The phonology of classical Arabic meter

Abstract

We propose a phonologically well motivated theory of metrics that avoids several problems (e.g. ternarity and center-headedness) with the traditional analysis of Arabic metrics (al-Xalīl †c. 791 ; Maling 1973 ; Prince 1989). We propose that the content of a metrical position is universally restricted to three prosodically motivated units : L, H, LL and that binarity holds at the levels of the verse foot and metron. This constrains the number of possible verse feet to nine and leads to the insight that the traditional Arabic verse feet are in reality metra (pairs of verse feet). The different degrees of popularity of the Arabic meters (cf. corpora in Vadet 1955 ; Stoetzer 1986 ; Bauer 1992), we argue, can be understood as a direct function of rythmic well-formedness. The best meters are all iambic (Ewald 1825 ; Jacob 1967 [1897] ; Fleisch 1956), the rhythmic advantage being that they contain no rhythmic lapse (Kager 1993), an important constraint in Arabic phonology and morphology generally (Fleisch 1956 ; McCarthy and Prince 1990). Relative rhythmic well-formedness is formally expressible under a simple constraint-based analysis (cf. Prince and Smolensky 1993)