hey will, good idea, although the question is how do we work out what might have been a normal grk intonation unit for this type of genre (i.e. on what basis could we modify the punctuation of the text as presented by the modern editor).
i can think of three possibilities right now which could serve as a basis for this (none particularly convincing though):
(a) if an ancient grk style critic has described the relative lengths (or given syllable counts) of several connected cola, you could extrapolate from this an intonation unit to adopt if your e.g. text comes from the same genre (i am thinking in particular of the first few chapters of demetrius on style, where he does give syllable counts and general rules for cola length),
(b) a study of any punctuation in papyrus and inscriptions in a text from the same genre as your e.g. text could help, and
(c) if typology of languages shows that intonation units tend to have about X syllables in the context and genre of your e.g. text. by the way will, your typology e.g. above concerning pitch was v interesting: has someone studied this further in the context of latin pitch accent i wonder?
cheers 