[tied] Re: the glottalic theory

--- In cybalist@..., Miguel Carrasquer <mcv@...> wrote:

> On Thu, 14 Nov 2002 15:05:43 -0000, "Richard Wordingham"
> <richard.wordingham@...> wrote:
>
> >--- In cybalist@..., Jens Elmegaard Rasmussen <jer@...> wrote:
> I was trying to establish a link between tones and consonantal
> laryngeal settings, which can be arranged in the sequence:
>
> /h/ ~ aspirated --> voiceless --> breathy voice (murmur) --> slack
> voice --> modal voice --> stiff voice --> creaky voice

(laryngealized)

> --> /?/
>
> The frequency goes from low to high as the vocal chords go from far
> apart (/h/) to close together (/?/) [at least for the voiced part of
> the sequence: breathy --> creaky], so a link with low/high tone

would

> be natural.

I suspect we may need to be careful about mixing consonant phonations
and prosodic features. Thurgood's article (
<http://www.csuchico.edu/~gt18/Papers/Vietnamese&tonogenesis.pdf> -
drat the '&'!) makes the point that it's the overall voicing
(generally register) rather than the initial consonsnts themselves
that matter. In terms of the initial consonants, the sequence has
been postulated by Hyman and Schuh as, low to high:

breathy, voiced obstruent, sonant, voiceless unaspirated, voiceless
aspirated, implosive.

Note that the two aspirations are at either end! This is universal
1253 at http://ling.uni-konstanz.de:591/Universals , a massive
harvest of assertions. (A lot of the assertions have counter-
examples although none are yet noted against them at the site.)

There is a significant risk that the association of tone and
consonant is *not* 2-way. Greenberg proposed (no. 766 at the
Universals' site), 'A voiced injective stop ... has an effect
identical with or more similar to that of ordinary breathy or voiced
consonants, i.e. it does not lower tone'. I assume 'injective'
= 'implosive'.

> We have cases of tone becoming laryngeal setting [glottal stop]

(e.g.

> Latvian),

at syllable boundaries!

> and of laryngeal setting becoming tone (e.g. Panjabi).

and many other languages!

Why do you want to start from tones? A register contrast, creaky
versus normal, might be ideal for your purpose.

> >Incidentally, is having /?t/ and /?k/ but not /?p/ plausible?

With

> >voiced pre-glottalised plosives, the natural gap seems to be to
> >lack /?g/ rather than /?b/, which does not suit us at all.
>
> In the labials, the gap (if any) is usually in the voiceless sphere
> ([ph], [p'], [p] missing), while in the velars/uvulars it's usually

in

> the voiced sphere ([G], [g] missing). I see no reason to think it
> would different for pre-glottalized plosives.

Looking at data from New Guinea and the Solomons, I get the
impression that with voiced pre-nasalised consonants, it's the
labials that go missing. As an extreme example, Rennellese has [B],
[l]/[d], but [Ng]. Several languages show [nd] and [Ng] in their
Swadesh wordlists, but no [mb]. No-one contradicted me when I
remarked on this pattern on the Austronesian list.

Richard.