Phonological Variation in Child-Directed Speech is Modulated by Lexical Frequency

Area of Science:

• Linguistics
• Developmental Psychology
• Phonetics

Background:

• Child-directed speech (CDS) is hypothesized to simplify phonology for infant learning.
• Korean noun stem-final obstruent variation provides a specific phonological context to test this hypothesis.

Purpose of the Study:

• To investigate if child-directed speech (CDS) exhibits a higher proportion of canonical pronunciations than adult-directed speech (ADS).
• To determine if lexical frequency mediates the rate of canonical pronunciations in CDS.

Main Methods:

• A word-teaching task comparing mothers' speech to infants versus adults.
• A follow-up experiment examining adults' pronunciation of high- versus low-frequency words.
• Analysis of high-frequency monosyllabic words to control for lexical frequency.

Main Results:

• Mothers used canonical pronunciations more frequently when speaking to infants than to adults.
• Adults produced canonical pronunciations more often for high-frequency words.
• No evidence of simplified phonology in CDS when lexical frequency was controlled.

Conclusions:

• The higher rate of canonical forms in CDS is mediated by word usage frequency.
• The apparent didactic function of CDS phonology may be a byproduct of using familiar words.
• Word frequency is crucial for understanding the nature of CDS.