Print

Print


Alternatively, you could mark the semivowel version with e.g. a breve
(although there's no breve + l or r so maybe use an underdot for those
still) and leave the vowel versions unmarked: alĭo 'other', păter
'father' etc.

I must say that while you've done well to only use ascii values, it
makes the orthography a bit opaque. f v w x y c z are a bit odd.

On 6/5/14, Alex Fink <[log in to unmask]> wrote:
> On Thu, 29 May 2014 23:40:22 +0300, Rusiok <[log in to unmask]> wrote:
>
>>Yes, 9 letters (h, i, l, m, n, q, r, u, ') have got 2 values each: syllabic
>> and sonorant one.
>>They are vowel near obstruent letters (b, c, d, f, g, j, k, p, s, t, v, w,
>> x, y, z) and
>>are consonant near vowel one (a, e, o).
>
> You could analyse the system differently: instead of (for example) labelling
> one phoneme /e/ and another underlying /?_j/ with an allophone [e], you
> could call these both /e/ but with two different types of morphophonemic
> behaviour.  Adopting that point of view, one might find reasonable a
> spelling like <a e o> for your current _h ' q_ and <ä ë ö> for your current
> _a e o_; the diaeresis says "this vowel does not in fact turn into a
> sonorant consonant when next to another vowel", which indeed is basically
> the normal function of the diaeresis.  Perhaps you could even leave off the
> diaeresis on instances of your current _a e o_ which never come next to
> another vowel, since if I understand you right the distinction is invisible
> there.
>
> Alex
>


-- 
(This is my signature.)