Deictic but Not Conventional Gestures Predict Children's Vocabulary One Year Later

Background: Children refer to objects with their hands (e.g., point at cat) before they can produce verbal labels for these objects (“cat”; Bates et al., 1979). Importantly, the onset of such deictic gestures predicts the onset of similar spoken words in typically developing (TD) children, showing a strong positive relation between early deictic gestures and early words (Iverson & Goldin‐Meadow, 2005). Children with autism spectrum disorder (ASD) show difficulties in early gesture use, particularly in pointing (Mundy et al., 1986); they also often show prolonged delays in producing words and use fewer words compared to TD children.

Objectives: In this study, we ask how the delays and difficulties that we observe in the vocabulary development of children with ASD are related to gesture production. We hypothesized that early production of deictic gestures would be more affected than the production of other gesture types (e.g., conventional, iconic) and that variations in deictic gesture production but not other gesture types would be related to later language.

Methods: We tested this question by observing 23 18-month-old TD children and 23 30-month-old children with ASD—comparable to TD children in productive vocabulary, as they interacted for 20 minutes with their mothers in a semi-naturalistic observational protocol (Communication Play Protocol, CPP). We coded the types of gestures children produced; these included deictic gestures (e.g., pointing at an object to indicate an object), conventional gestures (e.g., nodding the head to mean ‘yes’, extending an open palm next to an object to indicate ‘give object’) and iconic gestures(e.g., moving hand forcefully to indicate ‘throwing’). Iconic gestures were extremely rare in our data; therefore we only focused on deictic and conventional gestures in our analysis. In addition, we assessed children’s spoken vocabulary—both tokens and types of words produced, during a second CPP performed one year later.

Results: We found that children with ASD showed significant deficits in their production of deictic, but not conventional gestures: compared to TD children, fewer children with ASD produced deictic gestures (56% vs. 100%, X²(1) = 12.78, p<.001), also producing them at significantly lower rates (Kruskall–Wallis, H(1)=12.49, p<.001). Importantly, the production of deictic gestures predicted the size of children’s vocabulary both for word tokens (r =.69, p<.01) and word types (r=.68, p<.01) one year later; but no such predictive association was found for conventional gestures and vocabulary size.

Conclusions: These results show that deictic gesture is a fundamental aspect of the language learning process in children with ASD—as it is in TD children, predicting children’s spoken language development. Our results further suggest that it is not gesturing per se, but the production of a particular gesture type, namely deictic gesture, that serves as a stepping-stone for subsequent vocabulary development. Children’s deictic gestures may play this important role by helping children establish a joint focus that the caregiver can then elaborate with language.