The Great Proto-Indo-European Thread

[KathTheDragon]

Gah, I knew I'd have to look round more.

[Sleinad Flar]

As promised, some side-notes to Herr Hatting's excellent post. Note that I'm not attempting to give an alternative hypothesis here, just going to play the devil's advocate.

What is absent is the e-grade thematic present, which in other IE languages is the most far-spread form; but we have other thematic formations attested in Anatolian (zero-grade thematic, -sk'e/o-).

While there are forms in -e/a- in Hittite after a zero grade root, none of them match with certainty to tudáti-forms elsewhere (I'm using tudáti as short-hand for "oxytone zero grade simple thematic presents", and bhárati for "barytone e-grade simple thematic presents"). There are matches with statives in -(é)h₁-, and some are univerbations with -d^h(é)h₁-. The Anatolian demi-god Melchert seems to suggest this is the origin of the Hittite -e/a- conjugation, and there is no relation to the PIE simple thematic presents.
This would of that ALL simple thematic presents are absent in Anatolian, not just the bhárati-type. Which is doubly strange, as the derived thematic suffixes *-ye/o- and *-sk'e/o- are very common.

I don't think it very likely that the e-grade would vanish without a trace in Anatolian, while the "rarer" types remained, if it would have been already the "basic" type in IE when Anatolian branched off; I think it's more likey that the e-grade thematic class either did not exist or was still rare at that point, which means that it's a late development. I think that the original thematic inflection had the zero-grade, and that the e-type originally was formed from Narten roots (which have /e:/ grade in the full grade and /e/ in the zero grade); the thematic e-grade-present was then formed from non-Narten roots later in analogy.

All plausible, but what would be the starting point of adding -e/o- to athematic roots? What was it's function?

As there is also no trace of the subjunctive in Anatolian, although there are thematic presents, I find it a priori more reasonable to assume (keeping in mind that Anatolian is attested so early) that the thematic presents are not derived from the subjunctive, but that the subjunctive is a re-purposed thematic present. This would also be more in line with observable developments in other languages - there are many examples of simple presents that have become subjunctives or future tenses, but I wouldn't know any example where a subjunctive has become an indicative present tense.

Depends on the role of this 'subjunctive'. The handbooks seem to suggest that the PIE subjunctive was primarily a future tense, with some modalities like wish, demand etc. mixed in. In that light, I don't think a development from "I am going to carry" or (even better) "I am about to have carried" to "I am carrying" is that strange.
In this light, maybe the use of this 'future tense' was taboo in Anatolian, it was like tempting the gods. To be on the safe side, they deleted everything which looks like a future/subjunctive as well, including the bhárati type. The tudáti type might have gotten a free pass, as at this point it was indistinguishable from eh₁-statives.
(I am aware that all of this is just a train of thought, it's not even a hypothesis. I am also aware that this is a very weak counter point )

My assumption is that the late PIE had -se/o- presents that became used to express intended or prospective Actions; in some IE languages these formations have survived as prospective / future / desiderative tenses / moods; when formed from -H₁es- "to be" and from verbs that had s-aorists, the formation X-se/o- became re-interpreted as Xs-e/o- and then the newly formed -e/o- Suffix spread.

The precise form of this "desiderative" suffix is heavily debated: some think it was *-h₁s-, others think it's simply *-s-. Some think it was thematic, others (e.g. LIV) athematic. Some even think it didn't exist at all, and these were just s-aorist subjunctives (the reduplicated forms in Old Irish and Vedic would be a case of parallel development). I would be very hesistant to use this uncertain suffix as a basis for analogy, and there's grave danger of wishful thinking and circular reasoning.

Arguments for this:
1) All languages that have the "classical" -e/o- subjunctive also have at least traces of -se/o- futures / desideratives
2) All languages that have the "classical" -e/o- subjunctive also have at least traces of the s-Aorist

There might be a trace of the s-aorist in the Hittite 3sg preterite ending -s. Of course, this point is heavily debated: while pre-Anatolian PIE certainly didn't have a productive s-aorist, this ending might have been the precursor of the s-aorist.

3) AFAIK, there is at least one language family that has s-Aorist and -se/o- subjunctives, but not -e/o- subjunctives (Celtic)

The Old Irish f-future has been attempted to be explained as *-b^hw-e/o-, i.e. as subjunctives to the root *b^huH- (parallel to the Latin future).

So I think it's no accident that both both the s-Aorist and the -e/o- subjunctive are absent in Anatolian and Germanic - the interplay between s-Aorist and -se/o- prospectives / desideratives was a step in the development of the -e/o- subjunctive.

I hope (and suspect) you're not suggesting here that the e/o-subjunctive and s-aorist didn't exist in the ancestor pre-Germanic PIE as well. Germanic isn't exactly known for its archaic features! If anything, this would be a good argument that Anatolian lost the subjunctive: if Germanic can do it, why not Anatolian?

[Salmoneus]

The IE discussion has gone beyond my knowledge, but I'm very curious about the claim that "the Anatolian demi-god Melchert" tells us something about vowel alternations.

Who is this Melchert, and is his name Indo-European? Because it looks very much like the god normally known as Melqart, the Ba'al of Tyre - originally a full god, but in later years increasingly equated with the Greek demigod Herakles. Melqart, to my knowledge, is considered a typical Semitic god, possibly related etymologically to Moloch, Milcom, and other MLK gods. Melqart seems to have been around as a Semitic god since 1000BC at least. Are you suggesting that Melqart actually originates in Indo-European culture?

If not, how can a name presumably originally borrowed from an unknown Semitic language into an unknown Anatolian language and then again into whichever language it's attested in historically possibly tell us anything concrete about inherited vowel alternations in Anatolian?

[WeepingElf]

Who is this Melchert, and is his name Indo-European?

This guy

[R.Rusanov]

I reckon PIE had a lot of cases which collapsed into just a few in each branch, but with different endings contributing (if that makes sense).

Like bhagomos and bhagoyos might have been allative and essive or smt and one branch may have taken the one for its case and the other, the other.

[KathTheDragon]

A better solution to me would be a small number of core cases (at least nom, acc, voc and gen), with the remainder supplied by periphrastic constructions, which later underwent univerbation differently in each daughter. On that note, is it plausible to have synthetic singulars but periphrastic plurals?

[jal]

Ever tried EIEOL?

I hope it's more up-to-date on linguistics than on geopolitics: "Armenian is the official language of what in recent memory was the smallest republic of the former Soviet Union, now the southernmost republic of the Commonwealth of Independent States."

Who is this Melchert, and is his name Indo-European?

I'm pretty sure Sal is pulling your legs here :)


JAL

[hwhatting]

What is absent is the e-grade thematic present, which in other IE languages is the most far-spread form; but we have other thematic formations attested in Anatolian (zero-grade thematic, -sk'e/o-).

While there are forms in -e/a- in Hittite after a zero grade root, none of them match with certainty to tudáti-forms elsewhere (I'm using tudáti as short-hand for "oxytone zero grade simple thematic presents", and bhárati for "barytone e-grade simple thematic presents"). There are matches with statives in -(é)h₁-, and some are univerbations with -d^h(é)h₁-. The Anatolian demi-god Melchert seems to suggest this is the origin of the Hittite -e/a- conjugation, and there is no relation to the PIE simple thematic presents.

I'll have to look the "tudáti" type in Anatolian in more detail, then - my Information was just that the type was present. As "tudáti" is relatively rare in the other IE laguages as well, the lack of matches may just be an artifact of conservation of the type in the individual IE languages.

This would of that ALL simple thematic presents are absent in Anatolian, not just the bhárati-type. Which is doubly strange, as the derived thematic suffixes *-ye/o- and *-sk'e/o- are very common.

Indeed.

I don't think it very likely that the e-grade would vanish without a trace in Anatolian, while the "rarer" types remained, if it would have been already the "basic" type in IE when Anatolian branched off; I think it's more likey that the e-grade thematic class either did not exist or was still rare at that point, which means that it's a late development. I think that the original thematic inflection had the zero-grade, and that the e-type originally was formed from Narten roots (which have /e:/ grade in the full grade and /e/ in the zero grade); the thematic e-grade-present was then formed from non-Narten roots later in analogy.

All plausible, but what would be the starting point of adding -e/o- to athematic roots? What was it's function?

That's a whole another question, as they say. Looking at things like the Baltic 3rd Sg. in -a, Greek 3rd Sg. -e-i, and 1st. Sg. -o-H₂-(V), I assume that the thematic verbs originally belonged to the perfect / mediopassive complex, then acquired, at least partially, -mi endings, and what was originally part of the ending became re-interpreted a verbal stem (the thematic vowel). This is not an original idea, I've seen that argued elsewhere, just can't remember where.

Depends on the role of this 'subjunctive'. The handbooks seem to suggest that the PIE subjunctive was primarily a future tense, with some modalities like wish, demand etc. mixed in. In that light, I don't think a development from "I am going to carry" or (even better) "I am about to have carried" to "I am carrying" is that strange.
In this light, maybe the use of this 'future tense' was taboo in Anatolian, it was like tempting the gods. To be on the safe side, they deleted everything which looks like a future/subjunctive as well, including the bhárati type. The tudáti type might have gotten a free pass, as at this point it was indistinguishable from eh₁-statives.
(I am aware that all of this is just a train of thought, it's not even a hypothesis. I am also aware that this is a very weak counter point )

Indeed, I hope you don't mean this too seriously. As I said, I know of no examples where a future or subjunctive tense was re-purposed as a present tense (I don't talk about individual forms like Spanish eres here, but about shifts in usage for a category), but there are plenty of examples where a simple present became a subjunctive or future tense.

My assumption is that the late PIE had -se/o- presents that became used to express intended or prospective Actions; in some IE languages these formations have survived as prospective / future / desiderative tenses / moods; when formed from -H₁es- "to be" and from verbs that had s-aorists, the formation X-se/o- became re-interpreted as Xs-e/o- and then the newly formed -e/o- Suffix spread.

The precise form of this "desiderative" suffix is heavily debated: some think it was *-h₁s-, others think it's simply *-s-. Some think it was thematic, others (e.g. LIV) athematic. Some even think it didn't exist at all, and these were just s-aorist subjunctives (the reduplicated forms in Old Irish and Vedic would be a case of parallel development). I would be very hesistant to use this uncertain suffix as a basis for analogy, and there's grave danger of wishful thinking and circular reasoning.

I think there were several parallel formations - certainly it was not a well-formed category like later in Vedic; also you have things like the Greek future tense or Old Latin faxo etc. that don't fit into the Vedic pattern. I wouldn't put too much weight on what LIV reconstructs (as opposed to what they document), as they are clearly in the Rix-Cowgill tradition and try to fit the documented forms into that framework. One of the formations in the -se/o- group seems indeed to have been-H1se/o-, which, surprise, looks like a tudáti Formation from H₁es- "to be". Forms like this, or the –sye/o- formations, cannot be explained as subjunctives of s-aorists; what I assume that happened is that in those languages where s-aorists developed, the –se/o- forms and the s-aorists started to influence each other and the differences in stem-formation between were either eroded, making the –se/o- formation being seen as a sub-category of the s-aorist (“aorist subjunctive”), or, where the formation did not have a parallel s-aorist or was different in formation from the s-aorist, it became a separate category (e.g. Greek future tense).

There might be a trace of the s-aorist in the Hittite 3sg preterite ending -s. Of course, this point is heavily debated: while pre-Anatolian PIE certainly didn't have a productive s-aorist, this ending might have been the precursor of the s-aorist.

I think the latter interpretation of the facts is better – when Anatolian split off, there was no s-Aorist yet, only a form that served as a starting point.

3) AFAIK, there is at least one language family that has s-Aorist and -se/o- subjunctives, but not -e/o- subjunctives (Celtic)

The Old Irish f-future has been attempted to be explained as *-b^hw-e/o-, i.e. as subjunctives to the root *b^huH- (parallel to the Latin future).

Last time I looked (ca. 25 years ago), there were still doubt whether such a proto-form would lead to the outcomes attested in Old Irish, but let’s assume that’s really so. What we could have here as well as a subjunctive is a present tense formed from a root with a punctual meaning. That would actually be another pathway for the subjunctive to arise – present tense formations from punctual (aorist) roots that would, like Slavic present tenses of perfective verbs, have future / non-actual reference.
And look here: while browsing for some back-up on Latin faxo (about which I wasn’t sure any more whether it was attested) I found this paper, which argues exactly for such a pathway. Much more elegant than what I was trying for, and it gets rid of what bugs me most about the current Rixian consensus, namely, deriving thematic presents from the subjunctive, but, instead, like I prefer in line with general typology, arguing that the development was the other way round. I could still try to salvage my ideas about association between s-aorists and –se/o- formations by assuming that they played a supporting role in the rise of the subjunctive, but that's not so important.

I hope (and suspect) you're not suggesting here that the e/o-subjunctive and s-aorist didn't exist in the ancestor pre-Germanic PIE as well. Germanic isn't exactly known for its archaic features! If anything, this would be a good argument that Anatolian lost the subjunctive: if Germanic can do it, why not Anatolian?

Whether Germanic is archaic or not depends a lot on your model of PIE morphology. I have at least seen some trees where Germanic splits off after Anatolian and Tokharian; if Germanic indeed split off early, absence of features in Germanic can as well be a retention as an innovation. As long as the Brugmannian view of PIE prevailed, there was no choice but to assume that Germanic had lost a great amount of PIE features; integrating Anatolian into the model, we’re free to assume that things absent in Germanic never were there, as it split off before the full development of the Graeco-Aryan system. The classical assumption, e.g., is that Germanic replaced something like the Aorist by a new past tense based on the perfect, but do we have any evidence that Germanic ever underwent the step from aspectual root meanings and aspects expressed by variation of stem morphology to a formal split into a dual aorist and present system? If you derive the simple thematic present from the subjunctive, yes, then Germanic needs to have lost it; but if the subjunctive is originally a thematic present, Germanic never needs to have had it. Germanic certainly is nearer to Graeco-Aryan than Anatolian is – it has developed the IE perfect and the thematic present is the predominant formation. But we shouldn’t simply shrug off the absence of a feature in Germanic with a “probably lost” as long as the presence of a feature in PIE is not confirmed by Anatolian.

[Sleinad Flar]

Ever tried EIEOL?

I hope it's more up-to-date on linguistics than on geopolitics: "Armenian is the official language of what in recent memory was the smallest republic of the former Soviet Union, now the southernmost republic of the Commonwealth of Independent States."

Who is this Melchert, and is his name Indo-European?

I'm pretty sure Sal is pulling your legs here
JAL

Apparently the Commonwealth still exists, and Armenia is still a member. Besides, 1991 is pretty recent in my book.

Yes, I figured as much and indeed had a good laugh.

(Ninja-ed by hwhatting, more stuff to read!)

[Sleinad Flar]

(I'm going to answer to hwhatting in points here, otherwise this post will take up too much space with all its quotes):

1/2/3: I believe it was Jasanoff.
Another funny thing about the Hittite thematic presents is that they are mi-presents, not hi-presents as one would expect. Of couse the -mi part could have been simply added to original *-o-H > -a-.

4: Come to think of it, present > subjunctive is pretty common, e.g. in Tocharian, where subjunctive stems mostly correspond to present stems elsewhere (but sometimes also to root aorist subjunctives, e.g. *g^wem-e/o-). I was thinking of the reverse, and thought of Latin edo, optative edim (just like faxo, faxim), but this could simply be a case of thematicalisation.

5. Don't get me wrong, I believe the desiderative s-formations and the s-aorists are quite distinct as well. However, I don't think there's enough reason to posit -h₁s- (one has to explain why the laryngeal was lost after a plosive), let alone a relation with 'to be'. I also think the desiderative formation is athematic in origin, as relics like faxo, faxim and the si-imperatives show.

6. Agreed.

7. Excellent article.

8. The problem is, this would put Germanic even more basal in the tree than Tocharian, which does show the split present-aorist (with the aorist merging with the perfect to form the preterite, and even showing a pretty established s-aorist) and subjunctive (some of which corresponding to subjunctives elsewhere). Also problematic, the thematic optative corresponds with Graeco-Aryan (-o-i(h₁)-) instead of Italic (-a:-) and Tocharian (-ih₁-). The alternative would be to leave the tree model alone, and go to a wave model instead; some developments simply never reached to peripheral IE dialects. In such a model I'm willing to except that Germanic (which after Anatolian and Tocharian is as peripheral as it gets) never had the aorist/present-split.

[KathTheDragon]

1/2/3: I believe it was Jasanoff.
Another funny thing about the Hittite thematic presents is that they are mi-presents, not hi-presents as one would expect. Of couse the -mi part could have been simply added to original *-o-H > -a-.

Funnily enough, I'm currently reading through Jasanoff's Hittite and the Indo-European Verb, in which he presents his h₂e-conjugation theory. He alludes to a possible origin for thematic *-oH as *-o-h₂e, as thematisation of his original h₂e-conjugation verbs is the norm outside of Anatolian. The new thematic 1s suffix would then spread analogically to all other thematics. I don't think that a thematic 1s from *-oH-mi is tenable, as it would appear that the thematic vowel was actually in the e-grade before *-mi.

[Salmoneus]

Not only was I not joking, but I'm so ignorant about comparative middle-eastern religious scholarship that I don't even see what the joke was. Is a Melchert/Melqart connexion considered inherently ridiculous?

[Terra]

(1) In Greek, how common is the -s- perfect/aorist?
(2) In Greek, how common is the reduplicated perfect/aorist?

Today, I methodically went through Leiden's Italic/Latin Etymological dictionary, and counted 83 verbs with -s- perfects, and 26 with reduplicated perfects. I can list them all if anybody wants. I admit that I probably missed some, especially those that are verbs that are derived from nouns, because it's easy to check the verb's perfect form when it's the lemma and not buried halfway into the entry under a lemma that's a noun.

(I don't have Leiden's Greek Ety dictionary. If I did, I'd count them myself. If anybody has a pdf of it, I'd be greatly interested.)

[Sleinad Flar]

@KathAveara: It depends: it's -emi in verbs like utemi and the ske-verbs, but -ami in verbs like zinnami and the iye-verbs (sometimes also the ske-verbs).

@Sal: The way your post was written was pretty funny. But apologies if you didn't intend it that way.
PS The reason I'm calling Craig Melchert a demi-god is because he is the master in his field.

@Terra:
1. The s-aorist is the productive aorist formation. It isn't called the first aorist for naught.
2. Likewise, almost all perfects show reduplication, and most also add -k- in the perfect active. The reduplicated aorist is rare though.

(Is that Leiden dictionary the Beekes one: http://lib.freescienceengineering.org/v ... ?id=910115 ?)

[Terra]

(Is that Leiden dictionary the Beekes one: http://lib.freescienceengineering.org/v ... ?id=910115 ?)

Yes, I can't seem to download the file though. I click on the book's cover (a broken image), and then nothing happens.

[KathTheDragon]

@KathAveara: It depends: it's -emi in verbs like utemi and the ske-verbs, but -ami in verbs like zinnami and the iye-verbs (sometimes also the ske-verbs).

True, but didn't some instances of *e also become a? And there's nothing preventing the a from coming from *ŏ.

[Salmoneus]

Oh, sorry, completely misunderstood. I didn't realise YOU were being facetious when you called him a demi-god, and I didn't realise Elf WASN'T being facetious when he linked to a guy with that name.

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

[Sleinad Flar]

Yes, I can't seem to download the file though. I click on the book's cover (a broken image), and then nothing happens.

Hmm... worked fine when I made the link, but appears to be broken now.
Try this instead: http://www.filefactory.com/file/68oudfs9t789/n/EDoG_rar .

[Sleinad Flar]

True, but didn't some instances of *e also become a?

According to this only in some environments (enT, eRCC, eKsC and of course next to the appropiate laryngeal(s)).

And there's nothing preventing the a from coming from *ŏ.

True, long and short o merged anyways.

[Terra]

Yes, I can't seem to download the file though. I click on the book's cover (a broken image), and then nothing happens.

Hmm... worked fine when I made the link, but appears to be broken now.
Try this instead: http://www.filefactory.com/file/68oudfs9t789/n/EDoG_rar .

Wow, thank you very much! I've been looking for this. I have the Italic and Celtic ones if you want them.

[hwhatting]

Another funny thing about the Hittite thematic presents is that they are mi-presents, not hi-presents as one would expect. Of couse the -mi part could have been simply added to original *-o-H > -a-.

I think that simply different IE languages went different ways in merging the -o/e- paradigm with the -mi paradigm. Hittite -ami goes certainly back to *-o-mi, and there's also Vedic -a:mi, which is traditionally explained from *-oH-mi , but it could as well be from *-o-mi with Brugmann's law. In that case, we would have different degrees of "mi-fication" of the e/o paradigm, with Greek (and Baltic) at one end (3Sg. -e (-o), 1Sg. -o-H₂V), Anatolian and Vedic (at least partially) at the other end (-eti, -omi) and the other IE languages in the middle (-eti, but -o-H₂V). (Also thanks to Kath for making some of the same points.)

4: Come to think of it, present > subjunctive is pretty common, e.g. in Tocharian, where subjunctive stems mostly correspond to present stems elsewhere (but sometimes also to root aorist subjunctives, e.g. *g^wem-e/o-).

You see, the Tocharian verbal system is my big blind spot, I really need to start brushing up on that. I knew it had shifts from present to subjunctive, but don't know much about the details. What you describe looks to me like it simply did the shift from present to subjunctive wholesale, while other IE languages (Germanic, Balto-Slavic) didn't do it at all or (originally) split it along aspectual lines, with the present from aorist stems being used for future / hypothetical situations and the presents from non-aorist stems being used as indicative, which then later led to the rise of the subjunctive as a separate category and to the formation of subjunctives also from non-Aorist stems.

I was thinking of the reverse, and thought of Latin edo, optative edim (just like faxo, faxim), but this could simply be a case of thematicalisation.

That's what it is, I mean, edo is only half-thematic anyway.

5. Don't get me wrong, I believe the desiderative s-formations and the s-aorists are quite distinct as well. However, I don't think there's enough reason to posit -h₁s- (one has to explain why the laryngeal was lost after a plosive), let alone a relation with 'to be'. I also think the desiderative formation is athematic in origin, as relics like faxo, faxim and the si-imperatives show.

Well, as I said, I think there were several formations, including with and without *-h₁-, thematic and athematic, with and without -y-, some of which were differentiated by role in the individual IE languages , while in other languages only one type survived and in others they were eliminated entirely or never even became a separate categoy (E.g. in Anatolian).

8. The problem is, this would put Germanic even more basal in the tree than Tocharian, which does show the split present-aorist (with the aorist merging with the perfect to form the preterite, and even showing a pretty established s-aorist) and subjunctive (some of which corresponding to subjunctives elsewhere). Also problematic, the thematic optative corresponds with Graeco-Aryan (-o-i(h₁)-) instead of Italic (-a:-) and Tocharian (-ih₁-). The alternative would be to leave the tree model alone, and go to a wave model instead; some developments simply never reached to peripheral IE dialects. In such a model I'm willing to except that Germanic (which after Anatolian and Tocharian is as peripheral as it gets) never had the aorist/present-split.

In general, I'm "wavy" on this matter, and I certainly see Germanic more as a peripheral language that has not shared in some innovations than as one that had a Graeco-Aryan structure and than lost half of the features. On Italic - as Slavic also has the thematic optative in -o-i(h₁) and Italic is normally nearer to Graeco-Aryan than both Germanic and Slavic, it is possible that in this case, Italic has had that formation and replaced it. But it's also possible that absence of the -o-i(h₁) thematic optative is a retention and that not adopting it would be one of the shared features of Tokharian and Italic (besides r-mediopassives and the merging of aorist and perfect into a past tense). I don't remember right now - what is the attestation of the -o-i(h₁) optative in Celtic?

[vec]

Oh, sorry, completely misunderstood. I didn't realise YOU were being facetious when you called him a demi-god, and I didn't realise Elf WASN'T being facetious when he linked to a guy with that name.

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

Hahahaha this whole episode is incredibly funny

[Sleinad Flar]

Wow, thank you very much! I've been looking for this. I have the Italic and Celtic ones if you want them.

De Vaan and Matasovic? Already got 'em, but thanks anyway.

In general, I'm "wavy" on this matter, and I certainly see Germanic more as a peripheral language that has not shared in some innovations than as one that had a Graeco-Aryan structure and than lost half of the features. On Italic - as Slavic also has the thematic optative in -o-i(h₁) and Italic is normally nearer to Graeco-Aryan than both Germanic and Slavic, it is possible that in this case, Italic has had that formation and replaced it. But it's also possible that absence of the -o-i(h₁) thematic optative is a retention and that not adopting it would be one of the shared features of Tokharian and Italic (besides r-mediopassives and the merging of aorist and perfect into a past tense). I don't remember right now - what is the attestation of the -o-i(h₁) optative in Celtic?

Pretty non-existent. The athematic optative in -ih₁- is very weakly attested as well (one possible form in Gaulish IIRC). Anyways, the thematic optative is clearly secondary to the athematic one (just like the thematic subjunctive in -e-e-/-o-o- is secondary to the simple subjunctive in -e-/-o-), which pretty much points to an innovation, esp. since no optatives survive in Anatolian.
Slavic peripheral? It's about as "core" as it gets IMO, both geographically and in shared features with Greek and especially Indo-Iranian.
(The Tocharian conjugational system is AWESOME by the way; almost as insane as the Old Irish one. Check it out here when you have time to spare)

Hahahaha this whole episode is incredibly funny

IKR?

[KathTheDragon]

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.

[KathTheDragon]

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