Re: [tied] Abstractness (Was Re: [j] v. [i])

From: Piotr Gasiorowski
Message: 22707
Date: 2003-06-05

----- Original Message -----
From: Glen Gordon
To: cybalist@yahoogroups.com
Sent: Thursday, June 05, 2003 5:15 PM
Subject: Re: [tied] Abstractness (Was Re: [j] v. [i])


>> A phonologically plausible scenario could be as follows: an original
(pre-PIE) triangular inventory, /i, a, u/, developed into a square
inventory, /i, E, O, u/

> Hmm, I don't like that theory. For me to explain why there is *k versus
*q, I need to have one distinctively low vowel set apart from the rest. Is
this a common vowel system? Is it more common
than a centralized system of /I, &, a/?

The most widespread vowel systems consist of two parallel sets of high and
(where relevant) mid vowels with a [+/- front] contrast plus a single low
vowel (which is usually central and unrounded). A system like /i, a, u, &/
is also typologically plausible. Both vertical and qudrangular systems (the
latter with an equal number of [+front] and [-front] phonemes) are
relatively rare, though well-attested.

A three-term system may evolve into a five-term by splitting /i/ into /i, e/
and /u/ into /u, o/. This happens frequently, but doesn't seem to have
happened in pre-PIE, since it would have produced ablaut patterns like *i/*e
and *u/*o, rather than *e/*o. It follows that the split of old *a into
phonemes that eventually yielded *e and *o must have happened at some point.
It's hard to tell without discussing concrete proposals whether *a is of
later origin, or whether the whole non-high triad arose at once (in which
case PIE *a would represent the original quality surviving in appropriate
contexts, or whether there was first a split of /a/ into /a, &/ (the latter
[-low, -high, -front, -round]) and a later split of /&/ into /e/ and /o/.
All these developments look possible.

Piotr