The Concurrent Association Between Responsive Maternal Language and Children's Expressive Language Skills Is Moderated by the Specificity of the Mothers' Utterances

Background:   Previous research with typically developing children has shown that labeling objects in the child’s current focus of attention benefits vocabulary acquisition (Shimpi and Huttenlocher, 2007). Similarly, longitudinal research in autism has shown that undemanding maternal language is associated with children's subsequent language gains (Siller and Sigman, 2002, 2008).

Objectives:   This study aims to investigate whether the concurrent relationship between responsive maternal language and the language skills of children with autism is moderated by the presence or absence of specific object labels.

Methods:   The sample included 68 children with autism ages 32-82 months (M=57.19 months, SD=12.4 months), with expressive language IQs ranging from 5 - 81.8 (M=29.7, SD=18.5). Children’s language skills were assessed using the Mullen Scales of Early Learning (MSEL). Expressive language IQ was calculated by dividing Expressive Language Age by chronological age and multiplying by 100. To collect maternal language samples, mother-child dyads were videotaped for ten minutes with a toy set and given instructions to play as they normally would. A two-minute window of the play interaction was coded for responsiveness of maternal utterances using The Observer® (NOLDUS). The coding system is designed to evaluate two aspects of maternal speech: Demanding vs. Undemanding and General vs. Specific. Demanding utterances are those that aim to elicit a change of behavior from the child, either through correction or overt redirection; Undemanding utterances consist of those that mom uses to describe her own behavior, reinforce or comment upon the child’s behavior. In addition, maternal utterances were coded as Specific or General. Specific utterances are those that use nouns to label tangible objects; General utterances use pronouns in the place of noun labels. Utterances were coded as Specific if at least one item was labeled with a noun. Three independent coders established strong inter-observer reliability for all measures (ICC range =.73 to .87).

Results: While demanding specific utterances have no significant association with expressive language (MSEL), r=.047 (p>.05), a significant negative association was found between demanding general utterances and expressive language (MSEL), r= -.401 (p<.01). Conversely, undemanding general utterances have no significant association with expressive language (MSEL), r=.080 (p>.05), but a significant positive association was found between undemanding specific utterances and expressive language (MSEL), r=.354 (p<.01).

Conclusions:   These findings suggest that, for children with autism, the association between responsive maternal language and children’s expressive language is moderated by the presence or absence of specific object labels. The direction of the relationship between specific object labeling and children’s expressive language is unclear; whether children’s expressive language skills influence maternal language use, or the other way around, should be the subject of future longitudinal studies.