permissible IE roots?

[Miguel Carrasquer Vidal]

Rick Mc Callister <rmccalli at sunmuw1.MUW.Edu> wrote:

>	Gamkrelidze and Ivanov 1995: 120-22 give the following rules for IE
>roots with stops (that is, if my notes are correct):
>	1. No 2 consonants of the same order can occur in the same root
>	2. Series II * III can't occur in the same root
>	3. Series I & II almost never occur in the same root exceptions
>include *bhak'- "share, portion, etc", *bhbk'o- "beech", *k^'egh- "branch"
>(Germanic and Baltic only) [BUT *k^'egh- seems to violate rule 1]

Gamkrelidze and Ivanov explicitly allow combinations k^ .. k, kw
.. k etc. (if I remember correctly, they take the occurrence of
such roots with velar + palato-velar as evidence for the separate
status of the palato-velars).

>	problematic roots in Watkins: *ka:dh- "to shelter, cover", *kagh-
>"to catch, seize", *kaghlo- "pebble, hail", *kak-1 "to enable, help",
>*kekw- "to excrete",
>	Question:
>	Does the addition of other consonants or velar/palatal differences
>allow exceptions to G/I rules? e.g. *ghelegh- "type of metal", *kenk- with
>other consonants, *kekw-, stembh-, steigh-

See above for palatal/velar/labiovelar.  The presence of *s-,
which forces a following stop to be voiceless [presumably sd >
st, sdh > st(h)], allows for exceptions to the rule.  The rule
does not work across morpheme boundaries (e.g. *ghelegh- is
clearly extended from *ghel- "yellow") [admittedly there is a
risk here that any exceptions to the rule can be explained away
as "root extensions", making the theory unfalsifiable].

It's interesting that Watkins' exceptions all involve the
problematic vowel *a.

=======================
Miguel Carrasquer Vidal
mcv at wxs.nl