2007/02/09

Review of Blood of the Isles / Saxons, Vikings, and Celts by Bryan Sykes

I just finished reading Blood of the Isles,1 by Bryan Sykes of Oxford University. Sykes styles himself a genetic archaeologist, and in this book he spends as much time on the historiography and mythography of identity in Britain and Ireland as he does on interpretation of DNA tests. The result is a book that should be included as a text in any general survey history of the British Isles. It is the first history that uses genetic material as its primary data set, and it works very well, coming from a scholar with interests as well-rounded as Sykes. He outlines the way migrations to the Isles have been thought of, from the polemics of sixth century monks to the absurdities of nineteenth-century racists, sets out hypotheses based on the more reasonable theories, then tests them through data, separately for Ireland, Scotland, Wales, and England. The conclusion, and it is a very important step for our understanding of history to take, is that the British and Irish are basically one people, the same as they were when they settled the Isles from Spain as the ice receded, with only minor impact from the names that history records, Celts, Romans, Saxons, and Vikings. The impact of these peoples, genetic in part, was mostly cultural and historical.

Yet it suffers from one important fallacy, the same that the most brilliant minds in genetics still make. Anyone who understands the numbers involved, and Sykes clearly does, is able to see that a male resident of Aran, let's say, of matrilineal clan Ursula (a woman who lived in Greece something like 45,000 years ago), and carrying the Y-chromosome of Somerled, the Gael who drove the Vikings from the Isles, does not just have these two as ancestral lines. In a population that includes hundreds or thousands of lineages (all populations) which does not limit marriage to within both patrilineal and matrilineal clans (no population I'm aware of), most lineages in the society will become ancestral to most individuals, given only a few generations. I have no doubt that Sykes understands this, but it's curious that he doesn't emphasis it. To me it is the *most* fascinating part of genetic genealogy. Genetic genealogy has the potetntial to forge a quick connection to both local and worldwide communities, not just a single or dual marker of heritage.

That man from Aran is not just descended from Ursula, but from all mitochondrial lineages surviving in his population and from many mitochondrial lineages now extinct, but which lasted long enough to be counted among his ancestors. The same is true of all current and all marginalized Y chromosome haplotypes. He carries his descent from Ursula and Somerled in his genes, true, and there's something very exciting about that, but it's his invisible lineages, traceable only through testing uncles or cousins, close and distant, that reveal his true nature as an individual who encompasses in his DNA the entire diversity of his society, both other common haplotypes and very rare ones. Traceable genes may not be in his mitochondria or Y chromosome, genes may not even have been inherited from these people at all, but his ancestry is far, far more complicated than a simple story that goes to the ends of the Mediterranean and then back to Africa. It all goes back to Africa, of course, but by 2lots routes, not 2.

I find it difficult to believe that Bryan Sykes has not thought this fallacy through. In Blood of the Isles, as throughout his popular work, Sykes strives to preserve the individual nature of data collected from his subjects' DNA, preferring to let stories, rather than statistics, emerge from the data. The first woman to carry a particular mDNA haplotype becomes a clan mother (his The Seven Daughters of Eve) instead of a number, and ubiquitous Y chromosome DNA lineages are traced, quite believably, to historical and mythological kings, warlords, and other such headmen. These are stories people can quite easily append to family histories that are much closer in time. Descent from Jasmine, the Syria-born matriarch of the Neolithic agricultural migrations into Europe, from Genghis Khan, or from Somerled, is a much more romantic notion than identity with R1B or O3 or H. It's certainly easier to talk about at the family reunion. Yet the essentially communal nature of genealogy brings in another question. Every nuclear family is made up of individuals of different matrilineal clan lineages, every extended family of many matrilineal and patrilineal ones. This is as easy a notion to understand as that a family can have many surnames. So why can't genetic genealogy explore it.

If I weren't confident that Sykes has science's best interests at heart, I would be cynical and say that he perhaps believes that Oxford Ancestors' coffers will be more easily filled if people feel they are getting a real sense of identity from their tests that their neighbors don't yet have. Perhaps this is true, and perhaps more people will send in cheek swabs to Oxford Ancestors or to Spencer Wells's Genographic Project if they think the not insubstantial money they are paying can tell them "where they come from", rather than one of the countless places. I repeat, every one of the thousands of test subjects that make up the data for the book descend not only from their assigned clans, but from all. Sykes acknowledges this when he writes that the Oisin clan was found in both Norway and Scotland, so that Oisin men from the Orkneys are not necessarily descended from Picts/Celts; he does this by wryly noting that the Vikings did not test their crew's Y-chromosomes. Advance this warning farther, and truly think about his subjects as individuals, and you start thinking about their fathers' mothers as well, their mothers' fathers, and every other line that passes through ancestors of more than one sex.

If Sykes's new view of British and Irish history exposes the poisonous notion of separate Saxon and Celtic peoples as a lie, a more nuanced view of genetic genealogical data can go even farther. Sykes tells human stories by bringing the individual line out of the data, a very important step, because it brings us away from very simplistic notions that would have peoples, not individuals, splitting off on the way out of Africa, perhaps never to reconcile. The next thing we need to do is put those individuals back into their societies by first putting them back into their families. Then, and only then, can genetic genealogy tell an accurate human story. It's one we all can, and need to, understand.

The U.S. title is Saxons, Vikings, and Celts, which is a strange change for a book which goes about proving that the migrations of Saxons, Vikings, and Celts aren't very significant contributors to British and Irish genotypes.

3 comments:

Adam Lawson said...

If Sykes really believes that the British and the Irish are one people then he should read this scientific paper:

Evidence for an apartheid-like social structure in early Anglo-Saxon England

Ruby, J. M. said...

I don't know if Sykes has read it, but Oppenheimer has. He isn't convinced.

"The problem is that the English resemble in this way all the other countries of northwest Europe as well as the Frisians and Germans. Using the same method (principal components analysis, see note below), I have found greater similarities of this kind between the southern English and Belgians than the supposedly Anglo-Saxon homelands at the base of the Danish peninsula. These different regions could not all have been waiting their turn to commit genocide on the former Celtic population of England. The most likely reason for the genetic similarities between these neighbouring countries and England is that they all had similar prehistoric settlement histories."

afwife said...

Thank you for your excellent review! As a caveat to your criticism on the one bloodline fallacy, I'd like to point out a key part of the book you brought up at the end of the review. Despite other groups invading the isles many times, they made little impact on the genetic legacy of the population who lived there. My point being, widespread intermarriage and interbreeding between ethnic and cultural groups is a fairly new phenomenon. In pre modern times, many, if not most people lived their entire lives in a relatively small geographic area, and it's likely a family line would continue for many generations without much deviation away from their genotype or their local cultural group. Look at the Cheddar man! A living relative of that ancient deceased man was found living just a few miles away from where he had been buried for thousands of years. Of course, you are correct. A person's family tree is filled with thousands of names, especially once you go back several hundred years, but after a while those many names begin to insect and interweave, and you'll find the expansion isn't exponential. Hanks for the review!