May 12, 2008...2:07 pm
The Genius of Kinship: Human Kinship Systems and the Search for Human Origins
Thank you, Kambiz, for letting me introduce my new book to the Anthropology.net community.
The story behind The Genius of Kinship is an interesting one. In 1991, then a student of history at the St. Petersburg State University, I wrote a course paper on the traditional social organization of the Shoshone Indians as could be gleaned from ethnographies and trappers’ accounts. Why would a Russian student be interested in the Shoshone Indians is an entirely different story to be told on a different occasion. Let’s just say I was researching Shoshone Indians because they were not widely known in the Russian ethnological literature. My advisors apparently noticed my interest in pre-industrial social structures, and recommended that I explored Shoshone Indian kinship structure in greater detail next year. I poured over literature on kinship studies in Russian, French and English for a few months and then looked at Shoshone kinship again. I was struck by their logical consistency and by the fact that this elegant simplicity was not mentioned anywhere in the basic literature on kinship. Typical case studies came from Australia, Southeast Asia, Oceania, Sino-Tibetan languages, but not from North America.
I thought that was puzzling: kinship studies, as we all know, were founded in the mid-19th century by the American lawyer, Lewis Henry Morgan, on the basis on Iroquois and other North American Indian tribes/nations. The birth of kinship studies coincided with the birth of anthropology as a romantic quest for the origins of Western civilziation. But by the end of the 20th century American Indian kinship structures are nowhere that prominent. Possessed by a pioneer’s zeal, I ventured into kinship terminologies around the world and initially amassed a database of over a thousand kinship nomenclatures from many linguistic families. In 1997, I defended my research as a Ph.D. dissertation at the Peter the Great Museum of Anthropology and Ethnology in St.Petersburg, and in 2001 I published it as a book entitled The Phenomenon of Kinship. Without a particular premeditation, I followed in Morgan’s early footsteps when he wrote The Systems of Consanguinity and Affinity of the Human Family (1870) and conceived of kinship terminologies as a source of information about ancient human population dispersals. (Morgan as a famous social evolutionist emerged with the publication of Ancient Society in 1877 when he attempted to explain the diversity of human kinship structures as a matter of stages in the progressive maturation of humankind.) Over and over again, I caught myself thinking that American Indian kinship structures are unique and can provide a missing link for the evolution of Old World kinship structures.
When I came to the U.S. in 1997 as a Ph.D. student in Anthropology at Stanford, I faced another puzzling irony of history: it’s not just American Indian kinship structures that have been eclipsed from the anthropological agenda, American anthropologists were not doing kinship studies at all. As Sylvia Yanagisako said upon learning about my Russian research, “But nobody does this stuff here anymore.” Truth be told, she herself was part of a “revival” of kinship studies in the U.S. in the late 1980s but more along the lines of gender, with “kinship” being scowled at as a spurious Victorian invention. For some inexplicable reason, she was sceptical of kinship systems, structures, lineages, and especially terminologies. Speaking to other feminist professors at Stanford’s Department of Anthropology (later Department of Cultural and Social Anthropology) such as Jane Collier and Carol Delaney, I couldn’t figure out where all the good old kinship studies went. Where were Lowie, Kroeber, Radcliffe-Brown, Malinowski, Fortes, Levi-Strauss, Dole, Murdock, Tax, Scheffler, Lounsbury, Dumont, Allen, Barnes, Trautmann, Tyler, Kronenfeld, componential analysis, generative analysis, equivalence-rule analysis and other proud representatives of the anthropological tribe? Yanagisako, Delaney and Collier all referred me back to David Schneider who allegedly “proved” that “kinship” was a malignant excresecence on the body of the discipline manifesting all the imaginable vices from racism and colonialism to the masculine bias.
It didn’t make much sense: coming out of a former Soviet country with all its anti-racism and anti-colonialism and anti-capitalism-with-its-severe-exploitation-of-women-and-children, I still felt okay about kinship. Of course, it’s a tough field, not for everyone, but anthropology and kinship are inseparable. You can critisize, develop new theories, change paradigms, but still groom the central concept of the discipline. That’s how I felt.
Across the Main Quad at Stanford, another group of American anthropologists was setting up a different anthropology department called “Anthropological Sciences.” Jim Fox was teaching anthropological linguistics, Joanna Mountain population genetics, Merritt Ruhlen Greenberg’s multilateral comparison, Bill Durham general evolution. There was no kinship studies either, but at least Tom Trautmann once came in with a talk, Hill Gates asked me about the Russian kinship theorist, Mikhail Kryukov, and Joanna Mountain heard about African “segmentary lineages.” Needless to say, the Anthropological Sciences people were very much into out-of-Africa theory of human evolution. Correspondingly, they were supporters of Clovis-I in the Americas. I took classes with Joanna Mountain and worked in her genetics lab. She was a student of Luca Cavalli-Sforza. I also heard wonderful presentations from Richard Klein on African fossils, Peter Underhill on Y chromosome, Marcus Feldman and Lev Zhivotovsky (a Russian geneticist from Moscow) on autosomal markers, and Joe Greenberg on the peopling of the Americas. When Greenberg passed away in 2001, Christie Turner flew in from Arizona for the memorial conference, only to reiterate the “consensus” between his odontology and Greenberg’s linguistics as a rock-solid proof of a recent origin of American Indians. (By 2005, Matsumura and Hudson in “Dental Perspectives on the Population History in Souteast Asia” //
American Journal of Physical Anthropology 127 deconstructed Turner’s celebrated Sundadonty category as a result of relatively recent admixture, thus depriving his general theories of much of their power.) In 1997-1998, Tom Dillehay’s Monte Verde was coming into spotlight, and John Rick grudgingly accepted Dillehay’s dates, with a caveat that “Tom probably mixed up the strata” but now it’s too late to disprove his Monte Verde tome.
So, I was caught in a cross-fire: on the one hand, feminists and post-structuralists “proved” that kinship studies was the unfortunate invention of the confused Cro-Magnon male; on the other hand, archaeology and genetics from across the Quad “proved” that humans came from Africa some 50,000 BP and peopled the Americas no earlier than 11,500 BP (okay, 12,500 BP but Dillehay must have confused the layers). In 1986, Greenberg tried to endorse the latter view linguistically with his tripartite division of American Indian languages into Amerind, Na-Dene and Eskimo-Aleut. (This classification was eventually rejected by the actual specialists in American Indian languages, and with it went down the linguistic counterpart of the out-of-Africa model.) But there was a gap between molecular genetics and linguistics, namely demography, social structure, marriage practices, residence patterns and kinship terminologies. Only this sociocultural bit of evidence could imbue our human origins story with necessary realism. I thought kinship studies could definitely furnish this missing link, and whether human kinship is about “biology” or about “culture” was an utterly secondary matter. I was contemplating terms like “idenetics” and “gignetics” (from Greek gigno ‘to give birth’, a cognate of gen-) to dub this mature state of kinship studies in the 21st century.
Human origins and dispersals research in the 1980s-1990s was driven by “physical” disiplines, those being archaeology/paleontology and genetics. Sociocultural data, meaning linguistics, kinship systems, mythology, were lagging behind. As late as 1980, Robert Austerlitz published a paper in a highly-specialized linguistics periodical called Ural-Altaische Jahrbucher calling attention to the fact that American Indians harbor much more linguistic diversity than the Old World. In 1992, Johanna Nichols published Linguistic Diversity in Space in Time and concluded that our perspective on early human languages comes from America and Australia/Oceania and not from Africa and Europe. Sociocultural anthropology was supposed to contribute kinship studies to the growing interdisciplinary effort, but it forsook it for the sake of abstract ethical polemics. This is all the more bizarre and unfortunate, since anthropology had been working with worldwide databases of kinship terminologies and forms of social organization long before blood groups were discovered, never mind mtDNA sequenced. (In 1967, Murdock sampled 800 kinship terminologies for the patterns of sibling nomenclature that gave a pretty good overall resolution of what patterns are found on what continents.) And the futility of post-modernist moralizing was vividly manifested “across the Quad,” where any reflexivity was tantamount to heresy. Both versions of anthropology looked equally sterile to me.
As a true patriot of Stanford’s Anthropology, I escaped the split of the department into Cultural and Social Anthropology and Anthropological Sciences by migrating for two quarters to the University of Chicago. I listened to terrific Terry Turner on the Kayapo, the late Kostas Kazazis on Indo-European historical linguistics and Balkan dialectology, flamboyant Michael Silverstein on evidentiality, and cowboy-hatted Ray Fogelson (once Yanagisako’s professor at the University of Washington) on American Indian studies and psychological anthropology. I saw legendary Marshall Sahlins from behind and narrowly missed Eric Hamp and Paul Friedrich. And lo and behold, Tom Dillehay himself was there teaching “Andean Prehistory” for half-a-year from the University of Kentucky. For his class, I wrote a paper suggesting that a crucial piece is missing from our human origins research (namely, kinship systems and related linguistic typology), that the out-of-Africa theory is quick and premature, and that if we look closely at American archaeology we won’t be able to see the exact process by which an adventurous group of Siberian hunters colonized the Americas. Symptomatically enough, Cavalli-Sforza assumed a relatively recent entry into the Americas in order to substantiate his claims of an African origin of modern humans; genetically, American Indians and Africans are polar opposites, hence, if we know that America was peopled late, then we can be sure that Africa was the cradle. Looking at Pleistocene archaeology with a critical eye leads one to believe that the data can’t exactly justify using America as an inverted yardstick for Africa. Clovis tools are found everywhere in the Americas but not in Siberia; microblades are found everywhere in Siberia from 20,000 BP but in America they never penetrated south of the Vancouver Island.
Dillehay gave me an A-, and put me in touch with a fellow out of California by the name of Alvah (Pardner) Hicks who’s been trying for a good decade to convince people to look at America as a possible homeland of modern humans. Hicks was in the midst of the 1990s hoopla around the peopling of the America: he attended conferences, dated skulls in South America, corresponded with Tad Schurr, Dave Meltzer and Lou Binford, buttonholed Lyle Campbell and Emoke Szathmary and summarized a myriad of human origins-related scholarly articles for the Mother Tongue readership (kinda doing blogging before blogging became popular). He tried his best to at least make people consider the possibility that American Indians could have migrated into Siberia at the end of the Ice Age. (Franz Boas talked about it a hundred years ago after the Jesup expedition.) But scholars simply refused to listen to Hicks: he brought to the table wacky ideas and he didn’t have a Ph.D. Nevertheless, Dillehay put me in touch with Hicks probably because I had one extra Ph.D. to give away.
When I returned to Stanford in 1999, I rushed to the library to update my kinship terminological database. The Russian library resources are no match for the Stanford ones, and the several years I spent at Green library comparing kinship terminologies from America, Africa, Oceania, Australia and Eurasia were totally worth it. I dug into obscure Brazilian Ph.D. theses, old French dictionaries of rare Austroasiatic languages and Joe Greenberg’s own collection of African language studies. In the end, I assembled a database of some 2500 languages, diligently assembled a comprehensive bibliography and screened this sample for a bunch of typological markers, such as self-reciprocal terminology, “Crow-Omaha,” sibling nomenclature, formal morphology, etc. Some of these typological/polysemic markers were well-known in the literature, others I had to describe anew. When the ordeal was over, I realized that my initial findings were reinforced and bolstered. American Indian kinship terminologies are archaic, while African kinsip terminologies are transformed.
This is the central thesis of The Genius of Kinship. But it’s not the only one. I introduce the reader into the history of kinship studies within and outside of anthropology, compare the ideas about kinship held by Morgan, Darwin and Lyell and conclude (very much in a post-structuralist vein) that our 19th century ideas about human kinship influenced our ideas about human origins. We looked into archaeology and paleontology for answers to the questions of where we came from, while outright dismissing evidence from living human populations (language, kinship, folklore). We were interested in what was left behind in a garbage pit, not in what was passed down to the next generation. We sought our origins in Neandertals, while letting American Indians pass into oblivion. We bypassed American Indian linguistic diversity and grammatical uniqueness and declared them a “recent” population in virtue of the fact that no fossil hominids were ever found in the Americas.
The Genius of Kinship is not meant to be an advocacy for an out-of-America theory of human origins and dispersals. This is a gigantic task. Rather, it’s a revival of kinship studies in anthropology in conjunction with the recent advances in linguistics, psychology, sociology and historiography, a nitty-gritty typological analysis of a large sample of kin terminologies and its application to the prehistory of such accepted language families as Na-Dene, Austronesian and others. One of the results of this revival is a suspicion that the 150 years of finagling with anthropological knowledge (are Indians savages? shall we continue with kinship or shall we switch to gender? only archaeology can furnish reliable data, let’s look at Neandertals, languages are too difficult to comprehend, while we need something tangible to look at, etc.) has resulted in a confused picture of human origins and dispersals in which fundamental assumptions remain unproven, while every new piece of evidence is either swept under the carpet or instantly reinterpreted to fit the consensus.
For instance, the original mtDNA paper, namely “Radiation of Human Mitochondrial DNA Types Analyzed by Restriction Endonuclease Cleavage Pattern” published by Johnson, Doug Wallace and Cavalli-Sforza in the Journal of Molecular Evolution 19 (1983) clearly showed that American Indians have the highest frequency of the ancestral human mtDNA “morph combination” (That was a restriction-site analysis, but still it provided a foundation for all subsequent research, the only difference being that at some point the tree was flipped around and the Africans were declared the oldest population.) The tree topology in this early paper bears close resemblance to the map of human blood types, with American Indians being preponderantly type O. Now, match Johnson et al. (1983) with Ward et al.’s (1992) intriguing paper on extensive mtDNA diversity among the Nuu-Chah-Nullth exceeding that of African !Kung, and ponder as to why American Indians are widely considered to be a young population. In order to explain away an inconvenient fact, Ward et al. had to resort to an argument that American Indians brought this diversity with them from Siberia. Now that pre-Clovis coprolites attest to the antiquity of mtDNA A and B lineages in North America (see Gilbert et al. Science 320 [5877], 2008), even The Onion is smart enough to ridicule Ward et al.’s logic when it writes: “How can we be sure that some ancient nerd didn’t just carry an already thousand-year-old petrified turd with him when he crossed over the land bridge from Asia?”
Or, take Edward Vajda’s recent discovery of a linguistic connection between an isolated Siberian language, Ket, and Na-Dene in North America. The system of verbal prefixes is better preserved in Na-Dene than in Ket, the Ket sibling terminology is radically transformed from its Na-Dene prototype (The Genius of Kinship, p. 325). Or, the fact that the Kets’ neighbors, the Selkups and the Evenkis have a subtype of the purely American Indian mtDNA haplotype A2 (Tamm et al. Beringian Standstill and the Spread of Native American Founders. PLoS One, September 2007, no. 9). Isn’t it a genetic illustration of Boas’s and Hicks’s “back-migration”? While we find traces of American Indian genes in Siberia, the reverse isn’t true: we haven’t found such a close variant of a Siberian gene in the Americas. Or, if one reads Russian and opens up Vladimir Napolskikh’s dissertation on the Earth-Diver myths in America and Siberia, a simple scheme shows that all the archaic variants of this widely-spread motif are in North America, while all the derived ones are in Siberia.
There’s a puzzling contradiction in our data (if this data is looked at through unbiased interdisciplinary lenses), namely that “physical” disciplines such as archaeology/paleontology find lots of support for out-of-Africa, while “ideational” disciplines such as linguistics and kinship studies have Africa as a secondary spread zone and not a homeland. Which data should we trust? Can we ever hoe to build a robust theory on the basis of archaeology’s always-fragmentary-and-accidental evidence? Or, as Eldridge ad Gould famously claimed, such a theory can only be built on the basis of data coming from living biota? How do we now that the currently popular out-of-Africa interpretation of mtDNA and Y chromosome data is not simply a theoretically possible scenario and and the adaptation of a new system of information to suit the existing archaeological/paleontological consensus but a true description of unique population events? All our population genetic maps show the world at 1492, when Africa was probably genetically most diverse, but by 2008 America is arguably the most diverse continent: what scenario of human evolution would we develop hundreds of years from now provided that we wouldn’t know upfront that America was peopled from Europe, Asia and Africa after 1492? We dismissed America as a “New World” and its inhabitants as an Asian offshoot back in the 16th century, which is long before any scientific evidence has been accumulated, but can we reject the possibility that our modern archaeological, genetic, linguistic and ethnological data is consistent with two opposite “single-origin” scenarios? If so, can we rationally adjudicate between the two scenarios without demolishing one as “wacky” and then using the rubble to lionize the other?
113 Comments
May 12, 2008 at 6:12 pm
“While we find traces of American Indian genes in Siberia, the reverse isn’t true…”
“How do we now that the currently popular out-of-Africa interpretation of mtDNA and Y chromosome data is not simply a theoretically possible scenario and and the adaptation of a new system of information to suit the existing archaeological/paleontological consensus but a true description of unique population events?”
Do you really understand that all Native American haplogroups (both Y-DNA and mtDNA) are subclades of Asian ones? If there is Q in America, in Eurasia there is not just Q but also it’s upstream relatives: P and its other derivates (R), K and its other derivates (K1, K2, K3, K4, L, M, NO, PxQ), F and its other derivates (F1, F2, G, H, IJ), etc. If you go upstream by the tree you need to look at Africa, of course. The same applies for C3 and the mtDNA haplogroups.
Eurasian genetics are a subset of African ones, and Native American genetics are a subset of Eurasian ones. Claiming the opposite makes absolutely no sense, sorry. This does not contradict nor is contradicted by back-migration, either from Eurasia to Africa or from Beringia to mainland Siberia.
I think you really fail to understand the basics of haploid genetics: that SNPs are so rare that with all likehood only happened once in all Human history (prehistory included, of course). That Native Americans have the same upstream SNPs as Eurasians and Africans, that their whole haploid genome is just a subset of those.
It’s just primary school concepts:
1. The American set is a subset of the Eurasian one
2. The Eurasian set is a subset of the African one
The opposite is not true.
“There’s a puzzling contradiction in our data (if this data is looked at through unbiased interdisciplinary lenses), namely that “physical” disciplines such as archaeology/paleontology find lots of support for out-of-Africa, while “ideational” disciplines such as linguistics and kinship studies have Africa as a secondary spread zone and not a homeland. Which data should we trust?”
For me it is very clear. Kinship is cultural, like language or religion. Bones and codons are real stuff, not subject to easy manipulation by the always creative human mind.
Anyhow, more complex kinship structures do not need to mean older ones, the same that more complex architecture does not mean older buildings. Normally it is the opposite in fact. I really can’t believe that someone is challenging the quite well estabilished OOA model only on the grounds of theories of kinship.
May 12, 2008 at 8:22 pm
Luis,
Thank you for you comment. Please revisit Table 3 and Fig. 6 in Johnson et al. 1983 paper (attached to my post) and you will see that the original mtDNA tree was rooted in the restriction site combination found at 100% of American Indians with the decreasing frequency thereof among Asians, Europeans and Africans. The tree topology is very similar to the well-known blood type map.
You’re making a common logical mistake: you assume something to be true (because bones and codons can’t lie - period) and then dismiss everything that patently contradicts it as a product of a creative mind. An ideal theoretical model fits all kinds of evidence, no matter what their provenance is. Example: genetically Basques are similar to other Europeans, while linguistically they are very distinct. In this case, linguistics preserves the traces of an ancient population process better than genetics because genetics is subject to forces that linguistics is absolved of (and vice versa). The same with the Ket language: genetically they are like other Siberians, but linguistically they are distinct and, now we know, related to Na-Dene.
The Genius of Kinship is not an advocacy for the out-of-America model of human dispersals. Hence, I wouldn’t try to shoot it down just because it contradicts your beliefs about out-of-Africa. It’s a presentation of a systematic source of information located between linguistics and population genetics that contains an invaluable account of human prehistory. It needs to be taken into account on its own terms. After this is done, meaning The Genius of Kinship is read and digested, a single-origin alternative to OOA has to be tested against all the data (archaeology, genetics, kinship studies, linguistics). So far OOA was tested only against Multiregional Evolution, and out-of-America shares all its advantages over the latter by virtue of the fact it’s another single-origin model.
This is your school exercise in subset reasoning: OOA won over Multiregional in the single-origin category. Now it has to compete with out-of-America in the exact-continent-of-origin category.
May 13, 2008 at 3:46 am
As I was thinking more about Luis’s comment, I figured we’re dealing with two different demographic/population scenarios here. The Out-of-Africa theory assumes that the African population has been steadily accumulating size and allele/gene diversity over a long period of time, spitting out on a number of occasions daughter populations that colonized the globe. These daughter populations went through a bottleneck, lost the original African genes and developed new ones along the way. They’ve never recovered from the bottleneck and never exceeded Africans in gene diversity. At the next junction between Asia and America, this situation repeated itself: American Indians have never recovered from a bottleneck and they lost all the most frequent Asian genes (specifically, those belonging to macrohaplogroup M). These two bottlenecks that progressively obliterate the most frequent parental genes from a daughter population seem to be rather odd.
Basing on Native American population realities (Neal and Ward and other old-school geneticists wrote extensively about it between 1970s and 1990s), an alternative demographic/population scenario would assume that the parental population has remained in the original state of dispersed demes charcaterized by high mobility, frequent loss of lineages and low effective and demographic population size. This is consistent with America having the largest linguistic diversity in the world, the latter reflecting this highly fragmented population. As these small demes exited America, they expanded in size in response to the demographic pressure of colonizing the globe and achieved a high level of gene flow and lineage retention. As America was geographically isolated, there was no gene flow between the New World and the Old World. Africa ended up in the long run the most demographically dominant. America has preserved the ancient demographic condition, and not suprisingly Cavalli-Sforza advised other geneticists to use Brazilian tribes as models of original human population structure. In one of such papers built on Cavalli-Sforza’s method (Zhivotovsky, Estimating Divergence Time with the Use of Microsatellite Genetics Distances: Impacts of Population Growth and Gene Flow. Molecular Biology and Evolution 18 (5), 2001), South American Indians are considered “a reference for microsatellite variation in an ancient African ancestor because their population size is low and might be compared with that estimated for an African ancestor.”
Consequently, the nesting structure of the human genetic trees referred to by Luis, in which Asians are subsets of Africans, and American Indians are subsets of Asians, represents the result of a long-term evolution but doesn’t accurately reflect the actual process. If we assume the OOAf model, then the differences in the levels of linguistic diversity between Africa and America will remain aberrant. If we assume the OOAm model, then things look more logical, with both linguistics and genetics falling into their proper places. Kinship structures further corroborate the OOAm scenario, as Old World kinship structures, especially in Australia/Oceania and along the Pacific Rim, less so in Africa and Europe, show residues/survivals of that ancient demographic/population condition which is more fully preserved in American Indian kinship terminologies.
May 13, 2008 at 7:07 am
Genetic and linguistic diversity can seem greater due to greater isolation in different regions. However, this diversity might just involve subsets of populations, for example genetic haplogroups, from elsewhere that migrated into the region.
May 13, 2008 at 1:47 pm
You realize, of course, that New Guinea has the highest level of linguistic diversity in the world. Most likely the result of prolonged isolation on the part of a great many different populations and most certainly NOT because homo sapiens originated on that island.
There is no possible interpretation of kinship systems per se that can tell you where the concept started, just as there is no possible interpretation of language families per se that can tell you where language originated. I’ve worked for years on the musical evidence and I can tell you that there is no possible interpretation of the various musical families per se that can tell you where music originated.
Since the genetic research does, at least in principle, provide exactly that sort of evidence, kinship studies, historical linguistics and comparative musicology need to follow the lead of population genetics, not the other way ’round. As far as I’ve been able to determine, the musical evidence I’ve studied is in fact consistent with OOA. So, apparently, is the linguistic evidence.
If you find the kinship evidence to be inconsistent with OOA, then you can certainly feel free to contest it’s validity on that basis, rather than attempt to create a whole new theory of human origins on the basis of kinship alone. It can’t be done.
Frankly I’m disappointed, because I thought you were about to argue for the importance of kinship studies in the context of the new anthropological paradigm, based on the new genetic research. You could make a very strong point, because so much of the genetic research depends on an understanding of kinship, not only the different concepts and vocabularies, but also the different practices, based on kinship, by which different peoples select appropriate mates. Many of the differences between mtDNA and Y results can be explained on such a basis, yet kinship studies are rarely considered in most pop. genetics studies I’ve seen.
You could make an important contribution if you were willing to accept a more modest role. Which would not prevent you from contesting OOA on the basis of your kinship research.
Most meaningful research is based on a process or set of processes rather than a lightbulb going off over someone’s head.
May 13, 2008 at 2:35 pm
Thanks Victor.
1. You argue too opposite things at the same time: kinship studies are rarely, if ever, considered in population genetics research and genetic research hinges on a certain understanding of kinship, but kinship studies can only follow population genetics in the questions of prehistoric human dispersals. What’s the point of doing kinship studies if population genetics has already provided us with a correct picture? You dismiss 150 years of kinship research for the sake of a recent mtDNA and Y chromosome paradigm. This is not prudent. I believe, on the contrary, that every system of information, if properly analyzed, yields unique results. Then you compare across those unique results to weed out unwarranted assumptions. Kinship systems and terminologies don’t contradict genetic evidence, but they do contradict a particular interpretation of this evidence, namely the OOAf model of human dispersals. OOAf is based on certain assumptions about demography and Old vs. New World antiquity that haven’t been tested. Love it or leave it.
2. Papua New Guinea is not “more” diverse linguistically than the Americas. They are comparable and it depends how you look at it. If you judge diversity by the size of the geographic area, yes, but, then, for instance, America harbors all 6 possible types of word order; the same for kinship terminologies. American Indian linguistic diversity is associated with pretty archaic kinship structures, which leads me to believe that it may be the function of age, and not of linguistic evolution run amok in the last 10,000 years. But overall you are of course right: both America and Papua New Guinea (but not Africa!), are the examples of linguistic differentiation consistent with an ancient human demographic/population model. Outside of the Americas, it’s Australia, Oceania and the Pacific Rim in general that show this pattern in languages and kinship terminologies.
3. Kinship systems and terminologies have been sliced by scholars for at least 150 years. The available databases are considerable. I did additional typing of new variants and focused on those paterns that show lineal transformation over time, but overall my analysis stands on the shoulders of many generations of scholars. If we ever develop the same level of granularity in our musicological research, I would love to see the results. Folklore is problematic for many reasons, but sometimes transcontinetal similarities are striking. Every system of information requires its own training and its own patience, but methodologically truth is the common denominator behind multiple sources of data. Archaeology and genetics cannot furnish a complete picture without forcing us to apriori dismiss other disciplines as incoherent.
4. As for modesty, I think honesty is more humble: if your evidence doesn’t fit the consensus, stay with the evidence but be able to argue rationally and rescind statements that are being falsified by new evidence. I don’t feel compelled to rescind anything so far. And my book is full of rather modest applications to low-level language families.
May 13, 2008 at 4:36 pm
“You argue too opposite things at the same time: kinship studies are rarely, if ever, considered in population genetics research and genetic research hinges on a certain understanding of kinship, but kinship studies can only follow population genetics in the questions of prehistoric human dispersals.”
Clearly both fields need one another. Which is why I’d rather see someone like you collaborating with geneticists instead of spending all your time trying to convince everyone you’re not a crackpot, which is probably going to be your destiny for the next few years. The problem is that, for all the challenges pop. genetics still faces, theirs is the only field that holds any real promise when it comes to tracing human origins, because they are studying lineages directly — not indirectly via the various ways in which they are represented. Certain problems involving the relation of mtDNA and Y will probably require some help from kinship studies, for sure. But even without that help, they have compiled a mountain (no pun intended) of evidence strongly suggesting that they are on the right track.
Any theory you come up with will ultimately have to be tested against their results, not vice-versa. So why not work with them rather than fight them?
I’m not saying you are necessarily on the wrong track, just that you are probably going about things the wrong way.
>What’s the point of doing kinship studies if population genetics has already provided us with a correct picture?
It’s not necessarily correct. If you are able to convincingly demonstrate that your kinship research is inconsistent with their results, then they will be forced to pay attention to you. But if you simply dismiss their results and insist that your kinship research is all that is needed, then they’ll simply ignore you, along with everyone else.
May 13, 2008 at 5:00 pm
Victor, thanks again for your thoughts. You are right in principle. A few clarifications. Geneticists know that their results don’t sit well with American Indian linguistics. Their temporary alliance with Greenberg fell through because he was dismissed by all other linguists. Professional linguists don’t read what Cavalli-Sforza writes about a putative isomorphism between his gene trees and language trees because he uses unproven megaphyla. Then, geneticists are different. There’re genetic labs, such as David G. Smith’s at UC Davism, that work on small piecemeal research that I have a great interest in and a great respect for. For instance, I can see the same specificity and antiquity in Saami kinship terminology as the geneticists see in their mtDNA genes. I just think that OOAf is a premature solution that is based on unproven assumptions. Only a coalition of disciplines representing biology, society and culture can decide where humans came from, not one single discipline.
Collaborating with linguists is pretty easy for me. However, collaboration across such vast fields is still problematic because people are up to their ears in their own stuff with their own deadlines and deliverables.
I slightly regret that people snatch on my out-of-America ideas instead of thinking with me about what constitutes my real focus, namely human kinship systems, interdisciplinary alliances, the nature of anthropology as a discipline, and historical reconstructions without exotic solutions. But I look like a crackpot only because we’ve invested too much belief in such theories as OOAf or Clovis-I. We’re glued to them as if they were a Bible. In reality they impede our real progress. For me OOAm is a valid hypothesis to undertake, a strategic possibility, a good way to look at the data, and OOAm and OOAf are good to play against each other to make sure we don’t create false beliefs that will take hundreds of years to dismantle. But neither should become a matter of belief transcending logic and facts.
May 13, 2008 at 7:43 pm
One part of my problem with your approach is that I’m skeptical when you claim the ability to determine the antiquity of a kinship system — and by implication the antiquity of the culture as a whole — from its kinship terminology per se.
I’m especially sensitive to such claims because for a long time musicologists assumed that they could do the same for musical styles. Melodies with one or two notes, or “tumbling strains” regarded as basically “pathogenic,” were assumed to be archaic precursors of more “advanced” musical practices, involving more complex melodic types, leading ultimately to the most complex and sophisticated music of all: contrapuntal polyphony — which was generally assumed to have been developed only in Medieval European high culture.
As more and more field studies and recordings became available it became increasingly clear that this couldn’t possibly be the case, because in fact a great many indigenous hunter/gatherer groups in various parts of the world sing and play together in polyphonic styles that are indeed contrapuntal and in fact remarkably sophisticated in many ways. Others on the other hand, have much simpler musical styles with no polyphony.
If we were to consider only the music in itself, therefore, there would be no way to determine which style was truly archaic, and thus representative of the oldest form, because the most “archaic” societies sing in so many different styles.
It was only when I began studying the genetic research that a pattern began to emerge that made sense and enabled me to come up with a hypothesis, which, if not necessary the correct one, is at least coherent — and testable.
I’m very curious, nonetheless. And since I can’t afford to buy your book, I’m wondering if you could tell us what features of a kinship system are, according to your theory, consistent with the greatest antiquity.
May 13, 2008 at 9:10 pm
My two cents worth:
German Dziebel wrote, “we’ve invested too much belief in such theories as OOAf or Clovis-I. We’re glued to them as if they were a Bible”. I guess most people here now know I believe the beliefs actually come from the Bible. The OOAf theory allows us to still believe species come from just one couple (or virtually so). It’s easily possible to interpret all the evidence as showing that various human subspecies have been interbreeding since H. erectus times, but that’s another argument.
Both German and Victor bring up the comparison between New Guinea and America. What about the Caucasus Mountains? We consistently find the greatest genetic and linguistic diversity in mountainous or heavily forested regions. The reason is obvious if you think about it. Less intergroup contact. Africa has a largely more homogeneous habitat. Human groups have been able to expand greatly, obliterating earlier diversity. I suggest this accounts for the apparent less diversity in Africa than in either New Guinea/Southeast Asia or America.
May 14, 2008 at 2:05 am
Just briefly:
“These daughter populations went through a bottleneck, lost the original African genes and developed new ones along the way.”
They lost nothing, sorry. The original mutations leading to the nearest African branch of the Eurasian group are still there. Every single Eurasian male has all SNPs between macrohaplogroup CT (CR) and BT (CR) (and Y-DNA “Adam”, the root of the tree) too. Africans instead are more variegated, as they can belong to macrohaplo A, B or CT (specially E).
This clearly indicates that Eurasians are a subset of Africans. It’s impossible that it’s the other way around.
The same applies to mtDNA. And certainly it is the same with Native Americans, who are a subset of Eurasians. Native Americans are all (Y-DNA-wise) CF (C and F, recently discovered they were once one single lineage by Karafet or Underhill, not sure right now). In fact they are only C3 and Q, subsets of C and F (or K, or P) respectively. They don’t even have other important haplos of Eurasia, like D, IJ, NO, etc. (and some of them are frequent in their vicinity: in East Asia). They just have two Siberian clades - and that’s all (at least for Y-DNA, that is slightly simpler to explain).
So Native Americans are not even a subset of Eurasians, but a subset of Siberians. Would Eurasians be a subset of Americans, then all Eurasian haplos would be downstream of Q and C3 - and that’s not the case at all: the opposite is true instead.
You are founding all your delusion in a single old paper of the times when human genetics was in its baby stage (1983! That’s pre-Cavalli-Sforza!). And holding to it as the proverbial burning nail (not sure if this analogy exists in English but in Spanish it does, implying desperate stubborness).
The same with the Ket language: genetically they are like other Siberians, but linguistically they are distinct and, now we know, related to Na-Dene.
I’m not sure about Ket genetics right now but their “Ostyak” cousins, the Selkups (Samoyedized something-else) are one of the most American-like populations of Siberia, with large ammounts of haplogroup Q. Ket language is the only survivor (100-500 speakers) of the Yenisean language family, in other times widespread in parts of Siberia.
AFAIK , it’s not like they are unrelated genetically either.
So far OOA was tested only against Multiregional Evolution, and out-of-America shares all its advantages over the latter by virtue of the fact it’s another single-origin model.
I think it’s self-evident but kind of proof would you need to be persuaded of OOA? Americans being a subset of Eurasians and these being a subset of Africans seems more than sufficient evidence to me.
I think you are too much into formal kinship and ignoring biological one.
“Basing on Native American population realities (Neal and Ward and other old-school geneticists wrote extensively about it between 1970s and 1990s), an alternative demographic/population scenario would assume that the parental population has remained in the original state of dispersed demes charcaterized by high mobility, frequent loss of lineages and low effective and demographic population size. This is consistent with America having the largest linguistic diversity in the world, the latter reflecting this highly fragmented population. As these small demes exited America, they expanded in size in response to the demographic pressure of colonizing the globe and achieved a high level of gene flow and lineage retention. As America was geographically isolated, there was no gene flow between the New World and the Old World. Africa ended up in the long run the most demographically dominant.”
This is, with all due respect, the weirdest idea I have ever read. Basically you are saying that magically (no apparent reason for that, as America and Africa are comparable in size, and Eurasia is somewhat larger but also colder on average) Americans went through massive bottlenecks and so did Eurasians to less extent, while Africans never experienced any bottleneck at all - not even when crossing Hormuz strait, the deserts of Arabia and Bab-el-Mandeb. Right?
You are saying that a Noah’s Ark of American genetics crossed Beriningia in Western direction and that The Flood (or some other hyper-massive catastrophe) left America virtually depopulated (only way to explain the few surviving lineages, more consistent with the typical Beringia-to-America model). And that the same story (second Noah’s Ark and second Flood) happened between Eurasia and Africa again later on.
Don’t you realize that is totally ilogical? Isn’t it a lot simpler that a small subgroup of Africans crossed Bab-el-Mandeb (or surrounded the Red Sea, but that seems less likely nowadays) and later Hormuz Strait. In that case, only the lineages of the founding fathers and mothers would survive (and later branch out). And that is exactly what we see. And the same regarding America.
Were there minor back-migrations? Certainly. But as they did not find a virgin land, they never became really prominent. And exception to this was thought once to be haplogroup E but nowadays, after DE* has been found in Nigeria, it’s accepted that the E clan remained in Africa all the time, only spreading out in the Mesolithic.
You have to build up a really mythical story to justify your hyothesis. The OOA model is just plainly logical instead.
…
The main point you seem to bring to account is linguistic diversity. But we know that in older times Eurasia was much more diverse than it is now. Really a handful of quite well studied late prehistorical events can account for all that loss of linguistic diversity:
- Afroasiatic expansion (Mesolithic in Africa and later Semitic migrations in lowlands West Asia, the latter documented historically c. 4000-3500 BCE).
- Indo-European expansion (explained well with the Kurgan theory). Parallely, Uralic expansion across NE Europe.
- Chinese Meso-Neolithic expansion southwards, still occurring in historical times.
- Turkic expansion since the time of the Huns or before.
And, in addition, general localized homogenization because of trade and empires.
In sud-Saharan Africa also the loss of linguistic diversity can mostly be attributed to the expansion of the Niger-Kongo branch (Bantu expansion). Anyhow, considering how widespread is Y-DNA E, it’s only logical to think in some linguistic homogeneization related to this dominant (but not exclussive) male ancestry. It does seem like the E clan (and its subclans) spread around in the late Paleolithic and Mesolithic homogenizing somewhat nearly all Africa, but not erasing the diversity nevertheless (macro-clans B and A still exist).
In comparison with Europe and Africa, America was underdeveloped (Neolithic, Chalcolithic at most in some areas), so there were few groups that could become dominant as happened in Eurasia. Additionally we really don’t know how “Amerindian” languages releate to each other - at least not like we know most Eurasian and African ones.
Linguistics is also underdeveloped in this region. Greenberg (first real classificator of African languages) suggested that all non-Inuit non-Na-Dene languages belong to a single superfamily. This is controversial but, if real, it will reduce your “huge linguistic diversity” to just 3 groups. It would certainly be very consistent with the Beringia model, assuming that Na-Denes and Inuits arrived in later waves (from Beringia or otherwise in Siberia anyhow).
But in any case, linguistics cannot be the reference, as it has only a limited time depth and accuracy. Also people can learn new languages somewhat easily but cannot change their genes. We know of languages dissapearing every other day, not by physical extinction of the peoples that spoke them but because of cultural assimilation. We also know looking at history how languages diverge. A Northern Spaniard can have difficulties understanding a Southern one or a Mexican and certainly Hadrian would be unable to understand but a handful of words from either of them. The Chaclolithic proto-Indoeuropean who moved from the steppes into what is now the land of Saxony would certainly not understand a word of the verses of Bertolt Brech (nor any other modern IE language for the case). Languages diverge fast and die easily, they cannot be the foundation of any model of humankind dispersal - specially when we have much better tools available.
May 14, 2008 at 2:06 am
“Just briefly…”
LOL - I tried (initially). Sorry. :-)
May 14, 2008 at 10:19 am
Very good discussions. I won’t be able to comment on all the points made by Victor, Luis and TerryT, but here’re just a few attempts.
To Victor: Any data coming from living populations has chronological uncertainties. This concerns genetics, linguistics and kinship studies. Archaeological and geological evidence are the only ones that can be dated directly. this uncertainty diminishes with the increasing granularity of description but won’t disappear altogether, I think.
Relative chronology is nevertheless important. In linguistics, some very respectable scholars , such as Johanna Nichols or Sergei Starostin, will even say that glottochronology, if words are selected using stringent criteria, can give solid results. I’m very cautious about glottochronology but relative chronology coupled with typological and stock diversity in both kinship studies and linguistics cannot be dismissed. To Luis’s point, some language familes and subfamilies (like Romance) show rapid change, others (like Lithuanian) have been highly conservative. The kinship of Ket with Na-Dene may mean that languages can be very conservative and very stable over time (this is Ed Vajda’s contention as well), unless we prove that there was a direct migration across the Bering Strait in the last 5,000 years. So far we haven’t found any evidence of that, and the end of the Ice Age remains the only closest point in time during which Siberia and North America was in contact. This suggests that certain language families can be very old.
Kinship terminological evolution is established on the basis of the following diachronic universals:
1. “Dravidian” (symmetrical-prescriptive) systems that systematically link affines and consanguines (mother’s brother equals spouse’s father, etc.) change to “non-Dravidian” systems. “Dravidian” systems should better be called “Amazonian” systems, as the students of Amazonian kinship noted that they conform to the ideal type better than the kinship systems of the Dravidian-speakers. Africa is completely lacking Dravidian/Amazonian systems.
2. In the same way as “cross” kin categories are linked to affinal categories, “parallel” kin categories (such as father’s brother and mother’s sister) are linked to adoptive categories (step-father, step-mother). In later kinship systems adoptive categories and parallel consanguineal categories are kept separate.
3. Alternate-generation equations (grandfather equals grandson, mother’s brother equals sister’s son, etc.) disintegrate yielding terms for the polar kin categories. There’s a tremendous amount of diversity within “alternate-generation equivalence” systems, but Australian aborigines and North American Indians show the most complete sets of those equations. Similar forms exist in Munda, Uralic, Papua New Guinean and Dravidian terminologies. Khoisans in Africa come the closest to exhibiting this feature, but they are more derived than Australian and North American systems.
4. Sibling sets fall between two poles: those systems that lexicalize sex-of-speaker, sex-of-relative and relative age (yielding up to 8 terms) vs. those that have only one (e.g. Swahili ndugu ’sibling’) or only two terms (Eng brother/sister; sibling is an artificially constructed term). Out of some 4,000 possible combinations of the 3 parameters, languages consistently lexicalize only a dozen types. These types have language-family and continent-specific distributional regularities and can be linked into a “tree” (Fig. 7 in “The Genius of Kinship”). Africans end up on the “transformed” end (Niger-Congo in the middle), while American Indians, Papua New Guineans and some CircumPacific groups such as Hmong-Mien, Koreans and Ainu end up on the ancestral end of the tree. The linearity of change has been tested on all available evidence (Austronesian sibling set evolution is the one that was meticulously described in the early 1980s by both linguists and anthropologists).
5. Traditional typological markers such as Bifurcate Merging, Generational, Bifurcate Collateral and Lineal remain valid markers. In the 1950s, Dumont and others tied them to the “Dravidian” equations; I further clarified their connection to the other parameters (described above under points 2-4). Bifurcate Collateral is associated with strong Alternate Generation equations. Again, this feature is found in Australia, Papua New Guinea and the Americas but not in Africa.
6. Kinship terms and terminologies display variation on the morphological side as well. Descriptive systems (namely, thos comprised of kinship terms such as “father’s brother”, “mother’s sister’s son” in which the exact genealogical connection is decsribed literally in the surface realization of the term) have long been considered secondary. “Archaic” kinship systems (again, Australia, Papua New Guinea, America) that tend to disregard genealogies don’t have them. Descriptive terms are widely spread in Africa, the Causacus and Europe. In Africa, even siblings are called decsriptively (”mother’s son”, etc.)
7. Pragmatic variation is less of a formal and trackable feature in kin terminologies, but even here there’s a sharp division between Australia, Papua New Guinea and America, on the one hand, and Europe and Africa, on the other. (Nichols, by the way, showed the same continental opposition in grammatical features) For instance, inalienable possession (further linked to the grand division of world languages into head-marking vs. dependent-marking) and vocativity are strongly expressed in the “archaic” zones and almost non-existent in the “transformed” zones.
These seven diachronic universals (1-6 are the most important) yield a myriad of forms and subtypes around the globe. Universals 6-7 are intimately tied to the global language typology (Greenberg, van Driem, Nichols, among others), while universals 1-5 are kinship-specific with the ties to marriage forms, demography, etc. Proto-kinship systems for every language family can be described in terms of these parameters.
Kinship studies isn’t a master key to world secrets, but just an overlooked type of evidence with a strong position between linguistics and population genetics and with a long history of painstaking research behind it and a large global database. And once again no system of evidence is exempt from weaknesses. Kinship studies doesn’t have an objective chronology, it’s subject to selective forces of demography and social organization; archaeology is hopelessly divorced from modern-day populations, and it’s their prehistory that we’re trying to understand; population genetics is subject to demographic parameters and suffers from a lack of objective chronology. Linguistic connections can be confused by borrowing and coinicidences. Folklore is easily borrowed, etc.
To Luis: Hence, no single system of evidence can provide us with an ultimate answer, and your continuing apology for population genetics creates another mythological scenario for human prehistory. No matter how many times you use the word “prove” in your presentation of genetic data, it doesn’t add strength to your argument. All real “proofs” are interdisciplinary. If American Indians are a subset of Siberians, then this should be immediately evident in their kinship systems and in their languages. I can assure you that OOAf doesn’t translate into anything outside of genetics. Even archaeologically there’s no evidence of an “expansion” out-of-Africa. In the cases of America, it would’ve been easy to see because the process would’ve been very very recent, but the truth is there’s no evidence for American Indian kinship and languages being a subset of Siberian kinship and languages. Neither there’s evidence for Asian/Pacific languages being a subset of African languages. If you’re prepared to show how exactly typological and stock language diversity disappeared in Africa, or how American Indian grammatical features appeared from a Siberian linguistic soure, I’d be delighted to see it.
Your perspective forces you to dismiss kinship and languages altogether; mine allows integration of all systems of evidence into one.
My reference to the Johnson et al. 1983 paper (by the way, Cavalli-Sforza is one of the authors) is justified by the fact that its results have never been overturned, that the same tree is used in more recent papers by Alan Templeton, and that any search for alternatives requires revisiting some critical old stuff. The tree offered in Johnson et al. 1983 makes a complete sense from the kinship and linguistics perspectives. What a coincidence!
The fact that OOAf is built on certain demographic assumptions is no secret to anybody. Masatoshi Nei was one of those who wrote about it. Population genealogy, as described in the language of gene markers, and demography are related.
Let’s look at mtDNA: lineages L1, L2 and L3 are African-specific. They aren’t found anywhere outside of Africa, including Australia and Andaman islands, i.e. places that are thought of as possible first stops on the out-of-Africa journey. Haplogroup M is spread everywhere is Asia and Siberia but for some reason it didn’t make it into the Americas? Why would the two bottlenecks eliminate precisely the “oldest” and the most frequent markers in the daughter populations? Why would American Indians pick out rare Asian (A, C and D) and rare European (X) lineages but still pick them from both macrohaplogroups found outside of Africa (M [C and D] and N [B, A and X])? It’s just too much selectivity and sharpshooting on the part of a group of Siberian hunters.
Once again, I have a deep respect for genetics, and hope things will straighten out but we have to juggle between different systems of refernce without committing too much to one single view. We’re dealing with empirical data of radically different provenance, and not with objective experiements, hence we can’t afford an assumption that something transcends interpretation and is true by virtue of simply being a fact.
If OOAf proves to be true, it will be fine with me, just as long as it translates into other disciplines outside of genetics and helps me, as a specilaist on human kinship systems, to make sense of my data. As of now, looking out-of-Africa mangles my data beyond repair.
You of course may retort that my demographic scenario doesn’t make genetic sense. I’ll definitely think more about it, but be advised that all systems of genetic information concur that, on a worldwide scale, low diversity in the Americas is coupled with very high FST values (especially in South America) suggesting that American Indian demographic evolution has been unique for at least 10,000 years. What prevents us to think that it’s been the same for as long as humans existed as a separate species as soon as we can see that this population structure translates nicely into high levels of linguistic diversity and archaic kinship structures?
To Terry: you’re making an excellent point about geography. African geography may have been conducive to greater gene flow and lineage retention in Africa, hence to the greater levels of diversity, as well as to the aberrant proliferation of languages in such areas as the Caucasus, Papua New Guinea or California
May 14, 2008 at 12:37 pm
More on Luis’s comments:
Regarding Greenberg’s classifications, the only reason it worked in Africa is because he based his studies on the earlier work by Westermann’s, who used traditional methods. Still, it’s not without its problems even there (see Roger Blench’s work, including the proposed connection between Nilo-Saharan and Niger-Congo, the recent origin of the Pygmies, etc.). The classification of American Indian languages into Amerind, Na-Dene and Eskimo-Aleut has been rejected once and for all. It can be revived only under the assumption that Amerind is the oldest human language megaphylum, but any unity of American Indian languages beyond the first-order families is hard to illustrate linguistically. In terms of kinship, Eskimo-Aleut, Na-Dene and the rest share unique similarities vs. the Old World families (I believe these similarities are very old), hence Greenberg’s theory is falsifiable on all fronts. See (Weiss-Bolnick, Deborah A., (Schultz) Beth A. Shook, Lyle Campbell, and Ives Goddard. 2004. Problematic Use of Greenberg’s Linguistic Classification of the Americas in Studies of Native American Genetic Variation. American Journal of Human Genetics 75 (3): 519–522) for an example of how inconsistent genetic results are with the language situation in the Americas.
May 14, 2008 at 7:08 pm
Thanks very much, German, for the very interesting explanation of the kinship evidence and your interpretation of it. While I remain skeptical regarding aspects of this interpretation, I have a feeling the database you’ve compiled will prove to be of real importance, regardless of whether or not any of your more ambitious hypotheses eventually pan out.
I find it especially interesting that you’ve been able to attach a kind of historical “topology” to various types of kinship structure, on the basis of what sort of system is most likely to succeed what other system. While I’m wondering how much of this is based on rigorous analysis and how much is simply assumption, I agree that this is an approach well worth exploring.
Assuming, however, that your inferences regarding the difference between archaic and more recent kinship systems are essentially on track, I’m still wondering whether there might be another set of assumptions at work that you may be taking for granted.
Since, as I assume you’d agree, all contemporary societies stem ultimately from the same ancestral root population, all societies must be regarded as essentially of the same “age”. Therefore, the presence of archaic kinship systems among certain groups in one particular part of the world does not necessarily mean this region is necessarily where “modern” humans first developed. It could mean that certain groups may simply be more conservative than others. In other words, there is nothing necessarily archaic about a lineage that has stayed put as opposed to one that has migrated thousands of miles.
Archaic systems of thought, belief, expression, subsistence, etc. can be found in many different parts of the world, among many different types of indigenous peoples. As I see it, that’s more likely to be a measure of their relative isolation (and perhaps their collective conservatism), than their innate “archaic-ness.” Thus, if the Saami happen to have a more archaic kinship system than other peoples living in the same general area, I’d say that’s most likely because they’ve been more isolated and less subject to external influences and pressures. The same could be said for both the New Guinea highlanders and a great many Amerindian groups, especially in S. America. That doesn’t make such groups any “older” than any others. It just means that they have managed to preserve more archaic survivals than their neighbors.
As far as Africa is concerned, the great majority of sub-Saharan African peoples have not been isolated, at least not since the Bantu expansion of ca. 3,000 years ago. So one wouldn’t expect them to have retained the most archaic kinship systems — or language families. Bantu may in fact be among the more recently developed language families. And, like Indo-European, it undoubtedly replaced a great many far more archaic languages. The same is probably true of most Bantu musical styles, which also show signs of more recent provenance. This tells us nothing about whether “modern” humans originally migrated from Africa or not. It is simply a reflection of certain historical contingencies.
May 14, 2008 at 7:11 pm
…some language familes and subfamilies (like Romance) show rapid change, others (like Lithuanian) have been highly conservative.
I agree that language evolution does not necesarily happen at “constant mutation rates” (against what some linguists would like). Probably isolation makes evolution slower and expansion and creolization faster. Still Lithuanian is not PIE and a hypothetical PIE speaker would not understand Lithuanian at all. And it’s been “only” some 5-6000 years, a very short period in comparison with “modern” humankind (estimated some 150,000 years, probably more).
Where does IE stans in the global languages tree some 5,000 years before? We just have no idea. There are two or three of hypothesis and that’s all. We just cannot reconstruct the linguistic reality of the Upper Paleolithic, nor in most cases that of Neolithic times (Eurasian chronolgy). There’s an “event horizon” for linguistics beyond which we cannot see anything - like in a black hole.
The kinship of Ket with Na-Dene may mean that languages can be very conservative and very stable over time (this is Ed Vajda’s contention as well), unless we prove that there was a direct migration across the Bering Strait in the last 5,000 years.
Haven’t read enough on the issue (too recent) but I fancy that the 5,000 years figure is arbitrary. Dene-Yeniseic is a superfamily including two families: Yeniseic and Na-Dane. I know of no age estimates for divergence but I suspect that figure is ultra-conservative and somewhat capricious.
Still Na-Dene peoples might belong to a second Berigian wave that would have arrived after the main group (vaguely grouped as Amerindians). They seem the main group to have the second important Sibero-American Y-DNA clade: C3, what suggests a somewhat differentiated origin in any case. If Amerindians coast-migrated (say) c. 15,000 years ago, Na-Dene could well have arrived some 10,000 years ago (maybe bringing the famous Clovis tech with them). Inuits (the third wave) are a quite recent arrival but they still seem to be original from Beringia, fitting wholly in the Beringian-American restricted genetic genealogy.
Kinship terminological evolution is established on the basis of the following diachronic universals…
I really appreciate your class on kinship systems but I strongly suspect that kinship structures appear and disappear as societies evolve. The only really fundamental kinship terms are mother, son, daughter, brother, sister - and maybe father, if it’s known at all. Then guess that direct or extended relatives of the parental lines may be convenient (though you can also use terms like brother for cousin and father for uncle, etc.), and then you can fancy all kind of terms for all sort of possible kinship structures. But reality tells us that all options (from the simplest to the most elaborate) are equally possible and there is no real clear pattern.
I wonder if what you percieve as oldest types of kinship aren’t but the special (regional) forms that appeared in the circumpacific region in the time of human dispersal and not the truly original ones. For truly original anthropological stuff I always look to Pygmies and (specially) !Kung and these are a lot simpler - as are their societies.
…sibling is an artificially constructed term
Of course: all words are artificially constructed. Languages are always human-made (artificial) after all. There’s nothing humans make that is not artificial. A flintstone burin is not less artificial than a microwave oven , the word “mum” is not less artificial than “telescope”.
Africans end up on the “transformed” end…
I fear you are the one putting the tree upside down most likely. Why on Earth would be such a complex kinship system as that of Papuans be more ancient than the rather simple one of !Kung? I think that the complex system must be created by “intelligent design” (i.e. the human mind in action, trying to configure society according to some pre-concieved ideas), while the simpler system is more natural/spontaneous/intuitive/fresh. I also suspect that kinship systems can wildly change when socio-economical imperatives demand it (and mentality change has already taken palce somewhat).
If I don’t recall badly from the classics (I only know of Morgan via Engels, I admit - but anyhow) transitions from one system of kinship to another were rather easily done by “decree” in a single generation in many cases. This mostly refers to matrilineal to patrilineal but is still suggestive on how these social constructs can change easily.
In brief, while I think kinship studies are interesting, I don’t think you can build much upon them (as you are trying to).
Hence, no single system of evidence can provide us with an ultimate answer, and your continuing apology for population genetics creates another mythological scenario for human prehistory. No matter how many times you use the word “prove” in your presentation of genetic data, it doesn’t add strength to your argument.
I think I did much more than to say “prove” like a parrot. I explained in quite detail how genetics have reconstructed, so far not disproven or even put in question, trees of human parental lineages. These trees have a very obvious root in Africa and a main branch in Asia.
For interdiscplinarity, this is in excellent correlation with archaeology (precursor hominins, earliest H. sapiens remains, etc.) and biology (closest living relatives: gorilla and, specially, chimpanzee and bonobo). This is pretty good hard science interdisplinarity and you are just trying to interfere with humanistic disciplines, with what can only be described as speculative philosophy.
The alternative being “bones and codons” or wild speculation on human-created words, I strongly prefer the first (and so do most people, logically). But anyhow, I don’t really believe that linguistics or kinship studies contradict the OOA model at all (they do not have the time-depth nor the degree of certainty provided by biology and archaeology to challenge anything), it’s just your quite skewed (and I dare say: fundamentalist) interpretation of them which does.
Your perspective forces you to dismiss kinship and languages altogether; mine allows integration of all systems of evidence into one.
I just consider them too unclear and unable to penetrate the depths of human prehistory. But I don’t dismiss them for what they are worth, what I question is your viewpoint of them being able to do so and of doing precisely in the sense you suggest.
But certainly they seem to lack weight in comparison with biology and archaeology. They are still interesting though.
Linguistic studies for instance do suggest some correlation with less time-deep archaeological and genetical findings. For instance it is very possible to correlate Y-DNA R1a (some would argue that all R1) and the Kurgan theory (mostly archaeological) with the expansion of Indoeuropean languages, Y-DNA E3b and some archaeological cultures spwaning from Nubia with the spread of Afroasiatic languages, Y-DNA Q, the Dene-Yenisean language superfamily and the Beringia model for the peopling of America (2nd wave?) - just to mention some of the more clear ones. We can even study genetics more in depth and find (mostly mtDNA) haplogroups that seem to correlate with Aurignacian, Gravettian, Magdalenian or Neolithic demic movements in Europe… but at this time depth linguistics is not anymore of much help. You can still find some words that could well have a common Neolithic origin in West Asia (words meaning things like ram or town maybe) but that’s about all. Beyond Neolithic or Epipaleolithic we are really lost in what refers to languages.
I am not sure how kinship applies to all that, as I never really gave it such importance, but if your kinship models conflict with all that (including linguistics it seems), then you are probably in need of a radical review of your ideas. You seem to be watching the birds fly and conclude from it that the theory of Gravity is wrong.
… lineages L1, L2 and L3 are African-specific. They aren’t found anywhere outside of Africa…
Absolutely wrong, sorry. L3 is found all outside Africa: M and N (and all Eurasian, Oceanian and American natives belong to either of these macrohaplogroups) aren’t but subclades of L3. Just that, for convenience (and historical accidents, like Eurocentric bias), geneticists use shorter names (they could well be called, quite logically, L3a and L3b as well). This is most basic Wikipedia-level knowledge: no breaking news here, really.
Someone who is trying to put all the field of genetics upside down and boasts of having two PhDs should know that very well.
If OOAf proves to be true…
Does my clarification of this fundamental misconception of you constitute enough evidence? I hope so.
Kind regards,
Luis.
PS- Regarding Greenberg’s classifications…
Sure. I know that Greenberg’s proposal for America is most controversial. But I also have the strong feeling that we are now in regards to Native American languages in a stage like when in a different geography Germanic, Hindi and Greek were considered totally different language families., or when Berber and Semitic were thought to have no connection whatsoever (other than loanwords). It may well still be the case, as linguistics advances with the more strict (but difficult) methodology, that Native American languages show quite less diversity than the one you believe. Only time will say.
May 14, 2008 at 8:56 pm
To Luis: Most of what you wrote is the re-phrasing of your belief in genetics and disbelief in kinship or languages. I can’t make you change these assumptions. Neither I can make you change your assumption of what constitutes a proof: a consensus within a closed group of scholars of the same discipline, or an objective set of correlations between one discipline and another representing a common historical process. But time will. In the meantime, here’s what occurred to me as I was reading your comments.
You believe that paleontology (early hominids), taxonomy (gorillas in Africa) and a certain interpretation of population genetics form a system “proving” that modern humans expanded from Africa. Instead of simply reiterating the current consensus in population genetics, you should elaborate on the correlations between genetics, paleontology and taxonomy more, showing how they actually feed into each other. I believe that languages, kinship studies and a different interpretation of population genetics point to America as an autochthonous population, and also an isolate (an autochthonous population is an isolate as well). Methodologically, there’s a tighter connection between kinship, language and genetics, than between genetics, paleontology and evolutionary taxonomy. The former triad describes from different angles the same population process that happened within a manageable amount of time and on the basis of exhaustive global databases coming from virtually the same populations; the second triad is less integrated because paleontology/archaeology deals with accidental finds with dubious relation to living populations, and taxonomy is a very broad perspective on the whole diversity of nature and can’t provide any clues into most recent processes (the fact that gorillas live in Africa doesn’t mean that modern humans came from Africa, at least because other expansions out of Africa happened between 6 million years and 50,000 years ago).
You seem to follow the 19th century logic of evolutionary research: there’s a generic category of “man”. This “man” is related to “apes.” We need transitional forms to prove it: fossils furnish one category of transitional forms, non-Western peoples furnish the other category. Darwin is a noble representative of this logic but I also think we should go beyond that. In the 21 century we could focus on living human diversity, assess it from different angles, and understand where this whole diversity came from. Then we can go outside of our species and make further connections.
Again, humans may have come from Africa (or from America, or elsewhere), but the current ways of proving this alienate a lot of useful evidence from living human populations. Hence, OOAf is unconvincing.
May 14, 2008 at 9:25 pm
More on Luis’s comments: You claim that macrohaplogroups M and N are somehow the same thing as L3. I didn’t find the Wikipedia articles you’re relying on, but here’s a quote from a more specialized source: “Although 2 mtDNA lineages with an African origin (haplogroups M and N) were the progenitors of all non-African haplogroups, macrohaplogroup L (including haplogroups L0-L6) is limited to sub-Saharan Africa.” Or, “Only two mtDNA macrohaplogroups (M and N) and their derivatives persisted in non-Africans after the migration of modern humans out of Africa. macrohaplogroup L is geographically limited to Sub-Saharan Africa and has been divided into haplogroups L0-L6″ (Gonder et al. Whole mtDNA genome sequence analysis of ancient African lineages. Molecular Biology and Evolution 24 (3), 2007). These quotes show that L, M and N are distinct (macro)haplogroups. What’s happening here is that African lineages show a bunch of mutations in parts of genome that show no mutations whatsoever outside of Africa. One way to make sense of the evolution of this diversity is to assume that that African allele diversity is the direct reflection of population age. But another way is to suspect that the genome has “hot spots”. The same goes for languages: some language families evolve slow, others change fast. To say that M and N are “subsets” of L3 is like to say that all languages outside of Africa are subsets of Niger-Congo simply because Niger-Congo, of all language families, has the greatest number of distinct languages. Or, that Caucasians are a subset of blacks because the range of fluctuation of body size in Sub-Saharan Africa is greater than outside of Africa. M is indeed found in northeastern Africa (M1), but there’s a disagreement as to whether M1 came from India or M haplogroup shows too much homoplasy that you can’t really resolve this question.
May 14, 2008 at 10:10 pm
To Victor: I like your observations on “age.” Very true. I am also glad you can differentiate between two levels of argument: my thoughts on OOAm do no harm to my empirical analysis of human kinship systems. The same concerns some of the great genetics papers that are written by scholars believing in OOAf. In the case of the Saami, the fact that their kinship terminology provides an outgroup for all other Uralic kinship structures coupled with the fact that their mtDNA make up is unique has a certain bearing on the classification of Uralic languages. The traditional classification placed Finnic and Saamic at the tip of the Uralic tree, with Samoyedic being closest to the root, or the proto-language, while more recent research (independently of my kinship and of mtDNA) indicated that Volgaic, Balto-Finnic and Samoyedic may be equidistant from the Uralic proto-language. “In the standard binary classification, the number of proto-languages between Proto-Uralic and a Sámi language, for example Inari Sámi, is higher than the number of proto-languages leading to a Samoyed language. As the first steps, Inari Sámi derives via Proto-Eastern Sámi from Proto-Sámi, while the earlier stages are known as Proto-Finno-Sámi…, Proto-Finno-Volgaic, Proto-Finno-Permian, Proto-Finno-Ugrian, and, finally, Proto-Uralic.” But: “Indeed, by comparing material from any two of the nine basic branches, including pairs such as Sámi and Finnic, or even just Mansi and Khanty, we reach a level of reconstruction that is very close if not essentially identical to Proto-Uralic” (Salminen, Tapani. 2002. Problems in the Taxonomy of the Uralic Languages in the Light of Modern Comparative Studies. // Lingvisticheskii bespredel: Sbornik statei k 70-letiiu A. I. Kuznetso-voi. Ss. 44–55. Moscow: Izd-vo Moskovskogo universiteta).
It’s not a tree anymore, but rather “leaves” fallen from a tree. This metaphor is applicable to population genetic models: instead of connecting all human populations into a linear tree that supposedly accurately reflects the historical process of budding off, we should think of them as independent populations with varying demographic histories, hence with varying levels of diversity, and with different intensity of expression of features common to all. These archaisms will eventually lead us back to the most isolated “population,” that being an autochthonous one. I agree that the fact that the Saami have an archaic kinship system doesn’t mean that the Uralic family originated in Scandinavia, but it does suggest that the Saami didn’t participate in the major expansion of Samoyedic and Volgaic branches, and that their position on the map is a retention of a different distribution of a bunch of languages related to Saami. mtDNA suggests that these para-Saamic languages may have been present in Western Europe and that they migrated north with the northward recession of the glacier.
Another great paper in historical linguistics pertaining to the pitfalls of genealogical classifications and tree models (or “subsets” in Luis’s terminology) is Garrett, Andrew. 2006. Convergence in the Formation of Indo-European Subgroups: Phylogeny and Chronology // Phylogenetic Methods and the Prehistory of Languages, edited by Peter Forster and Colin Renfrew. Pp. 139–151. Cambridge: McDonald Institute for Archaeological Research.
May 15, 2008 at 9:23 am
Those sentences are plainly wrong (taken literally) or at least quite misleading (if we take them in a context of specialists’ communication, where everybody knows what is being talked about). It only makes sense if you interpretate, as the author undoubtedly meant, that L3 in this context means L3(xM,N). It is a normal usage of the term L3 but not a precise one.
Using the term “macro-haplogroup” for L is anyhow extremely slippery. L(xM,N) is not a haplogroup because it is paraphiletic. The term paragroup or para-haplogroup is used in these cases with much better sense.
From Wikipedia - Haplogroup: The mitochondrial haplogroups are divided into 3 main groups, which are designated by the 3 sequential letters L, M, N. Humanity first split within the L group between L0 and L1. L1 gave rise to other L groups, one of which, L3, split into the M and N group. There is also a quite nice clickable tree with the same structure.
From Wikipedia - Human mitochondrial DNA haplogroup:
Descendants of haplogroup L3
* Haplogroup M
* Haplogroup N
Some more professional references:
1. Graphs and maps:
- http://www.mitomap.org/mitomap-phylogeny.pdf (huge mtDNA tree but already obsolete in some aspects: lacks L4-L7 lineages for instance)
- http://www.stats.gla.ac.uk/~vincent/images/skeleton07-08-02.jpg (mtDNA skeleton by Vincent Macaulay, a notable geneticist)
- http://www.scs.uiuc.edu/~mcdonald/WorldHaplogroupsMaps.pdf (a quite handy, even if not the most precise maybe, global reference for anyone interested in human genetics: global maps and simplified trees for Y-DNA and mtDNA haplos)
Papers:
- (”The making of the African genetic landscape“, A. Salas et al, 2002).
There is much more literature, of course, specially for the Y-DNA lineages (most handy reference anyhow: http://isogg.org/tree/ISOGG_YDNATreeTrunk.html) but I think that for mtDNA that is the most basic.
I am not lying nor I am confused on this issue: M and N are subclades of L3. And a parallel scheme happens at the Y-DNA level, with C, D and F being part of a larger Afroasiatic macroclade known as CR (or CT since a few weeks ago). I have been following as much as possible this issue of human historical genetics since many years ago, having discussed it in many spaces with other more or less knowlegeable people. I know what I’m talking about here even if I am not a professional geneticist myself.
Please, don’t fool yourself with the casual language of some imprecise geneticist. The reality is very different and well documented.
May 15, 2008 at 11:00 am
Luis, then how come the sole African representative of macrohaplogroup M, namely M1 (found both in East and West African samples), is claimed to be the result of a back-migration into Africa? If M1 was the same thing as L3, then how can it migrate back into itself? If the whole world is Africa, then evolution is at a stalemate.
See Gonzalez et al. Mitochondrial lineage M1 traces an early human backflow to Africa. BMC Genomics 2007 (8).
There’re other recent papers on the issue. The back migration of M1, U6, N and J into Africa, all the way into Mali and Tanzania, has been argued for by several labs. And now recall the controversy around the YAP+ lineages that account for more than a half of African Y chromosome diversity and that may have antecedents in Central and South Asia. This issue isn’t resolved yet, though.
Yes, geneticists do use the word “subset” to describe tree topology. But it doesn’t mean “identity”. You could say that humans are a “subset” of Neandertals just because Neandertals have a whole string of mutations where humans have none. Since all species are related, and the older species tend to vary more than the younger ones, they end up on top of the tree, and the variation within younger species can be said to be a “subset” of the variation within older populations, but in actuality the younger ones don’t just reapeat a part of the older ones but only partially overlap with them. But for interspecific trees the word “subset” isn’t used, because in this case “humans” will become a “subset” of apes, while in fact the correct way to describe it is that our ancestors at some point converged with their ancestors.
You have to understand that tree-building is a theoretical exercise; it’s not immediately present in empirical data. What geneticists are saying is that since Africa shows a lot of allele variation and we need to somehow tie it back to non-Africa, we assume that certain mutations happened before others, and some of the mutations are not recorded in the available sequences. L3 is not identical to M or N. L3 is the way to link M and N to L0, L1 and L2. The whole L thing (call it para-) is African-specific, and is sufficiently distinct from M and N found everywhere around the globe, including Africa. Johnson et al. 1983 noticed exactly that: you either stick with the most common, pan-human markers, and then the root seems to lie outside of Africa; or you stick with the divergent lineages found only in Africa, and then the root is in Africa.
The question is where L (or more exactly, L0, L1 and L2) comes from. If it’s so diverse, then we may assume that it’s the oldest human clade (but then shouldn’t we find the same restriction sites active in Australia, Andaman Islands, or in other archaic pockets outside-of-Africa, in the same way as we find M and N lineages in Africa?); or we may suspect an admixture with H. erectus or whoever else; or we may hypothesize a combination of varying rates (hot spots) and demographic history (free gene flow, large effective population size, no lineage loss, etc.) I tend to lean toward the latter explanation, and it’s not because of ignorance. The diversity of L is not unique. M shows “a dazzling array of basal lineages” in India (see Sun et al 2006). A similar situation occurs in Australia, where a whole bunch of local lineages within the M and N macrohaplogroups have been described.
If African genetic diversity was simply a function of age, I would love to see how this translates into my data and into linguistics. That would help a lot, and I don’t care who is right. As of now, there’s a disconnect that I think needs to be openly discussed and not tacitly dismissed. If the American Indians simply preserved certain archaisms vis-a-vis the Asians, but Africans scored high on archaic characters, I wouldn’t care, but the whole kinship and linguistic spectrum is skewed away from Africa into Australia/Papua New Guinea/Pacific Rim/America. The situation is the exact opposite from the current consensus found in genetics, and I thought it was fascinating and intriguing. No need to take sides, just develop theories that are targeted toward explaining a puzzling situation but not explaining inconvenient data away.
See my most recent response to Victor regarding the problems with tree-building in linguistics, with examples from Uralic and Indo-European. A simplistic approach to tree-building results in certain languages/populations ending up at the tips of trees as supposedly the last ones to branch off (Saami in the traditional classifications of Uralic languages or American Indians in the recent genetic research). Why? Because scholars don’t know upfront what to look at, what is secondary local variation, what is the result of parallel evolution or coincidence (e.g., the satem languages within Indo-European are not a genetic grouping, but the result of convergent processes of palatalization that happened independently in the eastern “branches”; Grassman’s law in Greek and Indic describes two separate events, while on the surface they look like common inheritance, etc.), and what constitutes the ancient common fund, they tend to quickly generate trees based on unweighed/superficial characters. Then these trees are said to represent a real historical process.
May 15, 2008 at 11:32 am
Further thoughts: it’s also instructive to look at African click languages as a parallel to the African genetic situation. The Khoisans are as specific phonologically as all Africans are genetically. Knight and Mountain, who I know well from my days at the Anthropological Genetics lab at Stanford, once argued that since clicks mirror genes and we know that African genes are the most ancient, then we could say that the early human language was a click language, and that the Khoisans are the living testimony of that. It kinda jibes with our stereotype of ancient humans being odd and primitive like children who use clicking sounds for fun. But no serious linguist can accept that. Why? Because clicks are Khoisan-specific. The only other instance of a click language is Lardil, a secret language in Australia, but it can’t be used for comparison, since it’s an artificial language created for ritual purposes. The whole diversity of world phonologies shows no traces of clicks whatsoever. We need some kind of transition from clicks to the rest of world sounds in order to even to start thinking about clicks being as old the human language capacity. The only answer that makes sense to a professional linguist (and even Joe Greenberg used to say “I know how clicks evolved”) is that clicks are a local Khoisan innovation caused by isolation and other social factors. Christie Turner once said the same thing regarding the so-called “Bushman canine,” a dental character found only among the Khoisan-speakers but nowhere else in the world: “If modern humans had originated in and spread out of Africa, we would expect the Bushman canine to be more frequent than it is outside of Africa.”
All these themes are expounded on in “The Genius of Kinship.”
May 15, 2008 at 7:18 pm
German, the issues you are raising take us back to an era of anthropology, (and comparative musicology as well) where all sorts of very puzzling and contradictory ethnographic, linguistic, musicological and archaeological evidence was being tossed around both in favor of certain grand theories of human origins and evolution and against them. There is no lack of this sort of evidence both for and against just about any such theory anyone could think of. At a certain point everyone just gave up on that sort of thinking because it was leading nowhere. There was a ton of very compelling evidence, but it was impossible to interpret in any consistently meaningful way — and above all it was impossible to formulate a hypothesis based on any of the old theories that could be tested with any degree of rigor and objectivity.
At around the same time, very sadly, the postmodernists came along and turned anthropology (and comparative musicology) into a very mean spirited and also misguided puritanical inquisition, so it actually became dangerous to ones career to delve anymore into such issues at all.
What has changed in recent years is largely due to the verysurprising and unexpected findings of the geneticists, led by people like Cavalli-Sforza, Rebecca Cann, Alan Wilson, et al. Thanks to their efforts it is now possible to tie the ethnographic, linguistic, musicological and archaeological evidence to research that is far more objective, rigorous and reliable than anything that had even been imagined in the past. Is it perfect? No. Are there problems in devising these phylogenetic trees? Do they contain errors? Are they based on samples that may be too small or too biased? The answer to all the above is: yes. Nevertheless, such trees are based on a trulyrigorous and exciting new approach to human history that can indeed be tested and is being tested by people who for the most part have open minds and are not attached to any particular agenda. Their research is part of an ongoing process and is continually being critiqued and refined.
The points you have been making in your comments here are certainly interesting and meaningful but they look very much like an attempt to revive the old type of theorizing and revive the old pointless disputes that gave all this sort of thing such a bad name in the past. I have the impression that you are very knowledgeable about kinship and kinship terminology as well as historical linguistics, but any meaningful theory based on such knowledge has to be formulated in such a way that it can be tested. The new genetic research for the first time offers us the means to test such theories with a reasonable degree of rigor. It seems to me that you are rejecting the genetic research out of hand simply because you assume in advance that your theory will fail that test. Fair enough. But you must then find some other means of rigorously testing your theory, rather than simply dismissing whatever doesn’t appear to fit it. And if you can’t come up with any such test, then you are simply one more person out of the many thousands who’ve come before you who’s come up with an interesting theory that might or might not be true. As I see it, that’s the old way of thinking, the way that led nowhere and there’s no point in continuing any more along that path.
May 15, 2008 at 8:54 pm
Victor: Very eloquently written but I can’t see how it all applies to my research. I just can’t see OOAf and the recent peopling of the Americas from my data, from linguistics data, or from any other ethnological data that can be brought to the round table. Should I just stop right there, and not go into genetics at all? There must be people in science who could, with various degree of success, move between the data to see what fits and what doesn’t. Otherwise, one discipline usurps control over a certain set of questions and there’s no way to test if the theories are true. Archaeology has been treating the peopling of the Americas as its home turf for a century and arrived at a theory that was supposed to be super rigorous, namely Clovis I. But now it gets falsified almost every 5 years, from Monte Verde to Oregon excrements. There’s nothing in the archaeological record that warrants the idea that humans came to teh Americas 11,500 BP; no finds doesn’t mean no humans. But we were trained to believe that this rigor permeates the Pleistocene-Holocene archaeology of the American-Siberian interface. Rigor is great but rigor worship is not. We should clearly distinguish what we can demonstrate beyond reasonable doubt and what we cannot. Neither the recent peopling of the Americas, nor OOAf are currently demonstrable. And that’s fine: we have tons of evidence now, let’s just integrate it as we go along, family after family, region after region, and then we’ll see where we all came from. But nay! Naive 19th century science gets revived over and over again: let’s build grand theories before we can prove anything. And watch living languages die because we just don’t see what we can make of them. My OOAm theory is a way to preserve a critical spirit on this high level of the argument.
Now you can see, Victor, that I would rather read your well-written piece as a critique of Cavalli-Sforza et al., who try to penetrate new great data with the blunt tools of 19th century reasoning. I agree with your critique of post-modernism, but my reminiscences of my life at Stanford were intended to communicate the fact that both Cultural and Social Anthropology and Anthropological Sciences approach anthropological knowledge from two different perspectives, two different methodologies and are prone to commiting two different kinds of errors. OOAf theorists refuse to acknowledge that there’s an interpretive model working in their minds beyond what actual data shows. Post-modernists refuse to take their own principles literally and deteriorate into ethical formalism. I believe that we’re embedded in culture and hence absorb ideas through communicative channels that are very subjective. But then I believe you can always work toward lifting this veil of subjectivity, and that’s the modest purpose of science. It’s not whether this lifting can ever be final, it’s the process that matters, for it creates a better picture of the world.
I’m not afraid of having my ideas tested. “The Genius of Kinship” contains a different histioriography, a different demographic scenario, a different migration scenario and a different primary source of information. I didn’t try to re-do genetic trees, but I did come up with a tree-like structure of the evolution of sibling nomenclatures as the easiest kinship data to slice that way. Clear differentiation along all the critical points of a debate is what allows ideas to be tested. I may not have succeeded in explicating it in a way that makes it appealing to scholars who could test it, but the intention is clearly there. And now time will decide.
Recent history shows that science didn’t replace religion in public consciousness, and as we all know there’re a lot of people who believe in the Bible rather than in evolution. Science and religion simply divided the world into the spheres of influence. It means there’re insurmountable cultural differences between people. For instance, you’re criticizing me for something I can’t even accept as applicable to me. I appear to be the person who dismisses genetics - not true; I can see where a possible alternative was surpressed early on in mtDNA research - true. I appear to be the person who is afraid to have his theories tested - not true; I believe general theories have to be tested against data from multiple disciplines, not just one. What can I do to share at least the same reference frame with you? Agree on OOAf? But what if the root of the problem will turn out to have nothing to do with OOAf?
May 15, 2008 at 9:40 pm
then how come the sole African representative of macrohaplogroup M, namely M1 (found both in East and West African samples), is claimed to be the result of a back-migration into Africa? If M1 was the same thing as L3, then how can it migrate back into itself?
Oh my…!
M1 is not “the same” as L3 it is a subclade of a subclade (M) of L3. It is different from other L3 subclades and, yes, it is thought to represent a back-migration from West Asia at a later date than OAA.
There’re other recent papers on the issue. The back migration of M1, U6, N and J into Africa, all the way into Mali and Tanzania, has been argued for by several labs. And now recall the controversy around the YAP+ lineages that account for more than a half of African Y chromosome diversity and that may have antecedents in Central and South Asia. This issue isn’t resolved yet, though.
Sure, the fine detail may be not wholly clear. For instance Maca-Mayer suggests that U6 back-migrated via Ethiopia in spite of being more diverse in the Iberian peninsula (her own data). The YAP debate is apparently cleared anyhow since DE* has been found in Nigeria (and also because E is soooo diverse in sud-Saharan Africa). DE is now mostly believed to have coalesced in Africa and only a branch of it (D) originally being part of the OAA migration.
But nothing of this challenges the fact that the highest diversity at high levels of the tree is in Africa and not anywhere else. Y-DNA has two main baranches (A and BR) that are exclusively or most diverse in Africa, BR has two branches, B and CR, of which only CR is more diverse out of Africa. CR itself has two branches CF and DE (YAP), the first one almost exclusively Eurasian and the latter divided in two branches: one Eurasian (Asian only actually) and the other almost African only. The high diversity for Eurasia only appears at lower levels of the tree than in Africa.
Exactly the same happens with mtDNA: L0, L1*, L2, L3(xN,M), L4, L5, L6 and L7 are only found in Africa. M and N are subclades of L3, which is a subclade of L1. M and N are thought to have coalesced in Arabia or South Asia at the time of the OOA migration.
So the OOA clades were: C, D and F (Y-DNA) and M and N (mtDNA). The simplest model suggests that they all were together in Arabia or South Asia (where they coalesced before branching out and expanding further). Of course there can be alternative theories (seevral migrations, migration via Palestine…) but all seem to require an ancestral African urheimat anyhow.
You could say that humans are a “subset” of Neandertals just because Neandertals have a whole string of mutations where humans have none.
No way! Neanderthals and H. sapiens are separate sets on light of all known research. Both are subsets of a more ancient species certainly (it used to be known as H. erectus but now there is varied terminology) but not subsets of each other, not at all.
… but in actuality the younger ones don’t just reapeat a part of the older ones but only partially overlap with them.
That is not a subset but an intersection. It’s rare in biology but you can well say that the claimed introgressions of Neanderthal or other archaic species into modern humans acount for that. It would be a minimal intersection in any cas