Post-classical history


THE STORY THAT Harold Godwinson was killed at Hastings by an arrow in the eye is part of the historical legacy that every Macaulayan schoolboy is supposed to know. But is it any more than a story and does it deserve credence at any higher level than that of Alfred with the burnt cakes, Canute commanding the sea or Robert the Bruce and the spider? Until recently historians accepted the story of Harold and the arrow in an alarmingly uncritical way, doubtless concurring with the tradition in the spirit of Aristotle of an event which, even if it did not happen, could have happened, might have happened and even should have happened. But history is not drama: it obeys the laws of contingency not necessity; it goes its own stubborn and irreducible way regardless of plausibility. The historian, as opposed to the novelist, playwright or romancer, needs to ask, not is this the way Harold’s life should have ended, or does this lend the battle of Hastings an extra patina of glamour, but only: did it actually happen? As soon as we ask this question, the difficulties inherent in the story begin to proliferate.

There are three main views on the battle of Hastings and the death of Harold. The most credulous is that the Normans won the battle by a feigned retreat of the entire army and that Harold was killed with an arrow in the eye. The middle view, which I espouse, is that the battle was a more complex and nuanced affair than that and that Harold was killed by Norman knights. But it is only fair to mention a third view, what we might call the standpoint of Humean scepticism or, as they say nowadays in the media, ‘nobody knows anything’. A very good example is this statement by the great scholar of eleventh-century England, Professor Frank Barlow: ‘There is no acceptable story of how the Normans won the battle. The simple truth may be that they were still losing it, or at least had achieved no decisive advantage, when, to their surprise, the English fled, Harold having fallen, unrecognized by his foes, in some skirmish.’

The initial problem is one of sources. By far the best source for 1066 is the Anglo-Saxon Chronicle, but this tell us very little about the battle of Hastings other than that the Normans were victorious. Space does not permit one to delve into the provenance and authenticity of manuscripts and other materials, but suffice it to say that there is near-unanimous agreement among students of the battle of Hastings that the prime sources are fivefold: the Gesta Guillelmi Ducis Normannorum et Regis Anglorum of William of Poitiers; the Gesta Normannorum Ducum of William of Jumièges; the Historia Ecclesiastica of Ordericus Vitalis; the Carmen de Hastingae Proelio of Guy, Bishop of Amiens; and the Bayeux Tapestry. These are the only five artefacts composed by contemporaries of the events described.

William of Poitiers’s work is thought to be particularly valuable, for he was chaplain to William the Conqueror. Born about 1028, he did not accompany the Normans on their cross-Channel invasion in 1066 but spoke to those who had fought there when compiling his ‘official biography’ of William (written between 1074 and 1077), which is ill-concealed hagiography. Jumièges’s work, too, which is about Norman history in general and not specifically about William, dates from the 1070s.

The works of William of Poitiers, Jumièges, Ordericus and Guy of Amiens all describe the death of Harold but say nothing whatever about an arrow in the eye, or anywhere else for that matter. Jumièges and Ordericus also introduce a massive complication into the story by saying that Harold died, not as night was falling, as William of Poitiers would have it, but in the morning, during the first Norman assault on the English shield-wall. The first writer to state that Harold was killed by an arrow is Baudri of Bourgeuil, author of a work entitled the Adelae Comitissae, written between 1099 and 1102, but even he says nothing about an arrow in the eye. The first appearance of that story is not until 1125 (i.e. sixty years on from the battle), in William of Malmesbury’s De gestis regum Anglorum, an entertaining but not particularly accurate history of the English kings from Alfred the Great to 1125.

There is general agreement that the Bayeux Tapestry is an important source for the Norman Conquest, though far less unanimity on what that importance consists in. The so-called tapestry is 20 inches high and 230 feet in length and contains seventy-three scenes forming a more or less linear narrative. On a backing of bleached linen embroiderers worked in five colours of wool over a pre-sketched series of tableaux with stem and outline stitches. It is usually considered that the tapestry originally contained a further nine feet of post-Hastings scenes, dealing with William the Conqueror’s entry into London and his coronation. The consensus among scholars is that the tapestry was made by English embroiderers at Canterbury to a commission by Odo of Bayeux, William’s half-brother, to show him and his friend Eustace of Boulogne in the best possible light. It was probably completed in the late 1070s and may even have been ready for the dedication of Bayeux Cathedral in 1077.

It has often been alleged that the Bayeux Tapestry shows clearly that Harold died with an arrow in his eye. Plates 71 and 72 show two figures, one with an arrow in his eye, another being hacked down by a Norman horseman with a sword. The words ‘interfectus est’ (was killed) appear over the figure being scythed down. The common-sense interpretation of this would seem to be that the figure with the arrow in his eye is one of Harold’s last defenders and that Harold is the man falling under the sword. To get round this objection, those who are determined to adhere to the old story that Harold was killed with an arrow in his eye have advanced a number of ingenious arguments. Some of these revolve around arcane theories to do with the stitching of the tapestry: one plausible version is that there may be vestigial stitch marks by the head of the second figure, suggesting that the tapestry has been tampered with and that these allegedly phantom stitches originally represented the shaft of an arrow in the eye of the second, falling figure also. A popular argument of the arrow-in-the-eye faction is that both figures are Harold. To quote a recent apologist for this point of view: ‘It is a sort of cartoon strip representation of him being first hit by an arrow and secondly being finished off by a cavalryman.’

The Carmen de Hastingae Proelio is at the core of the entire argument about the arrow in the eye, for much depends on the value we place on it as a source. The most convincing view is that it was written by Guy, Bishop of Amiens, and is a product of the years 1070–71, thus making it the earliest of all sources and one drawn on by both William of Poitiers and the Bayeux Tapestry. This is a view that has been argued persuasively by the editors of the most scholarly edition of the poem, Catherine Morton and Hope Munz, by the doyen of scholars of eleventh-century England, Professor Frank Barlow, and the distinguished Cambridge academic Elizabeth van Houts. Those of the anti-Carmen school, led by R. H. C. Davis, want to date the poem to about 1100, to claim that the author was not Guy of Amiens, and to make it derivative from William of Poitiers and not vice versa. This might be the place to say that I personally can see no validity in any of their ideas and objections.

The importance of the Carmen is that it does not just, like Poitiers, Jumièges and Ordericus, say nothing about an arrow in the eye; it tells us in detail exactly how Harold died. When the Normans finally broke through the Saxon shield-wall at around 4.30 p.m. on that fateful Saturday, 14 October 1066, some of them dispersed to plunder even before the battle was finally won. William, though, saw Harold and a few housecarls still holding out at the top of Battle Hill. While ordering a charge by his knights on the housecarls, he handpicked a ‘hit’ squad to kill Harold. Although William had previously boasted that he would meet Harold in single combat, he knew from his Brittany campaign in 1064 of Harold’s phenomenal physical strength; baulking at the task, he sent four of his most prominent knights instead. These were Walter Giffard, Guy de Montfort, Hugh of Ponthieu and Eustace of Boulogne. They fell on Harold and overpowered him: one struck him in the breast after piercing his shield; another cut off his head; the third ran him through the belly with a lance; and the fourth cut off his leg (which may well be a euphemism for castration).

Given the barbarity of the age, this seems a very plausible ending for Harold. What do the defenders of the arrow-in-the-eye version have to say about the Carmen version? Mainly, it seems, they try to discredit the story by discrediting the Carmen itself: the principal argument is that the poem cannot be by Guy of Amiens (even though Ordericus Vitalis specifically mentions that a poem sounding very like the Carmen was written by Guy of Amiens) since it contains legendary and allegedly chanson de geste material that must come from a later era. All of the detailed points adduced can be refuted quite easily, but what concerns us here is what is said specifically about Harold’s death. The best thing is to cite a leading arrow-in-the-eye theorist about the Carmen version. Here is R. Allen Brown: ‘Had William, duke of the Normans with only three companions, attacked the heavily defended headquarters of the English army – which is what the alleged exploit amounts to – to kill the king and thereby take the crown, far from being hushed up as Morton and Munz will have it, the feat of arms would have been bruited abroad in every court and chanson in Latin Christendom and beyond. Meanwhile, as it seems to me, the whole improbable incident recorded by the Carmen goes far to condemn that source itself.’

Here is an amusingly tendentious defence of traditional orthodoxy. Note the phrase ‘heavily defended headquarters’ (the last word in particular insinuating a prepared position instead of the tattered remnants of a shield-wall). As is made very plain in theCarmen, in his last moments on earth Harold, far from being ‘heavily defended’, had with him a handful of housecarls who were sustaining crossbow fire and cavalry attacks at the very time William’s quartet beset Harold. William would certainly not have wanted ‘bruited abroad’ the fact that he had shirked single combat with Harold and that it had taken the Normans four picked men to kill the Saxon king. The final sentence is a classic of petitio principii. The incident is only ‘improbable’ in the light of a hidden premise: that theCarmen is a later work, adulterated by mythical accretions from jongleurs and others. The ‘improbability’ of the incident – itself a proposition based on the unacceptability of the Carmen as a major source – is then used circularly to argue for the unacceptability of the Carmen as a source.

Since this is not a learned article about the credentials of the Carmen, I can only add the ‘anecdotal’ point that my own researches and those of others continue to underline the high credibility of the Carmen. To take just one example, in a ground-breaking 1993 article in Mariner’s Mirror, designed to show the difficulties in William’s cross-Channel operation in 1066 from the viewpoint of the professional mariner, Christine and Gerald Grange comment: ‘As we have worked on this study, we have become increasingly convinced of the status of the Carmen as source . . . the author of the Carmen understood the importance of wind force and direction to people whose main way of travel was by water, and the sailor’s obsession with this if he is going out to sea. Above all, theCarmen is not only consistent with other contemporary accounts of the crossing, it is also consistent with the weather conditions which may be expected to have prevailed that summer and gives an insight into Duke William’s mind at a time of considerable stress. Only the Carmen recognizes the danger of a lee shore; only the Carmen describes a weather pattern which would be consistent with a period of unfavourable winds, followed by a weather system which would have brought the required southerly wind; only theCarmen records the moonless night of the crossing. Finally, while both William of Poitiers and the Carmen give an insight into the Duke’s mind during that crucial wait following the disaster of Saint-Valéry, only the Carmen recognizes the depth of despair which he must have felt.’

If the best source for the battle of Hastings tells us in detail how Harold was killed, the next best (William of Poitiers) says nothing about an arrow in the eye, and the two other prime sources (Ordericus and William of Jumièges) say nothing about an arrow, this leaves only the Bayeux Tapestry to be acounted for. As we have seen, some commentators, notably C. H. Gibbs-Smith, have used the Bayeux Tapestry as their principal evidence for arguing that Harold was not killed with an arrow in the eye. The entire argument from the visual evidence of the Bayeux Tapestry is inconclusive, but the argument at this level is anyway misconceived, for it assumes that the tapestry is a work of naturalistic representation, whereas it is very clear that many scenes in the 73-plate sequence are iconographic rather than documentary.

The Bayeux Tapestry becomes more mysterious the more one looks at it. To treat it naturalistically is absurd, for at one point the Norman soldiers on their boat seem to be bigger than the boat itself. Among the many unsolved questions arising from the tapestry are: who is the lady Aelfgifu who seems involved with a sexual scandal with a priest; why do Harold and Edward the Confessor appear more kingly than William himself; why are the English in the shield-wall shown with kite-shaped shields when this would have made it impossible for them to swing their two-headed battle-axes? Most importantly there is the problem of the animal motifs. For most of the tapestry there are top and bottom margins which contain illustrations of birds and mammals, both real and mythical, sometimes linked to fables of the Aesop type, but sometimes more enigmatic. It is assumed that these are cryptic comments on the main action, but the exact hermeneutics of these animal figures has proved elusive. What we can say with certainty is that there is a very clear allegorical slant to the tapestry.

If there are no good sources for the arrow-in-the-eye story, how did it get started in the first place? Here I would tentatively like to advance a threefold argument. The first part has to do with Harald Hardrada, who led a Norwegian invasion of England, which was defeated by Harold Godwinson on 25 September 1066 at Stamford Bridge near York. The orthodox view is that we know far less about Stamford Bridge than about Hastings, but, bearing in mind Professor Barlow’s comment above, it may be that the reverse is the case. There has long been controversy about the precise historical value of the Icelander Snorre Sturlusson’s Heimskringla or Sagas of the Norse Kings. It is true that Snorre often makes silly mistakes of detail, but I have found him to be surprisingly accurate. The main outlines of Snorre’s account of Harald Hardrada’s sojourn in Byzantium from 1034–43 are confirmed in sources on which all Byzantine scholars place a high value, notably the history by Cedrenus and the Chronographia of Michael Psellus. Even more impressive is the history of Hardrada as king of Norway from 1045–1066, where Snorre is confirmed in detail by other saga sources. Snorre is particularly good at making sense of confused sequences of events where the other saga writers veer off into fantasy.

It is usually considered that Snorre Sturlusson’s account of the battle of Stamford Bridge contains confused material unconsciously imported from accounts of Hastings. I think this is a misreading of the evidence. It is just as likely that William of Malmesbury’s account of Hastings contained transmogrified aspects of Stamford Bridge. In this connection Snorre’s statement that Harald Hardrada was killed by an arrow in the windpipe comes to the fore. What more natural than that a twelfth-century chronicler like Malmesbury should have heard the story that in 1066 ‘King Harold’ was killed by an arrow and should then have seen the story apparently confirmed in the Bayeux Tapestry?

The second part of the argument concerns the baffling statements by Ordericus and William of Jumièges that Harold was killed early in the battle of Hastings. We know that Harold’s brother Leofwine was so killed and, again, it is conceivable that he was struck down with an arrow. In the confusion between Harold and Leofwine the arrow motif may have survived to resurface later in bowdlerized form. We need not take seriously the scene in the Bayeux Tapesty showing Leofwine and Gyrth falling together; this is simply the neat symmetry of two brothers falling together beloved of an allegorist. The Carmen makes it clear that Gyrth was slain in a separate encounter, and indeed says it was William who killed him – unlikely from the documentary point of view but understandable ‘compensation’ from a French author who would show William in a bad light over the death of Harold but did not want to be accused of anti-Norman propaganda in general.

My final argument concerns the Bayeux Tapestry. If we accept that this is iconographic rather than naturalistic, we may even allow that the tapestry shows Harold with an arrow in his eye and still maintain comfortably that this is not how he died. It was a staple of medieval lore that kings were often miraculously blinded by an an arrow for offending against the divine order. In the Old Testament Nebuchadnezzar blinds the Jewish king Zedekiah for breaking his oath. Harold’s blinding therefore could either represent in general his punishment for ‘perjury’ in 1064 – when he supposedly took an oath to back William as next king of England when Edward died – or more narrowly divine vengeance for the alleged role of the Godwin family in blinding Edward’s brother Alfred in 1036. Such an interpretation fits the meaning of the tapestry at every point. The similarity of the tapestry images to those in the Roda Bible and other early-eleventh-century sources supports this idea, as does the main motif of the tapestry story: that the English defeat at Hastings was punishment for Harold’s perjury. The moral is an ‘eye for an eye’ in an almost literal sense.

The blinding of Harold as shown in the tapestry – and if this is what is shown – would not be intended to denote an actual event but would be meant to inculcate a moral lesson: that God punishes those who do not keep oaths. My contention is that William of Malmesbury heard stories of a King Harold being killed by an arrow in 1066, misread Harold Godwinson for Harald Hardrada, conflated the death of Leofwine (who may also have been killed with an arrow) at Hastings, then saw the story apparently confirmed on the Bayeux Tapestry. Such was the fons et origo of the story every schoolboy used to know. In following William of Malmesbury’s uncritical lead, later historians have unwittingly embraced the credo of the most unscrupulous kind of journalist: ‘Never let the facts interfere with a good story.’

If you find an error or have any questions, please email us at Thank you!