June 18, 2013

Mellars 2013: second round

As I mentioned before, I have already got copies of the controversial study by Paul Mellars et al., which argues for a very late colonization of Eurasia. It includes some aspects not dealt with in the first round, when I could only access the supplemental material. 

Paul Mellars et al., Genetic and archaeological perspectives on the initial modern human colonization of southern Asia. PNAS 2013. Pay per view (6-month embargo) → LINK [doi:10.1073/pnas.1306043110]

Maybe the most important is the very striking visual comparison between proto-LSA African microlithic industries and post-UP South Asian microlithic ones:

While it is maybe easy to dismiss the patterns drawn on ostrich shells in Africa and South Asia as not really looking the same at all and therefore likely coincidence, the visual comparison of the industries is much harder to reject. It does indeed pose a mysterious apparent link similar to others that are hard to explain like the similitude between Chatelperronian and Gravettian (not so long ago treated together as "Perigordian") or the hammering insistence by some rather marginal academics on the similitudes between the SW European Solutrean culture and the (much more recent) North American Clovis industry. 

Sure: impressive and intriguing. But when it comes to chronology the Mellars hypothesis seems to fail terribly. While the African microliths are pre-LSA and therefore from before ~49,000 years ago in all cases, the South Asian ones only show up mostly since c. 34-38,000 years ago, more than ten millennia later. Mellars makes this figure 40-35 Ka and then just 40 Ka for the following graph, which in fact misrepresents Petraglia's model and data in a key issue (see below):

It must be emphasized here that Petraglia's data and model, at least for what I know it, implies an hiatus between c. 110 Ka and c. 80 Ka BP, hiatus for which there is no archaeological data of any kind in South Asia. Therefore neither side graph should suggest continuity to the past before ~80 Ka, allowing at most for a highly hypothetical dotted line (as in Petraglia 2010):

Also there is nothing in Petraglia's work that could suggest discontinuity at the Toba ash layer, as suggested by Mellar's version, rather the opposite: continuity is very apparent in Jwalapuram:

Jwalapuram industries (from Petraglia 2007)

Quite conveniently Mellars ignores Petraglia's data again, which suggest continuity before and after microlithism in Jurreru Valley and then also finds a transition towards UP ("blade and bladelet", as well as "backed artifacts") technologies since c. 34 Ka BP. 

But regardless, I'm pretty sure that Prehistory-savvy readers have already noticed a major issue in all this chronology: we are talking of dates that are almost 20,000 years after the colonization of West Eurasian by H. sapiens with "Aurignacoid" technologies, which are dated to before 55 Ka BP in Palestine (OSL), to c. 49 Ka BP in Central Europe and to c. 47 Ka BP in Altai (C14 calibrated). 

And those who are also familiar with Eurasian population genetics are by now shaking their heads in disbelief and claiming to heaven and hell alike. Because West Eurasians derive, at a late relative date, from Tropical Asians and therefore, if our core ancestors were already separated before 55 Ka BP, there is just no room for the Tropical Asian (and Australasian) expansion that must have preceded the Sapiens colonization of the West Eurasian Neanderlands.

(Those unfamiliar with the basics of Eurasian population genetics, see here).

So there is no way that the Out of Africa migration could be dated to just c. 55 Ka BP, as Mellars does (after grabbing hard the burning nail of conjectural coastal sites now under the sea, which would have to account for some 15-20,000 years of Eurasian prehistory on their own).

In fact it is also impossible from the viewpoint of Australian chronology, which again needs to go after the settlement of Tropical Asia but surely before that of West Eurasia. 

So, regardless of the striking visual comparison between African and Indian industries, which is no doubt the "bunny in the hat" here, the Mellars hypothesis simply doesn't stand. 

Was there another cultural (surely not demic) flow from Africa to South Asia c. 40-35 Ka BP? Maybe. Or maybe it is just one of the many hard-to-explain coincidences in stone industry design. But whatever it is, it just cannot be the Out-of-Africa migration, unless one is ready to accept that Aurignacian and related European rock art, as well as Australian rock art, for example, are the product of archaic homo species (something that I am sure that Mellars won't admit to: it just goes against his "modern human behavior" prejudices). And, even then, it just doesn't add up either.

PS- Petraglia himself finds Mellar's alternative model untenable. From ABC Science (emphasis mine):

... Professor Michael Petraglia, an archaeologist from the University of Oxford disputes Richards' and Mellars' argument. 

Petraglia says there's not enough evidence to rule out an earlier colonisation before the eruption of Mount Toba. 

"The research reported by Mellars and colleagues is riddled with problems," he says. 

Petraglia says that the similarity between tools used in Africa 60,000 years ago and those from Asia dating to around 35,000 years ago is not a consequence of direct migration.

"These toolkits are separated in time by more than 20,000 years and distances exceeding several thousand miles." 

He questions the evidence supporting a migration along the coast. He says that surveys of ancient shorelines have not revealed any evidence for human settlements anywhere along the Indian Ocean shore between 55,000 and 50,000 years ago.

He also says genetic dating should be treated cautiously. 

"Most geneticists will admit that genetic dating of the out-of-Africa event is tenuous, at best. Published genetic ages for out-of-Africa range anywhere between 45,000 to 130,000 years ago.

Petraglia says his team is currently conducting archaeological fieldwork in Arabia, India and Sri Lanka they expect will show that the story of human dispersal from Africa is complex.

"What we can agree on is that little research in these key geographic regions has been conducted and much more evidence needs to be collected to support or refute the different theories," says Petraglia.


  1. I was going to point out something quite obvious for me, but Petraglia already did: "These toolkits are separated in time by more than 20,000 years and distances exceeding several thousand miles."

    I can add that, if you dont care to make an artificial selection of one.by.one picked lithic pieces, and you put them in a figure, you can fabricate similar comparisons from many other places and chronologies (Neolithic, Mesolithic, even from what maju's referes as Aurignacoid assemblages)

    1. I'm glad that you also see it this way. It seems that, other than the authors themselves, this paper has persuaded very few people.

      "... what maju's referes as Aurignacoid assemblages"...

      I know that "Aurignacoid" is a controversial concept and that you particularly are not too keen to accept it. However if something supports its genuineness is that the various industries grouped under this umbrella clearly belong to the same space-time and therefore seem to indicate the same broad process of AMH expansion in a geographically (and genetically) quite homogeneous region (West Eurasia). The particulars of each techno-culture grouped under this "umbrella" term and their connections with each other are of course debatable.

      In any case, what I find most fascinating in this paper (but is quite overwhelmed by the controversial aspects of it) is the fact that microlithic technologies existed in Africa long before they showed up in Eurasia (the oldest one seems to be indeed from India) and even (West Eurasian) Upper Paleolithic and (African) LSA, being maybe a precursor to this. It raises many questions in fact, not just about the microlithic industries themselves but also about the true origins of "mode 4" or Upper Paleolithic blade-based technologies.

      I really hope that more research comes out on this matter of the origins of LSA, whose chronology and particulars are not too well defined for what I know. I suspect as very plausible that, even if LSA is confirmed to be derived (to some extent only?) from West Eurasian UP, this can also be conceptually influenced by African microlithic ideas.

    2. Not to polemize, but as a general comment: there are many different types of "microlitism", the one suggested by Mellar's adhoc figures would be just one. I'll propose not to overdimension it. It seems to me that this could easily be a "microlitism" associated with just one very specific "functional niche", which is the composite-armored spears or (on later periods) arrows.

      Then, it represents just a very small (yet probably significative) part of the productive/economic system of those societies.

      In my view, that very specific "kit" can easily appear and dissapear, depending on the social and economic requirements of the groups, and the capability for innovation of the whole society/culture.

      Supposing a H-G "society" which dominates the lithic production as a whole concept, which lives in a non static (changing) social and climatic enviroment, and has an economic system favouring a certain structure of hunting groups and tactics, I believe that: (re) invention (thus, innovation whitin their specific socio-historical context) is the most pausible explanation for those microliths.

      So, i think: this is probably what happened along the eurasian (pre)history. The concept "microlithic modular pieces of hunting weapons" was re(invented)in different groups and times.

      In fact, i think there's a clear example of it (re-invention): The (western Europe) mesolithic microlithic traditions (perdurating into the Neolithic), precisely associated with very similar composite-armored hunting weapons.

      On the other hand, ther's another thing to be said: Probably, instead of limiting our scope to pursue this "rabbit-out-of-the-hat", we should look also at the other 99% of lithic production of those human groups, and the restant 100% of other tools in other materials (as bone). Also, the structure of their setlements and their distribution on the landscape. Comparing the whole seems a more reasonable that just comparing which is "strinking" similar. If only, to prove or falsify that the "striking" similarity is not just an artifact of researcher volition.

    3. In total agreement, Millán. You explain your position particularly well on this occasion also, what is appreciated.

  2. Maju,

    I agree with your assessment, completely.

    which are dated to before 55 Ka BP in Palestine (OSL), to c. 49 Ka BP in Central Europe and to c. 47 Ka BP in Altai (C14 calibrated).

    These dates are about 5,000 years earlier than those I remember - any references?


    I agree. Everything I have read points to the fact that there is no 1:1 correlation between AMHs and blade technology - at least not until ~30,000 - 40,000 ya. As I have posted elsewhere in regard to this article, even some of the very first opportunist AMH Europeans used crude stone technologies - because they didn't need more effort (likely, their hunting targets, environments, and strategies were so different from Neanderthals that initially, the prey wasn't afraid/ aware of them).

    1. For Emirian: http://www.aggsbach.de/2010/10/the-initial-upper-paleolithic-of-the-negev/

      A child burial was found at Taramsa-1 dating to this time (c.55 k.a BP): “The poorly preserved bones were those of a subadult ‘anatomically modern human’ similar in appearance to the Mechtoid populations of the north African Epipalaeolithic., referenced to Midant-Reynes 1992/2000 p.37.

      There are other (later) dates and interesting references in that blogpost and some of them strongly suggest a link between Emirian and Bachokirian, which other references consider an Aurignacian or Bohunician variant.

      For diverse Aurignacian, proto-Aurignacian and other European Aurignacoid (as well as other MP-UP transitional cultures') I'm using a table suggested by Millán and John Hawks, which is very up to date: direct download. The oldest date as of now for European Aurignacian (since a recent revision) is some Swabian site dated to ~49 Ka calBP, followed soon by the "true Aurignacian" or Aurignacian I of Istallosko (Hungary) c. 48 Ka calBP, as well as some others.


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).