Objectives: This study investigates how age, gender, and regional background influence speech rate, articulation rate, and disfluency patterns in elderly Korean speakers using large-scale spontaneous speech data. It explores how these sociolinguistic and physiological factors interact to shape speech tempo and seeks to establish empirical baselines for age-related changes in Korean elderly speech.Methods: The analysis was conducted on spontaneous speech data from 2,941 female and male speakers aged 65 and above across four regions of Korea. A total of 136,009 utterances (452,865 sentences, 6,726,921 Eojeols), amounting to 1,113.45 hours of speech, were examined. Speech rate, articulation rate, and various types of disfluencies including pauses were derived from annotated corpus data and statistically analyzed using linear mixed-effects models.Results: Individual variability accounted for approximately 43–45% of total variance, yet significant demographic and regional effects remained. Age-related slowing was primarily driven by increased pausing and disfluencies, while articulation rate remained stable across elderly groups. Males articulated faster but paused more, resulting in slower overall speech rates, whereas females maintained smoother continuity and faster overall rates. Regionally, Chungcheong speakers showed the slowest speech rate; Jeolla, the fastest articulation; Gyeongsang, rapid speech through reduced pausing; and Gangwon, slow articulation with minimal disfluency.Conclusion: Speech tempo in elderly Korean speech is shaped by the interaction between articulation speed and disfluency frequency rather than by motor decline alone. Age, gender, and regional identity reflect distinct temporal strategies balancing physiological capacity, cognitive control, and sociolinguistic norms. These findings underscore the value of large-scale spontaneous speech analysis for understanding temporal variation in language use in later life.
Choi et al. (Tue,) studied this question.