The existing accounts pose many problems, one of the most important of which is the degree of authority of the sources that have survived, combined with the impossibility of knowing what has been lost, and what light it might have cast on what we now have. The source closest in time to the battle and most generally relied on by historians (William of Poitiers) has survived in a single seventeenth-century text, with material missing at both ends, which was in turn printed from an original, now lost. This gives some idea of the scale of the problem; who knows how corrupt the text may have become between the eleventh and the seventeenth centuries? In addition to the depredations of time there is the silence imposed by political censorship in the period immediately following the conquest. It would not have been prudent for any English chronicler to write down his true feelings about the events of 1066 for many decades after them (although in fact one or two took the risk, as the Waltham Chronicle indicates), nor for any Norman chronicler to write about them other than sycophantically, as most of them did.
None of the surviving sources was written by an eyewitness of the battle. This is not necessarily a disadvantage. Even in periods much nearer our own day, it has generally proved impossible for those taking part in a battle to have more than a very partial, confused impression of what went on. A dispassionate overview could hardly be expected from anyone below the rank of commander-in-chief, and even he could speak only of the actions of his own side. He could only guess at the imperatives that underlay the strategy of his opponent. William of Poitiers (see d. below) excuses himself from mentioning the deeds of all the Normans who took part by making the very fair point that even an eyewitness could hardly have followed everything. This means that all the sources that have survived depend, directly or indirectly, on memory. The memories of two bystanders who witness a road accident are likely to differ, even a very short time after the event; how much more unreliable must be the memories of those who had taken part in so complicated and agonizing an event as a battle lasting many hours.
Some of the sources listed below were written so long after the event that in a different context one might hesitate to trust them at all. In general, testimony as late as that, for example, of William of Malmesbury (see j. below) might not be taken very seriously. But we do not know on what sources William based his account, and that goes for many others of similar date. They might have been good, they might have been bad. In William’s case, his version is corroborated in enough cases to make one hesitate to discard his testimony on uncorroborated details too rashly.
With these provisos, let us look at the most important of what has survived more or less in order of date, in so far as any accurate dates can be established.
a. The Anglo-Saxon Chronicle The Chronicle, which was reputedly founded by King Alfred towards the end of the ninth century and is written in Old English, was maintained in various monasteries in England, the different surviving versions being distinguished now by letters. The only versions that were still being kept up in 1066 were at Abingdon (C); Worcester (D – this is generally known as the northern version since at that time the episcopal sees of Worcester and York were closely linked but it was more probably written at Worcester); and Peterborough (E), which was probably written at Canterbury, at least during the period with which we are concerned here, and reveals much more detailed information of what was going on in the south of England. It is a feature of the Chronicle generally that different versions tended to include material relevant to the place where they were being maintained that is not found in others, and that each has its own particular bias. E, for example, is notably supportive of the Godwin family; C, by contrast, is anti-Godwinist. Each at various points has irritating silences at points of interest; other sources may tell us that important events happened during these silences, and we can only guess at the reasons why the Chronicle does not notice them. When a version was continued after the conquest, as E was, it may be assumed that some care had to be taken over what was or was not recorded. Thus, it is never safe to assume that, when nothing is recorded in the Chronicle, nothing happened. Examples could be adduced to prove that this was not the case. Of the versions that were being kept up in 1066, E survives only in a twelfth-century copy. C, perhaps surprisingly, does not mention Hastings at all, but ends with the battle of Stamford Bridge, the entry being followed by a short note in a different, later and much less educated hand on the valour of the lone Norwegian who held the bridge while Hardrada’s army got itself into fighting order. The entry in the D version has been quoted in the Introduction. E, after a brief description of Stamford Bridge, continues:
In the meantime Earl William came by sea to Hastings on Michaelmass day and Harold came from the north and fought with him before all his host had arrived, and there he fell, and his two brothers, Gyrth and Leofwine, and William overran this land.
We have no way of knowing when the entries in any of the three versions were actually written; it could have been the following month, the following year or several years later. It is possible that the 1066 entry in the D Chronicle was written fairly soon after the year’s end since it finishes after the account of William’s coronation and return to Normandy with the enigmatic words ‘May the end be good when God wills it’.
b. Carmen de Hastingae Proelio Attributed (though this has been hotly disputed) to Bishop Guy of Amiens, this is believed to have been written around 1067 (the date too has been disputed) and certainly not later, if the attribution is correct, than Guy’s death in 1074 or 1075. This verse narrative is the most controversial source of all. The attribution is based on the twelfth-century statement of Orderic Vitalis (see h. below) that Guy was the author of a poem in the style of Virgil and Statius describing the battle, and on words in the opening of the poem, ‘L. W. salutat’, which have been interpreted as ‘Wido [i.e. Guy] greets Lanfranc’ – Lanfranc, Abbot of St Stephen’s, Caen, was appointed in 1070 Archbishop of Canterbury. However, this is slender evidence and the claims of theCarmen to be a contemporary record have been convincingly challenged by R. H. C. Daviscxviii who places its date about the middle or second half of the twelfth century. One of his most persuasive arguments is that details in the Carmen not also found in the chronicle of William of Poitiers (see d. below) to which it has marked similarities, such as the legend of Taillefer, are usually found only in later sources such as Henry of Huntingdon and Wace. Another is the sheer incredibility of this version of the death of Harold at the hands of four knights, led by Duke William, which certainly bears all the hallmarks of embroidery on the original story well after the time when all the alleged participants were dead. The elements of the Carmen that seem more credible could all have been derived from earlier accounts such as William of Poitiers’. Davis’s arguments, however, have been equally convincingly disputed by E. M. C. van Houts,cxix who supports the early date of the poem and its association with the bishop. To some extent, the jury is still out on this one. There may, as Orderic Vitalis says, have been a poem on the battle by Bishop Guy of Amiens, and this may be it, but there is some doubt. Since there is a reasonable case for accepting that it is by the bishop, it has been retained in the list of sources, though without accepting its reliability on any uncorroborated points of importance. Guy of Amiens was not a Norman, and one might therefore expect to look to him for a more impartial account of events than might have come from a Norman chronicler; but the bishop was a native of Ponthieu, a county that had been under Norman rule for some years by 1066; his nephew Guy was the Count of Ponthieu who captured Harold on his shipwreck and handed him over to William. His younger nephew, Hugh of Ponthieu, brother of Guy, took part in the battle where his deeds and gallantry lose nothing in his uncle’s account. Hugh was certainly not the only participant in the action whom the bishop would have known, so there is no doubt that he would have had eyewitness accounts to draw on. Assuming that he is indeed the author, the questions are to what extent he felt himself bound by strict accuracy in a poem clearly intended for entertainment, in which he might well have felt that he had a greater degree of artistic licence than in a sober prose chronicle, and how far we can distinguish fact from fiction in his work. He starts his narration at St. Valéry to which William had moved his fleet before the invasion, and where he was weather-bound by contrary winds. He follows him across the Channel, reports the exchange of embassies with the king and then gives a fairly detailed description of the battle, in which he describes many natural features of the field very much as they are known to have been (he speaks, for example, of a hill and a valley on Harold’s side of the field, and of land too rough to be tilled), though his sequence of events is rather confused. He reports, additionally, the fact that the English stood in such dense formation that the dead could not fall, and he also reports the retreats, feigned or real, of the Normans that lured some of the English out of their defensive lines. Finally, he tells of the four knights, Duke William, Guy’s nephew Hugh of Ponthieu, Eustace of Boulogne (brother-in-law of King Edward and another kinsman of the author) and a knight called Giffard (or Gilfard), who cut King Harold down at the foot of his standard. It is this report, above all, that discredits the poem as a reliable account of the battle. Against the improbability of this version of the king’s death must be set the fact that, if Guy were the author, he knew at least three of these knights well; all survived the battle, and would thus have had the opportunity of hearing and if necessary, challenging his version of their deeds. The account then follows William to London and to his coronation on Christmas Day. What he does not cover is any of the period preceding Harold’s accession, though he refers in passing to Harold’s oath to William. He eulogizes William throughout, comparing him to Caesar in the opening lines. His description of the ravaging of the country around Hastings after William’s landing demonstrates his general attitude and tone:
Guarding the shore and fearing to lose your ships, you protect them with ramparts and pitch camp there. You repair the remnants of earlier fortifications and set guards to protect them. With peace, indeed, but little ground acquired, your men go out and devastate and burn the land – behaviour which, since the stupid people reject you as king, is not to be wondered at. It is entirely just that they should perish and come to naught.
c. Gesta Normannorum Ducum William of Jumièges’ chronicle ends in 1069, from which it has been assumed that it would have been completed by about 1071. It appears to have been a popular work, for many manuscripts of it have survived. It is a history of the Dukes of Normandy mainly based on an earlier work, but with a final section on the conquest; it is dedicated to the duke and, as might be expected, is openly partisan. The author narrates the story of Edward the Confessor’s undertaking that William should be his successor, the story of Harold’s oath and the breaking of it, and the appearance of Halley’s comet. He describes the assembly of William’s army and fleet (3,000 vessels), the crossing to Pevensey, the battle and its aftermath, and the duke’s coronation in London. His account has been somewhat discredited by his apparent statement that Harold fell in the first attack, which would be difficult to believe, since it would make nonsense of the generally accepted version that the English only broke, late in the day, on hearing of his death; but this peculiarity is generally accepted to be the result of error in an early translation of the text.
d. Gesta Guillelmi Ducis Normannorum William of Poitiers wrote this chronicle from 1073 to 1074. His military experience gives some added authority to his references to military matters, but he is shamelessly sycophantic of the duke. Unlike the work of William of Jumièges, of which there are many manuscripts, only one version (and that incomplete) has survived, beginning with the death of Cnut in 1035 and ending in 1067; but what is assumed to be part at least of the missing section is supplied by Orderic Vitalis (see h. below) who based his own Historia Ecclesiastica on it and tells us that it originally ended in 1071. Other evidence shows many of William’s statements to be slanted or simply wrong, causing the reader to regard with some suspicion those on which we have, unfortunately, no cross-check, such as Duke William’s dealings with the Pope in 1066. His knowledge of classical literature enabled him to move easily among the works of Virgil, Lucan, Sallust and Caesar, selecting the historical episodes with which William’s exploits could advantageously be compared, but also arousing the suspicion that certain of these could have been manufactured or adjusted to allow such comparisons to be cited. There are clear points of correspondence between his work and the Carmen; one of them has undoubtedly borrowed from the other, unless both used the same sources, oral or written, now lost. The tone throughout has been described by a partisan of the Norman side as displaying ‘the arrogance of success and the brutality of triumph’.
e. Vita Ædwardi Regis This is a biography of King Edward the Confessor by an anonymous cleric, possibly Flemish or Lotharingian and possibly a monk of St Omer (he refers to the English as ‘that race’ but shows little sympathy for Edward’s Norman friends), written at the instance of his queen Edith, daughter of Earl Godwin, to celebrate the achievements of the house of Godwin. This stops short at the king’s death, before Harold’s short reign and the invasion itself, though a degree of foreknowledge of future events indicates that it was concluded after the conquest. It is primarily of interest here for information about the Godwin family, in particular the characters of two of the main players, Harold and Tostig, whom the writer probably knew and whom he describes with some shrewdness. It must be assumed that for much of the content he was dependent on Edith, who commissioned it, which may explain its decidedly pro-Tostig stance as he appears from it to have been her favourite brother. Its date is uncertain; it was probably finished before Edith’s death (1075), but what is not certain is whether the first and most interesting part was written before the battle, the outcome of which is hinted at but not actually mentioned. It may all be post-1066. The manuscript itself has been dated c.1100.
f. Historia Novorum in Anglia Written between c.1095 and 1123 by Eadmer, a monk in the abbey of Christ Church, Canterbury, most of this history concerns later events than Hastings. Eadmer does not describe the actual battle, but the short account of Harold’s visit to Normandy and the subsequent conquest is illuminating, especially when compared with the account of, for example, William of Poitiers. Like almost all the English chroniclers, he sees the English defeat as God’s retribution for Harold’s broken oath. His version is the nearest we can get to an English perspective after the Anglo-Saxon Chronicle; but Christ Church was a notably pro-Godwinist establishment. On the other hand, it is possible that, in Bishop Æthelric, who had been a monk of Christ Church and whom he certainly consulted over his life of St Dunstan, he had access to the memories of a kinsman of Earl Godwin and may therefore have had information about his sons not known to other writers (see here above).
g. The Chronicle of Florence of Worcester (sometimes known as the Chronicle of John of Worcester) This forms part of a continuation of the world chronicle of Marianus Scotus (d. 1082) but its most valuable part is the section on English history from 450 to 1140. Based largely on a version of the Anglo-Saxon Chronicle, now lost, he adds information that supplements surviving versions that, especially when it concerns events of relevance to Worcester, cannot be easily disregarded, although his work, which must have been concluded after 1140, was written later than other sources. It is, for example, on his authority that it is accepted that Harold was crowned by Archbishop Ealdred of York; since Ealdred had formerly been Bishop of Worcester, and was succeeded there by Bishop Wulfstan under whom John worked (and who was almost certainly present at the ceremony), this is better authority than any of the Norman sources, which assert that the ceremony was performed by Archbishop Stigand of Canterbury, then under papal interdict.
h. Historia Ecclesiastica Orderic Vitalis (c.1123–24), a monk of half-English, half-Norman parentage, did not write until the twelfth century, like Florence of Worcester. A Benedictine monk born in Shrewsbury of an English mother and a Norman father, he was sent by the latter in 1085 to the monastery of St Evroult in Normandy, where he seems to have spent the rest of his life. His work was intended, like Bede’s, to be concerned with church affairs; his description of the Norman Conquest (in book 3) is a distraction from this objective. His aim, he said, was to tell the simple truth, impartial between English and Normans, and he does go some way to correct the bias of William of Jumièges and William of Poitiers on whose work he relied.
i. Roman de Rou Maistre Wace wrote about a century after the event (c.1160–70), in Norman French, and has therefore generally been dismissed as unreliable, but he spent much of his life in the neighbourhood of Caen in the company of those who could give him first-hand accounts; his estimate of the number of ships (696) taken by William to England, for example, given to him, as he tells us, by his father, is now regarded as more likely to be accurate than the more inflated figures in some of the earlier accounts, for example William of Jumièges; on the other hand, he is clearly writing for entertainment rather than as a sober historian, and it is unwise to accept his version of events unless they can be corroborated elsewhere. He is less likely to be reliable on purely English events for which he must have been dependent on hearsay and legend. He includes the story of Harold’s brother, Gyrth, attempting to persuade Harold to let him lead the English forces, since he had sworn no oath, and also the story that the envoy sent by William to Harold before the battle threatened the English with excommunication. In his version of Harold’s visit to Normandy, however, he gives the versions of William of Jumièges and William of Poitiers, but also the version of Eadmer, and says frankly that he does not know which is truthful. He may well be correct in many of his statements, possibly he often is; but we have no check.
j. Gesta Regum William of Malmesbury composed his chronicle about 1125 apparently at the request of Henry I’s queen (the great-granddaughter of King Edmund Ironside and the last survivor of the old royal family of Wessex). William had at least some of the instincts of a true historian; he compares his sources and points out contradictions in the evidence. But we have no means of knowing in many cases how authoritative some of his sources were. Though he is thought to have had some English blood, he writes avowedly as a Norman (‘to whom [i.e. Normans] I am strongly bound, both by my descent and for the advantages I enjoy’) and, though he regards Harold as a usurper, as his Norman loyalties oblige him to do, he is very divided in his account of him, telling us that Harold would have governed the kingdom with prudence and with courage, had he come by it lawfully. He is also the main source of the story that Harold did not share out the booty from Stamford Bridge before marching south, which caused many men to desert from him. This occurs also in Geffrei Gaimar’s L’Estoire des Engleis (1136–37), who may have taken it from William of Malmesbury or (since the two books were written very close together in terms of dates) may have had a different source; but Gaimar’s tendency to the creative embroidery of his material inspires little confidence in him as an independent witness.
k. King Harald’s Saga in Heimskringla Snorre Sturlason’s saga includes an account of the battle of Stamford Bridge that immediately preceded Hastings. Snorre was writing two centuries after the event, relying on the songs and legends of the court poets who were his predecessors, and has been generally dismissed as an unreliable source; but it has been pointed out that the most striking inaccuracies in his account (for example, English names and family relationships) are things of which an Icelandic writer might reasonably be ignorant, while his account of the battle of Stamford Bridge and the English tactics at it (other than the English use of cavalry in the battle) have in some details been corroborated elsewhere. The chief problem with his account is that, by the time he wrote, in the thirteenth century, the legends that had attached themselves to both Stamford Bridge and Hastings a fortnight later had become intertwined and confused and it is difficult now to untangle them. The one point that should be remembered is that the Norwegian versions of the battle, on which Snorre built his saga, derive from the memories and stories of the Norwegian survivors who took Hardrada’s body back with them, and therefore have a validity independent of the later English sources.
l. Bayeux Tapestry This is the most extraordinary source of all: a piece of embroidery that provides a cartoon history of the Norman Conquest from roughly 1054 to the closing hours of the battle (the final panels are missing but probably originally concluded with William’s coronation). For many years this was thought to have been commissioned by Odo, Bishop of Bayeux and half-brother of William, and this is generally accepted although it has been suggested that Count Eustace of Boulogne (who is certainly shown as playing a disproportionately large part in the battle) may have been responsible for it. It seems to have been designed to tell the story from the Norman perspective and it mainly corroborates the versions given by William of Jumièges and William of Poitiers. But it is almost certainly of English workmanship and was probably executed at Canterbury, a noted centre of English embroidery and a place with which Odo, as Earl of Kent, had close connections; and in someways the subtext of the story it tells is strangely ambivalent, as has been seen. It is noticeable, for example, that though William’s very spurious claim to the English crown depended on his inheriting it directly and immediately from Edward the Confessor, so that it was essential to it that Harold should always be portrayed as a usurper,Harold is always described in the Tapestry after the death of Edward as ‘Harold Rex’ and is, on the whole, shown sympathetically and with dignity. In the Domesday Book, on the contrary, he is referred to exclusively (apart from one or two oversights) as ‘Earl Harold’. It is interesting that he makes more appearances in the Tapestry than William: the story it tells is, in a very real sense, the Aristotelian tragedy of Harold, of the great man brought down by a fatal flaw. We know that the Tapestry was exhibited annually in the cathedral of Bayeux to celebrate the Feast of the Relics, assumed to be the relics on which Harold swore; this accounts for the scene of the oath being shown on the Tapestry as Bayeux, although William of Poitiers places it at Bonneville. Thus the Tapestry may be read as a religious lesson to the illiterate faithful on the fatal consequences of perjury. It may also be seen as the work of a designer who did not see the issue in quite such black and white terms as his patron.
It is vital to emphasize that we actually do not know precisely when the Tapestry was worked; a good case has been made for its origin in the years before the fall and disgrace of Bishop Odo in 1082. This is important, if only because of the tendency of historians and other writers (among whom I include myself) to use details in the Tapestry as evidence for and confirmation of details in written sources, when in fact, if it were of a later date, it might have been influenced by them. If, as a purely hypothetical example, the Tapestry had been the work of the embroiderer of the hanging depicting the battle that, according to Baudri de Bourgeuil (see below), hung in the chamber of Duke William’s daughter, or if both had been influenced by a common source, now lost, it would hardly be surprising that it appears to show Harold being killed by an arrow, since Baudri is one of the few sources to give this detail; if it had a totally independent origin, both it and Baudri would have much more credibility and would reinforce each other’s evidence. But, aswith so much else, we don’t know when it was worked or by whom; we can only guess.
There are other sources that mention the battle, mostly only incidentally, but that give details not provided elsewhere. The Latin poem of Baudri, Abbot of Bourgeuil, Adelae Comitissae (c.1100), addressed to William’s daughter, speaks of a hanging in her chamber depicting her father’s victory over the English and asserts that Harold was killed by an arrow; his account is corroborated in a history of the Normans (c.1080) by Amatus of Montecassino, who says that Harold was hit in the eye by an arrow. On the other hand, it has been suggested that Baudri’s hanging is an imaginary re-creation of the Bayeux Tapestry, which he might have seen at Bayeux or elsewhere; if this were true, his corroboration of the king being struck by an arrow would be valueless. There are also the later accounts of Henry of Huntingdon (Historia Anglorum, c.1123–33) in which the battle is covered fairly briefly, and the Chronicle of Battle Abbey (probably late twelfth century), both of which preserve fragments of information that may be authentic among much that is almost certainly not. The Battle Abbey chronicler, for example, is the only source to give the information that the Normans first spotted the English army emerging from the woods when they paused at Hedgland on Telham Hill to arm themselves. This is exactly the kind of local tradition about a local holding that is likely to have been remembered locally and to have been unknown to other chroniclers. On other matters, the Battle Abbey account is of little use. There are, too, the annals of Nieder-Altaich, which mention that in 1066 the Aquitainians fought and defeated the English in a naval battle; whether this is a confused version of Hastings itself, or corroboration of the statement in the E text of the Anglo-Saxon Chronicle that Harold launched a sea attack on William in 1066 shortly before Tostig’s raid on England cannot now be established.
There is one further factor that should be taken into account in the consideration of these sources. All of them, with the exception of Henry of Huntingdon (a cleric but not a monk) and Snorre Sturlason and just possibly the Bayeux Tapestry, are monastic productions. This has important implications. This was the age of the trial by ordeal, and the justice of a cause was proclaimed by the outcome of the trial. All these authors would have believed devoutly in the divine justice, manifested by the outcome of the trial. English misfortunes and defeats are commonly recorded throughout the Anglo-Saxon Chronicle, from Alfred’s wars against the Danes onwards, with the formula ‘as God granted it for the sins of the people’. Harold is reported by many of these historians to have exclaimed the night before the battle that God would judge between William and him. To those who wrote about the event afterwards, on both sides, it would appear clear that He had done so.