> I have always tried to explain ASD not as a one-dimensional spectrum but as a set of spectra (oversimplified: one spectrum for each trait) that when combined form a picture that describes a persons autistic behavior.
The term "spectrum" gets used in two different senses (1) a a continuum of severity, stretching from "normal" individuals, to the "mild" cases, through to the "severe" ones (2) to refer to the heterogeneity of different symptoms, such that two people might have a very different mix of symptoms despite having "equal severity". I think both ways of understanding the term are legitimate; some people insist that only sense (2) is correct and sense (1) is wrong, but in fact you'll find both senses being used in the research literature.
In some ways it is justifiable to view ASD as a two-dimensional construct: DSM-5 ASD is defined in terms of two domains, the social communication domain and the RRBI+sensory domain. And the DSM-5 says that a diagnosis should provide a level number from 1 to 3 for each of those two domains, and in principle you could have a different level in each domain. (That said, my personal impression is few clinicians actually do that in practice.)
ASD is not unique in being a "spectrum". There is also the schizophrenia spectrum (DSM-5 has a whole section called "schizophrenia spectrum and other psychotic disorders"), the obsessive-compulsive spectrum (DSM-5 didn't end up including that concept in its main text, although it was being considered for inclusion, a fact which the front-matter of the DSM-5 briefly alludes to), the bipolar spectrum (not mentioned in the DSM-5, but widely discussed in the literature), among others. In fact, for many different psychiatric/neurodevelopmental disorders, you'll find people in the literature adopting a spectrum-approach to them. But in the popular consciousness, the word "spectrum" gets treated as something specific to ASD rather than the much more general psychiatric/psychological concept which it is.
> I have similar worries about the idea of a hard boundary between ASD/non-ASD, mostly because I believe that the concepts of "neurotypical" and "neurodiversity" are social constructs we have created because the majority of people behave in a similar way, and everyone who does not is an outlier. We gave the majority behavior a name "neurotypical" and that created an "us vs them" polarization.
I agree with you, but I have another concern as well – the evidence for the "neuro" part just isn't there. The distinction between the "neurotypical" and the "neurodiverse" is actually clinical – it is based on the presence or absence of clinically significant dysfunction, in other words it is based on how well you function in contemporary society, not on any individualised study of brain structure. There is actually genuine neurodiversity among "neurotypicals". For example, most right-handed individuals activate left perisylvian regions of the brain during language processing, but a small minority of right-handed individuals activate right perisylvian regions instead. That small minority are an example of genuine biological neurodiversity. But nobody is going to call them "neurodiverse" unless they have some clinically significant problems with functioning in society, and that example of neurodiversity has no known clinical significance. (Indeed, almost nobody in that small minority would know they are in it, because unless you get someone to do language tasks under MRI, you can't tell whether they belong to it.) The "neurotypical"/"neurodiverse" distinction is fundamentally a clinical distinction masquerading as a neuroscience one. Hence I worry that the us-vs-them polarisation that you mention is both socially harmful and scientifically illiterate