The oddities of Basque

[Ketsuban]

If we're speaking of the form written 犬, I wouldn't be terribly disinclined to entertain the idea that it could be an Indo-European borrowing

Maybe, though I find it a little awkward that the Old Chinese form 犬 *kʰʷeːnʔ looks more like PIE *ḱwṓ than Tocharian ku, which is what actually got far enough east to enter China's zone of interest. 蜜 *mit by contrast is an extremely good match (e.g. Tocharian B mit "honey").

Apparently Chinese is the only part of the family to reflect a final -n in this word.

[hwhatting]

Is actually a word meaning “apple” or “fruit” derived from the Nakh-Dagahestani lexeme attested in Eastern Caucasian?

Not necessarily so, but it would provide the "missing link" for IE *mah₂l-o-

A potential semantic development is not much of a missing link

I was referring to phonology, not semantics, because the IE word isn't directly linked to the 'apple' Wanderwort.

So what is the exact role of that "missing link"? Do you assume PIE loaned the N-D word for "warm" and then derived meh2lo- "apple" from it? Or do you think EC and IE are related?

Problematic" means "it doesn't fit in the mainstream theory".

Well, yes. The more rigorous a theory, the more things it needs to leave unexplained. But in my view, that's better than handwaving and have ad-hoc explanations. The ad-hoc-ness is what I don't like about Kloekhorst's assumptions of an s-mobile.

(š is not a regular continuation of laryngeals; Kloekhorst assume an s-mobile here, which is also ad hoc and doesn't seem to be attested in other branches).

As a matter of fact, he rejects any relationship between šam(a)lu- and *ab(ō)l-, which it's too bad.

He does, yes. Just to be clear, the s-mobile is what he assumes for "eye" and "nail".

As I don't know much about Proto-Uralic, how does the Uralic form show evidence for a laryngeal?

Just as in Greek, initial *o- in Uralic (and incidentally also u- in Basque) would point to something like *h₃

Is that your own ad-hoc theory, or do is there a list of correspondences for this somewhere?

That looks basically like "everything can come from everything". It would be better if you applied the rigour and skepticism you apply to other people's theories also to your own theories.

Of course I do.

Well, you don't do it here.

Apparently, there is some variation. The most common reflex of PIE laryngeals in loanwords into Uralic is AFAIK zero, but there appear to be a few words where a laryngeal appears to be reflected by *k, *x or *š. See this chart for "Indo-Uralic" sound correspondences - which IMHO quite clearly shows that the bulk of the items are loanwords from IE into Uralic rather than inherited from Proto-Indo-Uralic or whatever.

Thanks!

Of course I disagree. Having both 'ditch' (or similar) and 'piglet' derived from the same lexeme is quiet suspicious to be true. On the other hand, the connection between the IE and Sinitic words for 'dog' seems too evident to be a chance resemblance.

You mean "pig"? Anyway, there is something called homonymy - *perk'- "spotted, multicoloured" and *perk'- "depression, trough" may not be the same lexeme, but two that have become homonyms - actually, you yourself have theorised that the latter may be a loan, which would make it a similar case to Gernan Strauß "bouquet" (native word) vs. Strauß "ostrich" (loan). And who says that the Sinitic "pig" words Nort mentioned, if they have something to do with this, aren't a direct or indirect loan from IE?

[Rounin Ryuuji]

If we're speaking of the form written 犬, I wouldn't be terribly disinclined to entertain the idea that it could be an Indo-European borrowing

Maybe, though I find it a little awkward that the Old Chinese form 犬 *kʰʷeːnʔ looks more like PIE *ḱwṓ than Tocharian ku, which is what actually got far enough east to enter China's zone of interest. 蜜 *mit by contrast is an extremely good match (e.g. Tocharian B mit "honey").

Apparently Chinese is the only part of the family to reflect a final -n in this word.

I assume ku is the form from Tocharian B specifically (for which we have evidence of loans into Old Chinese) and not Tocharian A (for which I'm not sure if we do or don't)? Either way, I might also entertain that it's a much earlier borrowing from an Indo-European language, though I suppose that begins to become a bit geographically dodgy.

I might also add that you've chopped off the bottom of the post, which notes that the word also incidentally resembles several areal dog bark onomatopoeia (Mandarin wang, Cantonese wou, Japanese wan), and that I think it could easily be a coincidence of two onomatopoeic words happening to come out similar. The domestication of dogs being, to my understanding, of a time depth much greater than any etymology we can reliably reconstruct, I doubt very much there is some *kwe~*kwo wanderwort underlying the two forms.

[Ketsuban]

I assume ku is the form from Tocharian B specifically (for which we have evidence of loans into Old Chinese) and not Tocharian A (for which I'm not sure if we do or don't)?

It's both.

I might also add that you've chopped off the bottom of the post, which notes that the word also incidentally resembles several areal dog bark onomatopoeia (Mandarin wang, Cantonese wou, Japanese wan), and that I think it could easily be a coincidence of two onomatopoeic words happening to come out similar. The domestication of dogs being, to my understanding, of a time depth much greater than any etymology we can reliably reconstruct, I doubt very much there is some *kwe~*kwo wanderwort underlying the two forms.

I chopped that off because I didn't have anything substantive to say about it (other than that to observe that the Old Chinese form doesn't much resemble those onomatopoeia, at least to my eyes) but I'll note that the root appears to go back to Sino-Tibetan, so if it is onomatopoeic it's a pretty gnarly time-depth.

[Rounin Ryuuji]

Words */kwen/ (earlier reconstructions point to a form */kwə(n)/) and /wɑŋ/ are both fairly similar in structure, beginning with a labial or labialised consonant, and ending in a sequence of a common vowel (in the case of */kwən/ and /wɑŋ/, a not dissimilar-sounding vowel; central or low vowels seem common in dog onomatopoeia) and a common nasal. Yes, there presumably is a lot of time-depth on it, but I could see the Proto-Indo-European and Sinitic forms both descending from incidentally similar onomatopoeia as easily as I could see them being related.

[Nortaneous]

If we're speaking of the form written 犬, I wouldn't be terribly disinclined to entertain the idea that it could be an Indo-European borrowing

Maybe, though I find it a little awkward that the Old Chinese form 犬 *kʰʷeːnʔ looks more like PIE *ḱwṓ than Tocharian ku, which is what actually got far enough east to enter China's zone of interest. 蜜 *mit by contrast is an extremely good match (e.g. Tocharian B mit "honey").

Apparently Chinese is the only part of the family to reflect a final -n in this word.

I assume ku is the form from Tocharian B specifically (for which we have evidence of loans into Old Chinese) and not Tocharian A (for which I'm not sure if we do or don't)? Either way, I might also entertain that it's a much earlier borrowing from an Indo-European language, though I suppose that begins to become a bit geographically dodgy.

TA and TB may not even have been distinct at the beginning of Tocharian-Sinitic contact, and some Tocharian loanwords into Sinitic have sources only attested in TA.

The resemblance of the two proto-forms is probably coincidental - a borrowing into IE would have to be pre-satemization, and a borrowing into ST would have to be widespread and involve a form that doesn't resemble the Tocharian, and depending on its ST distribution may not even make chronological sense as coming from pre-satemization Indo-Iranian or whatever else could've made it that far east.

If not transmitted directly, I'd expect intermediates, and the Proto-Kra-Dai form... OK, the PKD form is something like *kʰuma, so maybe there's something that could be done with that? And AFAIK that root isn't attested in Austronesian...

[Talskubilos]

(š is not a regular continuation of laryngeals; Kloekhorst assume an s-mobile here, which is also ad hoc and doesn't seem to be attested in other branches).

I'd link (but not necessarily in a direct way) Hittite šam(a)lu- 'apple' to IE *sam- 'summer' (seemingly not native), in turn corresponding to Semitic *ħamm- 'to be hot; warm'.

As I don't know much about Proto-Uralic, how does the Uralic form show evidence for a laryngeal?

Just as in Greek, initial *o- in Uralic (and incidentally also u- in Basque) would point to something like *h₃

Is that your own ad-hoc theory, or do is there a list of correspondences for this somewhere?

Thinking it over, back vowels /o, u/ in Basque and Uralic would be the result of regressive assimilation to the following /m/. In fact, besides the forms I've already mentioned, there's also umao, umau, umo 'ripe, seasoned'.

Taking into account the Semitic form, for the Basque and Uralic words I'd reconstruct a protoform *Hɑmæ-lV ~ *Hɑmæ-nV, where H stands for a "laryngeal". On the other hand, Basque uda 'summer' would derive from an unsuffixed variant *Hɑmæ.

Is actually a word meaning “apple” or “fruit” derived from the Nakh-Dagahestani lexeme attested in Eastern Caucasian?A potential semantic development is not much of a missing link

I was referring to phonology, not semantics, because the IE word isn't directly linked to the 'apple' Wanderwort.

So what is the exact role of that "missing link"? Do you assume PIE loaned the N-D word for "warm" and then derived meh2lo- "apple" from it?

That's right. What puzzles me is apparently the "laryngeal" underwent metathesis twice, one in EC and other in IE.

Or do you think EC and IE are related?

In an earlier post, I quoted an old article by Sergei Starostin where he proposed a number of loanwords (which he called "isoglosses") from EC to IE, including *porḱ-o- 'piglet'. IMHO, this linguistic stratum is related to the Kurgan people, i.e. the nomadic shepherds of the Pontic Steppes.

[hwhatting]

I'd link (but not necessarily in a direct way) Hittite šam(a)lu- 'apple' to IE *sam- 'summer' (seemingly not native), in turn corresponding to Semitic *ħamm- 'to be hot; warm'.

Thinking it over, back vowels /o, u/ in Basque and Uralic would be the result of regressive assimilation to the following /m/. In fact, besides the forms I've already mentioned, there's also umao, umau, umo 'ripe, seasoned'.

Taking into account the Semitic form, for the Basque and Uralic words I'd reconstruct a protoform *Hɑmæ-lV ~ *Hɑmæ-nV, where H stands for a "laryngeal". On the other hand, Basque uda 'summer' would derive from an unsuffixed variant *Hɑmæ.

Are you making the sound correspondences up as you go along in order to justify the connections you see or do you have a system of sound correpondences with examples worked out? If the latter, maybe you could share it?

Do you assume PIE loaned the N-D word for "warm" and then derived meh2lo- "apple" from it?

That's right. What puzzles me is apparently the "laryngeal" underwent metathesis twice, one in EC and other in IE.

Maybe because the words are lookalikes and actually don't have anything to do with each other? By the way, you now assume a sematic development deriving "apple" from "warm" in two cases; is that development actually attested somewhere?

Or do you think EC and IE are related?

In an earlier post, I quoted an old article by Sergei Starostin where he proposed a number of loanwords (which he called "isoglosses") from EC to IE, including *porḱ-o- 'piglet'. IMHO, this linguistic stratum is related to the Kurgan people, i.e. the nomadic shepherds of the Pontic Steppes.

OK, I have to get around to read that article some time.

[Zju]

I'd link (but not necessarily in a direct way) Hittite šam(a)lu- 'apple' to IE *sam- 'summer' (seemingly not native), in turn corresponding to Semitic *ħamm- 'to be hot; warm'.

Thinking it over, back vowels /o, u/ in Basque and Uralic would be the result of regressive assimilation to the following /m/. In fact, besides the forms I've already mentioned, there's also umao, umau, umo 'ripe, seasoned'.

Taking into account the Semitic form, for the Basque and Uralic words I'd reconstruct a protoform *Hɑmæ-lV ~ *Hɑmæ-nV, where H stands for a "laryngeal". On the other hand, Basque uda 'summer' would derive from an unsuffixed variant *Hɑmæ.

Are you making the sound correspondences up as you go along in order to justify the connections you see or do you have a system of sound correpondences with examples worked out? If the latter, maybe you could share it?

Indeed, all of **Hɑmæ/lV/nV look like arbitrarily constructed words to get the initially desired result. And of course, we can drop and add consonants as much as we want to get from **Hɑmæ to uda.

More examples and sound changes are called for, or it's all chance resemblances. Not that uda resembles šam(a)lu-.

[Travis B.]

With how fast and loose Talskubilos is playing with his supposed sound changes and loaning, much the less Wanderwörter, he can relate anything to anything no matter how remote. Based on all this, I can really not take anything he has been saying in this thread seriously.

[Rounin Ryuuji]

I have a broadly similar view, and might add to it a note that he tends to be rather ill-mannered if you disagree with him.

[alice]

With how fast and loose Talskubilos is playing with his supposed sound changes and loaning, much the less Wanderwörter, he can relate anything to anything no matter how remote. Based on all this, I can really not take anything he has been saying in this thread seriously.

Sounds like it's time to put it to a vote: is he cast or forged?

[Raholeun]

I have a broadly similar view, and might add to it a note that he tends to be rather ill-mannered if you disagree with him.

Certainly I lack the knowledge to make authoratative claims about Proto-Indo-European historical phonology, but the methods of Talskubilos do seem a bit too liberal.

However, I feel he has been quite easy-going given the pushback received.

2¢

[Rounin Ryuuji]

Well, I did say he exhibits bad manners, not a bad temper.

[Skookum]

Do you assume PIE loaned the N-D word for "warm" and then derived meh2lo- "apple" from it?

That's right. What puzzles me is apparently the "laryngeal" underwent metathesis twice, one in EC and other in IE.

Maybe because the words are lookalikes and actually don't have anything to do with each other? By the way, you now assume a sematic development deriving "apple" from "warm" in two cases; is that development actually attested somewhere?

A similar semantic path is possibly attested in Salish. Proto-Salish had a root *qʷ’al ~ *qʷ’ay with meanings ranging from "to scorch, ashes, black, roast; ripen; berry". For example, see Lushootseed qʷ’əl "burn, warm, bake, cook, ripe", sqʷ’əláɬəd "berry" (-áɬəd meaning "food"). But given the varied meanings it seems possible that there were originally two homophonous roots, one meaning "burn, black, roast, etc" and another meaning "ripe", which became conflated in descendant languages.

[Rounin Ryuuji]

It's a... vaguely plausible semantic development, which is partly-attested in Japanese — the element 焼く、焼き (yaku, yaki) which means "burn, cook, heat up", note 日焼き hiyaki, which can mean "sunburn" and "suntan", and also "discolouring owing to exposure to the sun", but 照り焼き teriyaki, a style of cooking (the first element means "shine, gleam"); I could see this verb developing a meaning ripen, since it can refer to multiple forms of change by exposure to heat, flame, or sunlight, but that extension hasn't happened yet, and would probably require quite a deal of time depth, or some other verb coming in to mean "burn, bake, cook" to push it to mean something more specific.

This said, I don't think the Indo-European words for "apple" were derived from a word for "warm" in this way. The leap is just too great to be plausible without any further explanation.

[Skookum]

It's a... vaguely plausible semantic development, which is partly-attested in Japanese — the element 焼く、焼き (yaku, yaki) which means "burn, cook, heat up", note 日焼き hiyaki, which can mean "sunburn" and "suntan", and also "discolouring owing to exposure to the sun", but 照り焼き teriyaki, a style of cooking (the first element means "shine, gleam"); I could see this verb developing a meaning ripen, since it can refer to multiple forms of change by exposure to heat, flame, or sunlight, but that extension hasn't happened yet, and would probably require quite a deal of time depth, or some other verb coming in to mean "burn, bake, cook" to push it to mean something more specific.

This said, I don't think the Indo-European words for "apple" were derived from a word for "warm" in this way. The leap is just too great to be plausible without any further explanation.

Yeah the more I think about it, the more I can see multiple paths of evolution, maybe something like "burn" > "black" > "darken" > "ripen". In some languages, reflexes of *qʷ’al ~ *qʷ’ay apparently mean "done", presumably in the context of "finished cooking/ripening", so I could see "roast" > "finish cooking" > "be ready for eating" > "ripen" as another development. Or even a development very similar to the one you mentioned in Japanese. But this is pretty far off topic so I'll leave it there.

[Nortaneous]

OK, let's look into the dog thing a little more.

PIE: *ḱwō, *ḱwon-
Proto-Hmongic: *hmaŋ C (Ratliff)
Proto-Kra-Dai: *kʰumaː (I forget)
Proto-Sino-Tibetan: *d-kʷəy-n (Matisoff)
Proto-Rgyalrongic: *k.naʔ (Hsiu)

But:
Proto-Karenic *θwi' (Luangthongkum)
Proto-Khmeric *cɔː, ckɛː (Sidwell & Rau)
Proto-Pearic *c(ɔ)ʔ (Headley)
Proto-Monic *clur (Diffloth)
Proto-Hmong-Mien *qlAu B1 ~ *qluwX (whatever Wikipedia's list is from)
Proto-Nicobarese: *ʔam (Sidwell)
Proto-Tungusic *ŋinda (Starostin)
Burushaski huk
Kusunda agəl
Japanese inu

So you could imagine some form like *kʷə́{m|n}a that was borrowed into PIE, Proto-Hmongic, and PKD, and then at various times into or within ST to account for the irregular correspondences there - but once you're there, it's tempting to continue on and imagine connections to the Proto-Nicobarese (via *qam?), Proto-Tungusic (an original *kʷəmda > *ginda > *ŋinda), Japanese (via cognacy to Proto-Tungusic following Starostin), and even Kusunda (*a-kən-?), Burushaski (*kʷʰə́~kʷən?), and anything else. And once you get there, you have to back away.

But the words still look similar, don't they? Is there any way to reduce that? Well, Japhug preserves signs of earlier Rgyalrongic (potentially also Tibetic) *kɯ- and *qa- prefix in animals, so the proper comparandum is just *naʔ. The Hmongic and Kra-Dai forms could be related, especially once you consider that every other proto-language of the region has been revised into di- or at least sesquisyllabism, although the correspondence of Hmongic *-ŋ to Kra-Dai *-0 is hard to explain. (Kra-Dai *a *aː correspond to Austronesian *ə *a, so the length in KD is expected.) But connecting this *kuma(ŋ) to PIE *ḱwon- is harder - why does the PIE form have an initial palatal and a root-final n?

Proto-Salishan *qʷʼal is clearly also related, although the semantic connection is difficult.

[Talskubilos]

Indeed, all of **Hɑmæ/lV/nV look like arbitrarily constructed words to get the initially desired result. And of course, we can drop and add consonants as much as we want to get from **Hɑmæ to uda.

Isn't it that way reconstructions are made?

*Hɑmæ > Basque uda 'summer'
*Hɑmæ-nV > Basque umao, umau, umo 'ripe, seasoned' ~ udare, udari 'pear'

The thing is /m/ was dropped in Paleo-Basque, apparently leading to a prosthetic /d/ (I mentioned other examples before), but in some varieties nasality was retained, hence the above doublet.

Do you assume PIE loaned the N-D word for "warm" and then derived meh2lo- "apple" from it?

That's right. What puzzles me is apparently the "laryngeal" underwent metathesis twice, one in EC and other in IE.

Maybe because the words are lookalikes and actually don't have anything to do with each other?

I don't think so.

By the way, you now assume a sematic development deriving "apple" from "warm" in two cases

Not exactly. As I mentioned before, the sequence would be: 'warm (season)' > 'ripe (fruit)' > 'apple/pear'.

[Skookum]

Indeed, all of **Hɑmæ/lV/nV look like arbitrarily constructed words to get the initially desired result. And of course, we can drop and add consonants as much as we want to get from **Hɑmæ to uda.

Isn't it that way reconstructions are made?

No