Dan Seriff wrote:

All very interesting and coherent.

> Imperative, example root t-r-zc [t-r-D] - "telling, narrating":
> The imperative is an interesting form, due to its shape (taCCC-), with
> all three root consonants in proximity. Many bizzare sound changes take
> place in this form (as is about to be demonstrated with /r/).
>         tatrzcaa [tAttDA:]                      tell! (masc. sg. subject)
>         tatrzc÷ [tAttDo]                        tell! (fem. sg. subject)
>         tatrzc÷zxa [tAttDoZA]                   tell! (pl. subject)
> This shows one of /r/'s many guises:
>         /C[stop]r/ -> [C:] / _C[any]
>         i.e., when between a preceding stop and any other consonant, /r/
> disappears, but not before geminating the stop

The lengthening here is very normal when there is loss of some
kind.   However, geminate consonants are basically defined by
a break in syllables lying between them.  That means that (if I
understand your data aright) syllable breakdown would have to
be something like

             tatrzcaa [tAttDA:]:            tAt.tDA:
             tatrzc÷ [tAttDo]:               tAt.tDo
             tatrzc÷zxa [tAttDoZA]:     tAt.tDo.ZA

The only thing slightly weird I find about this is that it seems to
require that you have a voiceless onset stop clustered with a voiced
fricative ([tD]) -- something that, if not impossible, I'd bet would
be very rare in the phonologies of the world, because it's basically
an affricate without agreement in voicing.  You could solve that
easily in a number of ways, e.g. by devoicing the [D] in such

That's really just a nitpick, though.

Thomas Wier | AIM: trwier

"Aspidi men Sai˘n tis agalletai, hŕn para thamn˘i
  entos am˘mŕton kallipon ouk ethel˘n;
autos d' exephugon thanatou telos: aspis ekeinŕ
  erret˘; exautŕs ktŕsomai ou kaki˘" - Arkhilokhos