Re: [tied] Texts in PIE

----- Original Message -----
From: "Glen Gordon" <glengordon01@...>
To: <cybalist@yahoogroups.com>
Sent: Saturday, April 19, 2003 12:36 PM
Subject: Re: [tied] Texts in PIE


> We've already talked about the reinterpretation of the satem isogloss: First uvular *q and *kW merge to *q, then both *k and *q are fronted respectively to *k^ (> *s^) and *k in order to reestablish natural patterns of markedness.

This is falsified by two Satem branches: Albanian (where *[q] and *[kW] developed differently before front vowels, so that e.g. *[qi] > *[ki] > [k^i] while *[kWi] > *[tSi] > [si]) and Armenian (where *[q] seems to have been fronted alongside *[k], e.g. *leukos [leuqos] > loys 'light'). Apparently *[k] was fronted first, vacating the plain velar slot, and *[kW] was then delabialised, filling the vacancy, whereas *[q] could merge with either, depending on the dialect. In most cases it fell together with *[kW], becoming *[k], but in Albanian this merger was later than the palatalisation of (labio-)velars before front vowels.

Piotr