Study title

This dataset consists of audio-recordings of naturalistic speech addressed to infants as well as their inventory of words when they were 9, 15 and 21 months of age. All infants were growing up in English-speaking families in Edinburgh, UK. The questionnaire was completed by a parent at each data collection point. The form is based on the UK Communicative Development Inventory (CDI) with slight modifications to include Scottish words (e.g., 'bairn', 'dinnae').

In infant-directed speech across languages and communities, many words are substituted with register-specific lexical items such as 'choo-choo' and 'tummy'. The main purpose of this project is to examine the hypothesis that such baby-talk words (BTWs) are functionally motivated by their typical phonological properties, including reduplication (eg, 'wee-wee', 'din-din'), lack of clusters (eg, 'tummy' cf. 'stomach') and recurrent ending patterns (eg, the /i/ ending in 'doggy', 'kitty'etc.). Words with these phonological characteristics are thought to be easier to detect and learn, offering an entry point to lexical learning. There are two major predictions that follow from this hypothesis: (1) infants should detect and learn novel words faster when they have phonological characteristics associated with BTWs than when they lack such properties; (2) because infants who are initially exposed to more learnable words are likely to break into lexical learning earlier than others, they should acquire more words during the initial stages of lexical development. These predictions will be tested through laboratory-based word learning experiments and longitudinal analysis of the relationship between BTW input and infants' naturalistic lexical development.