Shirley Ann Brown
The Bayeux Tapestry and the Song of Roland
1067, Guy of Amiens described in his Carmen de
Hastingae Proelio how a certain "mimus" rode before the assem-
bled French troops at Hastings and juggled with his sword. The purpose of
this bravado performance was to hearten the French and terrify the English.
An infuriated English knight rode forward to rid the field of this arrogant
intruder, but he was swept from his horse by the lance of "Incisor-ferri"
and, losing his head, became instead the first trophy of the battle.1
sixty years later the incident had assumed a different dimension when
William of Malmesbury indicated that the story of Roland was sung before
the French at Hastings to serve as an example of valour to those who were
about to face a fight that could end only in victory or death.2
It is not
surprising that Wace, writing between 1160 and 1174, combined the two
stories, and that the juggling knight "Taillefer" was said to have had the
Song of Roland
on his lips as he faced the army.
All that this proves, of course, is that the story of Roland's valor and death
at Roncevaux was a popular model of military heroism during the twelfth
century. It is difficult to imagine how the poem, as we have it, could have
been performed while two armies were poised and ready to attack each
Nevertheless, the Song of Roland has been associated with the Normans
and Hastings by many subsequent historians. The references in the poem to
Norman activities in England, Scotland, Italy, and Sicily have tended to
reinforce this association.4
Moreover, the values reflected in the Song of
belong not to the eighth, but to the late eleventh century, as the
actual events and personalities of the original occurrence were trans-
formed to carry contemporary relevance.
On the other hand, the embroidery popularly known as the Bayeux
Tapestry would appear to present a straightforward visualization of the
events leading up to and culminating in the English defeat at Hastings on
October 14,1066. The Tapestry begins with King Edward advising Harold
Godwinson to undertake a journey to Normandy, presumably to reaffirm to
Duke William that he, William, had been promised the English crown if
Edward were to die childless. It now ends with the English defeat at
Hastings, but unfortunately this end is badly damaged, and we have no
340 Olifant/Vol 6, Nos. 3 & 4/Spring & Summer 1979
indication of how much of the fabric has been lost.
The Tapestry has been described as one of the most reliable historical
sources for the years between 1064 and 1066, providing us with an indica-
tion of the political relationships between Normandy and England.5 In a
series of recent papers, I have taken issue with this viewpoint and have
tried to indicate that the Bayeux Tapestry was not meant to be a visual
recording of a sequence of actual events in the order in which they had
occurred. Instead, I think that the narrative of the Tapestry can be under-
stood and appreciated only if it is seen in the context of contemporary
literature, including both historical writings and the developing chansons
I have proposed that the creation of the Bayeux Tapestry was a composite
undertaking and included the employment of a person whose responsi-
bility was to determine the nature and sequence of the narrative which was
to be visualized. It has become apparent to me that this person was aware of
the several accounts of the Norman Invasion which were circulating in the
1070s and that his narrative reflects a deliberate choice taken from these
This person seems to have relied mostly upon French and Norman
sources, altering these when necessary to create a version more favorable to
his purpose. He has carefully modeled sections of his story on the Carmen
de Hastingae Proelio of Guy of Amiens and the Gesta Guillelmi of William
of Poitiers.6 Details were taken from the anonymous Vita Aedwardi7 and
the Gesta Normannorum Ducum of William of Jumièges.8
The Tapestry reflects the character of William of Poitiers's reconstruc-
tion of events by emphasizing the legitimacy of William of Normandy's
claim to the English throne, based upon Edward's designation of William
as his heir and Harold's oath of fealty. Specific instances of borrowing from
the Gesta Guillelmi can be seen in the inclusion of Harold's journey to
Normandy as King Edward's emmissary,9 the Breton campaign against
Conan in which Harold participated,10 the importance attached to Harold's
oath-taking,11 Harold's coronation with Archbishop Stigand emphasized
as celebrant,12 the dispatching of English spies to Normandy after the
crowning,13 the messenger sent to William at Hastings by Rodbert fitz-
Wimarch,14 and William personally delivering the pre-battle harangue to
Brown/The Bayeux Tapestry and the Song of Roland 341
The account of the embroidery also closely follows Guy of Amiens's
poem, which was almost contemporary with the Battle of Hastings. Most of
the battle scenes at Telham Hill correspond with the description of the
fighting in the Carmen, and the depiction of Harold's death appears to be a
direct borrowing from the poem.16 The inclusion of the Count of Boulogne,
a non-Norman, as a close companion-at-arms of Duke William, relies upon
The Vita Aedwardi furnished the scene for the death of Edward the
Confessor, including the details of the people present, their physical
placement, and their emotional reactions.18
The person responsible for the Tapestry may have based much of his
narrative on the stock of literary descriptions of the Norman Invasion to
produce a piece of Norman propaganda, but he was equally aware of other
literary materials. The borders of the Tapestry include a series of Aesopian
fables, and the influence of the chanson de geste can also be discerned.
Although historical writings determined the incidents and details found in
the Bayeux narrative, the basic approach was demonstrably influenced by
the techniques and sentiments found in epic poetry. The chanson which
immediately comes to mind is the Song of Roland, which is generally
accepted as having assumed its present form by the end of the eleventh
One of the fundamental principles expressed in both the Roland and the
Bayeux Tapestry is that of "consilium et auxilium."19 In the poem, none of
the leaders, whether it be Charlemagne, Marsile or Blancandrin, acts with-
out the counsel of his nobles and advisors. In the Tapestry, Harold accepts
the English crown on the advice of the Anglo-Saxon nobles, William
confers with his brother, Odo of Bayeux, before he orders ships to be built
for the invasion of England, and the three brothers, William, Odo, and
Robert of Mortain can be seen in council after the feast at Hastings. In a key
scene, William, Odo and Eustace of Boulogne lead the charge together in
the face of the fleeing younger knights.
Both the Bayeux Tapestry and the Song of Roland emphasize the
strength and binding qualities of ideal feudal loyalty: the treachery and
deception of the traitors in each case was a breach of feudal responsibili-
ties. It is to provide the evidence of a binding agreement between Harold
and Duke William that the first half of the Bayeux Tapestry presents the
story of Harold's visit to Normandy.20 Harold is shown as the trusted
342 Olifant/Vol 6, Nos. 3 & 4/Spring & Summer 1979
emissary from King Edward. He strays into the territory of a hostile Guy of
Ponthieu, from whom William rescues him, as William of Poitiers says, "at
great cost and by threats." Although it is not directly indicated in the
Tapestry, the Norman writings state that Harold was to have reaffirmed
William's succession to the English throne.
The Tapestry then adds a more personal dimension to the relationship
between the two men. Harold accompanies William on a campaign against
the rebellious Conan of Brittany and exhibits remarkable personal courage
and strength by rescuing several men from the sands of the treacherous
river Cousenon. The expedition has a much more successful conclusion in
the Tapestry than William of Poitiers allowed, for Conan is shown handing
the keys of Dinan over to the Duke. In the Gesta, no actual confrontation
took place and the results were indecisive.
This episode, in which the suppression of an unimportant local rebellion
assumes a heroic character, was slanted by the Tapestry "librettist" to
allow the bravery and success of the Norman army to be indicated early in
the narrative and perhaps to foreshadow their later success in England.
Perhaps meant as an example before Harold's eyes of how a rebellious
vassal could be easily subdued by the Duke of Normandy, this endeavor
results directly in the bestowal of arms and armor upon Harold, by William
himself. In other words, Harold, having demonstrated his valor and recog-
nizing the superiority of the Normans, takes upon himself the responsibi-
lities of a vassal. The solemn culmination of the decision is Harold's oath
on the relics at Bayeux. The preposterous conditions of the oath are de-
tailed by William of Poitiers and it was assumed by the Tapestry designer
that anyone looking at the scene would be aware of the implications.
Thus it is the double trust of Edward and William that Harold breaks
when he consents to be crowned King of England. He allows himself to be
swayed by the bad advice of the English thegns and by his own greed when
he fails to refuse the crown. Here "consilium" and "auxilium" go com-
pletely awry. Harold betrays his double feudal bond by breaking the most
sacred tie that can bind two men when he does not ultimately support
William's claim to the English throne.
We are now in the realm of the ideals of feudal relationships and not
dealing with the practices of the "real" world. It becomes obvious that the
Bayeux Tapestry story tampers with actualities to present us with certain
conventions which coincide with literary types. The plot-line is carefully
Brown/The Bayeux Tapestry and the Song of Roland 343
constructed so that the depth of Harold's betrayal is emphasized, and his
coronation, which is accompanied by the omen of the comet, becomes the
turning point of the narrative. The consequences are then unavoidable. His
actions must inevitably bring about the battle which results in his own
dismemberment and carries his family and supporters to destruction.
Within this framework, the outcome is never in question, and the entire
story is raised above the level of simple narration.
C. R. Dodwell has pointed out that the characterizations of Harold God-
winson in the Tapesty and of Ganelon in the Song of Roland are remarkably
parallel. It is perhaps coincidental that they were both brothers-in-law to
their sovereigns. Both are powerful nobles who have spent many years in
faithful service to their kings21 and who are entrusted with important and
dangerous missions in hostile territory. They are both handsome and
knightly in bearing, brave and active in battle, and intrepid when ultimate-
ly faced with death. They have served their lords well, but both betray their
trust as ambassadors and vassals for personal reasons: Harold for a crown
and Ganelon for revenge. Both had to meet death because of their acts:
Ganelon is torn asunder by horses and Harold is hacked to pieces in battle.
The families of both men paid the price for their kinsmen's respective
treasons. In the Bayeux Tapestry, Leofwine and Gyrth, Harold's remaining
brothers, are slaughtered. In fact, the only English casualties singled out for
identification are the three Godwinsons. In neither Harold's nor Ganelon's
case were the personal reasons for their actions allowed to excuse the depth
of their perfidy. Ganelon and Harold are treated as individuals whose
treachery was on a grand scale. They both played for high stakes, and,
failing, both managed to bring down whole armies by their actions. But
unlike the clear portrait of Ganelon in the Song of Roland, the actions of
Harold in the Tapestry have to be augmented by a knowledge of contem-
porary literature to be fully appreciated.
In the Tapestry there is no hint of the belittling criticisms of Harold that
fill the pages of the Carmen and the Gesta Guillelmi. The portrait of
Harold's character found in the writings of Guy of Amiens and William of
Poitiers is far less favorable than that found in the Tapestry; these two
sources seem thus to have been used by the "author" of the Tapestry as a
framework for events rather than for characterization. The French writers
emphasize Harold as the epitome of both moral and military cowardice. It
is to the hero-villain of the French epic and perhaps to the Song of Roland
that one must look to fill in the missing components for the inspiration of
344 Olifant/Vol. 6, Nos. 3 & 4/Spring & Summer 1979
It is almost predictable that there be correspondences between the
images created for Bishop Odo in the Tapestry and Archbishop Turpin in
the Roland. Turpin is not depicted as a prelate in the mould of a Lanfranc or
Anselm but as a feudal vassal who performs his religious duties as part of
his function as a loyal follower. He takes pride in his support of his lord, is a
valued counsellor in time of war, and when leading troops into battle he
personally kills great numbers of the enemy with great relish and fitting
epithets. He willingly dies fighting for his feudal values.
Odo of Bayeux did not perish at Hastings, of course, but the character
given him in the Tapestry is similar to that of Turpin in the Roland. Odo is
twice depicted seated next to his brother William: first when the decision is
made to build the invasion fleet and then when the strategy in England is
being formulated. Although we know that the Bishop of Bayeux contribu-
ted a large number of ships to the fleet, there is no evidence in the historical
writings and documents that he was an actual advisor of the Duke during
the English campaign.22 Recent research has discovered Odo's signature
on early ducal charters and lists of participants at ecclesiastical councils in
Normandy. Apparently, from the beginning of his career, he was allied
with William's ambitions to expand and consolidate ducal power. Imme-
diately after the Conquest he was made Earl of Kent and the most powerful
tenant-in-chief in England. His bishopric remained that of Bayeux in Nor-
mandy, while Lanfranc was made Archbishop of Canterbury. Odo was a
political animal by nature and belonged to the secular Norman Church
which was in the hands of a warrior nobility.23 In the Tapestry, the portray-
al of Odo as William's main counsellor at Hastings is probably a general
reference to his allegiance to his brother—albeit with a slight embroidering
of the facts.
William of Poitiers wrote that two bishops, Odo of Bayeux and Geoffrey
of Coutances, went along to Hastings as spiritual advisors to the invading
force.24 Geoffrey does not appear in the Tapestry. Odo appears four times in
the narrative, and in three of these he is labelled as Episcopus, so that there
is a repeated reference to his ecclesiastical status. But the only instance of
his performing anything remotely resembling a priestly function is when
he says grace at the feast after the forced march to Hastings. He is shown as
much less involved in spiritual matters than was Turpin.
Brown/The Bayeux Tapestry and the Song of Roland 345
Although we can be almost certain that Odo was present at Hastings, we
cannot be certain of the amount of actual physical action in which he
participated, if any. Odo is never described in contemporary sources as a
warrior-bishop and he is not mentioned as taking part in the Battle of
Hastings as described in the Carmen de Hastingae Proelio. There is no
indication that he had received early military training or that he took a
personal hand in suppressing revolts in conquered England. Nevertheless,
the Tapestry shows Odo in the thick of battle, when the young knights were
routed by the rumor that William had been killed. The bishop, brandishing
a club, brings up the rear and helps to regroup the fleeing men. Ahead, the
Duke turns to expose his face as an assurance that he is still alive and
fighting. The Duke's action is described by both Guy of Amiens25 and
William of Poitiers,26 but neither mentions Ode's presence. The decision to
depict Odo actually fighting at Hastings seems to create a deliberate paral-
lel with the heroic Turpin image from the Song of Roland. Perhaps this was
to relate Odo more closely with the popular epic figure.
In epic technique, the foe must be of a status fit to meet the hero, and in all
ways Harold is a fit adversary for William. In the Tapestry, Harold is
consistently depicted as "Dux Anglorum," a title which was not Anglo-
Saxon usage and to which he was not entitled, for he was Earl of Wessex.
This provides Harold with a rank equal to that of the Norman leader, who is
always labelled as "Dux Normannorum." The only element of the epic
lacking in the encounter between the two leaders was personal combat.
William of Poitiers says that indeed Harold had refused the offer of person-
al combat as a means of determining who would be King of England,27 but
no trace of this idea is to be found in the Bayeux Tapestry. No cowardice is
permitted of either William or Harold.
In addition to the emphasis on treachery and betrayal, the characteriza-
tions of the two main figures, and the rule that only equals may be pitted
against each other, there are similarities in descriptive preferences seen
between the Song of Roland and the images of the Bayeux Tapestry. Two
major battles are undertaken in each case, and much attention is given to
the details of the military expeditions. In the embroidery we see the prepar-
ations for war, the building and stocking of the ships, the sailing to Eng-
land, and the manner in which contemporary fighting was engaged. The
leader harangues the soldiers before the fighting, and pennons with flap-
ping tails are carried aloft as the armies go to meet each other. The same
love of action and brutality pervades both the Roland and the Tapestry,
where the borders are littered with the bodies of the fallen and armor is
346 Olifant/Vol. 6, Nos. 3 & 4/Spring & Summer 1979
being stripped from the corpses while the fighting is still going on. Delight
is taken in the excitement and gruesomeness of battle to the death. The
emphasis on fighting and strategy, the valiant last stand of the hero, the
relish in gory detail, and the general excitement are all techniques often
repeated in the chanson de geste.
In both genres, women play a very minor role. None is given an active
part in the Tapestry in which only three appear in the narrative. The only
one who stimulates the imagination is Aelfgyva, who is identified, but
about whom we know absolutely nothing. Queen Edith is at Edward's
deathbed, as described in the Vita Aedwardi. The nude females who
appear in the borders of the Tapestry lack the delicacy of the Roland
women: Aude who gracefully dies upon hearing of Roland's death, and
Bramimonde who becomes a Christian upon the death of her husband
Instead, much more attention is lavished on the horses which were so
essential to the knights' success. While there are only six women to be
found in the entire Bayeux Tapestry, there are 202 horses and mules. Much
delight is found in depicting horses in all poses and having great strength
and courage. Each has the heart of a "Veillantif" and not a few become
casualties in the fighting.
One of the most remarkable aspects of the Bayeux Tapestry is that the
story is told from a determinedly secular viewpoint. There is no indication
of piety on the part of any of the figures. When Harold is shown entering the
church at Bosham, it is probably to indicate the hypocrisy of his devotion,
for the church had become part of the Godwin holdings through trickery.
Bishop Odo of Bayeux says grace before a banquet but otherwise seems
remote from the priestly calling. There is no indication of William's pray-
ers and the relics of which so much is made by William of Poitiers and Guy
of Amiens. There are no angels, saints, or heavenly visions. In its surpris-
ingly pure secularity, the Tapestry surpasses the Song of Roland in which
the heroes pray often to God, the King blesses his men, and angels visit
Charlemagne to instill courage and valor into Christian hearts.
This intense secularity must be deliberate in the Tapestry narrative, for
the literary sources were much more oriented to the idea of God's will. For
instance, William of Poitiers constantly praises Duke William's piety,
saying that the Duke had himself assisted at Mass just before the Battle of
Hastings began, and had worn the Bayeux relics around his neck while
Brown/The Bayeux Tapestry and the Song of Roland 347
fighting.28 Guy of Amiens tells us about the prayers, masses, and pro-
cessions sponsored by the Norman Duke at St-Valéry just before the fleet
sailed for England.29 The Pope was reported to have sent a banner to show
his support for the Norman cause,30 and although some see this papal
gonfanon in several places in the Tapestry, I do not. Nor do I see traces of a
"crusading" aspect in the manner in which the Battle of Hastings is
It is my contention that the Bayeux Tapestry was not meant as a church
decoration and was not meant to be exhibited in Odo's cathedral at Bayeux,
as is popularly accepted. Because of its completely secular character,
instances of bawdiness in the borders, and the many deliberate correspon-
dences with secular literature, I suggest that it was meant to be displayed in
a secular setting. Its physical properties (size and proportions, small visual
details, etc.) would make it best suited for display along the long unbroken
walls of a Great Hall in a palace.32 As a secular object, the Bayeux Tapestry
would have made a striking setting in a space in which the chanson de
geste, or more particularly, the Song of Roland was recited. Both were
meant for the same warrior audience and used many of the same devices,
namely, a story-line which emphasizes the necessity of a just war, ideal
feudal loyalty and its betrayal with the dire consequences, the over-
whelming concern with battles fought at great costs to both sides. Both are
based on actual incidents which are treated rather freely in order to in-
crease the moral message.
The many correspondences with a variety of literary sources were, I feel,
meant to reflect the learning of the person who was responsible for the
invention of the program for the Tapestry, and, indirectly, to indicate the
erudition of its patron. The former was probably one of the young clerics
whom Odo had supported and educated, and this was a marvelous oppor-
tunity to exhibit his skill and appreciation. I think that the consonance
between the Tapestry and the secular song of deeds was deliberate, and
further, that the underlying tone of the Bayeux narrative is related to the
sentiments found in the Song of Roland.
But the question remains. What does the correspondence between the
two works indicate? Did the author of the narrative used in the Bayeux
Tapestry have a direct knowledge of the Song of Roland as we know it, and,
if so, does this indicate that the poem had been formulated before the
Tapestry was made, a date which certainly must fall between 1067 and
1097, or, as I think, between 1083 and 1087?33
348 Olifant/Vol. 6, Nos. 3 & 4/Spring & Summer 1979
The answer to the question must be approached with caution. The
person who created the narrative for the Bayeux Tapestry was undoubtedly
familiar with the Song of Roland through some unidentified tradition,
because of seemingly deliberate correspondences. The alterations he made
to the information found in the other sources he used appear to support this
conclusion. At least the sections of the epic which deal with the Roland-
Ganelon confrontation have their parallels in the Bayeux Tapestry. Even
the possibility that the two works both took form during the 1080's lies
within the realm of plausibility. But because there are not exact text-image
correspondences between the Song of Roland and the Bayeux Tapestry,
proof that the narrative author of the embroidery had specific knowledge of
the early Roland text now known unfortunately still eludes us.
Shirley Ann Brown
York University, Toronto
1The Carmen de Hastingae Proelio of Guy, Bishop of Amiens, edited and tran-
slated by C. Morton and H. Muntz, (Oxford, 1972) vv. 398-409.
2William of Malmesury, (Rolls Series, 1887-89), p. 302.
3Wace, Roman de Rou, edited by H. Andresen, vol. II, p. 348, vv. 8035-8040.
4David Douglas, "The Song of Roland and the Norman Conquest of England,"
French Studies, 14 (April 1960) 2, pp. 99-116.
5David Douglas and George Greenaway, English Historical Documents, 1042-
1189, (New York, 1953), p. 217.
6William of Poitiers, Gesta Guillelmi ducis Normannorum et regis Anglorum,
edited and translated by Raymonde Foreville, (Paris, 1952).
7Vita Aedwardi Regis qui apud Westmonasterium requiescat, edited and trans-
lated by Frank Barlow, (Edinburgh, 1962).
8William of Jumièges, Gesta Normannorum ducum, edited by Jean Marx, (Rouen
and Paris, 1914).
9William of Poitiers, I, 41.
10Ibid., I, 45.
11Ibid., I. 42.
Brown/The Bayeux Tapestry and the Song of Roland 349
13Ibid., II, 4.
14Ibid., II, 10.
15Ibid., II, 15.
16Guy of Amiens, vv. 533-550.
17lbid., vv. 525-526.
18Vita Aedwardi, II, f. 55-57.
19Jessie Crosland, The Old French Epic, (New York, 1951), pp. 4-8.
20Charles R. Dodwell, "The Bayeux Tapestry and the French Secular Epic," The
Burlington Magazine, 108 (November 1966), 764, pp. 549-560. The following is a
development of some of Dodwell's suggestions.
21See Vita Aedwardi, Book I, which emphasizes the rôle of the Godwins in
securing peace and prosperity in Edward's kingdom.
22David R. Bates, "The Character and Career of Odo, Bishop of Bayeux (1049/50-
1095)," Speculum, 50 (January 1975), pp. 1-20.
23David Douglas, "The Norman Episcopate before the Norman Conquest," The
Cambridge Historical Journal, 13 (1957) 2, pp. 101-115.
24William of Poitiers, II, 14.
25Guy of Amiens, vv. 441-464.
26William of Poitiers, II, 17-18.
27Ibid., II, 12-13.
28Ibid., II, 14.
29Guy of Amiens, vv. 52-77; 98-99.
30William of Poitiers, II, 3.
31Charles Gibbs-Smith, The Bayeux Tapestry, (London, 1973), pp. 184,187.
32Because of iconographie reasons which are too involved to explain here but
which will be included in an article on which I am currently working, I have
suggested that the Bayeux Tapestry was commissioned by Odo of Bayeux as a gift to
his brother William, perhaps for the ducal palace at Rouen. I think it was produced
350 Olifant/Vol. 6, Nos. 3 & 4/Spring & Summer 1979
while Odo was in prison in Rouen on charges of treason, and that its purpose was to
remind William of the contribution of his brother to the victory at Hastings with the
aim of a reinstatement into good grace.
33Odo died in Palermo in 1097; he had spent the years 1083-1087 in custody in