Inferential language in high-function children with autism.
Even chatty kids with autism need step-by-step lessons on metaphors, hints, and irony.
01Research in Context
What this study did
Richman et al. (2001) watched 12 high-functioning kids with autism who spoke in full sentences.
They gave short stories and then asked what the character really meant or why they spoke.
The team scored answers for basic facts, metaphors, and hidden intentions.
What they found
Kids got every basic fact right, showing solid vocabulary and grammar.
They fell apart on metaphors like "heavy silence" and on hints like "It’s cold in here" meaning shut the window.
In short, words were fine, but reading between the lines was not.
How this fits with other research
Petit et al. (2025) later used tablet games and found trouble only with scalar implicatures (some but not all) while metaphors were now okay. Their finer tools update the broad 2001 claim.
Vierck et al. (2015) saw the opposite: Mandarin-speaking autistic kids passed scalar implicatures. The clash likely comes from language and task differences, not a true failure to replicate.
Song et al. (2024) added irony to the list, showing second-order theory of mind is the bottleneck. Together the papers map which inferences break and why.
Why it matters
Your fluent-speaking clients may sound mature yet miss social cues every day. Pinpoint the exact inference type before writing goals. Use visual scripts for metaphors, teach scalar cues with pictures, and train partners to speak literally when needed.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Test one metaphor and one hint in play: if wrong, model the meaning and rehearse three times.
02At a glance
03Original abstract
Despite average verbal intelligence, high-function children with autism have social comprehension deficits that are expressed by how they use and understand language. In this paper, we explored the general hypothesis that high-function children with autism make some, but not all, of the pragmatic inferences necessary for successful communication, even when they have the ability to perform noninferential language tasks. We contrasted the ability of 8 high-function children with autism (each with Verbal IQ > 70) and typically developing children to use and understand: pragmatic inferences about given or presupposed knowledge in mental state words; pragmatic inferences about new or implied knowledge in mental state words; bridging inferences essential for coherence; elaborative inferences involved in enriching a communication by means of figurative language; and the intentional inferences involved in speech acts. High-function children with autism could define words and identify multiple meanings for ambiguous words. In understanding words for mental states, they made inferences from mental state verbs to given or presupposed knowledge. However, they failed to infer what mental state verbs implied in context; to make inferences about social scripts; to understand metaphor; and to produce speech acts, all of which are inferences that are the basis of successful social communication because they elaborate meaning or convey intentions.
Journal of autism and developmental disorders, 2001 · doi:10.1023/a:1005661613288