[Histling-l] Asymmetry in voiceless stop lenition

[jess tauber]

I don't know if this is relevant, but in Yahgan obstruents are not treated equally by articulatory position.

The system of the Yahgashaga dialect in the mid-19th century consisted of p, b, t, d, ch, j, and k, g.

In word final position or before another obstruent-initial syllable all these were fricativized (after dropping following vowel): p/b>f, t/d>hr, ch/j>rri (?), k/g>x. There is also r>sh.

However, there was a 'flux' of these forms in initial position, between t and d, ch and j, and k and g.
But p and b did NOT alternate- rather p alternated with f. There is some evidence that b went with w. It is possible that at some more abstract level r/sh fit this latter pattern, but there aren't any initial r's in the language.

Jess Tauber
phonosemantics at earthlink.net