4 - The typology of rounding harmony

Summary

Rounding harmony typology

Rounding harmony is a phonological process whereby certain vowels surface as rounded under the influence of a neighbouring rounded vowel. What is striking about rounding harmony is the fact that the simplest possible statement – ‘a vowel must be rounded when preceded by/followed by a rounded vowel’ – fails to characterise the great majority of rounding harmony systems. In most cases, conditions referring to tongue body position (height and/or backness) are imposed on either the triggering element, the target, or both. I argue that this interaction among vowel features renders traditional rule-based accounts of the typological patterns nonexplanatory. Within a constraint-based framework such as Optimality Theory (Prince and Smolensky 1993), however, the interaction of rounding with these other phonological dimensions can be modelled in a straightforward manner that allows for the characterisation of all attested rounding harmony patterns, while making falsifiable predictions regarding the logically possible but cross-linguistically unattested patterns. Central to my analysis is the claim that phonological systems are organised around principles of articulation and perception. These principles are encoded in the formal grammar as Optimality-Theoretic constraints.

The goals of the chapter are as follows: (1) to exemplify the range of attested rounding harmony patterns, (2) to identify the perceptual and articulatory principles that give rise to these patterns, and (3) to propose a formal model to characterise the role of these principles in grammar. In the final section, I outline the results of a recent experiment involving loanwords in Turkish. The experimental results indicate that the model of rounding harmony developed here, while motivated by evidence from typology, is also an appropriate model of individual grammars.