Introduction

Sitting in a near-empty restaurant in southern California, Heather talked with Adam, Lee, and Ned, three autistic leaders of the area's oldest and largest activity group for autistic adults. They had all just finished roller skating at a nearby rink and were enjoying a late lunch. Heather asked Adam, a man in his early thirties and the founder of the group, what he thought about the common use of labels to describe how an autistic person functions. In particular, Heather wanted to know how he and the others in the group understood the notion of an "autism spectrum."

Adam explained that popular understandings of the autism spectrum as a line that runs from low to high functioning can be useful. "We choose to interpret these terms negatively," he said, "and that's pointless." Pursuing this line of reasoning, Heather asked Adam how labels like "low functioning" and "high functioning" might be useful. But Ned interjected "everyone's autism is different." Lee then asserted "functioning labels have some use. Most of us in the group are high functioning, and that's how we can do stuff like go roller skating and take ourselves out to lunch. And some of us, like Steph, are low functioning. It's not like she's any less important, though." Ned and Adam seemed surprised by Lee's statement. Adam turned to Lee: "Steph isn't low-functioning…. She drives, works part-time, can have a regular conversation. She's definitely not low functioning." Ned agreed with Adam as Lee rolled her eyes and said, "Oh well, whatever, you know what I mean."

This interaction is drawn from fieldwork Heather conducted with autistic adults in the United States and online (where much autistic interaction and advocacy now takes place). It encapsulates a set of debates that were frequently encountered during this fieldwork—and that are pertinent to contemporary autistic communities more broadly. 1 In these debates, the notion of an autism "spectrum" becomes a focus of explicit identification, reflection, and contestation. Our point of departure in this article is that true to its derivation from the Latin specĕre, meaning "to look or see," the "spectrum" strongly influences what we see as autism. Originally an "outsider," etic concept developed for diagnosis and classification, the spectrum metaphor is now an "insider," emic term of selfhood as well as a basis for group identification (and disidentification). Given that metaphor "serves very specific purposes in autism discourse" with regard to understanding causes, responses, and communities (Broderick and Ne'eman 2008, 459; see also Danforth and Naraian 2007; Waltz 2003), it is highly significant that the spectrum has become the dominant metaphor for conceptualizing autism worldwide. This has fundamental consequences for notions of disability, diversity, and normality—not least because, as we note in the conclusion, the spectrum metaphor has spread to many domains beyond autism itself.

The expanding reach of the spectrum is possible because the processes of reconfiguration and transformation that led to the dominance of this metaphor with regard to autism are not unique to autism. Rather, they lie at the heart of many far-reaching shifts in contemporary culture and politics. Our analysis is informed by a growing challenge to the notion of a linear spectrum, a challenge that originates in communities of autistic persons and their allies. This challenge inspires us to reconsider the autism spectrum in light of another spectrum—the Kinsey Scale, a "spectrum" for conceptualizing sexual orientation that first appeared in 1948 but has been critiqued since the 1970s. In comparison to the Kinsey Scale, the autism spectrum is both more recent and more widely accepted. How might responses to the Kinsey Scale contribute to rethinkings of the autism spectrum?

This is a question about the cultural and political implications of metaphors and conceptual models. It is a question about the necessity of categorization for human experience and the tensions that always exist between categories and lived realities—including the way that categories shape lived realities. In the world of autism community and advocacy, one frequently encounters variations of the adage "If you've met one autistic person, you've met one autistic person." The problem is that the phrase "autistic person" in this sentence could be replaced with "lesbian," "Norwegian," "plumber," or any other category of human distinction. It is a question of the interplay between categorization and generalization on the one hand, and selfhood and individuality on the other. And it is precisely because this question is not unique to autism that drawing on analogous debates can prove so productive.

Our use of sexual orientation as the specific debate to illuminate the autism spectrum is facilitated in part by the long and rich history of intersections between disability and queerness. Historically, in dominant Western traditions queer sexuality was considered a form of disability. The American Psychiatric Association (APA) only removed homosexuality from its Diagnostic and Statistical Manual of Mental Disorders (DSM) in 1973, and forms of transgender identity remain pathologized. The classic analysis of this intersection remains that of Michel Foucault, who explored how the emergence of the "homosexual" in the nineteenth century related to an "incorporation of perversions" and a "new specification of individuals" (1978, 42–43).

There is by now a robust literature in disability studies addressing questions of sexuality (e.g., Kafer 2013; Kulick and Rydström 2015; Liddiard 2014; McRuer 2006; McRuer and Mollow 2012). However, this article is not about the sexuality of autistic people. Instead, our analysis takes place at the level of discourse: we draw on critiques of the Kinsey Scale to analyze the "autism spectrum" as a diagnostic tool, a model of selfhood, and a site of fervent debate. To develop this analysis, we first discuss the general issue of conceptual models and their implications. Second, we explore the history of the autism spectrum and its effects. Third, we examine the history of the Kinsey Scale, focusing on critiques of the scale and the positing of alternatives like the Klein Grid. In the final section, we draw on this history of the Kinsey Scale to imagine how autism might be understood if reconfigured in a similar manner.

With regard to this reimagining, and throughout the article, we interweave our conceptual analysis with ethnographic insights gleaned from autism communities themselves. In this regard we build on one of the strongest traditions in anthropology, which is treating ethnographic materials as a source of theory, not just data in isolation. It is by placing these ethnographic insights into conversation with our own conceptual work that we can most effectively craft our reimagining of autism beyond the "spectrum" metaphor. As will be clear, this reimagining is intentionally provisional. Our goal is not to replace the spectrum metaphor with another "better" metaphor; nor is it even to claim that the spectrum metaphor is irretrievably flawed. Metaphors, like language more generally, are contextual heuristics that can never perfectly capture social reality. But perfection is not what we seek. Rather, we seek to build on the insights of autistic persons who wish to forestall the naturalization of "the spectrum," and contribute to a discussion of what alternative frameworks might bring to questions of social justice, ability, and human flourishing.

Metaphors and their implications

We seek to place ethnographic attention to autistic experience into conversation with work in cognitive linguistics and philosophy of mind that explores the social implications of conceptual models and frameworks, particularly metaphors and categories (e.g., Bowker and Star 1999; Gibbs 2008; Hacking 2002; Lakoff 1987; Ortony 1993; Sweetser 1990). These are not simply window dressing on the workings of the mind. They are pivotal to sociality: "culture consists in the way analogies are drawn between things, in the way certain thoughts are used to think others" (Strathern 1992, 33). For instance, metaphors typically work by extending inferences from a "source domain" to a "target domain" (see Lakoff and Johnson 1980). An example: in many languages including English, emotions are metaphorically construed as containers. One can thus be "in a good mood" or have difficulty "getting out of" sadness. Here, containers are the source domain, which (as is typically the case) is more concrete than the target domain of emotions. One entailment from this source domain is that one is not normally inside two containers at once. This might make it harder to imagine experiencing two emotions simultaneously—compared to, say, a metaphorical construal of emotions as flavors, colors, or landscapes.

A detailed review of this rich body of scholarship lies beyond the scope of this article, but two insights are particularly relevant to our argument. First is the basic idea, noted above, that humans use inferences from one category to conceptualize other categories. This process is universal but historically and culturally specific, and usually takes place at an implicit level of tacit knowledge. Second, at issue is not just concepts and terms, but selfhood and society. For instance, Ian Hacking has explored how "social change creates new categories of people, but the counting is no mere report of developments. It creates new ways for people to be" (2002, 100). What Hacking termed "dynamic nominalism" means that "a kind of person came into being at the same time as the kind itself was being invented. In some cases, that is, our classifications and our classes conspire to emerge hand in hand, each egging the other on" (106). This foregrounds how metaphors and categories can involve analogies from source to target domain, but also from etic categories of outsider analysis to emic categories of experienced personhood. Hacking has frequently pointed out how homosexuality exemplifies this process (e.g., 2007), noting how Foucault pursued this line of analysis as well:

As defined by the ancient civil or canonical codes, sodomy was a category of forbidden acts; their perpetrator was nothing more than the juridical subject of them. The nineteenth-century homosexual became a personage, a past, a case history…. Nothing that went into his total composition was unaffected by his sexuality. (Foucault 1978, 43)

Many scholars have built on Hacking, Foucault, and others to explore queer manifestations of dynamic nominalism, including work one of us has published on the implications of metaphors other than "coming out of the closet" for gay subjectivities (Boellstorff 2005). More broadly, exploring the mutual constitution of people, things, and classifications highlights the interplay between (1) expert models of "normal" and "abnormal" social reasoning and (2) the imposition, appropriation, and transformation of those metaphors and categories through lived experience. This interplay between (1) and (2) often involves struggle, debate, and invention, and is emphatically not unique to homosexuality. That is why the comparatively well-documented and well-theorized example of homosexuality provides a useful exemplar as we examine contemporary debates regarding autism and autistic difference.

The emergence of the autism spectrum

Given the current hegemony of the autism spectrum, its history is strikingly shallow. Archival documents and folktales clearly indicate that persons we might now classify (and might classify themselves) as "autistic" have been found in cultures worldwide, though not always seen as a single category and treated with everything from veneration to benign neglect and even ostracization or violence. The origins of the contemporary medical definition of autism date to the 1940s, particularly the formulations of Leo Kanner and Hans Asperger (Silverman 2012; Singh 2015). In the wake of their work, autism was originally predominantly construed as a singular psychological disorder. Instances of autism obviously differed (like instances of, say, depression, diabetes, or epilepsy), but that difference was understood in individualized terms. Classification was binary: one was autistic or not.

Such psychological explanations (often taking the form of a "refrigerator mother" hypothesis that blamed mothers for autism) "were the prevailing ideas about the origins and course of autism treatment in both professional and popular social discourses well into the late 1970s" (Singh 2015, 23). They have a continuing impact, as revealed by the importance of the DSM, a document intended to help psychologists diagnose disorders, to the understanding and politics of autism even to the present day. Autism was first included in the DSM in its third edition, which appeared in 1980. But at the same time, by the 1980s the predominantly psychological explanations proposed by Kanner and Asperger and carried forward by others (e.g., Bettelheim 1967) were being challenged by explanations in terms of biology (including environmental, physiological, and neurological factors). One significant strand of these challenges has taken the form of a "focus on investigating autism through genetic and increasingly genomic lenses" (Singh 2015, 1).

But these genetic framings of autism continue to coexist with psychological understandings, not least because "without any reliable biological indicators to detect autism, the classification of autism relies on observation of behavioral characteristics" (Singh 2015, 25). Because such understandings tended to view autism as a single domain, differences encountered clinically with some degree of shared symptomatology were classified as new, independent domains, leading to the rise of five distinct "pervasive developmental disorders": autistic disorder, Asperger's syndrome, childhood disintegrative disorder, Rett syndrome, and pervasive developmental disorder-not otherwise specified (PDD-NOS) (Willemsen-Swinkels and Buitelaar 2002; Johnson and Meyers 2007). In other words, these shifting classifications did not destabilize the binary diagnosis of autistic or not autistic; instead, the number of binarisms proliferated. One consequence of this, for instance, is that a diagnosis of "Asperger's syndrome" might protect someone from the supposed stigma of autism altogether (Singh 2011, 237).

Parallel to these shifts in clinical and scholarly understandings of autism was the deinstitutionalization of many forms of "mental retardation" starting in the late 1960s. These parallel developments each shaped the other, transforming relationships between medical authorities, parents, advocates, and autistic persons themselves. Dominant construals of autism began to be challenged: "confusion and disagreement in the diagnosis of autism is widely recognized in the field" (Schopler 1978, 137). In this context "new categories began to be differentiated within a new institutional matrix that replaced custodial institutions—community treatment, special education, and early intervention programs" (Eyal et al. 2010, 3).

Extremely significant in this regard was Lorna Wing's development in the late 1970s of the idea that autism could be understood in terms of a "spectrum." As the mother of an autistic child and a medical researcher, Wing bridged the worlds of science and advocacy (Wing and Gould 1979; Wing 1993). Thus "[Wing's] real ingenuity lay in the invention of the autism spectrum as a device for translating and aligning the multiple interests of… autism researchers, clinicians, therapists, parents of severely afflicted autistic children, and parents of near-normal autistic adolescents, and ultimately also these adolescents and children themselves" (Eyal et al. 2010, 225).

For over a decade, the spectrum metaphor was just one of many interventions into clinical frameworks for understanding autism. This can be seen, for instance, in the fact that as late as 1993, Wing could state that that her "triad of social impairment [could be] referred to as the 'autistic continuum' (or, perhaps more appropriately, the 'autistic spectrum'" (Wing 1993, 70)). That Wing placed "autistic spectrum" in scare quotes reflects how at this point the spectrum was still incompletely consolidated as a dominant model. However, one year later, the fourth edition of the DSM (DSM-IV) first explicitly recognized Autism Spectrum Disorder (ASD). In 2013, the fifth edition (DSM-5) further formalized the autism spectrum by abandoning the previously dominant rubric of five "pervasive developmental disorders" mentioned above. One of these, Rett Syndrome, was removed from the DSM-5 entirely due to the discovery of its genetic cause. The others were subsumed under the ASD classification (Mayo Clinic Staff 2014).

It is important to note that a spectrum approach can be understood as operating at two levels, with assignment to a particular diagnostic classification and to a position within a spectrum category. A primary goal of a clinician or diagnostician is to link a classification or diagnostic label with a set of observable traits, and in the case of an autism spectrum condition, to situate a patient according to the perceived intensity of their disability experience within a hierarchy of function (Krueger and Bezdjian 2009). The spectrum approach entails situating diagnosed patients along a continuum that joins what were multiple but ostensibly related classifications (Willemsen-Swinkels and Buitelaar 2002). It thus allows for sites of overlap between once disparate classifications.

Some have argued that the spectrum enables physicians and patients to develop a more comprehensive understanding than a "discrete entities" metaphor (e.g., Lutz 2013). However, others have contended that the metaphor only appears to reduce diagnosis-associated stigma, when in fact it reifies a hierarchy of ability and a linear ranking of autistic need. Specifically, it can leave those labeled "severe" subjected to ableism and rendered invisible, framed as requiring intervention or cure in order to restore or grow culturally prized capacities, social value, and legitimacy. Conversely, those placed on the other pole (i.e., "mild") are often understood to need few (or no) therapeutic interventions. They can be variously pardoned for their near-normal "quirks," denied assistance, or stigmatized because of their diagnostic link to autistic people "on the other end."

But the impact of the spectrum metaphor is not limited to the clinical arena: "the consolidation of the spectrum diagnostic label into one overarching autism-related diagnosis… [has had] implications on social discourse and, therefore, on identity formation" (Jones et al. 2015, 1493). For instance, "Asperger's syndrome," which had been not just a clinical diagnosis but term of self-identity (Singh 2011), technically no longer exists; it persists only as an informal description. Thus many groups for persons identifying as having Asperger's syndrome "see the removal of the Asperger's diagnosis as a potential threat to their identity. These groups fear that the positive identity associated with being Asperger's will be overshadowed or stigmatized by the broader definition of ASD" (Singh 2015, 27). In her ethnographic research on autism, Bagatell found "a hierarchy of power, with those considered high functioning awarded positions of power" (Bagatell, 2007, 421); as one of her informants put it, "the rift in the autism community basically is between what we label high-functioning and low-functioning people" (Bagatell, 2010, 46). One key manifestation of this rift is with regard to treatment: "Sue Rubin, a low-functioning autistic and subject of the documentary Autism is a world, emphasizes that whereas high-functioning autistics tend to be against a cure, low-functioning autistics generally hold the opposite position" (Ortega 2009:430; see Fein 2016, 187–89).

Over time and continuing to the present day, then, the development of the "autism spectrum" has had implications for identity and community, not to mention medical interventions, legal regimes, and political rights. But as Jones notes above, for any of these implications to be possible, first and foremost the "spectrum" represents a transformation in "social discourse." This epistemic shift is indeed what Foucault would term a new "discourse" of autism—a new set of institutionally embedded norms and practices that determine a regime of truth, where "truth" is understood as "a system of ordered procedures for the production, regulation, distribution and circulation of statements" (Foucault 1977, 14). This discourse "naturalizes" the spectrum as an ostensibly self-evident feature of the world (Yanagisako and Delaney 1995). It enables us, for instance, to speak of "the spectrum nature of disability" (Kapp et al. 2013, 68, italics added), and even an "autistic phenotype" where the cause of multiple forms of autism is seen to originate in genetic and neurological programming (Orsini 2009; Singh 2015). In her ethnographic work with autism communities, Fein discusses a fascinating example of the spectrum and binarism metaphors clashing:

At one meeting, the father of a young man diagnosed with Asperger's ventured a speculation that rather than there being a sharp line distinguishing those on the autism spectrum from "neurotypicals," perhaps we are all on the same "human spectrum." The group facilitator firmly corrected him. "There really is a difference between people on the spectrum and those who aren't […] It's in the brain, it's wiring." (2011, 42)

These debates over the classification of autism and its consequences are part of a contemporary "autism matrix" (Eyal et al. 2010) that includes not just researchers and clinicians, but parents, advocates, and autistic persons themselves. In June 2015, in a popular public Facebook group for adults with Asperger's syndrome (which as noted earlier, is now an informal category not recognized by the DSM), Thomas observed a young autistic woman named Sasha discussing her struggle to interact with extended family members. Sasha was frustrated with their repeated claims that she just "doesn't seem autistic" or "just [has] Asperger's and real autism is more severe." Her efforts to explain about "Asperger's inclusion in the autism spectrum" seemed ineffective: "The autism label might suggest that I should be severely debilitated, but that doesn't mean that's how it actually is. I don't look like a stereotypical autistic person…. There's more to autism than 'seeming autistic'." Note how Sasha challenged the claim that her autistic identity is inauthentic because insufficiently "severe."

This reveals that Sasha and her family members were all using the spectrum metaphor. It has entered popular discourse such that family members could challenge Sasha not qualitatively—in terms of a missing trait or symptom—but quantitatively, in terms of position on an imputed spectrum of function. Such impact of the spectrum in autism discourse is not limited to everyday interaction: it appears in the domain of advocacy as well. Explaining the significance of the spectrum for autistic individuals, the self-identified autistic rights activist Transasta wrote on the blogging website Tumblr that she used the notion of spectrum "as in 'we're all affected in different ways and have different needs, strengths and weaknesses associated with autism' not spectrum as in 'some autistics are ~worse~ than others.'" Here, Transasta illustrates a common phenomenon among autism advocates: the drawing of a distinction between the medicalized use of the spectrum as a diagnostic tool that renders difference as deficiency, versus a social-movement use of the spectrum as a coalition-building tool that renders difference as variation of an underlying sameness. These dynamics are further revealed by another advocate, CorinneDuy, who responded "Yeah, the spectrum does kinda imply a linear scale through which one could easily be judged better or worse. Wouldn't 'set' be better? Like, all kinds of stuff are related because it's in the Autistic Set, but in no way classifiable as more or less than the other." In this interaction, Transasta and CorinneDuy are pushing toward a more expansive view of the autism spectrum—one that is de-stigmatized and de-medicalized—but they also propose alternatives to the spectrum itself.

What a spectrum can do

Later in this article, we place the insights of persons like Sasha, Transasta, and CorinneDuy into conversation with our own analysis of the Kinsey Scale and its legacies in order to rethink key assumptions of the autism spectrum. At this point in our analysis, however, we discuss one key entailment of the source domain of "spectrum" as it has manifested in the target domain of "autism." As CorinneDuy noted above, a spectrum encodes phenomena along a variable assumed to be linear. For instance, a thermometer is a spectrum of temperature: if something is measured to be 80 degrees Celsius, that does not mean that the first 60 degrees are of one kind and the final 20 degrees another. A Likert scale (used for many psychological measures) is often used to describe pain. If a doctor asks "rate your pain on a scale from 1 to 10," the assumption is that the kind of pain marked by "3" is the same kind of pain as that marked by "8," but less intense.

One of the most striking aspects of the autism spectrum is that the variable it measures is now almost universally presumed to be "function." In this understanding, to be "high functioning" is to be closer to a "normal" or "neurotypical" person in ability, and "low functioning" less so. The spectrum metaphor thereby frames autism in terms of a "deficit model" (Nicolaidis 2012), such that persons who are low functioning straightforwardly have a lack of that which persons who are high functioning have to a greater degree. This reverses the polarity of the original variable of the autism spectrum, impairment (e.g., Wing and Gould 1979; Wing 1993), where "high impairment" meant what is now more commonly termed "low functioning" and vice versa.

The grounding of the autism spectrum in function has important consequences. Within online autistic communities and especially amongst autistic self-advocates, struggles against medicalization hinge on a specific claim: namely, that autism is accepted a legitimate form of neurological diversity by autistic persons themselves. Two kinds of "neurodiversity" are simultaneously at play in this discourse: first, diversity between autistic persons and non-autistic persons (or "neurotypicals"), and second, diversity within the category of autism. With regard to the first form of diversity, some observers have noted how notions of neurodiversity, even when the "diversity" in question is understood in terms of a spectrum, can instantiate a new dichotomy between "neurodiverse" and "neurotypical":

some self-activists reject the distinction between "low" and "high-functioning" autism, and consider differences across autistic populations as variations of degree […] This brings about a paradoxical situation: while neuroplasticity helps account for neurodiversity, neurodiversity advocates tend to homogenize neurodiverse brains and minimize their differences so as to support their claims for the existence of a brain-based autistic identity. (Ortega 2009, 439, 441)

Since the late 1990s, resistance to medicalization has relied on autistic persons who can speak and write in ways that are legible to wide audiences, but also on the assumption that such "high functioning" autistic persons can represent the experiences of "low functioning" autistic persons as members of a shared community of diagnosis (rather than, say, as advocates).

This understanding of a shared autistic subjectivity is possible only in terms of the second form of diversity noted above: the idea of diversity within a singular category of autism. While the spectrum is not the only imaginable metaphor for conceptualizing this diversity, it has been the primary metaphor that has emerged in the worlds of autism since the 1980s. And while this metaphor can deconstruct the binarism of autism/normal (because there is a spectrum), it can also reify this binarism by opposing the neurotypical to the neurodiverse. The spectrum metaphor has not eliminated the tendency to frame neurodiversity as occupying a Gaussian extremity relative to a "normal" distribution of neurosocial experience. The metaphor, therefore, does not necessarily challenge a belief that the normal or neurotypical is superior or ideal (Davis 2013). In addition, it reconstructs a new binarism via "the very idea of separating the autism spectrum into categories of 'low-functioning' and 'high-functioning'" (Silverman, 2012, 327). This second binarism has gained salience and has several crucial entailments of its own, including political entailments:

It would be hypocritical to subsume all forms of the autistic continuum under the "high-functional," and then happily consider that autism is a lifestyle. These are the challenges faced by the neurodiversity movement, in which, in the US especially, the search for new and non-pathologized forms of identity and community stands in tension with reductionistic identity politics. (Ortega, 2009, 427)

This is a fascinating characteristic of spectra as metaphors: they challenge dichotomy by framing variables as gradients or continuua (as in the example of temperature), but with two poles (hot and cold, for instance) that can emerge as a new dichotomy. One consequence of this process with regard to autism has been the dynamic by which parents and caregivers frequently clash with autistic self-advocates who claim to speak for autistic family members who communicate in non-normative ways. Arguments for neurodiversity and demedicalization are sometimes challenged by these parents and caregivers, who contend that such arguments are relevant only for "high functioning" autism. For many of these parents and caregivers, the right to pursue intensive therapeutic interventions or cures is paramount. Additionally, they challenge the idea that they cannot access the minds, intentions, and desires of their "low functioning" autistic loved ones, patients, and clients.

These complex debates over categories and their consequences in the shadow of the spectrum metaphor can be illustrated by a conflict between two bloggers Thomas encountered on Tumblr in April 2015. An anonymous user who claimed to have a non-speaking autistic sibling wrote to Clara, a self-identified autistic person, explaining "My brother is a non verbal autistic. He doesn't understand a word anyone is saying. He can't do anything by himself. When we are out of the house he just screams and cries. Yeah, autism does need a cure." Clara responded "If he doesn't understand a word anyone is saying, how do you know he doesn't want to be autistic? Because that sounds a lot more like you don't want him to be autistic. And people like me speak for your brother more than people like you do." Here, Clara asserted that her diagnostic status granted her greater insight into autistic experience, even with regard to persons on a very different part of the spectrum (and thus with radically different life paths and capabilities).

Throughout fieldwork, examples like those presented above illustrated how the spectrum metaphor has become a point of contention. These challenges to the spectrum are still often implicit, but a growing number of advocates, as well as medical experts, are noting the limitations of the metaphor and speculating what alternative conceptions of autism might bring to the table (e.g., Burgess 2016, Fein 2016). Scholars have also joined this debate: "the [spectrum] image is problematic: spectra are linear and autism isn't. The metaphor suggests that you can arrange autistic people on a line, from more to less" (Hacking 2009a, 47). Here, Hacking identifies the conceptual limitation of the spectrum we noted earlier: the metaphor frames a variable in terms of a unilinear gradient, such that an increase in "one direction" is a simultaneous decrease in the "other direction." To be "more hot" is simultaneously to be "less cold," but it is less evident that to be "more high functioning" is simultaneously to be "less low functioning"—nor, as we discuss below, is it self-evident that "more homosexual" is simultaneously to be "less heterosexual." Scholars have also explored alternative metaphors, like a "sphere of humanity" that allows for "respecting differences without the implication of a value based where on the spectrum one falls" (Sarrett 2012, 29). In this regard, what might alternative frameworks for sexuality have to offer these debates—and their political and social implications?

The Kinsey Scale and its alternatives

Across a range of constituencies, knowledge communities, and advocacy agendas, the naturalization of the autism spectrum has been challenged, not least by autistic persons themselves. However, these challenges are relatively recent and have not been thoroughly reviewed in the scholarly literature. It is for this reason that before moving even further into contemporary debates over the autism spectrum, it will prove highly illuminating to explore how another spectrum, the Kinsey Scale, has been critiqued for over 40 years. Autism and sexual orientation are obviously distinct domains and we do not mean to conflate them. Rather, we begin from the fact that they have long been linked by forms of medical discourse. For instance, the DSM has served as the authoritative standard for classifying (and stigmatizing) not just autism but homosexuality. Furthermore, as Foucault noted the original, pathologizing discourse of homosexuality "made possible the formation of a 'reverse' discourse: homosexuality began to speak on its own behalf, to demand that its legitimacy or 'naturality' be acknowledged, often in the same vocabulary, using the same categories by which it was medically disqualified" (1978, 101).

This historical path from deviant homosexuality to LGBT rights and community shares features with many other domains. This is only one reason why autistic persons will sometimes speak of "coming out" as autistic, a metaphor borrowed from the domain of sexuality that appears with regard to many other "disabilities" (Bagatell 2007; Meyerding 1998, Samuels 2003; Swain and Cameron 1999). Sexual orientation does not hold some definitive answer with regard to autism: the domains cannot be artificially mapped onto one another. But nor can they be artificially severed. The history of classification with regard to homosexuality can provide specific examples regarding the limits of spectrum metaphors, as well as possibilities other metaphors might foster. These possibilities have concrete consequences for how we might imagine autistic selfhoods, communities, and rights.

Because our analysis focuses on the autism spectrum, our foray into the history of the Kinsey Scale must be brief. Fortunately, there exists a robust literature on the Kinsey Scale and the milieu of sexology in which it emerged (e.g., Terry 1999). Alfred Kinsey (1894–1956) was one of the best-known mid-twentieth-century sexuality researchers. His notoriety stemmed in part from his training as a biologist (whose research was primarily on wasps), and the empirical approach he brought from that prior work.

Kinsey's turn to the study of sex began in the 1930s. A half-century earlier, the invention of terms like "homosexual" and "heterosexual" reflected a shift from a dominant religious discourse of sexuality to a "psychological, psychiatric, medical category" (Foucault 1978, 43). Crucially, the rise of this medical discourse of sexuality entailed the creation of discrete, perduring categories of selfhood: "The sodomite has been a temporary aberration; the homosexual was now a species" (ibid.). In his landmark first major publication on sexuality, Sexual Behavior in the Human Male, Kinsey (and his coauthors) directly challenged this framework: "Males do not represent two discrete populations, heterosexual and homosexual. The world is not to be divided into sheep and goats. It is a fundamental of taxonomy that nature rarely deals with discrete categories" (Kinsey, Pomeroy, and Martin 1948, 639, emphasis added). However, Kinsey did not just reject this framework—he suggested an alternative: "the living world is a continuum in each and every one of its aspects" (ibid., emphasis added). In other words, Kinsey offered a spectrum as an alternative to the "container" metaphor of discrete entities. While all historical comparisons are partial and imperfect, there are striking parallels between this shift and the shift 30 years later from "pervasive developmental disorder" to the claim that autism represents "a continuum of severity rather than discrete entities" (Wing and Gould 1979, 26, emphasis added). In both cases, a spectrum is presented as a more accurate alternative to discrete domains.

But Kinsey also turned the spectrum into a rubric for measurement: "While emphasizing the continuity of the gradations between exclusively heterosexual and exclusively homosexual histories, it has seemed desirable to develop some sort of classification which could be based on the relative amounts of heterosexual and homosexual experience or response in each history" ibid., p. 656). This is the famous "Kinsey Scale," though this name for the scale was coined by other researchers (Kinsey referred to it as the Heterosexual-Homosexual Rating Scale; see Figure 1). The scale has integers from 0 to 6, with 0 representing "exclusively heterosexual," 6 representing "exclusively homosexual," and the intermediate numbers representing what could be termed gradations of bisexuality.

In its original use by Kinsey and his associates, the Kinsey Scale had complexities not always appreciated in its subsequent popularization. For instance, ratings were applied primarily by researchers based on interviews (not by the interviewees themselves), and the scale originally included an "X" which denoted asexuality or lack of sexual response (see Weinrich 2014a). Regardless, the subsequent popularization was unexpected and unprecedented: from its mention in the Cole Porter song "Too Darn Hot" from the 1948 musical Kiss Me, Kate, to bestseller status (Sexual Behavior in the Human Male and Sexual Behavior in the Human Female sold around 750,000 copies), to global impact (the two books were translated into 13 languages), Kinsey's research—and the Kinsey Scale developed as part of that research—deeply shaped discourses of sexuality from the second half of the twentieth century onward.

Given that this impact of the Kinsey Scale was far from limited to academic circles, it is not surprising that the scale was soon taken up in the sexual revolutions of the 1960s and 1970s, particularly by the LGBT movement. By the late 1970s and early 1980s, both activists and scholars had become aware that while the scale was an improvement over the earlier notion of homosexuality, heterosexuality, and bisexuality as discrete categories, it had limitations, including:

1. The scale's sole variable of sexual experience entails that sexual practices match up with identities and fantasies.
2. The scale's linearity entails that "more homosexuality" means "more heterosexuality."
3. The scale encodes a presumption that "male" and "female" are self-evident: it contains no provision for transgender or intersex identity, nor does it contain any provision for sexual practice with (or desire for) persons who are not cisgendered.
4. The scale has no temporal component, entailing that sexual orientation is unchanging through the lifespan.

Activists and scholars responded in multiple ways to these (and other) limitations of the Kinsey Scale, including notions of fluidity and polyamory, and by modifying the Kinsey Scale itself (for example, by using two scales, one for behavior and one for identity). Within our limits of space, we would like to briefly discuss one notable example of such a reworking of the Kinsey Scale: the Klein Sexual Orientation Grid ("Klein Grid"). This rubric was developed by the sexologist Fritz Klein in the 1970s, originating in his sense that the Kinsey Scale "was useful in understanding degrees of homosexuality or heterosexuality, but as a sex researcher I found it far too simple" (Klein 2014, 342). The Klein Grid has taken many forms over the years, but for our purposes we can examine it as first published in 1980.

As can be seen in Figure 2 above, the Klein Grid consists of a table of cells in which respondents (not interviewers, as with the original Kinsey Scale) enter "numbers ranging from 1 ('Other sex only') through 4 ('Both sexes equally') to 7 ('Same sex only')" (Weinrich 2014b, 317). The Klein Grid thus, in essence, contains Kinsey Scales within it, but applies them in new ways. First, the Kinsey Scale's focus on sexual behavior appears as only one (Row B) of seven variables that include not just attraction and fantasy (Rows A and C), but emotional and social preference (Rows D and E), self-identification (Row F), and "Hetero/Gay Lifestyle" (Row G).

A second difference is that the 1–7 numbering system is applied not only to these several variables, but over time (marked as columns on the grid). While the original Klein Grid had only temporal columns for "past" and "present," it could in theory be made more fine-grained (e.g., ten-year increments). The basic insight captured by even the past/present distinction is that these variables "change over a lifetime—quite drastically for some….We must understand and accept the fact that sexual orientation is a dynamic life process, not a fixed position" (Klein 2014, 342). In other words, not only can sexual orientation change over time; it might do so differently across different variables (for instance, someone's self-rating in terms of sexual behavior (Row B) might shift from 1 to 6, but might shift only from 3 to 4 in terms of sexual fantasies (Row C)). A third column allows respondents to identify an "ideal" rating for each variable.

Overall, then, the Klein Grid "move[s] beyond the univariate conceptualization of sexual orientation reflected in the Kinsey Scale itself, and the categorical notion of orientation that continued to be reflected in the use of the scale" (Galupo et al. 2014, 406). This is not to say the Klein Grid is without limitations. There is obviously an "apples and oranges" concern regarding its seven variables. They are not really comparable on a single scale, and Klein's own explanation of them (and why there are seven, rather than six, eight, or some other number) was unclear. Key conceptual presuppositions shaped the grid in a manner not unlike the Kinsey Scale; for instance, Klein "did not seem to have perceived the possibility of measuring attraction to men and attraction to women on separate scales" (Weinrich 2014b, 318).

For the purposes of this article, our goal is not to investigate these limitations of the Klein Grid. Nor is it to review the many revisions to the Klein Grid that have appeared since 1980, nor discuss other alternative models for understanding sexual orientation (e.g., Anders 2015; Weinrich 2014c). The Klein Grid does not represent an endpoint or solution to the problem of conceptualizing sexual orientation—but it does represent a clear improvement.

The Kinsey Scale emerged as a challenge to an earlier model of "discrete" entities for sexual orientation, offering in its stead a model based on the spectrum metaphor. Models like the Klein Grid emerged in response to limitations of that metaphor. The key point is that with regard to autism, it is as if this second stage never happened—or better, has not yet happened. The "first stage" has taken place: the continuum of the autism spectrum has supplanted an earlier framework based on discrete entities of "autistic" and "normal." However, no "second stage" has thus far definitively taken place. What would a "Klein Grid" for autism look like? Despite its inevitable limitations, what possibilities might it present for rethinking autism experience, advocacy, community, and politics? In the following section, we engage with the work of autism activists and researchers to chart the outlines of alternatives to the autism spectrum. We link this to our own theoretical analysis, drawing inspiration from debates over models for sexual orientation. Our goal, as noted in the introduction, is not a definitive model or perfect metaphor for autism. Language and theory more generally is never adequate to its object. Rather, we seek to open a conversation regarding emergent possibilities for rethinking autism—and the consequences of such a rethinking for how autism is understood, diagnosed, and valued.

From sexuality to autism

Metaphors, as noted earlier, involve entailments moving from a source domain to a target domain. But metaphors themselves can travel as well. The notion of "coming out" as gay was "patterned on the debutante and masquerade balls of the dominant culture" (Chauncey 1995, 7); before that, the metaphor was transferred from the "confessional" relationship in Freudian psychiatry between therapist and patient, where it was borrowed still earlier from the Catholic confession (see Foucault 1978, 57–65). The spectrum represents, at least in part, another instance of such travel—in this case, from the domain of sexual orientation to domain of autism. This reflects links between homosexuality and autism both as historically medicalized discourses, and through explicit political analogies:

[R]esistance to neuro-normative pressure is reminiscent of responses to heteronormative mores of the (not so distant) past, where claims that homosexuality was an "illness" or "disorder" to be cured were clearly the result of professional and popular prejudice and misunderstanding. Many [autistic persons] feel responses to their own difference are similarly misplaced. (Davidson, 2008, 798; see also Bumiller, 2008, 968).

One strong link between homosexuality and autism in this regard has been their shared pathologization in the DSM. As one advocate with autism put it, "Stop worrying about the latest changes to the DSM criteria, and just remove autism from the DSM entirely, like homosexuality was rightly removed years ago" (Nick Walker, cited in Silberman, 2012).

As we have indicated, there is nothing inherently wrong with the spectrum metaphor. But given these discursive links, how might debates over the Kinsey Scale "travel" to activist and scholarly work questioning the metaphor's limits with regard to autism?

The unilinear Kinsey Scale was seen as an advance over the "discrete categories" rubric that framed homosexuality and heterosexuality as distinct domains (with bisexuality ignored, downplayed, or set apart as a third domain). By transcending that discrete categories framework and implying that almost everyone had some homosexual potential, the Kinsey Scale allowed for new understandings of what had been dismissed as "situational homosexuality" (for instance, in prisons); it also made thinkable new kinds of LGBT community and politics. In an analogous manner, the autism spectrum has been seen as an advance over the "discrete categories" rubric not only because it can facilitate claims to community between persons situated differently along it. Like the Kinsey Scale, it additionally entails the potential inclusion of all humanity: "although neurodiversity is most important to people who identify as being on the spectrum, it also has the potential to enrich society and change how we understand ourselves and other people" (Bumiller 2008, 982).

However, recall that a limit of the Kinsey Scale is its framing of the "spectrum" of sexual orientation along a single variable (for instance, behavior or desire). We find an analogy with the autism spectrum in that the "function" in "high functioning" is typically assumed to be the same "function" in "low functioning," just to a greater degree. In this regard it is interesting to consider how the Klein Grid's delinking of unilinear rubrics for sexual orientation is not unlike the claims of many autistic activists and allies that autistic experience is highly variable, with social, communicative, and perceptual impairments shifting in character (not just intensity) depending on time and context (see, e.g., Kim 2013; Nicolaidis 2012; Zeliadt 2016).

As a thought experiment, consider the general interventions of the Klein Grid in comparison to the Kinsey Scale. The grid replaced the unilinear variable of behavior with seven variables ranging from fantasy to social preference. What if, with regard to autism, we replaced the spectrum's unilinear variable of "function" with seven variables? What might they be? And what would be the consequences of columns like "past," "present," and "ideal?" How might such an "autism grid" more accurately represent autistic experience beyond the poles of "high" and "low"? How would it transform diagnosis? What social and political possibilities might it facilitate? What might be some limitations of that alternative metaphorical framework?

Our goal is obviously not to provide answers, but to contribute to a discussion regarding the possibilities questions like those above make conceivable. And such contestations of the "autism spectrum" are not limited to scholarly debates. Within autism communities, activists, advocates, and others have challenged the clinical accuracy and political implications of the spectrum metaphor. Consider the following informal online assessment developed by Leif Ekblad. Intentionally termed an "Aspie Quiz" in recognition of the Asperger's syndrome category that remains an identity category despite its elimination in the DSM, the assessment consists of 150 questions about a range of topics regarding sensory perception, social interaction, and behavior. The results from the scoring of these questions are presented in a graph (see Figure 3).

In place of the spectrum's single variable of "function," the Aspie Quiz includes variables like "perception," "talent," "communication," and "attachment." This assessment was explicitly designed as "a way to score neurodiversity traits in a novel way, independent of diagnostic categories" (Ekblad 2013, 12). Ekblad's conceptualization of neurodiversity reflects a conceptual framework that—like the Klein Grid with regard to sexuality—allows for multiple variables. Like the advocate's notion of an "Autistic Set" we discussed earlier in this article, Ekblad's quiz illustrates how autistic difference might be understood beyond the spectrum metaphor. This reflects longstanding suggestions for "multiaxial" diagnostic approaches to autism (e.g., Cantwell and Baker 1988), but moves that multiplicity beyond clinical assessment, into realms of selfhood and society.

These various rethinkings indicate that while no generally accepted (or even broadly debated) analogues to the Klein Grid yet exist, there is a growing sense that the autism spectrum has significant limitations—clinically, socially, and politically. In a manner analogous to the Kinsey Scale, the linearity of the autism spectrum obscures interactions between cross-cutting experiential domains, which may operate on multiple axes and be highly variable. When working within the entailments of the spectrum metaphor, it is difficult to conceptualize such multidimensionality—and that difficulty has consequences for subjectivity, community, diagnosis, and rights.

Formally in the clinic and informally in everyday social engagements and self-analysis, we are running up against the limits of a metaphor that makes it hard to represent how multiple domains imbricate to produce a unique autistic or "neurodiverse" subject. Clinicians, advocates, and autistic persons themselves increasingly challenge any idea that the spectrum can be divorced from the implication of a hierarchy of social value measured in terms of a single variable of "function." Among autistic activists there is growing resistance to the clinical and everyday use of "functioning labels." They are seen as tools of bigotry used against those labeled as "low functioning," and a way to deny services and accommodations to those labeled "high functioning."

The challenge to these labels is especially visible in online autistic communities. For instance, during her fieldwork Thomas encountered the Tumblr user ASDLife stating that when told he must be "very high-functioning autistic," he "always wonders whether it's because they actually think that they have any… comprehension of how well I may or may not function in different areas—or if they simply can't imagine that an autistic person who they'd consider low-functioning could be so… human?" ASDLife asserted that under standard criteria he might qualify as low functioning due to how his impairments shaped everyday activities. But in online environments, the idea encountered from some interlocutors that he must not be low functioning centered on the idea that his abilities to communicate via the Internet must correlate with advanced social skills, and even the ability to live independently. In his challenge to these conflations, ASDLife positioned himself as disrupting the smooth line of a spectrum. We can see another example of such a challenge in the writings of Amy Sequenzia, a popular non-speaking autistic activist who has been labeled "low-functioning":

I am one of those autistics who were said to be hopeless… The "experts" said I was "too low-functioning to learn." Of course, they were wrong…. But I am still non-speaking and I look very disabled. I also need a lot of help with things that are considered simple by most people. Maybe that's why the "low-functioning" label stuck. It happens to a lot of autistics like me. The assessment is incomplete and based on parameters that were created for non-autistics, by non-autistics, not taking into consideration the neurological differences of autistics. (Sequenzia 2013)

The question these activists pose is: how might we situate an autistic person whose proficiency in one area does not line up with their impairments in another? And more generally, how might alternatives to the autism spectrum open up alternative imaginings of ability, function, and human flourishing?

Conclusion

This article has been an exercise in conceptual alternatives. By drawing on lessons learned from models for sexual orientation, and also from the work of autism activists and advocates, we have worked to imagine autisms otherwise. But while this analysis is focused on ways of thinking, it has very real social and political consequences, and is particularly timely in two key ways. First, while the spectrum metaphor is currently hegemonic, from all quarters we see growing dissatisfaction with its limitations and increasingly active efforts to understand autism differently. These alternatives are not just rhetorical flourishes or helpful descriptions: they are incipient discursive shifts with powerful cultural implications. The spectrum model shapes how clinicians, researchers, patients, and the general public theorize the ways autism works. How might other metaphors aid in understanding experience, reducing stigma, and accessing medical, educational, and economic resources?

A second reason we consider our analysis to be timely is that in the wake of the publication of the DSM-IV in 1994, the spectrum—with all its entailments—rapidly moved from the domain of autism to a whole host of other psychological conditions, so that it is now possible to speak of the schizophrenia spectrum, the social anxiety spectrum, the posttraumatic stress disorder spectrum, the obsessive compulsive spectrum, and so on (e.g., Moreau and Zisook 2002; Schneier et al. 2002). Indeed, by the early 2000s some researchers were claiming that the spectrum metaphor could be used with regard to "aggression, personality, obsessions and compulsions, anxiety, depression, paranoia, sleep, concern with one's bodily appearance, panic attacks, impulsivity, and virtually every other symptom listed in DSM-IV" (Maser and Patterson 2002, 869, emphasis in original).

This claim that the spectrum metaphor could be extended into, in effect, all domains of the self has been presented as a challenge to a "categorical" nosology that framed disorders in either/or terms (Maser and Matterson, 2002, 855). As we have seen, such dissatisfaction with classification based on discrete domains goes back to the 1970s with regard to autism, and the 1940s with regard to sexual orientation. In the case of sexual orientation, a second wave of conceptual innovation represented by interventions like the Klein Grid highlighted shortcomings of the spectrum metaphor. But with regard to these other domains from aggression to impulsivity, little consideration of these limits has appeared to date, not to mention attempts to forge conceptual alternatives. As a result, we hope that our analysis in this article not only contributes to rethinking autism, but to the preemptive rethinking of the spectrum metaphor's shortcomings for understanding mind and society in the broadest sense.

Even with regard to autism, emerging frameworks remain overshadowed by the hegemony of the "autism spectrum." Despite increasing awareness of the metaphor's limits, a shift parallel to the shift from the Kinsey Scale to the Klein Grid has not fully taken place. The spectrum persists as a productive force that does not merely shape autism as a discursive object parsed in the realms of medicine, public policy, education, and family. The spectrum also continues to influence how autistic people understand themselves, their relationship to the category of normalcy, and their connections to others who share their diagnosis. We wish to emphasize once again that no metaphor is perfect and all have something to offer. As discussed in this article, autistic people continue to creatively employ the spectrum in everyday experience and activism. This reflects the complex ways in which the spectrum can support shared identity and community-building efforts.

But the metaphor can also reinforce notions of deficiency, and we have discussed how a growing number of autistic persons, as well as their doctors, advocates, and kin, now highlight shortcomings of the spectrum as a descriptive and analytic framework founded in a unilinear variable of function. This serves as a powerful reminder of the ways psychiatric models are shaped by cultural logics regarding what kinds of lives qualify as "functional." Mainstream conceptualizations of value, difference, and power can become framed in terms of perceived oppositional positions, between psychiatrist and patient, expert and non-expert, neurotypical and neurodivergent, "high functioning" and "low functioning." In turn, the highly medicalized language of the linear spectrum—with its implied hierarchy of social value based on normative assumptions of function—can be reproduced in everyday social exchanges beyond medical texts and clinical encounters.

With growing activism among self-advocates who deny that the spectrum accurately reflects the lived experience of autism, as well as increasing scholarly dissatisfaction with the metaphor's accuracy and utility, it may be that the autism spectrum will experience the same fate as the Kinsey Scale—a dethroning from naturalized "truth" to merely one conceptual approach, one with specific advantages and flaws. It behooves us all to attend carefully to these debates and transformations, not least because our ways of thinking are largely tacit, taking the form of a "common sense" few think to question. Explicitly addressing the discourses that help constitute our social worlds can open new possibilities for selfhood, for society, and for the crafting of our collective futures.