A special geminate like ff in maffa ‘pillow’, which does not exist in Standard Japanese (SJ) is found in the Antoh dialect in Fukui Prefecture. This paper deals with the synchronic and diachronic phenomena concerning the ff geminate in addition to bb and ss. This paper has the following purposes:

(1) to illustrate the correspondence of geminates between the Antoh dialect and SJ,

(2) to investigate the historical development of the geminates in the Antoh dialect,

(3) to point out that the manifestation of the geminates in the Antoh dialect is similar to that of the Miyako-jima dialect in the Ryukyuan language, and

(4) to propose that the explanation for the processes of the geminate in this dialect gives a suggestive source to the discussion on the processes of sound changes in the Ryukyuan language.

In this paper we test the hypothesis that Ikema, a dialect of Miyako Ryukyuan, has a three-pattern accent system, where three accent classes, Types A, B, and C, are lexically distinguished, in contrast with previous studies which have claimed that it has a two-pattern accent system. The results of our analysis confirm the existence of three distinct accent classes. The three-way distinction can only be observed in quite restricted conditions, including when nouns followed by one or more bimoraic particles precede a predicate. The results also reveal that Type A words are few in number, indicating that Type A words are in the process of merging with Type B.

To elucidate the neural basis of stuttering, brain activation for reading words was compared between adults who do and do not stutter (AWS/ANS) with functional MRI. Japanese native speakers read aloud familiar (F), unfamiliar (U) and pseudo- (P) words of 4 or 5 syllables. P contained much fewer native syllable sequences than F or U. F primarily activated the left angular/supramarginal gyri (lAG/SMG), U Broca's area, and P the left ventral premotor/motor areas (lvPMA/MA), respectively, in ANS. AWS showed lower activation in lAG/SMG and Broca's area, but higher activation in lvPMA/MA, implying that AWS cannot read native syllable sequences as efficiently as ANS.

The place of articulation in Japanese moraic-nasals is known to vary in accordance with the following sound and is assumed to be uvular in word final position. This paper describes the articulatory variability of word-final moraic-nasals based on the X-ray microbeam speech production database in Japanese. The results depict substantive inter-speaker variability in three of the four words examined. Moreover, 75% of the data examined were deemed unlikely to be uvular nasals and the lips were likely to be closed in 40% of the data. The results do not support the claim that word-final moraic-nasals are uvular.