On Wed, 6 May 2009 11:59:31 -0700, steve rice <[log in to unmask]> wrote: >My a priori project Pesu uses phoneme segregation to isolate morphemes: since /s/ belongs to a class that begins the final syllable (and /p/ does not), the boundaries are clear regardless of stress. And -n marks a bound morpheme, clarifying the word boundaries for compounds. It's not foolproof, of course--slurred pronunciation can still lead to incorrect resolution--but it makes it easy to speak distinctly, which should be good enough. Do you have a web site for Pesu? Google found nothing, and the List archive returned just snippets. Do I guess it right that Pesu morphemes are of the form <class I consonant>V(<class M consonant>V)(<class F consonant>V)? My project has very few polysyllabic morphemes so far, but if they grow too many I'll probably redesign everything and pick a way to segregate the second syllable: a few reserved syllables or a few reserved consonants or one reserved vowel (the "e" vowel of the 4-vowel system I talked about some time ago). How does your -n mark work? Is it head-final, <complement>n<head>? But even with long and self-segregating morphemes, frequent compounding will generate ambiguity. If not stress/pitch/contour, maybe parenthesis operator morphemes could help, they function very similarly. I seem to recall lojban and a few other langs have them.