When infants utter their first words around 12 months they have already learned many of the intricacies of their native language through exposure to speech. For more than 50 years, researchers have tried to identify the powerful learning mechanisms conspiring to achieve such tremendous phonological, lexical and syntactic achievements in such a short time. Some features of speech perception in infancy derive from the uniquely human predisposition for language, others from general learning tools present in other species. Disentangle human-specific from general processing skills is crucial to understand the origins of language. This project investigates one of those features: the ability to assign specific linguistic roles to different categories of sounds composing speech.

Indeed there are two speech sound categories in the worlds’ languages: consonants are short, discrete and typically more various category than vowels, which are usually longer, louder and continuous. According to the Consonant/Vowel (or C/V) functional asymmetry, a ‘division of labor’ is found in most languages as well as in speech processing in adults (and to some extent in infants) whereby consonants are more important for lexical information (the consonant bias, or C bias) and vowels for conveying structures, i.e prosody and syntax (the vowel bias, or V bias). These biases are particularly illustrated in Semitic, but also in non-Semitic languages, e.g. in the ablaut phenomena in Germanic languages: vowel variations in the same consonantal semantic root generate different grammatical variants, as in sing, sang, sung in English; singen, sang, gesungen in German.

Various hypotheses have been proposed to explain these biases. While some consider they could help infants discover the words and structures of their language very early on, potentially from birth , others have argued that they are a consequence of the acquisition of the structure of language, and are thus acquired too late to contribute to language acquisition . The question remains whether they are specifically human mechanism of language acquisition, or they develop because the asymmetry is present in the linguistic input? A related question follows: why have languages evolved to display this universal division of labor?

A review of the empirical data on the C and V biases in adults and infants led to the hypotheses that this project aims at testing: (1) the C/V asymmetry plays an essential role in infants’ quest for words and syntactic structures in the speech input; (2) it derives nonetheless from ancestral perceptual biases whereby some types of sounds are instrinsequely better targets for identifying word-forms, and for extracting and generalizing abstract rules from a stream of elements. We will therefore test the contribution of the C and V biases to early lexical and structural acquisition in infants, and their presence in a non-linguistic species, i.e. rats.

Date: 1 May 2016 - 30 April 2018