--- In
cybalist@yahoogroups.com, "Glen Gordon" <glengordon01@...>
wrote:
> I'd like to contribute what I can if possible, but how can I if
you haven't
> formulated what your idea is about? [...] I think criticism is
still
> earned,
> this time for presenting an idea to vague to comment on.
Well, fair enough, but don't expect me to make empty guesses. I also
wish we knew more, and some just can't be bothered with discussing a
matter of which we know so little.
> Were you suggesting a sound change that caused two very seperate
> forms *-u and *-ro-?
Yes, I don't see how that could be wrong. Except that "very
separate" is a subjective statement.
> Something like syllabic *-r > *-u while thematic
> *-r-o- remained unchanged? Is that it?
No. I don't know how it was exactly, but I would not assume exactly
that. The whole matter is restricted to thematic formations, so I
cannot have *-u- deriving from an athematic form. I suspect all
three allomorphs *-ro-, *-u-, and *-i- represent the same underlying
form. Maybe we will some day be able to see how *-ro- produced the
other two under specific conditions of reduction, or, in my opinion
more probably, we will find out how some common older form different
from all three developed into each of the three.
> I think you mentioned something about *-u and *-ro- appearing in
> different phonetic contexts. Can you elaborate on this? Did you
mean
> that the choice between one suffix or the other is dependent on the
> phonetics of the root involved, somewhat like the choice of s- and
> l-genitives in Etruscan? If so, describe these phonetics. Maybe I'm
> missing something.
Yes, I did just that. You can find it by cutting up-thread in my
discussion with Miguel, you'll find a slightly updated list of
examples based on Brugmann's Grundriss. If I go look for my older
posting now, this message will disappear, so please understand that
I can't be more specific. The posting is hardly a week old. My
contention was that there is a far greater and clearer complementary
distribution between ro- and u-stems in terms of root structure than
could have been brought about by chance. Actually it is *very* clear
in Brugmann's lists although it is not a point made by Brugmann and
cannot be suspected to have been a disturbing principle in his
selection of examples.
The visible principles of distribution are these: The simple
adjective has *-ro- if the zero-grade root contains a vowel
(meaning, this time, also /i/ and /u/ and the little subphonemic
vowels sustaining "syllabified" laryngeals), while it has *-u- if
the zero-grade root has a syllabified sonant (other than semivowel).
There are too few examples having anything else to allow me to say
what they should have. When such adjectives are used as the first
part of compounds the ro-type replaces -ro- by -i-, while the u-type
keeps its -u-. If the ro-type - or any other thematic type - is used
as the first part of a juxtaposition of instrumental + finite verb,
the instrumental has the form *-ri-H1 (or other *-Ci-H1), the
classical example being Skt. kru:ri:- + forms of kr- 'do' or bhu:-
'become' meaning 'make/become bloody'. The analysis of the matter
has progressed to a point where it appears safe to regard the form
of the adjective as an original instrumental. Functionally it is a
predicative use of the instrumental just as in Slavic.
Now, how do we account for that? Well, forms enterering into
composition are plainly reduced, cf. *n-, *dwi-, *H1su- and many
others which never show full-grade. That rings a bell, for the
reduced forms of the thematic vowel (quoted as *-o- when not
reduced) are precisely *-u- and *-i-. A fine hierarchy is seen in
the variants of *-tó- (accented form) which are *-tu- (immediately
following the accent) and *-ti- (final of compounds): Vedic kr-tá-
'made', kár-tu-m 'to make', havís.-krti- 'the making of a
sacrificial meal'.
Since ro-adjectives (descriptively) become u-adjectives when the
root contains a sonant /l, m, n, r/, the only obvious explanation
for the u-variant is that is reduced because of the uncomfortable
presence of two such sonants in the same (phonological) cluster.
That would combine loss of /r/ and the change of /o/ to /u/ under
the heading "reduction" which tallies well with the appearance
of /u/ under conditions of reduction elsewhere. And the reduction of
*-ro- to *-i- is even clearer, for *-i- is certainly known to be the
reduced form of the thematic vowel in many formations.
However, the burning question os of course how the exact rules of
phonetic change were. Perhaps that just can't be answered: If there
are too few other forms of an undisturbed history lying around in
the lexicon to reveal how the changes were we can't solve the
problem. There are further ramifications of the problem that may
bring important evidence into play, but minds will differ as to
whether this pays off.
Jens