Re: [tied] Re: PIE conjugations

Hi, Miguel,

Hope you enjoyed your stay in Poland. I raised the problem of PIE conjugations in order to show that what even very safe-looking fragments of our PIE reconstruction may suddenly be found to stand in need of revaluation. What Adams does in the EIEC is certainly a little premature. A new type of present related -- as far as its inflections go -- to the traditional "perfect" and to the Hittite hi-presents has probably come to stay; but the revolutionary new look of the traditional thematic conjugation is not yet sufficiently established to be included in a book addressed not only to students of IE but to the general reader as well. I shall reserve my judgement on these new tendencies till the publication of Jasanoff's forthcoming book on PIE conjugations. It's not only the present but also the preterite system and the relation between the two that should be reconsidered.

Some of the complications can't be explained as easily as you propose. For example, to get from *-eti to Greek -ei you need to assume palatalisation to *-esi (one would expect Doric to preserve *-eti) and the analogical restructuring of the now-opaque ending (along with that of te 2sg.) to bring it in line with the aorist. If the proto-Greeks were so eager to level out present and preterite endings, why didn't they do something about the first person sg. (e.g. add a nasal to -o: as in Slavic or go the whole hog and replace the old ending with *-oin, parallel to past-tense -on)? Explanations that depend so heavily on analogy are difficult to constrain properly and must remain highly conjectural.

As for the 1sg. *-o:, the chain *-omwi > *-omwu > *-owu > *-o:u > *-o: is hardly preferable to *-o-h2, though the latter has its problems too. Even if one accepts your labialised /mw/ and the vowel modification it causes for the sake of the argument, the unmotivated loss of the nasal (who would have wished to remove the most transparent element of a 1sg. ending?) and the lack of tangible support for *-o:u make this derivation look somewhat fanciful.

----- Original Message -----
From: Miguel Carrasquer Vidal

To: cybalist@egroups.com

Sent: Wednesday, January 10, 2001 3:12 AM

Subject: Re: [tied] Re: PIE conjugations

On Sat, 23 Dec 2000 10:50:29 +0100, "Piotr Gasiorowski" <gpiotr@...> wrote: [regarding:] >Singular Plural >1 bhéroh2 bhéromes >2 bhéreth2e? bhérete >3 bhérei bhéronti >Many thanks, Mark. Of course, /th2e/ = *th2a. The question whether there should be a final *-i in the first two persons is really moot. Hittite endings (-(ah)hi, -ti, archaic -he, -te < *-h2a-i, *-th2a-i) I have a slight problem with this. The Hittite cuneiform for <he> and <hi> are the *same* sign (#335 in Ruester & Neu's list). True, there is a <he2> (#113), but it can also be read as <hi2>. The same problems apply to the signs <te> (#249) [also read <de4> or <ti7>] and <ti> (#37) [also read <di3> or <te9>]. Finally, we have sign #312 which is both <de> and <di>, but also <ti4>. Ruester & Neu fail to point out which of the syllabic readings are found in Old or Young Hittite, but I find it far from obvious that the Hittite forms should necessarily and exclusively be reconstructed as *-h2ai and *-th2ai. A perfectly good case can be made for *-h2i and *-th2i (with lack of palatalization as in *dhi > ti as opposed to *ti > (z)zi and *di > si) instead of or beside *-h2ai and *-th2ai. The shape of the past tense 1sg. *-hun shows that *-m was added to *-h2 directly, not to *-h2a. >suggest it should be there (and also in the plural: -meni/-weni, while the stable *-o: of the 1sg in the non-Anatolian branches points to *-o-h2 (with *-o- on the analogy of -o-m?). Vide infra. Regarding the earlier posting: >The details differ somewhat from author to author, but *bHere- is conjugated more or less like the rest of Class II (*bHeroh2(i), *bhereth2ai, *bHerei). I'm not sure about Adams's favourite endings; I'd have to go to my institute library to check that, and that won't happen until early next millennium. But Mark has a copy of his own and if we ask him politely, maybe he'll be good enough to check the details in the article on the "Proto-Indo-European Language" and tell us how Adams conjugates *bHere-. (Would you, Mark?) Anyway it's clear that *bHere-si and *bHere-ti are innovated (they don't occur in some branches in the thematic conjugation), while "*bHero:" is an old form. I'm confused as to what exactly a 2sg. *-eth2ai (or *-eth2a) is supposed to explain. It might explain the Tocharian 2sg. ending in *-t, but that is not an exclusively thematic form, which makes it less relevant to the discussion of the PIE thematic present. The "problematic" forms of the 2/3sg in the thematic conjugation (which do not have *-esi and *-eti) are the Greek and Balto-Slavic forms. They are: 2sg. 3sg. Greek: -eis, -ei Lith.: -i, -a OCS: -es^i, -etU Now the Greek forms are easily explained if we depart from the traditional thematic reconstructions. We have: present past *-esi *-es *-eti *-et, which by regular phonetic development become: *-ei *-es *-esi *-e. The -s from the past tense is restored in the 2sg. present (but in the wrong place): *-eis *-es *-esi *-e, and then, by analogy with past tense *-es, *-e, the present tense forms become *-eis, *-ei. This certainly makes more sense than explaining -eis from *-eth2ai (which cannot be done). As to the Balto-Slavic forms, the 2sg. is indeed mysterious. The Lithuanian form goes back to *-ei, while the final -i in OCS -es^i must also go back to *-ei. The most likely hypothesis is that the lack of *-s in Lithuanian was brought about by a re-analysis of 2sg. *es-sei "you are" as *es-ei. The OCS -i is likewise transferred from <esí> "you are", and in fact all Slavic lgs. except OCS testify to a form *-es^I, with short *-i. Whatever the raison d'être of Balto-Slavic 2sg. *-ei in the verb "to be" (cf. also OPruss. asmai "I am", more anon), it has little to to with the *thematic* conjugation ("to be" is athematic). Remains the problem of Slavic *s^. Usually this (and its unpalatalized counterpart *x) derive from PIE *s after *i, *u, *r and *k, but of course, this being the thematic conjugation, the preceding sound is necessarily *e. The only other possibilities are the cluster *sj (as in <s^iti> < *sju:- "to sew") and the PIE unitary phoneme *sw (as in <s^estI> "six"). Maybe a case could be made for *-es-jI in conjunction with 3sg. *-e-tU, i.e. postfixed pronouns *yos and *tos, but, frankly, 2nd person deixis for the pronoun *yos is unexpected. I prefer to see Slavic *-eswi as evidence for the pronominal origin of the PIE active conjugation (postfixed *-mu and *-tu, with developement *-tu > *-tw > *-sw > *-s in the Auslaut). The Balto-Slavic 3rd.p. sg. *-e (Slav. *-e-tU with postfixed pronoun [sometimes lacking], as opposed to expected -tI < *-ti in the athematic conjugation; Lith. -a ultimately from *-e, apparently due to the i-stems) can be seen as a B-S innovation for now. It certainly does not point to *-ei. In summary, I find it far from clear that *-esi and *-eti are "innovated", based on the available evidence. The main reason for wanting to see the Greek and Balto-Slavic forms as pointing to something similar to the "stative" conjugation (*-eth2a(i), *-e(i), or Beekes' *-eh1i, *-e) is surely the 1sg form *-o:, which is found practically everywhwere. It's certainly tempting to see it as reflecting someting with *-h2, the characteristic stative/perfect/ middle ending, but unfortunately *-oh2 doesn't quite convince. For one thing, the vowel is wrong (we'd expect *-eh2 > -a:). For another, where did the *-i go? PIE *-o: is mysterious, but explaining it as *-eh2i, with somehow *-o- for *-e-, and somehow loss of *-i, raises more questions than it answers. Since above I've argued that the PIE thematic present endings in the 2 and 3 sg. were almost certainly *-esi and *-eti, as we'd expect them to be (except that *-esi should be *-eswi, as shown by Slavic), let's see how far we can get by assuming expected *-omi (or, actually: *-omwi) for the 1sg. As it happens, it gets us quite far in comparison with *-eh2i, although unfortunately not quite where we'd want to get. What I suggest is *-omwi > *-omwu, with umlaut of *-i to *-u after the labialized consonant, as we've seen in the loc.pl. *-su (< *-sw-i). Then, loss of the nasal element, giving *-owu, which gives *-o:u as it does in the loc.sg. of the i-stems (*-ei-i > *-e:i) and the u-stems (*-eu-i > *-eu-u > *-e:u [or: *-ou-i > *ou-u > *-o:u]). This *-o:u, to the best of my knowledge, would account for most of the attested forms, except that we don't find variant forms with -au (-a:u) in Avestan or Vedic, unfortunately. So is there really no reason to postulate a paradigm like *-eh2(i), *-eth2(i), *-e(i) for PIE? Actually, there is. Except that the *-e- is not the thematic vowel, but the conjunctive marker *-e-. The "stative" differed, in my view, from the "active" by a different order of agglutination of the constituent elements. Whereas in the active we have ROOT-[conjunctive *-e/o-]-[thematic *-e/o-]-personal endings-[*-i marker], in the "stative" we have ROOT-[conjunctive *-e/o-]-personal endings-[thematic *-e/o-]-[middle markers]-[*-i-marker]. Thus, the perfect endings *-h2a, *-th2a, *-e, are in my view *thematic* [*-h2-e, *-th2-e, *-0-e], whereas Hittite -hi, -thi, -i are, at least partially, *athematic* [*-h2-i, *-th2-i, *-0-i]. The stative conjunctive would have had [athematic] *-e-h2, *-e-th2, *-e-(t), [thematic] *-e-h2-e, *-e-th2-e, *-e-(t)-e. Now, as I've argued elsewhere, PIE *-t became *-h1 (but not in Anatolian, nor where analogically restored as in 3sg. past *-et due to present *-eti) as is shown most clearly in the ins.sg. *-et (thus in Hittite) > *-eh1, or variants like *met- ~ *meh1- "to measure". If we consider 2sg. stat. conj. *-eth2 and apply the rule, we would get *-eh1h2, which (as long as we cannot give uncontroversial phonetic values to the laryngeals) might have been simplified to *-eh1 or *-eh2. In the latter case, we have a source for the a:-conjunctives and a:-imperfects of e.g. Latin (-a:+m, -a:+s, etc.). The former case (1sg. *-a: < *-eh2, 2sg. *-e: < *-eh2 < *-eth2, 3sg. *-e: < *-et or *-ee) is actually attested in the Latin future (< conjunctive) of the i- and C-conjugations (audiam, audie:s / emam, eme:s). Finally, generalization of *-e: would be the source of the e:-conjunctives (e.g. Latin) [partially also from normal thematic *-e-et] and e:-pasts [partially also from normal thematic *-et > *-eh1, where not analogically restored] (e.g. Latin, again, as in audi-e:-bam, em-e:-bam, and Balto-Slavic, e.g. the Slavic imperfect in -e:(j)a:xU < -e:-e:xU etc. which is parallel to the Latin one, except with a different form of the verb "to be"). This stative conjunctive may even be the explanation we were looking for for the mysterious Balto-Slavic *esei "you are". We have (perhaps in order to avoid thematic-looking athematic conjunctives of the <ero:>, <eris> kind) a "stative" conjunctive: *es-e-h2 > *esa: + *-i > *esai (> OPr. asmai) *es-e-th2 > *ese: + *-i > *esei (> OCS. esi, OPr. asei) *es-e-t > *ese: + *-i -- But for the -i, these forms are parallel to Latin audiam, audie:s, audie:t etc. ======================= Miguel Carrasquer Vidal mcv@...