Characterizing first and second language rhythm in English using spectral coherence between temporal envelope and...

Area of Science:

• Linguistics
• Phonetics
• Speech Science

Background:

• Speech rhythm is crucial for language perception and production.
• MacNeilage's frame/content theory provides a framework for understanding speech rhythm.
• Quantifying rhythmic differences between first and second language speakers is essential for linguistic research.

Purpose of the Study:

• To investigate and characterize the rhythmic differences in English speech between native American English speakers and native Mandarin speakers.
• To apply MacNeilage's frame/content theory to analyze speech rhythm.
• To evaluate the effectiveness of spectral coherence measures in differentiating language groups.

Main Methods:

• Collected speech data from 19 native American English speakers and 19 native Mandarin speakers.
• Computed spectral coherence between temporal envelope and mouth kinematics to operationalize rhythmic frame.
• Calculated spectral centroid, spread, rolloff, flatness, and entropy of the coherence.
• Utilized a binary logistic regression model for group differentiation.

Main Results:

• Spectral coherence measures effectively characterized rhythmic differences between native and non-native English speakers (A' = 0.71).
• Native English speakers exhibited significantly higher spectral centroid and spread compared to Mandarin speakers.
• Native English speakers showed significantly lower spectral flatness and entropy than Mandarin speakers.

Conclusions:

• The study successfully identified distinct acoustic correlates of speech rhythm differentiating native English and Mandarin speakers.
• Spectral coherence analysis offers a viable method for quantifying speech rhythm in linguistic studies.
• Findings contribute to understanding cross-linguistic phonetic variations and second language acquisition.