I'll comment on this paper after I read it.
Mol Biol Evol (2010)
doi: 10.1093/molbev/msq247
Extended Y-chromosome investigation suggests post-Glacial migrations of modern humans into East Asia via the northern route
Hua Zhong et al.
Genetic diversity data, from Y chromosome and mitochondrial DNA as well as recent genome-wide autosomal SNPs, suggested that mainland Southeast Asia was the major geographic source of East Asian populations. However, these studies also detected Central-South Asia- and/or West Eurasia-related genetic components in East Asia, implying either recent population admixture or ancient migrations via the proposed northern route. To trace the time period and geographic source of these Central-South Asia- and West Eurasia-related genetic components, we sampled 3,826 males (116 populations from China and one population from South Korea) and performed high-resolution genotyping according to the well-resolved Y-chromosome phylogeny. Our data, in combination with the published East Asian Y-haplogroup data, show that there are four dominant haplogroups (accounting for 92.87% of the East Asian Y chromosomes), O-M175, D-M174, C-M130 (not including C5-M356) and N-M231, in both southern and northern East Asian populations, which is consistent with the proposed southern route of modern human origin in East Asia. However, there are other haplogroups (6.79% in total) (E-SRY4064, C5-M356, G-M201, H-M69, I-M170, J-P209, L-M20, Q-M242, R-M207 and T-M70) detected primarily in northern East Asian populations, and were identified as Central-South Asian and/or West Eurasian origin based on the phylogeographic analysis. In particular, evidence of geographic distribution and Y-STR diversity indicate that haplogroup Q-M242 (the ancestral haplogroup of the native American-specific haplogroup Q1a3a-M3) and R-M207 probably migrated into East Asia via the northern route. The age estimation of Y-STR variation within haplogroups suggests the existence of post-Glacial (∼18 thousand years ago, kya) migrations via the northern route as well as recent (∼3 kya) population admixture. We propose that although the Paleolithic migrations via the southern route played a major role in modern human settlement in East Asia, there are ancient contributions, though limited, from western Eurasia which partly explain the genetic divergence between current southern and northern East Asian populations.
Link
98 comments:
The Chinese have been shown to be more divergent towards Southeast Asian populations when amongst other Northeast Asians. It seems reasonable that Korea and Japan would have more northern migration elements thus making them further removed from SE Asians.
Its too bad they didn't test more populations from Korea and Japan.
The study tested for the rare haplogroup M335, which is a brother clade of R1b1*, having M343+ V88- M73- M269-. This haplogroup was first reported in Turkey in Cinnioglu (2004). I think FTDNA introduced testing for it recently, and that effectively a few people have turned up M335+, always in the Middle East to Central Asia region. Well, this study found a single sample of M335, and it was from Yunnan, in southeast China, on the border with Vietnam. Furthermore, since we already have a few M335 samples, a modal can be observed for this cluster, with several distinguishing values. The study included the haplotype data for the M335 sample, and its STR values effectively indicate it's a perfect candidate for the M335 cluster. In other words, this isn't an error, it really is a case of R1b1* sample (technically R1b1c per ISOGG) found way out in southeast China, on the border with Laos and Vietnam.
Remember when the big Karafet study of Indonesia came out earlier this year? It found 2 samples of R1b1* in Bali, Indonesia. That study didn't include haplotypes. R1b1* in Karafet's study could well have included M335 (they didn't test for M335). At the time, I pointed out that R1b1* is so exceptionally rare in both India and West Europe, the only reasonable candidates for explaining West Asian y-dna in Bali, that it seemed extremely unlikely that it would be due to such a historic reason. This M335 haplogroup, unlike other R1b haplogroups, has always had a notable eastern slant to its distribution, stretching from Levant and east Turkey, to Iran, Oman, and finally Kyrgyzstan and Kashmir in India. So it's very interesting that this very anomalously eastward leaning R1b haplogroup has now been confirmed in one of the most unlikely eastern places imaginable for an R1b lineage: southeast China. I'm strongly suspecting those 2 Bali R1b1* samples may belong to M335. If only we had the haplotype data, I could determine if they belong to the M335 cluster at a glance. I understand Karafet tested for haplotypes but didn't include them in the study?
And there's something else about R1b in East Asia from this study. They found 5 R1b1b2 (M269) samples from East Asia, and I mean really East Asia, not Xinjiang or Uygurs. Their haplotype data was also included and yep, they're all R1b1b2. The interesting thing is that 3 of the 5 have 393=12, indicating likely membership in the eastern branch of R1b1b2 (L11-, also known as ht35), typical of Anatolia, Middle East, Iran, and southeast Europe. Some time ago I had studied the yhrd samples to find any cases of R1b1b2, and I found 2 samples, I think from China and Korea, both of which had 393=12. So in total we have 7 R1b1b2 samples from China/Japan/Korea, 5 of which have 393=12. That's too much to be a coincidence, considering that in West Europe R1b1b2 has 393=12 at a rate of less than 5%. So no, these R1b1b2 lineages in China and beyond are not due to recent historic causes, from West Europe. The plot thickens.
The authors eem very sure that haplogroups D-M174, C-M130 (not including C5-M356) entered East Asia from SE Asia. There is quite a possibility that these two haplogroups represent migration into East Asia via the northern route more ancient than the post-Glacial one of 18 thousand years ago.
aargiedude, thanks for explaining that in a language I could understand.
The paper is clearly making a huge effort to retrofit new data into an established consensus. And sometimes the consensus shows glaring problems. There're no underived CF (BTW, this paper's phylogeny is a bit outdated by showing C and F as separate clades) and D lineages in West Asia, hence there's no support for the "southern route" in the first place. All the earliest C and D lineages are in (South)East Asia, and all the downstream ones, including MNOPS, L, J, G, H are there, too. Only a subset of these descendants of CFDE is found in West Asia, so the northern route must have been followed in the southwestern direction and not in northeastern direction.
The authors compare Y-DNA with mtDNA ("Several mtDNA haplogroups (H, V and X), showing signatures of post-Glacial expansions (Torroni et al. 2001; Reidla et al. 2003; Derenko et al. 2007), were detected in NEAS populations (Yao et al. 2004; Yang et al. 2008)" but again the case of X1/X2 in West Asia and Northern Africa, in the context of having most of the basal upstream N clades in East Asia, America and Australia, suggests that X1/X2 was carried into West Asia/Africa, rather than out of this area.
The authors even found E in East Asia, which is otherwise the most common haplogroup in Africa. But it's hard to say if it's underived, or they simply didn't test for all the known E markers. I can see how it can be underived: D, DE and E are all found in East Asia. If proved true, it would mean a definitive Asian origin for the major African clade. The problem is that East Asian E comes from such admixed groups as Hui and Uighur, so it can be a recent importation from West Asia. Again, the authors don't make it clear which way it is. And they strangely omit in Table 1 the dates for E in East Asia.
The authors separate C5 as being of "northern" origin and part of a later migration to East Asia from its sister C3 which is one of the most common markers in East Asia. They claim C3 is "southern" by origin. I think those two can't be separated from each other: they are all part of the original expansion of C1, C2, C3, C4 and C5 lineages that have an "eastern" distribution, with C5, C1 and C3 being mostly northern and C2 and C4 being southern. C5 is just the most "western" of the bunch.
"Although the CSA- and WE-related Y-chromosome haplogroups account for only 6.79% of East Asian males, it is noteworthy that 69.55% of them belong to two haplogroups Q and R, which probably represent early migrations from CSA and WE into NEAS via the northern route."
It would be good to see if R-P25 pops up in ancient remains in the New World. the current consensus is that all the instances of R-P25 in the Americas come from recent European admixture but neither Zegura et al. 2004, nor Malhi et al. 2008 reported any of the additional markers to determine whether R-P25 in America is in fact the most common West European R1b1b2 (R-M269) or underived R-P25. One of the main reasons why Zegura et al. rejected the indigenous origin of R-P25 in the Americas is because it's not found in Asia. Well, now it is, and per aargiedude those Asian R's are rare but old.
In any case, it's rather odd that Q and R, the two most geographically widely spread haplogroups are so far downstream from CFDE but they chose a unique northern route - to traverse the globe.
In the earliest studies of YAP+ Y-DNA, this clade's maximal STR variance has been found in samples from Japan. Without any preconceptions, the unbiased conclusion should be that DE-YAP has originated in Japan.
Ebizur,
I don't know exactly which early publication you're referring to, but I'm aware of "Out of Africa and Back Again: Nested Cladistic Analysis of Human Y Chromosome Variation
M. F. Hammer,* T. Karafet, et al. (1998), with a couple of shorter antecedents that placed the origin of YAP+ chromosomes in Asia. Not specifically in Japan but in (East) Asia.
In Altheide and Hammer, 1997 ("Evidence for a Possible Asian Origin of YAP' Y
Chromosomes") there's an interesting passage: "Circumstantial evidence lends some support to the
Asian-origin hypotheses. For example, African populations
are inferred to have had larger long-term effective
sizes than either Asian or European populations (Relethford
and Harpending 1994; Harding et al. 1997; Jorde et al. 1997). Also, theoretical results based on restricted gene-flow models indicate that ancestral haplotypes are
more geographically widespread than derived haplotypes created by more recent mutations (Templeton et al.
1995). However, in this case the derived 3A haplotype is found in many widespread populations from western, central, southern, and eastern Africa, where the ancestral 3G haplotype is absent (Hammer et al. 1997). Thus, the loss of 3G by drift in African populations is less
likely than either the loss of 3G in a small founding population migrating to Africa or the loss of 3A in Asian populations."
What this reasoning suggests is that a small founding population in East Asia migrated to Africa and exploded. Interestingly enough, as Dienekes noted some time ago, the YAP+ mutation emerged in East Asia at a time when America was already settled. No Y-DNA E or D have been detected in the New World. At the same time, their sister clade, namely hg C, is present in the New World.
Interestingly, among the haplogroups that are YAP-, hg Q and R dominate in that putative "northern" expansion. And they are wide spread in Northern Eurasia-America. I was wondering if Q and R can be still shown to be closer related to African YAP- haplogroups, namely A and B. If YAP+ chromosomes are rare in Asia but frequent in Africa, YAP- chromosomes are rare in Africa but common in Asia. There's a nice symmetry in here, I think.
"There're no underived CF (BTW, this paper's phylogeny is a bit outdated by showing C and F as separate clades) and D lineages in West Asia, hence there's no support for the "southern route" in the first place".
German, i agree the southern route is not well supported by all facts yet. However there is a third alternative to a southern route and the American origin you seem to support. It is based on several sociological assumptions that might be or not be confirmed by paleoanthropology:
1.early specialisation in Africa (A sabana, B jungle, C and derived coastal),
2.rule of decision when moving: when possible better keep specialisation than change, and when possible better move to free territory than to territory occupied by similar species,
3.patrilineal clans social organization and secession of younger clans (that is, not random secession).
The alternative route i´m thinking about is a coastal C route around Africa, continue the coastal route up trough south Mediterranean, Levant and Black sea up to North-caucasus (or maybe Levant). There a heavy occupation by Neanderthals force to “negociate” some kind of alliance with them and adapt to a new ecosystem. Then throught South-Siberia to north East Asia. From there to south up to India, and from there IJ to middle east and K and derived, to central Asia-Tibet. The likely Xinjiang-Tibet Branch M-N-O, following river routes (Yangtze, Mekong) occupies east-asia and Oceania respectively. Since the south siberia traversal was done in pre-LGM times all C-D lines has disapeared there. This explains why C-D has not beeing found in the supposed south coastal route, when there the LGM would had not caused disappearance of these lines. In summary a spiral-like north route which explains some of the pieces of the whole puzzle (some Y-haplogroups distribution, expansion to America, Neands admixture etc...) which do not square.
To my knowledge this route is, if not likely at least plausible. That is, AFAIK there is not known fact which makes it impossible. Am i wrong ? However i see some drawbacks that should be clarified: was it easier at this time to traverse coastal West Sahara than the coastal South Arabic peninsula, regarding for instance the availability of water ? Why Neands had not occupied far-east if Sapiens could occupy it at the same time ?
Interesting approach.
"Was it easier at this time to traverse coastal West Sahara than the coastal South Arabic peninsula, regarding for instance the availability of water ?"
I don't know. You have to check some literature on that.
I agree with early ecological specialization (whether in Africa or elsewhere) and I agree with the decision making model you've provided: occupy unoccupied territories first. Hence, an extreme southern coastal route, an extreme northern Arctic/Subarctic tundra route and a foot-of-mountain-range route (all, in essence, biogeographic refugia) are likely as the smartest ways to first colonize the globe. This assumes that mainland was colonized by some earlier hominin occupants who had been around longer, hence didn't need to stick to the extremes of biomes anymore.
I disagree with the patrilineal nature of early social organization. There's a general understanding on the part of social anthropologists that the earliest social organization was bilateral, hence flexible and easily adaptable to the exigencies of environment and the fission-fusion demography of the earliest foraging demes. Later migrations into already inhabited areas were associated with matrilocality and matrifocality, as men were engaged in external warfare with natives. Patrilocality is associated with internal warfare and is more appropriate for a population that is not expanding into a new territory. This is what the Divale theory says.
"In summary a spiral-like north route which explains some of the pieces of the whole puzzle (some Y-haplogroups distribution, expansion to America, Neands admixture etc...) which do not square."
I like it more than the straight southern route theory that's for sure.
"Why Neands had not occupied far-east if Sapiens could occupy it at the same time ?"
Neanderthal DNA was reported from South Siberia. We may find more traces of Neanderthals in East Asia, although Asian Homo erectus could've been a formidable competitor for Neanderthals to make too deep of eastward inroads.
Yes, besides genetic Y-chromosome and social anthropology data many other variables (mt-DNA, paleoclimatic,etc...) must be taken into account to square the whole puzzle. The southern route does not satisfy me either.
Bilateral i/o patrilineal, ok. I do not think this invalidates the possibility of a North-Saharan route but i have to think how this new information fits into the whole picture. Independently of this route issue i´m looking for an explanation of why some Y-chr. lines tends to predominate in some geographical areas. Neolotic elite polygamy could be one explanation; the old patrilineal social organization was the other.
Regarding Neands. presence in South Siberia and Erectus in Far East, thanks for the info. There should had been some Erectus replacement then in Far East.
Regarding where the North-Saharan route might have started in Africa i would say that the most possible place is West-Africa, maybe Senegal. More concretely, C could have been a B secession specialised in coastal ecosistems. This and the Sahara might explain its absence in Africa. I wonder then how the distribution of haplogroup E fits...
Basal paragroup E-M96 have been found in Bantu, Pygmies and, interestingly, in two individuals from Saudi Arabia.
See Saudi Arabian Y-Chromosome diversity and its relationship with nearby regions, by Khaled K Abu-Amero et al. This jibes well with the finding of the same paragroup in East Asia. I think E branched off from CFDE and came to Africa through the Middle East. Others believe that DE originated in the Middle East, and then D went east and E went to Africa. Still others trace everything Eurasian out of East Africa, which is the opposite from your West African focus. I think the paucity of any out of Africa signatures in the Middle East, while effective population size in Africa and the Middle East is great, suggests that it's less likely that some would-be Eurasian lineages got lost in Africa and the Middle East. It's more likely that some would-be African lineages got lost in Asia.
"The authors separate C5 as being of 'northern' origin and part of a later migration to East Asia from its sister C3 which is one of the most common markers in East Asia".
I read that once before, long ago. I hadn't heard any more about it so I thought it had been dismissed, but if true it would tend to invalidate any use of C's distribution as evidence for any 'southern coastal migration'.
"Without any preconceptions, the unbiased conclusion should be that DE-YAP has originated in Japan".
I certainly think it makes sense that E (or DE) is an immigrant into Africa. I probably wouldn't go so far as to place its origin as far east as Japan, which I doubt that Ebizur does either. But still it's interesting.
"Others believe that DE originated in the Middle East, and then D went east and E went to Africa".
I think that is the most likely situation. In fact I'm reasonably sure that just one Y-hap emerged from Africa: CT. It diverged in the region between the Middle east and Northern India, E in the west, C and D in the northeast, F in the southeast. E moved back into Africa replacing many A and B Y-haps, F moved into and through India as far as SE Asia, and C and D moved across Central Eurasia as far as Japan, then moved south: D over land and C via the coast to Australia. In SE Asia F gave rise to MNOPS, members of this haplogroup expanding once more from that region: M and S across Wallacea to New Guinea, NO north through China, and P back west through India to emerge onto the steppe, where it split. R moved west (although it seems to have actually formed in India) and Q moved northeast to reach America. Some time later C3 followed on into America.
"I think that is the most likely situation. In fact I'm reasonably sure that just one Y-hap emerged from Africa: CT. It diverged in the region between the Middle east and Northern India, E in the west, C and D in the northeast, F in the southeast. E moved back into Africa replacing many A and B Y-haps, F moved into and through India as far as SE Asia, and C and D moved across Central Eurasia as far as Japan, then moved south: D over land and C via the coast to Australia. In SE Asia F gave rise to MNOPS, members of this haplogroup expanding once more from that region: M and S across Wallacea to New Guinea, NO north through China, and P back west through India to emerge onto the steppe, where it split. R moved west (although it seems to have actually formed in India) and Q moved northeast to reach America. Some time later C3 followed on into America."
Now, you just need to prove all that, Terry. There are no A and B in the Middle East, there are no Cs or Fs or Ds in Africa, hence it's unlikely that CT originated in Africa. There are no underived Cs or Ds in the Middle East.
"I certainly think it makes sense that E (or DE) is an immigrant into Africa. I probably wouldn't go so far as to place its origin as far east as Japan, which I doubt that Ebizur does either. But still it's interesting."
We're making progress - at least you're interested. But do you need to put the origin of the CDE clade in the Middle East, if you have East Asia with at least D, DE and C represented there? And maybe E, too.
"There are no A and B in the Middle East"
Which suggests just a small subsample of it managed to emerge from Africa. It left no surviving descendants as it moved further east.
"There are no underived Cs or Ds in the Middle East".
Which also supports the idea that just a small subsample of it managed to emerge from Africa. C and D formed far from Africa. Interestingly F-derived haplogroups are found along what might be a route east from Africa: G, J and I seem likely to have originated somewhere in the Zagros/Caucasus mountains.
"But do you need to put the origin of the CDE clade in the Middle East, if you have East Asia with at least D, DE and C represented there? And maybe E, too".
I have mentioned elsewhere that it looks to me as though they originated on the Iranian Plateau. But I met with severe opposition to the idea.
Carpetanuiq. Although your comments don't appear here they do appear in my email.
"This is surely going too off topic, so this is my last comment about this issue in this thread".
I don't think it's too far off topic. The topic is 'Post-glacial migrations of humans into East Asia', although we have moved to Pre-glacial migrations.
"the region lacks of autochthonous clades of haplogroups M and N the authors suggest that the area has been a more recent receptor of human migrations".
To me that is possibly the most obvious objection to the 'Great Southern Coastal Migration Theory'.
"I would like to know how ancient, frequent and variant is N in South Asia".
N itself is not at all common in South Asia but its derived haplogroup, R, is very common indeed.
German said,
"We're making progress - at least you're interested. But do you need to put the origin of the CDE clade in the Middle East, if you have East Asia with at least D, DE and C represented there? And maybe E, too."
What do you mean by "the CDE clade"? There is no published data to support a grouping of haplogroups C, D, and E as opposed to haplogroup F. According to the currently accepted phylogeny, haplogroups C, D, and E may be grouped together only in opposition to haplogroups A and B. Please refrain from inventing terms like "the CDE clade" without any evidence to support your opposition to the generally accepted nomenclature, which at present endorses the grouping of haplogroup C and haplogroup F together as subclades of C,F-P143 and the grouping of haplogroup D and haplogroup E together as subclades of DE-YAP.
Y-DNA Haplogroup E definitely has been found in East Asia (or at least in Japan and China) in addition to some other areas that may be included in a broader definition of East Asia (e.g. Thailand, Tuva), but it has not been made clear whether these cases of haplogroup E are closely related to extant West Eurasian or African types of haplogroup E (and therefore may reflect recent admixture) or whether they are highly divergent from any Western type of haplogroup E (and therefore may have a long history in East Asia). I recall that Fornarino et al. have found Y-DNA haplogroup E1b1b1-M35(xE1b1b1a-M78) in several individuals in a sample of Tharus from eastern Nepal, but E-M35 is already quite far removed from the root of haplogroup E.
"What do you mean by "the CDE clade"? There is no published data to support a grouping of haplogroups C, D, and E as opposed to haplogroup F. According to the currently accepted phylogeny, haplogroups C, D, and E may be grouped together only in opposition to haplogroups A and B. Please refrain from inventing terms like "the CDE clade" without any evidence to support your opposition to the generally accepted nomenclature, which at present endorses the grouping of haplogroup C and haplogroup F together as subclades of C,F-P143 and the grouping of haplogroup D and haplogroup E together as subclades of DE-YAP."
I may have been slightly inaccurate in my spelling of the CDEF or CDF clade, but I didn't mean to imply that F has been firmly shown as being an outgroup for CDE. This is however a possible resolution of the CDEF clade (see Underhill and Kivisild 2007, Fig. 8c). Although Underhill and Kivisild opted for the DE vs. CF and for an African origin of the YAP+ mutation, by Chiaroni et al. 2009 (Underhill is one of the co-authors) the situation became less clear to them, as it's admitted that E could have returned to Africa ("It is also interesting to sum the distributions of different haplogroups descending from the same mutation, as for example D and E, which both descend from DE-YAP, the first mutation that split into the E branch that PERHAPS RETURNED TO AFRICA (or arose there), whereas the other branch, D, is found today mainly in the Himalayas and Japan." p20177).
The possibility of a back migration into Africa (also in Chandrasekar 2007 and Shi 2008) corresponds to scenarios b and c in Fig. 8 of Underhill and Kivisild. Scenario b groups DEF vs. C and scenario c groups CDE vs. F. In Karafet 2008, C and F were linked, so, yes, within the CDEF clade, DE is opposed to CF. But since E is the only African-specific clade, it's logical to suggest that it's nested within a larger non-African clade. The same situation is found in mtDNA, with a subset of non-African lineages such as N1, U6, M1 entering Sub-Saharan Africa at an early date. L6 may also have non-African origin, and it's at the root of the whole L3'4'5'6 clade.
What I find entirely unconvincing in existing Y-DNA phylogenies is that A and B, which are YAP-, supposedly give rise to DE, which is YAP+, and CF, which is again YAP-. If YAP- continues unchanged from A/B to C/F, then they should form the primary split in the phylogeny. It makes much more sense to group African A and B with non-African C and F and then African E with non-African D. That's what I suggested above in light of the fact that Zhong et al. hypothesize a northern (in which Q and R dominate) and a southern migration.
The next observation that I think is important is that the only two regions that are free of YAP+ are America and Oceania, and hence are most "unsullied" of all. It means they've been in isolation for a very long time.
"-DNA Haplogroup E definitely has been found in East Asia (or at least in Japan and China) in addition to some other areas that may be included in a broader definition of East Asia (e.g. Thailand, Tuva), but it has not been made clear whether these cases of haplogroup E are closely related to extant West Eurasian or African types of haplogroup E (and therefore may reflect recent admixture) or whether they are highly divergent from any Western type of haplogroup E (and therefore may have a long history in East Asia). I recall that Fornarino et al. have found Y-DNA haplogroup E1b1b1-M35(xE1b1b1a-M78) in several individuals in a sample of Tharus from eastern Nepal, but E-M35 is already quite far removed from the root of haplogroup E."
This is exactly what I wrote above, Ebizur. Thanks for the Tharu reference.
""There are no A and B in the Middle East"
Which suggests just a small subsample of it managed to emerge from Africa. It left no surviving descendants as it moved further east.
"There are no underived Cs or Ds in the Middle East".
Which also supports the idea that just a small subsample of it managed to emerge from Africa. C and D formed far from Africa. Interestingly F-derived haplogroups are found along what might be a route east from Africa: G, J and I seem likely to have originated somewhere in the Zagros/Caucasus mountains.
"But do you need to put the origin of the CDE clade in the Middle East, if you have East Asia with at least D, DE and C represented there? And maybe E, too".
I have mentioned elsewhere that it looks to me as though they originated on the Iranian Plateau. But I met with severe opposition to the idea."
Terry, you make all these assumptions of extinctions of genetic lineages. But then you use exactly the same assumption for languages. You adhere to the extinction ideas every time there's no evidence for out of Africa. And mind you, there's no archaeological evidence of an out of Africa exodus either. So you have gaps in the evidence across all the disciplines. I get your logic but we do need evidence to consider something firmly established.
"you make all these assumptions of extinctions of genetic lineages".
Surely it's obvious that some lineages have become extinct. No examples of CDEF* have been found, other than the four single individual haplogroups and DE*. No examples of CF* have been found. As well as those problems there are many examples of a whole series of mutations along many lineages that have left no descendants other than those with more derived chromosomes.
"And mind you, there's no archaeological evidence of an out of Africa exodus either".
Which is what I was drawing attention to earlier ('Which suggests just a small subsample of it managed to emerge from Africa'). The fact that today just one subset of Y-hap B and two subsets of mtDNA L3 survive outside Africa is hardly proof that just one man and two women emerged from that continent. On the other hand it is hardly proof of any sort of substantial migration either. It simply proves the emergence of three or four highly successful haplogroups.
Carpetanuiq. Again your comments don't appear here although they do appear in my email.
"A pure coastal south route (corn of Africa-->south arabia-->south iran-->India) makes more dificult to explain such admixture".
Very difficult indeed.
"Climate changes cause expansion or contraction of ecosystems; peoples expansion or contraction follow them until they are forced to adapt".
And that has always been the case since H. erectus first emerged from Africa. Hence it is an explanation for why we do not see a complete cline of haplogroups leading from either Africa (the standard model) or from America (German's model).
"Surely it's obvious that some lineages have become extinct. No examples of CDEF* have been found, other than the four single individual haplogroups and DE*. No examples of CF* have been found. As well as those problems there are many examples of a whole series of mutations along many lineages that have left no descendants other than those with more derived chromosomes."
Of course, lineages go extinct. This is understandable. But when the key pieces of evidence on which a theory is based are declared extinct, it becomes suspicious. One possible reason why no examples of CF* and CDEF* have been found is because the phylogenies that predict that they should have existed are not exactly accurate.
"Which is what I was drawing attention to earlier ('Which suggests just a small subsample of it managed to emerge from Africa'). The fact that today just one subset of Y-hap B and two subsets of mtDNA L3 survive outside Africa is hardly proof that just one man and two women emerged from that continent. On the other hand it is hardly proof of any sort of substantial migration either. It simply proves the emergence of three or four highly successful haplogroups."
Terry, you simultaneously claim that only a limited number of original lineages survived in the course of an exodus out of Africa and that only a limited number of derived lineages emerged as a result of the migration. But this is precisely where the data is unambiguously unsupportive on both sides: there are no African lineages outside of Africa, there's a plethora of non-African lineages outside of Africa and there're no early versions of non-African lineages in Africa.
"Hence it is an explanation for why we do not see a complete cline of haplogroups leading from either Africa (the standard model) or from America (German's model)."
I can't see Carpetanuiq's comments but I've always thought that there's a cline between Africa and America that's visible in all genetic distance analyses.
"Terry, you simultaneously claim that only a limited number of original lineages survived in the course of an exodus out of Africa and that only a limited number of derived lineages emerged as a result of the migration. But this is precisely where the data is unambiguously unsupportive on both sides"
The contradiction is easily solved if we're prepared to admit the OoA haplogroups formed hybrids with humans already outside Africa. That would explain why only a limited number of lineages were involved.
"The contradiction is easily solved if we're prepared to admit the OoA haplogroups formed hybrids with humans already outside Africa. That would explain why only a limited number of lineages were involved."
This admission only leads into a different contradiction. You claim that African humans admixed into a pre-existing hominin population outside of Africa, while all other proponents of hybridization suggest that it's hominin populations outside of Africa who admixed into modern humans. In any case, you gotta have some African-specific lineages attested outside of Africa to make your version of the hybridization argument. For instance, certain Amerindian-specific markers (found in North and South Amerindians) such as C3 are also found in (North)east Asia. This can be interpreted as Late Pleistocene-early Holocene admixture. But there's nothing of that sort between Africa and Eurasia. On the other hand, those who argue for admixture from archaic hominins into modern humans outside of Africa should pay more attention to such facts as 1) there were several migrations of Eurasians into Africa from 40K on, as attested by mtDNA M1, U6 and N1, in the very least. This may have accounted for the attested high diversity of Africans vs. Eurasians; 2) an admixture from archaic hominins would have made humans outside of Africa more diverse than humans in Africa. This is not the case.
"You claim that African humans admixed into a pre-existing hominin population outside of Africa, while all other proponents of hybridization suggest that it's hominin populations outside of Africa who admixed into modern humans".
There's very little difference between those two options. To me the most obvious position is that we have had gene flow through our species from when H. erectus/habilis/Australopithecus first left Africa. In fact I'm sure that gene flow had been going on long before then. Haplogroup distribution can tell us a great deal about this flow but they are by no means the complete story.
"1) there were several migrations of Eurasians into Africa from 40K on, as attested by mtDNA M1, U6 and N1, in the very least".
And I'm pointing exactly that out to Maju at his blog.
"2) an admixture from archaic hominins would have made humans outside of Africa more diverse than humans in Africa".
Not if, as you have just reminded us, the descendants of those archaic humans had then introduced that diversity back into Africa.
Terry, I just realized that with the discovery of basal E-M96 in Saudi Arabia (Abu-Amero, Saudi Arabian Y-Chromosome diversity and its relationship with nearby regions), it becomes possible to compare Y-DNA E with mtDNA L3'4'5'6. L6, as the most divergent lineage in the group, is attested in Yemen (at high frequencies) and Saudi Arabia. The high penetration of the mtDNA L3'4'5'6 clade into Sub-Saharan Africa matches well with the high frequencies of Y-DNA hg E in Sub-Saharan Africa. The intriguing part is that mtDNA geneticists haven't identified a precise correlate of Y-DNA D in Asia, a correlate that would be characterized by a unique relationship with a major African clade. L3'4'5'6 must be directly related to some select M lineages in West and East Asia. Maybe the strange deep-rooting M23 in Madagascar should be considered in this regard. http://www.biomedcentral.com/1471-2164/10/605. In any case, there's a lot to be desired from mtDNA phylogenies. Y-DNA clearly shows a major expansion of non-African lineages in Sub-Saharan Africa, probably correlated archaeologically with the incoming modern humans at 45-40K.
Correction: "mtDNA L3'4'5'6" read "L2'3'4'6."
German said,
"The intriguing part is that mtDNA geneticists haven't identified a precise correlate of Y-DNA D in Asia, a correlate that would be characterized by a unique relationship with a major African clade."
The presence or absence of an identifiable mtDNA correlate of a Y-DNA haplogroup is utterly irrelevant to resolving the question of the Y-DNA haplogroup's origin. What is the mtDNA correlate of Y-DNA P in the Americas (derivatives of Q1a3-M346) and Y-DNA P in Europe (derivatives of R1-M173)? What is the mtDNA correlate of Y-DNA C in Australia (C4-M347) and Y-DNA C in northeastern Asia and the Americas (C3-M217)? What is the mtDNA correlate of Y-DNA N (mostly N1c1-M178) in the circum-Baltic region and Y-DNA N (mostly N1*-LLY22g) in Southern China?
The lack of an identifiable mtDNA link between Sub-Saharan Africa and East Asia at the same time depth as the Y-DNA link between Sub-Saharan Africa and East Asia (DE-YAP) does not have any bearing on the question of the region of origin of DE-YAP.
Try telling Maju that, and then stand back.
I haven't been able to see any mtDNA Ls that might have moved back to Africa with Y-hap DE yet, but I've just started looking.
Ebizur:
"The presence or absence of an identifiable mtDNA correlate of a Y-DNA haplogroup is utterly irrelevant to resolving the question of the Y-DNA haplogroup's origin."
"The lack of an identifiable mtDNA link between Sub-Saharan Africa and East Asia at the same time depth as the Y-DNA link between Sub-Saharan Africa and East Asia (DE-YAP) does not have any bearing on the question of the region of origin of DE-YAP."
Ebizur, you just repeated the same meaningless statement twice. Correlating two haploid phylogenies is a useful exercise for a simple reason that humans live not in unisex communities but in families and men and women tend to migrate together. Certain evolutionary dynamics specific to men and women certainly apply, as for example, the rates of matrilocality and patrilocality but they shouldn't interfere with gross regional and continental patterns. What you seem to be advocating for is a chaotic pile of abstract mtDNA and Y-DNA lineages on the computer screen without a unified social and demographic process underlying them on the face of the earth.
And the similarities between mtDNA and Y-DNA haplogroup distributions abound.
1. Outside of Africa all mtDNA haplogroups cluster into two macro-ones: M and N. All Y-DNA haplogroups do the same: DE vs. CF.
2. In the same way as there're are no Y-DNA C lineages in Africa, there're no indigenous mtDNA N sequences in Africa (U6 and N1 come in later); in the same way as there're no Y-DNA M sequences in Africa (M1 comes in later), there're no indigenous Y-DNA F sequences in Africa. In both cases, there's a huge geographic gap between the putative African source for modern humans and the earliest offshoots of the major Eurasian mtDNA and Y-DNA clusters. Or in other words, in both cases (South)East Asian lineages aren't derived from West Asian and South Asian lineages.
2. Khoisans are outliers in both systems: Y-DNA A and mtDNA L0 is at high frequencies among them.
3. Y-DNA E is pervasive in Sub-Saharan and Northern Africa, so does the L2'3'4'6 clade, especially L2 and L3. E* was found in the Arabian peninsula, so was L6.
4. mtDNA S in Australia and mtDNA haplogroup A in North America and Northeast Asia are part of macrohaplogroup N. Y-DNA C4 in Australia and C3 in North America and Northeast Asia are both part of C.
5. mtDNA R parallels Y-DNA MNOPS. mtDNA U in Europe (part of R) parallels Y-DNA R (part of P). In Oceania, Papuans are rich in mtDNA P (part of R) and in Y-DNA MNOPS (part of CF), while Polynesians are rich in mtDNA B (also part of R) and Y-DNA C2, as well as various MNOPS lineages (part of CF). In East Asia, mtDNA B and F (again part of R) again correlate with Y-DNA NO (again part of MNOPS).
A comparison between Y-DNA and mtDNA phylogenies is also necessary to be able to cross-check one source of information about human prehistory against another. And sometimes a Y-DNA phylogeny may hold a clue to place in the mtDNA phylogeny that wasn't resolved correctly. Some discrepancies between Y-DNA and mtDNA phylogenies that you listed above may stem precisely from that. For instance, the mtDNA CF clade means an MN clade in mtDNA terms. But mtDNA scholars presented the evidence in such a way as if M and N are independently derived from L3. But then how come mtDNA M and N lineages are attested together in such distant parts of the globe as Australia and South America? Using clues from Y-DNA, a better mtDNA phylogeny should group MN together, with L2'3'4'6 coming out it in the same way as African Y-DNA E comes out of the CDEF clade.
"Try telling Maju that, and then stand back.
I haven't been able to see any mtDNA Ls that might have moved back to Africa with Y-hap DE yet, but I've just started looking."
Terry, instead of headbutting with Luis (I'm getting tired of reading your mutually destructive 500-comment long discussions on Leherensuge), check out this guy, especially his constructive thoughts on phylogenetic tree reversals in light of the Denisova cave discovery.
http://rokus01.wordpress.com/2010/03/27/denisova-cave-and-the-mystery-of-the-mtdna-phylogenetic-tree/#comment-271
"The presence or absence of an identifiable mtDNA correlate of a Y-DNA haplogroup is utterly irrelevant to resolving the question of the Y-DNA haplogroup's origin".
I agree completely. In fact I'm sure that the continued desire to connect particular Y-haps with particular mtDNAs has led many of us to the wrong conclusions. And it's generally been noted that Y-haps seem to be younger than mtDNAs. The mtDNA is less mobile than the Y-haps.
"does not have any bearing on the question of the region of origin of DE-YAP".
I'm not sure if you are aware of it but I have been having an argument with Maju at his blog as the that region. To me it seems fairly obvious that Y-hap E is an immigrant to Africa, probably entering that continent as a mixture of DE* and E*. The haplogroup moved east as a mixture of DE* and D*.
By the way:
"What is the mtDNA correlate of Y-DNA P in the Americas"
It's obvious that the first Y-hap into any region must be accompanied by an mtDNA of course. But after that the correlation can become very ill-defined.
"But mtDNA scholars presented the evidence in such a way as if M and N are independently derived from L3".
I've certainly been under the impression that they were. Do you have evidence for their common derivation?
"1. Outside of Africa all mtDNA haplogroups cluster into two macro-ones: M and N. All Y-DNA haplogroups do the same: DE vs. CF".
Yes, but neither of the mtDNA lines correlate with just one of the Y-hap lines DE or CF, and vice versa. Although further down the line Y-hap C seems to cluster with mtDNA N around the periphery of Y-hap and mtDNA M's cluster. This is conventionally explained as being the product of C and N's rapid migration through India and then beyond that of F and M. But to me this explanation seems very unlikely to be the case. I largely agree with your comments in numbers 2, 2, 3, 4 and 5 (you have 2 twice, but I understand what you're getting at).
"But then how come mtDNA M and N lineages are attested together in such distant parts of the globe as Australia and South America?"
The obvious explanation is that those regions were settled relatively recently as compared to much of Eurasia. So the haplogropups had become mixed. This explanation applies especially for your numbers 4 and 5. As I said, 'It's obvious that the first Y-hap into any region must be accompanied by an mtDNA of course. But after that the correlation can become very ill-defined'.
From German's link:
"Is it possible that here we have evidence that points to a third species, next to Neanderthal and AMH? A species, that might have been as civilized as a AMH, or a beast our ancestors didn’t breed with, or anything else that didn’t involve 'us' so we can understand?"
Interestingly the human species has much less haplogroup diversity that do most other species. But I have no difficulty in accepting there were multiple mtDNA lines around until long after 'modern' haplogroups emerged (and I'm sure the evidence shows they emerged from Africa). However it by no means follows that the human species is descended only from bearers of those 'modern' haplogroups. Although the authors seem to accept, without question, that modern humans are a product of a single relatively recent expansion I'm also sure that bearers of the 'modern' haplogroups bred with people bearing those 'ancient' haplogroups. The interesting thing is how those modern haplogroups came to replace those ancient ones. I'm sure the difference was not genetic, rather it was cultural.
And it seems unlikely that the ancestors of those 'modern' haplogroups had remained in Africa over the two million years or so from when members of the original Homo erectus/habilis/Australopithecus had left. The haplogroups' ancestors could well have recently moved back in to Africa from Eurasia, only to emerge yet again.
""What is the mtDNA correlate of Y-DNA P in the Americas"
It's obvious that the first Y-hap into any region must be accompanied by an mtDNA of course. But after that the correlation can become very ill-defined."
I don't know (yet) what the mtDNA correlate of Y-DNA P is in the Americas but I do correlate mtDNA C and D (high frequencies in the Americas, especially in South America) with Y-DNA Q. Notice the interesting parallelism in two very unlikely regions: mtDNA D in Saami and Y-DNA Q in Norwegians (Saami substrate?) as well as mtDNA C and D in India next to Y-DNA Q5.
"In fact I'm sure that the continued desire to connect particular Y-haps with particular mtDNAs has led many of us to the wrong conclusions. And it's generally been noted that Y-haps seem to be younger than mtDNAs. The mtDNA is less mobile than the Y-haps."
Some of these problems may stem not from the inherent differences in the evolution of Y-DNA vs. mtDNA lineages but in the way people define haplogroups, construct phylogenetic trees and deal with homoplasies. Regarding the higher mobility of Y-DNA haplogroups, some cases directly contradict this observation: mtDNA M1, N1, U6, possibly X1/X2 penetrated northern and eastern Africa. There are no apparent correlates in Y-DNA for these small inroads. The correlation emerges, however, if we look at the whole mtDNA L2'3'4'6 lineage against the whole Y-DNA DE lineage. And what we notice is that L2'3'4'6 must be in sister relationship with only a subset of Eurasian haplogroups. And you wrote a couple of days ago: "I haven't been able to see any mtDNA Ls that might have moved back to Africa with Y-hap DE yet, but I've just started looking."
""1. Outside of Africa all mtDNA haplogroups cluster into two macro-ones: M and N. All Y-DNA haplogroups do the same: DE vs. CF".
Yes, but neither of the mtDNA lines correlate with just one of the Y-hap lines DE or CF, and vice versa."
Apologies, but I have to revise my earlier statement. I now think that Y-DNA DE does not have a correlate in current mtDNA phylogenies. mtDNA M correlates with Y-DNA F and mtDNA N correlates with Y-DNA C. Y-DNA DE is mtDNA L2'3'4'6 in Africa (E* and descendants) plus several mtDNA lineages (think of Y-DNA D* and descendants) hidden under the umbrella of M macrohaplogroup, e.g., M31, M32 in Andaman islands, M7, M8 in Japan, M9 in Japan, Tibet, SE Asia, etc.
"The obvious explanation is that those regions were settled relatively recently as compared to much of Eurasia. So the haplogropups had become mixed."
We discussed it earlier, Terry. I don't think evidence bears out the admixture hypothesis. mtDNA outside of Africa is very "flat," so we only see basal offshoots (e.g., hg A straight off of mhg N, M29'Q and M42 straight off of mhg M). You could say that hgs C and B came to America later (because they are seen as derived, respectively, from M7 and R*) but they are so deeply entrenched in South America and hg B is so rare in Siberia, so again this scenario has problems.
And since there's a parallelism between Y-DNA and mtDNA here as well (America has members of mtDNA M* and N* lineages and Y-DNA C* and F* (Q) lineages, Australia, too, has members of mtDNA M* and N* lineages and Y-DNA C* and F* (R*, unless from European admixture), the admixture hypothesis becomes less and less likely.
What I think is truly interesting is that Sub-Saharan Africa and Europe are the only two macroregions/continents in which we do NOT have this co-presence of two major East Eurasian clades. Europe only has mtDNA R* (off of N*) and Y-DNA F*, Sub-Saharan Africa has only one basal mtDNA M* lineage, namely M1 (plus mtDNA X1/X2, U6, Y-DNA R1b1a all of which look more derived) and then the huge Y-DNA DE clade, with absolutely no mtDNA C* lineages.
"Do you have evidence for their common derivation?"
No, I don't. But we don't have evidence for their common derivation in Africa either. At least outside of Africa both of them are attested in the same populations.
"Interestingly the human species has much less haplogroup diversity that do most other species."
I thought that Neandertals - whether separate species or not - are less diverse than humans. At least the following figure suggests so: http://johnhawks.net/graphics/green-2010-figure-3.jpg. Just like wolves are less diverse than dogs. Both cases seem to suggest that diversity may be inversely related to the age of a species/population.
"mtDNA M1, N1, U6, possibly X1/X2 penetrated northern and eastern Africa. There are no apparent correlates in Y-DNA for these small inroads".
There is, if you consider the possibility that Y-hap E originated outside Africa. Again the Y-hap would have traveled much further than the mtDNAs. But:
"I now think that Y-DNA DE does not have a correlate in current mtDNA phylogenies. mtDNA M correlates with Y-DNA F and mtDNA N correlates with Y-DNA C".
I agree with that.
"I thought that Neandertals - whether separate species or not - are less diverse than humans".
Presumably they were very inbred as a result of their long isolation. That's probably the main reason they died out, or were so completely absorbed.
"Just like wolves are less diverse than dogs".
I'm not sure that is true. Although I suppose it could be if you compare just one subspecies of wolf with the whole dog species. You may then find that to be the case if dogs descend from several species of wolf.
""I now think that Y-DNA DE does not have a correlate in current mtDNA phylogenies. mtDNA M correlates with Y-DNA F and mtDNA N correlates with Y-DNA C".
I agree with that."
Well, this attests to a sharp discrepancy between mtDNA and Y-DNA phylogenies. In mtDNA phylogenies non-African lineages are nested within one single African clade L3, in Y-DNA phylogenies the majority of African lineages are nested within a largely non-African clade CDEF. (note, however, that mtDNA L3 is closer to M and N than it is to L0 and L1). mtDNA is more hierarchical in Africa, Y-DNA is more hierarchical outside of Africa. If we further look at such diploid systems as X chromosome, we'll discover that the earliest branch will be exclusively non-African (with highest frequencies in America, as in the case of the B006 lineage in dystrophin gene http://books.google.com/books?id=X88O8C3ZHvMC&pg=PA76&lpg=PA76&dq=X+chromosome+B006&source=bl&ots=mjlR3EUv2d&sig=PKQLXov-E5YfIT1JqXoBrXXwMQg&hl=en&ei=TQOiTO6rLMGC8gboo925CQ&sa=X&oi=book_result&ct=result&resnum=6&ved=0CDgQ6AEwBQ#v=onepage&q=X%20chromosome%20B006&f=false).
This means, as we progress from mtDNA through Y-DNA to X chromosome African genes become a subset of non-African genes. Once we understand that mtDNA phylogenies are possibly inaccurate we can use insights from other phylogenies to locate the area of the problem. And I think we should look closely at M31, M32 in Andaman islands, M7, M8 in Japan, M9 in Japan, Tibet, SE Asia in search of a direct link to African L2'3'4'6.
"I'm not sure that is true. Although I suppose it could be if you compare just one subspecies of wolf with the whole dog species. You may then find that to be the case if dogs descend from several species of wolf."
Both of us seem to be right: "Interestingly, different breeds share many more linked loci than wolves do, supporting the notion that dogs went through a genetic bottleneck during domestication. Even so, village dogs harbor more genetic diversity than the gray wolf, perhaps because they managed to keep their populations large enough to avoid inbreeding, unlike the relentlessly persecuted wolf." (http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1000452)
"Well, this attests to a sharp discrepancy between mtDNA and Y-DNA phylogenies".
Why is that a problem?
"mtDNA is more hierarchical in Africa, Y-DNA is more hierarchical outside of Africa".
In general mtDNAs are older than Y-haps. Y-haps are more mobile, and seem to replace each other on occasions.
"If we further look at such diploid systems as X chromosome, we'll discover that the earliest branch will be exclusively non-African (with highest frequencies in America"
That's no problem as far as I'm concerned. To me the haplogroups just represent the equivalent of two more genes that have spread through the human population. And I keep telling Maju that diversity does not equal origin. Other factors can give rise to diversity. Lack of selection, for one.
"And I think we should look closely at M31, M32 in Andaman islands"
To me the evidence is overwhelming that the Andamans were settled relatively recently, as compared to Australia and New Guinea for example.
"village dogs harbor more genetic diversity than the gray wolf, perhaps because they managed to keep their populations large enough to avoid inbreeding, unlike the relentlessly persecuted wolf".
That's what I suspected.
""Well, this attests to a sharp discrepancy between mtDNA and Y-DNA phylogenies".
Why is that a problem?"
It's a problem if we want to use either as a source of information about the prehistory of populations, not genes. If mtDNA and Y-DNA are inherently incompatible, then how can be deduce population history from them? If, on the contrary, they are well correlated, then we're in a good spot to use them to infer population history.
"In general mtDNAs are older than Y-haps. Y-haps are more mobile, and seem to replace each other on occasions."
mtDNA mutates faster than Y-DNA.
"diversity does not equal origin. Other factors can give rise to diversity. Lack of selection, for one."
I agree. It's the same as with the lack of archaeological finds (read: diversity = 0) in a region: it doesnt' necessarily mean the absence of humans. The relaxation of a selective constraint is likely the cause behind extensive genetic diversity in the Old World as compared to the New World. As populations expand and grow, selection weakens.
"To me the evidence is overwhelming that the Andamans were settled relatively recently, as compared to Australia and New Guinea for example."
You're making the old mistake of interpreting the absence of evidence as evidence of absence. The fact that the earliest archaeological culture in the Andaman islands is some 6,000 years old doesn't mean this is the horizon of humans settling the islands. (Compare, the fact that up until 2005 we didn't have a single chimp fossil doesn't mean chimps are recent and derived from, say, Pygmies.) It just as likely we haven't found earlier signatures of human presence in the Andamans. From the genetics perspective, the first occupation of the Andamans is likely much older than 6,000 years BP because they harbor Y-DNA D lineages, which belong to an early stratum of genetic diversity in Asia (pre-MNOPS).
"It just as likely we haven't found earlier signatures of human presence in the Andamans".
True, as is the case in many (if not most) parts of the world.
"From the genetics perspective, the first occupation of the Andamans is likely much older than 6,000 years BP because they harbor Y-DNA D lineages, which belong to an early stratum of genetic diversity in Asia (pre-MNOPS)".
The Toalean (which is the first archeological presence on the Andamans) is more than 10,000 years old, perhaps 15k. So that clears up one problem. But to assume that the Y-DNA lineage D is ancient in the Andamans is not justified. Even if it originated somewhere within SE Asia there's no reason to assume it reached the Andamans immediately after it formed. And seeing that the Toalean arrived in the Andamans at some time it's reasonable to suppose it left some signature of its arrival. D is the only Y-hap there, or certainly the main one.
"The relaxation of a selective constraint is likely the cause behind extensive genetic diversity in the Old World as compared to the New World".
Expansion and utilisation of so far unexploited resources is probably the main reason for rapid expansion, and lessening of drift. But why would this factor have been so important in Eurasia but not in America with the first human arrival? Under the standard hypothesis the small number of original haplogroups in America diversified considerably soon after they arrived. So it fits a rapid expansion soon after their arrival. Similarly in other parts of the world that earlier forms of human had been unable to reach.
"If mtDNA and Y-DNA are inherently incompatible, then how can be deduce population history from them?"
I'm not sure that we can. The spread of one or other haplogroup is usually associated with the spread of some culture, but we can't automatically assume memebrs of the opposite sex accompanied that expansion. Certainly not over the whole of the expansion. To me it is far more likely that haplogroups have always moved through pre-existing populations, as seems to be assumed for post-Neolithic expansions. However there is a great reluctance to see Paleolithic haplogroup expansions as being a product of the same process.
"The Toalean (which is the first archeological presence on the Andamans) is more than 10,000 years old, perhaps 15k."
I think the earliest archaeological finds in the Andamans date back to only 6000-2000 years BP. Cooper, Z. (1993). The origins of the Andaman Islanders: local myth and archaeological evidence.
Antiquity 67, 394–399. The tools are reminiscent of the Toalean culture in SEAsia but the associated dates are much younger.
"But to assume that the Y-DNA lineage D is ancient in the Andamans is not justified. Even if it originated somewhere within SE Asia there's no reason to assume it reached the Andamans immediately after it formed..."
I like the idea of Andamanese arriving to the islands from SE Asia. This is my reverse coastal route idea. The rest of your argument poses some problems. If hg D arrived in the Andamans relatively recently meaning at a time when the MNOPS clade was well diversified in SE Asia, then the question is: why didn't they only selected basal hg D but not any of the younger lineages? Correspondingly, why don't the Andamanese show any signs of such a wide-spread SE Asian lineage as mtDNA hg B, but only an M lineage derived from the root of M*?
"Expansion and utilisation of so far unexploited resources is probably the main reason for rapid expansion, and lessening of drift. But why would this factor have been so important in Eurasia but not in America with the first human arrival? ..."
From an out-of-America perspective, a small group of Eurasian hominins such as Neandertals went through a bottleneck and migrated to the New World. Although we need more ancient DNA to ascertain that we are related to Neandertals via the American Indian outlier, some incipient correspondences are highly suggestive of this. E,g., the two Neandertals sampled for blood groups showed blood type O, which is the most frequent blood group in the Americas. Shovel-shaped incisors are most frequent in Neandertals and Amerindians. In the New World, the original founders diversified within the inherited set of genes and phenotypical traits (comp. the high diversity of blood group O subtypes in America) but maintained their basic inherited blueprint under selection. These selective constraints were lifted as they expanded out of the New World to recolonize the Old World - now as Homo sapiens sapiens - and to replace the Old World hominins, including the so-called "anatomically modern humans" in Africa. This Old World expansion resulted in the progressive decay of the original "blueprint," as evidenced by the decreasing frequencies of blood group O, shovel-shaped incisors, etc. in different parts of the world.
Also note the following passage made recently by John Hawks: "One might expect Neandertal-like morphology to show up at some low level. Of course morphological features are polygenic, so that phenotypic resemblance falls much faster than genic identity. And Holocene populations have continued to evolve. Maybe early Asian skeletal remains like the Upper Cave skulls (ca. 11,000-20,000 years old) actually reflect that Neandertal heritage to a greater extent than recent samples." And we know that the Upper Cave skulls show a rather close relationship with Paleoindian skulls.
In essence, the limited genetic diversity of Amerindians is the continuation of the limited genetic diversity of Neandertals, but just like Neandertals are older than AMH in Africa, Amerindiands are older than Homo sapiens sapiens in Africa.
See also my latest comment on Rokus blog http://rokus01.wordpress.com/2010/03/27/denisova-cave-and-the-mystery-of-the-mtdna-phylogenetic-tree/#comment-275 for other evidence of high frequency of "archaic introgressions" in Amerindians.
"To me it is far more likely that haplogroups have always moved through pre-existing populations..."
I think that putting together all data sources from culture to biology will allow us to get at the bottom of the population process involving real people in prehistory, while the use of a single source (including a single genetic system) will always give us a partial perspective. By identifying correlations between mtDNA and Y-DNA trees we are doing an important part of this integration work. Otherwise, it's all agnostic.
"why didn't they only selected basal hg D but not any of the younger lineages?"
There could be any number of reasons for that. Why did not all members of MNOPS arrive in Europe? D was possibly established on the nearby mainland and adopted the technology necessary to reach the Andamans.
"why don't the Andamanese show any signs of such a wide-spread SE Asian lineage as mtDNA hg B, but only an M lineage derived from the root of M*?"
That M lineage was probably established on the mainland somewhere nearby, perhaps Burma. The same M haplogroup's presence in India is quite possibly an immigrant there from futher east.
"a small group of Eurasian hominins such as Neandertals went through a bottleneck and migrated to the New World".
In theory, then, being the first such species there they should have multiplied and expanded through the whole of America. That's usully what species do when they first arrive in a new region. Either that or become extinct fairly rapidly.
"the two Neandertals sampled for blood groups showed blood type O, which is the most frequent blood group in the Americas. Shovel-shaped incisors are most frequent in Neandertals and Amerindians".
I have no doubt that Neanderthals, and other humans, have made more of a contribution to our genes than is commonly accepted. In the case of indigenous Americans it's likely to be 'archaic' Central Asian populations. The emphasis has usually been on how different we are to archaic humans, rather than considering how similar we are. Most love to imagine that we are far superior to all other species, including our closest relations.
"I think that putting together all data sources from culture to biology will allow us to get at the bottom of the population process involving real people in prehistory, while the use of a single source (including a single genetic system) will always give us a partial perspective".
Quite. But to assume that Y-haps always correlate with mtDNA is likely to hamper our view of the other connections. Y-hap and mtDNA do not always expand together.
"In theory, then, being the first such species there they should have multiplied and expanded through the whole of America. That's usully what species do when they first arrive in a new region. Either that or become extinct fairly rapidly."
Dah, they did. That's why we have 140 language stocks there. Most of the times, it takes time and isolation for languages to emerge.
"D was possibly established on the nearby mainland and adopted the technology necessary to reach the Andamans."
We need support for that, namely surviving mainland D that was the source of Andaman D.
"That M lineage was probably established on the mainland somewhere nearby, perhaps Burma. The same M haplogroup's presence in India is quite possibly an immigrant there from futher east."
Well, as far as I remember, members of the M31'32 clade are found in India.
"In the case of indigenous Americans it's likely to be 'archaic' Central Asian populations."
This is all very very speculative, Terry.
"Well, as far as I remember, members of the M31'32 clade are found in India".
But there's plenty of M in SE Asia so M31'32 could easily be a back migration.
"Dah, they did. That's why we have 140 language stocks there".
It's unlikely that they 'multiplied and expanded through the whole of America' because there is so little evidence for their presence. And the evidence that is present is controversial.
"It's unlikely that they 'multiplied and expanded through the whole of America' because there is so little evidence for their presence. And the evidence that is present is controversial."
Wrong logic, Terry. Law, at least in "civilized countries" has a nice principle called the "presumption of innocence." This principle is just as valid in the field of human origins. The fact that there're indigenous people in America and there are 140 language stocks among them is sufficient evidence for great antiquity of human presence in the New World. Unless convincingly proven otherwise. So far, despite the 150 years of deliberate searching for the recent origin of American Indians in Asia, there's no clear archaeological or paleobiological signatures illustrating their migration. But we have hundreds of years ahead of us to find traces of their presence in the New World prior to 15,000 years. Archaeological evidence of early human presence is growing on every continent, it's just the rate of find accumulation is different - dramatically different between Europe/Africa and the Sahul and America.
Terry, you are still there? I've been reading Melanesian mtDNA Complexity, by Jonathan S. Friedlaender, and realized that mtDNA B4 likely corresponds to Y-DNA C2, while mtDNA E to Y-DNA O. Surprised? Friedlaender criticizes the origin of the Polynesian motif from Taiwan both in terms of dates and in terms of phylogeny (only one np difference between the two, and what is it's a hot spot?). B4 is much older than the archaeological timeline for the spread of agriculture from Taiwan. I suspect that it's mtDNA E and Y-DNA O that represent the recent migration from Asia, while B4 and C2 are traces of the original Pleistocene migration. It's just for some reason B is not diverse enough, hence molecular clock places it in early Holocene, which is suspiciously young, although still much earlier than Neolithic. Just like highland Papua New Guinea lacks B, it lacks C. Just like C2, B4 is at high frequencies in Polynesia, which represents a late expansion from Near Oceania. So, it looks like mtDNA and Y-DNA are in full agreement with each other in Oceania.
"and realized that mtDNA B4 likely corresponds to Y-DNA C2"
That has been obvious to me for at least 10 years. That combination is the main East Polynesian grouping.
"mtDNA E to Y-DNA O"
Not so sure of that one. To me E is older in SE Asia than is O. E most likely arrived with some less derived F Y-hap.
"B4 is much older than the archaeological timeline for the spread of agriculture from Taiwan".
I've been claiming that at Maju's new blog. It probably arose in the Philippines, or Taiwan anyway.
"It's just for some reason B is not diverse enough, hence molecular clock places it in early Holocene, which is suspiciously young"
If you look at the Phylotree you see that B is actually almost as old as the diversification of R from N.
"Just like highland Papua New Guinea lacks B, it lacks C".
Which indicates that the two haplogroups bypassed it on their way out to Polynesia.
"So, it looks like mtDNA and Y-DNA are in full agreement with each other in Oceania".
I remember someone complaining that you can't judge the whole history of humanity by what happened in Australia. In fact my interest in haplogroup distribution began when I started unravelling the origin of the New Zealand Maori. The layers just peel back like an onion.
"That has been obvious to me for at least 10 years. That combination is the main East Polynesian grouping."
In the literature, however, it'd commonplace to find an assertion that incoming Austronesians carried with them to Polynesian an Asian mtDNA pool but not an ancient Y-DNA pool. The Y-DNA pool got picked up in Melanesia from pre-existing populations. This is because B4 is found in Taiwan, while C2 or K* is/are not. Then the argument is made that Austronesian and proto-Oceanic societies were matrilineal and matrilocal, hence they absorbed indigenous males en masse. I just feel that B4 and C2 are Austronesian lineages through and through.
""Just like highland Papua New Guinea lacks B, it lacks C".
Which indicates that the two haplogroups bypassed it on their way out to Polynesia."
They bypassed it but not because they came to the area thousands of years later. Does the data bear out a scenario by which Austronesians have been in Wallacea and Near Oceania since the time of the first colonization of the Sahul? They just stayed behind, while future Australians and Papuans went further down. And then in Holocene Austronesians just expanded back north and out east and received an Asian input from up north in the form of Y-DNA O and mtDNA E.
"I just feel that B4 and C2 are Austronesian lineages through and through".
I'd agree regarding B4. However I'm sure C2 was picked up from southern Wallacea after the initial Austronesian expansion. So the asserion:
"assertion that incoming Austronesians carried with them to Polynesian an Asian mtDNA pool but not an ancient Y-DNA pool".
Is partly correct. The B4 is Asian, or northern Wallacean. But to say that no Asian Y-DNA was involved is incorrect. Y-DNA O3a is Asian and made it well out into the Pacific, at least as far as Western Polynesia.
"Does the data bear out a scenario by which Austronesians have been in Wallacea and Near Oceania since the time of the first colonization of the Sahul?"
No. Some new mixture happened there.
"They just stayed behind, while future Australians and Papuans went further down".
Australains and Papuans have a different set of haplogroups, and different from each other what's more. So three different movements across Wallace's line, I'd guess. Some Papuan haplogroups did make it to Australia though (Y-hap K and mtDNA P).
"I'd agree regarding B4. However I'm sure C2 was picked up from southern Wallacea after the initial Austronesian expansion. So the asserion:
"assertion that incoming Austronesians carried with them to Polynesian an Asian mtDNA pool but not an ancient Y-DNA pool".
Is partly correct. The B4 is Asian, or northern Wallacean. But to say that no Asian Y-DNA was involved is incorrect. Y-DNA O3a is Asian and made it well out into the Pacific, at least as far as Western Polynesia."
If C2 was picked in Wallacea from extinct Papuans, we would expect to find it among surviving Papuans. I may be asking for too much, though.
O3a did make it out to Near Oceania, but much later. I don't deny that there was Holocene gene flow from East Asia into Near Oceania, associated with some Asian cultural inventory, but I'm checking the possibility that B4 and C2 represent offshoots of the earliest, Late Pleistocene, migration from East Asia to Oceania roughly contemporaneous with the colonization of Papua New Guinea and Australia. Under this scenario, some Melanesians and Polynesians are "purer" than Southeast Asian Austronesian speakers. They just expanded into Remote Oceania recently, but otherwise avoided Holocene East Asian influence due to their location in geographic refugia east of the Wallace line. I'm just not entirely comfortable with attributing C2 to admixture with Papuans if Papuans don't have this lineage in the first place. Why don't we then say that B4 was absorbed by Austronesians from Papuans, too.
"If C2 was picked in Wallacea from extinct Papuans, we would expect to find it among surviving Papuans".
C2 is not Papuan. The C2 in New Guinea almost certainly came in with Austronesian-speaking people, especially as it's concentrated in the Bird's Head region. If anything C2 is closer to Australian haplogroups.
"I'm checking the possibility that B4 and C2 represent offshoots of the earliest, Late Pleistocene, migration from East Asia to Oceania roughly contemporaneous with the colonization of Papua New Guinea and Australia".
I don't think so. The distribution suggests those haplogroups were confined to Wallacea, especially the south (C2) or northern Wallacea, especially the Philippines (B4) until the Austronesian expansion. If they were contemporaneous with the arrival in New Guinea or Australia there's no reason why they wouldn't have reached either country, or both.
"otherwise avoided Holocene East Asian influence due to their location in geographic refugia east of the Wallace line".
It would be difficult to make that case. Their expansion east of Wallace's Line has all the characteristics of Neolithic expansion.
"I'm just not entirely comfortable with attributing C2 to admixture with Papuans if Papuans don't have this lineage in the first place".
I don't think we could call the Southern Wallaceans 'Papuans'.
"Why don't we then say that B4 was absorbed by Austronesians from Papuans, too".
And B4 is definitely not 'Papuan'. It's probable there were four different groups in the region: Papuans, Australians, south Wallaceans and north Wallaceans.
By "Papuan" I mean all non-Austronesian and all non-Australian indigenous, pre-Neolithic populations. It's a blanket term commonly used in the literature. So, to re-phrase: I'm not entirely comfortable with attributing C2 to admixture with a Papuan or Wallacean substratum if extant Papuans don't have this lineage in the first place and all Wallaceans in the Philippines speak Austronesian languages. We just have a double gap in evidence here: no evidence for non-Austronesian languages in Wallacea and no evidence for a source of C2 among Papuans.
"If they were contemporaneous with the arrival in New Guinea or Australia there's no reason why they wouldn't have reached either country, or both."
Why not? Some people move forward, others stay behind. Otherwise, we would have all of humanity living in New Zealand. C2 is not found in Australia, C4 is not found in Melanesia or Polynesia, which means they are two distinct arms of a single ancestral population that existed in Southeast Asia 40-50K years ago. It's just C2 expanded out east into Remote Oceania relatively recently. Otherwise, it's like saying that America was colonized from 1492 as a result of the Neolithic expansion of Indo-Europeans.
"Their expansion east of Wallace's Line has all the characteristics of Neolithic expansion."
See, when it comes to kinship systems and languages, Austronesian Oceanic typological features and kinship systems cluster with Papuans and Australians, not with Sino-Tibetan, Tai-Kadai or Austroasiatic. So they are distinctly non-Asian. Same for, say, dentition when it comes to biological traits. The aborigines of Taiwan are Sinodonts just like Chinese or Koreans, Melanesians, Polynesians, Soutehast Asians, Negritos are Sundadonts. It's of course possible that they are all "areal" features meaning that Austronesian speakers absorbed them from neighboring Papuans along with Papuan Y chromosomes but it's not the most parsimonious solution.
"By 'Papuan' I mean all non-Austronesian and all non-Australian indigenous, pre-Neolithic populations. It's a blanket term commonly used in the literature".
True. Certainly it seems as though some sort of 'Papuan' phenotype was spread through SE Asia, Australia and New Guinea before the spread of the 'Mongoloid' phenotype.
"I'm not entirely comfortable with attributing C2 to admixture with a Papuan or Wallacean substratum if extant Papuans don't have this lineage in the first place"
Certainly not in New Guinea, but the haplogroup is common in Wallacea, especially in the south. It seems to have spread from there, almost certainly with the expansion of the Austronesian-speaking people. Eventually moving out into the remote Pacific.
"no evidence for a source of C2 among Papuans".
It seems almost certain that C2 originated in Tengarra, or nearby.
"all Wallaceans in the Philippines speak Austronesian languages".
As do most through Wallacea and SE Asia. But almost certainly introduced from further north. In other words the pre-existing population has adopted the language, although I'm sure some of the islands were (or had become) unoccupied by Austronesian times.
"no evidence for non-Austronesian languages in Wallacea"
True. But the fact that we can recognise an Austronesian language group suggests its expansion is recent.
"Austronesian Oceanic typological features and kinship systems cluster with Papuans and Australians, not with Sino-Tibetan, Tai-Kadai or Austroasiatic".
I'd guess that those kinship systems have spread from the Chinese Neolithic. Although I strongly suspect the Austronesian languages are also a product of the Chinese Neolithic the Austronesians may have either adopted Papuan systems or held onto an earlier system from SE Asia.
"The aborigines of Taiwan are Sinodonts just like Chinese or Koreans"
Almost certainly there was no 'single migration' of the Chinese Neolithic, just as I'm sure there was no 'single migration' of modern humans out of Africa. The further north you go in Wallacea the more obvious are Mongoloid characteristics. People in Timor look much more 'Papuan' than do people further north. It is extremely likely that a 'Mongoloid' phenotype has spread south through a 'Papuan' phenotype.
"But the fact that we can recognise an Austronesian language group suggests its expansion is recent."
Yes and no. Your logic is valid and it's what most linguists would agree with. One thing to keep in mind is that certain language families exhibit surprising structural or lexical conservatism. Austronesian, especially the Malayo-Polynesian branch is one of these families. Athabascan in North America is another. Hence, Na-Dene, for instance, is sometimes dated at 12,000 years, although, judging by its internal diversity, it looks young.
"Hence, Na-Dene, for instance, is sometimes dated at 12,000 years, although, judging by its internal diversity, it looks young".
Almost certainly looks young because it is much younger than 12,000 years.
"One thing to keep in mind is that certain language families exhibit surprising structural or lexical conservatism. Austronesian, especially the Malayo-Polynesian branch is one of these families".
Unlikely to be so. The original diversity of the eastern Pacific Austronesian languages beyond the Admiralties cannot be more than 5-6000 years. And they're extremely diverse.
I have a series of essays explaining the situation in the Pacific up at 'remotecentral'. I've used the situation in Polynesia as part of the evidence against creationists and evolution-deniers, so it has that perspective. But it will help anyone to understand the settlement of the Pacific. The first one is here, and links to the next in the series at the end:
http://humanevolutionontrial.blogspot.com/2009/06/human-evolution-on-trial-eastern.html
The one most relevant to the present discussion is actually this one though:
http://humanevolutionontrial.blogspot.com/2009/06/human-evolution-on-trial-pacific.html
Obviously I could now improve the essays because of the more extensive evidence available, but really nothing much has changed.
In fact more recently Maju has discovered a big expansion of mtDNA B, almost simultaneously with mtDNA R's expansion, and presumably part of that expansion. I've at last convinced him that mtDNA R originated in SE Asia, not South Asia. So we can assume that B too has a long history in the region.
B's earliest expansion stretched through much of East and SE Asia from the north and east of the western shore of the Wallacean lake. It did not cross Wallace's line till much later. And then just a small sub-section of that B was involved in the Austronesian-speaking people's expansion out into the Pacific. B4a1a1a makes up the vast bulk of B haplogroups beyond the Admiralty Islands, and so had presumably formed by the time the Austronesians had reached there. It seems very likely that B4a1a originated in Taiwan:
http://mbe.oxfordjournals.org/content/27/1/21.abstract
Some time ago Maju, Ebizur and I discussed the distribution of Y-DNA C through Australia, New Guinea and SE Asia. From the comprehensive data Ebizur supplied it seems that C2 developed in Tengarra and C4 developed in Australia. But C didn't make it to New Guinea at all until the Austronesian expansion introduced it to the Bird's Head. Various C2 haplogroups then wandered off into the Highlands, and others spread onwards through the islands north and east of New Guinea, and out into the Pacific to new Zealand.
So it seems that Y-DNA O1's role in all this was merely to carry mtDNA B4a from Taiwan into Wallacea (perhaps 7000 years ago) where they ran into O1's relation O3, who had also recently arrived in the SE Asian islands from the north. The language of the O haplogroup (probably O1's) came to dominate, but much else of B4a's southern culture and genes survived. The Os then introduced mtDNA B4a to their new friend from the south, C2, and off they all went: out into the Pacific.
On the way out to New Zealand O1 dropped off first, then O3. Only C2a1 made it all the way. So the Polynesians are a complex mixture of people: Y-DNA O from the Yangtze, mtDNA B4a from South China and Y-DNA C2 from Southern Wallacea.
"Hence, Na-Dene, for instance, is sometimes dated at 12,000 years, although, judging by its internal diversity, it looks young".
Almost certainly looks young because it is much younger than 12,000 years.
"One thing to keep in mind is that certain language families exhibit surprising structural or lexical conservatism. Austronesian, especially the Malayo-Polynesian branch is one of these families".
Unlikely to be so. The original diversity of the eastern Pacific Austronesian languages beyond the Admiralties cannot be more than 5-6000 years. And they're extremely diverse."
Terry, you're amusing with your categorical assertions regarding ages of language families. For Na-Dene, see Kari, "The Concept of Geolinguistic Conservatism in Na-Dene Prehistory" in most recent volume "The Dene-Yeniseian Connection." For Malayo-Polynesian see Donohue and Denham, "Farming and language in island Southeast Asia." (In fact, I communicated it to you some time ago - did you read it? it's available on the web). On p. 227 it says "From the presence and location of the proposed Central-Eastern Malayo-Polynesian subgroup,
which is seen as the eastern split of Malayo-Polynesian, as well
as the presence of more proposed subgroups in the south than
the range of possible dispersal centers for Malayo-Polynesian, based solely on linguistic evidence, would extend as far east as the Pacific. Consequently, it is clear that there is no linguistic evidence for an
orderly north-to-south dispersal of Malayo-Polynesian languages from Taiwan. The linguistic evidence for Malayo-Polynesian presents us with additional methodological challenges. The lexical conservatism of the family is remarkable (Blust 2000b)." The whole paper is interesting.
"It seems very likely that B4a1a originated in Taiwan:
http://mbe.oxfordjournals.org/content/27/1/21.abstract"
Again, Terry, I communicated to you a paper by Melanesian mtDNA Complexity, by Jonathan S. Friedlaender (available on the web for free) in which the Taiwan origin is questioned. Even in the Tabbada paper, a back migration from Indonesia to the Philippines is entertained to explain the distribution of B4a1a1. This would leave us with a gap between Taiwan and Indonesia.
An overlooked pattern, namely the phylogenetic and chronological correlation between Y-DNA C*/C2 and mtDNA B*B4 needs to be explored. I also see an old connection between mtDNA B*/B4 and P (Papua New Guinea, Melanesia, Australia) in the Pacific (parallel to C2/C4).
"In fact, I communicated it to you some time ago - did you read it?"
No. I couldn't download it at the time, but I've just managed to. So I'll read it when I have more time.
'a back migration from Indonesia to the Philippines is entertained to explain the distribution of B4a1a1".
Possible. In fact I'm sure the development of the Austronesian language and culture was a very complicated process.
"the presence of more proposed subgroups in the south than
the range of possible dispersal centers for Malayo-Polynesian, based solely on linguistic evidence, would extend as far east as the Pacific".
Certainly the development of the Austronesians is far more complex than a simple diffusion from Taiwan. But that doesn't alter the fact that the huge diversity of Austronesian languages to the east of the Northern Solomons can only have begun around 5000 years ago. So that branch of Malayo-Polynesian can hardly be said to be geolinguisticly conservative.
"But that doesn't alter the fact that the huge diversity of Austronesian languages to the east of the Northern Solomons can only have begun around 5000 years ago."
There's no doubt about it. But I do think that from the present evidence we cannot exclude the possibility that the area from Borneo to New Britain could be the ultimate source of Austronesian expansion at different times to different places: northwest to Andaman islands, southwest to Madagascar, northeast to Taiwan, and southeast to Polynesia. The northern Mongoloid component didn't bring the Austronesian language with it but was rather absorbed by the in-situ Austronesian populations that came to the region in Pleistocene together with Papuans and Australians.
The problem with the existing out-of-Taiwan linguistic phylogeny is that it's built on an assumption of progressive geographic separation of daughter languages without taking into account the process of in-situ piling up of parental languages. By the latter I mean the existence of an in-situ dialect chain with roots in the Pleistocene that experienced layers of innovations spreading across its range and creating nodes without any massive population movements. Those nodes would be obliterated with every successive innovation sweep. Instead of postulating Austronesians obliterating pre-existing Papuan populations but "sucking out" all their Y-DNA C2 lineages from them (a major logical problem for me), why can't we imagine these replacements to have been administered by the speakers of Malayo-Polynesian speakers against other Malayo-Polynesian speakers? In the latter scenario, we have continuity and replacement going on at the same time, thus validating both the uniqueness of C2 in Malayo-Polynesians, its absence in extant Papuans and the absence of any vestiges of non-Austronesian speakers in Wallacea.
I've managed to read just some of your link so far. There are no surprises in the paper.
"But I do think that from the present evidence we cannot exclude the possibility that the area from Borneo to New Britain could be the ultimate source of Austronesian expansion at different times to different places"
The authors are absolutely convinced that the Austronesian language originated in Taiwan. In other words it is not indigenous to island SE Asia or Wallacea.
"The northern Mongoloid component didn't bring the Austronesian language with it but was rather absorbed by the in-situ Austronesian populations that came to the region in Pleistocene together with Papuans and Australians".
I agree totally in a mixture. In fact in one of my essays I wrote:
"So the people who moved into the perhaps temporarily vacant Island Southeast Asian sub-point of the human star became a hybrid of two kinds of people. These were the Melanesian or Australian Aboriginal-looking people who, at the time, lived as far north as South China and the people from North China or Japan who brought technologies introduced there at the end of the ice age. A mix of people from the Australian and East Asian points of the human star (map 1). In other words there was not simply total replacement of one population by another. The present Island Southeast Asian population lies in a hybrid zone between two other different populations"
And earlier here I wrote:
"So it seems that Y-DNA O1's role in all this was merely to carry mtDNA B4a from Taiwan into Wallacea (perhaps 7000 years ago)"
So I accept that the influence of Taiwan was not overwhelming. Back to your comments:
"The problem with the existing out-of-Taiwan linguistic phylogeny is that it's built on an assumption of progressive geographic separation of daughter languages without taking into account the process of in-situ piling up of parental languages".
The only thing I would change in my essays ( a change I've known is necessary for some time) is to greatly reduce my proposed language backflow from the Philippines to Taiwan. The backflow actually originated much further south, and is presumably associated with Y-hap C2. In another of those essays I wrote:
"The development of the improved boating technology in Island Southeast Asia was probably the result of a complicated series of population movements in the region".
In agreement with the authors of the article.
Back to you:
"By the latter I mean the existence of an in-situ dialect chain with roots in the Pleistocene that experienced layers of innovations spreading across its range and creating nodes without any massive population movements".
The pre-Austronesian agricultural substrate in the region is the Hoabinhian. I agree that this is not a product of the Chinese Neolithic, and may even have been the source for it. But the authors don't mention the Hoabinhian at all. And the people almost certainly didn't speak and Austonesian or Austroasiatic language.
"why can't we imagine these replacements to have been administered by the speakers of Malayo-Polynesian speakers against other Malayo-Polynesian speakers?"
Unlikely. Originally anyway. The authors of link you provided express surprise that the language diversity seems to be only about 3500 years ago. To me this is a little on the young side, but perhaps not much. There must have been an earlier non-Malayo-Polynesian language through the region. And the same language, or group of languages, was present before the Austroasiatic language expansion.
"Instead of postulating Austronesians obliterating pre-existing Papuan populations but 'sucking out' all their Y-DNA C2 lineages from them (a major logical problem for me)"
I don't see why it's a problem. C2 seems to have originated in the southern region of Wallacea. So it's part of the backflow from the south that the authors see. It was not originally Austronesian-speaking.
The authors do admit (bottom of page 13 and top of page 14) that the Austronesians are probably a product of increased trade. What they don't say is that this trade is almost certainly a product of improved boating technology. such an improvement would most likely lead to the complicated language distribution the authors see through the region.
"the absence of any vestiges of non-Austronesian speakers in Wallacea".
Possibly because it had become uninhabited by the time of the Austronesian expansion.
"The authors are absolutely convinced that the Austronesian language originated in Taiwan."
Yes and no. They don't dispute it de jure but they do duspute it de facto.
"Consequently, the impression of a graduated dispersal of Malayo-Polynesian languages south from Taiwan does not hold up and must be rethought. If we restrict our inquiry to the linguistic evidence alone, there was a rapid, multidirectional, and multimodal propagation of Malayo-Polynesian languages across most of ISEA, from southern Indonesia to the Batanes Islands in the north, without
direction or hierarchy (see fig. 2B; Ross 2005). Can we ascertain a center for the dispersal? On purely linguistic grounds, the answer must be no. It is true that the presence of nine of the primary subgroups of Austronesian in Taiwan unambiguously
indicates that Taiwan is the homeland for the language family
and that, ultimately, the Malayo-Polynesian languages are the
southern branch of the higher Austronesian node, but this says
nothing about the internal relationships, dispersals, and later
history of Malayo-Polynesian."
The theory of out-of-Taiwan origin of the Austronesian family, therefore, is derived from formal linguistic factors ("shared innovations," which btw can be product of convergence as in the case of a couple of languages in Taiwan that share some Malayo-Polynesian innovations but not others) but not from linguistic geography. We supposedly know where the homeland of Austronesian was but we don't know where the home of Malayo-Polynesian was, although the latter is supposed to be younger hence better preserved in the "linguistic record." So, scholars who work on dispersals are forced to accept it but it's not the most "actionable" theory out there.
Then, if you look at Ilia Peiros's work (e.g., http://books.google.com/books?id=yq89xWtkguoC&pg=PA177&dq=peiros+formosan&hl=en&ei=72XRTNOaCcSAlAe-9JCzDA&sa=X&oi=book_result&ct=result&resnum=1&ved=0CCgQ6AEwAA#v=onepage&q=peiros%20formosan&f=false), you'll see that the 9 Formosan branches could in fact be collapsable into one, so we may look at Austronesian phylogeny as simply divided into the Formosan and the Malayo-Polynesian branches.
Finally, if you add Ongan and its possible (and plausible) connection to Proto-Austronesian (and not Austroasiatic! and not Malayo-Polynesian) into the mix, then the out-of-Taiwan theory becomes even more of a problem. South China could still work - as a mid-point between Taiwan and Andaman islands - but it's unlikely that an inland population would have the boating skills to colonize all of the Pacific and parts of the Indian Ocean.
"The pre-Austronesian agricultural substrate in the region is the Hoabinhian. I agree that this is not a product of the Chinese Neolithic, and may even have been the source for it. But the authors don't mention the Hoabinhian at all. And the people almost certainly didn't speak and Austonesian or Austroasiatic language."
You cannot be so sure, Terry. Hoabhinian is something I would look into very carefully. Especially in light of the Ongan connection.
"C2 seems to have originated in the southern region of Wallacea. So it's part of the backflow from the south that the authors see. It was not originally Austronesian-speaking."
You have to have C2 attested in non-Austronesian speaking peoples to make such a claim. As of now, C2 is the core of the Austronesian Y-DNA pool and the one dominating in the most recently colonzied areas, namely Polynesian. Unless C2 was under selection, I can't see how it could be picked up from a substratum and then driven to fixation in the most recent branch.
"It is true that the presence of nine of the primary subgroups of Austronesian in Taiwan unambiguously indicates that Taiwan is the homeland for the language family and that, ultimately, the Malayo-Polynesian languages are the southern branch of the higher Austronesian node"
The authors appear to accept completely that idea without question.
"We supposedly know where the homeland of Austronesian was but we don't know where the home of Malayo-Polynesian was"
But it's reasonable to suppose it began somewhere near Taiwan. The Philippines seems an obvious choice.
"but this says nothing about the internal relationships, dispersals, and later history of Malayo-Polynesian."
It does seem from the article that there was much movement around island SE Asia after that first emergence however.
"Hoabhinian is something I would look into very carefully. Especially in light of the Ongan connection".
There is evidence of a connection between the Andamans and the Toalean of island SE Asia. It's sort of pre-Hoabinhian. The connection is actually evidence fro a relatively late settling of the Andamans, not some anciet drop-off point on the great southern coastal migration.
"As of now, C2 is the core of the Austronesian Y-DNA pool and the one dominating in the most recently colonzied areas, namely Polynesian. Unless C2 was under selection, I can't see how it could be picked up from a substratum and then driven to fixation in the most recent branch".
I don't see that as really much of a problem. Haplogroups are very seldom intimately connected to language. For example it seems likely to me that Indo-European was spread first by Y-DNA R1a. But that haplogroup is far from common in Western Europe where Indo-European alnguages are almost universal.
"The authors appear to accept completely that idea without question."
They were unaware of the Blevins discovery at that time. The Ongan-Austronesian discovery hands a major blow to the out-of-Taiwan theory. Plus as you can see from reading Peiros there's a diversity of opinions. I'm not an Austronesianist, hence I can't fully evaluate the out-of-Taiwan theory on formal linguistic grounds, but judging from the situation in Indo-European (the languages I studied more intimately), Uralic and Tibeto-Burman linguistics hierarchical genealogical trees of linguistic relatedness over time give rise to "flatter" models. This flattening has already affected the Malayo-Polynesian node, as you can collect from the Denham-Donohue paper. Let's see what the next few years will bring.
"There is evidence of a connection between the Andamans and the Toalean of island SE Asia. It's sort of pre-Hoabinhian. The connection is actually evidence fro a relatively late settling of the Andamans, not some anciet drop-off point on the great southern coastal migration."
If Toalean is pre-Hoabinhian, then it gives a much deeper time horizon to proto-Austronesian. I completely agree with you that the Andamans is not a drop-off point in a coastal migration out of Africa but the presence of Y-DNA D in the Andamanese seems to be in a fairly good correlation with the presence of Y-DNA C in Austronesians, both of Pleistocene, pre-K* origin. I know you explain hg D as a later drop-off from coastal Southeast Asia. This is possible but not the most parsimonious.
"Haplogroups are very seldom intimately connected to language."
In this case I was talking about populations, not languages. If C2 was absorbed from a pre-"Austronesian" (forget the linguistic label for a second) population, of which we don't have any evidence but which apparently had this lineage at high frequencies, but then achieved high frequencies in a most recently derived population, namely Polynesians, then it's odd unless selection enables first the survival of C2 in the midst of Asian-derived genes and then it's propagation back to high frequencies.
As a general point, I think haplogroups are always connected to languages. It's just hard to pin down which haplogroup(s) was/were originally connected to which language grouping. Again, taking the data at face value it looks like the original spread of what would become "Austronesian language family" is associated with the emergence of Y-DNA C2 off of the C root at a point when C and D were still very common and geographically contiguous. This ancient lineage survives and thrives only in the southern region running from southern Wallacea eastward through Polynesia. No other Y-DNA lineage seem to be so uniquely associated with the populations speaking Austronesian languages.
At least this is the story that comes through Y-DNA. mtDNA is different, I admit but I do see much resemblance between the phylogeography of mtDNA B4 and Y-DNA C2, as I wrote earlier.
"I know you explain hg D as a later drop-off from coastal Southeast Asia".
No. To me it seems most likely that D moved south down through the hill country bordering Tibet and Western China, before it was able to cross the open water to the Andamans. Maju places D further south than I do, but it's very difficult to make a case for it being ancient island SE Asia.
"the emergence of Y-DNA C2 off of the C root at a point when C and D were still very common and geographically contiguous".
I'm sure they had parted long before they mat up again in island SE Asia. And reached that region by different routes.
"No other Y-DNA lineage seem to be so uniquely associated with the populations speaking Austronesian languages".
I'm not so sure of that. I don't think Y-DNA C reached Madagascar at all. And C2, which is specifically the Eastern Austronesian C Y-DNA, is not found west of Central Indonesia. To me O1 is very much associated with the early stages of the Austronesian expansion. Then O3 moves east with it (accompanied by C2), and O2a moves west.
"If C2 was absorbed from a pre-'Austronesian' (forget the linguistic label for a second) population, of which we don't have any evidence but which apparently had this lineage at high frequencies, but then achieved high frequencies in a most recently derived population, namely Polynesians, then it's odd unless selection enables first the survival of C2 in the midst of Asian-derived genes and then it's propagation back to high frequencies".
If, as seems very likely from its distribution, C2 originated in Southern Wallacea it most certainly wasn't 'in the midst of Asian-derived genes'. It was quite separate. They arrived later, from the north. To me it's no surprise that a recently absorbed haplogroup should be able to take over.
"Let's see what the next few years will bring".
Amen.
"I don't think Y-DNA C reached Madagascar at all. And C2, which is specifically the Eastern Austronesian C Y-DNA, is not found west of Central Indonesia."
This only means that the colonizers of Madagascar lost it. Micronesia and Indonesia have C*, which is again not found in Taiwan.
Micronesia: 1/16 = 6.3% C-M38*(xM208, P33)
Indonesia
Maluku Islands: 10/36 = 27.8% C-M38*(xM208, P33)
Sulawesi (West Province): 6/48 = 12.5% C-M38*(xM208, P33)
Timor: 4/12 = 33.3% C-M38*(xM208, P33)
Sumba: 202/352 = 57.4% C-M38*(xM208, P33)
Flores: 81/485 = 16.7% C-M38*(xM208, P33)
Mentawai: 1/74 = 1.4% C-M38*(xM208, P33)
Central-Eastern Indonesian (Maluku + West Sulawesi + Timor + Sumba + Flores) mean percentage C-M38*(xM208, P33): 29.5%
Murray P. Cox, Alan J. Redd, Tatiana M. Karafet et al., "A Polynesian Motif on the Y Chromosome: Population Structure in Remote Oceania," Human Biology, October 2007, v. 79, no. 5, pp. 525–535.
So, what this shows us is that Malayo-Polynesian languages are associated with at least two Y-DNA C lineages: C2 in the east and C* (C7?) in the west. You can postulate the massive absorption of indigenous males to account for this, but where's the evidence for the existence of those non-Austronesian, pre-Austronesian populations? It's possible that those indigenous populations spoke Austroasiatic languages but it's purely hypothetical. As the data stands, Austronesian is associated with a stack of haplogroups from Pleistocene ones to Holocene ones. The deepest node C is not attested in Papua New Guinea (hence it couldn't come from Papuan gene flow/substratum in Melanesia), but it is attested in Australia. The most parsimonious solution to this is that the expansion of Austronesian languages began in all directions after a lengthy period of their existence as a dialect chain between Wallacea and New Britain. In other words, Austronesian could be just another kind of Trans-New Guinean family, an old expansive Pacific population with ultimate roots in pre-10K foraging lifestyle and not a more recent Neolithic agricultural population from Taiwan.
"This only means that the colonizers of Madagascar lost it".
What evidence do have for that? It's doubtful Y-DNA C ever reached there.
"Micronesia and Indonesia have C*, which is again not found in Taiwan".
And it probably spread with the Austronesians. But it doesn't automatically follow that the language originated amoung them. After all Y-DNA R1b has been responsible for the spread of many Indo-European languages, but not many would claim it originated with them.
"what this shows us is that Malayo-Polynesian languages are associated with at least two Y-DNA C lineages: C2 in the east and C* (C7?) in the west".
Only as far west as Mentawai, off Sumatra. And not to mention the Austronesian association with Y-DNAs O1, O2, and O3 through much of its distribution?
"You can postulate the massive absorption of indigenous males to account for this"
Not particularly massive. In any Austronesian population, apart from in Polynesia itself, individual Y-DNA Os (sometimes more than one of them) represent a far greater proportion of Y-DNAs than does Y-DNA C.
"where's the evidence for the existence of those non-Austronesian, pre-Austronesian populations?"
The authors of the paper you linked to were sure that Austronesian had a relatively young expansion. That seems sensible to me, as if it was very ancient we would be unable to see the connections anyway. People were living through much of island SE Asia from fairly ancient times. Austronesian must have spread originally from just a single region within that spread. The evidence overwhelmingly suggests Taiwan. But where it came from before then is debatable.
"It's possible that those indigenous populations spoke Austroasiatic languages"
And it's very possible that many of them didn't. Some other language is more likely.
"As the data stands, Austronesian is associated with a stack of haplogroups from Pleistocene ones to Holocene ones".
Do you find that surprising? If its spread is relatively recent surely that's what we would expect.
"The most parsimonious solution to this is that the expansion of Austronesian languages began in all directions after a lengthy period of their existence as a dialect chain between Wallacea and New Britain".
The authors of the paper you provided say it's not very lengthy at all. They admit (bottom of page 13 and top of page 14) that the Austronesians are probably a product of increased trade. What they don't say is that this increase is almost certainly a product of improved boating technology. They argue that Taiwan was not included in pre-existing trading patterns until just 4000 years ago (bottom of page 16). Their problem becomes, 'how did the language from Taiwan come to replace previous ones?'. Not a huge problem. It has happened many times before, in many parts of the world.
"Austronesian could be just another kind of Trans-New Guinean family, an old expansive Pacific population with ultimate roots in pre-10K foraging lifestyle and not a more recent Neolithic agricultural population from Taiwan".
It's very difficult to support that theory.
"Only as far west as Mentawai, off Sumatra. And not to mention the Austronesian association with Y-DNAs O1, O2, and O3 through much of its distribution?"
Remember C* (C7?) is all over the Philippines, both Negrito and non-Negrito, at low frequencies. So, it pretty much covers all the corners of the Malayo-Polynesian world, with the exception of Taiwan and Madagascar.
As for Y-DNA Os, of course, they came from Asia, but why are we so sure they brought the Austronesian languages with them? This Mongoloid migration could've been absorbed by the indigenous Austronesian-speaking population, and the fact that Os decrease from northwest to southeast, while Cs decrease from southeast to northwest (Cs at Sumba and Timor are 33 to 57%) may mean that Os stopped short of penetrating deep into the long-term Austronesian refugias between Wallacea and New Britain, and hence this Asian/Mongoloid migration cannot account for the eastern areas of the Austronesian spread.
"Not particularly massive. In any Austronesian population, apart from in Polynesia itself, individual Y-DNA Os (sometimes more than one of them) represent a far greater proportion of Y-DNAs than does Y-DNA C."
Again, Sumba and Timor show appreciable frequencies of Cs.
"Their problem becomes, 'how did the language from Taiwan come to replace previous ones?'. Not a huge problem. It has happened many times before, in many parts of the world."
Evidence for these replacements is often anecdotal. African Pygmies is another case of supposedly language replacement but again there's little to no evidence for earlier Pygmy tongues. In Europe at least, there are place names that Vennemann argued come from a Vasconic substratum. But even here he doesn't invent a whole new language that was lost without a trace but connects it to an existing surviving member of this once-larger Vasconic family. This is one of the requirements for postulating language replacement: find its traces in the languages of colonizers and then connect them to an existing family.
"It's very difficult to support that theory."
The out of Taiwan theory has its problems, too.
"The out of Taiwan theory has its problems, too".
But not as many, and mostly smaller ones.
"Remember C* (C7?) is all over the Philippines"
As I understand things it's several different Cs, not named yet. And they are spread right around the South China Sea and not just confined to the Malayo-Polynesian region. In fact Polynesian C is specifically C2. I agree Y-DNA C is an early arrival in the region though.
"This Mongoloid migration could've been absorbed by the indigenous Austronesian-speaking population"
Could have been. But Austronesian has all the characteristics of having expanded recently. So its diversity is hardly Paleolithic.
"the fact that Os decrease from northwest to southeast"
Not really. It's just that very little crosses Wallace's Line.
"Cs decrease from southeast to northwest (Cs at Sumba and Timor are 33 to 57%)"
Specifically C2. And the huge amount of evidence Ebizur provided shows that C2 expanded north and east from the region of Sumba and Timor about the time of, or a little before, the apparent Austronesian expansion. But it certainly didn't take part in the Austronesian expansion in any other direction. So it's not fundamentally Austronesian.
"may mean that Os stopped short of penetrating deep into the long-term Austronesian refugias between Wallacea and New Britain"
But Austronesians penetrated only to offshore regions, which were presumably uninhabited until they arrived. The languages are found only in a thin, discontinuous coastal strip in New Guinea and are far from common in the Northern Solomons. The Austronesians were a coastal and island people. Their expansion fits well with theories of improved boating which was transfered to other groups as they passed them by. Members of these groups appear to have then followed along behind. We find New Guinea haplogroups making it as far as Fiji for example.
"and hence this Asian/Mongoloid migration cannot account for the eastern areas of the Austronesian spread".
It can really. O3 reached Western Polynesia, and is even present in Central Polynesia. O3 is almost certainly 'Asian/Mongoloid'.
"Again, Sumba and Timor show appreciable frequencies of Cs".
And 'appreciable frequencies' of O1. Almost as much in fact.
"he doesn't invent a whole new language that was lost without a trace"
Surely it's reasonable to suppose that languages related to New Guinea languages were spoken across the region, possibly as far as the Andamans. It's doubtful that the New Guinea languages sprang up on that island without having brought in some ancestral language.
"As I understand things it's several different Cs, not named yet. And they are spread right around the South China Sea and not just confined to the Malayo-Polynesian region. In fact Polynesian C is specifically C2. I agree Y-DNA C is an early arrival in the region though."
It's possible that those are different Cs, and each one of them is different from C2. But this will only leave us with a Pleistocene situation exactly mirroring the Holocene situation: there are several branches of Y-DNA O and there are several branches of Y-DNA C found in Austronesian speakers. In both cases, they are not exclusive to Austronesian speakers. Only C2 is.
It's noteworthy that Tai-Kadai speakers have Y-DNA F*, D (just like Andamanese but maybe a different sub-branch), C* and C3 (!). They were all detected on the Hainan island (Hlai or Li language), which was apparently peopled during the last LGM. None of them are found in Taiwan. And Tai-Kadai languages are considered to be related to Austronesian languages (sometimes even closer to Malayo-Polynesian languages than to Austronesian as a whole). So we have Andamanese, Tai-Kadai, Malayo-Polynesian, Formosan and Malgasy clusters. Only Formosan and Malgasy are associated exclusively with Holocene Y-DNA lineages, the rest have a strong Pleistocene heritage in them. In Andamanese, a branch of D prevailed over others because they were apparently shielded from the massive Holocene migrations. In Polynesians, a branch of C prevailed, likely for the same reason.
"O3 reached Western Polynesia, and is even present in Central Polynesia. O3 is almost certainly 'Asian/Mongoloid'."
I agree with the last statement. Could it have been brought to Western and central Polynesia after these areas had been already colonized?
"And the huge amount of evidence Ebizur provided shows that C2 expanded north and east from the region of Sumba and Timor about the time of, or a little before, the apparent Austronesian expansion."
Then it's hard for me to understand in what form had those Cs existed in ISEA, Wallacea and Melanesia before Austronesians picked them up. They weren't all just embryonic C*, they were all C2, C7, Cx, etc. It sounds like Austronesians stripped the branch of all its leaves and then put their own, new mutations onto an old branch and thus expanded "recently." But this is impossible. Also, a population can't hop from having a C7 lineage to having a C2 lineage.
It would be much easier to imagine a population living in ISEA or south China 40K years BP and having a bunch of stratigraphically early Eurasian lineages such as C, F and D and then breaking up into Malayo-Polynesian, Andamanese, Tai-Kadai, Australian clusters that would settle in different places around the Pacific and the Indian Ocean. Later around 10K parts of these clusters would expand in size and territory and evolve new lineages, among which is O. These O lineages would recolonize the same broad territory without reaching to the extremes of the distribution of the original Pleistocene lineages (Andaman islands, Australia and parts of Melanesia were spared).
BTW, genetic dates for Austronesian lineages, Y-DNA and mtDNA alike, Os or Cs or Bs, are all much older than the time horizon offered by linguists and archaeologists for the expansion of Austronesians from Taiwan. They are usually of Late Pleistocene/Early Holocene age.
"genetic dates for Austronesian lineages, Y-DNA and mtDNA alike, Os or Cs or Bs, are all much older than the time horizon offered by linguists and archaeologists for the expansion of Austronesians from Taiwan. They are usually of Late Pleistocene/Early Holocene age".
Which indicates Austronesian was spread to some extent independently of haplogroups.
"But this will only leave us with a Pleistocene situation exactly mirroring the Holocene situation: there are several branches of Y-DNA O and there are several branches of Y-DNA C found in Austronesian speakers".
I doubt very much that O was present in SE Asia during the Pleistocene.
"They were all detected on the Hainan island (Hlai or Li language), which was apparently peopled during the last LGM. None of them are found in Taiwan".
Which indicates Hainan was not settled from Taiwan.
"And Tai-Kadai languages are considered to be related to Austronesian languages"
I agree with that, although many argue against it. I'm fairly sure they, along with Mon-Kmer and Austro-Asiatic, expanded into South China and SE Asia during the Holocene.
"Could it have been brought to Western and central Polynesia after these areas had been already colonized?"
Quite possibly. But that doesn't solve the problem of connection between Taiwan and the Austronesian origin.
"Then it's hard for me to understand in what form had those Cs existed in ISEA, Wallacea and Melanesia before Austronesians picked them up. They weren't all just embryonic C*, they were all C2, C7, Cx, etc."
A whole series of different Cs. In time I suppose we will finish up with C7, C8, C9 and so on round the South China Sea. C2 coalesced south of this, in Tengarra. C4 coalesced in Australia. But of all those Cs just C2 moved northeast and out into the Pacific. But Austronesians also moved west, without C2.
"It would be much easier to imagine a population living in ISEA or south China 40K years BP and having a bunch of stratigraphically early Eurasian lineages such as C, F and D and then breaking up into Malayo-Polynesian, Andamanese, Tai-Kadai, Australian clusters that would settle in different places around the Pacific and the Indian Ocean".
To some extent I agree. Although I wouldn't include MP and TK in that. And the peole did not spread far into the Pacific or Indian Oceans until just a few thousand yeasr ago.
"Later around 10K parts of these clusters would expand in size and territory and evolve new lineages, among which is O".
I'm fairly sure Os origin is futher north. It represents a southward movement into South China and SE Asia.
"I doubt very much that O was present in SE Asia during the Pleistocene."
I agree. What I meant is that Y-DNA O is a Holocene hg, and Y-DNA C is a Pleistocene hg, but in both cases we have several branches of these haplogroups associated with Austronesian speakers. And it's only C2, from the known C lineages, that is exclusively Austronesian. All the Os are found throughout Asia from South Siberia (O1) down into Oceania.
"I'm fairly sure Os origin is futher north. It represents a southward movement into South China and SE Asia."
Okay, this makes sense. It represents the (Northern) Mongoloid migration to SEA and Oceania. Some of these Os end up in western Papua New Guinea, among non-Austronesian speakers, as admixture with Austronesians. So, this process is clear to me. But where do we see proof that Austronesians absorbed their Cs from the Papuan substrate. We need to have Cs widely spread across Papua New Guinea and then appearing in Austronesians ar reduced frequencies to begin seeing this admixture.
"Which indicates Austronesian was spread to some extent independently of haplogroups."
One thing I'm certain about is that Y-DNA O was carried by Austronesians to some remote places but it didn't originate with them. Judging by the presence of a sister of Proto-Austronesian in Taiwan associated with hg D and by the presence of Cs in the Philippines, Wallacea, Melanesia and Polynesia I suspect that Austronesian is much older than the dates offered by linguists and archaeologists. This doesn't mean that Austronesians didn't move around and trade a lot in the past 5,000 years.
Also, if you read Blevins' "A Long Lost Sister of Proto-Austronesian?
Proto-Ongan, Mother of Jarawa and Onge of the Andaman Islands" (available for free online), on pp. 190-191 she writes: "A striking feature of PON and PAN is their conservative historical phonologies. Both are recognizable as sisters, and have changed little from PAO. I suggest that this conservatism is a consequence of two independent features of Proto-Austronesian-Ongan: first, its relatively simple (C)V(C) syllable structure; and second, its dearth of inflectional morphology. Inflectional morphology gives rise to multimember paradigms, and when paradigms exist, analogical changes (e.g., leveling, extension) are likely to follow. PAO did not have such a system and, as a consequence, analogical change has not muddied the clear waters of regular sound change. As a consequence of relatively simple sound patterns and isolating morphology, then, it appears that PON, PAN, and many of their daughters have changed little over thousands of years."
If conservatism is a feature of Austronesian languages including Ongan, then the extreme divergence of Taiwanese languages may be a derived feature.
Correction: "Judging by the presence of a sister of Proto-Austronesian in Taiwan associated with hg D..." to read " Judging by the presence of a sister language of Proto-Austronesian in the Andaman islands associated with hg D..."
"But where do we see proof that Austronesians absorbed their Cs from the Papuan substrate. We need to have Cs widely spread across Papua New Guinea and then appearing in Austronesians ar reduced frequencies to begin seeing this admixture".
I'm certain that the Austronesian Cs did not enter that language group from New Guinea. On the contrary the only C to enter New Guinea is C2, and it seems to have done so about the time of the Austronesian expansion. C2 was almost certainly absorbed from a south Wallacean substrate. It came to dominate the further Polynesian gene pool but didn't spread west.
"One thing I'm certain about is that Y-DNA O was carried by Austronesians to some remote places but it didn't originate with them"
I wouldn't be so sure that 'it didn't originate with them'. Certainly not with all of them, but very likely with O1.
"Judging by the presence of a sister language of Proto-Austronesian in the Andaman islands associated with hg D..."
As far as I know that connection is far from universally accepted. And any connection may be through a common substrate.
"If conservatism is a feature of Austronesian languages including Ongan, then the extreme divergence of Taiwanese languages may be a derived feature".
It is certain that Malayo-Polynesian diversified considerably once it had moved into the uninhabited regions of the Pacific. So it seems unlikely to me that it had been previously very conservative.
"I wouldn't be so sure that 'it didn't originate with them'. Certainly not with all of them, but very likely with O1."
How do you explain the presence of O1 in South Siberia? I thought you were saying that all Os in Oceania are migrants from East Asia. Plus, according to Li et al. O1 arrived in Taiwan from ISEA or the Philippines.
"As far as I know that connection is far from universally accepted. And any connection may be through a common substrate."
It just came out of the blue. Austronesianists haven't been prepared for such a turnaround. But it came from a very respectable linguistic institution and is remarkably solid from the point of view of meeting the requirements for establishing a linguistic relationship. Regarding a "common substrate", Blevins in that paper noticed a few striking similarities between Ongan words and some Philippine Negrito words that Reid described as potentially pre-Austronesian. In her opinion, however, they are not pre-Austronesian but ancient Austronesian-Ongan.
"C2 was almost certainly absorbed from a south Wallacean substrate."
Well, this substrate is not attested, by which I mean that you got to have a non-Austronesian speaking population carrying C2 somewhere in ISEA or Oceania to claim that it was absorbed by Austronesians from a Wallacean substrate to which that attested population was related. You can't use the same empirical observation as both an argument and a proof.
"How do you explain the presence of O1 in South Siberia?"
I didn't know there was any there. There is certainly a very small proportion in Northern China, but if O's expansion is related to the Yangtze Neolithic that is hardly surprising.
"Plus, according to Li et al. O1 arrived in Taiwan from ISEA or the Philippines".
The Chinese are very keen to minimise the possibility that there has been any historical expansion from China. That's why they're so keen to place O's origin as a whole in SE Asia rather than in China. They wish to demonstrate the Chinese are not an expansionist people, in spite of the evidence of the Han expansion over the last two thousand years. It's reasonable to suppose that the Han expansion is simply a continuation of more ancient processes.
"Austronesianists haven't been prepared for such a turnaround".
and not many have been convinced so far.
"you got to have a non-Austronesian speaking population carrying C2 somewhere in ISEA or Oceania to claim that it was absorbed by Austronesians from a Wallacean substrate to which that attested population was related".
Not really. The islands are rather small so total language replacement is very possible, in fact probable. Non-Austronesian languages certainly survive on the nearby mainland although these too appear to be related to Austronesian and to languages further north.
"I didn't know there was any there."
"Haplogroup O1a-M119 Y-chromosomes also have been found to occur at low frequency among various populations of Siberia, such as the Nivkhs (one of 17 sampled Y-chromosomes), Ulchi/Nanai (2/53), Yenisey Evenks (1/31), and especially the Buryats living in the Sayan-Baikal uplands of Irkutsk Oblast (6/13)." (Wiki)
"O's expansion is related to the Yangtze Neolithic that is hardly surprising."
Southern China and southern Siberia are thousands of miles apart, though, with no O1 in between.
"and not many have been convinced so far."
People tend to reject new ideas for a long time. Notably, nobody has shown Blevins to be wrong.
"The islands are rather small so total language replacement is very possible, in fact probable. Non-Austronesian languages certainly survive on the nearby mainland although these too appear to be related to Austronesian and to languages further north."
Non-Austronesian language have survived on islands as well. Too bad, they don't show indigenous C lineages to support your claims.
For instance, Unexpected NRY Chromosome Variation in Northern Island Melanesia, by Laura Scheinfeldt et al., shows frequencies of C lineages in island Melanesia. In New Ireland, Kuot is a Papuan language surrounded by Austronesian languages. All of them have C2-M38. In West New Britan, Anem is a Papuan language and it has C2b. The rest of the Papuan-speaking communities don't have C at all, while many Austronesian-speaking do. I would interpret it as gene flow from Austronesians to Papuans and not the other way around.
Also, of interest, from the same paper, "Any recent Southeast Asian component (associated with the development of Lapita) is most certainly associated
with haplogroup O3-M122, which has a remarkably low frequency in this sample series (18 of 685 samples or
2.6%). Even if all O haplogroups might have been introduced just with the immediate Southeast Asian ancestral component of the Lapita peoples, their frequency in our series totals just 8% (table 2), almost entirely restricted to
Oceanic-speaking populations. In our sampling strategy, we may have missed higher concentrations of the O haplogroups. For example, many Lapita sites in the Bismarck Archipelago are located on smaller or offshore islands and we did not sample those, except for Mussau (where no O samples were found)."
This is paralleled by the lack of mtDNA hg B in ancient Lapita samples.
So, how could agriculturalists almost completely replace their own men with indigenous ones, while obliterating local languages without a trace? I think the data shows that Austronesians were donors of BOTH more ancient and more recent haplogroups.
"Haplogroup O1a-M119 Y-chromosomes also have been found to occur at low frequency among various populations of Siberia"
Thanks for that information. Note though: 'low frequency among various populations'. They are sort of eratics.
"Southern China and southern Siberia are thousands of miles apart, though, with no O1 in between".
According to a map I have O1 is found in North China, so no real problem.
"Non-Austronesian language have survived on islands as well".
Only on islands that were almost certainly already occupied by the time of the Austronesian expansion. That has been accepted for many years. Although it has also been accepted that some of these non-Austronesian-speaking groups followed along behind the Austronesian expansion.
"Too bad, they don't show indigenous C lineages to support your claims".
I keep telling you that there were no C's anywhere in the Pacific, or even east of Southern Wallacea, until the Austronesian expansion. Certainly there were none in New Guinea or in the islands to its north and east. So if anything the absence of 'indigenous C lineages' supports my claims. And the claims of virtually every scientist involved in studying the prehistory of the region.
"frequencies of C lineages in island Melanesia".
They are a product of the Austronesian expansion.
"In New Ireland, Kuot is a Papuan language surrounded by Austronesian languages. All of them have C2-M38".
People were present in New Ireland before the Austronesians arrived. So it seems that the incoming Y-chromosome has replaced the indigenous one, but the indigenous language has survived.
"In West New Britan, Anem is a Papuan language and it has C2b".
Same situation.
"The rest of the Papuan-speaking communities don't have C at all"
That's the pre-Austronesian situation.
"many Austronesian-speaking do".
Which is exactly what we would expect.
"I would interpret it as gene flow from Austronesians to Papuans and not the other way around".
The evidence you've just provided suggests very strongly that the Austronesian C2 is intrusive to New Britain/New Ireland. C2 was absent in the original Papuan-speaking population.
"Any recent Southeast Asian component (associated with the development of Lapita) is most certainly associated
with haplogroup O3-M122"
Exactly what I've been telling you all along.
"which has a remarkably low frequency in this sample series (18 of 685 samples"
Again that supports what has been accepted for many years: 'some of these non-Austronesian-speaking groups followed along behind the Austronesian expansion'. O3 has been replaced by the expanding Melanesian haplogroups and the C2 that came in with O3. O1 did not form very much of the Austronesian expansion beyond Wallacea itself.
"Even if all O haplogroups might have been introduced just with the immediate Southeast Asian ancestral component of the Lapita peoples, their frequency in our series totals just 8% (table 2), almost entirely restricted to
Oceanic-speaking populations".
Again exactly as has been accepted for many years. By the time of the Lapita expansion the people involved had become a mixture. And the Lapita people dodged islands, such as New Britain, New Ireland and the Northern Solomons, that were already occupied. The article even hints a such a factor:
"many Lapita sites in the Bismarck Archipelago are located on smaller or offshore islands"
The Lapita expansion began in the Bismark Archipeligo, but almost certainly from the Admiralty Islands, not from any of the larger ones. It seems very likely that the Austronesian-speaking people found the Admiralties unoccupied when they arrived.
"This is paralleled by the lack of mtDNA hg B in ancient Lapita samples"
I wasn't aware of that. What haplogroup(s) is there? And how come B4 came to dominate further east?
"how could agriculturalists almost completely replace their own men with indigenous ones, while obliterating local languages without a trace?"
You've just gone to a huge amount of trouble to show that the local languages didn't die out where people were already present when the Austronesians arrived. However it seems as though incoming Y-DNA introgressed into some of those local populations. Hardly surprising.
"You've just gone to a huge amount of trouble to show that the local languages didn't die out where people were already present when the Austronesians arrived. However it seems as though incoming Y-DNA introgressed into some of those local populations."
Terry, you've been arguing all along that Austronesians picked up their C lineage(s) from indigenous Papuan (meaning any non-Austronesian, Plestocene in origin) populations. I showed you the data that demonstrates the opposite: from what we know, Austronesians were the donors of Cs (and Os) into indigenous Papuan populations. That's why the absurd historical accident of Austronesians absorbing Cs from a non-Austronesian substratum and expanding them, while replacing this substratum linguistically has never happened. The data seems to indicate that C2 is a native Austronesian language. In conjunction with the Ongan-Austronesian linguistic connection, it suggests that Austronesian is older than Chinese Neolithic, it exhibits unique lineages that stratigraphically (phylogenetically) belong to the earliest set of Eurasian lineages (Dx in Ongan, C2 and Cx in Malayo-Polynesian) (Karafet's stage 1 of the colonization of the Pacific), and it's earliest expansion signal ranged between the northwest (Andaman islands)- southeast (Wallacea, Near Oceania) axis.
"I wasn't aware of that. What haplogroup(s) is there? And how come B4 came to dominate further east?"
They just tested for 9bp deletion - an easy to spot marker that's shared by all B lineages in Asia and America -, and they didn't find it in their Lapita sample. See Hagelberg here (http://www.jstor.org/pss/49647) and here (http://onlinelibrary.wiley.com/doi/10.1002/elps.1150180907/abstract). I'll mention, in passing, that there's no hg B in the Andamanese either. I don't know how to best explain the absence of hg B in Lapita samples - aside from questioning aDNA results as unreliable - if not by suggesting that B*/B4 in Near Oceania is much older than 3500 years BP, that it didn't arrive straight from Asia to colonize Polynesia and that it had enough time to get drifted out in some places. Just like Y-DNA hg C*/C2 is old enough to have drifted out of the Borneo area from which Madagascar was later peopled.
"I don't know how to best explain the absence of hg B in Lapita samples"
Very strange and doesn't make sense.
"if not by suggesting that B*/B4 in Near Oceania is much older than 3500 years BP"
But it cannot possibly be older than that beyond the Northern Solomons.
"Just like Y-DNA hg C*/C2 is old enough to have drifted out of the Borneo area from which Madagascar was later peopled".
It's very unlikely that C2 was ever present in either of those places.
"you've been arguing all along that Austronesians picked up their C lineage(s) from indigenous Papuan (meaning any non-Austronesian, Plestocene in origin) populations".
Yes. From the pre-Austronesian population of Southern Wallacea.
"from what we know, Austronesians were the donors of Cs (and Os) into indigenous Papuan populations".
Yes. Into Papuan populations in New Guinea and islands north and east of it.
"I'll mention, in passing, that there's no hg B in the Andamanese either".
Which argues against a very deep connection between Ongan and Austronesian.
"it suggests that Austronesian is older than Chinese Neolithic"
It seems to me you are arguing in a circle. You're claiming that Austronesian is an ancient, slowly-changing language, and this shows that it's long-establ;ished in island SE Asia. You're also arguing that because Austronesian is long-estsblished in Wallacea it must be an ancient, slowly-changuing language. What if both suppositions are wrong?
"It seems to me you are arguing in a circle."
At this point, I'm just trying to understand if we've exhausted all other interpretations before settling on the recent out of Taiwan expansion, which requires additional awkward assumptions such as the absorption of indigenous lineages in southern Wallacea, the complete replacement of their languages followed by the expansion of those adopted lineages and their gene flow into the Papuans further out east.
"Which argues against a very deep connection between Ongan and Austronesian."
How does it argue against it? All most recent expansions of Austronesians such as to Madagascar involved mtDNA B and Y-DNA O. Ongan and Austronesian are related at a deep genealogical linguistic level and correspondingly Ongan and Malayo-Polynesians show the deepest Eurasian Y-DNA lineages, Dx and C2. I don't think this can be explained away. We often hear that linguistics and mtDNA genetics support replacement out of Taiwan, but it also appears that linguistics and Y-DNA genetics suggest continuity through Pleistocene.
"Ongan and Malayo-Polynesians show the deepest Eurasian Y-DNA lineages, Dx and C2".
But Ongan is just one of several Andaman languages, and it makes some sort of sense that Austronesian may have influenced the Southern Andamans substantially independently of any supposed common origin.
"out of Taiwan expansion, which requires additional awkward assumptions such as the absorption of indigenous lineages in southern Wallacea, the complete replacement of their languages followed by the expansion of those adopted lineages and their gene flow into the Papuans further out east".
Those 'additional awkward assumptions' appear to be what has happened. This is my take on the various Wallacean crossings, which fits all the evidence (I include the Philippines in 'Wallacea' because they are certainly not part of 'Sundaland):
To me it looks almost certain that Y-DNA C and mt-DNA N were the first to cross Wallace's Line, about 60,000 years ago. In Australia they became Y-DNA C4 and mtDNA S, O N13 and N14. None of these basal haplogroups reached New Guinea.
Closely related haplogroups are found west of Wallacea however. Various Y-hap Cs are spread around the South China Sea and into the Philippines (although they could be later arrivals in this latter region). C2's apparent coalescence in Tengarra suggests that the first Wallace's Line crossing to Australia was in the south, via Timor.
Some unknown time later, possibly as much as 20,000 years later or perhaps almost immediately, people were able to again cross Wallacea. Presumably during another ice age and period of lowered sea level. K-derived Y-DNAs S and M, and mtDNAs M27, M28 and M29/Q made it to previously uninhabited New Guinea. With them went the N-derived mtDNA P. This last haplogroup, along with a couple of Y-DNA Ks and mtDNAs M42, M14 and M15, also entered Australia at the same time.
The aticle you linked to suggests that over the period of lowest sea level there was a continuing improvement in boating technology and trading networks through SE Asia, including parts of Wallacea. Various haplogroups shifted around, presumably with the development of the Hoabinhian. There seems to have been a movement westward of New Guinea Y-haps into Sulawesi and Halmahera (and possibly the Philippines) and members of SE Asian mtDNA haplogroups such as M9 moved east.
But the big shift was when Y-hap O1 entered the picture, presumably via the Philippines. Perhaps those islands had been unoccupied until then. The post Dienekes recently put up on the Philippines showed the Wallacean haplogroups in unexpected places there, although they tend to be associated with presumed Aboriginal Philippine populations.
All the above people formed the substrate on which the Malayo-Polynesian languages were first imposed upon. So we could say that the Austronesians developed from the mix of people around Wallacea, from where they spread both east and west.
In Genetic Admixture History of Eastern Indonesia as Revealed by Y-Chromosome and Mitochondrial DNA Analysis, by Stefano Mona, there's a table (p. 1867) with frequencies of C* and C2 in southern Wallacea. There are 140 instances of C2 among Austronesian speaking groups against 19 instances of C2 in non-Austronesian-speaking groups. For C* the ratio is even wider: 48 against 1. Various Y-DNA O lineages are also more frequent in Austronesians vs. non-Austronesians but the gap is not as wide, and in the case O-M134 Pantar (NAN) exclusively has it.
Again, it looks like in southern Wallacea, just like everywhere else, Austronesians were the donors of C lineages into Papuan populations and not the other way around.
"But Ongan is just one of several Andaman language..."
Blevins established linguistic relationship between Jarawa and Onga (the two smallest but most divergent languages in the Andaman islands). Great Andamanese is a large collection of languages that apparently don't show a connection to Austronesian. See for details www.andamanese.net/Lg%20Sciences%20Vol%2031%20GA%20article.pdf. What's interesting is the fact that if Austronesians landed in the Anadaman islands relatively late we would've seen their recently derived mtDNA and Y-DNA lineages there. But we don't. What we find in Ongan and Jarawa, instead, are Y-DNA Dx, which is the same stratigraphic level as Y-DNA C.
"In Genetic Admixture History of Eastern Indonesia as Revealed by Y-Chromosome and Mitochondrial DNA Analysis, by Stefano Mona, there's a table (p. 1867) with frequencies of C* and C2 in southern Wallacea".
And the abstract starts off:
"Eastern Indonesia possesses more linguistic diversity than any other region in Southeast Asia, with both Austronesian (AN) languages that are of East Asian origin, as well as non-Austronesian (NAN) languages of likely Melanesian origin".
This contradicts what you have been saying. Another comment that contradicts your claims:
"both NRY and mtDNA data showed a complete lack of correlation between linguistic and genetic relationships, most likely reflecting genetic admixture and/or language shift ... a clear example of the lack of the often-assumed correlation between the genes and languages of human populations".
"in southern Wallacea, just like everywhere else, Austronesians were the donors of C lineages into Papuan populations and not the other way around".
The data above supports the idea that C2 was present in southern Wallacea long before the Austronesians reached it.
"What we find in Ongan and Jarawa, instead, are Y-DNA Dx, which is the same stratigraphic level as Y-DNA C".
But very distantly related to it. The Austronesian connection is problematic, but probably has a simple explanation. It's just that we haven't found it yet.
I've just gone back and I see that Dienekes has already blogged on the article above. I don't know if you've seen it:
http://dienekes.blogspot.com/2009/05/genetic-admixture-in-eastern-indonesia.html
"Eastern Indonesia possesses more linguistic diversity than any other region in Southeast Asia, with both Austronesian (AN) languages that are of East Asian origin, as well as non-Austronesian (NAN) languages of likely Melanesian origin".
This contradicts what you have been saying. Another comment that contradicts your claims:
"both NRY and mtDNA data showed a complete lack of correlation between linguistic and genetic relationships, most likely reflecting genetic admixture and/or language shift ... a clear example of the lack of the often-assumed correlation between the genes and languages of human populations".
"in southern Wallacea, just like everywhere else, Austronesians were the donors of C lineages into Papuan populations and not the other way around".
The data above supports the idea that C2 was present in southern Wallacea long before the Austronesians reached it."
How does this contradict what I've been saying? What "data above" supports the idea that C2 was present before Austronesians reached southern Wallacea? Terry, you can access the article for free and read it and look at the table I referenced. From the distribution of frequencies, it appears that C*/C2 is much more frequent in Austronesians than in non-Austronesians, which is a usual sign of admixture, an admixture from the former into the latter.
There's little doubt that Austronesians came in contact with non-Austronesians, just like Trans-New Guineans came in contact with non-Trans-new Guineans, etc. Austronesians is a distinct language family. It's also ultimately derived from (South)east Asia, just like Australians and Papuans. The problem is when and how. The data shows that the lineage that's usually considered of Pleistocene origin, namely C*/C2, appears to have been an inherent part of the Austronesian expansion and not a random lineage picked up from pre-/non-Austronesians. You and I just looked at frequencies in southern Wallacea and in island Melanesia.
Language shifts do occur just like gene flow, and in this part of the world correlation between languages and genes is weak. (This is not a rule, however.) But language shift and gene flow occur both ways, and the data shows that C*/C2 is the lineage that was absorbed by non-Austronesians from Austronesians.
"But very distantly related to it. The Austronesian connection is problematic, but probably has a simple explanation. It's just that we haven't found it yet."
Of course, if there's a "simple" explanation we'll find it. What's infinitely more interesting to me at this point, however, is the fact that Ongan is 100% Y-DNA Dx and some Polynesian groups are almost 100% C2x. There seems to be a clear parallelism between the fixation of a Pleistocene lineage in two extreme catty corner island points of Austronesian (para-Austronesian) spread. You wouldn't argue that Andamanese colonized Andaman islands 3000 years ago and replaced a pre-Andamanese population that used to be Y-DNA D. This is just not parsimonious. Then why do we need to argue that Austronesians picked up their Pleistocene lineage from pre-Austronesians?
Take the Ainu: they have Y-DNA C, D and O. O is clearly recent but C and D are of Pleistocene age. I wouldn't be surprised that there was an ancestral population in SEAsia that was 50% D and 50% C. This population gave rise to Andamanese, Tai-Kadai and Austronesians (let's leave Austroasiatic for now). As the ancestral unity broke apart, different parts migrated in different directions, new mutations (various Ks and Os) arose and slowly edged out the original Pleistocene pool. Its vestiges survived in a Wallacean/Near Oceanic and in an Andamanese refugia. Then, in Holocene, Austronesians expanded in several directions. So, in essence they are a Pleistocene population on agricultural "steroids" and not a Taiwanese icing on the SEA/Oceanic cake.
"the fact that Ongan is 100% Y-DNA Dx and some Polynesian groups are almost 100% C2x".
Argues against an ancient connection. We need more information on any connection between Ongan and Austronesian before we can jump to definite conclusions.
"I wouldn't be surprised that there was an ancestral population in SEAsia that was 50% D and 50% C".
I'd put that population well north of SE Asia. Somewhere round the China/Mongolia/Tibet border region. From there C and D moved east to Japan, and then C moved south to Wallacea and Australia, and D moved south through the Chinese hill country to Burma and the Andamans. Some small amount of D reached Sumatra.
"This population gave rise to Andamanese, Tai-Kadai and Austronesians (let's leave Austroasiatic for now)".
I'd include Austroasiatic, along with Sino-Tibetan, Ket, Yenesian and Na-Dene.
"Then, in Holocene, Austronesians expanded in several directions".
I think everyone agrees with that.
"So, in essence they are a Pleistocene population on agricultural 'steroids' and not a Taiwanese icing on the SEA/Oceanic cake".
To most people the evidence is overwhelming that the language emerged from Taiwan. However I agree that the Austronesian expansion consisted in the main of Pleistocen SE Asian populations.
Sorry. I didn't see your previous post.
"This contradicts what you have been saying".
I'm sure it has been you who has been saying no pre-Austronesian languages exist in the region.
"Austronesians is a distinct language family. It's also ultimately derived from (South)east Asia, just like Australians and Papuans".
But austronesian is not related at all to Australian or Papuan languages.
"The data shows that the lineage that's usually considered of Pleistocene origin, namely C*/C2, appears to have been an inherent part of the Austronesian expansion and not a random lineage picked up from pre-/non-Austronesians"
Incorrect. C is by no means present in all Austronesian-speaking populations. And of course C2 derives from a C* of some sort, presumably in the Pleistocene and most likely in Southern Wallacea. But that in no way proves that the Austronesian language first appeared in the same place and at the same time. As you say later:
"Language shifts do occur just like gene flow, and in this part of the world correlation between languages and genes is weak".
Exactly. There is no such thing as an 'Austronesian haplogroup', although if you're prepared to accept the language derives from Taiwan it is very easy to see which male haplogroup it was originally associated with.
"To most people the evidence is overwhelming that the language emerged from Taiwan."
You are making a logical mistake by confusing opinions and data. I'm looking at the data. Opinions, even if they are those of a majority, are of secondary importance to me. There are some dissenters, among linguists, such as Ilya Peiros who I quoted. Malayo-Polynesian languages are not attested in Taiwan, hence, if another Austronesian language appears somewhere else, and this could be Ongan, then Taiwan becomes one of the destinations for the Austronesian dispersal and not its ultimate origin.
""the fact that Ongan is 100% Y-DNA Dx and some Polynesian groups are almost 100% C2x".
Argues against an ancient connection..."
How can it argue against it? It's consistent with it, if not directly suggestive. There are populations out there, such as the Ainu, who have both C and D lineages. Also, compare mtDNA Q lineages in Melanesians and Polynesians (again, not admixture IMO) and Andamanese M31, M32.
"I'd put that population well north of SE Asia..."
Okay, it could be further up north, but why do you think that those carriers of D and C lineages weren't Austronesians? Again, C2 is uniquely Austronesian and non-Austronesians got it from them, and not the other way around.
"I'd include Austroasiatic, along with Sino-Tibetan, Ket, Yenesian and Na-Dene."
Austroasiatic is interesting indeed, especially Nicobarese, right next to the Andamans. Nicobarese is the core of the Austric proposal (see most recently Lawrence Reid) because it shows a set of unique grammatical matches with Austronesian (equally Malayo-Polynesian and Formosan). I can support it with my kinship data, too. So, we may have a cluster of very divergent subgroups in western SEA: Ongan, Nicobarese-Austroasiatic and Malayo-Polynesian. A group of them went to colonize Taiwan, maybe while it was still part of mainland, as boating terminology is not well-attested in Taiwanese languages.
""Then, in Holocene, Austronesians expanded in several directions".
I think everyone agrees with that."
But this doesn't mean that Austronesian didn't exist before. Not as the second largest family in the world but as a dialect chain with innovations spreading across it without much population movement.
"I'm sure it has been you who has been saying no pre-Austronesian languages exist in the region."
I've never said that there are no non-Austronesian languages in south Wallacea or in island Melanesia. What I was saying is that, if we find an old Pleistocene lineage in Austronesian, we shouldn't automatically attribute it to an extinct pre-Austronesian population. There are living non-Austronesian populations and the gene flow seems to have been going on between Austronesians and non-Austronesians since the Pleistocene. Even if we assume that Austronesians picked up some lineages (such as Y-DNA K) from non-Austronesians, it should've taken them more than 3,000 years to dilute their "Asian"/Mongoloid male ancestry by 90%.
What I believe Y-DNA C2 lineage in Wallacean, Melanesian and Polynesian Austronesians shows is that Austronesians entered this broad area rather early on. They expanded to Polynesia late, but otherwise they've been present east of the Wallace line since the Pleistocene.
"C is by no means present in all Austronesian-speaking populations. And of course C2 derives from a C* of some sort, presumably in the Pleistocene and most likely in Southern Wallacea. But that in no way proves that the Austronesian language first appeared in the same place and at the same time."
No, but all the data shows that C*/C2 is an Austronesian lineage. And, as you remember, there are other Cs in the Philippines (across Negrito and non-Negrito populations).
"There is no such thing as an 'Austronesian haplogroup'..."
The aboriginal Taiwanese don't have any Pleistocene-level lineages. It means they are more recent populations than Ongan or Malayo-Polynesians, although there may have been "Asian"/Mongoloid gene flow from Taiwan down south in the Holocene. Linguists just should take note. Or, Y-DNA is not a Pleistocene lineage.
"boating terminology is not well-attested in Taiwanese languages".
And that is very interesting, although could be explained if just one group of Austronesian-speaking Taiwanese had the technology and were able to move to the Philippines.
"The aboriginal Taiwanese don't have any Pleistocene-level lineages".
Probably means there was no-one there in the Pleistocene.
"They expanded to Polynesia late, but otherwise they've been present east of the Wallace line since the Pleistocene".
The evidence seems pretty convincing that no-one was able to move beyond a few of the northern Solomon islands until about 3-4000 years ago. So Austronesian certainly wasn't present too far east of Wallace's line before then either. Non-Austronesian languages are spoken in the few parts of Melanesia not occupied until that time, again arguing against a long-term Austronesian presence anywhere near New Guinea or Melanesia.
"But this doesn't mean that Austronesian didn't exist before. Not as the second largest family in the world but as a dialect chain with innovations spreading across it without much population movement".
It would be very difficult to make the claim that the Austronesian spread goes back as far as the Pleistocene. Sure, Austronesian must derive from some language spoken during the Pleistocene. But the fact we can see a relationship between the various Austronesian languages indicates its spread is much more recent, even if we allow the possibility that Taiwanese languages are a separate cline from the Malayo-Polynesian group. You may like to claim Austronesian is a very slowly-evolving language but that claim seems unlikely. Other Pleistocene languages in the region, especially those of Australia and New Guinea, are extremely diverse. So much so that it is impossible to see any relationship between most of the language groups. So how would a language isolated on various islands maintain a common thread through the whole of the period since the Pleistocene?
"Probably means there was no-one there in the Pleistocene."
Exactly. Taiwan was colonized by Austronesians in the Holocene. No doubt about it.
"So Austronesian certainly wasn't present too far east of Wallace's line before then either."
Apparently, not too far.
"But the fact we can see a relationship between the various Austronesian languages indicates its spread is much more recent, even if we allow the possibility that Taiwanese languages are a separate cline from the Malayo-Polynesian group. You may like to claim Austronesian is a very slowly-evolving language but that claim seems unlikely. Other Pleistocene languages in the region, especially those of Australia and New Guinea, are extremely diverse. So much so that it is impossible to see any relationship between most of the language groups. So how would a language isolated on various islands maintain a common thread through the whole of the period since the Pleistocene?"
Linguists working on Austric note that simple morphology and prosody (Austronesian and Austroasiatic languages are simple morphologically, unlike Indo-European with its elaborate suffixation and suprasegmental features) may be conducive to the longer preservation of genealogical relatedness between languages. This is one possibility. Another factor to keep in mind is that, while linguistic diversity does confer a certain antiquity to the region where it's found, there are cases such as Ainu, when there's only one language, hence diversity is close to 0, but there are all the reasons to believe that it's been around since the Pleistocene. This is because Ainu, a small population, evolved through several stages during which innovations were redistributed across all existing dialects. So, my interpretation of the combined genetic and linguistic data from the Pacific and SEAsia involves three stages in the evolution of Austronesian languages: 1) proto-Austroasiatic-Ongan-Austronesian unity somewhere in western (South)East Asia; 2) it's breakup, with a small Austronesian dialect chain establishing itself in (approximately) Wallacea, with subsequent innovations spreading by diffusion without any population movement; 3) expansion north, west and east from there in the early Holocene, with parallel gene flow from Asia back south in mid-to-late Holocene.
Post a Comment