| Summary: | A number of authors have examined consonant alternations in Inuit (e.g. Thalbitzer (1904), Ulving (1953), Kaplan (1982), (1985)) and the related phenomena of fortition and lenition in Yupik. Voiceless stops and voiced continuants regularly alternate in Inuit dialects.While most of these alternations, such as those between /p/ and /v/, could be explained as the spread or insertion of the features [VOICE] and [CONTINUANT], there exist some puzzling alternations, such as that between /t/ and /j/ in Eastern Inuit. This alternation appears to be more complicated than the mere addition of voicing and frication, which we would expect to yield [ð]. Why is there a change of place of articulation? It would be favourable if this /t/~/j/ alternation could be explained as part of the larger pattern of stop-fricative alternations. However, it is difficult to imagine a single phonological rule that would subsume the /t/~/j/ alternation.I propose that feature underspecification can be used to unify these alternations,both within individual dialects and across all dialects. In particular, I propose thatDresher (2002)’s Successive Division Algorithm can be used to create contrastive feature hierarchies to assign underspecified contrastive feature specifications to the phonemes in these dialects. Furthermore, I will argue that for all dialects these alternations center on the contrastive feature [CONTINUANT]; the feature [CONTINUANT] appears to be the underlying contrastive feature responsible for the (voiceless) stop – (voiced) continuant alternations across all dialects. While other (non-contrastive) approaches would need to specify different rules for the different alternations, particularly the /t/~/j/ alternation in Eastern dialects (and between /t/ and the other reflexes of Proto-Eskimo *ð in other dialects), my proposal can unify the alternations; stops alternate with their contrastively [+CONTINUANT] counterparts across all dialects. Conversely, I argue that in all dialects /t/alternates with the least contrastively specified continuant (i.e. a continuant unmarked for place of articulation). Languages and dialects differ in terms of how they implement these underspecified representations.
|
|---|