Binarity in Prosodic Morphology and Elsewhere

Abstract

This paper explores the role of binarity in prosodic morphology by proposing that all representations are maximally binary branching, as stated in (1).

(1) Binarity Hypothesis: All representations are maximally binary branching.

Our evidence comes from examining patterns in which fission (Integrity violations) and fusion (Uniformity violations) of segments satisfies morphological and phonological constraints: multiple reduplication, haplology, coalescence, and breaking. Where there appears to be 1:3 or a 3:1 mapping between input and output segments, we propose that this must arise from two separate 1:2 or 2:1 mappings (perhaps at a stem and word level). We illustrate that a number of seemingly complex patterns of multiple reduplication in Salish, Wakashan and Uto-Aztecan languages follow naturally from the Binarity Hypothesis.