Abstract

1. Introduction

This commentary outlines a typology of final laryngeal neutralization under the assumptions of ‘laryngeal realism’, an approach to feature representation which distinguishes structurally the two-way contrasts of ‘voice’ languages (Dutch, Polish, Spanish) from those of ‘aspiration’ languages (German, Somali, Washo). The typology in turn opens the door to a fresh test for what has emerged as a central debate about Evolutionary Phonology (EP), namely, the status of ‘unnatural’ developments, in particular, the possible occurrence of ‘final voicing’ patterns. While EP allows for the existence of such systems, Kiparsky (this volume) argues vigorously that the proposed cases should all be reanalyzed, and that final voicing is actually unattested because it is impossible in human language.