“It forced me to redefine autism.” When Douglas Biklen spoke these words, in 1990, about facilitated communication, he did not realize how insolent it sounded to the psychiatrists and psychologists who had labored in the field for decades. With nothing more than a degree in special education, Biklen was challenging “the received knowledge coming from almost fifty years of energetic and sustained research into the condition of autism.” To the established experts, it was appalling that Biklen, whom Eric Schopler called an “ideologue,” would dare claim the right to redefine autism.
And yet that was exactly what the experts had been doing since the time of Leo Kanner—debating one another and jockeying to advance their own definitions and redefinitions of autism: its shape, its scope, its causes. This continuing push and pull happened mostly out of sight of the lay public, which assumed that when people with PhDs and MDs used the word “autism,” they all meant the same thing. But that was never true, because the definition of autism has always been malleable—determined by consensus, or by whoever among the jockeying experts was most persuasive at any given moment. All the big ideas about autism over the years—from mother blaming to concepts with more solid science behind them, like mindblindness, weak central coherence, and the unifying spectrum—were still only hypotheses. Yet most were treated, at the height of their popularity, as objective fact. At least until the next big idea came along.
Given the elasticity of the definition, it has been easy for competing interests to cast autism as what they want it to be, which has at times led to enormous strife. Over the decades, autism has been seen as a behavioral, or a medical, or a psychogenic challenge. It has been blamed on mothers and it has been blamed on vaccines. It has been answered with “substitute mothering,” megavitamins, facilitated communication, slaps in the face, hugging therapy, and, in some African countries, with exorcism. At this writing, autism is viewed as occurring on a spectrum, a concept that has the virtue of promoting inclusiveness but has also caused acrimony between those who see autism as a tragedy and those who claim it as a gift and an identity.
For years, many of the most respected figures in the autism community have complained about the lack of a single, incontrovertible definition of autism. In 1968, for example, Michael Rutter lamented the “hopelessly confused state of affairs…more confusing than it need be.” A full ten years later, in 1978, Eric Schopler warned that scientists were critically misunderstanding one another’s work “because criteria for diagnosis are different.” One more decade after that, in 1988, Fred Volkmar and Donald Cohen wrote in frustration about the “long and controversial history of the concepts surrounding autism.” And in 1998, Volkmar and Ami Klin complained about “the confusion and the plethora of diagnostic concepts” surrounding autism, and, by then, Asperger’s syndrome as well.
One of the autism community’s favorite adages has long been “If you’ve met one person with autism, you’ve met one person with autism,” which captures the essence of the problem even as it affirms the complex variability of autistic traits. It is this enormous variability of autistic traits that makes defining autism so challenging. That was what hampered Victor Lotter’s effort to count people with autism in the early 1960s, when he had to draw a line to define autism that even he considered arbitrary, or at least subjective, given that it was based on only his own observations. Nevertheless, the subjective judgments of Lotter and other researchers had real consequences, especially in the field of epidemiology. Valid comparisons among populations in different eras and places are impossible when definitions are highly inconsistent. Lotter thought the lesson worth sharing in 1966, when he cautioned that the true prevalence of autism would remain impossible to determine until everyone agreed on what it was they were trying to count.
Debates about who counted as autistic were not only academic; they had a profound effect on the lives of individuals. After all, a person must be diagnosed with autism before he or she can gain access to support services.
Repeatedly, one researcher or another would design a set of criteria intended to bring an end to the chaos—proposing a definition or working model meant to stand as the last word on the matter of what is meant by autism. The effort put into this was enormous. The goal, however, remained elusive.
MILDRED CREAK HAD tried to bring order to chaos in 1961, when she published her Nine Points. These were quickly found wanting by the clinical world, which tore them to shreds. She likened the experience to feeling as if she were “throwing them to the wolves.”
Michael Rutter also tried to make sense of the mess in 1972, when he successfully campaigned to retire the diagnosis of “childhood schizophrenia,” which had long competed with the diagnosis of autism in children. It was “high time,” Rutter argued, for that troublesome concept to be “politely and respectfully, but firmly, put into its proper place in the section on ‘the history of psychiatry.’ ” Still unsatisfied, he produced a new framework for autism, which became known as the “Rutter criteria,” in the early 1970s. Based on Rutter’s and other researchers’ epidemiological and clinical studies, the criteria specified that autism manifested itself in three domains, in varying degree—through language impairment; social impairment; and restrictive, repetitive patterns of behavior. An inspiration for Lorna Wing and Judith Gould’s later “triad of impairments,” the Rutter criteria gained a following, and for several years thereafter they were often cited by researchers as the standard they relied on. Some of this acceptance was due, no doubt, to Rutter’s standing as one of academic psychiatry’s most lucid thinkers and writers. He was adept at defining the parameters of a topic, and he had a clear, unpretentious style that made his conclusions that much more forceful. “The question was not how to differentiate autism from normality. The hospital porter could do that without skilled assistance,” he once wrote, on the challenge of clarifying what autism looked like in children seen at the Maudsley.
By 1977, however, the National Society for Autistic Children was promoting its own definition, which added a fourth domain: sensory processing. Developed by UCLA researchers Edward Ritvo and B. J. Freeman, this definition made “abnormal responses to sensations” also central to the experience of having autism.
And then there was the American Psychiatric Association’s DSM, which constituted its own saga.
IN 1993, WHEN Fred Volkmar and his colleagues were trying to get Asperger’s syndrome into the DSM, the book had already been in print for more than forty years, with revised editions published whenever psychiatry changed its mind, which was surprisingly often.
After 1980, for example, the DSM was overhauled at least once a decade, with whole diagnoses heavily revised or dropped altogether, and new diagnoses added. Most famously, after much pressure from the gay community and its allies, and despite the vociferous objections of a contingent of mostly older psychiatrists, the DSM retired the diagnosis of homosexuality as a mental illness in 1972. Though this and other amendments were billed as the positive consequences of better research and understanding, they revealed the extent to which psychiatric diagnosis is influenced not only by scientific data but by cultural, political, and other factors.
Autism was not included in the DSM until 1980, in the edition known as DSM-III, which appeared nearly forty years after Leo Kanner published his first paper on the condition. It proved to be a protean definition—altered significantly in 1987, and again in 1994. Although it was changed only slightly in 2000, it underwent a dramatic rewrite, accompanied by much angst, in 2013. Throughout this process, the length of the definition waxed and waned. The number of words in its symptom checklist, originally about seventy, jumped to more than six hundred in one edition, shrank to three hundred in the next, and then expanded again, to almost nine hundred words, two editions later. The name kept changing too—from Infantile Autism to Autistic Disorder to Autism Spectrum Disorder. Most critically, the symptom checklist the DSM writers developed to identify those with autism kept being revised. In one edition, a patient had to match a minimum of 8 out of 16 criteria to receive the diagnosis. In another, he or she had to match at least 6 out of 12. In the second case, according to Volkmar and Brian Reichow at Yale, there were more than two thousand different symptom combinations that could produce an autism diagnosis. The addition of Asperger’s disorder as a separate diagnosis in 1994 further complicated the diagnostic picture.
Very often, the DSM’s zigzags on autism have reflected the almost philosophical divide between two camps of experts: the “lumpers” and the “splitters.” This tension was on display when Lorna Wing and Hans Asperger disagreed as to whether his cases and Leo Kanner’s were more alike than different, or the other way around. Asperger, it will be remembered, believed that he saw two distinctly different syndromes in his cases and Kanner’s cases. His insistence on splitting his and Kanner’s children into separate groups, according to key differences, made him a classic splitter. Wing, on the other hand, always spoke like a true lumper, arguing that virtually everyone with some trace or combination of autistic behaviors belonged to the same group, since the core mechanisms of autistic behavior operated in all of them in the same way, albeit to varying degrees. This thinking is what led to her construction of the highly influential “spectrum” model of autism. Today this quintessentially lumper idea is widely accepted, to the point where even popular culture appears to embrace it as unassailable truth.
But if history is a guide, that could change. As with any previous understanding of autism, the spectrum construct is not without drawbacks, and a swing back to the splitters is quite possible if, for example, further scientific inquiry determines that clusters of autistic behaviors that look similar in fact arise from divergent causes—environmental, genetic, or other. The study of cancer offers an analogy. For years “cancer” survived as an umbrella term for all cancers, until the field of oncology learned that different cancers are in fact radically different, each with its own signature and its own distinct biological process—and was thus forced down the path of the splitters. Autism too might turn out to be a plural entity. Researchers might discover that there are in fact autisms, each with its own genetic “fingerprint,” each significantly unrelated to the others. If that were to happen, the idea of a broad spectrum, which smudges those differences, would have less explanatory power, and might lose some of its currency in the conversation about autism.
Such lines of inquiry, however, would require a commitment to research into the biomedical aspects of autism rather than a focus on psychiatric or behavioral interventions. Yet at the beginning of the 1990s, there was virtually no biological research into autism, and the few scientists studying it had little financial support. That is, until four parents, living on opposite coasts of the United States, decided to do something about it.