Dalton Martha, Chasaide Ailbhe Ní
Phonetics Lab. CLCS, Trinity College Dublin, Dublin 2, Ireland.
Lang Speech. 2005;48(Pt 4):441-64. doi: 10.1177/00238309050480040501.
A comparison of the contour alignment of nuclear and initial prenuclear accents was carried out for the Irish dialects of Gaoth Dobhair in Ulster (GD-U) and Cois Fharraige in Connaught (CF-C). This was done across conditions where the number of unstressed syllables following the nuclear and preceding the initial prenuclear accents was varied from 2-0. This tests a variable peak hypothesis prompted by findings for other languages, that peak timing drifts as a function of the number of syllables preceding (the prenuclear) and following (the nuclear) accent. These data also test a second hypothesis that the L*+H dominant accent of GD-U might be viewed as being underlyingly the same as the dominant H* or H*+L accent of the CF-C dialect. According to this realignment hypothesis, the difference between these Ulster and Connaught dialects lies in the way that the melodic tier is aligned to the segmental tier: GD-U would be viewed as having a delayed realization of the peak relative to the Connaught dialect. Results do not support the variable peak hypothesis for Irish, as in either dialect, the peak appeared to be rather fixed across the three conditions examined (though not necessarily identical for prenuclear and nuclear positions). The results also militate against the realignment hypothesis, which rather than providing a more simple unifying account, would greatly complicate it. One reason is that there is a peak timing difference between the nuclear and prenuclear accents of CF-C, not mirrored in GD-U. Furthermore, even if one were to limit consideration to a single (e.g., prenuclear) context, a simple realignment of the accents in one dialect does not generate the appropriate contour in the other.
对阿尔斯特的高思多贝尔(GD-U)和康诺特的科伊斯法拉吉(CF-C)的爱尔兰方言进行了核重音和初始前核重音的轮廓对齐比较。这是在核重音之后和初始前核重音之前的非重读音节数量从2到0变化的条件下进行的。这检验了一个由其他语言的研究结果引发的可变峰值假设,即峰值时间会随着前核重音之前和核重音之后的音节数量而漂移。这些数据还检验了第二个假设,即GD-U的L*+H主导重音可能在本质上与CF-C方言的主导H或H+L重音相同。根据这种重新对齐假设,这些阿尔斯特和康诺特方言之间的差异在于旋律层与音段层的对齐方式:相对于康诺特方言,GD-U的峰值实现似乎有所延迟。结果不支持爱尔兰语的可变峰值假设,因为在任何一种方言中,在所研究的三种条件下峰值似乎都相当固定(尽管前核和核位置的峰值不一定相同)。结果也不利于重新对齐假设,该假设非但没有提供一个更简单的统一解释,反而会使其大大复杂化。一个原因是CF-C的核重音和前核重音之间存在峰值时间差异,而GD-U中没有反映出来。此外,即使只考虑单一(例如,前核)语境,一种方言中重音的简单重新对齐也不会在另一种方言中产生合适的轮廓。