[This is the text of a paper I gave at a very good conference in Aarhus in October 2003, which I haven't (yet) written up (the rest of the conference is published here). At some point I might yet work it up into a piece, although I am thinking of using some of it as an illustration in Escaping the Past. A counter to the various counters to my original article on feud (1998) is probably overdue. For now, I hope the points of contact with my previous post are clear - especially in the stitching up of time and the non-ending of narrative.]
Feuds make good stories and it is stories, you might say, that make feuds. I will therefore – as I imagine many of us will – start with a story. It is set in 1073 or 1074 in the Wolds of East Yorkshire, in the aftermath of a feast. Feasts were fine times for violence. The most famous alleged feud of the early Middle Ages – that of Sichar and Chramnesind – apparently originated at a Christmas party. So much for the season of peace on earth and goodwill towards men! The case that concerns me was on the other hand – according to some interpretations – the last recorded episode in a feud. One might expect feasts, with the gathering together of lots of people and the consumption of large amounts of alcohol, to have been the occasion for acts of hot-blooded violence such as begin periods of hostility or feud. ‘From one, an irascible ale-swiller, a man full of wine, a sword’s edge will thrust out the life on the mead bench; previous to that his words will have been too hasty’, as an anonymous Anglo-Saxon poet predicted (note, incidentally the reference to three types of alcohol in a single sentence). Be that as it may, in this case, the gathering together of a family for a feast presented its enemies with a golden opportunity to wipe them all out at once.
The sons and grandsons of a man called Carl had assembled for a feast at their manor of Settrington. One, however, stayed away. He, Sumerled, shared his principal estate with the man who organised the bloodshed in Yorkshire, Earl Waltheof of Bernicia. As the story’s most recent and most eloquent interpreter, Richard Fletcher, says, this is unlikely to be coincidence. Waltheof’s men, possibly coming by sea to the Yorkshire coast and thence riding west to Settrington, fell upon Sumerled’s brothers and nephews and butchered the lot of them. Or almost. One, Cnut, was spared ‘because of his innate goodness’. He was lucky. Few people in early medieval episodes like this were spared just for being nice.
Our primary source for this episode is an anonymous and very brief narrative written not long afterwards, called De Obsessione Dunelmi, ‘On the Siege of Durham’, a hugely misleading title although a siege of Durham does appear in the course of the story. Its author digresses from his tale, actually about the descent of six manors on the borders of Yorkshire and Northumberland, to tell us of the violent story that cumulated in this massacre. The story is well enough known to late Anglo-Saxonists but, for those of you who specialise in other eras, the tale, not very much abbreviated, goes like this:
Once upon a time there was a powerful and energetic earl called Uhtred who saved Durham from the Scots. Uhtred was married three times. His second marriage, to the daughter of one Styr Ulfsson, was contracted on condition that Uhtred kill Styr’s enemy, Thurbrand. Alas, when Uhtred came to swear allegiance to his new king (and old enemy), Cnut, Thurbrand Hold (or Thurbrand the Hold) and the king’s soldiers ambushed him and forty other chief men and killed the lot of them. Uhtred’s brother Eadwulf succeeded him in the earldom but when he died, Uhtred’s son (by his first wife) Ealdred became earl and killed Thurbrand Hold. Thurbrand’s son, Carl, and Ealdred then campaigned against each other until they were prevailed upon to become sworn brothers and go on pilgrimage together to Rome. Unfortunately, the ship was delayed by bad weather so whilst they waited, Carl entertained Ealdred at his home in the East Riding of Yorkshire, in Holderness. One day, whilst showing Ealdred around his estate – you can guess what’s coming – Carl killed Ealdred in (and from now on I quote the source)
a wood called Risewood and still today the place of his murder is marked by a small stone cross. Some time later, the grandson of Earl Ealdred, Earl Waltheof, who was the son of his daughter, sent a large band of young men and avenged the killing of his grandfather with the utmost slaughter. For when the sons of Carl were feasting together at their elder brother’s house at Settrington, not far from York, the men who had been sent caught them unawares and savagely killed them together, except for Cnut whose life they spared because of his innate goodness. Sumerled, who was not there, survives to this day. Having massacred the sons and grandsons of Carl, they returned home bringing with them much booty of various kinds.
Seen from Durham c.1100, this was a series of tit-for-tat killings, such as we might unproblematically call a feud. In the last episode, at Settrington, according to the writer, Waltheof ‘avenged the killing of his grandfather’. This all looks fairly straightforward but, unlike the story of Sichar and Chramnesind, which, although far from straightforward, everyone (except me) sees as a feud, there has been significant debate about whether the conflict (to use a neutral term) between Uhtred’s and Thurbrand’s families can in fact be called a feud at all. I want to contribute to that debate and use recent analyses of the story to highlight some of the problems involved in studying feud in the early Middle Ages.
In my past writings about feud I have expressed profound scepticism about the existence of feud, as we would understand it, in the early medieval west – or most of it anyway. This of course raises the definition of feud, something I expect we will spend much time doing during this conference! I feel strongly that we need a fairly precise – but not over-restrictive – analytical definition of feud, a definition which distinguishes feud from simple violent dispute or potentially violent hostility. It seems to me that much recent discussion of feud runs the risk of dissolving feud into any relationship of hostility, even into any potentially violent dispute. Thus far I suspect I am on solid ground. Where I suspect I may not be, and where I expect to be rigorously questioned by people specialising in periods with better evidence, is in that I think we also need to distinguish feud from vengeance. This point raises what is an absolutely fundamental problem in defining and studying the early medieval ‘feud’ (and I hope you can hear the inverted commas around feud there). The words that meant, or are translated as, ‘feud’ in the early Middle Ages did not mean ‘feud’ as we would understand it. Faida, faehþe and their cognates were generally rendered into Latin as inimicitia, perhaps hostilitas. I hope we can agree that enmity and hostility are not, necessarily, feud.
More to the point, the relationship rendered by these words was a one-way relationship. ‘Feud’, as rendered by these early medieval vernaculars, was not a two-way, reciprocal relationship. It meant the right, by one party, either to avenge, or rather to punish, an affront (actually, as far as I can see, it invariably meant a killing) through a reciprocal act, usually vengeance-killing, without (and this is the key) any fear of a subsequent revenge attack by the initial aggressors; or, probably more commonly, it meant the right to exact a compensation payment in lieu of such an attack, through the threat of such an attack. ‘Bearing the feud’ can mean the obligation to seek redress for a wrong, by one party. At the same time, to ‘bear the feud’ means to bear the threat of vengeance against you, or to bear the cost of the compensation payment. I have seen no use of the word to mean a lasting and reciprocal relationship of violence, and I have found very little evidence indeed of feud in the modern sense of lasting vendetta. Now, we may be happy to call this relationship feud, because early medieval people called it feud. The problem is that in English at least, what people today understand by feud is something rather different. And, more to the point, the phenomenon analysed as feud in other historical periods, or in anthropological fieldwork, is something rather different. In a sense that is the root of the problem and why we need tight conceptual frameworks. Most work on early medieval feud acknowledges the point about contemporary, medieval understandings but then understanding is badly hindered by the fact that modern notions of feud keep bleeding into the analysis.
I should make it clear that I am not sceptical about the historical or anthropological existence of the phenomenon of feud. I don’t adopt an unnecessarily stringent or exacting definition of feud but I have to say that I can find very little evidence that would satisfy even a minimal (if rigorous) definition of feud in the early medieval West.
The principal area of contention in response to my work appears to lie in a discussion of motivation: feud versus ‘politics’. Now, of course, this distinction is crude, neither hard and fast nor very clear – perhaps not even very useful. Nevertheless, it is an important issue, and what is at stake should become clear in the course of my discussion. Some writers, William Kapelle and Matthew Bennett for example, have argued that some violent early medieval encounters were not feuds but simply part of politics. The counter, needless to say, is to argue (as, for example, John Gillingham, Isobel Alfonso, Richard Fletcher, Christopher Morris) that you cannot separate feud from politics. In a sense they are quite right. But in a sense, too, this is not a very satisfying argument if you really want to analyse and understand feud in the early Middle Ages. I would argue that we do need some way of distinguishing feud from violent competition for resources between powerful groups or families that happen to have found themselves upon opposing sides in the past. At least we need to think about it, and conceptualise such differences. Many of my points run straight up against the frustrating lack of evidence in the early Middle Ages. Here I would like guidance from those who work on better documented eras. One lesson I have to underline, however, is that we must work from the (however vaguely) known to the unknown. So much writing about early medieval violence is beset by assumptions, cherished, long-held notions about the nature of the period. But feud as a central facet of early medieval life (as many think it was) has to be demonstrated, not assumed. It won’t do to argue that the fact that we cannot find dramas of repeated tit-for-tat violence running beyond two acts – one initial wrong, and one act of revenge – just because the data is patchy. And it won’t do, either, to argue that the reason some of our famous ‘feuds’ don’t, on close scrutiny, seem to be feuds is because the authors of the accounts of them were churchmen who did not understand the concept of feud.
There are a number of issues, most of which can be illustrated with reference to the story of Uhtred and Thurbrand’s families. The first, possibly the most important, is that of history and narrative. The point that historians shape the traces of the past into a story is often associated with the ‘post-modern’ turn in historiography but it has actually been made since the very earliest days of modern history. Thomas Carlyle said as much in the earlier nineteenth century. To be crude, we choose items from the information that has come down to us and connect them together to make a story. This is, of course, especially relevant to the study of feud, doubly relevant because the point also applies to many of our witnesses. Contemporaries, or near contemporaries like the Durham Anonymous, selected episodes from the past and linked them together to make a single strand of narrative – the story of a feud. But was it a feud? Or was it simply written up as such? The episodes in the Northumbrian feud sound unproblematic, as related by the Durham Anonymous. However in fact there were long gaps between them. If, as is usually assumed, the Thurbrand, enemy of Styr Ulfsson, and Thurbrand Hold are one and the same (no one appears to have wondered why, if they are the same man, the author only explains who Thurbrand Hold was on his second appearance – and one lesson of the story is that name-giving was confusingly unimaginative in eleventh-century Northumbria) then it took ten years for any violence to erupt as a result of Styr’s injunction to Uhtred to kill Thurbrand. Uhtred had remarried in the interim in any case and, furthermore, it was Thurbrand who did the only recorded attacking. Styr and Thurbrand might have been rivals but rivalry is not necessarily feud.
Then there must have elapsed another seven years or more before Ealdred exacted his revenge on Thurbrand. As William Kapelle noted the next episode is interesting in that it is described not as Carl trying to ambush Ealdred, trying to find an occasion to carry out his vengeance killing, but as a period when the two tried to do away with each other. And the period up to Carl’s killing of Ealdred was another long one. Ealdred’s murder dates to 1038, that is twenty-three years after his father’s murder of Ealdred’s father, and at least ten years after Thurbrand’s killing. As noted, the Anonymous proceeds from the death of Ealdred to say simply that ‘Some time later, the grandson of Earl Ealdred, Earl Waltheof, ... avenged the killing of his grandfather with the utmost slaughter.’ What might not be clear from this is that the Settrington massacre, the fourth and final episode of the Northumbrian ‘feud’, took place no less than thirty-six years after Ealdred’s murder. Indeed Carl killed Ealdred four years before Waltheof was even born. Thirty-six years is a long time – most of my lifetime. It took as long to happen in the eleventh century.
So: four acts of murderous violence spread over fifty-eight or fifty-nine years, with probably at least a decade between each of them. It might be argued that this is to limit discussion to the recorded acts of hostility but that is one area where I object that we cannot assume the existence of types of violence. For one thing, we do know about other acts of violence involving these people. The war (let’s call it that) between Ealdred and Carl has been mentioned. Another war is mentioned over the estates in Teesdale. Waltheof, ‘avenger’ of Ealdred, was the son of Siward, who married Ealdred’s daughter, as the Durham Anonymous notes. Now Siward, father of Waltheof, also happens to have killed Eadwulf, half-brother of Ealdred, in 1041. No feud appears to have ensued between Eadwulf’s family and Siward’s. This poses huge problems for the reconstruction of the De Obsessione’s story as a feud. It makes Waltheof switch sides – at best. The Durham Anonymous could say that when Waltheof had the sons of Carl killed at Settrington, he avenged his grandfather. From his point of view, as he told his story, it looked like that. It’s a nice narrative turn. But, firstly, it only looked like Waltheof had avenged his grandfather because the writer chose to forget that Waltheof’s own father had killed Waltheof’s great uncle, the brother of the man he was supposed to be avenging. And secondly, however the Anonymous saw it, it does not have any implications for what motivated Waltheof in 1073/74.
There was other murderous violence. The third son of Uhtred, Gospatric, was, like his father and elder brother, killed at court on the orders of Tostig, Siward’s successor as earl of Northumbria, in 1063. Gospatric’s son Oswulf killed Tostig’s crony Copsig in 1067 but no ‘feud’ between Uhtred’s family and the Godwinesons is mentioned. Oswulf was himself murdered by a robber, later in 1067. More to the point, when Uhtred was killed in 1015 by Thurbrand Hold, he was killed alongside forty other leading men. Why was it only his family that ‘feuded’ with Thurbrand’s? And, finally, note that neither Uhtred’s brother Eadwulf Cudel, nor either of Ealdred’s brothers appear to have done anything at all to avenge their murders. The Anonymous said that Eadwulf Cudel was a lazy cowardly man but not because he didn’t avenge his brother, but because he was afraid of the Scots. There seems to be a lot of selection from the evidence, selection from a sea of violence and killing, going on here in order to make this into a long, unilinear saga of murder and revenge. Like all historians, the Durham Anonymous chose how to emplot his story, and he did so in the tragic mode.
An alternative explanation has been proposed. The admittedly patchy sources for Northumbrian politics at this time allow the suggestion that the families of Thurbrand and Uhtred were on different sides during the turbulent early decades of the eleventh century. Uhtred’s family had control of the earldom of Bamburgh, or Bernicia, and appears to have been loyal to the old English royal dynasty. Thurbrand’s family, on the other hand, seems to be counted amongst the supporters of Cnut, and perhaps at the head of a Danish faction in Yorkshire. The first recorded action in the hostility between the two groups was Thurbrand’s slaughter of Uhtred and other English leaders. However, as has been pointed out, even by proponents of the feud interpretation, to have slain forty-one men must have required a substantial armed force. Indeed the sources say the first massacre was on King Cnut’s orders and that the milites who did the killing were his troops. In the broader English context of the events of 1016, and Charles Insley has correctly pointed out that we cannot see episodes like this in purely local light, Cnut was following a policy of the eradication of the principal supporters of his erstwhile enemy and rival for the kingship, Æthelræd II ‘Unræd’. Siward’s murder of Ealdred’s brother Eadwulf – unavenged, remember – was carried out in similar fashion, and so were other killings at the royal court (like that of the third brother, Gospatric), none of which appears to have provoked feud. Kapelle has suggested that the family of the Bernician earls was in revolt against the crown for a while. Carl on the other hand appears to have been an important royal official in Yorkshire, supporting the Danish royal house. The reconstruction is plausible and gives us a reason to see how the two could have spent some time campaigning against each other even without the blood that lay between them.
Carl’s eventual murder of Ealdred might have been part of politics on the wider stage. Another royal official in York, Siward, after all did away with Ealdred’s brother, apparently on royal orders. But it is equally possible that – at the same time – his actions might have been motivated, and justified, by Ealdred’s killing of Carl’s father. Finally, Waltheof sent his lads to slaughter Carl’s sons and grandsons. It is very unlikely indeed, for reasons touched upon above, that this should be seen as an act of feud. Again, national politics play a part. Waltheof and the sons of Carl had both found themselves in revolt against William I in 1069-70. Waltheof, however, had charmed William into forgiving him and even giving him his father’s earldom. This might not have gone down well with other former rebels. Possibly trouble was brewing again. Either way, it seems unlikely that we have to go back further than 1069 to understand the reasons for the Settrington massacre of 1073/4. Whether, like the Durham Anonymous, Waltheof justified his action as revenge for his grandfather, whether he added it as an afterthought, as icing on the justificatory cake, we shall never know.
The two motives – political expediency (or orders) and personal revenge – are not necessarily clearly separable. That is the essence of the argument against the ‘politics, not feud’ position. It is here, however, that we run into the really intractable issues. Even if we knew how Carl justified his killing of Ealdred, and the cross erected on the site suggests atonement, the past is always made in the present. Individuals select aspects from their knowledge of the past to justify their deeds in the present. Carl and Ealdred had been engaged in some sort of a struggle; Carl’s father had killed Ealdred’s father twenty years before, and ten years ago (let’s say) Ealdred had killed Carl’s father. These were all aspects that would make close friendship between the two unlikely in any society, feuding or otherwise. Yet they had chosen at one point to forget these past events and become sworn brothers. There was a choice. This choice lies at the heart of what Wallace-Hadrill called the ‘dormant feud’. Yet I think that the concept of a dormant feud is, analytically, deeply problematic.
In 1998 I wrote:
‘if, rather than having a continuous state of violence and enmity, we have a series of independent incidents taking place to solve immediate problems and the contingent selection from the past of particular episodes to justify or explain them, it does not seem analytically useful to link them into a spuriously continuous ‘chain’ of events and call that feud.’
In the first, Fletcher draws attention to the cross erected on the site of Ealdred’s murder. Ealdred left only daughters, three of whom, imaginatively, he called Aelfflaed. It was one of these that married Siward. Now, says Fletcher, anthropological and ethnographic as well as historical parallels show that women often play a part in keeping a feuding relationship going, goading their men-folk, producing artefacts associated with the murdered victims, singing songs about the wrongs done, and so on. The daughters, suggests Fletcher, erected the cross to remind their kin of the wrong done, which required vengeance. He uses that as a springboard for a nice discussion of female status and power in eleventh-century England. This is an interesting idea but it is not convincing. Firstly, if it was the daughters of Ealdred that erected the cross (no source says as much) and if it was their intention to keep the feud alive in their menfolk’s mind, it was both spectacularly ill-advised and astonishingly unsuccessful. As mentioned, neither of Ealdred’s brothers is recorded as lifting a finger to avenge him, and nor did either of his nephews, or his brother-in-law, or his son-in-law. If their memory of the feud was dependent on the Risewood cross then this is hardly surprising. It was erected in the middle of their enemy’s estates. One might suggest that it is unlikely that the relatives of Ealdred were permitted to make regular pilgrimages to a cross on the estate of their relative’s killer in order to whip the men of the party up into a vengeful frenzy. It seems to me far more likely that, unless Carl erected the cross in an act of penance, the cross was part of the settlement. Not only would it have cost money – a sort of compensation payment – if it was the sort of elaborate stone monument well enough known from that part of the world. It would also have been a permanent reminder to Carl and his family of his perfidy and wrong-doing. This might have been punishment enough. In any case, if feud there had been up to that point, this seems to have ended it.
Fletcher’s is a good try but unfortunately it doesn’t work. And alas there are precious few better examples of the maintenance of a feuding relationship. The Merovingian Queen Chlothild’s urging her sons to avenge her father by attacking the Burgundians is sometimes cited as an instance but, as Ian Wood has pointed out, there are huge problems with the story, not the least being the thirty years it apparently took Chlothild to realise that her parents needed avenging. Now there are references to similar mechanisms at work in the early Middle Ages, but they all concern hostility between larger groups – warfare between peoples. Frankish washerwomen are said to have sung songs commemorating past wars against the Saxons, for example. Now, you could assume that if these mechanisms were known at that level then they might have been known at other levels of violence in early medieval society. But that would be precisely that: an assumption. There’s no evidence. This is what lies behind the rest of Fletcher’s analysis (like that of many another commentator), an assumption that feud, as we understand it, was central to early medieval life and thought.
Briefly, we need to remember that the old idea that the extended family, or clan, was an essential feature of early medieval social organisation has been under attack for some time. Practical kin-groups were small. Furthermore, in spite of decades of detailed research into the reconstruction of early medieval aristocratic Sippe, there is little or no evidence that they acted as unified groups in politics. They fought each other and changed sides in entirely understandable ways, but ways which render the assumptions about early medieval people’s supposed subjection to the demands of feud entirely questionable.
Fletcher’s other approach, in arguing that the selected and scattered episodes of violence were a bloodfeud, is to argue that medieval texts, laws, show that feuding was a common feature of Anglo-Saxon life. If it shows nothing else, the story of Thurbrand, Uhtred and their descendants shows that these people had a concept of vengeance. But many, probably all, societies have a concept of revenge. To early medieval people God took revenge – ultio – vengeance was, after all, his – but the Lord did not, I suggest, feud. That would suggest an altogether heretical view of the Almighty. And the texts cited as showing the centrality of feuding are, as I have argued before and as I have suggested above, about vengeance, not about feud as in modern understanding.
There is only time for brief points. Essentially, early medieval law permitted vengeance. Even law codes like that of King Edmund of England, permitted vengeance, but, and here they drew upon a tradition that went back to the very last Roman laws in the west. They made precise grants, in specified circumstances, of the state’s right to vengeance. Only exceptionally energetic and powerful rulers believed they could prevent all acts of vindicatory violence and say that all punishment would be done by their officers. The early medieval Latin and vernacular words for feud meant the threat to take revenge. This threat was strategic. It brought in external parties to mediate. Specifically, for much of the period, it brought in the king’s officers. Thus, as I have argued before, ‘feud’ (in early medieval terms) was not a sign of absent state power and nor was its appearance in law a sort of abject admission of this fact, a weary acceptance of the fact that the only way to keep the peace was through a system of deterrence, of mutually assured destruction. Because the system brought in royal officers to adjudicate, it was a means of enhancing royal presence in the localities. This vengeance was structurally and analytically different from feuding. For one thing, it seems to have been the threat of violence that was used strategically to draw attention to the dispute and bring in third parties in an attempt to resolve it. If compensation was not paid and an agreement or consensus was reached that vengeance was justified, then a revenge killing was seen as punishment, and to have ended the dispute. It seems to me that this is importantly different from feud, where killing rarely ends the dispute, except perhaps where the whole of the opposing party is wiped out. Thus to suggest that ideas of feud permeated society, or that feud, as a threat of mutually assured destruction was used to keep the peace in a stateless society, badly misrepresent the early medieval evidence.
The conflict between Uhtred’s family and Thurbrand’s was, to use Paul Hyams’ phrase, feud-like behaviour. There were certainly elements about it that look like feud. They were, however, only parts of a complex mix and probably not the most important ones. Feuds make good stories and the Durham Anonymous wrote this one up into a gripping tale. But a hugely misleading one. He made a selection to emplot his narrative as a tale of feud. The actors in his story might themselves have selected events from the past to justify their deeds. The point is however that they had a choice. When other considerations made violence necessary and actions from the past helped to justify this, they chose these justifications. Most of the time, however, it seems that they chose not to make past violent acts the basis for present or future actions. These people were not imprisoned within a feuding culture, or bound by the demands of feud. To emplot the Northumbrian story as a feud masks the complexity of the situation. More importantly it denies the actors in the story the historical choices that they clearly had and the freedom to act that they evidently enjoyed.