Interaction of Knowledge Sources in Spoken Word Identification

Area of Science:

• Psycholinguistics
• Auditory Perception
• Speech Processing

Background:

• Spoken word identification relies on acoustic-phonetic cues and contextual information.
• Understanding how these factors interact is crucial for explaining speech perception.

Purpose of the Study:

• To investigate the interplay between acoustic-phonetic information and sentence context in spoken word identification.
• To determine the minimal acoustic-phonetic information needed for word recognition in different contexts.

Main Methods:

• Utilized a gating technique to present varying durations of spoken word segments.
• Manipulated the location (initial vs. final) and availability of acoustic-phonetic information within meaningful and anomalous sentences.
• Analyzed word identification accuracy and error patterns.

Main Results:

• Listeners could identify words with either initial or final acoustic-phonetic information alone.
• Less initial information was required for word identification in meaningful sentences compared to final information.
• Error analysis revealed an interaction between acoustic-phonetic cues and syntactic context.

Conclusions:

• Spoken word identification in sentences is an interactive process utilizing multiple knowledge sources.
• Meaningful sentence context significantly enhances word recognition, particularly through initial word information.
• Acoustic-phonetic information at the beginning of words is highly effective in facilitating rapid identification within coherent sentences.