argue in their 2022 Annual Research Review for the Journal of Child Psychology and Psychiatry,^[27] “Shifting from ‘normal science’ to neurodiversity in autism science”:

Ultimately, autism research has been characterised by a narrowness of perspective. By prioritising research on causation, we have failed to understand the nature of autistic people’s life experiences. Imposing nonautistic, medically driven priorities has had the grave consequence of diverting resources away from existing autistic people and the areas that matter most in their lives.

Speaking of medically driven priorities, Wikipedia — for sensible reasons — demands a higher standard of reliability for ‘biomedical information’. In short, “information that relates to (or could reasonably be perceived as relating to) human health” must be “based on reliable, third-party published secondary [and tertiary] sources, and must accurately reflect current knowledge.” That means that things like up-to-date review papers, textbooks and guidelines from health bodies are considered good sources about human health, while things like individual clinical studies are not. After all, biomedical studies read in isolation are often inconclusive in a way that can mislead laypeople, and it can be dangerous to make health decisions based on half-understood findings.

Of course, practically the whole point of the neurodiversity paradigm is that viewing autism and similar differences through a biomedical lens is often a mistake. Obviously not everything about autism is biomedical; like all humans, the experiences of autistic people have social, political and cultural dimensions. So who gets to decide what type, or standard, of evidence is appropriate where? For that matter, how do we decide how much biomedical information to include?

In principle, Wikipedians do their best to talk things through and arrive at a reasoned consensus on difficult questions like this.

Taking neurodiversity seriously has far-reaching implications for how we talk about autism and other kinds of neurodivergence, and also what is important to discuss. It suggests that it is a mistake to assume these differences are disorders or collections of deficits, for example; and that in many ways, a medical framing of something like autism is unhelpful. In line with the social model of disability, it suggests that it is more useful to talk about how best to accommodate neurodivergent people than to hunt for ‘cures’. The causes of autism might be of scientific interest, but they are not particularly interesting to most autistic people, our families or those who work with us; what matters more is understanding what it takes for autistic people to thrive.

Again, it is important to emphasise how mainstream these ideas actually are in 2024. One recent paper^[28] found that almost 98% of its 278 autistic subjects and 96% of its 226 non-autistic ‘autism community members’ had at least heard of neurodiversity, and a strong majority viewed it favourably. They generally endorsed societal reform, rejected efforts to make autistic people ‘normal’, and supported autistic leadership in autism-related matters.

Similarly, studies looking at community priorities for research^{[29][30][31][32][33][34][35]} (both among autistic people, and those who live and work with us) consistently find little interest in biomedical autism research. Instead, people want research on how to support autistic people’s mental health; access to education, healthcare, employment and wider society; how to improve communication; and better diagnosis.

It is safe to say, then, that the current structure of Wikipedia’s main Autism entry is not written to address what is important to most autistic people, their family members, or professionals working with them:

It is not clear who is this is for. It seems to follow a generic template designed for diseases, but autism is not considered a disease by reputable sources, and Wikipedia is not a scientific or medical encyclopaedia: its primary audience is not supposed to be scientists or medical professionals. People reading this entry are far more likely to be wondering if they or a loved one is autistic; or trying to understand what a recent diagnosis means; or hoping to learn what kind of thing can be helpful for autistic people.

The current entry is written as if it was there to provide health information, but it isn’t — and it shouldn’t be. In fact, given the higher standard of evidence that Wikipedia requires for biomedical information, much of this content probably shouldn’t be there at all, because it simply isn’t verifiable.

The main NHS page on autism^[36] provides an instructive contrast. For example, whereas Wikipedia spends eleven pages on ‘causes’ and has a whole separate entry on ‘causes of autism’, the NHS boils it down to this:

Nobody knows what causes autism, or if it has a cause. It can affect people in the same family. So it may sometimes be passed on to a child by their parents.

Then it has a quick bullet-point list of things it’s not caused by.

So if the autism coverage on Wikipedia isn’t doing what it ought to be doing, what is to be done? The obvious answer would be edit it — anyone can do that, after all — but any change that could be seen as controversial is likely to be reverted. The next step is to go to the Talk page.

Every entry on Wikipedia has a Talk page attached, where editors (which is to say, anyone who comes along, but especially people who are quite invested in Wikipedia) work to settle disputes and come to agreements. The current Talk:Autism page runs to 57 screens on my laptop, and that only takes us back to May of this year. Before that, the old Talk:Autism spectrum had 19 separate pages of archives. A look at the edit history for the Autism entry shows us that 500 edits have been made since last December. There really is very little consensus about anything to do with autism, and that really does present a problem for Wikipedia’s model of collaborative editing.

To help make sense of why Wikipedia’s autism pages are quite as bad as they are, let me describe a couple of my experiences of the site.

Back in early 2022, I poured many bafflingly stressful hours into what should have been a pretty uncontroversial change. You see, the main Autism entry at that point was not about what is officially called Autism Spectrum Disorder (ASD), but instead about what’s sometimes called ‘classic autism’: what used to be called Autistic Disorder, among other names, before the diagnosis was abolished and merged into ASD in DSM-5 in 2013. At one time, you did need to be careful to be clear whether you meant ‘Kanner’ type autism or the ‘Asperger’ kind, but people had been using ‘autism’ to mean both since long before all the variations were officially merged into one diagnosis. That was part of the problem, actually: much of the entry had obviously been written by people who were thinking about autism in general, not the obsolete sub-category.

Nine years after the merge, Wikipedia was almost unique in still filing ‘Kanner autism’ under just ‘Autism’. However, at least one long-term user clearly saw my retitling request as hostile, and tried to shut it down: declaring it ‘vandalism’, attempting to enlist help from the Wikipedia Discord and at least one of the editors who had contributed to the original article, many years before. My sense was that I had waded into an epic battle that long predated my involvement, in which an unknown number of veteran Wikipedians have made it their mission to fight the tide of autistic rights activists who sporadically try to depathologise its coverage.

This one editor insisted that most reliable sources didn’t agree with me; that this was biomedical information, and hence needed review articles or similar to justify any changes; that he had provided several citations showing I was wrong. All of this was objectively false: even his own citations either backed up my point or left it untouched, and eventually I prevailed, with the help of a couple of friendly editors, backup from a helpful professor, and fifteen or so citations — several of them providing completely unnecessary biomedical-quality verification.

This was an exhausting and frankly crazy-making process. It was also a fascinating glimpse into Wikipedia’s sticky internal workings. I’d had an account on the site since 2005, and I’ve made well over a thousand edits by now, but I had never really waded into any major stooshies before. When my edit-blocker tried to threaten me with an “Arb-Com discretionary sanction”, I had to go and look up Wikipedia’s formal dispute resolution processes to satisfy myself that this was a totally hollow threat. I should say here that Wikipedia insists on a strong presumption in favour of good faith in disagreements… which is nice, if sometimes hard to maintain. I actually do think this person was sincere, despite being so infuriatingly and doggedly wrong.

I and some other editors tried to plan a series of other improvements to the page. We did make some inroads, but it was a huge job, and we stalled — the article is ridiculously long, covering far too much ground for one entry, and while we were trying to build a consensus-based re-draft, people kept editing the entry as it was. The proposed outline we came up with is now buried over here.

I did rewrite the lead section to make it clearer and more concise, and to use neutral terms where appropriate, rather than language taking a pathologising view for granted, but most of these changes were undone before very long.

By this point I was noticing that spending time on Wikipedia was doing terrible things to my blood pressure.

Unfortunately, while the main autism entry is something like a sprawling catacomb of inadequately-sourced, unhelpful and occasionally downright harmful information, the entry on Applied Behavior Analysis (ABA) is in some ways even worse. Written like a sales pitch for a widely celebrated scientific discipline, only someone utterly unfamiliar with public discourse about ABA could imagine that this entry was anywhere close to neutral.

Of course, many people are entirely unfamiliar with ABA. If that includes you, you should know that ABA is a kind of psychological intervention which is widely practised in the USA and elsewhere, mainly on autistic children — many American health insurers actually refuse to pay for anything much else that’s autism-related.

ABA is very unpopular in online autistic communities.^[37] Many who have experienced it describe it as a kind of systematised abuse.^[38][39][40] It is also notably controversial among psychologists:^{[41][42][43][44]} several meta-analyses have failed to find significant evidence of any lasting benefit;^[45][46] grave concerns have been expressed about the systematic failure of researchers to investigate likely long-term harm;^{[47][48][49][50][51]} studies have overwhelmingly been authored by researchers with undisclosed conflicts of interest,^[52][53] likely resulting in substantial publication bias and motivated reasoning. Much of the work that initiated the field was unethical on a scientific and human level,^[54][55] conducted by some of the same people who pioneered pseudoscientific conversion therapy for gay people.^[56][57] ABA itself is sometimes described as conversion therapy for autistic people.^[58]

One of the USA’s major professional bodies for ABA, ABAI, spent many years taking money from the Judge Rotenberg Center — which gained international notoriety (and UN condemnation) for its systematic use of electric torture machines on disabled children — and inviting its representatives to give talks at their conferences defending these practices.^[59] Finally its members voted to oppose these practices in 2022,^[60] but not to stop inviting JRC to their conferences.

ABA is controversial. It is also a multi-billion-dollar industry. You may be wondering, at this point, how Wikipedia summarises this controversy…

ABA is considered controversial within the autism rights movement due to a perception that it emphasizes normalization instead of acceptance, and a history of, in some forms of ABA and its predecessors, the use of aversives, such as electric shocks.

This is kind of an interesting case study in how to mislead as much as possible in a few lines, without technically saying anything untrue.

ABA is considered controversial because it is objectively controversial, within the autism rights movement and also among scientists, parents and practitioners due to a generally correct perception that it emphasizes normalization instead of acceptance, and a history of, in some forms of ABA and its predecessors, the use of aversives, such as electric shocks, among many other reasons.

As an aside it is, as far as anyone can tell, not particularly controversial among autistic people, let alone the autism rights movement: mostly, we just hate it. It’s impossible to get guaranteed bias-free statistics on this sort of thing, but Chris Bonnello’s huge 2022 survey^[61] found that less than 4% of its 6,576 autistic respondents supported the use of ABA on autistic children. Among the 341 autistic respondents with direct personal experience of ABA, 62% were strongly opposed to this use.

Of course, despite the numbers, and Bonnello’s status as a well-respected author and speaker on autism with a diverse following, this survey doesn’t necessarily meet Wikipedia’s standard for a ‘reliable source’. Self-published blogs are usually only considered reliable “when produced by an established expert on the subject matter, whose work in the relevant field has previously been published by reliable, independent publications.” I would certainly consider Bonnello a subject matter expert, but not necessarily one that fits these precise criteria. Similarly, while his hugely impressive poll and analysis could be published in an academic journal, it hasn’t been, and maybe being able to cite it on Wikipedia is not quite a good enough reason to go through that process.

One editor told me that along with this survey, “advocacy groups are not neutral, reliable sources. That means no blogs from Neuroclastic, ASAN, or Thinking Person’s Guide to Autism. Or Autism Speaks.” To be clear, Wikipedia does not have a blanket ban on citing ‘advocacy groups’, and “reliable sources are not required to be neutral, unbiased, or objective. Sometimes non-neutral sources are the best possible sources for supporting information about the different viewpoints held on a subject.” Of course, many people would simply take this user at their word rather than checking.

Sources are required to have a reputation for fact-checking and accuracy; I would argue that Thinking Person’s Guide to Autism and ASAN have a better reputation for that among autistic people than any single newspaper, and several major academic autism journals — but again, whose gets to decide? Crucially, autistic people still have relatively few outlets in the mainstream media — most news articles about autism include no autistic voices, for example.

My experience editing the ABA page was even worse than with the main Autism entry. One editor who seems heavily invested in ABA would simply dismiss concerns raised, determinedly revert edits, and act very offended by any suggestion that they were somehow taking a side. Another persistently came back to revert edits without ever engaging. I did eventually manage to secure agreement to remove the long, barely-sourced list of supposed applications of ABA which are never mentioned in the body of the entry, and add a mention of the fact that its weak evidence base is one reason it’s controversial, but all of this was reverted when I came back a few months later. The current text (as of mid-October 2024) is almost identical to the version I had tried to change, back in early 2022.

Unsurprisingly, a whole string of editors have come along and tried to change this entry to better reflect the actual controversy, but all it takes is a couple of very determined people and maybe the odd vaguely reputable-looking source to wave at, and it becomes gruellingly difficult for anyone else to make any meaningful changes. It is easy to get worn out trying to fix one sentence, while the rest of an entry remains just as problematic.

In theory, Wikipedia’s dispute resolution procedures could help here, but my attempt was closed down after a couple of days because I missed a message I hadn’t been tagged in. I was too tired to try again, and nobody else has picked this back up.

The specific thing that sent me on the train of thought leading to this article was seeing Wikipedia’s entry on The Reason I Jump, which describes it as “a biography attributed to Naoki Higashida” and continues

The book alleges that its author, Higashida, learned to communicate using the scientifically discredited techniques of facilitated communication [FC] and rapid prompting [RPM], which raises suspicions about the book’s authorship.

“Researchers dismiss the authenticity of Higashida’s writings” is later stated as a simple fact; the general impression given by the article is that only credulous goofballs would think Naoki Higashida actually wrote this book.

I had a feeling that this was not a fair summary of the mainstream consensus on the authorship of this book. I knew that Higashida has been filmed writing entirely independently (including in a video linked from the article, as it turns out); that the director of the film loosely based on the book described conversing with him in a blog post;^[62] and that Higashida has made various public appearances, and published many more books. As a side note, it also turns out there isn’t even anything in The Reason I Jump about FC or RPM — his ‘alphabet grid’ system seems conceptually similar, but it is not clear that the same criticisms apply at all.

A quick search confirmed that while there certainly was some controversy around the time of its writing, it seems to be generally accepted by now that Higashida did in fact write his books.

I was reminded that back in 2019, NeuroClastic published a piece about Wikipedia, FC and RPM. I think it’s fair to say that these are controversial methods of communication,^[63] and there seem to be legitimate concerns about their potential for abuse. I have heard more than one autistic academic express worries about them, but it doesn’t seem to be true that there is a strong academic consensus that they are never valid. Indeed, from what I can gather, various nonspeaking autistic people have progressed from their use to fully independent communication, and have not disowned the communication attributed to them from before that transition.

One 2022 ‘perspective’ piece published by Frontiers in Psychology^[64] reviews the available evidence, and concludes that we “must consider the possibility that assisted typing is valid and offers a flexible communication tool for self-expression for certain individuals.”

Higashida was one of several notable autistic people whose Wikipedia entries were deleted (and archived on NeuroClastic). Others include Tito Mukhopadhyay, widely considered the author of eight books, who you can watch in conversation at CASY on YouTube; and the poet and long-term ASAN board member Amy Sequenzia.

I have never felt well-placed to reach a firm conclusion about either RPM or FC, but in excluding any information that suggests that anything similar could possibly be legitimate, Wikipedia is going way beyond anything resembling a mainstream consensus. It is excluding many nonspeaking advocates from consideration, because of a controversial technique that many of them only use some of the time. Again, all that this takes is a couple of sufficiently determined editors with an axe to grind, and some superficially plausible sources they can point to.

So far, my focus has been on matters around testimonial injustice — unfairness related to whose word is trusted. The other kind of epistemic injustice is hermeneutical injustice, which happens when someone’s experiences are not understood — including by themselves — because they, or others, lack the concepts needed to make sense of them. ‘Hermeneutical’ means ‘relating to interpretation’; the injustice comes in when people or populations aren’t able to access ideas that could allow them to better interpret their lives. For example, people may not recognise that they are being abused or exploited if they don’t know what that can look like. They might not realise that other people experience many of the same things — that their experiences have a name and follow a pattern, that other people have found helpful approaches to their problems.

Things like encyclopaedias have a powerful role to play in combating this kind of injustice, by providing people with the conceptual tools needed to make sense of their own lives and those of others. I imagine this is exactly the sort of thing Denis Diderot was getting at in the quote from Encyclopédie included in Wikipedia’s page about its own purpose — collecting knowledge “so that our offspring, becoming better instructed, will at the same time become more virtuous and happy.”

In order to benefit readers in this way, an encyclopaedia needs to be written with an eye to what is likely to be useful. This is where testimonial and hermeneutical injustice come together: the people best-placed to identify what is useful for people to know are the ones most affected by whether people know it.

This is a big part of why ‘Nothing About Us Without Us’ has always been a rallying cry for disability justice movements, among others; it’s why the UN Convention on the Rights of Persons with Disabilities demands that disabled people (and our representative organisations^[65]) should always be closely consulted with and actively involved in decision-making related to disability. It’s also why autistic-led autism training tends to be rated so highly.^[66][67][68]

Hermeneutical injustice is a huge issue for autistic people and neurodivergent and disabled people more broadly — as well as for relevant parents, carers and professionals. For example, many neurodivergent people will tell you that they thought they were stupid, or lazy, or just needed to try even harder, because they didn’t understand why they found things so difficult, when other people insisted they should be easy. Realising that a person is in fact disabled completely changes how they and others understand their challenges — especially in the context of the social model of disability, which suggests that when things are difficult for disabled people, it is usually because they are not set up for people like them.

Becoming aware of common aspects of autistic experience can help make sure you’re prepared for them and have some idea how they might be accommodated — or avoided. For example, meltdowns^[69] and shutdowns^[70][71] can be terrifying, even debilitating: very much more so if you don’t understand them, or what triggers them (sensory overwhelm, for example^[72]). Burnout is a danger every autistic person should be aware of.^[73][74] Monotropism has helped many people to make sense of a wide variety of autistic experiences.^[75] All of these have received some attention from researchers,^[76] but not as much as they should; all are well-known in autistic communities. Autistic burnout and monotropism do get a sentence each in the Autism article, and their own entries elsewhere. The rest aren’t mentioned at all.

Meanwhile, wrong ideas about autistic and disabled people can be immensely damaging, contributing to stigma, exclusion and negative self-image.^[77][78][79] Take, for example, the idea that autistic people are generally incapable of empathy. Despite all attempts to soften the blow by distinguishing cognitive from affective empathy, the fact is that people in general see a lack of empathy as a terrible thing. Believing that someone inherently lacks empathy, they are likely to assume that they don’t (and can’t) understand other people’s perspectives, and that any misunderstandings are usually their fault.

The concept of the Double Empathy Problem (described earlier) has the power to right these injustices, by opening up a completely different interpretation of what’s happening: one that, as it turns out, is well supported by empirical research.^[80]

The Double Empathy Problem does get a mention in the main Wikipedia entry on autism. It’s about 16 pages down, while the second sentence of the article opens with the statement that “Autism generally affects a person’s ability to understand and connect with others”.

I have met more than one person who was convinced they couldn’t be autistic after reading about it on Wikipedia, only to realise years later that they definitely were (and yes, they had this confirmed by specialists). Others have read it only after they were assessed, with a sinking, shrinking feeling.

A series of studies have shown that the development of positive autistic identity has real benefits for mental health.^[81][82][83] On the other hand, reading negative, dehumanising claims about yourself or your loved ones is obviously disheartening. It would be one thing if all of these claims could be shown to be objectively accurate, but that is far from the case; in many cases, the scientific case for them is in fact very shaky, and avoiding ableist language adds to accuracy,^[84] and comes closer to neutrality.^[85]

Our collective understanding of autism — and of neurodivergence more broadly — is evolving at great speed.^[86][87] That is largely thanks to the paradigm shift I have discussed, and the accompanying trend of actually listening to autistic people about our experiences.^[88][89] Wikipedia cannot be expected to be fully up to date, while things are shifting, and consensus is so elusive: no encyclopaedia focused on summarising established knowledge ever could be.

There is no excuse for Wikipedia being as badly out of date as it is, though, given the impact it has.

It is hard to say exactly how much of the issue is down to individuals, and how much could be fixed by clearer guidelines and other institutional fixes.^[90] The questions peppered through my text here are not purely rhetorical — they are mostly things left unanswered by guidelines as they stand.

As far as the content goes, my sense is that the main Autism entry probably needs to be gutted and planned out from scratch, with an eye to what most readers will actually benefit from reading. Some of its content could be moved into separate entries, as has already happened with ‘pathophysiology of autism’ (and should happen with the ‘causes’ section). There may be large chunks that could be lost entirely, particularly where biomedical claims are not adequately supported by up-to-date studies.

There seems very little to recommend the disease-based structure the article currently has, except that it is shared (to a point) with other entries about conditions and disorders. If Wikipedians were committed to keeping this structure for whatever reason, it might be wise to go the same way they have gone with transgender and gender dysphoria: have one article for the phenomenon in general, and a separate one specifically for the ‘condition’ as described by diagnostic manuals. This may be the kind of compromise that leaves everybody unhappy, though.

Several related articles bear the obvious marks of people with an anti-neurodiversity axe to grind, with an incredulity that goes far beyond balanced scepticism. This problem is less pronounced than it was a few years ago, perhaps partly thanks to one prolific anti-neurodiversity activist being banned in 2020, but in places it remains glaring. I do not expect Wikipedia to be generally neurodiversity-affirming as long as the rest of the world is not, but it is failing in its mission when it goes to the other extreme.

It could help if more people who know a bit about autism and neurodiversity spent a bit of time working together on Wikipedia’s coverage — perhaps reviving the dormant WikiProject Autism. The site undoubtedly has many autistic editors already — infodumping and correcting people who are wrong on the internet is autistic culture, after all — but not all of them know it, and most are more interested in editing other things anyway. Out of those with a special interest in this topic, many must have tried over the years to fix some of these problems, and been worn down by the resistance. Others have helped keep that resistance going.

As I said at the start, Wikipedia is the encyclopaedia that anyone can edit*. That doesn’t have to be you, but if you’ve read this, I hope you can see why someone needs to do it.

With thanks to David Jackson-Perry, Krista Maxwell, Avery Rowe, Helen Edgar and Sonny Hallett for helpful conversations and feedback.

*Two caveats are called for here: first, because anybody can edit Wikipedia, anything I describe or quote here may have changed by the time you look! Second, the main autism entry is ‘semi-protected’, apparently to prevent vandalism. The Talk page is open to anyone, but the article itself can only be directly edited by users with accounts that are at least 4 days old and have made at least 10 edits; if you’re interested in participating, you may like to go and improve some writing or make some corrections elsewhere on the site first!