On Thu, 14 Aug 2003 19:24:18 +0000, Richard Wordingham
<richard.wordingham@...> wrote:

>This should then apply to any proper subset of human phones, e.g.
>nasals or clicks.

Well, yes. Apart from the effects of sound changes in and out of the
subset (d -> n, n -> d, etc.).

The case of the clicks is worth special mention. It was (is) widely
believed that clicks are relics of a very early ("primitive") stage in the
evolution of human speech. See J.C. Catford "The Myth of the Primordial
Click" in Hegedu"s, Michalove, Manaster-Ramer "Festschrift for V.V.
Shevoroshkin". This would imply clicks are a "source" (they can be lost,
but never gained, except, as in the cases of Zulu and Xhosa, through
borrowing). Catford dismisses this idea, and presents at least one example
of a pulmonic source for clicks:

"A formerly well-known example is the notorious Breton bilabial click. This
is an adventitious momentary (and presumably velaric) ingressive click
occurring at the transition from m to k in <toram kwat> "let's cut wood".
A kymogram in Rousselot (p. 493) clearly shows momentary ingressive airflow
between the -m and the kw-. This was referred to by Vendryes (1923, p. 39)
and, of course, was seized on by van Ginneken and linked with the Welsh ll
as "survivances des clics" (1938, p. 18)".

>> To be sure, epenthesis is not the main mechanism (which I think is
>> word composition, operating at the lexical/semantic level, above the
>> phonological level).
>Word-initial Germanic *p and German /p/ come to mind as problems
>with the logic here. They weren't established by sound changes.

Another occasional non-phonological mechanism is borrowing.

Miguel Carrasquer Vidal