Clearly, there’s no archaeological evidence for an out-of-Africa migration. Scholars acknowledge it. But, lo and behold, they use nothing else but the peopling of the New World as the justification for this lack of archaeological evidence for an out of African migration. In a perverse logic, they absolve archaeology from the need to provide material evidence for the peopling of the Americas and rely on a speculative but firm consensus that this peopling event somehow happened. This speculative consensus is then made to look like “science” and is exported back as a yardstick for the archaeological standard needed to prove the out-of-Africa model. Seeing the gap in the archaeological record, they nod in the direction of craniology and genetics. But the study of the Hofmeyer skull that they quoted (Grine et al. 2007) actually showed that this 35,000-year-old South African specimen looks nothing like modern Khoisans, Pygmies or Negroids. It looks fully Eurasian (!) just like a bunch of Cromagnon skulls from the same period that are in fact related to modern European through straightforward morphological continuities. The genetic data they invoke, again, shows exactly the same biogeographic gap between African and Eurasian haplotypes with no “trail” connecting the two. For instance, mtDNA macrohaplogroup N is not indigenous in Africa, it expands in East Asia and shows up in Africa only in the form of derived clades such as N1 and U6. So we have a huge geographic gap between a putative African ancestor and the first non-African descendants. mtDNA macrohaplogroup M shows virtually the same pattern: it’s not indigenous in Africa, it expands in South/East Asia and it’s attested in Africa only in the form of a derived clade, namely M1.

So we have a perfect match between archaeology and genetics but the one that doesn’t support the out-of-Africa model but in fact directly contradicts it.

The cross-disciplinary support for the out-of-Africa model is therefore patched together with rumors and not facts. This, in turn, opens up a window of opportunity for an alternative model that is indeed based on facts, among which are the fact that there’s no archaeological evidence for the peopling of the Americas and there’s no archaeological trail for an out of Africa migration.

The obsession of prehistory buffs with skeletons and artifacts is rather noteworthy. There are two forms of cultural transmission: vertical (from generation to generation, from father to son) and horizontal (within a generation, from a friend to a friend). Languages are transmitted mostly vertically with some horizontal interference. Archaeological artifacts don’t belong to either type. They are objects that are left behind by the users (not passed down or along to other users) and deliberately collected by future non-users. Hence, archaeological artifacts constitute the most artificial datasets, whose relevance to the issues of origin, migration, evolution or common descent of naturally evolving human populations is rather tangential. Hence, going back to Nettle’s unfortunate article, benchmarking a language chronology by an archaeological chronology involves comparing apples and oranges, or better apples and tennis balls.

“Out of America – humans presumably descended from howler monkeys.”

Unlikely. But I would turn your attention to an interesting piece of research on human lice DNA. http://www.plosbiology.org/article/info:doi/10.1371/journal.pbio.0020340. Especially, the phylogeny on Fig. 2.

Let’s split the burden. We all have to prove something, otherwise science is meaningless. The fact that America was peopled at 15,000K and that Africa is the region with hominin continuities needs to be proven, too. People just forget about it.

As you admit that there are other factors (like language extinctions due to agriculture and other technology and economy-based factors or other factors like wars, climate, epidemics, etc.), we shouldn’t automatically assume that linguistic diversity increases with time. So the onus is also on you to prove your unproven claim.

Nettle’s counter-argument is no good because archaeological findings increase with time. On all continents. Some continents just accumulate these finds faster than others. It’s a plain fact. He assumes that they are fixed for good. Nichols’s argument (and I’m afraid you haven’t read her very carefully researched studies) predicts that more earlier findings are due to emerge in the New World. To question that linguistic diversity increases with time means denying the very basic principles underlying the discipline, namely that languages evolve through time and isolation. There are other factors, of course, but they haven’t been shown to be systematic. the onus is on you, onur, to prove this different.

Razib, it’s hard to identify what a linguistic mainstream is in this case. Greenberg’s Amerind has been rejected by all mainstream linguists, and I concur. Cavalli-Sforza’s attempt to map genes on Greenberg’s classification of world languages was criticized for using the wrong linguistic approach to justify genetic phylogenies. And I concur. An alternative approach to deep prehistory from the point of view of historical linguistics is Johanna Nichols’s “Linguistic Diversity in Space and Time.” In this book, she contrasted Europe and Africa with America, Asia and Sahul and concluded that our perspective on the earliest forms of human grammatical organization (specifically, head-marking with a long list of associated features) come from the latter and not from the former. This is entirely consistent with what I presented in “The Genius of Kinship” from the kinship studies perspective. And Nichols’s book was very well received.

Gcochran: The difference between you and me is likely only in the level of risk tolerance. Some people call the doctor after a single sneeze, others climb the Everest while blind. From the short note you sent along, I figured that you believe that Pleistocene extinctions were caused by humans. This is an old myth that you picked up from a trusted book, I understand. Please refer to more up-to-date research such as this one: Judy R. M. Allen et al., Last glacial vegetation of northern Eurasia. Quaternary Science Reviews, 2010, in which an ecological change is the key factor in Pleistocene extinctions. Human hunting was part of the process just because humans did hunt 10,000 years ago as they do now, but it’s not the cause for the extinction. To divine human prehistory on the basis of our putative destructive behavior is a mistake. There are two kinds of prehistoric myths: one attributes causes for a major observable fact to a natural catastrophe (Toba eruption, a supernova), the other to disproportionate human impact (megafauna extinctions in the New World, the mass replacement of foragers by agriculturalists in the Old World). I prefer to stay away from these stereotypes and only look at facts.

By the way: do you believe that the Sun rotates around the Earth? Not to be presumptuous, but I have a feeling you may look at Copernicus as another Renaissance idiot.

Matt: I agree there’s a danger of reifying “linguistic diversity” as an objective measure of time. And we do need to stay away from circular arguments. This is something to keep in mind. But one can’t, like Nettle does, assume that archaeology always provides the right picture, while linguistics has validity only if archaeology allows. Archaeology has a different pace of evolution: the finds are usually accidental and sporadic and don’t stop coming at the year 2010. Any find can challenge the beliefs “founded” on the previous find. There are plenty of examples thereof from recent and remote past. Linguistics directly works with a fundamental human faculty that differentiates us from hominins and it directly describes the extant populations whose origins we’re investigating. It needs to be explained in the context of human dispersals, evolution and adaptation. Archaeology works with circumstantial evidence. It benefits from the security of radiocarbon dating but by itself it furnishes no secure timeline for what we’re trying to understand. Modern human behavior, of which language is key, begins having identifiable archaeological correlates only at 45-40K throughout the Old World. This is close to the upper bound of linguistic diversity estimates in America in Nettle’s paper (and in Nichols’s paper that he tries to criticize).

Dave Chamberlin: Don’t you think that the expression “when the only tool you have is a hammer everything starts to look like a nail” best fits the archaeological/paleontological perspective on human evolution: if I don’t have a “record” of you, you’re not real? A series of comments on a blog post may make me argument look like a one-sided advocacy for one single parameter, namely “linguistic diversity.” It’s more complex than that. In “The Genius of Kinship” I identified several major lines of evidence converging toward some big theoretical alternatives to the out of Africa model. Plus what I think the levels of linguistic diversity reflect is a historical interaction of age, distance from homeland and population size. Africa and Europe are low in linguistic diversity because they are the furthest removed from the center of expansion, the colonization involved large expanding populations and they didn’t have enough time to accumulate enough new linguistic diversity after leaving the homeland.

For those who may be interested in some background, please check out http://anthropology.net/2008/05/12/the-genius-of-kinship-human-kinship-systems-and-the-search-for-human-origins/

more seriously, i haven’t followed this closely. my main objection to german’s thesis is simply that water is more likely to flow uphill with cultural phenomena (language diversity) than physical ones (archaeology, genetics). secondarily, i don’t believe that german’s thesis of deep commonalities of languages as he’s positing is currently in the linguistic mainstream, so it’s not even a symmetric comparison of archeology and linguistics.

‘s all i’ll say to that.

You’re welcome. Though his argument to me seems to run “There is no observable relationship between the archaeological record and linguistic diversity, in any region, except a slightly negative one; Therfore, as the archaeology is our only sure guide as to the time depth of habitation, we can’t infer anything about the time depth of habitation from linguistic diversity”.

That doesn’t seem very circular. You need some objective measure of time depth of habitation then you need to compare that to linguistic diversity before you can prove that linguistic diversity has any relationship with time depth of habitation. Otherwise you’d just be making a circular argument where you assume that linguistic diversity correlates with time depth of habitation without any empirical basis. This is true even if you have an empirical basis to make the assumption re: linguistic diversity over short time scales.

Let me guess: the dinosaurs committed suicide, too. And the continents don’t drift, right?

Who will rid me of this turbulent idiot?

Nom de Plume, you’ll be surprised to know that in early mtDNA studies all regional haplotypes radiated from what was referred to as “morph 1.” And this morph 1 was found at highest frequencies in the Americas, with all other populations showing a number of derived haplotypes that weren’t detected in American Indians. See, e.g., Fig 4 in “Mitochondrial DNA Polymorphism among Five Asian Populations”, S. Harihara // Am. J. Hum. Genet. 43:134-143, 1988. Then geneticists “re-thought” it and placed this formerly nodal haplotype all the down to the end of a MP tree (See Fig. 4 in that paper). But remained odd is that that haplotype was geographically most widely spread (highest frequencies in America, lowest in Africa, intermediary in Asia), so it couldn’t have been the most derived. That early tree looks identical to the one that DNA Tribes put together for the autosomes. See http://www.dnatribes.com/populations.html (Map of Interconnected Regions). I’m not affiliated with them in any way but I was struck how their tree is built around a primary split between the New World and the Old World and not between Africans and non-Africans.

No matter how outlandish it looks to you, John, I wouldn’t dismiss this possibility. Working under an assumption that America was peopled relatively recently, archaeology, over the past 150 years, hasn’t produced a consistent explanation of the origin of Paleoindian cultures from Asian Paleolithic (not surprisingly, scholars now entertain trans-Atlantic contacts to explain the Clovis phenomenon). We also now know that pre-Clovis existed. So, the door is wide open for alternatives. The data suggests that Africa and Europe have significantly less linguistic diversity than America, parts of Asia and Sahul. A simple out of Africa scenario of human migrations whereby a small subset of African populations left Africa to colonize the rest of the world flies in the face of this rather stubborn fact. How can linguistic (and other cultural forms of diversity such a kinship, myths, music, etc.) progressively increase from the supposedly oldest continent out, if populations that left Africa and then populations that left Siberia were small subsets of their original populations? What’s important is that this odd evolutionary process, whereby a small original population ends up exceeding the linguistic diversity in its parental population, while still showing clear signs of a genetic bottleneck, supposedly happened twice over the past 40-60,000 years. And the second time around the magnitude of this linguistic outburst exceeded the one that took place during the first wave of colonization. I can’t accept this. What I think may have happened is that at 12,000 years there was a backflow of Amerindian populations into Siberia and at 40,000 years there was a backflow of Asian populations into Europe and Africa (with European Upper Paleolithic deriving from such South Siberian sites as Kara-Bom that show similar technology at a slightly earlier date). This scenario of backflow admixtures at the two junctions where we see “bottlenecks” leaves us with a Sub-Saharan population, an Asian population and an Amerindian population at 40-60,000 locked in their respective continents.

America / more diversity = settled early; Africa / less diversity = settle later.

Another paper of interest is by Roger Blench: http://www.rogerblench.info/Linguistics%20papers%20opening%20page.htm#Amerind

I don’t know who gave you this “summary” but if you go to the original paper “The Settlement of the Americas: A Comparison of the Linguistic, Dental and Genetic Evidence” and open p. 494, you’ll see the date 20,000 for suggested “departure” from Asia, with other two migrations following at later points in time. If you open Wells’s “The journey of man”, p. 143 you’ll see the date of 12,000 for Amerind divergence with reference to Greenberg. On p. 333 of Greenberg’s “Languages in the Americas” you’ll find a direct tie-in of Amerind with Clovis and the dates of 11-12,000 YBP.

“The Nostratic hypothesis is controversial because it is rather weakly grounded in evidence , even though geographically historically plausible. Your theory seems to be equally conjectural but much less plausible. Calling it “areal” makes no sense; what kind of “area” is it that includes the Caucasus, New Guinea and Australia, and North and South America? It’s like you’re looking for the first Ur-language of Eden or the Noble Savage . Likewise, the long time frame you claim you need to explain the diversity of Amerindian languages works against the idea that correspondences between these three groups can be perceived after 40,000 years of separation.”

I don’t understand what “geographically historically plausible” means. Two solutions are compatible with it: borrowings that require geographic proximity and common descent. Parts of Nostratic, such as Altaic, is often dismissed as maybe borrowing-induced. I admit “areal” may be a misnomer but the very idea that similarities across vast distances may represent survivals from the time the ancestors of the speakers of these languages lived in close geographic proximity with each other (with two possible solutions, see above) is perfectly fine, I think. And in fact population genetics showed on a number of occasions that populations found in geographically distant areas can be genetically connected. And comparative method allows for the following assumption: while surface similarities disappear early on, deeper similarities may linger longer.

“There’s no difficulty in the idea that whole language stocks might disappear. Five of the six Yeniseian languages have gone extinct in the last two centuries, and Kettish has one fewer than 500 speakers left.”

Of course, they do but you can’t use an isolated example as a foundation for a theory. In America, hundreds of languages disappeared in the 19-20th century. While in Siberia, Yeniseian, AFAIK, is the only example. Again, empirically, the trend you are invoking is working against your argument.