Discussions about individuals with autism and related diagnoses are now pervaded by references to a putative "spectrum." Individual patients are said to be "on the spectrum." A range of developmental disabilities are referred to as "autism-spectrum disorders" or "ASD." The putative justification for this terminology is that there is great variation among those who share the diagnostic label and it is believed that "spectrum" effectively conveys this diversity of abilities, behaviors, and symptoms.
On the contrary, "spectrum" is seriously inadequate to the linguistic task for which it has been pressed into service. The chief reason is that a "spectrum," such as the electromagnetic spectrum, corresponds to the continuous variation along a single dimension. In the case of the electromagnetic spectrum, the variation is in wavelength. Individuals with autism and so-called ASD, however, exhibit variation along multiple, at least somewhat independent, dimensions, including: cognitive ability, extent of verbal activity, interest in social engagement, frequency and types of repetitive behavior, and areas of intense interest.
Such diversity cannot be properly represented as single magnitudes (more or less "autistic") continuously arrayed along a single axis. Proceeding as though it can be so described inevitably distorts thinking about the pertinent conditions and those affected by them.
A more useful metaphor than "spectrum" is "cloud." In other words, individuals with autism are better represented by points distributed in a multi-dimensional space than by points distributed along a one-dimensional line. Where the boundaries should be placed between those deserving a clinical diagnosis and those with related tendencies who nevertheless are not accurately captured by the label is a major challenge both intellectually and practically.
Therefore, it is disappointing when anyone with one or more traits generally associated with autism (or related diagnoses) is labeled as "autistic." The economist Tyler Cowen has written, in The Chronicle Review, about the advantages for academic life of being autistic. He speaks, for example, of the ability to focus intensely on one subject, thereby appearing to equate a restricted range of interests, a typical finding in individuals with autism, with the sort of intense concentration often necessary for serious scholarship.
This equating of a relatively narrow range of intellectual engagement and the ability to apply oneself to a particular task for extended intervals of time is dubious. Scholars capable of working on a book or article for several hours without significant distraction are not necessarily lacking in a broad range of interests overall nor are they necessarily exhibiting autistic behavior by pursuing their academic mission assiduously. The preceding does not deny that some individuals with autism or autism-related traits can produce academic work of high quality or even that some of their autism-associated traits might not be beneficial to their academic pursuits. However, caution is appropriate in inferring that the successes of such individuals can be attributed solely to the traits putatively connected with their autism.
Other hallmarks of autism are not necessarily and absolutely unique to autism as a clinical condition. A relative lack of empathy, although clearly found in most, or at least many (some authorities might say all), individuals diagnosed with autism, is not limited solely to those with autism. For example, narcissists and sociopaths also exhibit deficiencies in empathy.
Inevitably, given the mechanics of inheritance, the traits associated with autism can be found in parents and other relatives of affected individuals in widely varying combinations and to varying degrees. It is neither accurate nor productive to extend the label of "autism" to such individuals unless there is compelling evidence based on detailed knowledge of their full repertoires of abilities, tendencies, and behaviors.
The recent advances in associating genetic variants with autism and in characterizing the neurological correlates of autism notwithstanding, it remains true today that the understanding of autism is relatively primitive. None of the research that has come to my attention can justifiably claim to understand the etiology or pathogenesis, in terms of cells and molecules, of the conditions grouped under the "autism" label. The treatments, including the much-vaunted pharmacologic fruits of the revolution in "biological psychiatry," are almost completely empirical, as likely to be ineffective, or even cause harm, as to lead to sustained benefit.