On 2006-07-24 16:52, Jens Elmegård Rasmussen wrote:
> As for the abolition of the subjunctive structure, I believe that has
> quite a good chance of being correct. What has convinced me is the
> strange fact that it is not only the thematic stem formation of the
> verbs concerned that is missing, the particular verbs do not in fact
> occur in *any* form. If the thematic inflection of *bher-, *H2eg^-,
> *weg^h-, *pekW-, *dhegWh-, *leg^-, *seg^h-, etc. was just a
> morphological innovation which Anatolian did not share, then one would
> expect the verbs to show whatever stem-formation they had before they
> took on thematic shape. But the verbs just are not there.
But can you imagine the complete abolition of a root (let alone a class
of roots) so that it disappears together with its derivatives, no matter
how obscured? Let's imagine that the word <God> becomes taboo in English
after a revolution staged by fanatical atheists (or, for different
reasons, after a revolution staged by religious fundamentalists). As a
result, <god> (as a common noun), as well as <godly>, <goddess>,
<godmother> etc. might disappear as well, even if not officially
proscribed. But <giddy> and <gossip> would surely slip by, since it
takes a historical linguist rather than a sans-culotte (or a
blasphemy-hunting zealot) to spot the etymological connection.
Of course it's hard to be sure of anything when one faces something as
odd as the Anatolian situation, but I'm inclined to believe that the
generally high level of lexical replacement in Anatolian, the
"flattening" of the verb system (through the loss of the present/aorist
distinction), and the limited documentation of the Anatolian languages
sufficiently account for the absence of many inherited stems.
Actually, _some_ of the verbs in question do display root cognates in
Anatolian. For example, several formations derived from *leuk- are
attested in Hittite, confirming the reconstruction of a PIE root aorist
middle and the associated causative, but the expected "simple thematic
present" (or rather aor.subj.) *léuk-e/o- (as in Skt. rócate) fails to
turn up. This, in my opinion, militates against the total abolition
theory but is compatible with the hypothesis that it was just the
subjunctive that was lost, and the *CéC-e/o- subjunctives had not yet
become true presents by the time Proto-Anatolian split away from the
rest of IE. Note also that "pseudothematic" middles like CLuw. ziyar(i)
< *k^éj-e(/o-r) apparently were not abolished in Anatolian (Hittite just
used an innovated middle ending in <kitta(ri)> but didn't proscribe the
root).
> The Tocharian paucity of thematic presents is in my opinion different
> from this. As I see it, it was the usual fate of pre-Tocharian aorist
> subjunctives to become present indicatives, and Tocharian repeated the
> process with the subjunctive of the new aorist (s-aorist) which
> yielded the se/o-present. The small number of thematic presents in
> Tocharian represents then, I submit, the tenacious core presents that
> would not yield, the last ones to survive, not the first ones to be
> made.
Now, in Tocharian we do seem to find reflexes of *g^énh1-e/o-,
*gWém-e/o-, *léuk-e/o- functioning as subjunctives rather than presents.
This suggests to me a shift of function in which the above-quoted forms
are subjunctives that might one day become present indicatives --
queuing up for the status, so to speak -- while the descendants of
*h2ág^-e/o- and *bHér-e/o- (plus perhaps a few others) are already on
the other side.
> Under these views, both Anatolian and Tocharian are derivable from IE
> as we know it without any serious problems. For other reasons, the two
> branches may still be the first ones to split away from the common
> trunk, but it is not shown right here.
PIE "as we know it" is a tricky notion, since PIE is what we reconstruct
it to be. Anatolian and Tocharian evidence may force us to revise the
conservative reconstruction, which was essentially arrived at without
taking those "outgroups" into account. In handbooks, *bHér-e/o- is often
the first example of a reconstructible PIE verb stem, and will probably
continue to be so used at least until the dust settles and one of the
currently competing analyses gains general approval. I don't mind it at
all: we need a common frame of reference, even if it's provisional (like
everything in science).
Piotr