|
What do infants know about permissible sound sequences in their native language?| title | What do infants know about permissible sound sequences in their native language? |
|---|
| start_date | 2023/07/04 |
|---|
| schedule | 14h |
|---|
| online | no |
|---|
| location_info | salle Ribot |
|---|
| summary | Languages differ in their phonotactics – that is, the position and sequencing of adjacent and non-adjacent segments within words. My goal is to identify what infants encode into representations to acquire phonotactics, when, and how they might do so. To answer these questions, I will integrate findings from a meta-analysis (n=~1400), with experimental results from adults (n=~300), infants (n=~160) and machines.
I will first show that 4-month-olds are sensitive to a typologically common harmony pattern where non-adjacent vowels in a word agree in backness, independent of language experience. Thus, the learning of some phonotactic patterns is privileged, because it is supported by infants’ initial perceptual abilities, consistent with Attunement theories of perceptual development (Aslin & Pisoni, 1980; Aslin et al., 2002).
Many of the sequencing restrictions on segments though are unique to a given language, and therefore must be learned from the input infants receive. I will show that sensitivity to adjacent segmental restrictions is induced by 5 months of age. Then using machine learning I will evaluate prelexical (e.g., Adriaans & Kager, 2010) versus lexical (e.g., Thiessen, Kronstein & Huffnagle, 2013) hypotheses about phonotactic acquisition. |
|---|
| responsibles | Sackur |
|---|
Workflow history| from state (1) | to state | comment | date |
|---|
| submitted | published | | 2023/06/19 13:58 UTC |
| |
|