Word Order Understanding Guides Wh-Question Comprehension

Friday, May 13, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
M. Jyotishi and L. Naigles, University of Connecticut, Storrs, CT

Wh-questions are challenging for children with ASD. Prior research has shown delays in both production and comprehension of questions like “What did the apple hit?” and “Who pushed the doggy?”  in ASD compared with TD groups (1, 2). Some researchers have argued that children with ASD have particular difficulties with complex grammatical structures (3) while others have proposed that the impairments are related to pragmatics (4). We investigate the former possibility by examining whether early grammatical competence predicts later wh-question comprehension. One index of early grammar is word order; children who successfully process wh-questions must have abstracted the subject-verb-object (SVO) frame.  Therefore, children’s ability to successfully process SVO sentences could help them learn wh-question structures. 


We investigated the degree to which comprehension of SVO word order at 2 years predicted wh-question comprehension at 3-4 years in a large sample of children. 

Methods:  Language comprehension of 35 TD children (MAvisit1 =20.20 months) and 29 children with ASD (MAvisit1=33.16 months) was assessed via intermodal preferential looking every four months for two years in this longitudinal study.  At visits 1-2, children viewed the word order video, in which they were shown side-by-side reversible actions involving a costumed horse and bird, paired with Baseline trials (“we can see both”) and test trials (“The horse is washing the bird” vs. “the bird is washing the horse”) (3,4).  At visits 3-6, children watched the Wh-Question video, in which each horse-and-bird action was followed by trials in which the horse and bird appeared side by side. The test audios were e.g., “Where is the bird/horse?” for Where/Baseline trials, “What washed the horse?” for Subject wh-questions and “What did the bird wash?” for Object wh-questions. Children’s eye movements were coded off-line. The dependent measures were difference scores of proportion looking to the target during the test minus baseline trials; larger differences indicated stronger comprehension. Word order scores were combined across visits 1 – 2; wh-question scores were analyzed for each visit. 


In the TD group, a significant correlation was obtained between early word order and subject-wh-question at visit 5, r = .359, p < .05. In the ASD group, a significant correlation was obtained between early word order comprehension and object-wh-comprehension at visit 6, r = .381, p < .05. Stepwise regressions in which NVIQ (Mullen VR) and vocabulary (CDI:Communicative Development Inventory) were entered first revealed that word order was the only significant predictor of subject-question comprehension for the TD group, F(1,30) = 4.43, p = .044 (R2 = .129). For the ASD group, NVIQ and early word order (but not vocabulary) jointly accounted for a significant amount of variance in object wh-question comprehension (R2= .272).


Early performance on SVO word order predicted later wh-question comprehension in both group; thus, children’s understanding of wh-questions seems to be emerging from their knowledge of English sentence frames. These findings lend support to the hypothesis that the sparse and delayed wh-question use of children with ASD has grammatical roots, perhaps in combination with pragmatic roots.