An enhanced autosegmental-metrical theory (AM+) facilitates phonetically transparent prosodic annotation

Abstract

The application of autosegmental theory to a variety of tonal systems in the past 40 years has spurred a wealth of insight. The application of this theory to intonation languages spawned a family of transcription approaches known as Tones and Break Indices, or ToBI. This general theoretic framework, informed by metrical stress theory, has come to be known as autosegmental-metrical (AM) theory. Yet, a number of weaknesses and limitations have been noted, both at the theoretical and empirical levels. We argue that core limitations in the traditional AM theoretic approach can be traced clearly to a failure to consistently and transparently encode syntagmatic relationships in phonology. Building on the core insights of traditional AM theory, and drawing on empirical evidence about cognitive representations for pitch from phonetics, music cognition, music theory, and cognitive neuroscience, we propose a new theoretic approach, termed enhanced AM theory, or AM+. This proposal offers a theoretical clarification of syntagmatic elements in phonology. It is shown that attributing both syntagmatic and paradigmatic properties to tones provides a unifying account of multiple outstanding challenges in tone and intonation research that have not yet found a satisfactory explanation.