Theoretical problems with oversimplifying autistic diversity into a single category

Abstract

Mao (2023) presents a unified theory of language acquisition and processing which aims to bridge the gap between nativist and constructionist views on pragmatic competence. Mao argues that autism comprises a specific population which reveals how an Integrative Model of Pragmatic Competence works given apparent autistic difficulties in pragmatic competence but intact grammatical and lexical systems. Mao concludes that features perceived to be indicative of all autistic people (egocentrism and a lack of theory of mind) do not prevent linguistic competence. It is possible for internal modular components of language to function without recourse to intersubjective sociocultural engagement. It is argued this ultimately supports a nativist view over a constructionist view in relation to language acquisition, thus Mao proposes subsequent research should focus on neurobiological aspects of language acquisition and processing.

Our commentary will not focus on the broader linguistic debate about reconciling the positions between nativism and constructionism. Rather, we wish instead to focus specifically on the assumption Mao puts forward that autistic people constitute a homogenous category for studying the basic properties of human language. This assumption underscores theoretical problems within the linguistic ideas proposed which will be discussed below. It also has important moral and ethical implications in light of the way autistic people have been historically misrepresented as being entirely egocentric and unable to partake in authentic sociocultural life (Heasman & Gillespie, 2018; Ochs et al., 2004). Indeed, a more precise application of the double empathy theory (i.e., that there is a gap in understanding between autistic and non-autistic people due to two-way dispositional differences), highlights how double empathy is a relational construct and functions to critically examine how autism knowledge is produced. Thus, the double empathy theory can help to guard against the risks of unintentionally perpetuating overgeneralised deficit-framed stereotypes of autism. We therefore raise five issues with Moa's theory and suggest that a pathway forward to strengthen the theory would be to shift focus away from the complex label of autism and instead define populations for study based on specific linguistic measures relevant to the research question at hand.

First, Mao acknowledges that the label of autism describes a wide range of linguistic abilities from lacking functional language to competent verbal abilities. Such extremes in linguistic competence presents problems for the ambitious aim of developing a model of human language and its use that could resolve the divergent beliefs between constructionism and nativism. The problem is that while autistic people may demonstrate a range of linguistic abilities, the reasons why are multifaceted, as autism is not a linguistic diagnosis, it is a complex amalgamation of social, cognitive, sensory and interactional criteria. Across the autism spectrum there are vastly different competencies, developmental trajectories, associated learning abilities and co-occurring diagnoses that would greatly complicate attempts to create a unifying linguistic theory using this label alone. Moreover, there are notable differences in linguistic ability within this population, such as between autistic males and females, where females are observed to have a higher propensity for communicative and pragmatic skills (Sturrock et al., 2021). Therefore, any linguistic model that aims to develop hypotheses from studying autistic people needs to consider this complexity.

A second issue in theorisation results from how Mao presents ever-shifting definitions of autism to support a specific linguistic profile comprising two core features, an enduring insufficiency in intersubjective communication and lasting difficulties in mind-reading. This creates further problems as this “linguistic profile” is now no longer specific to autistic people. It is possible for someone to demonstrate difficulties in intersubjective communication but not be autistic, for example people who have been diagnosed with social pragmatic disorder but do not have any other criteria associated with an autism spectrum diagnosis. Indeed, there is also a wealth of literature on the difficulty non-autistic interlocutors experience with intersubjective aspects of language (e.g., the study of conversational repairs; Schegloff, 1992). What would have strengthened Mao's theory here is a more specific operationalisation of what is meant by intersubjectivity and how the criteria apply specifically to autistic people, as intersubjectivity has been used to describe a variety of different interactional situations (Gillespie & Cornish, 2010). Indeed, Mao appears to propose that immediate consensus is the only form of “successful” intersubjective communication, a view which contrasts with the more distal patterns of intersubjectivity observed between autistic interlocutors (Heasman & Gillespie, 2019b).

A third issue, related to the above point, is the claim that difficulties in mind-reading are a core component of autistic non-intersubjective language use. Mao makes a broad claim when describing autistic individuals as conducting self-sufficient thinking activities without mind-reading. To claim that autistic people completely lack the ability to understand others' intentions and thoughts is an oversimplification and inconsistent with contemporary evidence which highlights a much more nuanced picture of how autistic people interact with each other (Crompton et al., 2020), with familiar others (Heasman & Gillespie, 2018), and indeed, how non-autistic people struggle to imagine autistic minds (Sasson & Morrison, 2019). Mind-reading and theory of mind have received numerous critiques (Boucher, 2012; Dant, 2015; Yergeau & Huebner, 2017) which should be acknowledged within the rationale of the linguistic theory presented. However, Mao further conflates mind-reading with another construct of egocentrism, where autistic people are described as egocentric but capable of steady mental computations. This is a vague claim which is not indicative of the diversity of autistic subjectivity or consistent with autobiographical accounts of autism (Milton & Sims, 2016; Ridout, 2017). Egocentrism is a highly contentious construct as it very much depends on one's position within a social field as to whether speech can be classified as egocentric. Therefore, the use of this broad term when referring to autistic people has been debunked for many reasons, not least because it is not only autistic people that could be said to engage in egocentric speech (Begeer et al., 2016).

A fourth issue derives from evidence used to support arguments about autism. Dated references are used to support broad assumptions, such as the use of Baron-Cohen et al. (1985) to suggest that autistic people with high verbal ability have a very slim chance of understanding the intentions of others. This is an issue given the extent to which autism research has been enriched in recent years through both the neurodiversity paradigm and through the inclusion of autistic voice in research which had been previously “frozen out” of the literature (Milton, 2014). At other times, Mao argues by analogy with reference to savants, such as Stephen Wiltshire and William Christopher. Yet not all savants are autistic and savants themselves are an incredibly rare population. Savants are therefore not a suitable analogy for the wider autistic spectrum, nor as a basis for developing a universal theory of language.

These four issues concerned with accurately defining autism result in a fifth issue, the misapplication of the double empathy theory. Mao uses the double empathy problem as evidence to support the claim that linguistic pragmatic processing is fundamentally different between these parties. A closer reading of the theory, proposed by autistic sociologist Milton (2014), highlights that (1) the double empathy gap is a relational difference, not a deficit attributable to one party as Mao implies; (2) that the difference is caused by a range of factors far beyond linguistic processing style; and (3) the gap is amplified by harmful stereotypes about autism which circulate within science and popular culture.

The idea that (1) the double empathy gap is a dyadic relational construct rather than an individual one impacts strategies for reducing said gap. Specifically, it follows that reducing dispositional difference (or increased similarity) would lead to a reduction in the gap, and empirical research has shown this to be the case within a communicative context where autistic peer-to-peer communication is successful, but autistic to non-autistic communication less so (Crompton et al., 2021). Indeed, differences between autistic and non-autistic people can even be ameliorated by supporting shared interactional experiences (Chapple et al., 2021). However, Mao situates pragmatic difficulty as internalised to the autistic individual alone. Under this proposal, autistic parties sharing the same pragmatic difficulties would not be able to overcome a double empathy gap. Thus, the fit between double empathy and the linguistic theory Mao presents appears to not be aligned.

The idea (2) that double empathy is attributable to factors beyond linguistic features alone has already been addressed above. However, (3) the potential to reinforce negative stereotypes of autism through scientific endeavour does require attention, as has been highlighted through the study of looping effects and epistemic injustice (Chapman & Carel, 2022; Hacking, 1999). The representations of autism created through scientific inquiry can impact the way autistic people see themselves, often in negative ways given their focus on deficits (Milton, 2014). Autistic people navigate stigma at various levels while also historically having their voice and personal lived experience overlooked (Botha et al., 2022; Heasman & Gillespie, 2019a). For this reason, it is important for studies to take a careful and sensitive approach towards the topic of autism which recognises the potential for autistic people to read, research and engage with the work presented.

For Mao's linguistic theory to be taken forward, we would strongly recommend avoiding the heterogenous label of autism to illustrate a universal property of language acquisition. Rather, a more precise measure of verbal and pragmatic competence, specific to the research question at hand, would help to yield a more robust exploration of the ideas presented.