Japanese children with high-functioning autism spectrum disorder (HFASD) have relatively more difficulty responding to maternal Wh-questions (Wh-Qs) than yes/no questions (Y/N-Qs) when compared to typically developing (TD) children (Oi, 2010). A similar difficulty was found in non-echolalic English-speaking children with autism who are mildly retarded (Curicio & Paccia, 1987). Japanese Wh-Qs are different from English Wh-Qs in grammatical construction. Movement of Wh, inversion of verb-subject order, and use of substitutive or auxiliary verbs are not seen in Japanese. For Y/N-Qs, the pragmatics is different between the two languages (Tsuchihasi, 1983). Despite these differences, the relatively greater difficulty responding to Wh-Qs than Y/N-Qs can be hypothesized to be observed across languages.
We tested this hypothesis in Taiwanese children. Taiwanese is different from both Japanese and English in terms of having A-not-A questions (ANA-Qs). Taiwanese is also different from Japanese in terms of the pragmatics of Y/N-Qs (Chao, 1968). We asked whether these differences influenced the greater difficulty of response to Wh-Qs compared to Y/N-Qs.
We compared the response to maternal questions in conversations collected under a semi-structured setting in 12 children (CA 7-15 years, mean 10.5, sd 2.76) with HFASD with 12 TD children matched by age, gender, IQ, and MLU. The response data were coded by type of maternal question, type of response, and meshing of question and response by adopting a slightly modified version of the coding schema devised by Bishop et al. (2000). Maternal questions consisted of Wh-Qs (mean+/-sd 38.9+/- 8.68 for TD, 42.8+/- 7.63 for HFASD), Y/N-Q (mean+/-sd 24.6+/- 7.89 for TD, 29.1+/- 4.01 for HFASD), Ch-Qs (mean+/-sd 25.5+/- 6.17 for TD, 24.5+/- 4.96 for HFASD) and ANA-Qs (mean+/- sd 25.9+/- 9.41 for TD, 26.6+/- 2.64 for HFASD). A graduate student who was unaware of the objectives coded the data. Inter-rater reliability was determined by using a second coder who was also a graduate student. The Kappa coefficients were .84 for type of question, .86 for type of response, and .89 for meshing.
Taiwanese children with HFASD produced significantly higher proportions of inadequate and pragmatically inappropriate responses, to Wh-Qs and Y/N-Qs than the TD children. HFASD children didn’t show more relative difficulty with Wh-Qs than to Y/N-Qs when compared to TD children. No difference was seen for Ch-Qs and ANA-Qs between the two groups.
Similar to the Japanese study, Taiwanese children with HFASD showed difficulty in responding to maternal Wh-Qs when compared to TD children. Y/N-Qs were, however, also difficult for them. In contrast, the HFASD children responded to Ch-Qs and ANA-Qs as adequately as TD children. Taiwanese Ch-Qs and ANA-Qs might have functioned like Y/N-Qs in Japanese and English; both types expect the listener to choose one of two answers. The relative difficulty responding to Taiwanese Y/N-Qs could be explained by the tendency for these questions to be asked when the probability of the listener’s intention to do the actions was less than 50%. These results merit further investigation into why children with autism have difficulty responding to Wh-Qs across languages.
See more of: Cognition and Behavior
See more of: Symptoms, Diagnosis & Phenotype