At 2:03 pm -0600 15/4/02, Dirk Elzinga wrote:
>>p = {pi, pu}   b = {pE, pO}
>>t = {ti, tu}   d = {dE, dO}
>Do you mean 'd = {tE, tO} ?

Most certainly.

>You're not including /a/?

I'd prefer to keep /a/, but I don't see a neat way round without also
having [1].

>Personally, I don't think *I* would have a problem with <h> = {ni,
>nu}. Alternatively, you could restrict the number of nasal-initial
>syllables so that <m> = {mi, mu} and <n> = {ni, nu}, and not have
>{mE, mO, nE, nO} at all. It does break up the symmetry of the system,
>but you don't have to resort to numerals or unintuitive mappings.

I know - but I would like to keep the symmetry if possible.

>But what if you compound a front stem with a back stem? If the
>morphemes are disemous, then there can't be harmony across a compound
>word boundary; each root will have to be marked separately for vowel

Yes, I have thought of:
-i- = 1st morpheme front & 2nd morpheme front
-e- = 1st morpheme front & 2nd morpheme back
-u- = 1st morpheme back &  2nd morpheme back
-o- = 1st morpheme back & 2nd morpheme front

That would keep all the vowel symbols, except (so far) {a}, as "cements",
but it would mean I have to rethink {je, jo} and {we, wo}.

>Hmmm. The suggestion I'm about to give will probably conflict with
>the 'brief' goal, but here it is. You can use {i,u} as prefixes to
>indicate vowel quality, {-} to separate suffixes from stems and {=}
>to separate clitics from hosts. {+} can be used to separate
>compounds, with each root of the compound given its own {i} or {u},
>although I would probably prefer to not have orthographic compounds.

I'm trying, at the moment, avoiding need for (much) use of the shift key so
I'm keeping away at present from {+}.

>The reason I see this as conficting with the goal of brevity is that
>every morpheme will be introduced by a character which is not
>pronounced. If morphemes are 3 or 4 letters, that's a 20% to 25%
>increase in text; you may find that unacceptable. On the other hand,
>a morpheme on its own (not affixed or cliticized) will still have to
>indicate vowel quality in some way.

True - and I had considered prefixing i- and u- if a free morpheme occurred
without any affixed functional morphemes.

>I also notice that you have no syllables containing /a/. If a
>morpheme has no harmony prefix, could it default to /a/?

But that would either mean that, e.g. _p_ and _b_ become homophonous /pa/
or _p_ = /p1/

>I haven't been keeping up on the Lin grammatical notes, so I'm
>probably suggesting things you've already decided not to consider.

Not at all - at this stage very little has been ruled out and nothing
definitely ruled in.