This study examined the regional diversity of Standard Japanese (SJ) focusing on the accentual variation in nagara-clause. The results of our analysis utilizing the Corpus of Spontaneous Japanese indicated that a common principle underlies the variation in SJs spoken in the metropolitan area and its surrounding areas: the frequency of the variation, in which a pitch fall emerges anew, increases in response to the length of verbs preceding nagara. In contrast, no such correspondence is observed for SJs in the region of accentless dialects and Western Japan, but a positive preference for neutralization characterizes the latter. As for the Keihanshin area, the variation frequency is remarkably lower, suggesting that speakers refer to the rules of their native dialects in parallel with utilizing those in SJ.

Although Japanese foot is known as a rhythmic unit in the literature of metrics, spontaneous Japanese speech is considered to be mora-timed and the perceptual property of foot as a timing unit is unknown. The present study investigated whether Japanese foot is of psychological reality by using an identification task of moraic rests between odd-numbered or even-numbered mora nonwords. The results showed that the absolute threshold of moraic rest was shorter if nonwords were pentamoraic (five morae) than tetramoraic (four morae), suggesting that odd-numbered mora words were easier to be perceived with moraic rest because of the structural stability of foot. This offers behavioral evidence for foot as a perceptual timing unit in spontaneous Japanese speech perception.