Abstract
I have considered hypermeters from various points of view. In this paper I introduce the viewpoint of elision and investigate the differences between group A and B in Man'yoshu which produce hypermeters. I exclude elision in waka where certain problems of recitation may be involved. The method adopted in this paper is to compare the hypermeters in Man'yoshu with elision in prose. Both group A and B seem to have been used in reciting waka. The hypermeters in group B are only realized by meeting the following condition: they have the same linguistic forms or approximately the same ones as those observed in elision in prose; group A are basically unrelated to elision and all the relevant linguistic forms show hypermeters. Elision in prose is one of the phonological phenomena based on everyday language. Through extensive research into these phenomena, I demonstrate the following two key points: 1) the hypermeters in group B can be characterized as one of phonological phenomena, 2) the hypermeters in group A are located as a phenomenon produced by a certain reciting practice which is different from the one adopted in everyday language and exists as a distinct pattern.
