| Abstract: | The phonetically grounded constraint against voiceless stops after nasals (*NC) has diverse and well-attested effects, but some counterexamples have been cited. Tswana reportedly bans not voiceless but voiced stops after nasals ([bata] 'look for,' [mpata] 'look for me'). A closer look at the phonology of Tswana and a phonetic study suggests that this description is not accurate. For some speakers, no stops are ever voiced, including post-nasal stops. For others, the consonants in question turn out to be not stops but voiced sonorants, which become stops after nasals. Phonologically, the distribution of Tswana obstruents is governed by the interaction of well-known and independently motivated constraints on voicing in obstruents, positional faithfulness to word-initial position, and a process of post-nasal fortition. The study shows that Tswana provides no evidence against *NC. |
