Chilcotin flattening and autosegmental phonology

Abstract Pharyngealization is the quintessential characteristic of Chilcotin phonology, which is called ‘Chilcotin flattening’. This process consists basically of two rules, one that flattens (i.e., pharyngealizes) vowels, which is triggered by underlying flat consonants (i.e., one set of postdental sibilants, including affricates, and one set of postvelar obstruents), and another that harmonizes flat and nonflat sibilants (excluding neutral sibilants). Although both rules are based on the tongue root feature ([+/- R(etracted)T(ongue)R(oot]), Vowel Flattening (VF) spreads [+RTR] to vowels, while Sibilant Harmony (SH) spreads either [+RTR] or [-RTR] to nonneutral sibilants depending on the relative position of the trigger and target. While the autosegmental treatment of Chilcotin flattening illuminates the tongue root harmony in both rules and their interaction, it presents interesting problems that are difficult to deal with. Aside from the difficulty of defining the domain for Vowel Flattening, the theoretical problems include one that is concerned with association line crossing and another with the role of ‘opaque segments’ (a la Clements and Sezer 1982). Chilcotin flattening provides evidence that the opaque segments are not uniformly ‘spreaders’, ‘nonundergoers’, or ‘blockers’ and that they play different roles in different contexts.