Loanwords in (P)IE

[Talskubilos]

I hope so. You are smart; you can learn, and not all your ideas are misguided. It just seems as if you still have some things to learn about the methodology of historical linguistics.

I also hope you don't equal historical linguistics to the std theory of PIE and little more.

Indeed I don't. If you read my posts to the Great PIE Thread, you'd know that I have ideas that go beyond that. While I think that the standard theory describes Late PIE (i.e., the common ancestor of the non-Anatolian IE languages) quite well, I don't think it was the last word on it, and I do think that Early PIE (i.e., the common ancestor of Late PIE and Anatolian) was different in some regards. Currently, I am researching a hypothetical lost branch of IE in Bronze Age Western Europe, which is entirely my own idea. Also, as you know, my ideas about the further relationships of IE go well beyond the orthodox "we don't know" - they are just different from yours, and nobody can say who is right

My own approach gives the same priority to "horizontal" (i.e. loanwords and Wanderwörter) than "vertical" (i.e. inheritance) language relationships, so I I'm in favour of a classification of words and languages following biblioteconomy principles rather than genetic trees.

[bradrn]

‘Biblioteconomy‘ isn’t even an English word. Apparently in Spanish it means ‘library science’, which makes no sense in context.

[Talskubilos]

‘Biblioteconomy‘ isn’t even an English word. Apparently in Spanish it means ‘library science’, which makes no sense in context.

Yes, you're right. Library science or library economy is the discipline or art of classifying books and other written documents. Change 'books' by 'words' and hopefully you'd understand what I'm aiming for. For example, we could label an etymology according to chronology (Mesolithic, Neolithic, etc.), transmission type (inheritance, contact loanword, cultural loanword, Wanderwort) and so on.

[WeepingElf]

My own approach gives the same priority to "horizontal" (i.e. loanwords and Wanderwörter) than "vertical" (i.e. inheritance) language relationships, so I I'm in favour of a classification of words and languages following biblioteconomy principles rather than genetic trees.

So it seems as if we had been talking past each other all the time Surely, loanwords often have interesting stories to tell, and are worth studying, as they tell us which languages had contact with which, and there is hardly a language that doesn't have loanwords. But languages are classified by their inherited vocabularies and grammatical morpheme inventories. The numerous Norman French loanwords do not make English a Romance language, though these loanwords are actually useful for the study of the history and dialectology of French and the Romance language family, and are for this reason included in Romance etymological dictionaries. The problem with your approach is that it doesn't do a good job telling inherited words, loanwords and chance resemblances apart, as is necessary to classify languages and reconstruct their common ancestors, and some of your etymologies are so adventurous that they are likely misled by chance resemblances.

[Richard W]

From looking at a similar case in the Americas, there is a similar pattern of n-m in western North America and Mesoamerica, also crossing multiple families that have not been demonstrated to be related. What is the chance that this sort of thing is areal rather than genetic?

Unfortunately, establishing language-relatedness in the Americas seems to rely on unique cognates. So if A and B are sister languages and C is a cousin, but language B has been altered badly enough to makes its relationship to A uncertain, one can't establish that A and C are related. This is a major pain if languages diverged 20,000 years ago, and makes the best the enemy of the good.

My gut feeling is that the phenomenon is primarily genetic in origin, but case by case refutation might conceivably be possible. The only areal mechanism I can think of would start with a collapse of the pronouns, but even then I am not persuaded that a common set of shared new forms would be established. Historically, I don't think we have an example of a new set of forms being established across an area.

[WeepingElf]

I have decided to leave Talskubilos alone and refrain from picking apart his methodology or his etymologies (I hope I manage to keep this promise and don't get tempted to say something - if I should fail, please remind me). Because that's meaningless - it wouldn't change anything. The best way to deal with them is certainly to consider it all something like Talskubilos's conlang. So as long as he stays in this thread and doesn't molest the Great PIE Thread or other historio-linguistic threads on the ZBB, it is all fine.

[Talskubilos]

Surely, loanwords often have interesting stories to tell, and are worth studying, as they tell us which languages had contact with which, and there is hardly a language that doesn't have loanwords. But languages are classified by their inherited vocabularies and grammatical morpheme inventories. The numerous Norman French loanwords do not make English a Romance language, though these loanwords are actually useful for the study of the history and dialectology of French and the Romance language family, and are for this reason included in Romance etymological dictionaries. The problem with your approach is that it doesn't do a good job telling inherited words, loanwords and chance resemblances apart, as is necessary to classify languages and reconstruct their common ancestors, and some of your etymologies are so adventurous that they are likely misled by chance resemblances. (boldface is mine)

On the contrary, my own focus is on words, which can be inherited and borrowed multiple times along their history. For example, the ultimate origin of Latin faber 'smith' < Italic *fabro- < IE *dhabh-ro- is an onomatopoeia *tap/dab- of the hitting of metal.

[Talskubilos]

I have decided to leave Talskubilos alone and refrain from picking apart his methodology or his etymologies (I hope I manage to keep this promise and don't get tempted to say something - if I should fail, please remind me). Because that's meaningless - it wouldn't change anything. The best way to deal with them is certainly to consider it all something like Talskubilos's conlang. So as long as he stays in this thread and doesn't molest the Great PIE Thread or other historio-linguistic threads on the ZBB, it is all fine.

No problem, but remember I also created another thread specific for Basque. As some say, each linguist has his/her own theory.

[WeepingElf]

Fair. Historical linguistics is such a wide field that anyone working in it has to specialize. This is especially true for us amateurs, who have fewer resources at our disposal than the academic professionals. Champollion specialized in the decipherment of the Egyptian hieroglyphs; Ventris specialized in the decipherment of Linear B; etc. (Both would perhaps have taken on new challenges if they had lived longer, though.) Likewise, you are specializing in tracing loanwords; and I am specializing in the languages of prehistoric Europe, both Indo-European (such as the language of the Bell Beaker culture and its Bronze Age descendants) and pre-Indo-European (this specialization is of course of relevance to my conlang projects, which also means that I have to be careful to keep the hypothetical natlangs and the conlangs distinct, and restrict the information flow to one direction, namely from the former to the latter).

[Talskubilos]

Fair. Historical linguistics is such a wide field that anyone working in it has to specialize. This is especially true for us amateurs, who have fewer resources at our disposal than the academic professionals. Champollion specialized in the decipherment of the Egyptian hieroglyphs; Ventris specialized in the decipherment of Linear B; etc. (Both would perhaps have taken on new challenges if they had lived longer, though.) Likewise, you are specializing in tracing loanwords; and I am specializing in the languages of prehistoric Europe, both Indo-European (such as the language of the Bell Beaker culture and its Bronze Age descendants) and pre-Indo-European (this specialization is of course of relevance to my conlang projects, which also means that I have to be careful to keep the hypothetical natlangs and the conlangs distinct, and restrict the information flow to one direction, namely from the former to the latter).

My own speciality is etymology, i.e. the history of words, but I'm also interesting in Paleo-European languages, and so I study loanwords.

Germanic itself has a lot of non-IE lexicon, most of which has no apparent cognates in known language families. Only a few items have correspondences in Semitic, as e.g. crab ~Semitic *ʕa-k’rab- 'scorpion', but most of the lexicon of this origin is shared with other IE branches. For example, gable < *ghebhōl-, with cognates in Greek kephalê 'head' as well in Tocharian, but there's also *kapōl-o- (Old English hafola, Sanskrit kapá:la-), a doublet which I regard as loanwords from Semtic *gbl- 'mountain; boundary, border' (Arabic ʒabal 'mountain'), with a straightforawrd semantic shift.

[WeepingElf]

Fair. Historical linguistics is such a wide field that anyone working in it has to specialize. This is especially true for us amateurs, who have fewer resources at our disposal than the academic professionals. Champollion specialized in the decipherment of the Egyptian hieroglyphs; Ventris specialized in the decipherment of Linear B; etc. (Both would perhaps have taken on new challenges if they had lived longer, though.) Likewise, you are specializing in tracing loanwords; and I am specializing in the languages of prehistoric Europe, both Indo-European (such as the language of the Bell Beaker culture and its Bronze Age descendants) and pre-Indo-European (this specialization is of course of relevance to my conlang projects, which also means that I have to be careful to keep the hypothetical natlangs and the conlangs distinct, and restrict the information flow to one direction, namely from the former to the latter).

My own speciality is etymology, i.e. the history of words, but I'm also interesting in Paleo-European languages, and so I study loanwords. For example, Germanic has a lot of non-IE lexicon, most of which has no apparent cognates in known language families (although a few items, for example crab, have correspondences in Semitic).

Yes - such loanwords are the most important source of information about Paleo-European languages. There are, of course, also names of geographical objects which may be from such languages, but the drawback is that we don't know the original lexical meanings of such names - unless a correlation between a name element and a salient feature of the named objects can be discerned, as in the names of Central European salt production sites with the element *hal-, which thus probably meant 'salt' in whichever language those names are from (in this case, perhaps an IE language of the hypothetical branch I call "Aquan", as this is a plausible development of PIE *sh₂el-; but that is merely a hypothesis of mine that may be wrong).

[Talskubilos]

Yes - such loanwords are the most important source of information about Paleo-European languages. There are, of course, also names of geographical objects which may be from such languages, but the drawback is that we don't know the original lexical meanings of such names - unless a correlation between a name element and a salient feature of the named objects can be discerned, as in the names of Central European salt production sites with the element *hal-, which thus probably meant 'salt' in whichever language those names are from (in this case, perhaps an IE language of the hypothetical branch I call "Aquan", as this is a plausible development of PIE *sh₂el-; but that is merely a hypothesis of mine that may be wrong).

Somebody suggested the source could be simply Celtic, as the shift s- > h- is common there, but it could also be an etymological doublet with an alternance s ~ h like the one I'm going to explain:

Celtic *samo- 'summer' and English summer < *sm-ro- ~ Greek hēméra and Armenian awr 'day' < *h₂ēm-r- '(heat of the) day', which I'd link to Semitic *ħamm- 'to be hot; warm'. So we've got a root *HVmV 'heat' which can be either applied to the day (in southern latitudes) or the summer (in nothern ones). A derivative *HVmV-lV/HVmV-nV would be reflected (with methatesis) in NEC *mhalV/mhanV- 'warm' and IE *mah₂lo- 'apple', with a straightforward semantic drift from 'warm (season)' to 'fruit'. We've also got Hittite sam(a)lu- 'apple (tree)', Paleo-European *abVl- 'apple' (< *amHVl-?), Uralic *omena/omVrV 'apple' and Basque udare, udari, madari 'pear' < *uma-lV, um(a)o 'ripe, seasoned' < *uma-nV and uda 'summer' < *uma.

[WeepingElf]

Yes - such loanwords are the most important source of information about Paleo-European languages. There are, of course, also names of geographical objects which may be from such languages, but the drawback is that we don't know the original lexical meanings of such names - unless a correlation between a name element and a salient feature of the named objects can be discerned, as in the names of Central European salt production sites with the element *hal-, which thus probably meant 'salt' in whichever language those names are from (in this case, perhaps an IE language of the hypothetical branch I call "Aquan", as this is a plausible development of PIE *sh₂el-; but that is merely a hypothesis of mine that may be wrong).

Somebody suggested the source could be simply Celtic, as the shift s- > h- is common there,

I know this suggestion, but the shift s > h in Celtic is only post-Roman Brythonic, see the Old British river name Sabrina which is now Hafren in Welsh but Severn in English. The shift did not happen in Irish, nor in Continental Celtic (Paris is not on the †Heine). But the change happened independently at least thrice in IE: in Brythonic, Greek and Iranian (cf. Avestan haoma vs. Sanskrit soma), so why not a fourth time in a lost IE language in Bronze Age Central Europe?

Greek hēméra and Armenian awr 'day' < *h₂ēm-r- '(heat of the) day'

Nope! Greek h is from PIE *s (see above), never from PIE *h₂.

[Talskubilos]

Greek hēméra and Armenian awr 'day' < *h₂ēm-r- '(heat of the) day'

Nope! Greek h is from PIE *s (see above), never from PIE *h₂.

Are you sure? This is Mallory-Adams's reconstruction, although it would reinforce the link with the 'summer' words.

[WeepingElf]

Greek hēméra and Armenian awr 'day' < *h₂ēm-r- '(heat of the) day'

Nope! Greek h is from PIE *s (see above), never from PIE *h₂.

Are you sure? This is Mallory-Adams's reconstruction, although it would reinforce the link with the 'summer' words.

I just checked, and did not find it, and I just can't believe that Mallory and Adams missed such a well-known bit of handbook knowledge as the disappearance of laryngeals in every non-Anatolian IE language. Don't attribute your fancies to authors who don't have anything to do with them!

[Talskubilos]

Nope! Greek h is from PIE *s (see above), never from PIE *h₂.

Are you sure? This is Mallory-Adams's reconstruction, although it would reinforce the link with the 'summer' words.

I just checked, and did not find it, and I just can't believe that Mallory and Adams missed such a well-known bit of handbook knowledge as the disappearance of laryngeals in every non-Anatolian IE language. Don't attribute your fancies to authors who don't have anything to do with them!

FYI, it's on page 303 of their book The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World (2006): A Greek-Armenian isogloss for ‘day’ is *h₂ehₓ-mer-, a derivative of *h₂ehₓ- ‘burn’ (i.e. Grk ēmēra, Arm awr)

But you're right about Greek h- coming from *s-, although not always in the std way. Take for example hágios ‘holy’, hagnós ‘pure’, házdomai ‘to worship’, Sanskrit yájati ‘workships’ < *yag´- 'to honour, to workship' ~ Latin sanctus ‘sacred’, sacer ‘holy’, Hittite šākl(ā)i- ‘custom, rule; rite, ceremony’ < *sa(n)k- 'to sanctify'. NEC *=ǝqE ‘to raise, to be high’ points to the semantic motivation: victims of sacrifices were raised upon the altar, as today priests do with the communion wafer.

[Lērisama]

Nope! Greek h is from PIE *s (see above), never from PIE *h₂.

Are you sure? This is Mallory-Adams's reconstruction, although it would reinforce the link with the 'summer' words.

I just checked, and did not find it, and I just can't believe that Mallory and Adams missed such a well-known bit of handbook knowledge as the disappearance of laryngeals in every non-Anatolian IE language. Don't attribute your fancies to authors who don't have anything to do with them!

Wiktionary gives the same etymology, and cites Beekes, although I don't have access to it, and claims the h is probably by analogy with hespérā. It still seems a little suspect to me to reconstruct based only on Armenian and Greek when the latter requires an irregular development.

[WeepingElf]

Are you sure? This is Mallory-Adams's reconstruction, although it would reinforce the link with the 'summer' words.

I just checked, and did not find it, and I just can't believe that Mallory and Adams missed such a well-known bit of handbook knowledge as the disappearance of laryngeals in every non-Anatolian IE language. Don't attribute your fancies to authors who don't have anything to do with them!

FYI, it's on page 303 of their book The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World (2006): A Greek-Armenian isogloss for ‘day’ is *h₂ehₓ-mer-, a derivative of *h₂ehₓ- ‘burn’ (i.e. Grk ēmēra, Arm awr)

Thank you for the page reference; I've found it now. But let me explain what is going on here. Some Greek dialects had lost initial /h/ (this change is called psilosis), of whichever origin, others not. This caused some confusion, and in some words, an /h/ was "restored" that had not been there even before psilosis (an example of hypercorrection). But because many of these words had an initial laryngeal in PIE times, this may look like a preserved laryngeal which it actually isn't. Note that Mallory & Adams give the word without the initial /h/, because that's the etymologically correct, expected form here. So I apologize for accusing you of misattribution.

[Ketsuban]

I don't think it's from Semitic because I don't think Semitic and Indo-European were spoken anywhere near enough to each other for them to happen (and saying "Wanderwort" requires showing the links in the chain or you're just saying "it was loaned through Old High Martian") but you may be onto something in connecting Germanic *sumaraz and Celtic *samos with Greek ἦμᾰρ and Armenian awr which are otherwise isolated and a bit strange. (The competing hypothesis is that "summer" is related to "semi-", because the Sanskrit cognate can mean "half-year".)

That said,

A derivative *HVmV-lV/HVmV-nV would be reflected (with methatesis) in NEC *mhalV/mhanV- 'warm' and IE *mah₂lo- 'apple', with a straightforward semantic drift from 'warm (season)' to 'fruit'. We've also got Hittite sam(a)lu- 'apple (tree)', Paleo-European *abVl- 'apple' (< *amHVl-?), Uralic *omena/omVrV 'apple' and Basque udare, udari, madari 'pear' < *uma-lV, um(a)o 'ripe, seasoned' < *uma-nV and uda 'summer' < *uma.

I already took you to school over this before, and the fact you're still peddling it is frankly damning. "Apple" is solidly Indo-European, deriving from a metathesised variant of *meh₂l- (*meh₂l- > *h₂eml- > *h₂ebl-) motivated by association with h₂em-ro- "sour". Hittite šamalu- is s-mobile. Your Uralic *omena/omVrV is "Starostinian gibberish formed by smashing together Proto-Finnic *omëna and Hungarian alma (apparently assuming a metathesis from *omra with a magic vanishing vowel)"; both the Proto-Finnic and Hungarian words are Indo-European loans. The only reason you mention Basque is because madari is superficially appealing, but that's a prefix ma-; udare is clearly and obviously nothing to do with IE *meh₂- "grow, increase".

[Talskubilos]

I don't think it's from Semitic because I don't think Semitic and Indo-European were spoken anywhere near enough to each other for them to happen (and saying "Wanderwort" requires showing the links in the chain or you're just saying "it was loaned through Old High Martian") but you may be onto something in connecting Germanic *sumaraz and Celtic *samos with Greek ἦμᾰρ and Armenian awr which are otherwise isolated and a bit strange. (The competing hypothesis is that "summer" is related to "semi-", because the Sanskrit cognate can mean "half-year".)

I'm not sure if this word was borrowed from Semitic, but the connection seems likely to me.

On the other hand, there're quite a lot of proposed loanwords from this source, likely through the languages spoken by the farmers who colonized Europe in the Neolithic. Some of them have been borrowed more than once, thus generating doublets like the one I quoted before: *ghebhōl- (kephalê 'head', English gable) ~ *kapōl-o- (Old English hafola, Sanskrit kapá:la-), from Semitic *gbl- 'mountain; boundary, border' (Arabic ʒabal 'mountain'), with a straightforawrd semantic shift.

I already took you to school over this before, and the fact you're still peddling it is frankly damning. "Apple" is solidly Indo-European, deriving from a metathesised variant of *meh₂l- (*meh₂l- > *h₂eml- > *h₂ebl-) motivated by association with h₂em-ro- "sour". Hittite šamalu- is s-mobile.[/b]

I disagree. Hittite šamalu- is clearly related to the 'summer' words and this *(s)meh₂l- is a desperate attempt to keep the the ass of the std PIE model guarded, ignoring external evidence. But it's clear there wre several metathesis involved from the original *HVmV-lV/HVmV-nV: *(V)mHVlV/(V)mHVnV (Paleo-European *abVl- < *amHVlV? and NEC *mhalV/mhanV-) > *mVHlV (IE *mah₂l-o-).

Your Uralic *omena/omVrV is "Starostinian gibberish formed by smashing together Proto-Finnic *omëna and Hungarian alma (apparently assuming a metathesis from *omra with a magic vanishing vowel)"; both the Proto-Finnic and Hungarian words are Indo-European loans.

I think you suffer of IE-centrism, as the distribution of these words points to a Wanderwort which spread to several languages, including IE, but it didn't originate there.

The only reason you mention Basque is because madari is superficially appealing, but that's a prefix ma-; udare is clearly and obviously nothing to do with IE *meh₂- "grow, increase".

This prefix has no etymological value: uda < *HVmV and udare, udari < *HVmV-lV, with the nasal replaced by a prosthetic d, but it's kept in um(a)o < *HVmV-nV.