I've been arguing that the entire Masoretic notation relates to
liturgical chant and was never intended for speech (Richard Goerwitz
apparently expresses a similar opinion in his dissertation). In this
vein, we can now explain pausal elongation.
Penultimate vowel in pausal position is elongated to avoid
interrogative intonation. This is a well-known phenomenon to narrators
and opera singers. The simplest (and rather derided) way of canceling
interrogative intonation is reducing the ultimate vowel to near schwa.
The Masoretes adopted a better approach: they elongated penultimate
vowel so that the relative length of the final vowel is reduced
without making the vowel inaudible.
Incidentally, several singers told me that in their experience Hebrew
speakers are more prone to elongating final vowel, which turns the
phrase into a question. Whether this anecdotal evidence holds, I don't
know.
In interrogative sentences, pausal elongation would be unnecessary.