The Great Proto-Indo-European Thread

[Salmoneus]

Spoiler: it's from pre-Greek.

Doesn't matter what word you're looking up, it's always from pre-greek. And there's nothing about sound changes to greek, but there is a long chapter all about pre-greek.

[hwhatting]

Spoiler: it's from pre-Greek.

Doesn't matter what word you're looking up, it's always from pre-greek. And there's nothing about sound changes to greek, but there is a long chapter all about pre-greek.

Actually, before Beekes Ancient Greek already had two fully serviceable etymological dictionaries (Frisk and Chantraine). My impression is that Beekes mostly put out his own etymological dictionary to ride his Pre-Greek hobby horse.

[masako]

Another interesting article.

[kanejam]

Another interesting article.

[...] includes English, Spanish, French, Greek, Russian and Hindu?

[jal]

[...] includes English, Spanish, French, Greek, Russian and Hindu?

Probably an innocent typo (i next to u, Hindu existing word so no squigly red lines). Earlier in the article it correctly states "Hindi". I personally find this sentences the most troublesome: "There is the well known Indo-European language tree – ranging from Hindi to Russian to Spanish. But it’s also quite unclear how the languages spread to their present regions."


JAL

[KathTheDragon]

I've had a thought about a potential origin for the i/e-stem pronouns, whose ablaut is rather unprecidented and unparalleled anywhere in PIE. To me, that strongly suggests that it's not a regular development, but rather the product of analogy. The attested forms are as follows (using the interrogative stem *kʷi/e-, and disregarding the feminines and neuter plural):

Code: Select all

    m.sg   n.sg    pl
nom kʷis ~ kʷit  kʷeyes
acc kʷim ~ kʷit  kʷins
gen   kʷesyo     kʷeysōm
obl   kʷesm-     kʷey-

This is strongly reminiscent of i-stem nominal inflection, although the *-y- element found in the plural probably is a plural marker of some sort. Given that, we'd expect the acc.pl to have been **kʷeyns, but it's hardly surprising that we find an apparently i-stem form instead, probably based on the nom.pl and gen.pl forms, which do show the expected stem *kʷe-y-, but having been reanalysed as the full-grade forms of a stem *kʷi-. This new stem can easily have spread to the acc.sg, replacing original *kʷem ~ *kʷet, and from there to the nom.sg, replacing original *kʷes ~ *kʷet, which are all formed from the stem *kʷe-, still found in the genitive and oblique singular cases (apparently because i-stems have full grade here, too).
So, we can trace the i/e-stem inflection back to an older e-stem pronominal inflection, whence:

Code: Select all

    m.sg   n.sg    pl
nom kʷes ~ kʷet  kʷeyes
acc kʷem ~ kʷet  kʷeyns
gen   kʷesyo     kʷeysōm
obl   kʷesm-     kʷey-

Furthermore, this allows us to take the adjectival stem *kʷo- to just be the o-grade form of the stem.

Thoughts? Does this make sense, or am I spouting gibberish?

[Terra]

Spoiler: it's from pre-Greek.

Doesn't matter what word you're looking up, it's always from pre-greek. And there's nothing about sound changes to greek, but there is a long chapter all about pre-greek.

Yes, Beekes seems quite eager to declare a word as pre-Greek. Every-other word I look up seems to be pre-Greek.

***

Lately, I've been thinking about PIE's (and pre-PIE's) vowels, I want to see if I understand them correctly, so please correct me if I'm wrong:

At some point, pre-PIE had the GVC (great vowel collapse), which collapsed most mid vowels into /a/, thus reducing the system to only 3 vowels: /a i u/. It then turned stressed /a/ into stressed /e/, and unstressed /e/ into unstressed /o/. Unstressed /u/ became /o/. (Did unstressed /i/ became unstressed /e/?) Stressed /o/ was then added by analogy later. Where did stress come from in the first place though?

[KathTheDragon]

Lately, I've been thinking about PIE's (and pre-PIE's) vowels, I want to see if I understand them correctly, so please correct me if I'm wrong:

At some point, pre-PIE had the GVC (great vowel collapse), which collapsed most mid vowels into /a/, thus reducing the system to only 3 vowels: /a i u/. It then turned stressed /a/ into stressed /e/, and unstressed /e/ into unstressed /o/. Unstressed /u/ became /o/. (Did unstressed /i/ became unstressed /e/?) Stressed /o/ was then added by analogy later. Where did stress come from in the first place though?

According to (I think) WeepingElf, yes. I believe that's more or less how their theory runs.

[hwhatting]

@ KathAveara
I think that any theory that explains the singular forms of the interrogative pronouns by analogy to the plural froms is a dead end. The singular forms are much more salient, many IE languages don't even have plural forms, and the plural forms you cite can easily be explained as formed based on k^wis and k^wid in analogy with the i-stems.
Both in PIE and in the daughter languages, one frequently has "irregular" forms for the nominative and the oblique cases formed on analogy of more "regular" declinations. That seems to be the case here as well - the NOM and ACC are formed with an element *-i added to the original interrogative / relational element *k^w-, and the oblique cases are built on the *-es- element that is also found in other pronouns. See also the pronoun *is / *id, which has a similar structure.
On the whole, I think the declencion of *k^wi is quite young, and arose only when the indeclinable clitic was tuned into a full pronoun. IMO, this happened when ablaut already had become a morphological tool. The development was that -s, -m, and -d were added in the nominative, then the oblique stem was created on the analogy of *i/e-, and then some of the daughter languages created femal and plural forms based on analogy with the i-stems.

[WeepingElf]

Lately, I've been thinking about PIE's (and pre-PIE's) vowels, I want to see if I understand them correctly, so please correct me if I'm wrong:

At some point, pre-PIE had the GVC (great vowel collapse), which collapsed most mid vowels into /a/, thus reducing the system to only 3 vowels: /a i u/. It then turned stressed /a/ into stressed /e/, and unstressed /e/ into unstressed /o/. Unstressed /u/ became /o/. (Did unstressed /i/ became unstressed /e/?) Stressed /o/ was then added by analogy later. Where did stress come from in the first place though?

According to (I think) WeepingElf, yes. I believe that's more or less how their theory runs.

Yes, except for the "unstressed /u/ becomes /o/ and unstressed /i/ becomes /e/" bit. That did not happen, otherwise we'd have i~e and u~o ablaut, which is not observed! Also, not all unstressed /a/ became /o/ - many simply disappeared (zero grade). I still haven't found out when unstressed /a/ became /o/ and when it became zero.

As for /i/ and /u/: Where /a/ became /e/, they became /ei/ and /eu/, respectively, where /a/ became /o/, they became /oi/ and /ou/, and where /a/ went to zero, they remained /i/ and /u/.

The stress was penultimate in Pre-PIE, I think, but I have to admit that my idea leaves a lot of things unexplained, so it is not the whole story.

[KathTheDragon]

@ KathAveara
I think that any theory that explains the singular forms of the interrogative pronouns by analogy to the plural froms is a dead end. The singular forms are much more salient, many IE languages don't even have plural forms, and the plural forms you cite can easily be explained as formed based on k^wis and k^wid in analogy with the i-stems.

Makes sense, I suppose.

[sirdanilot]

So now people also assume the existence of 'pre-PIE', a single monolithic language entity spoken as-is with PIE as its only descending daughter 'language'?

When will it ever end...

[jal]

So now people also assume the existence of 'pre-PIE', a single monolithic language entity spoken as-is with PIE as its only descending daughter 'language'?

The whole reason for 'proposing' pre-PIE is that PIE was not a monolythic language. So depending on where you draw the line between dialect and language, pre-PIE is either the actual PIE, or pre-PIE is the pre-dialectal stage (if there ever was any - your previous arguments against such a posibility had some merit, though not quite in the way you went about it).

When will it ever end...

With pre-proto-world, obviously.


JAL

[sirdanilot]

Okay, that clarifies.

I do think people in PIE studies are slowly beginning to realize the non-monolithic nature of what they are studying, slowly but surely. As long as you see pre-PIE as a construct for historical comparison and reconstruction rather than as an actual monolithic spoken language at some point in time.

I think the difference in looking at these things is also because we already know what languages are indo-european and we have so many sources on them. If you do historical linguistics in South America it's an entirely different beast. Even of many living languages we have very little information, and this is even more true of the gazillions of dead languages. Any indo-europeanist would be completely lost if they had to do historical linguistics in south america.

[WeepingElf]

So now people also assume the existence of 'pre-PIE', a single monolithic language entity spoken as-is with PIE as its only descending daughter 'language'?

The whole reason for 'proposing' pre-PIE is that PIE was not a monolythic language. So depending on where you draw the line between dialect and language, pre-PIE is either the actual PIE, or pre-PIE is the pre-dialectal stage (if there ever was any - your previous arguments against such a posibility had some merit, though not quite in the way you went about it).

Another reason, of course, is finding out what else it is related to. Many scholars entertain the hypothesis that Uralic is the closest living kin of IE, so one would expect Pre-PIE to be more Uralic-like than PIE proper is. Considering what we see in the work of Rasmussen, Kortlandt and others, this seems to be the case, so Indo-Uralic at least is a plausible idea.

When will it ever end...

With pre-proto-world, obviously.

Yep

[Dewrad]

Any indo-europeanist would be completely lost if they had to do historical linguistics in south america.

While most linguists specialise in one field or another, there are several not insignificant historical linguists who have made valuable contributions both to South American historical linguistics and Indo-European historical linguistics. Lyle Campbell immediately comes to mind (also has contributed to Uralic historical linguistics).

[sirdanilot]

Yes Campbell is great. It is not impossible, of course, but historical linguistics in South America is certainly a different beast from IE because there is so little information compared to IE.

[WeepingElf]

Yes Campbell is great. It is not impossible, of course, but historical linguistics in South America is certainly a different beast from IE because there is so little information compared to IE.

One needs to keep in mind that IE is exceptional in a specific way: there are extensive ancient written records of several IE languages that bridge about half the time elapsed since the assumed break-up of PIE. This makes comparison and reconstruction as easy as if IE was not 6,000 but only about 3,000 years deep. And these ancient languages are indeed better known than many modern languages spoken in such remote places as Amazonia or New Guinea. The only other language family which is alike IE in this way is Semitic.

There may be language families in some parts of the world which are no deeper than IE but still yet undetected - simply because we know their members less well than Vedic or Hittite!

[Terra]

Yes, except for the "unstressed /u/ becomes /o/ and unstressed /i/ becomes /e/" bit. That did not happen, otherwise we'd have i~e and u~o ablaut, which is not observed! Also, not all unstressed /a/ became /o/ - many simply disappeared (zero grade). I still haven't found out when unstressed /a/ became /o/ and when it became zero.

Indeed, I'd like to know what determined whether the unstressed /a/ became 0 or /o/.

As for /i/ and /u/: Where /a/ became /e/, they became /ei/ and /eu/, respectively, where /a/ became /o/, they became /oi/ and /ou/, and where /a/ went to zero, they remained /i/ and /u/.

Again, I suppose there's no way to tell (from looking at just PIE, at least) whether an instance of /i/ goes back to /i/ or unstressed /ai/?

The stress was penultimate in Pre-PIE, I think, but I have to admit that my idea leaves a lot of things unexplained, so it is not the whole story.

Do you have an idea/theory/explanation for what caused a Pre-PIE plosive to become (using traditional terminology) unvoiced, voiced, or voiced-aspirated?

[Tropylium]

Any indo-europeanist would be completely lost if they had to do historical linguistics in south america.

While most linguists specialise in one field or another, there are several not insignificant historical linguists who have made valuable contributions both to South American historical linguistics and Indo-European historical linguistics. Lyle Campbell immediately comes to mind (also has contributed to Uralic historical linguistics).

FWIW I'm not sure if Campbell has "contributed to" historical Uralistics as much as just reviewed some of the results in English.

---

Is anyone aware of any good etymological overviews of PIE lexicon, btw? I have various handbooks, e.g. Mallory & Adams', and they certainly often list plenty of roots. But I'd be interested in seeing something that also analyzes roots in terms of their distribution. How do the numbers of "Western" or "West-Central" or "Indo-Greek" roots stack against roots that are found in at least most of the large and well-known subfamilies (Celtic, Italic, Germanic, Greek, Balto-Slavic, Indo-Iranian)? Or how much and what parts of the widely spread lexicon can be traced to Anatolian and Tocharian?

[Tropylium]

As for /i/ and /u/: Where /a/ became /e/, they became /ei/ and /eu/, respectively, where /a/ became /o/, they became /oi/ and /ou/, and where /a/ went to zero, they remained /i/ and /u/.

Again, I suppose there's no way to tell (from looking at just PIE, at least) whether an instance of /i/ goes back to /i/ or unstressed /ai/?

How much evidence is there for unstressed diphthongs in pre-PIE at all? A lot of the candidates for relatives of PIE (the entire Uralo-Yukaghir-Altaic belt; at least CK is more complicated; I don't know enough about EA, Dravidian and Karvelian) seem to do this thing with current or earlier initial stress, and more syllable types allowed initially, that is fairly typical for predominantly suffixing languages. These also often do not distinguish diphthongs from vowel + glide.

[hwhatting]

Is anyone aware of any good etymological overviews of PIE lexicon, btw? I have various handbooks, e.g. Mallory & Adams', and they certainly often list plenty of roots. But I'd be interested in seeing something that also analyzes roots in terms of their distribution. How do the numbers of "Western" or "West-Central" or "Indo-Greek" roots stack against roots that are found in at least most of the large and well-known subfamilies (Celtic, Italic, Germanic, Greek, Balto-Slavic, Indo-Iranian)? Or how much and what parts of the widely spread lexicon can be traced to Anatolian and Tocharian?

This seems to be work that still needs to be done... M&A, unfortunately, normally don't give full attestations for their roots. LIV and NIL do, but only contain subgroups of the roots. When will we get an update of Pokorny?

[Pabappa]

A bit offtopic I just wanted to mention that one of my newest conlangs is going to have /pʷ p t k kʷ/ and /d g gʷ/ with no /b/ or /bʷ/ and it made me think of PIE. The reason is that it gots its voiced stops from previously existing voiced fricatives, but the parent language's labial voiced fricatives were so weak that they turned into /ʕ/ and /w/ instead of something like /v/ and /vʷ/,. This could explain the lack of /b/ in PIE in a way perhaps a bit more elegant than simply assuming it unconditionally changed to /w/ .... perhaps there was a time when the plain voiced stops were all fricatives, and then they changed to stops (possibly also to laryngeals in some conditions), except that by this time /v/ had already merged to /w/, so there was no labial to produce a new /b/.

[R.Rusanov]

That's not more elegant

[jal]

That's not more elegant

Yeah, and as we all know, that's a good benchmark for anything PIE related...

JAL