> Subject: constructs and vowel length
> From: "M. Jean Balcaen" <jbalcaen AT chass.utoronto.ca>
> Date: Wed, 13 Oct 1999 1:13:59
> As my thesis topic, I have chosen to work on the question of the
> realationship between peosody, syntax, and phonology (whew!).
> Actually, I'm starting with monosyllables in construct and/or
> prosodically bound (maqqef) to a following word. A very odd pattern
> has emerged - in lexical monsyllables from geminate roots, [e] and
> [o] types regularly undergo compensatory lengthening of the vowel
> due to degemination of the final radical, subject to a whole variety
> of morphological and prosodic conditions. However, [a] type
> monosyllables steadfastly follow the syntax - the vowel is short in
> constructs, long otherwise, regardless of the prosody.
Actually, I don't think that the s.ere in _h.es._ "arrow" (stem
_h.is.s.-_) or the h.olem in _h.oq_ (stem _h.uqq-_) are due to
compensatory lengthening at all, but rather due to normal lengthening
of underlying /e/ and /o/ in main-stressed syllables.
One clue to this is to look at the rare words in which an underlying
consonant cluster survives as such word-finally. Here _qosht._ and
_nerd_ have h.olem and s.ere vowels (while _'ard_ has pathah., except
in pause), but obviously no compensatory lengthening has taken place.