Pragmatics in autistic communication
Observed pragmatic differences in autism may be framework variation rather than deficiencies
The following is an adaptation of my final research paper for PHIL 147: Philosophy of Language, taught by Professor Bernard Nickel with teaching fellow Justin Cavitt (Spring 2023).
Contents
1. Abstract
Observed pragmatic differences seen in autism spectrum disorder can be better explained as a difference in pragmatic principles rather than a deficiency in pragmatic skill. Although many studies have shown that communicative autistic people show lower performance in a variety of pragmatic tasks, including parsing ambiguity, understanding figurative language, and providing appropriate information in conversation, other work suggests that these results may not arise from an inability to grasp pragmatics. Recent studies indicate with autism have a coherent and consistent set of communication principles that differ from non-autistic communicators, and autistic communicators do not universally lack the ability to understand the pragmatic reasoning of non-autistic people. This evidence challenges a model where pragmatics is derived from processes like rational decision-making, though additional work is needed to support these objections.
2. Introduction
Autism spectrum disorder1, as described by the Diagnostic and Statistical Manual of Mental Disorders, has the essential feature of “persistent impairment in reciprocal social communication and social interaction”2. People with autism spectrum disorder demonstrate difficulties in understanding figurative language, parsing ambiguous meanings, detecting pragmatic violations, and providing appropriate information in conversation34. Although this evidence is considered indicative of pragmatic weaknesses in autism, additional work suggests that the results cannot be simply explained as inability. Current work supports a “double empathy problem” between autistic and non-autistic groups, where social impairments result not from social deficit, but a mismatch in perception and attitude that affects both autistic and non-autistic people5. The communication divergence underlying the double empathy problem may be rooted in a different set of pragmatic rules between the groups.
3. Pragmatics
Grice, in Logic and Conversation, established the foundations of pragmatics by proposing that conversation is rational and strategic, and communicators cooperate to encode more meaning into statements than the literal semantics of an utterance. Grice proposes that conversation follows maxims of quantity, quality, relation, and manner, which allow for listeners to draw implicatures6.
Implicatures are conclusions that can be drawn from speech that are not necessarily logical entailments (necessarily true). For example, statements that use “some” often produce the scalar implicature of “not all”, even though “some” only entails “not none”. This implicature arises because “all” is more informative than “some”, as the former is a strict subset of the latter.
However, the maxims do not fully explain all features of non-literal language, which are often dependent on social context, shared perspective, and accepted cultural norms. For example, there are gendered differences in usage of verbal irony or sarcasm, with authors attributing the more frequent usage of irony in men as indicative of a higher inclination toward risk-taking rather than a different set of pragmatic goals7.
These considerations do not completely rebut Gricean machinery, as the original framework allows for the incorporation of additional information for inference. It is probably uncontroversial to claim, for example, that someone will adopt a different set of pragmatics to parse the veiled, passive-aggressive language of a professional meeting compared to an open, straightforward, and private conversation with close friends. Better understanding of the factors that can affect the situational pragmatic framework used by communicators can further the understanding of the communication differences seen in autism.
3.a Pragmatics in autism
Although some autism researchers claim that “pragmatic language impairments are universal” in ASD8, this group may be better characterized as having pragmatic differences rather than deficits. A review of pragmatic studies in autism, conducted by Loukusa and Moilanen in 2009, found that autistic people (ranging from 6 to 57 years of age) consistently displayed “weaknesses” across a variety of pragmatic tests, including idiom comprehension, non-literal language interpretation, and contextual inference from stories9.
However, many studies included in the review find that pragmatic abilities are well-conserved for simpler pragmatic tests in both autistic children and adults. Additionally, a 2018 study of scalar implicatures found that autistic and non-autistic children had similar judgments when expressing agreement/disagreement with underinformative statements. Differences were noted when children were presented with a third, middling option (“I agree a bit” rather than “I agree” or “I disagree”), with autistic children preferring the stronger binary responses10. A similar study performed with adults drew similar conclusions, finding no significant differences between scalar implicature judgments between the autistic and non-autistic participants11.
The preservation of simpler pragmatics and divergence in more complex communication suggests two possible theories. First, consistent with prevailing views, pragmatics deficiencies are only visible at more complex tasks because the pragmatics skill or talent in autistic people is only sufficient to handle simpler inferences. However, these results could also suggest that autistic people have the same pragmatics machinery as non-autistic people, indicated by the similar performance in the straightforward, rationality-derived inference tests, and divergence in more complex tasks arises from differences in dealing with uncertainty when rational inference does not resolve all the ambiguity in the statement. These differences will be explored later, in Section 5.b.
4. Proposed explanations of pragmatic weakness are insufficient
This section will explore arguments against the first interpretation that autistic people lack pragmatics skill. The previously cited review by Loukusa and Moilanen identified two main streams of thought for explaining deficiency: weak central coherence and impaired theory of mind.
Weak central coherence proposes that individuals with autism tend to have difficulty interpreting information because of lack of an ability to integrate information from different sources and tendency to interpret utterances in isolation. This reasoning is sometimes also subsumed into a broader theory of executive dysfunction, which claims that autistic individuals may have impaired planning, ability to allocate attention, and mental flexibility. Additionally, some autism researchers propose that theory of mind is related to ironic understanding, claiming that pragmatic comprehension difficulty is related to difficulties with theory of mind in autistic individuals.
These current explanations seem contradicted by existing evidence that autistic people do possess these capabilities. For example, autistic individuals appear to score similarly to non-autistic controls on simple theory of mind assessments12. Additionally, some autistic people are competent at modeling the mental processes of non-autistic communicators, which will be explored later in discussions of masking.
4.a Shortcomings of weak central coherence theory
Weak central coherence is not consistent with other, non-verbal studies of autistic individuals. Lopez and Leekam contrasted the ability of autistic children and IQ-matched controls to extract information in different situations, finding that autistic children did not significantly differ from non-autistic peers in tasks of accounting for visual/verbal context and visual/verbal semantic memory. Autistic children only showed differences in an experiment of disambiguating homographs, suggesting a specific divergence in complex verbal stimuli that required using sentence context to disambiguate meaning13.
However, an eye-tracking study observed that in verbal contexts, autistic children use context to quickly resolve ambiguity, similar to non-autistic controls14. Differences may be related to specific semantic tasks that require an autistic listener to account for the behavior of a non-autistic speaker, rather than central coherence inability. However, it may be possible that weak central coherence theory could be modified to only suggest connectivity deficiencies in the auditory processing stream while allowing for retained performance in other areas, like visual processing.
Additional evidence in autistic adults also does not clearly support weak central coherence. A study found autistic men performed as well as non-autistic controls on central coherence tests, though the former group took significantly longer on the same tasks15. This evidence could support weak central coherence, but does not reject the theory that those with autism may have different intrinsic preferences for how they use contextual evidence and require more time to adjust to the “correct” standard answer of the test battery.
4.b Masking demonstrates autistic understanding of non-autistic pragmatics
Additional evidence of autistic “masking” supports a model of many autistic people being capable of identifying and executing appropriate conversational moves. Within the last decade, more studies have focused on how women and girls with autism can, with some effort, mask autistic traits and appear to be non-autistic16. One of the earliest studies on camouflaging, conducted in 2016, noted that many of the respondents reported “pretending to be normal”, with many practitioners overlooking an autism diagnosis due to their ability to hide autistic traits17. In many cases, autistic women may never receive a diagnosis or exhibit any observable functional impairments18.
Even with “successful” camouflaging, autistic women report high levels of stress, anxiety, and exhaustion19. The study of camouflaged autism remains under-studied, though progress has been made in the development of the Camouflaging Autistic Traits Questionnaire (CAT-Q), which reflects observations that autistic individuals may practice “normal” social indicators like body language and facial expression and closely observe and imitate non-autistic communicators20.
These results suggest a weak rebuttal of the assumption that pragmatic deficiencies are a hallmark of autism. For at least some autistic individuals, there is no underlying inability to imitate the mannerisms of non-autistic communicators. Possibly, distress with masking suggests that the pragmatics "skill" of this group is impaired, though
5. Different but coherent pragmatics in autism
The above evidence suggests that theories of autistic deficiency insufficiently explain observed autistic communication patterns. Additional evidence on autistic peer-to-peer communication further suggests that autistic people converge on a different pragmatic framework.
5.a Empirical evidence from autistic-to-autistic communication
A study by Crompton et al. found that groups of all-autistic and all-non-autistic communicators had similar levels of rapport and information retention, and a group of mixed autistic and non-autistic communicators saw decreased rapport and communication fidelity21. Crompton et al. used daisy chains, lines of people passing along a story, to measure how communicators retain and describe features of a scenario. Communication ability was measured using the proportion of details recalled, and rapport was measured as self-reported comfort with the teller and receiver of the story.
The equal effectiveness of communication between all-ASD and non-ASD groups indicates that autistic communication, rather than being a set of random perturbations of conventional communication, is consistent within a group of only autistic people. The study’s researchers explicitly endorse the double empathy problem of autism, which describes how autistic and non-autistic people have two different neurotypes that make cross-group communication difficult. These neurotypes entail differences in perspectives and preferences that affect an approach to social situations. There is a “double” problem because both groups have difficulty modeling the other, though we have seen with camouflaging that autistic people may be more successful in doing so due to overlying social and cultural incentives to blend in with the majority.
5.b Tentative description of an autistic communication framework
Outlining the exact differences in pragmatics for an autistic neurotype is difficult because empirical studies finding differences in pragmatic performance rarely interrogate deeper mechanisms. Aggregated personal accounts by masked autistic people provides some guidance, as these people are usually attuned to different social needs and have determined appropriate compensatory strategies.
Many autistic people learn to correct for being more warm and friendly, using habits like doing research on friends before small talk or asking people about their feelings and not talking about themselves. It is also common for autistic people to realize that they have hyper-fixations on specific topics and then restrict their discussion of these areas. Additionally, autistic people often mask as being more “humble” by pretending to not know answers to questions and not correcting people on facts22. Qualities shared by autistic people include having a deliberative processing style or processing the world “from the bottom up”, with less reliance on heuristics to make decisions23.
5.b.i Deviations from Gricean maxims
The traits suggest common preferences that may affect Gricean maxims. Preference for deliberation and truth over heuristics suggest that the boundary of “quality” information may be higher in autistic people. Difficulty with natural small talk suggests a higher standard for relevance. There also seems to be some movement of the boundary of relevance toward personal preference of topics, with masking autistic people realizing that what they find interesting is often not relevant for others. A penchant for deliberation could move the boundary on the maxim of quantity, with a tendency toward providing more information to resolve anticipated ambiguity. This difference would explain the differences between autistic and non-autistic children when resolving scalar implicatures, with the former preferring binary agree/disagree responses rather than partial agreement.
It seems that a cooperative principle governing autistic communication will incentivize precise language with a preference to derive the most practical rather than social or reputational benefit from conversations. Social or reputational benefit arises from sharing of friendly, materially irrelevant rapport, such as small talk, inquiries into personal life, etc., which often contains information with no expectation of retention.
5.b.ii Relevance theory in autistic communication
Outside of Gricean pragmatics, other theories may be able to account for pragmatic differences in autistic people. Relevance theory, a more formal theory of communication compared to Grice’s original outline, describes how expectations of relevance are enough to allow listeners to decode speaker meaning24.
Relevance is related to positive cognitive effect, or “a worthwhile difference to the individual’s representation of the world”, and the processing effort, or the amount of inference, perception, and memory required to understand the input. Wilson and Sperber propose that “Other things being equal, the greater the processing effort expended, the lower the relevance of the input to the individual at that time” (19, 251–252). The systemic cognitive differences in autistic people, such as less reliance on heuristics, could fit into a relevance theory framework that emphasizes processing effort.
5.c Burden of explanation for autistic pragmatics
Discussion of systemic differences in autistic pragmatics is also made difficult by unclear distinctions between autism and other social communication disorders. The most updated DSM-V, for example, made the move of aggregating many autistic disorders into a single spectrum diagnosis, while also separating out a separate diagnosis for Social (Pragmatic) Communication Disorder (SPCD).
SPCD is distinguished from ASD by a lack of repetitive and restricted behaviors (RRB), and would have previously fallen under some autism diagnosis prior to the DSM-V. Pushback against the ASD/SPCD division claims that the distinction between the two diagnoses is unclear, with some practitioners raising concerns about individuals with SPCD being denied ASD-specific support that would otherwise be helpful25. The line between ASD and SPCD will inform whether a full explanation of pragmatic differences in ASD should also be able to account for the cognitive differences that lead to ASD-characteristic RRB, or whether the differences in language network connections can largely be analyzed independently of these other behavioral manifestations.
For the former, a strength of the weak central coherence theory is that it partially explains why autistic people avoid stimuli like loud noises or crowd, as the inability to appropriately filter language can be related to a similar inability in non-verbal stimuli. If more evidence points to the latter hypothesis (that pragmatic differences are largely separate from other cognitive/behavioral differences), pragmatics may be better understood as a specific social network or cognitive adaptation that exists outside of cognitive processes like reasoning ability and information comprehension. This hypothesis, however, needs further support, possibly from cognitive science, to explain the apparent convergence in communication ability within groups of only autistic people.
6. Conclusions and future work
For verbal, IQ-matched autistic people, pragmatic deficit does not capture the full picture of communication divergence with non-autistic people. In many cases, simpler pragmatic applications and inferential tasks are consistent with non-autistic controls, suggesting reasonable comprehension of maxims and cooperativity. Cases of autistic people camouflaging, with some effort, with non-autistic society suggest not an inability, but an intrinsic preference against non-autistic communication. Additionally, communication performance within autistic groups suggests consistency in pragmatic differences.
Shared behavioral traits and decision-making of autistic people suggest a framework of communication that is more informative and precise, leading to different pragmatic principles and choices. Further work needs to be performed to investigate the experience of autistic people, especially masked autistic people, to delineate the specific pragmatic deviations. Answering this question may also raise considerations about how much of pragmatics can be derived from processes like rational decision-making versus a contribution from more specific social learning and practice.
Autism spectrum disorder (ASD), which is the current accepted diagnostic code in the DSM, will be referred to as autism as shorthand. This post will focus on verbally communicative autistic people whose intelligence is within normal range, often classified under the unofficial diagnosis of high-functioning autism (HFA). “Autistic” will refer to individuals in this part of the spectrum unless indicated otherwise. Although the term “Asperger’s” occurs in some of the cited studies, this diagnosis is discouraged by autism rights advocates and has fallen out of use.
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Association. Page 53.
Schaeken, W., Van Haeren, M., & Bambini, V. (2018). The Understanding of Scalar Implicatures in Children With Autism Spectrum Disorder: Dichotomized Responses to Violations of Informativeness. Frontiers in Psychology, 9. https://www.frontiersin.org/articles/10.3389/fpsyg.2018.01266.
Volden, J. (2017). Autism Spectrum Disorder. In L. Cummings (Ed.), Research in Clinical Pragmatics (pp. 59–83). Springer International Publishing. https://doi.org/10.1007/978-3-319-47489-2_3.
Milton, D. E. M. (2012). On the ontological status of autism: The ‘double empathy problem.’ Disability & Society, 27(6), 883–887. https://doi.org/10.1080/09687599.2012.
Grice, H. P. (1975). Logic and Conversation (pp. 41–58). Brill. https://doi.org/10.1163/9789004368811_003.
Colston, H. L., & Lee, S. Y. (2004). Gender Differences in Verbal Irony Use. Metaphor and Symbol, 19(4), 289–306. https://doi.org/10.1207/s15327868ms1904_3.
Volden, J. (2017). Autism Spectrum Disorder. In L. Cummings (Ed.), Research in Clinical Pragmatics (pp. 59–83). Springer International Publishing. https://doi.org/10.1007/978-3-319-47489-2_3.
Loukusa, S., & Moilanen, I. (2009). Pragmatic inference abilities in individuals with Asperger syndrome or high-functioning autism. A review. Research in Autism Spectrum Disorders, 3(4), 890–904. https://doi.org/10.1016/j.rasd.2009.05.002.
Schaeken, W., Van Haeren, M., & Bambini, V. (2018). The Understanding of Scalar Implicatures in Children With Autism Spectrum Disorder: Dichotomized Responses to Violations of Informativeness. Frontiers in Psychology, 9. https://www.frontiersin.org/articles/10.3389/fpsyg.2018.01266.
Pijnacker, J., Hagoort, P., Buitelaar, J., Teunisse, J.-P., & Geurts, B. (2009). Pragmatic Inferences in High-Functioning Adults with Autism and Asperger Syndrome. Journal of Autism and Developmental Disorders, 39(4), 607–618. https://doi.org/10.1007/s10803-008-0661-8.
Loukusa, S., & Moilanen, I. (2009). Pragmatic inference abilities in individuals with Asperger syndrome or high-functioning autism. (Same as reference 8).
López, B., & Leekam, S. R. (2003). Do children with autism fail to process information in context? Journal of Child Psychology and Psychiatry, 44(2), 285–300. https://doi.org/10.1111/1469-7610.00121.
Interpretation of some studies on autism, including this one, are complicated by a now-outdated distinction between high-functioning autism (HFA) and Asperger’s syndrome. Asperger’s syndrome was retired as a DSM diagnostic code in 2013 and folded under the broader category of autism spectrum disorder.
Hahn, N., Snedeker, J., & Rabagliati, H. (2015). Rapid Linguistic Ambiguity Resolution in Young Children with Autism Spectrum Disorder: Eye Tracking Evidence for the Limits of Weak Central Coherence. Autism Research, 8(6), 717–726. https://doi.org/10.1002/aur.1487.
Walęcka, M., Wojciechowska, K., & Wichniak, A. (2022). Central coherence in adults with a high-functioning autism spectrum disorder. In a search for a non-self-reporting screening tool. Applied Neuropsychology: Adult, 29(4), 677–683. https://doi.org/10.1080/23279095.2020.1804908.
Though masking is not limited to women, the behavior has been first and primarily studied in women. Masking appears to be more characteristic of autism in women, which may partially explain the apparent gender ratio of autism.
Bargiela, S., Steward, R., & Mandy, W. (2016). The Experiences of Late-diagnosed Women with Autism Spectrum Conditions: An Investigation of the Female Autism Phenotype. Journal of Autism and Developmental Disorders, 46(10), 3281–3294. https://doi.org/10.1007/s10803-016-2872-8.
Allely, C. S. (2018). Understanding and recognising the female phenotype of autism spectrum disorder and the “camouflage” hypothesis: A systematic PRISMA review. Advances in Autism, 5(1), 14–37. https://doi.org/10.1108/AIA-09-2018-0036.
Ibid.
Hull, L., Mandy, W., Lai, M.-C., Baron-Cohen, S., Allison, C., Smith, P., & Petrides, K. V. (2019). Development and Validation of the Camouflaging Autistic Traits Questionnaire (CAT-Q). Journal of Autism and Developmental Disorders, 49(3), 819–833. https://doi.org/10.1007/s10803-018-3792-6.
Crompton, C. J., Ropar, D., Evans-Williams, C. V., Flynn, E. G., & Fletcher-Watson, S. (2020). Autistic peer-to-peer information transfer is highly effective. Autism: The International Journal of Research and Practice, 24(7), 1704–1712. https://doi.org/10.1177/1362361320919286.
Price, D. (2022). Unmasking Autism: Discovering the New Faces of Neurodiversity. Harmony/Rodale. Page 42.
Ibid, page 14.
Wilson, D., & Sperber, D. (2002). Relevance Theory. Proceedings of the Tokyo Conference on Psycholinguistics, Tokyo, Japan.
Brukner-Wertman, Y., Laor, N., & Golan, O. (2016). Social (Pragmatic) Communication Disorder and Its Relation to the Autism Spectrum: Dilemmas Arising From the DSM-5 Classification. Journal of Autism and Developmental Disorders, 46(8), 2821–2829. https://doi.org/10.1007/s10803-016-2814-5.