Voiced ejectives (was Simafira : Phonetics)

Thomas Weir wrote:
>  Jesse Bangs wrote:
> > I don't think voiced ejectives exist, and I'm pretty sure that
> > they're physically impossible.
They are - implosives (glottalic ingressives) may be voiced OR
voiceless, but ejectives (glottalic egressives) may only be voiced.
That said, some languages have reflexes of ejectives with a very
short voice onset time (the time between the release of the stop
and the start of the following vowel) which may sound similar to
voiceless unaspirated stops and in some cases the glottalic
mechanism has weakened to the point that it has been replaced by
creaky voice, so, although these stops may on historical or
comparative grounds be labelled ejectives, they are no longer
truly ejective. In other languages the delay of the onset of
voice is much greater (we're talking about milliseconds here).

This has been documented by J. C. Catford for the various languages
of the Caucasus - all possess ejectives, but in some languages
the ejectives have much briefer voice onset or have weakened to
creaky voiced stops (IIRC Kabardian falls into the latter group,
while the Adyghe dialect Bzhedugh has retained more fortis ejectives
with longer lag times - necessitated, perhaps, by the presence of
four series of stops in Bzhedugh: voiceless aspirated, voiceless
unaspirated, voiced, and ejective, while Kabardian has only three:
aspirated, voiced, ejective)

>  FWIW, England, in her _A Grammar of Mam, a Mayan Language_
>  writes (p. 26):
>
>  "The glottalized occlusives are unit phonemes which contrast with
>  plain occlusives plus glottal stop.  Phonetically, /tz' ch' tx' ky'
>  k'/ are ejectives, with glottal closure occuring simultaneously or
>  preceding the onset of oral closure.  /b' q'/ are implosives.
>  /t'/ varies between ejective and implosive, depending on the
speaker.
>  ejective and implosive, depending on the speaker.  The imploded
>  bilabial /b'/ is always voiced in initial or medial position, but
>  is devoiced finally; /t'/ is voiced initially and medially, and
>  devoiced finally; /q'/ is partially voiced initially and medially,
>  and also devoiced finally."
Either this refers to the creaky voiced stop that replaces ejectives,
or to an implosive. At least one language, Lendu, a Nilo-Saharan
language spoken in the Congo, is reputed to have a phonemic contrast
between voiced and voiceless implosives, as well as 'vowelless'
syllables (syllabic /z/ and trill /r/, which contrast between 'clear'
and 'dark' (produced with advanced tongue root vs neutral or retracted
tongue root, similar to the vowel systems in a number of African
languages).

The reason true voiced ejectives are not possible is that in an
ejective the glottis is closed during the production of the stop,
which prohibits the vibration necessary for voicing. I can't recall
why voicing is possible with implosives, but dimly recall that it
is connected with the ingressive nature of those segments.

A few Khoisan languages (Zhu'/oansi, !Xoo) have what are called voiced
ejectives, but these are actually prevoiced ejectives [bp', dt', gk']
or sequences of voiced stop plus (voiceless) ejective. The glottis
vibrates at the beginning, then closes to produce the ejective.

Regards,
Paul

Reply