Nearly every piece of research on the phonology of Mixtec languages
makes reference to the couplet, a term used to describe the shape of
canonical roots, and to explain the domain of numerous phonological
processes. Despite its ubiquity, scholars disagree over whether the
couplet should be defined in prosodic terms as a bimoraic foot, or in
morphological terms as a root. In this paper, we describe and analyze
three patterns in San Martín Peras Mixtec that are commonly attributed
to the couplet in the Mixtec literature, namely the distribution of
laryngeals, tones, and nasality. We show that, in San Martín Peras
Mixtec, the foot is necessary to explain the distribution of laryngeals
and rising tones, while the root is needed to explain the domain of
phonotactic restrictions on nasality. We conclude that the Mixtec
couplet does not correspond to a single grammatical category across all
varieties, and that the couplet in San Martín Peras Mixtec is an
emergent category that arises out of the interaction of (at least) the
bimoraic foot and the morphological root.