Re: Random |mormon| :-) (was: my proposals for a philosophical

John Cowan wrote:
>Andreas Johansson scripsit:
>
> > I meant, of course, that Tairezazh can't have non-initial /m/.
>
>Arrgh, and I thought "can't have non-final /m/" was such a k3wl konstraint,
>er, constraint.  AFAIK there are no natlangs that put more restrictions on
>initial consonants than on final ones.  A challenge!
>
> > Non-initial [m] occurs, but only as an allophone of /n/ occuring before
> > labial consonants. Eg _tshenp_ [tSEmp] "forest".
>
>Then perhaps it is better to say that there is only one final nasal,
>which is given the same place of articulation as the following consonant,
>or [n] if there is none.
That works beautifully if we use an andesque description with different
phonemic sets in different positions. But if we insist that this non-initial
nasal phoneme has to be identified with on of the contrasting initial pair
/m/ and /n/, it clearly must go with the latter, and the native (and
romanized) orthography assumes a such identification.

>  This is what happens in Italian, although for
>historical reasons both "n" and "m" are used in spelling.  But the
>nasal in "inferno", e.g., gets labiodental articulation.  Does this
>work across morpheme/word boundaries, as it does in Italian?
Always across morpheme boundaries, usually across word ones. Prefices turn
formerly initial m- to -n-, too.

                                             Andreas

_________________________________________________________________
MSN 8 with e-mail virus protection service: 2 months FREE*
http://join.msn.com/?page=features/virus