This data collection contains data from three experiments looking at word learning in toddlers aged 16 to 23 months, supporting a published paper (Floccia et al. 2014). Each child was presented with 8 trials in which she has to learn two made-up words differing by one consonant or one vowel. All experiments with toddlers were conducted either using object manipulation tasks, or computer-controlled procedures with eye-trackers in both sites. The research found that English toddlers at 16 or 23 months did not show evidence of a consonant bias when tested in an interactive word learning situation, contrary to well-established results in French toddlers at the same age. It has been proposed that vowels and consonants play distinct roles in language, with vowels tending to carry intonation and grammar and consonants conveying the meaning of words (Nespor et al., 2003). Whilst this distinction appears relatively clearcut in experimental behaviour with adults its developmental origin is a matter of considerable conjecture, with far reaching implications to our understanding of language acquisition. Three main hypotheses have been suggested to account for this asymmetry. The first hypothesis posits that it is a property of the linguistic system, whilst the second indicates that the relative roles of consonants and vowels is driven by the disparities in their acoustic properties. In the final hypothesis the asymmetry arises from differences in the stochastic distribution of vowels and consonants for a particular language. In this project we will directly compare the processing of vowels and consonants between adult, toddler, and infant listeners of English and French. As there are considerable differences in the distribution and acoustic properties of vowel and consonants between these languages we will be able to judge the relative merits of the competing hypotheses by tracking the asymmetry at different stages of language acquisition.

Face to face interaction. Children are from families who volunteered to take part in Babylab research. They are invited for this study if they are monolingual and aged +/- 3 weeks around the target age. Each participant was allocated a participant number; demographic information is retained (gender and age at the moment of testing). Only the data of the children who contributed to 6 out of the 8 trials were retained for analysis. Detailed methods information is available in the published paper (see Related resources). Similar experiments were carried out with adults and are available in dataset 851381 (see Related resources).