Infants' acquisition of word segmentation

Words are a central building block of language, and segmenting words out of continuous speech is a challenging step in early language acquisition. Indeed, in spoken language there is no acoustic equivalent to the spaces that separate words in written language: spoken words are not separated by pauses. Infants start developing word segmentation abilities during the first year of life, relying on their sensitivity to both phonological and syllable co-occurrence cues to word boundaries.

One phonological cue is vowel harmony: In many languages (e.g. Turkish,  Hungarian, Somali), vowels within a word must share some articulatory property, for instance they must all be pronounced either with rounded or with spread lips. Adults speaking a language with vowel harmony are known to rely on this phenomenon as a word boundary cue (Suomi et al.1997). More recent research has shown that 7-month-old English-learning infants are also sensitive to vowel harmony as a cue to word boundaries, despite the fact that English has no vowel harmony process (Mintz et al. 2018). We will extend this research by pitting vowel harmony against consonant harmony.

Whereas vowel harmony is extremely widespread, consonant harmony is cross-linguistically rare. Yet, there is much evidence that consonants are more important than vowels in lexical processing. In statistical word segmentation paradigms, for instance, adults who speak a language without harmony preferentially track transitional probabilities across  consonants, not vowels (Bonatti et al. 2005). We will test 7-month-old French-learning infants on both vowel and consonant harmony, to examine which of the two is preferentially used as a cue for word segmentation.

Overall, this study will shed light on the question of infants’sensitivity to phonological regularities regarding vowels and consonants in word segmentation, and of potential differences between adults’ and infants’ processing mechanisms.

Depending on the duration of the internship, the student will be involved in several or all stages of the project, from experimental design and preparation of the stimuli and script, to data gathering (in the ENS babylab) and analysis.

The student should have an interest in language acquisition but a background in (psycho-)linguistics is not required. A working knowledge of French is necessary to interact with infants and their parents.

• Bonatti, L., M. Pena, M. Nespor & J. Mehler (2005). Linguistic constraints on statistical computation. Psychological Science 16, 451-459.
• Mintz , T., R. Walker, A. Welday & C. Kidd (2018) Infants' sensitivity to vowel harmony and its role in segmenting speech. Cognition 171, 95-107.
• Suomi, K., J. McQueen & A. Cutler (1997). Vowel harmony and speech segmentation in Finnish. Journal of Memory and Language 36, 422–444.
• Thiessen, E. & J. Saffran (2003). When cues collide: Use of stress and statistical cues to word boundaries by 7- to 9-month-old infants. Developmental Psychology 39, 706–716.