One of the peculiarities of the synthesis of 19th and early 20th historical linguistics and biological anthropology was the perception by many British thinkers that the English, as the scions of the Anglo-Saxons, were fundamentally a different race from the Celtic nations to their west, the Welsh and Irish, and the Scots to the north (yes, I know the Scottish nation emerged is a mix of Celtic and Anglo-Saxon elements which were preponderant at different times and periods). In other words English nationalists would characterize their own race as a branch of the German peoples. English was a Germanic language, and the linguistic chasm emphasized more starkly a distinction from the Celts who inhabited Britain prior to the arrival of the Germans, and gave the island its name before they were marginalized and pushed to the “Celtic fringe.”
The historical context of this does not need to be elaborated in detail. The Emerald Isle’s integration into the United Kingdom of England, Wales, Scotland and Ireland was always a difficult affair. This was due in large part to religion (the lack of an effective Irish Reformation may have had other structural causes); the Irish were a Roman Catholic populace at a time when Roman Catholicism and loyalty to the monarchy were presumed to be contradictory. In 1800, before the potato famine and the English demographic explosion, Ireland accounted for one third of the population of the United Kingdom (I do not put much stock in the linguistic difference, as the Welsh speaking regions were firmly Protestant and so not perceived to be sources of equivalent dissension despite their cultural marginality). With the rise of taxonomic science what was a crisp social chasm was reconceptualized as a biological and evolutionary gap along the Great Chain of Being.
In the twentieth century the tide turned, today most scholars would assert that the shift from Celtic to Anglo-Saxon speech and culture in what became England was a matter of emulation, not genetic replacement. Personally I suspect that the pendulum has swung too far, but it does show how strongly influenced by fashion these sorts of preconceptions are.
Modern genetics can clear up the confusion to some extent. A new paper in The European Journal of Human Genetics surveys samples from Dublin, the south & southeast of England (the heart of Saxon Britain), Aberdeen, Portugal, Bulgaria and Sweden. Population structure and genome-wide patterns of variation in Ireland and Britain. I’ll just focus on the figures of interest in relation to the questions I aired above.
I’ve added some labels to figure 1, but it’s pretty obvious what it’s depicting. Each point is an individual. CEU = Mormons from Utah. This is mostly a British origin sample, but I assume its overlap with Swedes is indicative of the European immigration to Utah by early Mormon converts, some from Scandinavia.
And here is what economists would term a more stylized figure from the supplements:
These figures are showing what we know from other studies on European genetics; the largest component of variation seems north-south (at least until you start pushing into Russia where a simple European wide pattern starts to break down), and the second component is west-east. This is more evident in the frappe plots, where you see the individuals within the populations broken down by K ancestral groups.
Again, from the supplements:
The above figures require a little art in their interpretation. Remember that the PC charts are just representing the biggest components of independent variation within the data set. As for the frappe results, they don’t always represent real ancestral populations in a straightforward manner. Or at least we have no independent checks on what was going on ten thousand years ago in Europe. So below are the pairwise Fst values. Remember, these compare the proportion of between group genetic variation across the pairs. The print is small, so let me just tell you that the Fst value for England-Sweden is twice a large as England-Ireland. In other words the English of the south and east of England are closer to the Irish of Dublin than they are to the Swedes.
Ideally the Swedes would not be the reference population for the Germans of yore. Rather you’d want Frisians, Danes and Saxon Germans. From what I’ve seen in the other results on European genetics Swedes have been somewhat influenced by the Finns, who are genetically peculiar, so that might understate the German affinity of the English as some of the distance might be due to the Fennic component in the Swedish gene pool. But I’ve seen other studies which lead me to infer that the peoples of the Isles share more than not, and the English share more ancestors with the Irish and Scottish than they do with the Saxons over the sea.
H/T: Dienekes
Citation: Population structure and genome-wide patterns of variation in Ireland and Britain, doi: 10.1038/ejhg.2010.87
Razib Khan’s degrees are in biochemistry and biology. He has blogged about genetics since 2002, previously worked in software development, is an Unz Foundation Junior Fellow and lives in the western US. He loves habaneros.
June 24th, 2010 at 5:25 am
[...] This post was mentioned on Twitter by Ron Simon, Geoffrey Dyson, Maggie, Al Poe, World Amazing Things and others. World Amazing Things said: The English & Irish, together again | Gene Expression: One of the peculiarities of the synthesis of 19th and early… http://bit.ly/cwoK12 [...]
June 24th, 2010 at 6:54 am
Finally, there is genetic evidence of who the English is. Dominant culture spreading is not always result of genetic spreading. It had happen many times and places like India, Middle East, China, south america. ect. Indo-european culture itself is an example. At end, dominant rulers want submissive subjects (or domestication to your verbal command like domesticated animals) who can produce wealth for the lords.
June 24th, 2010 at 7:40 am
It it does show how strongly influenced by fashion these sorts of preconceptions are….
I can imagine the dynamic. For teaching purposes people in the field (broadly defined) want to have a definite answer on most questions, excluding things under active investigation at the time. And most people in the field won’t have studied the particular question (e.g. emulation v. replacement) and wont want to. So just to make things simple a standard answer will be reached by consensus, and in order to change the consensus serious new work will have to be done. It saves labor.
June 24th, 2010 at 10:19 am
Apologies as I’m not too well up on the genetic charts. I re-read Stephen Oppenheimers Origins of The Britsh recently (being Irish I’d rather not being lumped in with British but he does address this) and I think he said the same thing.
So please correct me if I’m wrong. His claim is; something like 80% of Scots, irish and Welsh have more genetic markers that match the Basque region (R1b14 i think), with small neolithic contributions from the meditteranean but alot of cultural exchange with the language possibly coming from here.
South and East England especially differs in that it received more of an input from across the North Sea (doggerland and all that)during the Mesolithic and Neolithic and this continued up until Roman times, and the Saxon invasion was probaly an elite male replacement but possibly more Angle and Jute, who were probably more Scandanavian than German based.
And the term Celt as an ethnic lable for Scots, irish and Welsh is a mislabel/glorified myth.
June 24th, 2010 at 10:58 am
His claim is; something like 80% of Scots, irish and Welsh have more genetic markers that match the Basque region (R1b14 i think), with small neolithic contributions from the meditteranean but alot of cultural exchange with the language possibly coming from here.
i think this hasn’t panned out. oppenheimer was looking at one locus, the male lineage. total genome examination seems to show a somewhat larger north-south, not east-west, gap. the neolithic part is really confused right now as well; r1b1 may not have been typed correctly. all of the rest of the stuff, oppenheimer was being heterodox, and i don’t think that’s panned out either.
June 24th, 2010 at 11:02 am
It’s a great pity that Oppenheimer’s conclusions are suspect – I spent my own money on a copy of that book.
June 24th, 2010 at 11:27 am
From what I remember he did look at mitochondrial but the main focus was Y chromosome.
So what you are saying is there may be more of an ifluence from the North Sea region (ultimately a Balkan or Uralic ice age refuge)? Or some things aren’t just straight forward.
June 24th, 2010 at 11:32 am
everything i say is really uncertain. but yes, from the total genome stuff it does look like there’s more affinity of northern peoples than atlantic peoples. but obviously the british are a synthesis, we just don’t know the weights too well right now. but the total genome stuff doesn’t place the british that close to the spaniards.
June 24th, 2010 at 11:56 am
But the idea of cultural similarities/difusion of Atlanitc sea board people as put forward by people like Barry Cunliffe would be solid enough?
When you say British, are you grouping Irish in there also? Is there anything worth reading on the topic?
June 24th, 2010 at 12:06 pm
But the idea of cultural similarities/difusion of Atlanitc sea board people as put forward by people like Barry Cunliffe would be solid enough?
yes. cultural and genetic. i tried not to interpret too much, but look at the 4th colorful figure. each horizontal line is an individual, and the colors represent ancestral components from hypothesized original groups which admixed.
When you say British, are you grouping Irish in there also? Is there anything worth reading on the topic?
first question, yes. second question, i think historical genetics of europe is in flux right now. perhaps a book will come out in a few years. right now if anyone says anything authoritative, assume they’re lying :–)
June 24th, 2010 at 12:17 pm
Thanks for answering my questions, this is a topic I find very interesting but have yet to get a grasp off.
Just a couple other points.
Do the graphs above use Dublin only data or other parts or Ireland, as Dublin was originally established by the Vikings and was the seat of power of the British would this skew the data as opposed to using data from the west coast?
Is using current political country boundaries probably an inaccurate way of looking at things (from trying to fit things into nice boxes like who are the Irish, who are the English etc) and would maybe taking things from more regional points of view be a better way.
And what in your opinion is the main barrier to getting a good overview of European genetics?
Thanks.
June 24th, 2010 at 12:37 pm
Do the graphs above use Dublin only data or other parts or Ireland, as Dublin was originally established by the Vikings and was the seat of power of the British would this skew the data as opposed to using data from the west coast?
i think this is an issue.
And what in your opinion is the main barrier to getting a good overview of European genetics?
coverage and sample size. i assume this will disappear in a brute force manner; regions will be covered, and sample sizes will increase.
June 24th, 2010 at 12:49 pm
Naughtius said:
Do the graphs above use Dublin only data or other parts or Ireland, as Dublin was originally established by the Vikings and was the seat of power of the British would this skew the data as opposed to using data from the west coast?
I have concerns too over the locations that the data was collected. The Viking settlement of Dublin was on the North side of the river Liffey, today the North side is the poorer side and populated by people from all over the country, especially the West. The South side is the wealthier side and was (is ?) a bastion of the Anglo-Irish Ascendancy – Merrion Square/Sandymount etc. From a previous study (Capelli?), the town of Rush, just to the North of Dublin has some Viking influence. So depending on which part of Dublin, you could get results that are reflective of the population substructure in the city.
I would have loved to have seen samples from:
West Kerry (Neolthic metal workers?), West Mayo (Early Neolithic – Ceide Fields) , South East Wexford (Normans), Leitrim (Vikings), East Donegal (Ui Neills), Golden Vale Cork (Elizabethan English) and finally West Laois (purest native Irish – never colonized).
Also Aberdeen in the North East of Scotland had little Viking influence, but probably had some influence from the Angles of South and Eastern Scotland.
June 24th, 2010 at 12:52 pm
though please note that the PC charts look pretty much like all the other ones with irish, british, scandinavian and iberian charts in terms of the relationships.
June 24th, 2010 at 12:55 pm
Well, I, too, bought Oppenheimer’s book, and it was a fun read.
I also bought Bryan Sykes’ (no recent relation, but shared origins in Huddersfield).
Bryan thinks the mitochondrial and Y analyses suggest somewhat different origins. Female “Celts” have little relation to continental/mediterranean Celts and are diffused throughout the British Isles. This may reflect a population that was already there when the so-called Celtic migration occurred.
On the other hand, there does seem to be a male gradient with some Saxon/Danish DNA in eastern England. But even there, Sykes’ concludes that the population is largely the descendants of very early migrations, say 6000 ya.
His samples are very small, but his results seem to conform better to this new paper than Oppenheimer’s.
June 24th, 2010 at 1:00 pm
fwiw, i think both sykes & oppenheimer probably understestimated the anglo-saxon contribution. it may be minor, but i don’t think it’s trivial. inferences from contemporary patterns to the past have turned out be somewhat sketchier than we’d have thought.
June 24th, 2010 at 1:03 pm
I think that this sort of thing is probably true in many places where an invader came and conquered, but the majority of the population remained indigenous. Such as: if a similar genetic study was done in Turkey, especially west Turkey, they would find that there are more Greco/Roman/Anatolian etc., genetic markers than Turkish.
June 24th, 2010 at 1:06 pm
Such as: if a similar genetic study was done in Turkey, especially west Turkey, they would find that there are more Greco/Roman/Anatolian etc., genetic markers than Turkish.
yes. when i looked at it looks as if around 10% of the turkish anatolian genome (or less) is of mongolian origin.
June 24th, 2010 at 2:27 pm
pconroy
I thought it would be Leitrim no-one would have went near not Laois, but nevermind.
But Leitrim as Vikings interesting, would that have to do with coming up the Shannon as it’s a small coastline or proximity to Donegal (fort of the foreigner and all that)?
Oppenheimer didn’t necessarily write off “Saxon” influence, he claimed it was part of a longer sustained movement of people.
June 24th, 2010 at 3:39 pm
One problem is that Dublin experienced a high level of British migration over the centuries and is probably not a good representation of the Irish population. It would likely have been better to sample populations in the west of Ireland which tend to have less British ancestry.
June 24th, 2010 at 3:53 pm
This is fascinting stuff. To me, the Germans and the English have facial characteristics different enough that I can sometimes pick them out. However, the Irish and the English have always looked very similar to me. I always have wondered why they don’t look more distinct.
June 24th, 2010 at 4:00 pm
Naughtius,
There was a paper some time ago on skull/facial morphology of the Irish, and the result was they identified broadly 2 fairly distinct types, the first covering much of Ireland, the second in the Midlands and North Midlands in particular – where a short faced more Scandinavian type prevailed. Y-DNA evidence also points to the 4 counties around Upper and Lower Lough Erne as having more Scandinavian type DNA.
Another thing worth mentioning is that Ptolemy mentioned a number of coastal tribes in Ireland, 2 of these the Cauchi/Chauchi and the Menapii were located in Wexford/Wicklow in the South East, and are the same names as tribes on the Saxon Shore (Litus Saxonicum) of modern day Belgium/Holland, who were battling the advancing Romans from the South and the advancing Germans from the East. These Belgae established the town of Menapia in Wexford. They were know in Gaelic as Monaig or Manach and later migrated to the Lough Erne, and would lend their name to Fermanagh ( Fir Monaig/Manach). This would be an alternative explanation for the Viking type skull/face and DNA.
June 24th, 2010 at 4:31 pm
yes. when i looked at it looks as if around 10% of the turkish anatolian genome (or less) is of mongolian origin.
The average Mongoloid ancestry among the Turks tested for the last Jewish study seems to be 5% (maybe less). But, of course, much more investigation is required to be able to say anything conclusive on this matter.
June 24th, 2010 at 5:10 pm
onur, there is sensitivity to region that one is sampling from.
June 24th, 2010 at 5:44 pm
onur, there is sensitivity to region that one is sampling from.
You may be right, but autosomal tests are newly beginning to be applied to Turks in genetic studies, even mtDNA studies about Turks are very insufficient. The only noteworthy Y-chromosome study about Turks is Cinnioglu et al.’s.
June 25th, 2010 at 2:17 am
[...] The English & Irish, together again | Gene Expression | Discover … [...]
June 25th, 2010 at 4:59 am
Btw, there are a lot of ethnic Turks who participated in 23andme’s autosomal ancestry test. In it, the average of the Mongoloid components of ethnic Turkish individuals is 2%, with significant variance in their Mongoloid component: some have no Mongoloid component, while some have as much as 8% Mongoloid component.
June 25th, 2010 at 6:32 am
pconroy
Interesting stuff, with the Wexford area Belgaic tribe, would the fact that there is a different gaelic dialect (I think so at least, but it may be more Waterford) there tie into this? I think Stephen mentioned this in Origind os the British, another thing that crossed my mind when reading about some of the tribe names was the Brigantes, I think he said there was a tribe of that name in North central England and somewhere on mainland Europe (according to wiki in the Wexford area aswell http://en.wikipedia.org/wiki/Brigantes), this name ties in with some goddess mythology. Would there be any link to Saint Brigid of ancient Ireland fame?
Crazy specualtion; the Gaelic for foreigner is Gall, now with one of the ideas being that Gaelic may have spread from the Gallicia area. Would there be any link between the two words?
Is there any good reading material out there for what you mentioned above.
It’s actually interesting that The Book of Invasions may have some of it’s stories grounded in some kind of reality.
June 25th, 2010 at 8:49 am
Btw, there are a lot of ethnic Turks who participated in 23andme’s autosomal ancestry test. In it, the average of the Mongoloid components of ethnic Turkish individuals is 2%, with significant variance in their Mongoloid component: some have no Mongoloid component, while some have as much as 8% Mongoloid component.
It is also worth noting that Turks’ eastern neighbors like Armenians, speakers of Caucasian languages, Kurds, Iranians (even just Indo-European speaking ones) and some northern Arabic speakers show Mongoloid components in similar amounts with Turks in the ancestry tests of companies like 23andme and deCODEme and in academic genetic studies.
June 25th, 2010 at 11:52 am
[...] 23andMe research article finally published. Dr. Dan MacArthur offers his take on a new PLoS Genetics paper which was published using 23andMe’s user base. Of course there’s already information coming out of 23andMe’s user community not getting into the academic literature, see this comment below. [...]
June 25th, 2010 at 12:31 pm
Naughtius,
I think the Belgae spoke a form of Germanic, as they came from the Saxon Shore and the word Saxon itself – like Viking later – just meant sea rover/raider, not necessarily an ethnic name, more an occupation. So to me Belgae=Saxon. The areas the Belgae settled in England became Essex, Wesex, Sussex and so on, and were clearly Saxon and Germanic speaking later on. It’s also interesting that in the area around the town of Menapia in Wexford, spoke a Germanic language of unknown provenance, Yola, up till the 1800′s.
There is also evidence that the Belgae – after they left the Saxon shore in droves, also invaded Western Ireland, and spread Northwards from there. See this book on Belgae in Western Ireland for more information, and a great collection of maps and photos.
June 25th, 2010 at 1:06 pm
In what context does O’Connor use the term Celt? Somoene speaking a celtic language or the traditional temr of iron age invader?
June 27th, 2010 at 2:49 pm
[...] 23andMe research article finally published. Dr. Dan MacArthur offers his take on a new PLoS Genetics paper which was published using 23andMe’s user base. Of course there’s already information coming out of 23andMe’s user community not getting into the academic literature, see this comment below. [...]
June 27th, 2010 at 10:27 pm
Ive seen these studies before, and I don’t understand why they would take from the Swedes instead of looking at Dutch and North German samples. These are the source areas for modern English, and Dutch and Low German are still much closer to English than any other languages. Granted, I would have expected more overlap with the Swedish samples but as you say, there’s some Finnish admixture there as well. It’s almost as if researchers don’t really want to test this idea for fear that their Anglo-Celtic synthesis won’t bear out. Why else would you deliberately ignore the source areas of the Anglo-Saxons, which are very well-documented? I do remember a coming across a fairly early study which claimed that in fact, the English are closer to the Danish than any other group, let’s see if I can find it.
June 28th, 2010 at 7:43 am
Maire,
I agree. Although Eastern Ireland is substantially English by ancestry, there are areas of Ireland where the older population survives, and they weren’t sampled.
Also, of all countries to sample, Sweden would be the worst one, due to Finnic admixture -why not Schleswig-Holstein, Frisian islands and Denmark. Likewise for Scotland, there were 4 nations there originally, Aberdeen would bias the sample towards the Picts only.
However against that, I’ve seen some evidence that the Beowulf may have been based on an epic from Eastern Sweden. Also, what became the Saxons, may have originally spread Westwards from Northern Poland and the Baltic area.
Interestingly, based on 23AndMe data, people who are an admixture of Irish and Northern Polish have very high overall matches with English people?!
Based on recent demography, and 200,000 Poles coming to Ireland, future Irish people will look even more “English”…