Zompist Bboard Again

Talskubilos wrote: ↑Tue Jan 11, 2022 7:33 am

keenir wrote: ↑Tue Jan 11, 2022 7:30 amUm, if the Celts were the masters of making wheels, why wouldn't the Gauls borrow from the Celts?

But Gauls were Celts, weren't they?

Ah, okay. I got confused; I became so accustomed to each name being unrelated ("the romans borrowed X&Y from the basques" etc), that I thought that was a pattern being continued. My bad; sorry...i shoulda gotten more sleep before hitting Reply.

Talskubilos wrote: ↑Tue Jan 11, 2022 5:20 am

Linguoboy wrote: ↑Mon Jan 10, 2022 4:00 pm

Talskubilos wrote: ↑Sun Jan 09, 2022 9:42 amFrom an external one, we've got Celtic loanwords into other IE languages such as Greek kanthós 'wheel rim'.

What's the evidence for this being a Celtic borrowing? The sense of "wheel rim" could have developed from the earlier meaning of "corner of the eye" which is attested in the works of Aristotle.

As a matter of fact, ancient Celts were known for their mastership at making wheels and wheeled vehicles, and so Latin carrus 'cart, wagon' was a Celtic borrowing, most likely from Gaulish. In Alinei's opinion, Latin rota 'wheel' was also from Celtic: The Celtic origin of Lat. rota and its implications for the prehistory of Europe.

That's not actually an answer to the question I asked. Michelin makes great tyres, but it doesn't follow from that that our word for "tyre" is a borrowing from French.

Linguoboy wrote: ↑Tue Jan 11, 2022 11:26 am That's not actually an answer to the question I asked. Michelin makes great tyres, but it doesn't follow from that that our word for "tyre" is a borrowing from French.

Though the logic would be faulty, it seems that our word ultimately does come from French.

WeepingElf wrote: ↑Tue Jan 11, 2022 8:58 am And @keenir - Gauls are Celts. Talskubilos is right on this point.

But in how many of the three parts of Gaul? However, Gaulish is a Celtic language.

Linguoboy wrote: ↑Tue Jan 11, 2022 11:26 amThat's not actually an answer to the question I asked. Michelin makes great tyres, but it doesn't follow from that that our word for "tyre" is a borrowing from French.

Actually, they had got also Ferrari.

2+3 Clusivity wrote: ↑Mon Jan 10, 2022 5:39 pmI do not subscribe to this, but it would make much more sense to look at something like the truncated PIE formant /*(H)wi-/ for two (c.f. "twenty" and "apart/again")

How does the std isolationist PIE model explain the relationship between the numeral '2', this formant and the above dual suffix?

Talskubilos wrote: ↑Sat Jan 15, 2022 5:54 am

2+3 Clusivity wrote: ↑Mon Jan 10, 2022 5:39 pmI do not subscribe to this, but it would make much more sense to look at something like the truncated PIE formant /*(H)wi-/ for two (c.f. "twenty" and "apart/again")

How does the std isolationist PIE model explain the relationship between the numeral '2', this formant and the above dual suffix?

Did you miss the "styrofoam cup with the bottom cut off" reply?

Or was there another reply?

Talskubilos wrote: ↑Sat Jan 15, 2022 5:54 am

2+3 Clusivity wrote: ↑Mon Jan 10, 2022 5:39 pmI do not subscribe to this, but it would make much more sense to look at something like the truncated PIE formant /*(H)wi-/ for two (c.f. "twenty" and "apart/again")

How does the std isolationist PIE model explain the relationship between the numeral '2', this formant and the above dual suffix?

The usual explanation is dissimilation from *dwi- when attached to a root starting in a dental (and note that every example attested outside of Indo-Iranian is in fact immediately followed by a dental!)

KathTheDragon wrote: ↑Sat Jan 15, 2022 10:48 am

Talskubilos wrote: ↑Sat Jan 15, 2022 5:54 am

2+3 Clusivity wrote: ↑Mon Jan 10, 2022 5:39 pmI do not subscribe to this, but it would make much more sense to look at something like the truncated PIE formant /*(H)wi-/ for two (c.f. "twenty" and "apart/again")

How does the std isolationist PIE model explain the relationship between the numeral '2', this formant and the above dual suffix?

The usual explanation is dissimilation from *dwi- when attached to a root starting in a dental (and note that every example attested outside of Indo-Iranian is in fact immediately followed by a dental!)

And then there's also the distinct possibility that all three things don't have any relation between themselves whatsoever.

It's certainly possible that *Hwi- is not a dissimilated variant of *dwi-, but it would be very surprising indeed if the prefixal part of "twenty" was not! Given that the stem *-(h₁)ḱm̩t- is the "decad" portion, itself likely related to *deḱm̩t "ten", and all the other decads do have an element transparently being or related to the corresponding unit.

KathTheDragon wrote: ↑Thu Jan 20, 2022 12:08 am It's certainly possible that *Hwi- is not a dissimilated variant of *dwi-, but it would be very surprising indeed if the prefixal part of "twenty" was not! Given that the stem *-(h₁)ḱm̩t- is the "decad" portion, itself likely related to *deḱm̩t "ten", and all the other decads do have an element transparently being or related to the corresponding unit.

I remember reading the opinion of a renowed IE scholar (Kortlandt? Schmalstieg? Can't remember who it was) that the decade system doesn't go back to PIE, but was developed in parallel in the daughter branches (which may have influenced each other in this) - that would explain the pesky differences in e.g. the composition vowels in the individual branches that other scholars like e.g. Szemerenyi have valiantly tried to reconcile. But "20" still looks strange enough, and exactly the fact that it doesn't have **dwi- seems to indicate that it is old.
Another explanation would be the sporadic development *d -> *h1 that has been proposed by some from the Leiden school.

hwhatting wrote: ↑Thu Jan 20, 2022 11:34 am Another explanation would be the sporadic development *d -> *h1 that has been proposed by some from the Leiden school.

This is essentially the same development as the dissimilation I noted - the best examples of the "Kortlandt effect" are exactly the ones where it can be blamed on dissimilation.

KathTheDragon wrote: ↑Thu Jan 20, 2022 10:07 pm

hwhatting wrote: ↑Thu Jan 20, 2022 11:34 am Another explanation would be the sporadic development *d -> *h1 that has been proposed by some from the Leiden school.

This is essentially the same development as the dissimilation I noted - the best examples of the "Kortlandt effect" are exactly the ones where it can be blamed on dissimilation.

Ah, okay, then we're talking about the same thing.

hwhatting wrote: ↑Fri Jan 21, 2022 5:14 am

KathTheDragon wrote: ↑Thu Jan 20, 2022 10:07 pm

hwhatting wrote: ↑Thu Jan 20, 2022 11:34 am Another explanation would be the sporadic development *d -> *h1 that has been proposed by some from the Leiden school.

This is essentially the same development as the dissimilation I noted - the best examples of the "Kortlandt effect" are exactly the ones where it can be blamed on dissimilation.

Ah, okay, then we're talking about the same thing.

Yes, but given the mild difficulties in actually identifying the residue of *d with *h₁ I prefer to talk about it as "dissimilatory loss of *d" rather than specifically *d > *h₁.

KathTheDragon wrote: ↑Fri Jan 21, 2022 8:17 amYes, but given the mild difficulties in actually identifying the residue of *d with *h₁ I prefer to talk about it as "dissimilatory loss of *d" rather than specifically *d > *h₁.

In the framework of the PIE glottalic theory (e.g, Gamkrelidze-Ivanov), this would be explained as tˀ > ʔ.

I'm deeply skeptical of all the supporting arguments for most versions of the Glottalic Theory, so I disregard this interpretation.

I am also deeply skeptical of the glottalic theory. If you ask me, the *D set may once have been voiced spirants, as they are reconstructed for Proto-Uralic and Proto-Eskimo-Aleut, and *h₁ was [h], so you get a debuccalization ð > h.