Do frequency and frequency-related measures signal turn completion? An exploratory corpus study

Area of Science:

• Psycholinguistics
• Computational Linguistics
• Sociolinguistics

Background:

• Conversational speakers utilize mental lexicon word frequency information.
• Word frequencies may function as turn-completion cues in dialogue.

Purpose of the Study:

• Investigate the role of word frequencies in conversational turn-taking.
• Determine if word frequency patterns predict turn transitions versus continuation.

Main Methods:

• Analysis of the Freiburg Multimodal Interaction Corpus (FreMIC).
• Comparison of word frequencies and related measures in question and storytelling turn-constructional units (TCUs).
• Application of mixed-effects models to analyze S-shaped frequency patterns and predict turn transitions (TT).

Main Results:

• Word frequencies in TCUs exhibit an S-shaped pattern, consistent with prior research.
• A more pronounced word frequency drop at the end of question TCUs predicts turn transition.
• Story TCUs show less pronounced frequency drops, predicting turn continuation.
• Turn-final question TCUs display a right-leaning distribution of nouns.

Conclusions:

• Word frequency patterns serve as crucial cues for listeners to anticipate turn completion or continuation in conversation.
• The study links turn-transition dynamics to social actions, specifically differentiating between questions and storytelling.