Language

1. Introduction

Mikhail Kissine's 2021 target article, 'Autism, constructionism, and nativism', brings up a very important question that the field of linguistics has barely begun to address: what are the implications of autistic people's atypical language development for the competing schools of thought about L1 acquisition? Rather than attempting to answer Kissine's question, this response instead attempts to point out its underlying assumptions, in the hope that a deeper understanding of the question will guide successful future research on the matter.

In a thread on November 27, 2020, Twitter user @autisticats (an account maintained by three autistic people named Jasper, Eden, and Leo) wrote:

Autistic people, like Deaf people, are a sociolinguistic group with our own culture and norms of communication. The communicative burden in conversation between autistic people and NTs [neurotypicals] should not fall solely on autistic people. Communication requires effort on all sides.¹

The authors provide several citations in support of their assertion (Bell 2007, Crompton, Ropar, et al. 2020, Davidson 2008, Kasirer & Mashal 2014, Morrison et al. 2020, Wu et al. 2014). This tweet represents a perspective that is largely missing from Kissine 2021—the perspective through which autism is considered to be a category of identity or a community of practice rather than a neurological deficit, a perspective that is often referred to as the neurodiversity paradigm.

2. The neurodiversity paradigm

For several decades, scholars of disability studies have distinguished between the culturally ascendant medical model of disability and a contrasting model in which disability is considered to be a socially constructed category (Shakespeare 2017). Increasingly, autistic people have strongly favored the social model of disability and the interpretation of autism as an identity rather than a disorder (Kapp et al. 2013, Woods 2017)—in the neurodiversity paradigm originated by autistic sociologist Judy Singer in 1998 (described in Singer 2016), autism simply refers to one [End Page e211] of several natural variations in human neurology. As autistic scholar Nick Walker (2013) puts it:

If you reject the fundamental premises of the pathology paradigm, and accept the premises of the neurodiversity paradigm, then it turns out that you don't have a disorder after all. And it turns out that maybe you function exactly as you ought to function, and that you just live in a society that isn't sufficiently enlightened to effectively integrate people who function like you. And that maybe the troubles in your life have not been the result of any inherent wrongness in you.

(Walker 2013:236)

If autism is considered as a marginalized identity rather than as an impairment or deficit, our entire approach to the question of autistic language acquisition should change, as Sterponi et al. (2015) suggest. Rather than defining autistic communicative competency in terms of how well a person succeeds at passing for neurotypical (an essentially harmful practice that autistic people call masking or camouflaging and which may be responsible for the elevated rate of autistic suicidality; see Cassidy et al. 2018, Wilkenfeld & McCarthy 2020), we should use the tools of sociolinguistics to characterize the differences between autistic and allistic (nonautistic) speech, to examine the role of language attitudes and the construction of identity in autistic social interaction, and to determine whether principles identified in the literature on multilingualism, translanguaging, and intercultural communication are applicable to the autistic/allistic interface.

Much of Kissine's argument rests on his assertion that the atypical communication development experienced by autistic people is the result of a deficient theory of mind or lack of skill at mind reading. However, more recent studies have critiqued this explanation (Gernsbacher & Yergeau 2019) and supplanted it with the double empathy problem identified by autistic scholar Damian Milton (2012), whereby difficulties in communication between autistic and allistic people are caused by mutual lack of comprehension of the other person's cognitive processes and cultural practices, rather than a defect on the part of the autistic person. In other words, autistics make poor guesses about what neurotypicals are thinking—and vice versa—because the two groups' perceptual and processing experiences are so different, not because autistic people inherently lack the ability to guess effectively (a conundrum that bears a passing resemblance to the gavagai thought experiment; see Quine 1969).

The double empathy problem has been repeatedly demonstrated experimentally: autistic people effectively transfer information among themselves (Crompton, Ropar, et al. 2020) and autistic people prefer the company of people with autistic traits (Granieri et al. 2020); both allistic and autistic participants rate interpersonal rapport higher in matched-neurotype interactions than in mixed-neurotype interactions (Crompton, Sharp, et al. 2020), and allistic people are ineffective at interpreting the behavior of autistic people (Sheppard et al. 2016). In fact, according to Sasson et al. (2017), allistics form negative impressions of autistics even when presented with a static image, when no linguistic cues at all are present. In this way, negative allistic evaluations of autistic communication may merit consideration in light of the findings of decades of matched-guise studies (originating with Lambert et al. 1960) that demonstrate comprehension difficulties in intercultural communication caused by the language attitudes of the hearer as much as by the linguistic proficiency of the speaker (e.g. Rubin 1992). To my knowledge, no one has yet completed a matched-guise study comparing autistic and allistic speech. But these studies suggest that, at the very least, autistic speech proficiency should be judged according to autistic standards rather than neurotypical ones.

The pragmatic misunderstandings Kissine identifies as autistic deficits—difficulties with metaphor, indirect requests, and irony—could likewise instead be attributed to intercultural miscommunication. For example, Sharifian (2014) identifies metaphor-based [End Page e212] communication difficulties between Aboriginal English and Australian English speakers, le Pair (1996) describes nonnative use of indirect requests by Dutch learners of Spanish, and Kim (2014), in her discussion of Korean speakers' difficulties in learning English-language irony strategies, notes that 'appropriate understanding of sarcasm in L2 is acknowledged to be a great challenge for L2 learners' (Kim 2014:193). The one-way difficulties in understanding Māori humor experienced by Pākehā New Zealanders (as described in Holmes & Hay 1997) may, as the Autisticats argued (see above), allow a useful comparison with the experiences of autistic people, who often report feeling like foreigners within their own native culture (Hacking 2009). This would be consistent with the finding of Ochs et al. (2004) that autistic children have less conversational success in discourse involving sociocultural indexicality. Further study of pragmatics in autism, as Salt (2019) proposes, should prioritize conversation analysis of intra-autistic communication to determine whether autistic-specific discourse norms can be identified in contrast with those of their neurotypical linguistic environment.

3. Neurodivergent identities and language acquisition

In identifying potential autistic discourse practices, it is important to remember that, like other social identities such as sexual orientation and physical appearance, autism is also heritable (Sandin et al. 2017). In fact, many autistics are raised in families where one or more members share their neurocognitive and sociocultural attributes—either because they are also autistic (so-called 'multiplex' families) or because they embody the broad autism phenotype (Gerdts & Bernier 2011), displaying autistic traits even if they are not recognizably autistic (Losh et al. 2017). Thus, at least some autistic children (perhaps up to one-third; see Sasson et al. 2013) acquire their L1 in an environment influenced by other autistic people. To my knowledge there has been no observational research performed on naturalistic L1 acquisition by the children of autistic parents, which could help determine whether autistic discourse norms are being learned by either autistic or allistic children.

A further complication to evaluating autistic language acquisition and linguistic capabilities is that distinctions between autistic and allistic language use may be less salient in writing (Newton et al. 2009, Dansereau & Flanagan 2019), an observation that calls out for further study. Sasson et al. (2017) found that allistics' negative evaluations of autistics occurred only when they had audio or visual information available; allistics did not negatively evaluate written transcripts of autistic speech. Autistic activist Martijn Dekker famously wrote that, for many autistics, the internet is 'what sign language is for the deaf' (Dekker 1999), and nonspeaking autists like author Amy Sequenzia and activist Mel Baggs have demonstrated that an inability to speak orally does not entail a lack of linguistic ability. An unknown yet apparently large percentage of autistic people are hyperlexic, demonstrating advanced and precocious literacy skills (Ostrolenk et al. 2017), and it is clear that our existing techniques for measuring autistic cognitive ability based on verbal ability are unreliable (Alvares et al. 2020, Bal et al. 2016, Prince 2010).

It is likely not a coincidence that the proliferation of orthographic messaging technology has correlated with the flowering of autistic culture and visibility. Yet both nativist and constructionist accounts of language acquisition rely on the basic notion that written communication is not language. Autistics who use augmentative and alternative communication (AAC) because apraxia prevents them from typing demonstrate similar communication practices to allistic typists (Jaswal et al. 2020), yet professionals are overall extremely reluctant to provide this assistive technology to autistic children [End Page e213] in the early stages of language acquisition (Dugan et al. 2006), a practice comparable to the audiocentric bias that Deaf activists have combated for decades (Eckert & Rowley 2013). Given the dearth of existing studies on autistic literacy development (Westerveld et al. 2016), we should devote more research to this topic and perhaps give serious consideration to the idea of literacy as an additional language modality, rather than merely a language technology, as suggested by scholars of multimodality (e.g. Hawkins 2018).

Kissine dismisses the idea that autistics and neurotypicals may use different neurocognitive mechanisms for language learning (Kissine 2021:e146), but this idea merits further examination in light of the fact that many holistic neurocognitive theories of autism—such as monotropism (Murray 2019, Murray et al. 2005), enhanced perceptual functioning (Mottron et al. 2006), hyperreactivity (Markram & Markram 2010), and hyperconnectivity (Supekar et al. 2013)—have direct relevance to traditional neurolinguistic understandings of language acquisition. If autistics are cognitively overwhelmed by real-time language's requirement of simultaneous parsing, interpretation, and response formulation, as autistic activist Sandy Yim (2009) hypothesizes, real-time written interaction (including captioning; see Zdenek 2015) may provide autistics with the opportunity to process language in serial rather than in parallel, easing the cognitive load.

4. Final thoughts

In sum, Kissine makes a good point that the debate between nativism and constructionism could greatly benefit from incorporating information from the study of autism. Unfortunately, much of the most helpful research in this regard has yet to be completed, largely because the field of linguistics, like the rest of our society, is vulnerable to institutionalized ableism (Bottema-Beutel et al. 2021). Linguists should do more to incorporate the perspectives of autistic adults (Gillespie-Lynch et al. 2017) in their research and should prioritize the work of autistic scholars on the principle of 'nothing about us, without us'. Despite these concerns, however, it benefits all of us and our field that Kissine has drawn our attention to questions about autistic language acquisition, and there is reason to hope that this will lead to more fruitful future research.