Robust Lexically Mediated Compensation for Coarticulation: Christmash Time Is Here Again.

Area of Science:

• Cognitive Science
• Auditory Perception
• Psycholinguistics

Background:

• Investigating the integration of high-level knowledge with sensory input in cognitive science.
• Distinguishing between direct top-down influences and postperceptual biases in speech perception.
• Lexically mediated compensation for coarticulation (LCfC) as a key test case for top-down influences in spoken word recognition.

Purpose of the Study:

• To provide strong evidence for top-down interaction in spoken word recognition.
• To address inconsistencies in previous LCfC findings.
• To validate computational models of speech recognition incorporating feedback mechanisms.

Main Methods:

• Extensive piloting of stimuli to ensure robust phoneme restoration and coarticulation compensation.
• Conducting a well-powered, preregistered experiment (N=40) using pretested stimuli.
• Performing a direct replication study (N=40) to confirm findings.

Main Results:

• Observed robust lexically mediated compensation for coarticulation (LCfC).
• Demonstrated reliable phoneme restoration and coarticulation compensation with pretested stimuli.
• Confirmed findings through a direct replication study.

Conclusions:

• LCfC effects are reliably observable with carefully selected and pretested stimuli.
• Results provide strong support for computational models of spoken word recognition that include top-down feedback.
• High-level lexical knowledge directly influences auditory perception.