Comprehension of Non-Canonical Sentences in Mandarin-Speaking Children at risk for Developmental Language Disorders

• In Event: 3-037 - Syntactic Development in Mandarin-Speaking Children

Integrative Statement

Difficulty comprehending spoken sentences is a hallmark of developmental language disorder (DLD), a condition that affects 7.4% of the general population (Tomblin et al., 1997). Compared to sentences with canonical subject-verb-object (SVO) order, sentences with noncanonical word orders (e.g., passive sentences) cause greater comprehension difficulties among English-speaking children with DLD than their typical-developing (TD) peers (Dick et al., 2004). In addition, the group with DLD failed to show age-related improvement in comprehending noncanonical sentences (Montgomery et al., 2017). This difficulty was attributed to the DLD group’s deficits in utilizing morphosyntactic cues to properly assign thematic roles to noun arguments (Marinis & van der Lely, 2007). In this study, we compared comprehension of canonical and noncanonical sentences among Mandarin-speaking preschoolers at risk (AR) for DLD and their TD peers. We asked two questions: 1) Do AR and TD groups differ in comprehending canonical (SVO), noncanonical ba- (S-ba-O-V), and passive bei-sentences (O-bei-S-V)? 2) Do AR and TD groups show improved comprehension over time?
Twenty-one AR children were selected from a cohort of 202 children who participated in a larger project on Mandarin language development. At-risk status was defined as below-average performance on a combination of vocabulary, sentence repetition, and teacher ratings of children’s language ability. Specifically, to be included at AR, children must have two scores that were below -1SD or one score that was below -1.5SD for their age group. A picture-sentence matching task was administered twice with a 6-month interval. In this task, children listened to semantically reversible sentences and chose one out of 4 pictures that matched the sentence. The stimuli depicted fantasy scenes (e.g., The chicken is hugging the frog) with both animate agents and patients. The distractors depicted the reverse action (e.g., The frog is hugging the chicken), a changed agent (e.g., The horse is hugging the frog), or a changed patient (e.g., The chicken is hugging the butterfly). Each sentence type contained 6 trials; the 18 trials were randomized.
A 2 (group) x 3 (sentence type) x 2 (time) repeated-measures ANOVA was conducted. There was a main effect of group with the AR children performing significantly worse than the TD children, F(1,40)=12.25, p=.001. A sentence type effect (F(2,80)=26.63, p<.001) was detected: children were less accurate in comprehending bei-sentences than both SVO and ba-sentences (ps<.001), but accuracy for ba- and SVO sentences was not significantly different (p=.08). A time effect was also detected, F(1,40)=12.86, p=.001. Children performed better at Time 2 than at Time 1. There were no significant interactions between these variables (ps >.20) (see Figure 1).
Even though ba-sentences are noncanonical and rare in the input (Ji et al., 2018), the SOV word order allows children to use the “first noun as agent” strategy (Evans & MacWhinney, 1999) and does not impose additional comprehension loads compared to SVO sentences. The challenges posed by bei-sentences are consistent with the syntactic account of language acquisition: the long-distance movement involved in the passive bei-construction (Huang, 1999) made it harder for children to make thematic role assignments.

Authors

• Li Sheng, University of Delaware
  Presenting Author

• Yue Ji, University of Delaware
  Non-Presenting Author

• Danyang Wang, University of Delaware
  Non-Presenting Author

• Ying Hao, University of Texas, Austing
  Non-Presenting Author

• Li Zheng, Nanjing Normal University
  Non-Presenting Author