Statistical learning of dependencies in children with specific language impairment: An exploratory study

2019 IJRSLL – Volume 8 Issue 2

Author/s:

Veeramani, Priyadharshini*
All India Institute of Speech and Hearing, University of Mysore, India (jayceechan7@gmail.com)

Rathinaswamy, Rajasudhakar
All India Institute of Speech and Hearing, University of Mysore, India (rajasudhakar.aiish@gmail.com)

Abstract:

Children with SLI (CwSLI) are identified to have a significant delay in language (mostly in expressive language), with preserved non-verbal abilities. There are different views put forth to account for deficits among CwSLI, which are namely, cognitive-based (deficits in working memory and procedural memory), language-based (morpho-syntax deficits) and perception-based (perceptual acoustic deficit in listening). Studies have concluded that sentences processing is relatively slow in CwSLI, compared to their normal peers. The ability to track dependencies between the syntactic constituents in a sentence structure (example- subject-verb agreements), is vital for language acquisition and processing. This study aimed to explore the statistical learning of dependencies among CwSLI, through a sentence processing task. The participants of the study included a total of thirty native Kannada-speaking children (15 children with SLI and 15 typically developing children) in the age range of 7-13 years. 40 Kannada sentences including short and long sentences with adjacent and non-adjacent dependencies were used in the study. The task was computerized and the stimuli were delivered to the participants, using Psychopy software (version 1.83), through laptop. Children were instructed to do syntactic judgement of the sentences, by clicking (mouse-click) on to the appropriate icon (tick icon for syntactically correct sentence and wrong icon for a syntactically wrong sentence), which were displayed on screen. Accuracy was the measured dependent variable. Results revealed CwSLI performed poorly (i.e. more errors) on processing sentences with non-adjacent dependencies than in adjacent dependencies, when compared to their normal peers. This provides insights into statistical learning among CwSLI.

Keywords: specific language impairment; statistical learning; sentence processing; dependencies; syntactic judgement

PDF

DOI: https://doi.org/10.5861/ijrsll.2019.3015

*Corresponding Author