Competing targets in English sibilant imitation.

Area of Science:

• Phonetics and Speech Science
• Acoustic Phonetics
• Speech Perception and Production

Background:

• Speech imitation is crucial for language acquisition and maintenance.
• Understanding imitation targets (normalized vs. raw acoustics) informs the perception-production link.
• Previous research has explored acoustic convergence but targets remain debated.

Purpose of the Study:

• To investigate whether speakers imitate English /s/ by converging to normalized or raw acoustic targets.
• To determine the role of acoustic properties in phonetic imitation.
• To explore the influence of baseline acoustic values on imitation direction.

Main Methods:

• Participants were exposed to altered spectral mean (SM) of English /s/ produced by a model talker.
• Acoustic analysis measured participants' SM shifts after exposure.
• Participant baseline SM was recorded to assess individual differences.

Main Results:

• Exposure to increased SM led participants to raise their own SM, matching the model's raw acoustics and the pattern of increase.
• Exposure to decreased SM resulted in shifts dependent on the participant's baseline SM.
• All participants consistently converged to the model talker's raw acoustic values.

Conclusions:

• Phonetic imitation can target raw acoustic values directly, not solely mediated by perceptual normalization.
• Speech imitation mechanisms may involve direct mapping of acoustic signals.
• Findings have implications for theories of the perception-production link and analysis of speech convergence.