Justifying a stress pattern

One project of mine that has never progressed much beyond the naming language
stage is Keshean (Kesheâras). One particularity it has is a system of stress
placement that seemed to make sense at the time but now strikes me as odd. Now,
certain real-world languages feature stress patterns that alo strike me as odd,
but can be explained, or at least compactly described, by moraic theory or the
like - eg. Latin, where stress falls on the second-to-last mora, ignoring the
final syllable. Perhaps someone can think of a similar compact description for
Kesheah stress. The noncompact description might be stated like this:

i) The stress goes on the last syllable if that contains a long vowel (or
diphthong) or ends in a consonant cluster.
ii) Failing that, the next long syllable to the left.
iii) If all nonfinal syllables are short, stress goes on the first syllable.

Some examples (colon marks long vowel, accent stress):
élshas
elshá:
elshaís
aréts
áreda
áredikas
stra:gá:nas
reáxtanas

('sh'=/S/, the rest more or less = IPA)

The oddity, of course, is that a long final syllable doesn't attract the stress
if it ends in a short vowel followed by a single consonant. The description
would be simplified if the final consonant, if any, of each word were ignored -
the stress rule would then be stress on the rightmost long syllable, or on the
leftmost in the absence of long syllables - but that seems very arbitrary.

Suggestions? Comments?

 Andreas

Replies