Reconstruction talk:Proto-Indo-European/h₂eryós
Latest comment: 10 years ago by Florian Blaschke
I'm not sure we should be listing roots reconstructed only according to fringe views such as the existence of "h₄" and the possibility of a PIE root derived from Proto-Afro-Asiatic. The refs listed seem to be from non-linguists. —Angr 11:20, 10 May 2013 (UTC)
- There's nothing fringe about these sources; they're all academically published and in the case of TOitPIEatPIEw it is co-written by Douglas Adams who's an Indo-European comparativist. What's fringe about "h₄"? It's right there in both TOitPIEatPIEw and THtWaLHBARftESStMW. The book cited for PAA is written by linguists too, and since the validity of Nostratic is less certain than proto-Afro-Asiatic, it's perhaps more likely derived from PAA rather than Nostratic. EliasAlucard / Discussion 12:15, 10 May 2013 (UTC)
- It's fringe because only a tiny minority of Indo-Europeanists believe in "h₄", and although PAA itself isn't doubtful, it's doubtful a PIE root could be borrowed or (worse yet) inherited from PAA. —Angr 12:27, 10 May 2013 (UTC)
- Why would that be "worse"? Proto-Afro-Asiatic palaeolinguistics places PAA in the Fertile Crescent, and PIE linguistic palaeontology places PIE in the Pontic-Caspian steppe; geographically, those two regions are not that far off from each other. There are proto-Semitic loanwords in PIE, and PIE loanwords in proto-Semitic (see *táwros and *kʷékʷlos as examples). Moreover, there's some actual genetic evidence in the case of mtDNA H5a being carried by an R1a Scythian and even today, H5a is carried at low frequencies throughout Poland, Finland and Ireland, and R1a-M17+ is undoubtedly the main and only proto-Indo-European Y-DNA, not to mention that R1a-M420* probably originated in the Middle East, probably in the Neolithic Levant. The pre-historic PAA and PIE populations were definitely biologically related one way or the other. That "h₄" isn't a majority view is ad populum basically, but if you have any academic sources arguing that it should be h₃, h₂ or h₁, that's fine. I personally don't have a problem with both h₄eryós and *ħər being cognates and going back to a common Nostratic source, but like I said, Nostratic is so and so in validity. EliasAlucard / Discussion 13:07, 10 May 2013 (UTC)
- Genetic evidence is meaningless in linguistics. Whether two populations are genetically related or not says absolutely nothing at all about whether the languages they speak are related. (For example, modern-day English-speaking Osage people are unquestionably genetically descended from people who spoke Osage 150 years ago, but English and Osage are not related.) I'll see if I can find what more mainstream Indo-Europeanists consider to be the root of arā-/aire/ārya, but my resources are limited to what I have on my own bookshelf. —Angr 14:03, 10 May 2013 (UTC)
- The Osage folks are not a good example, because we know they've gone through a language shift; they're not really representative of the proto-Germanics, much less so the proto-Indo-Europeans. In the case of PAA and PIE, their respective urheimats are located geographically close (but not too close) and as David Anthony has argued, PIE had some sort of linguistic contacts with proto-Semitic, proto-Kartvelian and proto-Uralic. Moreover, this is what Mallory wrote in 1989:
- 27 Szemerenyi (1977, 125-149) provides a thorough summary of all the arguments concerning the word arya- and concludes that it is not even Indo-European but a Near Eastern, probably Ugaritic, loan word meaning ‘kinsman, companion’. — Mallory, J. P., In Search of the Indo-Europeans, p. 276, 1989, →ISBN
- So, at least one linguist has argued that the root for Aryan was derived from an Afro-Asiatic branch. In 1997, Mallory/Adams in the Encyclopedia of Indo-European Culture, said it was "unlikely" that *h₄eryós came from the Ugaritic word (p. 213), however, they didn't go into details why it was unlikely or why they rejected Oswald Szemerényi's explanation. Genetic evidence can be used in support of linguistic evidence. In itself, genes don't tell much of a language, but if they can be linked to language groups/families, it's definitely a corroboration of some sort. For example, modern Afro-Asiatic speakers (excluding the obviously race mixed ones, such as Ethiopians) and Indo-European speakers, have high genetic similarity in the autosomal profile, and on the haplogroup level, it's the same story there. There's no question that proto-Indo-European and proto-Afro-Asiatic as language families, are more closely related than proto-Indo-European is with for example, Japanese. And if you look at the oldest clades of R1a, it's found in the Middle East and the same is true of R1b (also carried by Afro-Asiatic speakers, such as Assyrians and Chadic speakers; the latter carrying R1b-V88+), which has its highest diversity in Anatolia. Obviously, haplogroups in themselves can't determine if this and that word are cognates, but to completely rule out population genetics in understanding the origin of language families, that's an ignorant position, and I know linguists often ignore population genetics, but that's a mistake on their part. Anyway, yeah, please do dig up some additional sources, I'm curious on the origins of this "most loaded term", especially because it has been so politicised. EliasAlucard / Discussion 14:28, 10 May 2013 (UTC)
- Identifying it as a loanword is one thing, but implying that PIE is descended from PAA and that this word is inherited from PAA is very misleading. And if it is a loanword, there's really no reason to propose "h₄" for it, which is nothing more than an escape hatch for words that seem to start with postlaryngeal a- but don't have h in Hittite as would be expected if the PIE form started with h₂. But if it's a loanword, there's no reason not to just call it aryós or h₁aryós. Population genetics, like archeology, can provide some interesting parallels to linguistic hypotheses, but their evidence just isn't relevant enough to linguistics for them to be used either to support or to refute theories of linguistic relationships. —Angr 14:49, 10 May 2013 (UTC)
- I never said PIE as a whole is derived from PAA, not sure how you got that impression. I never proposed "h₄", this is what Mallory, Adams and Anthony, wrote it as. They do not think *h₄eryós is a loanword from proto-Semitic/PAA. Szemerényi has argued that it is a loanword from Ugaritic, and the main source I'm citing in this entry, are Allan Bomhard and John Kerns who think it goes back to Nostratic. This is how Anthony spells it, exactly *h₄eryós, and his book is published by Princeton, so I wouldn't call it a fringe source the first thing I do. EliasAlucard / Discussion 15:17, 10 May 2013 (UTC)
- It's the current wording of the etymology section that implies this word is inherited from PAA, and the only way language X can inherit a word from language Y is if language X is descended from language Y. I know you're not the one proposing the h₄ or the source of the loanword, I just don't think the authors you're citing are really mainstream Indo-Europeanists. Just because someone gets published by a university press doesn't mean his theories aren't fringe. —Angr 15:35, 10 May 2013 (UTC)
- No, just because there are a couple of loanwords in language X, that doesn't mean language X has to descend entirely from language Y. That would be absurd. There are typical "travel words" between different language families. The proto-languages obviously had less loans between each other than modern languages do, since contacts between the populations speaking these proto-languages was quite limited back then, but there were loans even back then. What you're suggesting is that proto-Indo-European was developed in total and complete isolation from other language families, which we know isn't the case. For example, the proto-Uralic word orja, is a loan from proto-Indo-Iranian. That's an example of a proto-language inheriting a word from a language from a different language family. Not mainstream Indo-Europeanist? Are you kidding me? Mallory's revised Kurgan theory is the most popular and accepted theory today on PIE urheimat. While he may not be the tip-top PIE linguist, Adams knows more about the linguistic stuff. Anyway, my point on genetics was, languages that develop in close geographic contact with each other do indeed share a lot of words, cognates and roots with each other. R1a and R1b are derived from a common ancestral source, namely R1*. EliasAlucard / Discussion 16:59, 10 May 2013 (UTC)
- As I said, the problem was the wording. Just saying "From Proto-Afro-Asiatic" makes it sound like a descendant. I've edited it to say "Possibly a loanword from Ugaritic, from Proto-Afro-Asiatic", making it clearer that it's held to be a loanword from Ugaritic into PIE. My point about genetics is just that if speakers of two language groups that are known to be related also happen to share genetic characteristics, that's great, but you can't say "these two language groups are probably related because their speakers share genetic characteristics" any more than you can say "these two populations are probably genetically related because their languages are related". Only linguistic evidence can be used to formulate linguistic theories, and only genetic evidence can be used to formulate genetic theories. —Angr 19:35, 10 May 2013 (UTC)
- I'm well aware linguistics and population genetics are two independent fields, but genes and languages can match. The exception are known language replacements, but Jamaicans speaking English and being more closely related to Zulu speakers than they are to Russians, is beside the point, because one cannot ignore historically documented language replacements. As for "h₄", I haven't found any academic source spelling *h₄eryós as *h₃eryós, *h₂eryós or *h₁eryós, but if you can find such sources, that would be great and interesting reading. Admittedly, I haven't read every book on PIE. EliasAlucard / Discussion 20:27, 10 May 2013 (UTC)
- The thing is, we have no way of knowing what language replacements may have happened prehistorically. If a previously Afro-Asiatic-speaking population adopted an Indo-European language in 3000 BC, we'd have no way of knowing it, but we'd now have AA genes in an IE-speaking people. *h₃eryós and *h₁eryós wouldn't give the right results in any language, and *h₂eryós is problematic for the Hittite forms. The most likely forms are *h₁aryós (with original a, which was rare but not nonexistent in PIE) and simply *aryós. —Angr 20:52, 10 May 2013 (UTC)
- The reason why you think we have no way of knowing about ancient language replacements, is because you're ignoring population genetics. Of course it's difficult to know for certain if Y-DNA J2/R1b carrying Armenians and/or Persians, spoke exactly an Afro-Asiatic language, but we can say with certainty that their pre-Indo-European language was replaced by incoming R1a-Z93+ Indo-Iranians at some point. Y-DNA R1a-M17+ is pretty much the only non-Indo-European male marker (in the case of Armenians, Urartian was probably replaced by Indo-Europeanised Balkan Y-DNA I2 and E-V13 males). See also here. I don't know about *h₁aryós and *aryós, I'm no expert on how to spell PIE roots, but isn't the 4 supposed to sound like an A or something in reconstructible linguistics? I'm asking out of curiosity to know, not saying it's so. Always wondered how those numbers affect pronunciation. EliasAlucard / Discussion 09:01, 11 May 2013 (UTC)
- Basically, h₁ doesn't alter the pronunciation of e, h₂ changes e to a, and h₃ changes e to o. In Anatolian languages, h₁ disappears while h₂ and h₃ become ḫ. The problem is that a very small number of words, including this one, seem to show a but don't have a ḫ in Anatolian. So a few people have come up with h₄ to be a laryngeal that colors e to a like h₂ but that disappears in Anatolian like h₁. Thus while *h₂eryós would have become something starting with ḫar- in Hittite, *h₄eryós became something starting ar-. The more popular view is that these words with a but with no laryngeal in Anatolian had a as the primary vowel already in PIE and then had either h₁ or no laryngeal at all, so that the origin of Hittite ar- is *h₁aryós or *aryós. The question boils down to which reconstruction you prefer: one in which the number of laryngeals is restricted but which allows roots to contain a vowel other than e, or one in which the number of laryngeals is greater but which permits only e as a root vowel. —Angr 09:42, 11 May 2013 (UTC)
- Cal Watkins in The American Heritage Dictionary of Indo-European Roots lists aryo- as the "self-designation of the Indo-Iranians" and suggests it may come from *h₁ar- "to fit together", the root that is the source of Germanic arm, Latin arma "arms", ars "art", and artus "joint", and Greek arthros "joint" and aristos "best". He puts aryo- in brackets, indicating that it is one of the roots "that are not, strictly speaking, Indo-European, because they are found in only one branch or are known to have been borrowed into the family from an outside source". Unfortunately I can't tell whether aryo- is in brackets because he considers it to be found only in Indo-Iranian or because he agrees with Szemerényi that it's an Afro-Asiatic loanword. But the "fit together" root he definitely reconstructs with h₁ and a-vocalism, not with h₄ and e-vocalism. —Angr 10:08, 11 May 2013 (UTC)
- Basically, h₁ doesn't alter the pronunciation of e, h₂ changes e to a, and h₃ changes e to o. In Anatolian languages, h₁ disappears while h₂ and h₃ become ḫ. The problem is that a very small number of words, including this one, seem to show a but don't have a ḫ in Anatolian. So a few people have come up with h₄ to be a laryngeal that colors e to a like h₂ but that disappears in Anatolian like h₁. Thus while *h₂eryós would have become something starting with ḫar- in Hittite, *h₄eryós became something starting ar-. The more popular view is that these words with a but with no laryngeal in Anatolian had a as the primary vowel already in PIE and then had either h₁ or no laryngeal at all, so that the origin of Hittite ar- is *h₁aryós or *aryós. The question boils down to which reconstruction you prefer: one in which the number of laryngeals is restricted but which allows roots to contain a vowel other than e, or one in which the number of laryngeals is greater but which permits only e as a root vowel. —Angr 09:42, 11 May 2013 (UTC)
- The reason why you think we have no way of knowing about ancient language replacements, is because you're ignoring population genetics. Of course it's difficult to know for certain if Y-DNA J2/R1b carrying Armenians and/or Persians, spoke exactly an Afro-Asiatic language, but we can say with certainty that their pre-Indo-European language was replaced by incoming R1a-Z93+ Indo-Iranians at some point. Y-DNA R1a-M17+ is pretty much the only non-Indo-European male marker (in the case of Armenians, Urartian was probably replaced by Indo-Europeanised Balkan Y-DNA I2 and E-V13 males). See also here. I don't know about *h₁aryós and *aryós, I'm no expert on how to spell PIE roots, but isn't the 4 supposed to sound like an A or something in reconstructible linguistics? I'm asking out of curiosity to know, not saying it's so. Always wondered how those numbers affect pronunciation. EliasAlucard / Discussion 09:01, 11 May 2013 (UTC)
- The thing is, we have no way of knowing what language replacements may have happened prehistorically. If a previously Afro-Asiatic-speaking population adopted an Indo-European language in 3000 BC, we'd have no way of knowing it, but we'd now have AA genes in an IE-speaking people. *h₃eryós and *h₁eryós wouldn't give the right results in any language, and *h₂eryós is problematic for the Hittite forms. The most likely forms are *h₁aryós (with original a, which was rare but not nonexistent in PIE) and simply *aryós. —Angr 20:52, 10 May 2013 (UTC)
- I'm well aware linguistics and population genetics are two independent fields, but genes and languages can match. The exception are known language replacements, but Jamaicans speaking English and being more closely related to Zulu speakers than they are to Russians, is beside the point, because one cannot ignore historically documented language replacements. As for "h₄", I haven't found any academic source spelling *h₄eryós as *h₃eryós, *h₂eryós or *h₁eryós, but if you can find such sources, that would be great and interesting reading. Admittedly, I haven't read every book on PIE. EliasAlucard / Discussion 20:27, 10 May 2013 (UTC)
- As I said, the problem was the wording. Just saying "From Proto-Afro-Asiatic" makes it sound like a descendant. I've edited it to say "Possibly a loanword from Ugaritic, from Proto-Afro-Asiatic", making it clearer that it's held to be a loanword from Ugaritic into PIE. My point about genetics is just that if speakers of two language groups that are known to be related also happen to share genetic characteristics, that's great, but you can't say "these two language groups are probably related because their speakers share genetic characteristics" any more than you can say "these two populations are probably genetically related because their languages are related". Only linguistic evidence can be used to formulate linguistic theories, and only genetic evidence can be used to formulate genetic theories. —Angr 19:35, 10 May 2013 (UTC)
- No, just because there are a couple of loanwords in language X, that doesn't mean language X has to descend entirely from language Y. That would be absurd. There are typical "travel words" between different language families. The proto-languages obviously had less loans between each other than modern languages do, since contacts between the populations speaking these proto-languages was quite limited back then, but there were loans even back then. What you're suggesting is that proto-Indo-European was developed in total and complete isolation from other language families, which we know isn't the case. For example, the proto-Uralic word orja, is a loan from proto-Indo-Iranian. That's an example of a proto-language inheriting a word from a language from a different language family. Not mainstream Indo-Europeanist? Are you kidding me? Mallory's revised Kurgan theory is the most popular and accepted theory today on PIE urheimat. While he may not be the tip-top PIE linguist, Adams knows more about the linguistic stuff. Anyway, my point on genetics was, languages that develop in close geographic contact with each other do indeed share a lot of words, cognates and roots with each other. R1a and R1b are derived from a common ancestral source, namely R1*. EliasAlucard / Discussion 16:59, 10 May 2013 (UTC)
- It's the current wording of the etymology section that implies this word is inherited from PAA, and the only way language X can inherit a word from language Y is if language X is descended from language Y. I know you're not the one proposing the h₄ or the source of the loanword, I just don't think the authors you're citing are really mainstream Indo-Europeanists. Just because someone gets published by a university press doesn't mean his theories aren't fringe. —Angr 15:35, 10 May 2013 (UTC)
- I never said PIE as a whole is derived from PAA, not sure how you got that impression. I never proposed "h₄", this is what Mallory, Adams and Anthony, wrote it as. They do not think *h₄eryós is a loanword from proto-Semitic/PAA. Szemerényi has argued that it is a loanword from Ugaritic, and the main source I'm citing in this entry, are Allan Bomhard and John Kerns who think it goes back to Nostratic. This is how Anthony spells it, exactly *h₄eryós, and his book is published by Princeton, so I wouldn't call it a fringe source the first thing I do. EliasAlucard / Discussion 15:17, 10 May 2013 (UTC)
- Identifying it as a loanword is one thing, but implying that PIE is descended from PAA and that this word is inherited from PAA is very misleading. And if it is a loanword, there's really no reason to propose "h₄" for it, which is nothing more than an escape hatch for words that seem to start with postlaryngeal a- but don't have h in Hittite as would be expected if the PIE form started with h₂. But if it's a loanword, there's no reason not to just call it aryós or h₁aryós. Population genetics, like archeology, can provide some interesting parallels to linguistic hypotheses, but their evidence just isn't relevant enough to linguistics for them to be used either to support or to refute theories of linguistic relationships. —Angr 14:49, 10 May 2013 (UTC)
- The Osage folks are not a good example, because we know they've gone through a language shift; they're not really representative of the proto-Germanics, much less so the proto-Indo-Europeans. In the case of PAA and PIE, their respective urheimats are located geographically close (but not too close) and as David Anthony has argued, PIE had some sort of linguistic contacts with proto-Semitic, proto-Kartvelian and proto-Uralic. Moreover, this is what Mallory wrote in 1989:
- Genetic evidence is meaningless in linguistics. Whether two populations are genetically related or not says absolutely nothing at all about whether the languages they speak are related. (For example, modern-day English-speaking Osage people are unquestionably genetically descended from people who spoke Osage 150 years ago, but English and Osage are not related.) I'll see if I can find what more mainstream Indo-Europeanists consider to be the root of arā-/aire/ārya, but my resources are limited to what I have on my own bookshelf. —Angr 14:03, 10 May 2013 (UTC)
- Why would that be "worse"? Proto-Afro-Asiatic palaeolinguistics places PAA in the Fertile Crescent, and PIE linguistic palaeontology places PIE in the Pontic-Caspian steppe; geographically, those two regions are not that far off from each other. There are proto-Semitic loanwords in PIE, and PIE loanwords in proto-Semitic (see *táwros and *kʷékʷlos as examples). Moreover, there's some actual genetic evidence in the case of mtDNA H5a being carried by an R1a Scythian and even today, H5a is carried at low frequencies throughout Poland, Finland and Ireland, and R1a-M17+ is undoubtedly the main and only proto-Indo-European Y-DNA, not to mention that R1a-M420* probably originated in the Middle East, probably in the Neolithic Levant. The pre-historic PAA and PIE populations were definitely biologically related one way or the other. That "h₄" isn't a majority view is ad populum basically, but if you have any academic sources arguing that it should be h₃, h₂ or h₁, that's fine. I personally don't have a problem with both h₄eryós and *ħər being cognates and going back to a common Nostratic source, but like I said, Nostratic is so and so in validity. EliasAlucard / Discussion 13:07, 10 May 2013 (UTC)
- It's fringe because only a tiny minority of Indo-Europeanists believe in "h₄", and although PAA itself isn't doubtful, it's doubtful a PIE root could be borrowed or (worse yet) inherited from PAA. —Angr 12:27, 10 May 2013 (UTC)
- Also, how can a word present in Common Indo-Iranian (i. e., about 1500 BC, and almost certainly already in Proto-Indo-Iranian, ca. 2500–2000 BC), and that this page suggests was inherited from PIE (at least 3000 BC), be a loanword from Ugaritic (attested from the 14th century BC on)? This is so obviously nonsense. It's a long time ago that I read his book and I don't have it here right now, but I strongly doubt that he makes such a ridiculous claim. That's why there should always be quotations for sources not easy to check online. --Florian Blaschke (talk) 11:22, 14 June 2014 (UTC)
