The Great Proto-Indo-European Thread

[WeepingElf]

What're the general opinions on Jasanoff's h2e-conjugation theory? It certainly does sound very plausible, and I'll probably run with it for a bit, to see where it takes me.

It appears sensible to me. There seems to have been a pattern in some ancestor of PIE that active verbs took *m-endings and stative verbs *h2e-endings. That stage may have been a true active-stative language. (Of course, I used that idea in Old Albic.)

[KathTheDragon]

Yeah, that's been my plan for my IE langs.

[WeepingElf]

Gamkrelidze and Ivanov also think that PIE was an active-stative language. However, their argumentation is not without problems.

1. They seem to think that the two sets of personal endings encoded the animacy of the object. This IMHO is nonsense. They encoded the semantic role of the subject.

2. They phantasize about two genitive suffixes, *-os and *-om, the former on inanimate nouns, the latter on animate nouns, and claim that it worked that way in Hittite. I consider this claim unfounded. None of the sources on Hittite I consulted confirmed it.

The reasons why Pre-PIE (not PIE at the time of breakup) may have been active-stative are IMHO these:

1. There are two sets of personal endings - the *m-set and the *h2e-set - of which the former was associated with active meaning and the latter with stative meaning.

2. Inanimate (neuter) nouns have a defective case paradigm, with an absolutive ("nominative-accusative") ending in zero for athematic nouns and in *-om for thematic nouns. No ergative case can be reconstructed for PIE; the Hittite ergative case probably is an innovation, and some other old IE languages avoid neuter transitive subjects by passivization. This looks like an animacy-based split between accusative marking on animate nouns and ergative marking on inanimate nouns, which could easily have emerged from an active-stative pattern where inanimate nouns may have lacked an agentive case.

BTW: The two sets of personal endings seem to have cognates in Uralic (at least in Hungarian and Selkup) as well as in Eskimo-Aleut.

[KathTheDragon]

Hittite ergative case probably is an innovation

IIRC, it was the something (ablative or instrumental, I think instrumental) case of n-stems, which generalised.

, and some other old IE languages avoid neuter transitive subjects by passivization.

IIRC, this is how Indic got its ergative.

BTW: The two sets of personal endings seem to have cognates in Uralic (at least in Hungarian and Selkup) as well as in Eskimo-Aleut.

Really? Sources?

[gach]

BTW: The two sets of personal endings seem to have cognates in Uralic (at least in Hungarian and Selkup) as well as in Eskimo-Aleut.

Really? Sources?

I'd be interested in specific proposed cognates as well.

[hwhatting]

Slavic peripheral? It's about as "core" as it gets IMO, both geographically and in shared features with Greek and especially Indo-Iranian.

Not peripheral like Anatolian or Tokharian, no, that's not what I meant; my point is more that I find Italic (or let's better say Latin) nearer to Graeco-Aryan than Slavic at least in some regards (reduplicating perfect, thematic long-vowel subjunctive, preservation of the mediopassive) than Slavic. My impression of Slavic is that, like Baltic, it was originally further away from Graeco-Aryan, but then shared some Graeco-Aryan developments under the influence of Iranian. Some issues:
1) Slavic has only a few remnants of the IE perfect (except for vědě "know", all of them are only recognisable by their ablaut grade and show the usual thematic endings), none of them reduplicating; as a category, the perfect does not exist and its remnants are part of the present tense system, not of the past tense system. It very much depends on your view of the development of the IE verbal system and of the exact position of Baltic and Slavic in the IE family whether you assume that Baltic and Slavic lost the "classical" reduplicating perfect or perhaps never developed it at all.
2) Baltic and Slavic have some athematic verbs with the expected mi-endings and thematic verbs with a mixture of mi- endings and endings that look like old mediopassive / hi-endings, but do not have an old IE mediopassive, neither according to the Graeco-Aryan model nor with r-based endings. Again, you can explain this as a loss of the mediopassive that only left traces in some (active!) thematic endings, but one could also assume that Baltic and Slavic simply went a different road when the e/o type verbs moved to the mi-conjugation and never developed the functional split into active and mediopassive endings. (That, of course, would make them very peripheral, as even Anatolian has that split)
3) The Slavic Aorist is a mixed bag, consisting partially of formations that have parallels in aorist formations in other IE languages (especially s-aorists), but also of imperfects / injunctives of present tense stems, and partially of formations that stem-wise look like the a- and e-preterites of Baltic, but with s-Aorist endings glued on. I assume that Baltic and Slavic inherited an aspect distinction from IE and an s-Aorist like formation to form perfective stems for imperfective verbs, but that the Slavic Aorist as a tense category was fully formed only under Iranian influence, changing from a system more similar to the Baltic one, while Baltic lost those formations.

[Pole, the]

Because 'Melchert' is actually a name that could easily be the name of, literally, an Anatolian demigod.

Well, maybe Craig Melchert is literally an Anatolian demigod, they just didn't worship him. And that's why they went extinct.

(Yes, I am joking.)

* * *

BTW: The two sets of personal endings seem to have cognates in Uralic (at least in Hungarian and Selkup) as well as in Eskimo-Aleut.

As far as I know, originally the Uralic language had only one conjugation paradigm, derived from personal pronouns (1sg *mV, 2sg *tV, 3sg *sV).
Later the conjugation drifted towards definiteness in singular (*-mV *-tV *-sV → Hung. -m -d -i/ja) and indefiniteness in plural (*-mVk *-tVk *-[nA]k, retained in archaic Hungarian). Definite plural used the 3sg pronoun to denote definiteness (*-sV-mV-k *-sV-tV-k *-sV-k, cf. Hung. -jUk -i/ja-tVk -i/ja-k).*
Singular indefinite paradigm was then recreated with some frequentative suffixes (thus Hung. -k -sz/l -∅).

And then we have a group of somewhat stative verbs ending in -ik — but those are not an active/stative retention, rather an innovation on the basis "they like it" → "it is pleasing". (See above.)

*) 1pl endings in contemporary Hungarian have yet another source.

[WeepingElf]

So Hung. 1sg. indefinite -k has nothing to do with PIE 1sg. stative *-h2e, as I thought, then. Given the uncertain nature of the PIE laryngeals, it is massively uncertain that there is any connection between PIE *h2 and Uralic *k anyway. Also, the meanings of the suffixes differ considerably, and the lopsided distribution of "Uralic" *-k (only Hungarian and Selkup, not even Ob-Ugric or North Samoyedic!) speaks against a Proto-Uralic origin. The second person suffixes are *-l in Hungarian and *-nt in Selkup - hard to reconcile with each other and with PIE *-th2e. So this was just wild, unfounded speculation, and I should better forget it.

The Eskimo-Aleut case is problematic, too. The personal suffixes there are highly fusional and difficult to disentangle. I have seen intransitive suffixes 1sg. *-k and 2sg. *-tk in a reconstruction by Miguel Carrasquer Vidal, but other sources don't confirm that, and it may have been wishful thinking on Carrasquer's side.

[Pole, the]

Interestingly enough, you have frequentative -l -k in English as well (albeit inproductive, v. scribe → scribble or tell → talk).

[jal]

Interestingly enough, you have frequentative -l -k in English as well (albeit inproductive, v. scribe → scribble or tell → talk).

Isn't that purely coincedental?


JAL

[Zju]

Interestingly enough, you have frequentative -l -k in English as well (albeit inproductive, v. scribe → scribble or tell → talk).

Isn't that purely coincedental?


JAL

Indeed, an onomatopoeic origin has been argued for -l^[1].

^[1]http://langevo.blogspot.com/2013/05/morphemes-are-forever.html?showComment=1367747083711#c3207882693051432378

[Pabappa]

On another note, why is the first person singular nominative pronoun reconstructed as *h₁éǵōh₂? (Specifically, why that particular ending?)

I cant answer you here, but is it really ōh₂? I thought that the vowel length would only arise after the laryngeal had elided away. If we assume that there was a stage at which long vowels didnt exist, then it would have to have at some point been h₁éǵeh₂, which can be simplified to h₁éǵh₂. Homeric Greek occasionally had a nasal ending, and it looks like Sanskrit does as well, which might suggest an even older form like h₁éǵh₂m, but I guess the consensus is that the -m was not used in the original language.

[vec]

Aesthetically, I simply cannot accept that a 1st person pronoun was h₁éǵh₂m.

[jal]

Aesthetically, I simply cannot accept that a 1st person pronoun was h₁éǵh₂m.

Of course, it needn't have been a pronoun originally. For all we know it meant "the one speaking" or "this person" or whatnot*, and it later developed into a pronoun.

* Yeah, I know we know it doesn't mean that, just giving an example.


JAL

[CatDoom]

On another note, why is the first person singular nominative pronoun reconstructed as *h₁éǵōh₂? (Specifically, why that particular ending?)

I'm curious as to where you're getting that reconstruction. According to Wikipedia, Beekes (1995) and Sihler (1995) both reconstruct the first person singular nominative pronoun as *eǵoH, with Beekes giving the alternate form *eǵHom, presumably to account for the nasal ending in early Greek and Sanskrit. Wiktionary, on the other hand, gives it as *éǵh₂, but doesn't cite a source. The monosyllabic reconstruction would, I think, help to explain the Gothic ik, at the very least.

[Sleinad Flar]

Quoth Beekes:

‘I’. Nom. Some languages have added *-ō (-e/oH?), others -(H)om (Skt. ahám < *h₁eǵHom). Old Avestan has azə < *h₁eǵ once. Hitt. ūk has its -u- from ammuk, OCS azъ for *jazъ < *ēg(H)om with lengthening of *e according to Winter-Kortlandt’s law.

So basically the only certain part is *(h₁)eǵ- (Beekes does recontruct the initial laryngeal, even though there is no direct evidence for it).

Hypothizing mood: *-ō/-oH and *-om look suspiciously like 1sg verbal endings, don't they?

[KathTheDragon]

On another note, why is the first person singular nominative pronoun reconstructed as *h₁éǵōh₂? (Specifically, why that particular ending?)

I'm curious as to where you're getting that reconstruction.

Me too.

[KathTheDragon]

Is it possible that the final *-m from Greek and Sanskrit might be the secondary 1s verb ending *-m? I know that both Greek and Sanskrit did renew their thematic 1s endings with -m-.

[R.Rusanov]

I thought eghom was originally a verb meaning smt like "I do" or "I stand". Cognate to agere and agitate.

[KathTheDragon]

I thought eghom was originally a verb meaning smt like "I do" or "I stand". Cognate to agere and agitate.

No, agere is from *h₂eǵ-.

[hwhatting]

Wiktionary, on the other hand, gives it as *éǵh₂, but doesn't cite a source. The monosyllabic reconstruction would, I think, help to explain the Gothic ik, at the very least.

The Germanic forms can also go back to something ending in a short vowel, as many of those were dropped word-final in Proto-Germanic.

[Chagen]

This reminds me of how some endings in IElangs don't match up with each other.

For instance, the 1P ending is -mus in Latin, -men in Greek, and -mas in Sanskrit. There's no one PIE suffix that could explain the sibilant in Italic and Indic but the nasal in Greek. It's the same way with *eghoh2/*eghom/*egh2. PIE has to be one of the most bizarre languages to ever have existed. Are there any cross-linguistic parallels to its ablaut system, for instance?

[R.Rusanov]

... have you heard of Arabic?

[KathTheDragon]

I can mostly explain IE ablaut as simple stress movement and vowel quantity changes. Nothing Semitic-esque is required whatsoever.

[vokzhen]

This reminds me of how some endings in IElangs don't match up with each other.

For instance, the 1P ending is -mus in Latin, -men in Greek, and -mas in Sanskrit. There's no one PIE suffix that could explain the sibilant in Italic and Indic but the nasal in Greek. It's the same way with *eghoh2/*eghom/*egh2. PIE has to be one of the most bizarre languages to ever have existed. Are there any cross-linguistic parallels to its ablaut system, for instance?

Well, PIE as we reconstruct didn't exist. I'm pretty sure not very controversial to say it's an artificially-unified dialect chain, where the differences between dialects can't be teased out, and as a result we have three different possible reconstructions of the 1S pronoun, and something like 20 different ways of forming the present. That's to say nothing of the time element of attempting to include languages over a 1000-year span as all "PIE," when for example we know Anatolian never had the ridiculous verbal system of PIE but we can't reconstruct past the complexity in the other languages to a stage that actually describes both.