Harmony Triggering as a Segmental Property

Harmony is a process by which some property of a trigger segment spreads onto one or more undergoer segments. In some languages, any segment that bears a harmonizing property will trigger harmony, while in others some segments bearing that property will trigger harmony while others will not. This paper examines such cases in the nasal vowelconsonant harmony systems of several Malayo-Polynesian. Previous researchers have claimed that these languages have phonological inventories include both triggering and non-triggering nasal consonants. I propose that the patterns of nasal harmony triggering in these languages are best analyzed by encoding a segment’s status as a trigger of harmony within its subsegmental representation. This analysis is implemented within the Gestural Harmony Model (Smith 2016a, 2016b, 2017), in which the assumed unit of representation is a modified version of the dynamically-defined gesture of Articulatory Phonology (Browman & Goldstein 1986, 1989). Within this model, a phonological grammar may shape a language’s inventory to include both harmony-triggering and non-triggering segments. The result is a new perspective on how a phonological grammar generates complex patterns of harmony triggering, as well as how such patterns arise through diachronic sound change. The paper is organized as follows. Section 2 introduces the patterns of triggering of nasal harmony in the Malayo-Polynesian languages Acehnese and Rejang. Section 3 reviews how the triggering of harmony is represented in the Gestural Harmony Model and proposes an analysis of the harmony triggering patterns in both of these languages within that framework. Section 4 suggests possible historical reasons for such patterns of contrastive triggering to emerge. Section 5 concludes.