The Great Macrofamily thread: Indo-Uralic, Altaic, Eurasiatic, Nostratic etc.

[dhok]

Muddying the waters--I don't think I saw Whim mention this, though I apologize if he did and I failed to catch it--is that Blackfoot is clearly Algonquian, but seems to have split off very early, so that the rest of Algonquian is itself a proper clade. However, Blackfoot is rather phonologically (and, to some extent, grammatically--I can't recall the details very well) innovative and it's not quite clear what it tells us.

So it's rather, in fact, as if you had:

--the Indo-Aryan languages (Algonquian)
--modern Persian (Blackfoot)
--Swedish (Wiyot)
--Portuguese (Yurok).

Note also that it's been proposed that Yurok and Wiyot may fit together in a "Ritwan" clade, though last I checked the consensus is that even if so they split apart so early that any shared innovations would be impossible to distinguish from retentions.

Furthermore, there have been attempted connections with some families of the Northwest Coast, particularly Salishan and Wakashan. I believe this is mostly a case of shared pronominal morphology (like 1st-person n-), but I don't have the time to check right now.

[Znex]

I suppose this question belongs here better than in other threads.

What's the rationale for m > b~w in proposed macro-language sound changes, eg. in Altaic and Indo-Uralic rather than the opposite: b > m~w? I would have thought b > m~w works better, as on one hand it's already attested in Turkic (for instance, *ben {1SG} almost unanimously > men), and in the case of IE it would help explain why *b is so infrequent in the PIE lexicon.

But I don't know whether either of these changes is more likely than the other crosslinguistically.

[Pabappa]

Is /b/>/m/ widespread, though, or sporadic? /b/ > /m/ in the neighborhood of a nasal might be common without it applying across the language.

[Znex]

Is /b/>/m/ widespread, though, or sporadic? /b/ > /m/ in the neighborhood of a nasal might be common without it applying across the language.

Yep, the Turkic example is b > m / _...N. If b > m~w is likely for PIE, I would think *w is the most common output.

[Whimemsz]

.

[WeepingElf]

I have made an observation about Mitian. Of the eight Mitian families, six show vowel harmony or vestiges thereof. There are two kinds of vowel harmony within Mitian, palatal (in Uralic, Turkic, some Mongolic languages and Yukaghir) and root retraction (in other Mongolic languages, Tungusic and Chukotko-Kamchatkan); in Mongolic, the two types of vowel harmony are cognate. Only two families show no trace of vowel harmony: Indo-European and Eskimo-Aleut. And these are precisely the two families which seem to have undergone massive reductions of the vowel inventory, which would have caused mergers of the two vowel harmony classes and thus obliterated vowel harmony completely. So, can we assume vowel harmony for Proto-Mitian? Or are the "reduced" vowel inventories of IE and Eskimo-Aleut archaic, and the other languages enlarged their vowel inventories by introducing vowel harmony?

[Nortaneous]

I have made an observation about Mitian. Of the eight Mitian families, six show vowel harmony or vestiges thereof. There are two kinds of vowel harmony within Mitian, palatal (in Uralic, Turkic, some Mongolic languages and Yukaghir) and root retraction (in other Mongolic languages, Tungusic and Chukotko-Kamchatkan); in Mongolic, the two types of vowel harmony are cognate. Only two families show no trace of vowel harmony: Indo-European and Eskimo-Aleut. And these are precisely the two families which seem to have undergone massive reductions of the vowel inventory, which would have caused mergers of the two vowel harmony classes and thus obliterated vowel harmony completely. So, can we assume vowel harmony for Proto-Mitian? Or are the "reduced" vowel inventories of IE and Eskimo-Aleut archaic, and the other languages enlarged their vowel inventories by introducing vowel harmony?

Vowel harmony is also known in Sino-Tibetan and Sahaptian. Some Qiangic languages have forms of vowel harmony that look a lot like what happens in Chukchi:

dʑa-ʁwasi 'elbow'
dʑaˤ-quˤ 'leg'
dʑaˤ-suˤ 'finger'

hanə 'two' - ɦaˤnəˤ-suˤ 'twenty'
ksi 'three' - ksiˤ-suˤ 'thirty'
gʐə 'four' - gʐəˤ-suˤ 'forty'

nu-la 'bring upstream'
nuˤ-staˤ 'pull out upstream'

Maybe that's just a coincidence - it happens in Arabic - but what are Sino-Tibetan languages doing with pharyngealized vowels in the first place? And in Nuosu, as far as I can tell, it developed contrastive vocalic +ARGH through creaky voice, and that harmonizes in compounds.

[mae]

-

[WeepingElf]

You couldn't assume vowel harmony for "Proto-Mitian" even if ALL "Mitian" branches had it, because that's not how reconstruction works. You need to show that the same specific alternation exists in cognate material, not the same kind of alternation in any material.

True. Unless one can show such cognates, such an idea is nothing more than a guess - perhaps useful as an idea where to look for what, but in itself meaningless.

[jal]

I recently learned that old-Mongolian did not have vowel harmony, so if that's true, that might be a nail to the coffin...


JAL

[WeepingElf]

Fair. I have since ditched the "Proto-Mitian vowel harmony" idea. Vowel harmony is exactly the kind of thing that is likely to spread areally, and the vowel harmony systems of the various Mitian languages are different enough from each other. For instance, there is nothing like the famous Turkic vowel cube in Proto-Uralic - just a vowel system that is reduced to a vertical 2-vowel inventory in non-first syllables, with each of the two vowels having a front and a back allophone according to the frontness of the first vowel.

[Nortaneous]

I recently learned that old-Mongolian did not have vowel harmony, so if that's true, that might be a nail to the coffin...

Janhunen reconstructs Proto-Mongolic with the harmonic pairs */a o u/ and */e ö ü/ and the neutral vowel *i, which he says is the merger of an earlier harmonic pair */ï i/.

[mae]

-

[Pabappa]

madagascar being Austronesian but with no austronesian languages in africa, india, etc is a good example, and that one is definitely real. there might be a few Dravidian langs in Pakistan. and within America itself there are some families like Algic that have a pretty lopsided distribution .... Algic is 99% eastern, with just two languages in California, and those two are themselves not closely related.

[Richard W]

Romani as Indic sounds a bit of a stretch. On closer examination, the question becomes 'From which part of India?'.

Hadza as Khoisan.

[Tropylium]

Basque–North Caucasian is probably not totally bonkers. (The proposal that it's moreover the same family as Dene-Yeniseian, though, kind of is.)

[Nortaneous]

Basque–North Caucasian is probably not totally bonkers.

?

[Moose-tache]

It doesn't bode well that "North Caucasian" is already bonkers.

[Nerulent]

I recently learned that old-Mongolian did not have vowel harmony, so if that's true, that might be a nail to the coffin...


JAL

Where in that video does it say that? It’s also not exactly a scholarly article. Old Mongolian didn’t have the specific labial harmony of some modern Mongolian languages but it certainly had the RTR vowel harmony.

[jal]

Where in that video does it say that?

Well, maybe it doesn't say so exactly, but here it says something about some harmony not yet occurring.


JAL