Is there a (socio) pragmatic language impairment in autism, or only a core language impairment?

 Cross-posted at

This piece, the latest in my series of posts on Morton Gernsbacher’s FC-friendly take on autism, critiques a 2012 article by Gernsbacher and Pripas-Kapit entitled “Who’s Missing the Point? A Commentary on Claims that Autistic Persons Have a Specific Deficit in Figurative Language.” This article takes issue with the common understanding that autism involves pragmatic language deficits.

Curiously, however, Gernsbacher and Pripas-Kapit’s primary claim is solidly non-controversial: If autistic individuals have difficulty comprehending language, they’ll also have difficulty comprehending metaphoric, idiomatic, inferential, potentially ambiguous, or otherwise complex language. This includes individuals with Asperger’s, who often have quite sophisticated vocabularies but still struggle with comprehension. And it holds, as well, for non-autistic individuals with language delays.

Gernsbacher & Pripas-Kapit cite several studies to back this up. First they cite Norbury (2005), who only implicates language delay in difficulty with metaphors, as opposed to diagnosis or ability to pass basic TOM tests. The single most predictive aspects of language, Norbury finds, is performance on the Test of Word Knowledge, which goes beyond the typical vocabulary tests to assess awareness of words with similar meanings, understanding of idiomatic phrases, ability to define words, and ability to state multiple meanings of homonyms and homophones.

Regarding idioms in particular, Gernsbacher & Pripas-Kapit also cite Norbury (2004). Norbury found that language skills—rather than Theory of Mind, autism diagnosis, or Weak Central Coherence—are the biggest determiner in whether subjects are able to deduce the meanings of unfamiliar idioms from context. Among language skills, the biggest factor was sentence processing. Norbury proposes that “[t]o remember and use contextual information efficiently, children must have a good linguistic understanding of the context.”

Regarding homonyms and homophones, Gernsbacher & Pripas-Kapit cite a remark by Giora, Gazal, Goldstein, Fein, & Stringaris (2012, same volume as theirs): namely, what makes a difference here is comprehension skills, not autistic traits.

But inferences, in contrast to idioms and ambiguous terms, are more problematic. Here Gernsbacher & Pripas-Kapit’s source is Norbury and Bishop, 2002. While Norbury and Bishop do find that language comprehension ability affects the ability to draw inferences from stories, they also found that individuals with autism had more difficulty with inferences than other groups of study participants, and that those difficulties were correlated with autistic behaviors as measured by the ADOS. (The ADOS, or Autism Diagnostic Observation Schedule, is a robust measure of core autism traits, particularly social deficits). Furthermore, the correlation between autistic traits and inferencing skills held despite the fact that most of the study’s inferencing questions did not require references to emotional states.

Norbury and Bishop note, by way of explanation, that the inferencing questions required participants to “integrate pieces of linguistic information within the text and with his/her own general knowledge.” They found, moreover, that most of the inferencing errors did not amount to a failure to draw inferences, but a drawing of incorrect inferences: inferences that might be correct outside the context of the story, and that might have reflected a difficulty on the part of autistic participants in suppressing irrelevant information from their personal lives. 

Gernsbacher & Pripas-Kapit then cite a paper by Young et al, which, they claim,

controls for language comprehension ability and reports no differences between autistic and typically developing children on a wide range of story comprehension processes, including drawing inferences. (94).

But that is not what Young et al. say. Rather, these authors find that, when asked questions about a story:

children with ASDs recalled the factual information stated in the story as well as the controls, but they had difficulty with the higher level language skills needed to make inferences and answer questions about concepts that were not specifically stated in the story. (70)

Young et al. explain that:

The inferential questions required the child to demonstrate insight into the reactions and the mental states of actors in the story. For example… questions such as, “Why did the boy smile when he heard the croaking sound?” and “Why was the frog sitting proudly with a mother frog? (70)

These differences emerged despite the fact that “[t]he participants in this investigation were matched on receptive language according to the CELF-3.” This, Young et al. note, “suggests that it was the pragmatic nature of the tasks, not comprehension weaknesses, that accounted for the significantly poorer performance of the children with ASDs.” Young et al. note that “most standardized language tests focus on semantic knowledge at the level of single sentences rather than requiring comprehension of a paragraph or short story.” But it is paragraph and story-level comprehension that especially tap into pragmatic skills in general, and inferencing skills in particular—such that a language asessment that includes paragraph and story comprehension is inevitably not just a measure of core language skills, but also a measure of pragmatics and inferencing.

Gernsbacher & Pripas-Kapit go on to discuss the connection between core language skills and the ability to make evaluative comments about mental and emotional states. Citing Norbury and Bishop’s (2003) study, they state that “core language abilities rather than pragmatic skills or diagnostic status” drive the ability “to provide evaluative comments, especially about mental or emotional states.” But that is not what the study finds. Instead, the authors, while finding “robust relationships between syntactic complexity and frames of  mind  references  in  both  the  control  and  clinical  groups” caution that one should not “overstate the case as these two measures are not entirely independent  because  of  the  tendency  for  mental  state  terms  to  occur  in syntactically complex constructions.”

Gernsbacher & Pripas-Kapit  go on to fault Ozonoff & Miller (1996) for only taking account of vocabulary, not comprehension, when discussing difficulty drawing inferences, understanding indirect requests, and understanding humorous endings. They point out, correctly, that vocabulary is a flawed proxy for comprehension. And they also point out that the same is true with verbal IQ—many autistics score high on verbal IQ yet struggle with comprehension.

The correlation between comprehension and figurative language is, in many ways, neither surprising, nor exclusive of the core socio-cognitive factors of autism.

First, it is well established that, especially in the early years of life, social behaviors like Joint Attention predict subsequent language learning (see here). Thus, language in general, and comprehension in particular, is necessarily correlated with Joint Attention behaviors.

Second, comprehension in particular, when it goes beyond vocabulary (as Gernsbacher & Pripas-Kapit proposes it should) to larger stretches of discourse, is naturally going to encompass elements of pragmatics—for example, inferencing. Even when we are not responding to a request to explicitly draw an inference, we make inferences all the time, and fill in all sorts of gaps, when we comprehend language, whether as listeners or as readers.

Third, when comprehension goes beyond individual words to utterances or tracts of discourse, it inevitably includes Theory of Mind skills like perspective taking (what is the speaker/writer trying to convey) and, often, making sense of conversations, dialogues, and social scenes.

Fourth, discourse comprehension also entails background knowledge about the world—and here, too, the social deficits of autism may play a role. Specifically, the reduced Joint Attention skills, or reduced tuning into other people’s statements or conversations, which are sources for a great deal of implicit learning about the world, reduces opportunities for acquiring the background knowledge that is essential to making sense of extended discourse. In particular, as Norbury (2005) argues, worldly background knowledge is crucial to making sense of many metaphors, evocative as they often are of real-world properties that aren’t spelled out in dictionary definitions (consider, for example, “lawyers are sharks”).

Finally, and relatedly, limitations in comprehension can also reduce opportunities to learn about the world, such that limitations in comprehension are to some extent self-fulfilling.

These many connections between comprehension skills and socio-cognitive skills are consistent with the general consensus among autism researchers about what underlies the widespread pragmatic difficulties in autism: namely, that autistic people have deficits in Theory of Mind skills. But Gernsbacher & Pripas-Kapit disagree. They cite Gernsbacher and Frymiare, (2005), who (as I discussed in an earlier post) blame failure on TOM tasks on linguistic deficits alone. As I noted in my critique, however, while language plays a role in TOM tasks, so does automatic, intuitive perspective taking, and automatic, intuitive perspective taking is a core component of Theory of Mind.

Gernsbacher & Pripas-Kapit also take issue with Happe’s (1993) assessments of figurative language comprehension, which, they point out, presuppose a great deal of general language comprehension. Here they are back on firmer ground.

Gernsbacher & Pripas-Kapit then turn to claims that autistic individuals lack empathy, and my second post on this paper, I will focus on what they say here.


Gernsbacher, M. A., & Pripas-Kapit, S. R. (2012). Who’s Missing the Point? A Commentary on Claims that Autistic Persons Have a Specific Deficit in Figurative Language Comprehension. Metaphor and symbol27(1), 93–105.

Giora, R., Gazal, O., Goldstein, I., Fein, O., & Stringaris, A. (2012). Salience and context: Interpretation of metaphorical and literal language by young adults diagnosed with Asperger’s syndrome. Metaphor and Symbol, 27(1), 22–54.

Norbury, C. F. 2005b. The relationship between theory of mind and metaphor: Evidence from children with language impairment and autistic spectrum disorder. British Journal of Developmental Psychology, 23: 383–399.

Norbury, C. F. and Bishop, D. V. M. 2002. “Inferential processing and story recall in children with communication problems: A comparison of specific language impairment, pragmatic language impairment and high-functioning autism”. In The poems of John Godfrey Saxe Vol. 37, 227–251.

Norbury, C. F., & Bishop, D. V. (2003). Narrative skills of children with communication impairments. International journal of language & communication disorders38(3), 287–313.

Ozonoff, S., & Miller, J. N. (1996). An exploration of right-hemisphere contributions to the pragmatic impairments of autism. Brain and language52(3), 411–434.

Young, E. C., Diehl, J. J., Morris, D., Hyman, S. L., & Bennetto, L. (2005). The use of two language tests to identify pragmatic language problems in children with autism spectrum disorders. Language, speech, and hearing services in schools36(1), 62–72.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s