Abstract
11 normally developing children of 20 to 24 months of age were taught a nonsense word under informative and increasingly routine conditions. Modelling was the teaching procedure. There were significantly more spontaneous productions of the new word in the informative condition than in the most routine condition for most of the children. This was also true for the familiar words presented. Significantly more language play in the form of repetitions was exhibited for the new word than with familiar words by 10 children, indicating that language play is a strategy for acquiring new words. Only 5 children used imitation of the model as an acquisition strategy. Modest support for the effects of informativeness is provided. Other motivations for word use are discussed as well as clinical implications of these results.