Barnes’s Defence of ‘Mere-Difference’

We will demonstrate what we think is a problem with the mere-difference view by addressing the argument Barnes offers in her paper mentioned above. Of course, as with most such loosely characterized views, there will be variants among defenders and disagreements with some concomitant claims. Nevertheless, we take it that most who hold the mere-difference view will be inclined to make claims that disability is analogous to features such as gender, sexuality, or ethnicity in morally relevant respects, that disability is not a ‘defect’ or departure from ‘normal functioning,’ and that the principal source of bad effects of disability are due to society’s treatment of disabled people rather than the disability itself.

In her paper, Barnes does not provide positive arguments in favour of the mere-difference view, but instead defends it against two objections that are taken by some to show that the view is a non-starter. The two objections are based on a pair of claims.

1. (1) If the mere-difference view of disability were correct, it would be permissible to cause disability.

2. (2) If the mere-difference view of disability were correct, it would be impermissible to ‘cure’ disability.

Each of these, coupled with an appeal to an intuition that Barnes (plausibly) grants is probably widely shared—the first with the intuition that it is wrong to cause disability and the second that it is permissible to cure disabilities—gives us an argument that the mere-difference view is not correct. Barnes attempts to undermine these arguments by considering a series of examples, a first to remove some irrelevant but tempting considerations from the field, then two pairs of seemingly analogous cases. She uses the first pair of cases to reject claim (1). She uses the second pair not to reject (2), but to undermine the intuition that it is permissible to cure disability. It will be worth our while to get these arguments laid out before moving to the task of evaluating them.

Barnes’s first example is intended to clear away a possible source of confusion. An obvious way to cause disability might involve Amy and Ben working with lasers in a lab. ^{Footnote 10} If Amy deliberately blinds Ben, we’re hardly likely to think she has done no wrong because ‘disability is a mere difference’ (and so Ben is no worse off). So, at least sometimes, causing disability is impermissible and, from (1), by modus tollens, we have an argument that the mere-difference view is incorrect.

Barnes plausibly diagnoses why this argument doesn’t work. She notes that this case introduces irrelevant complications that make its implications unclear. There is a big difference, after all, between being disabled and becoming disabled, and the transition involved in the latter is often a painful one. Cases like this one make it too easy to arrive at the conclusion that causing disability is wrong, since one is causing not only disability but also avoidable pain. One can easily agree that Amy has done something wrong without accepting that the wrong thing she has done is causing the disability per se.

Barnes, therefore, considers a modified version of causing disability. She calls the case ‘Disabled Baby.’ ^{Footnote 11} In it, what is in question is a person introducing a pre-birth genetic modification to a fetus that causes a baby to be born with a disability when it would not otherwise have been. In this case, there are no detectable transition costs, so, if there is wrongness here, it is not in the unnecessary pain of the transition. And yet, Barnes notes, most will hold that this genetic interference is morally unacceptable. Since this appears to show that causing disability is impermissible, we seem able to infer that the mere-difference view is wrong, too.

Barnes hopes to undermine this bit of reasoning by introducing a parallel case she calls ‘Baby Genes.’ Baby Genes is the same as Disabled Baby but for replacing “causes the baby to be born with a disability” with “causes the baby to be gay.” ^{Footnote 12} Here, she suggests, the intuition remains that the interference is morally wrong. But there is nothing wrong with being gay, so the problem in Baby Genes cannot be a matter of introducing a problematic difference in the baby. Being gay is, surely, a mere difference. Since the cases are taken by Barnes to be parallel in all other important respects, the source of the wrongness is plausibly the same in both cases. Her contention is that this shows that the problem in Disabled Baby is not causing disability, but instead is the same problem as lies behind the intuition of wrongness in Baby Genes, namely causing any significant, identity-determining feature of a person. Most of us hold, she notes, strong “non-interference intuitions” when it comes to such features, and it is the violation of strictures against interfering with others that causes us to recoil in both cases, rather than anything inherently problematic about either gayness or disability. ^{Footnote 13} She takes this to show that claim (1) is false.

In response to the argument based on claim (2), Barnes considers essentially the same two cases with the causation running in the other direction, so she gives them the apt names ‘Reverse Disabled Baby’ and ‘Reverse Baby Genes.’ In Reverse Disabled Baby, there is interference in utero with the genes of a baby to make her non-disabled where otherwise she would have been disabled. In Reverse Baby Genes, the interference changes the baby from gay to straight. Unlike the previous pair of cases that Barnes views as (for most people) intuitively similar, this time Barnes suggests that most people’s intuitions will run in the opposite direction, finding the latter change problematic while the former they might find laudable. Her contention, though, is that the cases should not be regarded differently.

Barnes chooses the example of being gay with a purpose. Many of the reasons that one might point to as justification for regarding Reverse Disabled Baby as problematic also applied to being gay until recently in the societies in which most of her readers live, and still hold in many places—being a person with those characteristics comes with risks of disadvantage and social stigma. She presumes that most readers will share her view that ‘curing’ someone of being gay is completely inappropriate. (We hope it goes without saying that we share them, but we do note that the results of the recent U.S. presidential election and its aftermath show that there is still some work to be done if we hope to justify confidence that this is a pervasive view in North America.) What’s more, from the present vantage point, most readers can see that the existence of social factors disadvantaging gay people in the not-too-distant past would not have made ‘curing’ someone of being gay appropriate in those not-too-distant times, either.

Barnes encourages us to think similarly of the Reverse Disabled Baby case. She suggests that the source of our sense that it is okay to ‘cure’ a disability in the way contemplated in the case is the same sort of prejudice as might have led us to say something similar about ‘curing’ being gay in the past. What she hopes to show is that the problem with the second argument is not with (2), but with the suggestion that intervening to prevent a disability is permissible.

This is the heart of Barnes’s defence of the mere-difference view in her paper. As noted, in the paper, she does not argue directly for the correctness of the mere-difference view. Instead, she defends the mere-difference idea against two arguments she takes to be the reasons people have for supposing the view to be clearly mistaken. We turn next to showing why we think that even this merely defensive case is mistaken and, as we will briefly suggest (without arguing in great detail), the same mistake undermines her positive case for the mere-difference view in The Minority Body. Barnes’s paper is otherwise filled with insightful remarks, and we surely are in broad agreement with her about most things to do with disability, including that “the intuitions of the privileged majority do not have a reliable track record to serve as reliable a guides to how we should think about the minority,” ^{Footnote 14} which as members of the privileged majority in this case makes us cautious about appealing to intuitions. Nevertheless, we think that Barnes’s core argument fails, and that its failure is instructive.

Indications of a Problem

It is striking that, in her discussion of the five cases, Barnes returns repeatedly to blindness and deafness as examples of the disability in question. Perhaps this is because these conditions, possibly along with autism, have a history of association with the most vocal advocates of versions of the mere-difference view. Yet, it is important to notice that Barnes wants her conclusion to be broader—in her paper, she says she is offering a defence of the mere-difference view for disability, not merely for a few select examples of disabilities.

What is the range of cases Barnes has in mind when she uses the word ‘disability’? Early in the paper she tells us: “For present purposes, I want to understand ‘disability’ as a term introduced by ostension. Think of paradigm cases of disability—mobility impairments, blindness, deafness, rheumatoid arthritis, achodroplasia, and so forth.” ^{Footnote 15} Let us look again at those analogies, but this time with other “paradigm cases” of disability in mind besides, for instance, deafness and blindness, the disabilities she appeals to in that part of her paper.

First, let’s look at rheumatoid arthritis, an example Barnes explicitly lists as a paradigm disability. First, recall Barnes’s discussion of Disabled Baby and Baby Genes. Is it clear that it is merely our ‘non-interference’ intuitions that tell us that it would be wrong to cause someone to have rheumatoid arthritis when otherwise she would not have? We think not. Conversely, in connection with Reverse Disabled Baby, it strikes us as especially implausible that it is merely a matter of prejudice to think there is something right about preventing it. Rheumatoid arthritis is a condition characterized not merely by mobility impairments but by significant pain and degeneration over time. A worthwhile way to voice the concern might be that this is a condition that comes not just with transition costs, but that inherently has costs. So, while the parallel to Baby Genes might show us that it is wrong for non-interference reasons to cause people to have rheumatoid arthritis when they otherwise would not, that’s not plausibly the only reason it’s wrong.

Moreover, while the range of any ostensive definition is going to be vague, it seems to us that if rheumatoid arthritis makes the list of disabilities, then perhaps so too should conditions such as multiple sclerosis, Parkinson’s disease, dementia, or amyotrophic lateral sclerosis (ALS). We find that with examples like these in mind the intuitions to which Barnes appeals are simply unavailable. For one thing, as noted, in conditions characterized by degeneration, there is no way to abstract away the transition costs as Barnes does when moving from the blinding-with-lasers case to the Baby Genes case.

Perhaps, then, it was a mistake of some sort for Barnes to include rheumatoid arthritis in her list of paradigm cases, and we should set these examples aside as diseases rather than disabilities. We think, though, that the problem is not resolved so easily. Consider disabilities that presumably appear on anyone’s ‘ostensive’ list, such as Down syndrome and autism. Since one area of important social progress in recent years has been the support made available for people with Down syndrome to build lives in the community—indeed, as one might optimistically describe the most successful situations, conditions evolving to the state where people create their own lives in the community—most people in many countries are familiar with some typical characteristics of people with Down syndrome. What may be less familiar are some of the complications that come with it: approximately 50% of people with Down syndrome have significant heart conditions, with 10-15% of them being serious enough to require life-saving surgery in the early years of life; people with Down syndrome are 10 to 20 times more likely to develop leukemia; there is a high prevalence of hearing and vision disorders, and a higher prevalence (and earlier onset) of macular degeneration, increased risk of early onset dementia, and shortened life expectancy. ^{Footnote 16} Autism likewise comes with various health risks. Of course, risk statistics are going to be open to debate for a group whose membership is imprecise, such as autism. Our point doesn’t depend on the numbers being very precise, though. Up to 40% of people diagnosed as being on the Autism spectrum suffer from a seizure disorder; estimates of the percentage with gastrointestinal disorders such as celiac disease, colitis, chronic constipation, or diarrhoea range between 5 and 85%; sleep disorders are common; and while not common there is an elevated risk of a range of conditions such as Angelman syndrome, fragile X syndrome, and others. ^{Footnote 17}

What happens if we have examples of disability of this sort in mind when considering Barnes’s discussion? Our contention is that, once again, her case no longer seems plausible.

First, it seems to us that these are paradigm cases of disability by anyone’s reckoning, and the accompanying conditions undermine the suggestion that the wrongness in causing these disabilities is nothing more than a violation of strictures against interference with others.

For the argument based on the converse examples, it is worth looking a bit more closely at Barnes’s discussion in which she seeks to undermine the intuitions, which she takes most people—or at least most of the readers she hopes to convince—to have, namely that it is acceptable to intervene to change a baby from disabled to not, but not to intervene to change a baby from gay to straight. She notes that when someone (a parent, say) considers making such an intervention it would be reasonable to note that there are risks that life would be, in various ways, harder for a disabled child. Such considerations, though, ought not to carry enough weight to tip the decision in favour of intervention. There are still many hardships that come with being gay, even in these days of radically changed social views about sexuality in the societies where Barnes seems to be assuming most of her readers live. Moreover, there once were, and in other places still are, much worse such hardships. Barnes’s suggestion is that even in these recently past times it would have been wrong to change a baby from gay to straight. What we ought to do instead, Barnes suggests, is struggle for a society that removes these hardships instead of defending an individual child by changing her in a way that steers her clear of them. Indeed, Barnes hints that there is a degree of political cowardice involved in defending an individual child rather than confronting the broader social issue. The point of the discussion is to invite readers to think the same way about changing a baby from disabled to not: to do so in order to remove hardships for the individual child is a reflection of prejudice, or at least of political cowardice. Here, though, we think it is instructive that Barnes’s discussion includes reference to disabilities such as deafness rather than other paradigm cases such as Down syndrome, autism, or rheumatoid arthritis. Her contention may have a certain degree of plausibility if the risks and hardships in question are things like not having ready access to resources and supports that will allow one to communicate effectively, for instance. It is highly plausible that this sort of lack of access to supports is a social failure, especially in a wealthy society, and so an injustice that calls for a political response. ^{Footnote 18} It is, to say the least, less clear that this is the right response when one has in mind protecting a child against a 20-fold increase in the odds of having leukemia. ^{Footnote 19}

Where the Problem Lies

We think that the considerations above show that the central argument of Barnes’s paper, as stated there, doesn’t hold up. We also think, though, that it is worth a more careful analysis of where the argument goes astray. In particular, our suggestion is that Barnes succumbs to a tempting but incorrect line of thinking that also shows up in other discussions of disability, with unfortunate effects.

The mere-difference view, and, more generally, most views that fall in the broader category we’ve been calling ‘social models’ of disability, rely upon a distinction between impairment and disability. ^{Footnote 20} Disability is seen as a relation between an impairment and a social context. This, as noted above, facilitates arguments that society is the source of disability and so ought to adapt in ways that eliminate the barriers that cause disablement. The ‘human diversity’ view of disability, and the mere-difference view in particular, suggests that the relevant variations associated with the disability are instances of human diversity analogous to height or athletic prowess, but also race and sexuality, and only become a disability when not accommodated by society. ^{Footnote 21} We think this distinction does not stand up to scrutiny for many of the instances where Barnes would want it to, for reasons that the problems with Barnes’s arguments reveal.

Consider again Down syndrome. According to the mere-difference view, it should be considered a matter of human diversity, one that should be accepted without an assumption that a cure is desired or desirable. What, though, are we to make of commonly co-occurring conditions, such as significant heart conditions? Since society did not cause the heart condition, it cannot remove it. Yet, of course, it may restrict people’s ability to live their daily lives as they desire, because of symptoms such as dizziness and shortness of breath. Is it acceptable to cure the heart condition of a person with Down syndrome?

Presuming the answer is ‘yes,’ it seems to us that advocates of the mere-difference view require another distinction. That is, they must distinguish ‘the Down syndrome itself,’ which is an example of human diversity that ought to be accepted and celebrated, from the heart problems, which are a medical condition that is neither an aspect of to-be-celebrated human diversity nor socially constructed. With this third category, one can view Down syndrome as a mere difference while also affirming that the heart condition is the appropriate target of medical care.

Alas, we think there is no principled line to be drawn between the variation and the medical conditions that frequently accompany it. On what grounds might we put different characteristics into one category or the other? The only likely criterion that occurs to us is whether the condition is one that we’re happy to say deserves to be cured. ^{Footnote 22} But if this is our criterion, then the mere-difference view of disability risks reducing to a tautology. Disability is a mere difference because those aspects frequently involved with any particular disability that deserve cure are not part of the disability, but are something else. Disability is not a bad difference because the bad parts don’t count as part of the disability.

A similar lesson can be drawn from another example, where the relation between the disability and the ‘co-occurring condition’ is of a somewhat different character. Consider a person who uses a wheelchair. This is perhaps the classic example that makes obvious why a social model of disability is plausible, since there are so many familiar examples of readily removable obstacles that society traditionally did not—but has recently begun to—remove for those who get around in wheelchairs. It seems clearly to be a disability constructed by society’s lack of attention to accessibility, and so that a person does not move around by use of her legs is perhaps one of the most plausible cases for the suggestion that a disability is a mere difference. But, once again, consider a case where the use of a wheelchair is required because of serious heart failure. Presuming that it is appropriate to attempt to cure the heart problem, to maintain the mere-difference view requires a sharp distinction between the mobility impairment and the heart condition that in this case is the source of it. But this careful parsing of what counts as a disability certainly doesn’t match up well with Barnes’s supposed ostensive definition of what disabilities are, since it counts as disability only the things that are not necessarily bad. But this again pushes the mere-difference view towards tautology: that what isn’t necessarily bad, is not necessarily bad.

There is another option for defenders of the mere-difference view, though it seems to us a counsel of despair. One might deny that matters such as heart failure, leukemia, degenerative conditions, and the rest are ‘bad differences’ and instead contend that they too are merely examples of human diversity. The harms involved in all these conditions, too, should be regarded as socially constructed. ^{Footnote 23} Indeed, one can find some who defend this sort of view by challenging the legitimacy of the notion of impairment. They do so by suggesting that the diagnostic criteria used to determine what counts as impairment are actually an expression of socially constructed values that narrowly define a range of ‘natural’ and ‘normal’ cases from the vast range of human diversity. ^{Footnote 24} Since the very notion of impairment is based on socially constructed norms, any ‘impairment’ is a potential candidate for the same sort of rehabilitation as has been experienced by other ‘conditions’ that were once pathologized. Left-handedness and homosexuality (to borrow two examples mentioned by Barnes) were once ‘conditions’ that were ‘diagnosable.’ That we no longer regard them as pathologies reflects the role of social values in drawing these distinctions and in how desirable human function is enforced.

We grant that this discussion makes an important point: (plausibly) diagnosis is never entirely value-free. We don’t think the point is of much use to the advocate of a mere-difference view in the present context, though. To use it for this purpose is to absurdly run together terminal cancer with left-handedness, and runs the risk of undermining the ability to define and talk about important things. ^{Footnote 25} Social values may well be involved in either sort of diagnosis, but to grant that much is not to grant that there is no crucial difference between the cases. Such an argument seems to make the mistake of supposing that sharing some feature involves sharing all relevant features. In the end, there are morally and philosophically crucial differences between things like disvaluing severe pain, loss of capacity for enjoyment of life, and early death, though social values lie behind this disvaluing, and things like the association of left-handedness with the devil. ^{Footnote 26}

While we can still allow for thoughtful critique of the role values play in the creation and application of diagnostic criteria, the ability to identify a condition such as heart failure as an objectively undesirable physical state is a reasonable requirement for a theory of disability to meet. As such, this alternative defence of the mere-difference view cannot provide a suitable response to our concerns.

We think this is real trouble for the mere-difference view. Either it depends on an implausibly sharp distinction between disability and accompanying medical conditions, or it denies there is any such distinction, which requires accepting absurd conclusions. Neither option adequately explains the role of impairments in a person’s life, and the lines between these categories are far more complex than the mere-difference view can account for.

While we will not go into all of the details of the positive defence of mere-difference that Barnes offers in her book, we want to point out that we think this same problem undermines a couple of its significant components.

At one point in the book, Barnes argues that things like chronic pain or a shortened lifespan might well be associated with disability, and yet this would be “perfectly compatible with the mere-difference view.” ^{Footnote 27} Consider an analogy: the shorter lifespan for males does not in itself make maleness a ‘bad difference,’ nor do the various “ineliminable … bad features of being female” make it a bad difference rather than a mere difference. ^{Footnote 28} The view is not that “no aspect of disability is a bad difference,” but that locally bad differences do not by themselves show that something is other than a mere difference. “One can maintain a mere-difference view of disability itself while still thinking that some features commonly associated with disability are bad-difference, and that disabled people would be better off without such features. Something can be neutral overall, but have aspects which are bad.” ^{Footnote 29}

But how is this supposed to be possible? Barnes points to Laura Hershey, whom she characterizes as a prominent advocate of the mere-difference view who nevertheless was an active supporter of medical research that would prolong the lifespan of persons with her own disability (muscular dystrophy). “What Hershey strongly objected to … was the idea that the goal of research was ultimately a ‘cure’ for the disability.” For, Barnes suggests, “chronic pain and shortened lifespan are a part of some people’s experience of disability, but we needn’t assume that they are an essential part of those disabilities.” ^{Footnote 30}

It is this distinction between what is ‘essentially part of a disability’ and things that are fit to be cured that we take the arguments above to have shown to be unprincipled, or at best so far unexplained. It is instructive, we think, that Barnes does not defend it, but simply appeals to it, when addressing the problem of ‘associated features’ that are clearly negative. We take this to be evidence that the problematic line of thought we have identified is tempting enough to pass unnoticed, especially among those who find its conclusion attractive.

The other line of mere-difference offered by Barnes is that a disability that comes with negative features might also come with compensating positive features as well. “The very same thing—blindness—that’s a local bad with respect to your ability to visually appreciate the faces of your loved ones can also be a local good for you with respect to other things you care about.” ^{Footnote 31} These might be less susceptibility to prejudice, immunity to vanity, liberation from certain cultural norms, or things like a special epistemological standpoint and so access to sources of knowledge not readily available to others.

Our problem with this line or argument is, again, that it seems to be one that is grounded in intuitions about some particular sorts of disabilities, but the conclusion is, it seems, intended to be broader. Note again the use of the go-to example of blindness, where the suggestion that there are ‘compensating benefits’ is already part of popular culture—so much so that Barnes has to distance herself from the ‘X-Men View’ that the disability comes with “enhanced abilities only available to the disabled.” ^{Footnote 32} In this case, the intuition that the down-side and the up-side of the various accompanying conditions attendant on the disability will often simply balance might be reasonable enough. But considerations like the ones offered above indicate to us that we will not find the suggestion that we should expect the pluses and minuses to balance out so reasonable when considering other conditions that also should count as disabilities. ^{Footnote 33}

When one supposes that a person’s experience can be neatly divided into the medical aspects for which a person can seek care, on one hand, and an impairment which, on pain of a just charge of creating a disability, society ought to accommodate as part of the natural range of human diversity, on the other, one assumes a kind of disentanglement of a disability that misrepresents the lives of real people. It does not acknowledge the fuzzy aspects of a diagnosis that are neither straightforwardly solely medical nor entirely social, and precludes anyone from admitting that these aspects are undesirable; a person’s honest account of her own life may be dismissed as an expression of a politically retrograde ‘bad disability’ ideology.

A Broader Lesson

We turn next, more briefly, to political matters. As we noted at the outset of this paper, it has long been recognized that adoption of a particular model of disability is, at least for most participants in the discussion, not a merely theoretical matter. From the start, the social model was advanced not merely as a more accurate characterization of disability, but as a view that made successful advocacy for social change more likely, and subsequent refinements within broadly social models have likewise been evaluated on such grounds. So one question on the table is whether the mere-difference view might be politically useful, and something we ought to espouse (or in other ways make use of) even while recognizing that it is not strictly correct. In that case, appeal to the mere-difference view would be an example of the ‘calculated overstatement’ that disability advocates sometimes use. ^{Footnote 34} Calculation of political usefulness is, of course, a fraught business, so our judgements on this score are ones offered with considerable unspoken hedging, but we think that, in the long run, the answer is ‘no.’

First, we do not see much plausibility in the suggestion that a mere-difference view will have appreciable rhetorical benefits when compared with more accurate, more nuanced accounts of disability that recognize, for instance, some of the complications described above.

Of course, the most fundamental rhetorical value of any model of disability will be in counteracting the naïve views many people first bring to the issue of disability, where disability issues are simply assumed to be fundamentally ‘medical’ matters, with the attendant bad effects (pity, stigma, lack of attention to removable barriers). But one does not need the mere-difference view to make this clear. The ability to make this case is a cardinal virtue of any model that pays attention to the social component of disability, and the case for valuing the ways in which society may be better because of the presence of the sort of diversity that comes with the presence of people with disabilities does not depend on claiming that such diversity is a mere difference. ^{Footnote 35}

Is the value of the view, then, in its simplicity? Perhaps attention to the nuances required to state a clearer version of the nature of disability would prevent advocates from generating the memorable slogans that stick in the minds of politicians and bureaucrats and so help make practical change. Again, though, we think this is probably a bad calculation. We think that considerations of the sort we described above, when identifying where Barnes’s arguments went wrong, are not abstruse nor unlikely to occur to people who think about these matters. Especially since mere-difference rhetoric is unlikely to be the only rhetoric decision makers will hear from people with disabilities and their allies, for reasons we will presently describe, the net effect of using slogans that will tend to strike people as untrue will be to diminish how seriously advocacy is taken. In short, more accurate rhetoric will be more effective for much the same reason that arguing that the net effects of a particular policy are positive tend to be more persuasive than arguing that it has no down-side whatsoever. Moreover, the public policies and changes in social attitudes one would likely use the mere-difference rhetoric to advance are equally easy to defend on the basis of a related human diversity view of disability that, unlike the mere-difference view, acknowledge that at least many disabilities are ones that are, on balance, a negative difference even in the estimation of most people living with them. Taken together, these two points undermine the idea that the mere-difference view is ‘productive overstatement.’

The most significant political risk of the mere-difference rhetoric, though, is its potential to have a negative impact on the making and retention of political alliances. Social and political progress normally depends on the creation of alliances and coalitions. ^{Footnote 36} If, however, many people with disabilities (and their families and friends) do not recognize themselves in the rhetoric of mere-difference, nor feel they are adequately represented by other extreme versions of the ‘it’s just an example of human diversity’ view, the rhetoric risks committing a common flaw of radical politics: cleaving to views that separate a politically ineffectual vanguard from the very people whose interests it claims to advance.

We do not know of any studies that have tried to quantify whether people with disabilities by and large identify with mere-difference rhetoric. In addition to some of the challenges we mentioned in footnote 33, there are also obvious questions about just who gets surveyed, and the possibility that the answer will vary by disability, or by severity within any particular grouping of disabilities. So evidence to back the suspicion we have just voiced must be indirect.

Some useful, if indirect, evidence of this sort can be gathered from a fascinating recent paper by Jennifer Sarrett. ^{Footnote 37} For the paper, Sarrett asked questions and monitored discussions in online discussion forums for people who regarded themselves as part of the neurodiversity movement. Her paper quotes liberally and informatively from the discussion, as well as providing useful summaries. While the question of ‘mere-difference’ is not directly raised, Sarrett identifies some trends in the discussion that are relevant to the present question. We will mention two: (1) even among people committed to the view that being on the autism spectrum is a manifestation of human diversity that, on the whole, ought to be celebrated, it is a common view that there is a “need for a certain level of impairment and life-long difficulties in order to be granted membership in this autistic community.” ^{Footnote 38} This, to us, given the problems we have identified with the mere-difference view, suggests that a version of the human-diversity view that eschews the mere-difference rhetoric will strike the discussion participants as better reflecting their views. And, (2), there is a strong suspicion about the legitimacy of people who attempt to use their own experiences to represent the community as a whole (for some, this is particularly so if the self-appointed spokespeople are self-diagnosed). This suggests to us a suspicion of those who might regard themselves as worthy members of a vanguard.

We have primarily used Barnes’s 2014 paper as a foil for our discussion, but at this juncture it is worth considering one aspect of the discussion in her book that we only briefly mentioned earlier. In the book, Barnes explicitly sets aside questions of developmental and psychological disability, choosing to focus on physical disability as a starting point, and she recognizes that even here there might be some ‘hard cases’ that are difficult to fit into the theory. This approach certainly differs from the official one in her paper, where Barnes opts for a notion of disability that involves attention to ‘paradigm cases,’ surely not all of which are physical disabilities. One might suspect, however, that something of this sort was behind her focus on, for instance, deafness and blindness when discussing the thought experiments in the paper. Perhaps ‘mere difference’ is more plausible as the core notion for physical disabilities, even if it will be less plausible as the core of a more general account?

We find this unlikely, and note that not all of our counterexamples to the argument were of developmental or psychological disability—for instance, we considered the case of someone in a wheelchair as a result of a heart condition. But we also think there is a more fundamental problem with the ‘start with the physical and work out’ approach.

Again, this more fundamental problem is in the politics of the approach. The approach is officially ‘ameliorative,’ and the project of defining disability is part of an ongoing political struggle on behalf of the disability rights movement. If the strategy results in a definition in which many people with disabilities do not see their own lived experiences reflected—as we fear is likely by starting with a selective minority of activists with particular physical disabilities for whom the idea that disability is mere-difference is plausible—the strategy will lead to a group of people whose attempts to speak on behalf of everyone are regarded as illegitimate by many and so treated with suspicion or hostility. This is not the best way to build solidarity with, for instance, people with developmental or psychological disabilities and their families or allies, nor with people with physical disabilities who do not see their differences as mere differences. There is every risk of running into the problem that confronts most attempts at vanguard politics: the people the vanguard wants to lead towards revolution don’t want to be led by the self-selected vanguard. ^{Footnote 39}