The Great Proto-Indo-European Thread's Sequel

Natural languages and linguistics
User avatar
Znex
Posts: 161
Joined: Wed Aug 01, 2018 10:59 pm

Re: The Great Proto-Indo-European Thread's Sequel

Post by Znex »

I don't know if this has been discussed before, but I'm currently looking at material hypothesising two pre-PIE changes (although the main focus is on the second) that may explain a number of semantic pairs in the IE lexicon, namely:
1. There was an early three-way split of labiovelars, similar to the later change in Pre-Greek.
2. There was an early simplification of initial clusters containing reconstructed initial laryngeals.

Some pairs suggested to be cognate as a result that stand out to me include:
  • *h₁egʷʰ- {to drink} and *píbh₃- {to drink} < **ph₁egʷh₃-
  • *h₁eḱw/u- {horse} and *peḱu- {cattle} < **ph₁eḱu-
  • *h₂enh₁- {to breathe} and *pnew- {to breathe} < **ph₂(e)nh₁-
  • *ḱwóns = *ḱwṓ {dog} and *wĺ̥kʷos {wolf} < **ḱh₃we-l/n-
  • *h₂ṓws~h₂éws- {ear} and *ḱlew(s)- {to hear} < **ḱlh₂éw-
  • *h₃nṓgʷʰs {nail} and *ǵónu {knee} < **ǵh₃énu, ǵh₃n-o/e-gʷH-
  • *h₃nṓmn~h₃némn-~h₃nmén- {name} and *ǵen- {know} < **ǵh₃n-, ǵh₃n-o/e-mn
  • *h₂rmos {arm} and *bʰer- {to bear,carry} < **bh₂r-, bh₂r-mos
  • *h₂erǵ- {white,bright} and *bʰerHǵós {birch}, *bʰerHǵ- {bright,shine} < **bh₂erǵ-
  • *h₂egʷʰ- {snake,creature} and *dʰéǵʰōm {earth} < **dh₂éǵHu-
Do you follow the reasoning in the article? What is your opinion on the matter?
User avatar
KathTheDragon
Posts: 783
Joined: Mon Jul 09, 2018 3:57 am
Location: Disunited Kingdom

Re: The Great Proto-Indo-European Thread's Sequel

Post by KathTheDragon »

There's an awful lot to talk about, and I'm not really awake enough to do so right now, so I'll sum up: there's little to no merit to this.
User avatar
WeepingElf
Posts: 1511
Joined: Sun Jul 15, 2018 12:39 pm
Location: Braunschweig, Germany
Contact:

Re: The Great Proto-Indo-European Thread's Sequel

Post by WeepingElf »

This reminds me of some of Starostin's "North Caucasian" etymologies Octaviano liked to cite: if two segments can't be made to match by any means, just posit a protoform that contains both, with one language deleting one and the other deleting the other. By such brute force, just about everything can be made to "match". Very convincing ;)
... brought to you by the Weeping Elf
My conlang pages
User avatar
KathTheDragon
Posts: 783
Joined: Mon Jul 09, 2018 3:57 am
Location: Disunited Kingdom

Re: The Great Proto-Indo-European Thread's Sequel

Post by KathTheDragon »

Having looked through it in more detail, that's an apt summary, all the more so since there's no attempt made to explain why any given form ends up the way it does in the daughters.
Kuchigakatai
Posts: 1307
Joined: Mon Jul 09, 2018 4:19 pm

Re: The Great Proto-Indo-European Thread's Sequel

Post by Kuchigakatai »

WeepingElf wrote: Tue Aug 20, 2019 3:48 pm This reminds me of some of Starostin's "North Caucasian" etymologies Octaviano liked to cite: if two segments can't be made to match by any means, just posit a protoform that contains both, with one language deleting one and the other deleting the other. By such brute force, just about everything can be made to "match". Very convincing ;)
If I could go back to the time when Octaviano was here, I would have a field day trolling him with etymologies of supposed Nostratic cognates going back to Zompist's Proto-Eastern.

For example, one day Octaviano mentioned that (as he reconstructed them) PIE *sol-wo- 'healthy, whole' (Lat. salvus, Grk. ὅλος) and PIE *h₂ʷolyo- (Germanic *allaz, Celtic *olyo-) were related to (as he reconstructs it) "Proto-North-Caucasian" *dzzo(w)ɫV. Surely all these simply reflect Proto-Eastern sōl 'strong', and are therefore cognates with Cuêzi sōle 'true' and Cad'inor zol 'strong'!
mae
Posts: 115
Joined: Sun Sep 09, 2018 11:00 am

Re: The Great Proto-Indo-European Thread's Sequel

Post by mae »

-
Last edited by mae on Wed Oct 16, 2019 10:54 am, edited 1 time in total.
User avatar
Znex
Posts: 161
Joined: Wed Aug 01, 2018 10:59 pm

Re: The Great Proto-Indo-European Thread's Sequel

Post by Znex »

mae wrote: Tue Aug 20, 2019 11:15 pm Not sure who Dziebel is but this is equally if not more ad hoc as it requires *gwh > b in branches where (some) labiovelars preserve their velarity, such as Skt., Armenian, and Celtic.
I was looking up Dziebel, and his arguments are more indefensible (see the comments). It's clear he's not really familiar with the IE language changes (eg. he supposes that satem changes occur already in PIE though the evidence relies on correspondences between *ḱ and *s; *ḱ is demonstrably distinct from *s in early satemic attestations).
Salmoneus
Posts: 1057
Joined: Thu Jul 26, 2018 1:48 pm

Re: The Great Proto-Indo-European Thread's Sequel

Post by Salmoneus »

Znex wrote: Tue Aug 20, 2019 11:47 pm
mae wrote: Tue Aug 20, 2019 11:15 pm Not sure who Dziebel is but this is equally if not more ad hoc as it requires *gwh > b in branches where (some) labiovelars preserve their velarity, such as Skt., Armenian, and Celtic.
I was looking up Dziebel, and his arguments are more indefensible (see the comments). It's clear he's not really familiar with the IE language changes (eg. he supposes that satem changes occur already in PIE though the evidence relies on correspondences between *ḱ and *s; *ḱ is demonstrably distinct from *s in early satemic attestations).
I must admit, "Proto-Indo-European was obviously an Iroquoisan-Siouxan creole spoken in the Pacific Northwest" is a new one to me. I think it deserves to win some sort of prize...
zompist
Site Admin
Posts: 2944
Joined: Sun Jul 08, 2018 5:46 am
Location: Right here, probably
Contact:

Re: The Great Proto-Indo-European Thread's Sequel

Post by zompist »

Ser wrote: Tue Aug 20, 2019 10:07 pm
WeepingElf wrote: Tue Aug 20, 2019 3:48 pm This reminds me of some of Starostin's "North Caucasian" etymologies Octaviano liked to cite: if two segments can't be made to match by any means, just posit a protoform that contains both, with one language deleting one and the other deleting the other. By such brute force, just about everything can be made to "match". Very convincing ;)
If I could go back to the time when Octaviano was here, I would have a field day trolling him with etymologies of supposed Nostratic cognates going back to Zompist's Proto-Eastern.

For example, one day Octaviano mentioned that (as he reconstructed them) PIE *sol-wo- 'healthy, whole' (Lat. salvus, Grk. ὅλος) and PIE *h₂ʷolyo- (Germanic *allaz, Celtic *olyo-) were related to (as he reconstructs it) "Proto-North-Caucasian" *dzzo(w)ɫV. Surely all these simply reflect Proto-Eastern sōl 'strong', and are therefore cognates with Cuêzi sōle 'true' and Cad'inor zol 'strong'!
6000 BCE magical portal confirmed!
User avatar
WeepingElf
Posts: 1511
Joined: Sun Jul 15, 2018 12:39 pm
Location: Braunschweig, Germany
Contact:

Re: The Great Proto-Indo-European Thread's Sequel

Post by WeepingElf »

Oh, things were fun when Octaviano was around here - we had a lot to laugh ;) But he seems to have calmed down a bit and no longer pursue his ideas: he didn't post to his blog since November 2018, and according to this FrathWiki page he no longer adheres to the Vasco-Caucasian hypothesis, but his user page on FrathWiki seems to show that he is still quite far in the loops.
... brought to you by the Weeping Elf
My conlang pages
Richard W
Posts: 1471
Joined: Sat Aug 11, 2018 12:53 pm

Re: The Great Proto-Indo-European Thread's Sequel

Post by Richard W »

There is a related hypothesis, that PIE had *pʷ etc. which mostly became labiovelars, but sometimes plain labials. The usual riposte to that is what we see is just spontaneous changes of labiovelars to labials.
User avatar
dhok
Posts: 298
Joined: Mon Aug 20, 2018 4:39 am
Location: The Eastern Establishment

Re: The Great Proto-Indo-European Thread's Sequel

Post by dhok »

Or perhaps vice-versa; *h₂ep- 'water' is attested on most of the periphery and in the most archaic branches (Hittite, Tocharian, Indo-Iranian) while *h₂ekʷ- is only attested in Europe (in Germanic, Latin and perhaps Lusitanic and Slavic.) I think it's consensus that they're probably variants of each other.

Italic is supposed to preserve both variants, the latter in Latin aqua and the former in Oscan, which is supposed to continue the p-variant for Reasons even though *p and *kʷ of course fall together...but there is also Latin amnis (not **agnis), so both roots are continued in Latin.

Celtic has *abū (OIr. aub/ab/ob, Welsh afon), an n-stem supposed to be from *h₂ep-hɜen-, with píbati-voicing. But that could easily be from *h₂ekʷ-, given that PIE *gʷ becomes /b/ in Celtic (cf. OIr. ben 'woman' < *gʷenh₂).

I don't think anybody believes *h₁egʷʰ- is related to the above, though. Laryngeals don't mutate into each other, and there's no reconstructible *T~*Dʰ alternation either.
User avatar
KathTheDragon
Posts: 783
Joined: Mon Jul 09, 2018 3:57 am
Location: Disunited Kingdom

Re: The Great Proto-Indo-European Thread's Sequel

Post by KathTheDragon »

dhok wrote: Thu Aug 22, 2019 8:49 pmI think it's consensus that they're probably variants of each other.
On the other hand, I came across this paper quite a while back, which presents an entirely other line of analysis, taking "*h₂ekʷ-" to be rather *h₁eh₁kʷ-, for which it finds quite a number of words across the family to support it.

It's also important to mention that Hittite ḫapa- cannot be from *h₂ep- with voiceless *p, as we'd expect a geminate pp in Hittite, and it's doubtful that any of the lenition laws can be invoked. So with a root *h₂ebʰ-, the Celtic is easy to explain without any doubtful voicing effects of *h₃, and also Latin amnis can be included.
User avatar
Howl
Posts: 56
Joined: Sun Jul 15, 2018 4:39 am

Re: The Great Proto-Indo-European Thread's Sequel

Post by Howl »

dhok wrote: Thu Aug 22, 2019 8:49 pm I don't think anybody believes *h₁egʷʰ- is related to the above, though. Laryngeals don't mutate into each other, and there's no reconstructible *T~*Dʰ alternation either.
KathTheDragon wrote: Sat Aug 24, 2019 4:15 am On the other hand, I came across this paper quite a while back, which presents an entirely other line of analysis, taking "*h₂ekʷ-" to be rather *h₁eh₁kʷ-, for which it finds quite a number of words across the family to support it.
And *h₁eh₁kʷ/*h₁egʷʰ would not be the only *T~*Dʰ doublet. See also this paper on possible consonant alternations in PIE. The problem with these consonant alternations is that we do not know if this is a real thing, and if so, what is going on.
User avatar
WeepingElf
Posts: 1511
Joined: Sun Jul 15, 2018 12:39 pm
Location: Braunschweig, Germany
Contact:

Re: The Great Proto-Indo-European Thread's Sequel

Post by WeepingElf »

Howl wrote: Sat Aug 24, 2019 6:49 amAnd *h₁eh₁kʷ/*h₁egʷʰ would not be the only *T~*Dʰ doublet. See also this paper on possible consonant alternations in PIE. The problem with these consonant alternations is that we do not know if this is a real thing, and if so, what is going on.
Interesting paper. I have been entertaining, for some time, the idea that there was a split of a single grade of stops, corresponding to the single stop grade of Proto-Uralic, into the two grades that we now know as "voiceless" and "voiced-aspirated" due to some sort of prosodic feature that bound to morphemes, which explains why "voiceless" and "voiced-aspirated" stops never co-occur in a root. Under such an assumption, there may be related words with different values of that prosodic features, which result in just such kinds of alternations.
... brought to you by the Weeping Elf
My conlang pages
Nortaneous
Posts: 1660
Joined: Sun Jul 15, 2018 3:29 am

Re: The Great Proto-Indo-European Thread's Sequel

Post by Nortaneous »

WeepingElf wrote: Sat Aug 24, 2019 7:29 am I have been entertaining, for some time, the idea that there was a split of a single grade of stops, corresponding to the single stop grade of Proto-Uralic, into the two grades that we now know as "voiceless" and "voiced-aspirated" due to some sort of prosodic feature that bound to morphemes, which explains why "voiceless" and "voiced-aspirated" stops never co-occur in a root. Under such an assumption, there may be related words with different values of that prosodic features, which result in just such kinds of alternations.
cf. Yabem, where tone is a property of the disyllabic foot and stop voicing is conditioned by tone, so all the plosives in a foot must agree in voicing: kó-tóŋ 'you give' vs. gò-ŋgòŋ 'you sit'
Last edited by Nortaneous on Sat Aug 24, 2019 8:01 am, edited 1 time in total.
Duaj teibohnggoe kyoe' quaqtoeq lucj lhaj k'yoejdej noeyn tucj.
K'yoejdaq fohm q'ujdoe duaj teibohnggoen dlehq lucj.
Teijp'vq. Teijp'vq. Teijp'vq. Teijp'vq. Teijp'vq. Teijp'vq. Teijp'vq.
User avatar
Howl
Posts: 56
Joined: Sun Jul 15, 2018 4:39 am

Re: The Great Proto-Indo-European Thread's Sequel

Post by Howl »

WeepingElf wrote: Sat Aug 24, 2019 7:29 am I have been entertaining, for some time, the idea that there was a split of a single grade of stops, corresponding to the single stop grade of Proto-Uralic, into the two grades that we now know as "voiceless" and "voiced-aspirated" due to some sort of prosodic feature that bound to morphemes, which explains why "voiceless" and "voiced-aspirated" stops never co-occur in a root. Under such an assumption, there may be related words with different values of that prosodic features, which result in just such kinds of alternations.
As you probably already know, Kortlandt has proposed something very similar in his paper on Indo-Uralic Consonant Gradation. It's a pity that his paper is very light on details. Also, the paper repeats the "Vowel Collapse" theory which I now think is total bullshit.
Last edited by Howl on Sat Aug 24, 2019 8:21 am, edited 1 time in total.
User avatar
WeepingElf
Posts: 1511
Joined: Sun Jul 15, 2018 12:39 pm
Location: Braunschweig, Germany
Contact:

Re: The Great Proto-Indo-European Thread's Sequel

Post by WeepingElf »

Howl wrote: Sat Aug 24, 2019 7:56 am
WeepingElf wrote: Sat Aug 24, 2019 7:29 am I have been entertaining, for some time, the idea that there was a split of a single grade of stops, corresponding to the single stop grade of Proto-Uralic, into the two grades that we now know as "voiceless" and "voiced-aspirated" due to some sort of prosodic feature that bound to morphemes, which explains why "voiceless" and "voiced-aspirated" stops never co-occur in a root. Under such an assumption, there may be related words with different values of that prosodic features, which result in just such kinds of alternations.
As you probably already know, Kortlandt has proposed something very similar in his paper on Indo-Uralic Consonant Gradation. It's a pity that his paper is very light on details.
Yes, I am aware that others have had the idea before me, and I am also aware of this paper in particular. The details need to be worked out, though.
... brought to you by the Weeping Elf
My conlang pages
User avatar
KathTheDragon
Posts: 783
Joined: Mon Jul 09, 2018 3:57 am
Location: Disunited Kingdom

Re: The Great Proto-Indo-European Thread's Sequel

Post by KathTheDragon »

I think in any case these alternations predate PIE far enough that it's not easy nor necessary to investigate them in depth yet. There are more pressing matters than a number of roots that look alike.
User avatar
WeepingElf
Posts: 1511
Joined: Sun Jul 15, 2018 12:39 pm
Location: Braunschweig, Germany
Contact:

Re: The Great Proto-Indo-European Thread's Sequel

Post by WeepingElf »

KathTheDragon wrote: Sat Aug 24, 2019 9:50 am I think in any case these alternations predate PIE far enough that it's not easy nor necessary to investigate them in depth yet. There are more pressing matters than a number of roots that look alike.
Sure. These alternations, if that's what they are, would date back to a stage before even the separation of Late PIE and Anatolian. Clearing this matter up may require external comparison - which throws up the question: Compare to what? Uralic seems to be the best choice, but there are probably not many cognate pairs to start with; and then there is the question where the two differ, which one was the innovative one - which would need even further external comparisons, leading us into the vagaries of Mitian, Eurasiatic, Nostratic or whatever kind of macrofamily. What may perhaps help would be the assessment of correlations between this "alternation" and other such things as accent/ablaut classes. Necessary to the comparison of living and attested extinct IE languages this is not; but is the question "How did PIE become as it is?" illegitimate? Certainly not! This is how science works: each answer to a question throws up further questions.
... brought to you by the Weeping Elf
My conlang pages
Post Reply