Monday, 18 September 2017

Some Further Discussion of the Article on Bj 581

Since writing my previous blog post, I have been prevented, for a variety of personal reasons, from engaging in any way with the discussions that have raged about this matter on social and news media. I do see this as a blessing in disguise. As I said then, I do not think the complex matters raised in the American Journal of Physical Anthropology article entitled 'A female Viking warrior confirmed by genomics' lend themselves to the reductions demanded by Twitter, or the selection and rewriting that are inevitable when the press come calling for quotes.

Why am I writing again?

Now that I am back in harness, I do however feel it is my duty to come with some kind of response, even if not on Twitter or in the media. This is not least because my blog post has, at the time of writing this, had some ~60K pageviews. This is exactly 15x as many pageviews as my next most viewed blog post and far exceeds any expectations I might have had when writing. Such a reach for a matter which is essentially about the academic minutiae and the conventions of academic discourse certainly deserves public acknowledgement. I will discuss some aspects of this response below.

A further reason for writing, which I would like to but cannot ignore, is that I have been publicly challenged, in the New York Times, no less, by 'Mattias Jakobsson, a geneticist at Uppsala University and a co-author of the paper, adding: “We would like to urge her to send her critique to a peer-reviewed journal.”' The second purpose of this blog post is therefore to explain why I will not be submitting my 'critique' to a peer-reviewed journal and why I think that this is an inappropriate challenge.

The worldwide response

First, the easier question. My blog post was entitled 'Let's Debate Female Viking Warriors Yet Again' and this in itself reveals that my aim in writing was indeed to stimulate debate. This seems to have happened, in spades, and I am delighted that that has happened. On Twitter, and in the responses to my blog, I have generally found the debate to be thoughtful and considered, even when I thought the contributor was misguided, or hadn't really understood what I was saying. Responses have come from both layfolk and academics, from supporters and opponents (though I am sorry when the discussion does turn into a case of 'for' or 'against'). I'm pleased to say it has certainly stayed fairly polite, unlike what I gather some of the Facebook responses have been (mercifully I am not on Facebook), or some of the responses BTL in the popular press as outlined by my colleague Howard Williams on his blog recently. Many of these were perhaps responding to the original article, or how it was presented in the media, rather than to my blog post in particular and thus do not concern me here (though see below on the responsibility of academics in this age of open access). In this way, I feel I have achieved my aim in writing the blog in the first place - the debate has taken place. I am particularly proud of when I became a Twitter Moment (until this week I didn't even know what a Twitter Moment was and am still not very sure) - its headline was Prof adds a grain of salt to the 'female Viking warrior' story. A grain of salt is pretty much how I envisaged my contribution, and not a big bag of sodium chloride.

I would also point out that I have with one exception not censored the comments published on my blog, even though some of them are getting a bit repetitive and some I consider misguided. There was just one response which I chose not to publish simply because, though witty, I thought it had no real relevance to the current debate. That comment section is now closed, though I am for the moment happy to entertain comments on this post here.

The challenge

Although generally polite, many of the responses, from archaeologists and scientists in particular, have been quite firm in declaring me wrong. These commentators have made the following points, among others:
  • I am not a scientist and therefore not qualified to evaluate the science behind the article
  • other than the new scientific results presented in the article, all the information in the Am J Phys Anthropol article was 'pre-established' and therefore no longer a matter for discussion
  • they would believe a 'peer-reviewed article' over a 'blogpost' any day
  • I am out of order to complain about established reference conventions in scientific/archaeological journals
The challenge, as noted above, for me to present my 'critique' in a peer-reviewed journal is misguided and the challenger has I think not read my original blog post carefully enough. I made it pretty clear there that my concern was not with their results, but with the quality of their argument in the interpretation of those results. This poor quality that I think I have identified relates to all of the points raised above:
  • I did not claim to have any opinions about the actual scientific analyses reported in the article and would never do so. My critique was partly about (a) the foundations of and the evidence used in the scientific analyses and (b) about the historical interpretations of the scientific analyses. I think this is clear enough in the blog post and if any readers have not picked that up, they should read it again.
  • the article, despite all of its scientific apparatus, poses an essentially historical question, and frames this question using vague, unexplained and unsupported references to narratives, poetry and historical documents. This means that the article chooses to interpret its scientific results in a historical/literary framework, without having had the courtesy to understand, or correctly cite, the long-standing discussions that have taken place within that historical/literary framework.
  • on peer review, see further below. I would just point out that I was not presenting any counter-argument to the published paper, for people to 'believe', but pointing out what I considered to be deficiencies in the argument of the published paper.
  • I explained in the previous blog post why I did not think that a referencing system designed for short scientific articles was valid when citing books of several hundred pages and stand by what I said there. And is it not a fundamental principle of science that results should be reproducible? This should also apply to the thought processes behind the arguments as well as what happened in the lab.
Peer review

All academics understand that peer review is both necessary and imperfect. I find it particularly ironic that commenters are claiming the superiority of the article because it has been peer-reviewed and attacking me for daring to critique it without the benefit of peer review, because I do not believe that the peer review process at Am J Phys Anthropol has done the authors any favours at all, other than giving them a huge audience for their work.

Forgive me if I have misunderstood, but I assume that the peer reviewer(s) for Am J Phys Anthropol are not well-acquainted with Old Norse literature and Viking Age and medieval Scandinavian history and therefore are unlikely to have picked up on the deficiencies of the article in these areas. I do wonder though why they couldn't at least recognise that the article might have had more force if it had avoided straying into these areas, and simply presented its scientific results for others to interpret. Whether or not a board game indicates an 'officer' is hardly a matter that a physical anthropologist can determine.

More seriously, I am surprised that the peer reviewer did not pick up on the fact that the supposed osteological analyses which these latest genomic analyses are supposed to confirm are not properly referenced in the article. I have already pointed out the fact that the article provides no indication of where these osteological analyses can be checked. Even a Swedish archaeologist generally positive towards the article recognises that it is a bit slim in the information it provides and states the following:
The plan of the grave shows which bones were well preserved. This should be enough to counter the charge that maybe the skeleton currently labelled Bj 581 is not in fact the one found in this weapon grave. This the authors should have written a few sentences about. I take their silence to mean that having already published her arguments about this elsewhere, Kjellström considers the issue uncontroversial.
Kjellström may consider the issue uncontroversial but are we just to believe her? Why couldn't the authors have simply provided a proper reference to where the osteological analyses have been published?

Since writing my critique, I have discovered that there is still some doubt about both the bones themselves and the plan of the grave as published in the article. These doubts have been expressed in a draft response by Fedir Androshchuk. This is clearly a draft and should be taken as such, but at the very least it suggests some caveats which the authors really should have cleared up properly before doing their scientific analyses. Other highly respected Viking Age archaeologists have also expressed doubts about some aspects of the analysis and interpretation. Again, these are in some cases quite specialist doubts which were perhaps not so easily picked up by the anthropological peer reviewer.

I stated in my original blog post that I did not have a considered alternative hypothesis for Bj 581, and may never have. There is therefore nothing to submit to peer review. However, I do feel I am qualified to come with a critique (and once again I repeat myself), not of the 'results' of the investigation, but of the quality of the argument and the nature of its academic discourse. I myself am often asked to peer review articles, books or projects that are primarily in Viking Age archaeology (though usually with some interdisciplinary aspect) and there seem to be plenty of people out there who consider me able to do this. Indeed I have indirectly heard from some such authors that they have respected and appreciated my critiques. It is my strong view that, in this age of open access and public engagement, academics have an even stronger responsibility than before to present the best possible research to the general public as well as to fellow academics. Which brings me to my final point.

Academic responsibility

My colleague Howard Williams, in another one of his blogs on this issue, points out that 'this has become a story about modern identities, and perhaps also about the crisis of academics attempting to be both digital public archaeologists and public intellectuals.' The original article had a very arresting title which overstates the case made in the article itself. The article is open access and was clearly designed for maximum worldwide public impact, as indeed it proved. To my mind this indicates all the more reason for the doubts, caveats and issues of interpretation to be brought to the fore in the discussion and not brushed under the carpet. Precisely because this is an article clearly intended to have maximum public and popular impact, it is entirely appropriate for it to be critiqued, by me and others, in the public domain of social and news media, and not in some peer-reviewed article I may or may not write within the next year or two and have published within the next five or ten. In an era of open access we do a disservice to our readers by leaving out the processes by which we arrive at our conclusions and just feeding them the sensational results. Although a bit of a shot in the dark as to its potential audience, my critique was indeed aimed at those readers of the article who may not have been sufficiently well versed in Viking Studies to see that there were some holes in the argument. I am content that many lay readers (or experienced academic readers in other disciplines) have understood this, but you can't win'em all.

P.S.

The last paragraph was going to be my final point, but there is one more thing worth mentioning. Many of the discussions of the original article, whether or not influenced by or in reaction to my blogpost, have turned on questions of gender fluidity, non-binary genders and similar matters, as for instance in a recent article in the Guardian, quoting Carolyne Larrington, and much of the Twitter and other discussion has turned on this matter. I would just point out that any such assertions still rely very heavily on various kinds of literary evidence, and that these texts should be subject to the same kinds of source criticism as the archaeological evidence. Interpretations of sagas are not set in stone, but in my experience few saga specialists have wanted to engage with archaeologists enough to help them work out what interpretations of these texts are plausible as evidence in conjunction with archaeological evidence when considering the Viking Age. There are many different kinds of relevant texts in Old Norse and other languages, and each genre has its own quirks and characteristics. All this, and the evolving context of literary study, has to be understood before these texts can be automatically transferred into more general historical or archaeological arguments. It's not an easy matter, and it's something I have been thinking about for most of my career, and occasionally expressed my views in writing on. It's also the kind of detailed study that some of my former PhD students have tackled, for example Roderick Dale on the berserkir and Teva Vidal on houses and domestic life. Both have been able to demonstrate the stratigraphy of certain sagas in ways that must please any archaeologist. Let us hope there is more such work forthcoming and that interdisciplinary dialogue, to which most Viking Age archaeologists of my acquaintance pay lip service, truly happens, in contexts which demand less disciplinary constraint than the Am J Phys Anthropol.




Saturday, 9 September 2017

Let's Debate Female Viking Warriors Yet Again

The Viking Twittersphere is currently alive with tweets about a new article with the arresting title of 'A female Viking warrior confirmed by genomics'.* The article concerns the interpretation of a particular burial at Birka, in Sweden. Since I do not think this is a matter which can be dealt with in the space of 140 characters, even when 'threaded', and one or two have asked what I think, then I have decided to put a few thoughts down here while the matter is fresh. I would however urge you not to read this until you have read the article first - it is open access and available to all, following the link above. Don't forget to read the 'Supporting Information' as well. I would also point out that this is a preview of an article not yet fully published.

To put my cards on the table, I will say that I have always thought (and to some extent still do) that the fascination with women warriors, both in popular culture and in academic discourse, is heavily, probably too heavily, influenced by 20th- and 21st-century desires. At the same time, I also think it is interesting to debate these matters and I am happy to do so (although not with the type of people who write UTL words to the effect of 'I just KNOW there were women warriors in the Viking Age'). I try to keep an open mind, but I also get very frustrated by what I consider to be academic discourse that seems to be mostly concerned with grabbing attention in order to facilitate further funding and/or claim 'impact'. And academic discourse in which topics that have been of concern to the humanities for decades if not centuries are suddenly somehow 'confirmed' by those gods, the scientists, without giving sufficient consideration of the 'non-scientific' evidence which inevitably raised the questions in the first place. (And here I wish we used the word 'science' in the same way as the Germans do Wissenschaft, which would make all evidence 'scientific', i.e. subject to reasoned analysis and argument). What I am really interested in is the quality of academic discourse - I'm very sensitive to what I consider to be shortcuts in an argument or sloppy use of evidence. We are all guilty of these at times, but I believe it is one of the functions of academic debate to point these out to each other, which is why I am writing this, even though I know many of the authors of this article and consider them friends.

One more caveat - I am not a scientist (in the English sense) and not qualified to comment on the natural scientific experiments carried out for the article I am about to discuss and their results. But I think I am qualified to discuss the ways in which the results are interpreted, and I am certainly qualified to comment on the way the authors of the article use textual evidence, and also how they interpret more general cultural historical aspects of the period, which is something I have been thinking about for about four decades. Hence my weighing in here.

My approach below is to work through the article, picking up points that I think are relevant to the quality of the argument. I have not yet spent enough time thinking about this particular problem to be able to offer a well-reasoned, holistic counter-interpretation. I am not even sure yet that I think the authors are necessarily wrong, or that it is my job to counter their arguments if I do. But I don't think they make a good case, and I would like to take an opportunity to point out some matters which I would like people to take into consideration before jumping to accept the conclusions of the article. I'm afraid too many people will just read the title of the article and not think about it more before endlessly retweeting it (you know who you are!) or making it go viral on Facebook. So here goes.

(1) I note that while the article has ten authors, they have chosen not to involve any specialist in language or texts, in spite of the fact that the article begins with reference to early medieval 'narratives about fierce female Vikings fighting alongside men', and concludes with a quotation from an Eddic poem in translation. The impression given is that the authors consider that no special expertise is required to handle this kind of evidence unlike bones, or DNA, or archaeological finds. The authors might argue that they cite people who do have such expertise, including myself. I would just point out that their primary reference to my work is to a semi-popular book published 26 years ago. (See also point 6., below). I would have thought they could have made the slight effort required to read what I wrote on the subject of women warriors in a recent monograph (The Viking Diaspora 2015, pp. 104-7), a less popular and more considered work. There (and elsewhere when I have written about such things) I do try to show that women warriors and/or Valkyries and/or shield maidens (they are all often mixed up) are not just 'mythological phenomena' as stated by the authors, but relate to a whole complex of ideas that pervade literature, mythology and ideology, without necessarily providing any direct evidence for women warriors in 'real life', which is what I take the current authors to be interested in. I do wish the authors would engage with these more subtle and complex interpretations, rather than just unthinkingly using texts both as the starting and the finishing point of their argument, without any indication of what narratives they have in mind, or even what kind, or any explanation of why a particular quotation might be relevant. An example of their sloppy thinking is when they claim that 'the material and historical records' both suggest that 'the male sex has been associated with the gender of a warrior identity' (a statement I think I understand, but it sounds awkward). This is to elide the nature of two very different types of evidence and does, in my view, a disservice to what they call 'historical records' (which may or may not be the same as the 'narratives' or 'mythological phenomena' referred to earlier). Needless to say, they do not specify what 'historical records' suggest this (or indeed what 'material records' do the same, whatever they are).

(2) Several times in the article the authors refer to an earlier article by the second-named author (Kjellström 2016)** which appears to be of great importance to their argument because in it she apparently provided 'a full osteological and contextual analysis', 'age and sex estimation results' and 'sex identification and a proper contextualisation' for the burial in question. The scientific analyses of the current article apparently arose out of a desire to confirm (as the title of the article suggests) these earlier results by scientific means. Having followed up the article in question, I can find nothing in it which explains why this osteological and contextual analysis suggests the deceased was a female - it's a rather general article summarising the author's osteological research on a large body of material which may well have included burial Bj 581, but does not say much about this particular burial. Without specifying its details, the earlier article does refer to a 'chamber grave furnished with fine armour and sacrificed horses' for which 'three different osteological examinations all found that the individual was a woman'. I suppose this is the grave under consideration in the most recent article, but interestingly, the author concludes that 'Whether these are not the correct bones for this grave or whether it opens up reinterpretations of weapon graves in Birka, it is too early to say' (the article was originally presented at a conference in 2013, not 2014 as suggested in the current article). This is because of problems arising from the fact that the graves were mainly excavated in the 19th century and there has been a certain amount of confusion regarding where various bags of bones came from. Extraordinarily enough, this is not even mentioned in the current article. It is admittedly covered, though fairly briefly, in the 'Supporting Information' to the current article, but I do think this element of possible doubt is crucial enough to have been mentioned in the main article, which is what most people will read - many will not even be aware of the status or significance of the 'Supporting Information', which contains both tables showing the scientific results and some discursive comments about sex and gender identities in Viking Age graves.

(3) Having concluded, to their own satisfaction, that the deceased in Bj 581 was indeed a female warrior, the authors go on to conclude, with very little discussion or justification, that she was 'a high-ranking officer', based apparently on the fact that the burial contained 'a full set of gaming pieces' which apparently 'indicates knowledge of tactics and strategy'. Another factor which may have led them to this conclusion, though it is not stated explicitly, is the fact that they determined that the individual was 'at least above 30 years of age'. By the end of the article, 'the individual in grave Bj 581 is the first confirmed female high-ranking warrior', because 'the exclusive grave goods and two horses are worthy of an individual with responsibilities concerning strategy and battle tactics'. All this seems to me to move rather quickly from evidence to speculation which is presented as fact.

(4) The authors also note that there were 'No pathological or traumatic injuries' observed on the skeleton. They point out that 'weapon related wounds ... are not common in the inhumation burials at Birka' and elsewhere, so apparently the 'warriors' of these graves were either so good that they were never injured, or perhaps they weren't really 'warriors' at all. According to the authors 'our results caution against sweeping interpretations based on archaeological contexts and preconceptions' - they do not seem to recognise that if they take this principle to its logical conclusion, the interpretation of this and many other graves as 'warrior' graves is thereby called into question. They can't have their cake and eat it too. They also say nothing about whether there was any indication on the bones of the kinds of activities one might expect a warrior to have engaged in, as strenuous physical activity might be expected to have left some traces, particularly if they were good enough to avoid injury to themselves.

(5) Although the authors point out that 'previous arguments have ... neglected intersectional perspectives' they do not really pursue alternative explanations regarding Bj 581 either. Was it possible, for example, for a biological woman to have been buried with a full 'warrior' accoutrement, even if she had not been a warrior in life? After all, archaeologists are always cautioning us that 'the dead don't bury themselves' and they often seem not to like interpretations in which the deceased's grave goods are taken as representing their roles in life. But such perspectives do not seem to be applied here - they want the woman to be a warrior, so the scientific analysis makes her a woman and her 'archaeological context' makes her a warrior. No doubt other explanations are possible, still assuming that the bones have been correctly assigned to the grave-goods, but discussion of such alternatives would rather detract from that arresting title, and would probably have ruled out publication in the American Journal of Physical Anthropology. The authors might have been better advised to keep this article to the purely scientific data, and leave the interpretation of it to other contexts which might have given them more space to reason more carefully.

(6) Finally, a bit of a rant against the prevalence of short name+date references in scientific and archaeological articles. Reference to an article in a scientific journal in this way is OK when the article is only a few pages long, as they often are. But referring to a 230+-page book as Name Date is cheating. The interested reader who may want to follow up the point being 'supported' by such a reference is faced with having possibly to read the whole book, or to work out from the index which of several possible sections of the book contain the information on which the referring authors rely. And one does sometimes get the impression that authors using such a reference system have not really read the work in question, at least not carefully or thoughtfully.

These are some of my caveats which I would dearly love people to take into account before tweeting all over the world about women warriors in the Viking Age. It's too easy to take the title of an article at face value and send it round the Twittersphere without further thought. I do know I'm banging my head against a brick wall, since I have blogged, spoken and written about these matters before and have come to realise that the emotional lure of the woman warrior, especially in the Viking Age, is too strong for reasoned argument.

Nevertheless, I am still happy to engage in this debate. And just in case there is any doubt, although this blog is ostensibly anonymous, my name is Judith Jesch and I am happy to acknowledge what I have written above - with this kind of direct critique of an article by people I know well, anonymity would be completely unethical. I did consider sending this piece to https://theconversation.com/uk so as not to be anonymous, but previous experience with them suggests that long and complex pieces don't really work there. Taking complex research to the general public inevitably involves a loss of complexity. But it shouldn't do in an academic journal, and it is in the end the academic arguments I am most concerned with. I do also like trying to explain complex academic arguments to those who don't normally engage with them, but that's another story.

* Hedenstierna-Jonson C, Kjellström A, Zachrisson T, et al. A female Viking warrior confirmed by genomics. Am J Phys Anthropol. 2017;00:1-8. https://doi.org/10.1002/ajpa.23308.

**Kjellström A, (2016) People in transition: Life in the Malaren Vallye from an Osteological Perspectve. In V. Turner (Ed.), Shetland and the Viking World. Papers from the Proceedings of the 17th Viking Congress 2013 (pp. 197-202). Lerwick: Shetland Amenity Trust.