Re: Sound changes - whither retroflex sounds and glottal stop?

Hey.

On 7/22/06, Patrick Littell <puchitao@...> wrote:
> The glottal stop can affect tone, if you want to introduce tone.  It
> can lead to either high or low, by different historical processes --
> I'll try to find some examples within a single language family if you
> want me to look it up.  I believe it will affect tone on the previous
> vowel, but I don't know the specifics.  (For example, did it need to
> be in the syllable coda to affect tone, or could it have been in the
> following syllable's onset?  I don't know.)
Goshute (a variety of Shoshoni) has nascent tone which developed from
a medial glottal stop. In Western Shoshoni, there are several pairs of
words which are distinguished solely by the presence/absence of /ʔ/:

[tsoʔapʰ] /tsoʔappɨh/  'ghost'
[tsoapʰ] /tsoappɨh/  'shoulder'

[siʔipʰ] /siʔippɨh/  'sheep'
[siːpʰ]  /siːppɨh/  'urine'

The Goshute cognate for 'ghost' is missing the glottal stop but has a
HL contour; 'shoulder' is the same as in Western Shoshoni (there is
also a "de-stridentization" of the affricate):

[tθóàpʰ]  'ghost'
[tθóápʰ]  'shoulder'

[sîːpʰ]  'sheep'
[síːpʰ]  'urine'

> Laryngealization of the vowel is another possibility.  I'm pretty sure
> the laryngealization contrast in the Totonacan languages is derived
> from an earlier alternation between V and V? (with the latter becoming
> laryngealized V).  I could look this up as well, if needed.
>
> Laryngealization or something like it can be a reasonable source for
> low tone later on, if you want it.  Straight from glottal stop to tone
> would most likely give you high, but if glottal stop goes to
> laryngealization, you'll get low instead.
This seems to be the likely route for the Goshute tonal contour; in
Western Shoshoni, the vowels surrounding the glottal stop can be
pretty creaky.

Dirk