Proceedings of the Annual Meeting of the Cognitive Science Society

Previous research has demonstrated a relationship between quantity of language input and childrens rate of languagedevelopment: Children who hear more words learn faster. This work takes on two mutually-constraining questions:(1) How should we define quality, and (2) what is the relationship between input quality and language development?We analyzed a longitudinal corpus of interactions between 50 children and their parents using four measures of lexicaldiversity: Type Token Ratio (TTR), Moving Average TTR, and two more recent measuresvocd-D and MTLD. We foundthat only MTLD gave a prima-facie correct characterization of childrens development, and parents MTLD was correlatedwith childrens over development. Results of simulations showed that MTLD was distinct from the other measures in itssensitivity to both lexical diversity and word order, suggesting that quality should be defined not just by diversity of words,but also by the variability of sentence structures in which they occur.