Phonotactic acquisition in healthy preterm infants

Abstract

Previous work has shown that preterm infants are at higher risk for cognitive/language delays than full-term infants. Recent studies, focusing on prosody (i.e. rhythm, intonation), have suggested that prosodic perception development in preterms is indexed by maturational rather than postnatal/listening age. However, because prosody is heard in-utero, and preterms thus lose significant amounts of prenatal prosodic experience, both their maturation level and their prosodic experience (listening age) are shorter than that of full-terms for the same postnatal age. This confound does not apply to the acquisition of phonetics/phonotactics (i.e. identity and order of consonants/vowels), given that consonant differences in particular are only perceived after birth, which could lead to a different developmental pattern. Accordingly, we explore the possibility that consonant-based phonotactic perception develops according to listening age. Healthy French-learning full-term and preterm infants were tested on the perception of consonant sequences in a behavioral paradigm. The pattern of development for full-term infants revealed that 7-month-olds look equally at labial-coronal (i.e. /pat/) compared to coronal-labial sequences (i.e. /tap/), but that 10-month-olds prefer the labial-coronal sequences that are more frequent in the French lexicon. Preterm 10-month-olds (having 10 months of phonetic listening experience but 7 months of maturational age) behaved as full-term 10-month-olds. These results establish that preterm developmental timing for consonant-based phonotactic acquisition is based on listening age (experience with input). This questions the interpretation of previous results on prosodic acquisition in terms of maturational constraints, and raises the possibility that different constraints apply to the acquisition of different phonological subcomponents.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 88 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.