There is a particular tendency in European history to try and pin down the ethnic statistics of a given area in a given time. Within my own field, the interest is in how much of Anglo-Saxo England was populated by ‘genuine’ Anglo-Saxons? Were there Celtic holdovers? Classicists and Celticists alike wrestle with questions such as ‘What is a Celt, anyway?’ ‘How widely spread across Eurasia were they?’ Tracking the movements of pre-literate peoples is difficult, and it is tempting to combine genetic evidence with archaeological.

I don’t really object to this sort of thing in principle. After all our sources describe ‘Saxons’ or the ‘Jutes’ or what have you, and it is pretty apparent that a ‘Celt’ -whatever that may be- is a very different thing from a ‘Roman’- whatever that may be, particularly in the late Empire. Celtic scholars still quibble (rage is a bit strong a word) over how to define Celticness; should it be linguistically defined? Should it be defined on the basis of material culture? Which groups/artistic techniques are ‘Celtic’? Defining an Anglo-Saxon is almost as difficult. The primary sources list three tribes of people, and there is archaeological evidence for some others, but it is all a bit higgledy piggledy and scholars are only human: we want solid, reliable facts.

Or, at least, as close as we can get to such things.

I sympathise when historians and archaeologists turn to the apparently-reliable ghoul of genetics to identify the ‘ethnicity’ of a particular corpse. Tell us if this is a Welshman, we might say. Or an Anglo-Saxon. Or a Roman rather than a Frank or a Goth or an Alamann or a Saxon or a… wait. What IS the difference between an Alamann and a Saxon, anyway? They are both German, after all. Both spoke similar languages, and seem to have had similar material cultures. What makes genetic evidence so sure-fire?

Obviously we can’t go with linguistic evidence. For one thing, skeletons don’t speak dead languages any better than living ones. For another, languages are memetic, and memes don’t rely on ethnicity to travel. The more important reason is that people subordinated and/or neighbouring another tend to learn the language of the other guy. This is pretty obvious, but it means that a dead guy speaking Greek in Egypt during the reign of Justinian is not necessarily a Greek. The Germanic invaders of Italy adopted Latin as their language, and it has evolved into Italian.

Material culture is similar. The richness of early modern Irish and Anglo-Saxon art seems pretty closely intertwined; pottery techniques can and are mimicked, and so forth. Archaeologists have been long aware of this. The basic problem is that a given corpse may well be clad in the jewellery and robes (and language?) of Ethnic Group A, but ‘genetically’ belongs to Ethnic Group B.

So, we use genetics to prove things one way or the other and hurrah! now we know.

Except. Do we? I am incredibly suspicious of genetic evidence. If a guy wears the clothes, carries the spear, and speaks the language of the Old English, does it matter that much of his DNA is Celtic? Clearly he did not belong to that social group in any meaningful sense. It could mean that he rose above his station, or was a hostage fighting alongside his captors, or any number of situations which would mean that he did belong to his ‘genetic’ group. It could mean that he no longer considered himself part of his ‘genetic’ tribe. Women tend to be taken by invading groups, after all. The colonisers of Iceland stopped off at Ireland to collect some wives- would anyone then claim via genetics that an Icelandic corpse is Irish? It would be silly.

Humans are not our DNA. We are as whatever we identify ourselves. In early medieval France, different peoples retained their own laws, but there was no genetic test. While I am sure there are examples of individuals straddling boundaries, so far as I am aware it was simply based on an individual declaring himself of a given people- or obviously being so, thanks to his parentage. He, or she, would have been identified based on those pesky things I just dismissed for being too flexible: material culture and language.

I have yet to say anything new here. Scholars who like to play with genetic evidence know all this. It is obvious if you think about it; of course genetic evidence must be taken with extreme caution.

Research from the latter half of last year sheds doubt on the very existence of ethnicities. I suspect the journalist may be overstating it slightly, but nonetheless it seems clear that ethnicity has more to do with social boundaries than with genetic disparity:

The boundaries used by individuals to distinguish themselves from members of other ethnic groups are generally cultural, linguistic, economic, religious and political. Heyer and her colleagues confirm the absence of common ancestry in a specific ethnic group; there were on average more differences between members of the same ethnic group than there were between groups.

Ethnic groups are not based on genetics. They are based on human interaction with one another.

Human interactiona do not leave much in the way of evidence. This is inconvenient, and genetics offers us an easy way out, a means of identifying ethnic groups over the long years of history. If this article is followed by others of its kind, then we must discard genetic evidence. At the very least, we must approach such evidence with new scepticism, and be prepared to ask questions of the relevancy to a particular find.

Frankly this can only be a good thing. If ‘Celt’ is a linguistic term as some -as I- think it is, why were we using genetics to find specimens anyway?

[I would also like to point readers to this post at A Corner of Tenth Century Europe as being relevant to the discussion.]