Monday, July 6, 2015

A Native American Linguistic Megafamily

A 2006 paper finds that the three North American language in a band from the Atlantic to the Pacific, immediately below the lands where Inuit and Na-Dene languages are spoken (i.e. Algic, Kutenai and Salish) are part of the same language family.  These language families combined were spoken in the lion's share of what is now the most densely populated part of Canada, near the U.S. border.

This finding would add to an increasing tendency of linguistics studying North American languages to fit languages previously described as language isolates or small language families into larger language families, a trend that has been particularly notable in cases involving languages spoken in what is now California.

19 comments:

terryt said...

I'll bet that German disagrees with the conclusions. However I have long thought the plethora of Native American languages surely must resolve into a relatively small number of language families.

Maju said...

Calling it "mega-" family is clearly excessive, "super-" (above, superior) is the usual term for this kind of findings, "macro-" (large) in the case of (usually just conjectured) large groupings of various families (for example Nostratic, so far a mere speculation). "Mega-" (mighty, huge) is never used. How would you describe Amerind then? Tera-family, Peta-family?

andrew said...

My view is that hyperbole in headlines is a tolerable sin, so long as the stories that readers are lured into reading are accurate.

bellbeakerblogger said...

This out today

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0130331

Maju said...

@BBB: is it really interesting? I quickly run over a strong inconsistency:

Out of the 832 haplotypes, 232 (27.9%) and 419 (50.4%) fell in the hunter-gatherer (HUNT; Hgs U, V) and farmer (FARM; H, T, K, J) groups, respectively. The remaining 173 samples represented haplogroups D, HV, I, N, R, W, X and Z.

For all we know, V is a "farmer" lineage (not detected anywhere prior to Neolithic, although it could still originate in UP Europe), while H is a "hunter-gatherer" lineage (detected in European populations before the Neolithic, both in Iberia and Karelia, maybe also Greece). Also I, N* (usually N1), W, X, etc. are all "farmer" lineages. HV should be considered Paleolithic probably, while D is clearly from the Siberian Bronze Age.

Seems junk-in: junk-out. Although maybe there's some raw data that still has some interest, the interpretation is clearly unacceptable.

andrew said...

@BBB Thanks for the tip.

I'm not the least surprised by the bottom line result, although I'm not sure I'd explain why I am not surprised in the way that they explained their results.

I'm also intrigued about a blog comment from one of the Tarim Basin ancient DNA researchers at Eurogenes referencing unpublished work showing that the R1a found in the Tarim basin mummies was a European rather than an Asian hg.

Maju said...

That last probably refers not to Tarim Basin but to Afanasevo R1a, which was shown not to be Z93, and hence plausibly (but not necessarily) of European origin.

andrew said...

Nope, the reference is to a Tarim Basin cemetery.

http://eurogenes.blogspot.com/2015/07/another-look-at-ancient-mtdna-from.html

Maju said...

I stand corrected. Thank you.

Ryan said...

This linguistic superfamily corresponds well with genetic evidence - it essentially corresponds to Reich's second inland migration.

I'd be curious if you have any thoughts on the recent paper on the supermassive blackhole CID-947.

http://www.sciencemag.org/content/349/6244/168.abstract

Can this be reconciled with modified theories of gravity? Would it be some sort of primordial black hole, or is there some other way to explain

Maju - Apologies for never doing that calculation re: Jewish origins. Been swamped. I saw a neat paper on IDB for Jewish communities and host populations - Basques and Ashkenazi Jews had the highest IDB sharing of any Jewish/non Jewish pair tested. Now, a lot of really interesting populations were omitted, but correct me if I'm wrong here - there's never been a large Jewish population in Euskadi lands has there? So this would have to be a signal from the Neolithic, correct?

Maju said...

There were some Jewish communities in the Basque Country prior to the Castilian invasion of 1512, also Muslims, later Huguenots. In general Basque independent states protected minorities. The most famous was Benjamin of Tudela, a Marco Polo style Basque Jewish traveler. AFAIK in the North (annexed to France) there was never any expulsion.

However I don't think that means any sort of Basque-Jewish relationship. I don't know about IBD but haploid markers just don't relate, certainly not Y-DNA.

capra internetensis said...

@Ryan

Are you referring to "Reconstructing Native American Population History"? That didn't include any Salishan or Kutenai samples, only Algonquian, and they clustered with the Athapaskan-speaking Chipewyan, so I don't see how it has any bearing on this proposal.

The recent Kennewick Man paper has some D statistics which are somewhat relevant. The Colville are mostly southern Interior Salish speakers (though some are Sahaptian), and two northern Interior Salish groups (both Shuswap) were also included. D(Yoruba, Colville; X, Karitiana) is approximately zero for most populations, including Central and South American ones, and (suprisingly) . It is significantly negative for the Shuswap, as you'd expect (though not by very much), but it is positive for Algonquin, Ojibwa, Chipewyan, and Tlingit - in fact it is most positive for Ojibwa, i.e. they are the *furthest* from the Colville out of the whole set (not counting Eskimos of course).

So I don't think that the genetic evidence, such as it is, particularly favours the hypothesis. However, the Northern North American (Na-Dene-related?) element in the Algonquians could be obscuring their connection to the Salishan people, if it exists. Something like IBD sharing might pick it out, or detailed analysis of Y chromosomes. There was nothing obvious in the ADMIXTURE analysis in this paper, but a North American targeted one with a large sampling of native groups (we should be so lucky) might find something.

Ryan said...

"Are you referring to "Reconstructing Native American Population History"? That didn't include any Salishan or Kutenai samples, only Algonquian, and they clustered with the Athapaskan-speaking Chipewyan, so I don't see how it has any bearing on this proposal."

Yes, I'm referring to that. No they did not include Kutenai or Salish, but the Kutenai and Salish have had the same distinctive genetic markers found on their Y DNA and Mitochondrial DNA. I'm trying to find a source but a bunch of papers on the genetics of fish and links to a post on Linear Population Model are confounding me at the moment. I believe it was mtDNA haplogroup X2 and Y-DNA haplogroup C3 that had links to the Kutenai and Salish groups though, as well as Algonquian groups. Found this off hand but haven't had time to digest it yet: http://ufdcimages.uflib.ufl.edu/UF/E0/04/28/69/00001/wilson_j.pdf

@Maju - That's what I thought. I'd note the second highest non-Jewish/Jewish IDB pairing was between Sardinians and Ashkenazi Jews. The third was Tuscans and Ashkenazi Jews. For non-Ashkenazi Jews, the closest non-Jewish pairing is between Greek Jews and Sardinians and between Turkish Jews and Sardinians.

What do Tuscans, Sardinians and Basques have in common? Relatively late use of non-Indoeuropean languages. So my thinking is that rather than some sort of massive Basque-Jewish or Jewish-Basque community, this signal is rather due to relatively less Indo-European admixture.

Now granted the number of non-Jewish populations included is rather low and the selection rather poor. French, Russian, Tuscan and Basque populations are the only European ones, and there are no North African samples either. But I thought it interesting none the less.

http://www.sciencedirect.com/science/article/pii/S0002929710002466#gr3

Maju said...

@Grey: the IBD "relationship" (fig. 3A of Atzmon 2010) between Basques and Ashkenazim is weaker than between Basques and Russians, so I never thought it could have any relevance at all. Also it's average total sharing and without consideration of segment length, what is also of little significance. It probably just indicates some European admixture in this population at their own founder effect (as distinct from other "Roman Jews") which surely took place in Marseilles around 700 CE. All the highest IBD relations of modern Jews correspond to Ashkenazim and the related populations are all surrounding Marseilles (Basques, Sardinians, French and North Italians). A similar situation happens with Italian Jews (French and North Italians) and in a "convoluted" way to Greek Jews (Sardinians and Basques, suggesting true Spanish Sephardic origin, at least in part - I've watched Sephardic Jewish Medieval style chant in Thessalonique and is plain Spanish, barely archaic what they used to speak, much as you could read in La Celestina or El Quijote). So in general all those European-Jewish lesser IBD relations must reflect European admixture. It'd be probably even more clear if there was a Portuguese or Spanish sample (Basques, Sardinians and French surely act as proxy for that primary source, at least among Sephardites).

Ryan said...

Agree with all your comments re: methods. I feel like there are some great ideas in this paper that would be so much more insightful if applied in a broader way.

"All the highest IBD relations of modern Jews correspond to Ashkenazim and the related populations are all surrounding Marseilles (Basques, Sardinians, French and North Italians)."

I think the relative ordering of this is interesting still. Re: Marseille - you'd think Tuscany would rank higher if this was a Occitan coast thing, no?

"So in general all those European-Jewish lesser IBD relations must reflect European admixture."

Must they? I agree that most would, but all? Why would Basque rank higher than French or Tuscan or Sardinian?

Maju said...

"Re: Marseille - you'd think Tuscany would rank higher if this was a Occitan coast thing, no?"

No. Why? Occitania has no particular relationship with Tuscany in terms paleohistorical nor historical, nor genetic either as far as I know. In the Paleolithic it was part of the Franco-Cantabrian region, in the Neolithic of the Cardium Pottery area, in the Chalcolithic (without copper but Chalcolithic anyhow) of the Chassey-La Lagozza complex (proto-Ligurian surely), participating also of Megalithism and Bell Beaker (SW province). Tuscany is unrelated to all that except Cardial. In population genetics, Tuscany also somewhat anomalously "oriental" even within peninsular Italy, what is usually attributed to Etruscan colonization bringing Anatolian or similar genetics (very diluted but noticeable).

"Must they? I agree that most would, but all? Why would Basque rank higher than French or Tuscan or Sardinian?"

Because Basques are (at least in some aspects) more closely related than all those to Iberians and surely also Occitans, neither of which is represented in the sample. So you should read "Basque" as "Ibero-Occitan proxy" almost certainly. My opinion anyhow.

capra internetensis said...

@Ryan

Thanks for the link, that looks really interesting

Ryan said...

@Maju - My issue with your Iberian hypothesis is that Basques aren't the top pairing for Sephardic groups. You'd think they'd have the clearest link to Iberia, no? I'm likely reading to much into this, but still.

@capra - glad you found it interesting.

andrew said...

"I'd be curious if you have any thoughts on the recent paper on the supermassive blackhole CID-947.

http://www.sciencemag.org/content/349/6244/168.abstract

Can this be reconciled with modified theories of gravity? Would it be some sort of primordial black hole, or is there some other way to explain"

Neither conventional dark matter theories nor modified theories of gravity have much to say about the size of the central black hole of a galaxy.

A primordial black hole is generally less than 3 solar masses, vastly smaller (and much less dense per volume within the event horizon) than this black hole which had a mass 10% of the entire galaxy v. 0.3%-0.5% which is typical today.

It strongly suggests that galaxies were assembled from the inside out, rather than in a parallel development to the galaxy itself as conventional wisdom suggested. Thus, first the black hole appears at the center and grows rapidly, and then gradually over billions of years the galaxy adds stars at the frontier while growth of the black hole itself slows dramatically after it gobbles up everything close to it.