Linguistic processing

Background

People with autism show a range of linguistic difficulties, across the entire IQ range. A common difficulty is with the use of sentence context in spoken language, and several studies have examined this experimentally. These studies have suggested that individuals with autism are unable to integrate linguistic information. A well-known study, for example, has shown that individuals with autism tend to read homographs with their common pronunciation (e.g. "a tear in a dress", rhyming with deer). However, there are other explanations for some of these results, besides an inability to integrate linguistic information. For example, in the case of the homographs, individuals with autism may find the common pronunciation so salient that they cannot inhibit it as they read the sentence out loud. Alternatively, people with autism may have difficulty accessing the rare pronunciation of the homograph, under the time pressures of the experiment.

On-line Word Monitoring

In order to test more directly the use of linguistic context, we have carried out a number of studies using a word monitoring procedure: where participants are asked to press a button as soon as they hear a pre-specified target word. By using this online task, we are able to tap into the automatic, obligatory processing of language, without giving the participants time to reflect upon their response.

Example sentences (Target = "PENCIL")

Normal prose

Semantically-anomalous prose

Random word order

In one study, three sentence forms were used, containing different amounts of sentence context. By comparing reaction times to the same targets across these three conditions, we are able to obtain a measure of participants' use of different forms of linguistic context. We found that, whilst children with autism and typical children showed similar word monitoring reaction times when targets were within a random jumble of words (demonstrating that the two groups performed the task in a similar way), the children with autism showed significantly less benefit from the inclusion of linguistic context. This directly demonstrates a difficulty processing linguistic context, rather than some other more general difficulty with performing the task.

Misleading Sentence Contexts

Example sentences (Target = "DOG")

Congruous sentence

Incongruous sentence

We also found a similar effect in a more subtle manipulation, this time involving word-monitoring for targets that occurred in either congruous or incongruous sentences. In this task, we would expect congruous contextual information to facilitate typical children's responses, but for their performance to be hindered when the target was preceded by misleading, incongruous contextual information. In contrast, if children with autism show reduced use of linguistic context, we would expect their reaction times to be similar in the two conditions. This was the result that we found: children with autism demonstrated significantly less use of sentence context than their developmentally-typical peers, and indeed showed a trend towards faster reaction times in the incongruous condition.

Further work

We are now planning a number of extensions to these studies. For example, we are developing a more tightly-constrained version of the latter experiment, which we will test in parallel with the children's visual embedded figures test. A direct comparison of analogous visual and linguistic tasks should prove useful in assessing whether the poor processing of sentence contexts is due to a linguistic deficit or a deficit in a central cognitive mechanism in autism.

Related reading

Jolliffe, T., & Baron-Cohen, S. (1999). A test of central coherence theory: linguistic processing in high-functioning adults with autism or Asperger syndrome: is local coherence impaired? Cognition, 71, pp.149-185.

Happé, F.G.E. (1999). Understanding assets and deficits in autism: Why success is more interesting than failure. The Psychologist, 12(11), pp.540-546.

Norbury, C.F., & 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. International Journal Of Language and Communication Disorders, 37(3), pp.227-251.

Happé, F.G.E. (1997). Central coherence and theory of mind in autism: Reading homographs in context. British Journal of Developmental Psychology, 15, pp.1-12.

Bell, S.J. (2004). The use of linguistic context by children with autism. Unpublished PhD thesis, University of Cambridge.

Bell, S.J., Davis, M.H., & Plaisted, K.C. (in preparation). Children with autism exhibit reduced use of semantic context in continuous speech.

Bell, S.J., Davis, M.H., & Plaisted, K.C. (in preparation). A linguistic analogy to the embedded figures test.

Witkin, H.A., Oltman, P.K., Raskin, E., & Kapp, S. (1971). A manual for the Embedded Figures Test. Consulting Psychologists Press, California.