Genitives in the Future

A while ago I quite quickly accepted the possibility of a contraction of Homeric Greek genitive ending -οιο to -ου, as seen in the most dialects.

Now I'd like to argue this once again! First of all, this is not the commonly expected contraction of any dialect; And I'd rather avoid explaining strange forms by irregular sound contractions. In this post I opted that -οιο seems to reflect a form PIE *-osio where -ου reflects PIE *eso or *oso. The first one would correspond with Sanskrit -asya, while the second form agrees with forms like the one we find in Gothic -is.

Of course our high sense of lumping everything together as some kind of Proto-Greek would make us want to delete once of these forms. But I think we might have to consider that maybe Proto-Greek at its very earliest stage already had dialectal variations, which may have even been carried over from Indo-European.

It might be worth noting that -οιο is not only seen in Homeric Greek, but also in Mycenean. Bonfante went as far as suggesting that Homeric text was originally a Mycenaean text in his article "Homer Text is Mycenaean"[1]. I would not go this far, throughout Homer there's such an enormous amount of evidence that the main dialect was Ionic, that it seems absurd to claim that it's Mycenaean just because of a Mycenaean-like genitive and evidence of a once existent waw (which all dialects must have had at some point, not just Mycenaean).

I'd sooner think that -οιο was loaned from Mycenaean because it is metrically a lot more pleasant to work with, for a dactylic hexameter. But that idea is sort of denied due to the almost exclusive use of -οιο rather than a metrical whim deciding whether to use -οιο or -ου.

So when we indeed believe that -οιο and -ου are from different etymological sources, which is in my opinion more probable than a sporadic contraction, especially because we find both *-osio and *-eso reflexes in other languages, then we can conclude that there is some kind of switch between *-si- and *-s-, this would be a bizarre assumption if it was only found in the genitive, but lo and behold, such a switch is found in the formation of the future as well!

In Greek, to create the future, you take the verbal root, and add a -s-e/o suffix to it. For example πίμπλημι 'to fill' is a reduplicated present of the stem πλη/πλα- Then the future is: πλήσω. The sigma, in some environments disappears, but this is a post-proto-greek development, and not very important for this discussion.

Now, let's have a look at the Sanskrit future! Sanskrit doesn't insert a -s-a- suffix, like you'd expect looking at the Greek form, but a -sy-a- suffix, and just like Greek it's added to the full grade of the verbal root. For example: tiṣṭhati तिष्ठति 'to stand' stem: sthā- स्था The future is sthāsyati स्थास्यति.

aha another s/si switch. What exactly does this mean? Why does it happen? Why does Greek have both a *s and *si variant for the genitive?

I think this is an indication that Greek and Indo-Iranian languages may have been a lot closer than we think. But why this *s *si alternation seems to be taking place is beyond me.

It might also be worth mentioning that Greek and Indo-Iranian are in fact the only two branches that have a future with *s. Most Indo-European languages don't have a future at all. Why do we reconstruct this future as something from PIE? Because one day someone came up with the arbitrary rule: If it's in a European PIE languages, and an Asian PIE language, than it must be PIE. This idea is nonsensical, it might be helpful to establish true PIE roots, but for grammatical information like this, such rules should not be applied.

Once again, I find it difficult to go as far as saying that Graeco-Iranian was once an actual separate branch, but there's so many unique features to this group of languages not found anywhere else that it's quite idiotic to take the rules found in these languages and propose them as Indo-European, rather than the Graeco-Iranian dialect. Proof of the things we find in Graeco-Iranian is technically absent in all the other languages. Just because both branches have been extremely conservative doesn't mean that non of their shared features are archaic, rather than shared innovations.

[1] Bonfante, Giuliano 'Homer Text is Mycenaean' JIES 1996

[EDIT] Thanks to Glen for pointing out my terrible misspelling of Giuliano's name. Hah. I should look up whether this article also has the typo, because I was fairly sure I directly copied it.

Comments

[Glen Gordon]

Wow, since I'm so narrowly focussed on Etruscan, my brain normally associates Giuliano Bonfante with Etruscans so it's strange to see references to his work on Greeks but of course he had to have acquired his Indo-European biases from somewhere ;) (Btw, you mispelled his first name. The 'i' should come immediately after the 'G' to soften the otherwise hard /g/ to /dʒ/, of course. Sorry for being nitpicky but Language Hat has diagnosed me with Asperger's so it's expected of me, hehe.)

As for the Greek future, isn't the very fact that it contains a sigma an indication that the form is from a future in *-s-ye/o-, not **-se/o-? Afterall, intervocalic sigma should disappear as it does in the Greek prefix eu- 'good' (< PIE *h₁esú-)), no? And if you admit that the loss of sigma is a "post-proto-greek development", then don't you effectively neuter any argument based on the Greek data that asserts **-se/o- as a future marker at the Indo-European stage?

"Because one day someone came up with the arbitrary rule: If it's in a European PIE languages, and an Asian PIE language, than it must be PIE."

Basically, yes. It's a throwback to the "centum-satem split" myth. Of course, with the Anatolian-Tocharian data as well as the fact that language evolves more like waves than into clean phylogenetic 'splits', we need to reconstruct something that represents an older stage that we can appropriately call "Proto-Indo-European" in the modern age. 

The stage at which *-s-ye/o- was synthesized into a future suffix appears to be a post-IE development. The component *-s- is the suffix that would become the sigmatic aorist (probably originally a stative suffix) and Anatolian evidence works against the idea that Proto-IE had yet developed a formal "sigmatic aorist". The second component of the future suffix is *-ye/o-, which is used on its own to either form denominal verbs or to convey the durative (i.e. the continuous present-future). I assume that since verbs in *-s- probably indicated states without a definable end originally and since *-ye/o- was already used for duratives that were by nature often also used for futures anyway, the two suffixes combined and gradually became specialized for a future tense now seperated from a present tense.

From what I understand, Proto-IE (properly addressing Anatolian data) must have maintained only a tensal distinction between <a href="http://books.google.ca/books?id=gWY7-DBWPW4C&pg=PA362&vq=%22contrast+with+the+systems+of,+say,+germanic+and+hittite,+with+a+primary+opposition+of+past+vs.+non-past+events.%22&dq=&source=gbs_search_s&sig=D6sVcVWCVulMYie9lnGeOBwOVTw">past and non-past</a>. Naturally, in such a system, it would not be possible for verbs with inherent "aorist semantics" (i.e. momentaneous actions) to have, by their nature, a strong tensal distinction at all. So, this subtle semantic imbalance would be the foundation for the more complex "triaspectual system" of the Greek-Sanskrit sprachbund. In that way, we account for all PIE conjugational systems.

As I keep on saying, Jasanoff and his musings into a revised reconstruction of the PIE conjugation system is really worth a careful read. :)

[PhoeniX]

[this is good] Yeah I should really get round to reading that now, I've been saying I should for ages now.

As for the sigma being there in Greek shouldn't be an indication of a *-s-ye/o-. Just like *osyo develops into *oyo* we'd expect a *y to be retained as a iota, rather than the sigma.


And if you admit that the loss of sigma is a "post-proto-greek
development", then don't you effectively neuter any argument based on
the Greek data that asserts

**-se/o-

as a future marker at the Indo-European stage?


Yeah, now you mention it, I've formulated it rather clumsily. What I meant to say was; at some point there must have been an *h marking the future in most forms, which is still retained in some cases, while *s functioned as the marked of the future if the root ended in a consonant. We usually propose that then the *s was generalised throughout the paradigm, especially imaginable if you realise that this *h becomes Ø in the context of the future.

I'd like to find out how Mycenaean futures worked, or if we have any attested forms at all. I don't know any Mycenaean, but I'll see if I can look into it.

Bottom line is though PIE *sy should give PGr. *i intervocally not *s. I'd like to unify these two forms though, but honestly have no clue how to do it right now, I'm open to any suggestions.

On a completely different note: You might find it interesting that I made a copy of a description of rituals concerning birds. Given your recent interest in Anatolian rituals and divination I thought you might be interested. I'll try to put up translations soon.

[Glen Gordon]

I think I goofed up somewhere along the way. It's "all Greek to me". Lol :P Oh well, let's start again with more coffee in hand. Mmm, coffee. Anyways, about futures in Mycenaean, perhaps this link might be useful to you which shows an ancient future in -s- (i.e. do-se = dōsei 'he will give')

 

Phoenix: "Given your recent interest in Anatolian rituals and divination I thought you might be interested. I'll try to put up translations soon."

Yes, great! I've just recently got sucked into Pre-IE again, but I have something still to write about concerning Etruscan haruspicy and Near Eastern connections but I've been putting it off and putting it off. It's a big topic and I'm not sure how I should approach it. So much to say, so little time.