Morphophonology-based stress variability in Lebanese Arabic

This paper concerns the systematic stress variability in two Lebanese Arabic morphological paradigms. The variation in these paradigms – Measure VII and VIII – has been noted in the literature, but this study is the first to treat the issue with empirical rigor, demonstrating nuance in the variation that has not been discussed to date. To do so, this study utilised 17,657 MVII(I) tokens in the casual speech of 272 speakers from over 425 hours of speech in Lebanese podcasts. The resulting data shows strikingly symmetrical and complex morphophonology-based variation that would appear best described by probabilistic models of grammar. The variation was found to be intra-speaker (token-wise) such that the variability itself appears embedded in individual speakers’ grammars, expanding our conception of how stress governance can behave in Arabic. Methodologically, this paper demonstrates the promise held in the analysis of large-volume audio corpora to treat Arabic variation patterns.