Trojan Origins and the Use of the Æneid and Related Sources
in the Old Icelandic Brut

Hélène Tétrel

JEGP, Journal of English and Germanic Philology, Volume 109, Number
4, October 2010, pp. 490-514 (Article)

Published by University of Illinois Press
DOI: 10.1353/egp.2010.0000

For additional information about this article

Access provided by Uniwersytet Warszawski (5 Mar 2014 06:57 GMT)

http://muse.jhu.edu/journals/egp/summary/v109/109.4.tetrel.html

Journal of English and Germanic Philology—October

© 2010 by the Board of Trustees of the University of Illinois

Trojan Origins and the Use of the Æneid and

Related Sources in the Old Icelandic Brut

Hélène Tétrel, Centre de Recherche Bretonne et Celtique,

Université de Bretagne Occidentale

There is no extant Old Icelandic “Saga of Aeneas” that we can use to com-
pare with other medieval romances that draw on the Aeneid. Nevertheless,
medieval Scandinavia was not untouched by Virgilian historiographical
tradition. Though Scandinavian responses seem not to have followed
the allegorizing tradition associated with the Aeneid, they were perfectly
aware of the historical possibilities offered by the tale of Aeneas’s jour-
neys. Indeed, the theory of origins that prevailed in Europe from late
Antiquity to the late Middle Ages was imported to Iceland in early times.
For this very reason, the tale of Aeneas and his descendants, founders of
several empires, became an obligatory chapter in the history of Iceland as
well as in the histories of Great Britain and France. Like their European
counterparts, historians writing in Old Icelandic felt the need to present
their respective Trojan lineages in a synchronic perspective, both because
it helped to establish, indirectly, the Scandinavian branches themselves,
and because it was a common feature in universal histories. Thus, Aeneid-
inspired episodes appear in historical contexts in Old Icelandic literature.
I shall refer in this study, firstly, to a very brief Aeneas section in the Ice-
landic compilation AM 764 4to, which also contains a partial translation
of Geoffrey of Monmouth;

1

secondly, to a similar section in the so-called

Veraldar saga (History of the World), in which there is no obvious knowl-
edge of Geoffrey;

2

and, thirdly, and the main focus of this study, to the

I wish to thank Professor Andrew Wawn for his valuable help at correcting my English.
1. Studied and edited by Svanhildur Óskarsdóttir in “Universal History in Fourteenth-

century Iceland, Studies in AM 764 4to” (unpublished PhD diss., University College London,

2000). The Brut section in AM 764 4to is edited on pp. 269–74 of this work. On fol. 11r,

l. 7 of the manuscript, a short account taken from the Aeneid links the fall of Troy with the

story of Britain’s settlement, but beginning with line 12, the text is interlaced with other

materials and only returns to Brutus on fol. 11v, l. 5, with the sentence: “Her hefr ad segia

af breta konungum” (Here begins the story of the kings of Britain). The text then translates

Geoffrey of Monmouth up to the wars of Casibellanus and then moves back to Roman his-

tory on fol. 12v, l. 40.

2. Veraldar Saga, ed. Jakob Benediktsson, STUAGNL, LXI (Copenhagen: Samfund til

Udgivelse af gammel nordisk Litteratur, 1944). The Troy-Aeneas section can be read on

pp. 44–47; the Aeneas section is found on p. 46.

two extended variants of a shortened Aeneid to be found respectively in
the longer and shorter versions of the Breta sögur (Saga of the Britons).

3

The present study seeks to show that these four variants are connected
and draw on similar sources.
As a consequence of the way in which the Aeneid was imported into
medieval Iceland, the variants of Aeneas’s story which find expression
in its ancient books naturally draw on late-Antique historical renderings
of Virgil’s poem, as well as on the poem itself. The Icelandic texts reflect
two categories of compendia that were based on the same story: the first
category linked Aeneas’s tale to the history of the first kings of Alba, start-
ing with Ascanius’s son Silvius, and from then on to the founders of the
Roman Empire. As well as drawing on Virgil, the late Antique chronicles,
which tell about the “Roman” lineages, include elements borrowed from
several Latin historians, poets, and mythographers but place them in a
general framework based, as was usual at that time, on Eusebius-Jerome’s
Chronicon. The other category of chronicle attached to the branch of
Aeneas that reached Iceland was dedicated to the founders of the British
Empire, starting with Silvius’s son Brutus: these texts are mostly based on
the Historia regum Britanniae but show knowledge of other materials as
well. Lastly, these two categories of chronicle, which were often copied
together in the same cycles, are always found in association with a Trojan
prehistory based on interpolated adaptations of Pseudo-Dares’s account
of the Trojan fall. Thus, it is appropriate to speak of a “Trojan-Roman-
Briton” context and also necessary to connect the different parts of the
cycle to understand how they came to Iceland, with the Aeneid serving as
one chapter in this cycle.
Among the sources on which the Icelandic texts draw for their inter-
polated paragraphs, an important part is played by Ovidian materials
(Ovid and indirect renderings of Ovid) associated with Servian comments
and less easily identifiable glosses. Those borrowings from the gram-
matical tradition might have been transmitted through various accessus
and sooner or later inserted to fit an historical background. Moreover,

3. For the Virgilian section, see Hauksbók, ed. Finnur Jónsson and Eirikur Jónsson (Co-

penhagen: 1892–96), pp. 230–38. Another and longer version of the Breta sögur is being

edited by Jonna Louis-Jensen, to whom I am indebted for her help with the transcription

(any philological mistakes are my own). I shall use manuscript AM 573 4to as a basis for

the longer version, augmented by the paper copy of Ormsbók, Stockholm Papp. nr 58. The

Aeneas section in AM 573 4to is lacunary; it starts on fol. 24r, but the narrative stops on

fol. 25v, with Aeneas’s visit to King Evander and the story of Nisus and Euryales. The text

on fol. 26r starts in mid-sentence. This first lacuna can be filled with Stockholm. Papp. 58,

fol. 138r, l. 6, to fol. 141r, l. 17. A second lacuna in AM 573 4to unfortunately deprives us

of the end of the Aeneas section; fol. 27v ends within the description of Aeneas’s last battle

against Turnus. The text of the Breta sögur only resumes in the Brutus section. Here, again,

the lacuna can be filled with Stockholm Papp. 58, fol. 151v, l. 13, to 152v.

Trojan Origins

491

the vernacular literatures drew little distinction between romance and
history, as far as the Aeneid and the “matter of Rome” are concerned.
As our knowledge of Anglo-Norman and French literature reminds us,
translations of Latin chronicles and romans antiques were kept together
in manuscripts: the widely circulated compilation that Paul Meyer called
the Histoire Ancienne jusqu’à César is one of the most famous examples of
this intricate vernacular tradition.

4

The medieval Breta sögur survive in two somewhat divergent versions.
The two principal manuscripts representing these two versions can be
dated to the fourteenth century. The first one is the work of the Icelandic
historiographer Haukr Erlendsson. Though this compilation, commis-
sioned and partly written by him, has come to us in a somewhat disordered
state, it is easy to see that his Brut was meant to be a part of a more general
history, starting from the destruction of Troy and tracing events down to
the later royal dynasties of Great Britain, eventually linking the latter to
the Norwegian dynasties and Scandinavian history. The other redaction,
which is nowadays represented by a lacunary manuscript and a paper copy
of another lost manuscript, is more detailed in several important aspects of
the story. Its rendering also shows more interest in lovestories and chivalric
episodes, and retains images of pagan gods when they are depicted in the
Latin text; though much damaged, this manuscript includes a fragment of
Valvens þáttr, a tale of Gauvain translated from Chrétien’s Conte du Graal.
Both Old Icelandic versions have extended the Virgilian section, which
can be read at the beginning of the Historia. The passage in Geoffrey’s
chronicle, starting with Aeneas’s travels when the hero left Troy, was merely
an introductory chapter to the Brutus section, which is the real beginning
of the Historia, immediately after the dedicatory epistles.
The aim of this essay is to investigate the relations between the two
versions of the Virgilian opening of the Old Icelandic Brut, their links to
the two other above-mentioned Icelandic texts, their links to the sources,

4. Paul Meyer, “Les premières compilations françaises d’histoire ancienne,” Romania,

14 [1885], 1–79. Jacques Monfrin describes the Aeneid section of the Histoire Ancienne (as

found in BN fr 20125) as follows: “Il reste que la mise en série chronologique et ‘histori-

que’ des romans répond à un besoin très général des esprits (je reste vague à dessein, car

le public est bien difficile à cerner) dans la seconde moitié du XIIe et au début du XIIIe

siècle. L’Histoire Ancienne jusqu’à César. . . est elle aussi une réponse à cette demande, et

une réponse très proche, dans sa structure, de celle qu’offraient les recueils de romans”

(“Les translations vernaculaires de Virgile au Moyen-âge,” in Lectures Mediévales de Virgile,

actes du colloque organisé par l’Ecole Française de Rome [Rome, 1985], pp. 189–249).

Christopher Baswell, Virgil in Medieval England (Cambridge: Cambridge Univ. Press, 1995),

pp. 9–11, distinguishes three categories of the medieval reception of the Aeneid: allegorical,

romance, and pedagogical. Belonging to the second of these, the “romances” find a natural

place in the historical compilations, especially regarding the Brut tradition. The accessus,

grammatical treatises, and commentaries belong to the third category. If this classification is

applied to the Old Icelandic reception, the first category is the only one not represented.

492

Tétrel

and the way they came to be connected to the Historia regum Britanniae in
Iceland. In this enquiry, a comparison with other vernacular works will
be made when relevant. In another paper, I sought to discuss a common
interpolation preserved by both Icelandic texts, in the Arthurian part
of the saga,

5

and to suggest that the longer and shorter versions of the

Breta sögur are based on the same source for the relevant part of the saga.
The present study will follow the same line. Nevertheless, its conclusions
regarding the sources of the Aeneas section relate exclusively to this sec-
tion, in advance of further investigations generating sufficient evidence
to confirm (or invalidate) a more general conclusion. As such, this inves-
tigation is an element in a more extensive exploration of the sources of
the Breta sögur and the reception of the Galfredian chronicle in Iceland.
The state of preservation of the Old Icelandic Brut is unsatisfactory, both
because it is lacunary, and because the oldest manuscripts are presumably
younger than the first Old Icelandic translation of the Historia. Hauksbók
is an Icelandic (or mostly Icelandic) compilation that is preserved as a
whole codex but corresponds to a collection of originally separate quires.
The three parts of the book are now known as AM 371 4to, AM 544 4to,
and AM 675 4to but have been assembled in one codex according to Árni
Magnússon’s own recommendations.

6

AM 371 4to contains the name of its

principal scribe and commissioner, the Icelandic historiographer Haukr
Erlendsson. Haukr’s aim was to copy various texts of an historical, scientific,
or doctrinal nature; these works were of varied provenance, some Icelan-
dic or Scandinavian, some foreign, mostly translated from Latin (though
a small portion of text is derived from Old English). The Breta sögur are
found in AM 544 4to. The hand has been demonstrated as being Haukr’s
own by Stefán Karlsson, who dated this part of the codex from 1302 to
1310.

7

Haukr’s version of the Brut is shortened but integral. There are oc-

5. Hélène Tétrel, “Arthur et le géant aux barbes: genèse et circulation d’un épisode fon-

dateur dans l’adaptation norroise de l’Historia regum Britanniae,” in Histoires des Bretagnes,

vol. 1: Mythes de Fondation (Brest: CRBC, forthcoming).

6. A detailed description of the three parts that constitute the codex can be found in Jón

Helgason’s introduction to his facsimile edition, Hauksbók, Manuscripta Islandica (Copen-

hagen: Ejnar Munksgaard, 1960), V. Elisabeth Ashman Rowe summarizes Hauksbók’s codi-

cological history in “Literary, codicological, and political perpectives on Hauksbók,” Gripla,

19 (2008), 51–75. Jonna Louis-Jensen has given a description of AM 573 4to and related

fragments and copies in the introduction to her edition of the

β-version of Trójumanna saga

(Trójumanna Saga, Editiones Arnamagnæanæ, A, 8 [Copenhagen, 1963]). For her survey

of Breta sögur, see Medieval Scandinavia, An Encyclopedia, ed. Phillip Pulsiano and Kirsten

Wolf (New York: Garland, 1993), pp. 57–58. See also Svanhildur Óskarsdóttir, “Genbrug i

Skagafjörður: Arbejdsmetoder hos skrivere i klostret på Reynistaður,” in Reykholt som makt-

og lærdomssenter i den islandske og nordiske kontekst, ed. Else Mundal (Reykholt: Snorrastofa,

2006), pp. 141–53.

7. Stefán Karlsson, “Aldur Hauksbókar,” reprinted in Stafkrókar (Reykjavík, 2000), pp.

303–9.

Trojan Origins

493

casional changes in order to reveal connections with Scandinavian history,
particularly with Norwegian royal dynasties.

8

This redaction includes the

Merlínússpá, a translation of the Prophetia attributed to the monk and scald
Gunnlaugr Leifsson, who is known to have died in 1218.
The other redaction of Breta sögur is represented by the badly pre-
served Icelandic vellum AM 573 4to, whose lacunae can be filled from
the seventeenth-century paper copy, Stockholm Papp. 58 fol. This pa-
per manuscript represents a version of Breta sögur which is longer than
Haukr’s and often closer to the Latin text. It now contains only a version
of the “Saga of the Trojans,” a much damaged copy of the Breta sögur
(particularly in the last paragraphs of the Arthurian section), and the
beginning of the Valvens þáttr. This manuscript reveals a clear interest
in the chivalric content of the text:

9

firstly, because it ends with Arthur’s

death at Avallo, and, secondly, because the text from fol. 63r directly con-
nects the chapter depicting Arthur’s death to the beginning of Valvens
þáttr. Moreover, it exhibits a general tendency to develop romance motifs
in the narrative. Though there is no way of knowing for sure if it once
included more Arthurian sagas, since it breaks off abruptly soon after the
beginning of Valvens þáttr, the context of reception does appear to be dif-
ferent from that of Hauksbók. AM 573 4to is obviously a gathering of two
specific sections: the first part corresponds to the pre-Arthurian part of
the Historia regum Britanniae, and is copied by the same hand. The second
part, copied by a second hand (starting on fol. 46r), begins with the story
of King Uther. The division of the work is probably not random: the first
hand seems to have been responsible for the pre-Arthurian chapters of
Geoffrey’s chronicle, while the other hand was responsible for Geoffrey’s
Arthurian chapters and extended them with the inclusion of Valvens þáttr.
Nowadays we cannot know whether this manuscript originally included
other Arthurian romances or the post-Arthurian chapters of Geoffrey’s

8. See A. G. Van Hamel, “The Old Norse Version of the Historia Regum Britanniae and the

Text of Geoffrey of Monmouth,” Etudes Celtiques, 1 (1936), 197–247; Stefanie Würth, Der

“Antikenroman” in der isländischen Literatur des Mittelalters. Eine Untersuchung zur Übersetzung

und Rezeption lateinischer Literatur im Norden, Beiträge zur nordischen Philologie, 26 (Basel

and Frankfurt a. M., 1998); Hélène Tétrel, “Filiations improbables du cycle troyen-breton en

Islande,” in Lignes et Lignages dans la littérature arthurienne, Actes du 3e colloque arthurien de

Rennes de novembre 2005, ed. Christine Ferlampin-Acher and Denis Hüe (Rennes: Presses

Universitaires de Rennes, 2007), pp. 245–58.

9. About the longer redaction of Breta sögur and its Arthurian material, see Marianne

Kalinke, “The Arthurian Legend in Breta sögur: Historiography on the Cusp of Romance,” in

Greppaminni. Rit til heiðurs Vésteini Ólasyni sjötugum (Reykjavík: Hið íslenska bókmenntafélag,

2009), pp. 217–30; Hélène Tétrel, “La Saga des Bretons: naissance et exploitation du mythe

arthurien dans les compilations pseudo-historiques de Scandinavie,” in Enfances Arthuri-

ennes, Actes du 1er colloque de Rennes de mars 2003 (Orléans: Paradigme, 2006), pp.

299–311.

494

Tétrel

text (in the latter case, the Arthurian tales would have been inserted in
the text and not added as an extension). The paper copy was made from
Ormsbók, a once well- known but now no longer extant manuscript, which
contained other “chivalry sagas.” Both witnesses of the longer redaction
will be used in the discussion that follows, with AM 573 4to as the prin-
cipal text.
There is evidence of an original link between Trójumanna saga and Breta
sögur. Jonna Louis-Jensen has shown that both the shorter and the longer
redactions are linked to the

β-version of the “Saga of the Trojans.”

10

Fur-

ther, we know that the “Saga of the Trojans” and the “Saga of the Britons”
as extant in Hauksbók are both in Haukr’s hand. It is difficult to draw any
firm conclusion regarding the intention of the scribes of AM 573 4to,
given the lacunary state of the copy. Concerning the link between the two
cycles, though, Jonna Louis-Jensen pointed out in her description of the
manuscript that “it was obviously the intention to have Bs in a separate
codex.”

11

This transmission, then, is strikingly comparable to the situa-

tion of the French anonymous thirteenth-century Estoire des Bretons in the
manuscript BNF fr 17177. Indeed, the author of the manuscript 17177
inserted a Brut in the middle of the Histoire Ancienne.

12

The last sentence

of Trójumanna saga emphasizes the link in Hauksbók: “en her eftir hefir
sogv fra Enea ok þeim er Bretland bygðv” (and here begins the story of
Aeneas and the men who inhabited/founded Britain).

13

And so does AM

573 4to, whether or not it was a sizeable compilation: “Her eftir hefiaz
upp Breta sogur fyrst fra Enea oc hans ætt monnum. Turno oc Bruto oc
Arturo oc odrum þeim er bygdu Bretland” (Here begin the stories of the
Britons; first, about Enea and his descendants, [and then] Turnus, Brutus,
and Arthur and all the men who inhabited/founded Britain).
It is remarkable that the reference is made not to the “Breta sögur” as a
whole but more specifically to “Aeneas,” his offspring, and “the men who
inhabited Britain.” The longer redaction is even clearer than Hauksbók:
“first, about Enea and his descendants.” The sentence refers more precisely
to the beginning of the interpolated Virgilian part of the Old Icelandic
Brut, rather than to the Brut proper. Indeed, a title added in postmedieval

10. Louis-Jensen, Medieval Scandinavia, pp. 57–58.

11. Louis-Jensen, Trójumanna saga, 1963, p. xxxii.

12. The oldest and best manuscript of the Histoire Ancienne, BN f. fr. 20125, does not retain

any trace of the Brut. It combines an adaptation of the De Excidio Trojae with a summary of the

Aeneid, and passes directly to the Roman section. Following the same structure, the scribe of

BNF f. fr. 17177 retained the Trojan-Virgilian section and the Roman section, but inserted

a translation of Geoffrey in the middle. Old Icelandic compilations, such as Veraldar saga

and AM 764 4to, have a similar organization, including, for the latter, a partial copy of the

Brut.

13. Hauksbók, p. 226; see also Finnur Jónsson, “Indledning,” Hauksbók, p. xcvii.

Trojan Origins

495

times on fol. 24r in AM 573 4to confirms this impression: “Sagan af Enea
hinum frœga eður Brettlands konga Sogur” (The story of the famous Enea
or histories of the British kings). This sentence, probably inspired by the
final sentence in Trojumanna saga, emphasizes the connection between
the matter of Troy and the matter of Britain but also sheds light on the
“Eneas” part of the cycle.
In her study, Julia Crick mentions a total of twenty-seven manuscripts
of the Historia regum Britanniae (vulgate and variant versions included)
containing a version of the De Excidio Trojae. According to her account,
this story of the fall of Troy is the single work most frequently associated
with Geoffrey’s Historia, a statement which is confirmed by Louis Faivre
d’Arcier, who adds a manuscript to the list of twenty-seven already estab-
lished.

14

Noting that the connection is “logical,” Crick poses the question:

“to what degree, then, is the association of these widely-circulated texts in
the manuscripts inherited or spontaneaous?” The remarks by Neil Wright
in the description of the eight manuscripts he used to establish his edition
of the First Variant Version make clear that the Trojan-Briton cycle effectively
existed in some of the first variant manuscripts.

15

The Historia regum Britanniae is known to have been widely translated in
Europe very soon after its composition. We do not know exactly when it
was translated into Old Icelandic for the first time, nor in what context it
came to Scandinavia. Several explanations for its transmission have been
proposed, all of them have taken into account the difficult question of
the relation between the Merlínússpá and the Historia regum Britanniae
itself. The question as to whether the Merlínússpá was directly translated
from a libellus Merlini or extracted from the Historia regum Britanniae has
been raised but remains unresolved. Whatever might have happened in
the process of transmission, there are good reasons for believing that
Gunnlaugr, the author of the Old Icelandic prophecy, knew the rest of the
Historia regum Britanniae, and that his prophecy did not remain isolated
very long.

16

14. Julia Crick, The Historia Regum Britannie of Geoffrey of Monmouth, vol. 4: Dissemination

and Reception in the later Middle Ages (Cambridge: D. S. Brewer, 1991); Louis Faivre d’Arcier,

Histoire et Géographie d’un Mythe, la circulation des manuscrits du De Excidio Troiae de Darès le

Phrygien (VIIIe–VXe siècles), Mémoires et Documents de l’Ecole des Chartes, 82 (Paris, 2006),

p. 151.

15. The Historia Regum Britannie of Geoffrey of Monmouth, vol. 2: The First Variant Version, a

Critical Edition (Cambridge: D. S. Brewer) 1988, p. lxxviii.

16. About the dating of Merlínússpá, see Henry Goddard Leach, “De Libello Merlini,”

Modern Philology, 8 (1911); Van Hamel, “Old Norse Version”; J. S. Eysteinsson, “The Relation-

ship of Merlinusspá and Geoffrey of Monmouth’s Historia,” Saga-Book, 14 (1953–55), 95–112;

Jakob Benediktsson, “Merlinusspá,” in Kulturhistorisk leksikon for nordisk middelalder, 11 (1966),

cols. 556–57; Sveinbjörn Rafnsson, “Merlínusspá og Völuspá í sögulegi samhengi,” Skírnir,

173 (1999), 377–419.

496

Tétrel

In the long process of editing the Historia regum Britanniae, Geoffrey of
Monmouth scholars have brought to light evidence concerning the exist-
ence of the variant version and its relationships with the other redactions
(all classified under the name “vulgate version”).

17

Over the last century,

scholars have shown that the Anglo-Norman poem composed by Wace,
our oldest Brut, is indebted to the variant version of the Historia regum
Britanniae.

18

Since its first publication by Jacob Hammer in 1951,

19

it has,

indeed, been possible to show that some of the discrepancies between the
Historia regum Britanniae (as previously edited in its main text by Faral and
Griscom

20

) and Wace’s Brut were not due to the Anglo-Norman redaction

but to the author of the variant version. Thus, some of the statements
made by Leo Waldner

21

and Margaret Houck

22

appeared to be explicable

in terms of there having been a Latin text more or less contemporary
with Geoffrey himself. The relation between the variant version and the
vulgate text has been discussed by several scholars since Jacob Hammer’s
edition. The most recent editor of the First Variant Version, Neil Wright,
offers a thorough exploration of these issues in the introduction to his
own edition.

23

A detailed comparison between both texts of the Breta sögur and the
vulgate and variant versions of the Historia regum Britanniae is beyond the
scope of this article. Familiarity with the texts led Stefanie Würth, who
translated the Trojan-Briton cycle from Hauksbók, to think that the author
of Breta sögur worked with the variant version or with a hybrid manuscript.

24

A closer observation of the first chapters of Breta sögur may confirm this
assumption, given the fact that the variant version of the Historia regum

17. Julia Crick nevertheless reminds us that the terminology is partially misleading: “It

should be stated at the outset that while the two main forms of variants identified have

been analyzed and textually examined, ‘vulgate’ remains a blanket term for the remainder

(some 190 manuscripts) and has as yet little textual meaning” (The Historia Regum Britannie

of Geoffrey of Monmouth, IV, 14).

18. See Robert Caldwell, “Wace’s Roman de Brut and the Variant Version of Geoffrey of

Monmouth’s Historia Regum Britanniae,” Speculum, 31 (1956), 675–82; “The use of sources

in the Variant and vulgate versions of the Historia Regum Britanniae and the question of the

order of the versions,” Bibliographical Bulletin of the International Arthurian Society, 9 (1957),

123–24.

19. Jacob Hammer, ed., Geoffrey of Monmouth, Historia Regum Britanniae, A Variant Version

(Cambridge, MA: The Mediaeval Academy of America, 1951).

20. Edmond Faral, La Légende Arthurienne, Etudes et Documents, 3 vols. (Paris: Champion,

1929); Acton Griscom and Robert Ellis Jones, The Historia Regum Britanniae of Geoffrey of

Monmouth (Oxford and New York, 1929; repr., Geneva: Slatkine Reprints, 1977).

21. Wace’s Brut und seine Quellen (Karlsruhe, 1914).

22. Margaret Houck, “Sources of the Roman de Brut of Wace,” in University of California

Publications in English, vol. 2 (Berkeley: Univ. of California Press, 1941), pp. 161–356.

23. The Historia Regum Britanniae of Geoffrey of Monmouth, vol. 2: The First Variant Version,

pp. xi–lxxviii.

24. Isländische Antikensagas, “Nachwort,” p. 311.

Trojan Origins

497

Britanniae is a demonstrably interpolated adaptation of the Galfridian
text. Neil Wright notes that the author of the First Variant Version “was
also interested in historical matters, for he had recourse to Bede and
Landolfus Sagax and tried on occasion to reconcile the Galfridian ver-
sion of events with these more orthodox historical authorities.”

25

As will

be suggested below, this is the case for the Old Icelandic translator(s) as
well, for they show knowledge of elements deriving from a version of the
Historia Romana.

26

The Breta sögur lack the first chapters of the Galfridian chronicle. If we
compare them with the vulgate version of the Historia regum Britanniae, we
note that both redactions of Breta sögur lack chapters 1 to 3 (or 4, depend-
ing on the version of the Historia regum Britanniae used for comparison).
These chapters of the Latin text contain (depending on the copy) the
dedication made by Geoffrey himself in three different forms. These dedi-
cations are notoriously missing in all copies of the variant version: since
this work was a rewriting, it did not need any reference to the dedicatee.
Chapter 5 of the Historia regum Britanniae, containing the description of
Albio, is also missing in the Old Icelandic texts, though it is found in the
vulgate and in the variant version.

27

Instead of these first five chapters,

the authors of the Breta sögur, like the author of the Histoire Ancienne, give
an extended narrative about Aeneas.
Geoffrey of Monmouth has kept elements of Aeneas’s travels in Chapter
6 of the Historia regum Britanniae. This summary deals with Aeneas’s suc-
cession, with his son Ascanius and the birth of Silvius, father of Brutus, in
just a few lines. In Breta sögur, the passage is considerably lengthened. As
edited in Hauksbók, this preamble summarizing the Aeneid corresponds
to pp. 231.1–238.25. The longer version, represented by the two above-
mentioned redactions, covers several pages and is twice as long as Hauks-
bók’s rendering.
This Old Icelandic “Aeneas’s saga” is written in ordo naturalis, as is to
be expected in an historical context, and consists of a hybrid rendering
of the poem. This rendering takes liberties with the Latin text: Princess

25. The Historia Regum Britanniae of Geoffrey of Monmouth, vol. 2: The First Variant Version, p.

viii.

26. Amedeo Crivellucci, ed., Pauli Diaconi Historia Romana, Fonti per la Storia d’Italia, 51

(Rome, 1914), p. 6. The constant borrowing among chroniclers makes the identifications

and attributions notoriously difficult; see the remarks by Crivellucci, p. xliii. A text of the

Historia Romana, established by Droysen according to the three layers of text (Eutropius,

Paulus Diaconus, and Landolfus Sagax) is edited in the Monumenta Germanica Historiae,

Auctores Antiquissimi (Berlin, 1879), II.

27. Julia Crick comments on the “separability of the opening chapters” and explains it by

their content: the dedications are interchangeable. She mentions only one manuscript of

the Historia, however, omitting §5. The Historia Regum Britannie of Geoffrey of Monmouth, IV,

pp. 100–1.

498

Tétrel

Camilla, for example, does not appear, though she is well represented in
other vernacular renderings of the Aeneid, whereas the story of Nisus and
Euryales plays an important part in the saga. Entire books are missing:
Book II, because it had become superfluous in a context where a Troy
section was already at hand, and Book VI, because it seemed irrelevant
in an historical context, as opposed to the allegorical tradition, where
it was of fundamental importance. Despite the difference in length and
treatment, the frame of the narrative is the same in both the shorter and
longer versions. Unlike many other Bruts, but as with the French Eneas,
this “romance” developed several episodes, the most striking of which was
the episode of Queen Dido.
The very beginning of the saga corresponds to Aeneid, Book III (the
first three hundred lines approximately), which tells of Aeneas’s trav-
els around Sicily, combined with a far-fetched and drastic summary of
Book V, in which Aeneas stays in Sicily and is welcomed by Acestes; the
paragraph is followed by an account of the last part of Book I, in which
Aeneas reaches Dido’s land; thus, there is a transition to the episode told
in Book IV. Then, the narrative takes Aeneas directly to Latium instead
of mentioning his stop in Sicily; it skips Book VI and then concentrates
on the wars in Italy as narrated in the second half of the poem.
From the moment when the Aeneid comes to its end and moves toward
the Galfridian text, the transition is created from elements which are
neither derived from the Aeneid nor borrowed from a Geoffrey text but
which come from late-Antique historical renderings of Aeneas’s journeys.
Since Aeneas was presented as the ancestor of both the Roman and the
British lineages, elements can be found relating to both Roman and Brit-
ish history. It is not surprising to find more connections with other Old
Icelandic texts in this part of the Breta sögur. Then, from the very moment
when the saga comes to the story of Brutus’s father Silvius, both the longer
version and Hauksbók start to follow the Historia regum Britanniae.
A rapid overview of the treatment of the British Islands, Anglo-Norman,
Anglo-Saxon, or Welsh in the edited Brut has not revealed any trace of a
comparable interpolation in the opening chapters.

28

Wace, who composed

28. The Oldest Anglo-Norman Prose Brut Chronicle, ed. and trans. Julia Marvin, Medieval

Chronicles, 4 (Woodbridge: Boydell, 2006). I have not been able to complete research on

the manuscript transmissions of Layamon’s Brut and the Brut y Brenhinnedd for this article.

I am aware that the edition of the Brut y Brenhinnedd by John Jay Parry does not represent

the only known version. According to Parry, other versions, like that in the manuscript Llan-

stephan I, from which Brynley Roberts edited his selection of excerpts, contain additions

that seem to derive from classical sources (Brut Y Brenhinnedd, Cotton Cleopatra version, ed.

and trans. John Jay Parry (Cambridge, MA: The Mediaeval Academy of America, 1937), p.

xi; Brynley Roberts, Brut y Brenhinnedd, Llanstephan Ms 1 Version (Dublin: Dublin Institute

for Advanced Studies, 1971).

Trojan Origins

499

a section (106 ll.) relating to Aeneas’s journeys, does not seem to have
made use of any other material than Geoffrey text(s), which he consider-
ably amplifies. The Munich Brut,

29

on the other hand, combines Geoffrey

and other Aeneas materials which differ slightly from the Latin narrative.
The author may have had knowledge of the Aeneid but does not show any
connection with the Old Icelandic texts. The French Bruts in prose offer
far better material for a comparison. The prose Bruts have been examined
by Géraldine Veysseyre, who studied and partially edited three of them.

30

They all have in common with the Old Icelandic renderings the fact that
they are parts of larger universal histories written in vernacular prose: the
thirteenth-century anonymous Estoire de Brutus contained in manuscript
BNF fr 17177 has been inserted (fol 82v-108) in the middle of the Histoire
Ancienne jusqu’à César, which covers fols. 13–189. Originally, this Brut was
a separate work, but it was subsequently inserted between quires IX and
X. The insertion was made with some deliberation, for it is signalled by a
rubric, and the last relevant page of the Histoire Ancienne has been cancelled
before the insertion.

31

This textual environment shows much closer simi-

larity to the Old Icelandic renderings than to the older French verse Brut.
Another work studied by G. Veysseyre, the Croniques des Bretons, found in
four manuscripts all dated from the fifteenth century, offers enough simi-
larities with the Old Icelandic texts to be mentioned as a parallel example;
this text was copied in order to become a part of a general history. The
translation of the Historia as read in the Croniques des Bretons is preceded by
a paragraph which its author calls a “preamble” summarizing the travels of
the three “generations” (lineages) of Trojan immigrants. The text mentions
a young “Priam,” son of the old king Antenor, and then Aeneas.

32

The title

announcing the preamble reads as follows: “Ci dessoubz commence un
petit preambule par maniere de thieume sus la translacion du livre de Brust
d’Angleterre”

33

(Here begins a small preamble to serve as an introduction

to the translation of the Brut/ chronicle of England).

29. Der Münchener Brut, ed. Konrad Hofmann und Karl Vollmöller (Halle,1877), pp. 3–7,

ll. 91–214.

30. Géraldine Veysseyre, “Translater Geoffroy de Monmouth: trois traductions en prose

française de l’Historia Regum Britanniae (XIII–XV)” (unpublished PhD diss., University

of Paris-IV Sorbonne, 2002; forthcoming).

31. Veysseyre, II, 2, transcription of BNF fr. 17177, fol. 82r.

32. This brief “preamble” could be compared to the summa historiae introducing an un-

known accessus ad Lucanum, whose existence was discussed by Dietrich Hofmann and Þor-

björg Helgadóttir to explain the relations between Rómverja saga and Veraldar saga (Dietrich

Hofmann, “Accessus ad Lucanum: Zur Neubestimmung des Verhältnisses zwischen Rómverja

saga und Veraldar Saga,” in Sagnaskemtun. Studies in Honour of Herman Pálsson, ed. Rudolf

Simek, Jónas Kristjánsson, and Hans Bekker-Nielsen (Vienna: 1986), pp. 121–51; Þorbjörg

Helgadóttir, “On the Sources and Composition of Rómverja Saga,” Saga-Book of the Viking

Society for Northern Research, 24 (1996), 203–20.

33. Veysseyre, III, 62. Thieume is a variant for theme. It is to be understood as a rhetorical

term meaning “small composition, introduction to.”

500

Tétrel

For its principal part, the thieume is devoted to Aeneas, whose story is told
over several paragraphs.

34

The narrative is composed in ordo naturalis and

consistently omits Book II. Book III is missing as well, but a summary of
Aeneas’s visit to Sicily (Book V) occurs after the paragraphs devoted to the
sojourn in Cartago (Book I-IV). Book VI is omitted, and the Italy/ Latium
section of the story is obviously emphasized. Some important variants of
the Breta sögur are not shared by the Croniques des Bretons, and the similari-
ties between the texts do not show evidence of any direct relation; on the
other hand, some others, as I shall try to demonstrate, tend to confirm the
impression that we are dealing with a similar context of transmission, and
that both texts rely on similar materials. This text strongly resembles the
Aeneas section in the Histoire Ancienne jusqu’à César as found in its main
manuscript, BNF f. fr. 20125.

35

Veraldar saga is an Icelandic universal history divided into six ages,
beginning with the Creation of the World, rather briefly treated and
drawing heavily on Petrus Comestor for the Genesis section. The Troy
section appears in the fifth age, synchronized with other reigns. Its ver-
sion of Aeneas’s journeys is based on the usual account of the three
“generations,” Helenus, Antenor, and Aeneas, and follows each of them
briefly. The last fugitive, Aeneas, is followed until he arrives in Latium,
fights Turnus, and marries the daughter of King Lavinius, before the text
passes to the Roman section: thus, Aeneas’s story is used as a prehistory
to the Roman empire. In his edition of Veraldar saga, Jakob Benediktsson
points to readings shared by the Trójumanna saga and Veraldar saga and
suggests that both had a similar model.

36

As I shall discuss in this paper,

the possibility of a shared source can be extended to a Trojan-Briton
cycle, and not only to the Trojan part of the story: the Aeneid part of
Veraldar saga does indeed share readings with the Breta sögur rendering
of Aeneas’s travels.
The last text to be included in our corpus is the Icelandic AM 764 4to,
a fourteenth-century manuscript. Like Veraldar saga, with which it has obvi-
ous links, AM 764 4to contains, together with other material, a universal
history. The Troy section is much shorter than the version to be found in
the Old Icelandic Brut and shorter than the Troy section in the Veraldar
saga; on the other hand, it has retained a partial story of the British branch
according to Geoffrey, to be found later in the manuscript, whereas Veraldar
saga ignores Brutus and his offspring. Apparently, then, the writer of AM

34. Veysseyre, III, 66–74.

35. BNF f. fr. 20125, fol. 150–76. On this manuscript, see also Jacques Monfrin, “Les

Premières compilations françaises d’histoire ancienne.”

36. Veraldar Saga, p. xlvii: “den eneste forklaring synes at vaere, at kilden til dette afsnit i

Veraldar Saga har været en udvidet bearbejdelse af Dares, af samme beskaffenhed som den

redaktion der ligger til grund for Trójumannasaga.”

Trojan Origins

501

764 4to had in mind the fact that Aeneas’s generation led to two different
branches and felt the need to mention it.
This apparently heterogeneous corpus raises several questions:

1) Do the shorter and the longer versions of the Old Icelandic Aeneas

section in the Breta sögur derive from the same source?

2) On what materials, besides the Historia regum Britanniae, do our

texts draw and what are the relations between Breta sögur and the
other texts of the corpus?

3) When were these materials brought to Iceland?

My aim is, first, to settle the first question once and for all; then, to deal
to some extent with the second; and, last, to open up ways by which the
third and most complex question can be investigated. I shall therefore
proceed to the analysis of a selection of variants which will, I hope, help
the discussion.
The first series of items on which I shall focus go back to an unknown
version of the Historia Romana used by several compilers in Iceland. It is
very likely that a form of this work came to Iceland along with other ma-
terials, but as yet it is difficult to decide whether the Galfridian material
in it (or them) was an important part or whether it was only inserted as a
summary of the first generations of the British kings, starting with Brutus
(which corresponds to what remains of it in Veraldar saga and AM 764
4to). On the other hand, the Historia Romana could have been transmit-
ted independently, for instance together with the Historia Langobardorum,
which is another chronicle by Paul the Deacon known to have circulated
widely. Many manuscripts have kept it together with the Historia Romana,
and it is known to have reached Iceland in the twelfth century.

37

To start with the weakest argument, we can observe that the duration
given for the reigns of Aeneas and Ascanius is more or less based on the
Historia Romana, and, therefore, does not always rely on the vulgate Histo-
ria regum Britanniae. According to both versions of the Breta sögur, Aeneas
reigned for three years over Italy, which corresponds to the number of
years mentioned in the Historia Romana

38

and at least one manuscript of

the First Variant Version of the Historia regum Britanniae, though the main
manuscripts of the First Variant Version specify “four years,” as does the
French Croniques des Bretons; the other redactions of the French prose
Brut do not identify the length of the reign, and neither does the text

37. Theodoricus Monachus, the author of the Historia de Antiquitate Regum Norwagiensium,

dated to the end of the twelfth century, refers to it directly in chapter 17 of Monumenta

Historica Norvegiæ, Latinske Kildeskrifter til Norges Historie i Middelalderen, ed. by Gustav Storm

(Kristiania, 1880; repr., Oslo: Norsk Historisk Kjeldeskrift-Institutt, 1973), p. 32.

38. Pauli Diaconi Historia Romana, p. 7.

502

Tétrel

in AM 764 4to, which is too short to go into such detail. According to
Hauksbók and Veraldar saga, Ascanius reigned for thirty-seven years: this
figure is not found anywhere else; he reigned for thirty-eight years ac-
cording to the longer version of Breta sögur (Stock. Papp. 58) and the
Historia Romana; thirty-nine years according to the Munich Brut (line 332);
thirty-four years in Wace’s Brut (as copied from the variant version of the
Historia regum Britanniae); and the Croniques des Bretons probably misread
the number thirty-four in making it forty-four. One might conclude from
the observations concerning the length of these two reigns that the Breta
sögur show correspondence with the Historia Romana on the first number,
while Veraldar saga and AM 764 4to do not contradict them. Regarding
the second number, the longer version of Breta sögur is again in agree-
ment with the Historia Romana; and the thirty-seven years of Hauksbók and
Veraldar saga could be understood as an interpretation of the thirty-eight
years (Ascanius died during the thirty-eighth year of his reign, and the
numbers thirty-seven and thirty-nine can be derived from the original
thirty-seven). However, such numbers are not in themselves particularly
compelling evidence, because of the high probability of scribal errors,
and they need to be corroborated by other evidence.
A second and more interesting element is a mistake regarding Turnus.
According to Virgil, Turnus was the son of the king of the Rutelians. The
vulgate version of the Historia regum Britanniae follows the Aeneid, but a
common variant reading is found in the variant version of the Historia
regum Britanniae, which in turn borrowed it from the Historia Romana.
This detail has been discussed by Leo Waldner, Margaret Houck, Robert
Caldwell, and finally Neil Wright.

39

In Wace’s Brut,

40

Turnus indeed is said

to be son of “Daunus, king of Tuscany,” This reading is shared by Hauksbók:
“Damius het konvngr er reð fyri Tvskania”

41

(A king called Damius reigned

over Tuscania). But the longer version of Breta sögur and Veraldar saga
have another variant: “Dullunus hét konungr; hann réð fyrir þar sem nú
stendr Rómaborg”

42

(A king called Dullunus reigned over the land where

Rome now stands). “Hann var s(onr) Davnvs er riki hafdi i þeim stad er nv
er Rvmaborg”

43

(He was the son of Daunus, who reigned over the place

39. See above, nn. 17–21.

40. Le Roman de Brut, de Wace, 2 vols., ed. Ivor Arnold (Paris: Société des Anciens Textes

Français, 1938–1940), I, ll. 50–51. Two of the prose-Bruts, ed. G. Veysseyre, share the same

reading (Estoire de Brutus, II, fol. 82v, and Croniques des Bretons, III, § VI). They might have

borrowed it from Wace. Jehan Wauquelin, however, describes Turnus as “roy des Ruteliens”

(IV, §86).

41. Hauksbók, p. 233.

42. AM 573 4to, fol. 25v; on fol. 137r of Stockholm Papp. 58, fol. is a similar reading.

43. Veraldar Saga, p. 46. According to the editor, variants from the B-class manuscripts

read: “Dacenus.”

Trojan Origins

503

where Rome now stands). Once again, AM 764 4to, in its brevity, does not
give any specific details regarding Turnus. Thus, the reading divides the
text into two groups. The first of these follows the variant version of the
Historia regum Britanniae: within this group fall Wace’s Brut; the French
prose Bruts (knowing that the latter might have been copied from the
former); and Hauksbók, alone among the Old Icelandic renderings. In
the second group are to be found AM 573 4to and Veraldar saga, neither
of which follows the variant version or the Aeneid or the vulgate version
of the Historia regum Britanniae, in a strict sense. I would be tempted to
regard this second reading as a correction of the first, because it shows
no sign of being a specific variant, but its author (presumably the same
individual) seems to have grown conscious of the inconsistency in respect
to the Virgilian narrative. One detail is certain at least: although their
wording and content are slightly different, Veraldar saga and AM 573 4to
share the same reading and differ from all the others. If my interpreta-
tion is correct, they rely on a version which emended a text based on the
Historia Romana (or the First Variant Version of Historia regum Britanniae),
which must, therefore, have been the original version of the Old Icelandic
Brut.
Another interpolation, which goes back to the Historia Romana but as-
sociates Hauksbók and Veraldar saga against AM 573 4to, is a hapax legomenon
in our corpus: it has no counterpart in Wace’s Brut or in any other Brut or
in the variant version of the Historia regum Britanniae. Quite the opposite: it
belongs to a non-British context, so to say, and usually appears exclusively
in the Roman section of universal histories (or, for a different purpose, in
grammatical-etymological paragraphs). The content of the interpolation
is as follows: at the mention of King Latinus’s daughter, both Hauksbók and
Veraldar saga insert a comment on the invention of the Latin alphabet:

Konvngr reð fyri Italia sa er Latinvs (het). dottir hans het Latina. hon fan

fyrst latinv stafrof ok af hennar nafni heita aller Latinv menn þeir er þa tvngv

kvnnv. (Hauksbók, 233.6–9)

(A king called Latinus reigned over Italy. His daughter was called Latina.

She invented the Latin alphabet, and all the Latins, men who speak this

language, are named after her.)

Eneas gek et eiga Lavinea dotvr Latinvs konvngs er latinvtvnnga er við kend

þvi at Nicostra(t)a kona hans fan latinv stafrof. (Veraldar saga, 46, 83–85)

(Eneas married Lavinea, daughter of King Latinus, after whom the Latin

language is named, because his wife, Nicostrata, invented the Latin alpha-

bet.)

This etymology, though totally disconnected from the Historia regum Bri-
tanniae and its context, can be traced back to Servius’s commentary on

504

Tétrel

the Aeneid,

44

which was frequently copied in medieval histories. Isidor’s

Etymologies and Paul the Deacon’s Historia Romana are among the most
frequent late antique works which contributed to its transmission in the
Middle Ages. But Servius has a text which differs from the two above-
mentioned passages in Old Icelandic: Nicostrata is called “Carmen” “quia
carminibus vaticinabatur” (because she prophesied in verse) and is King
Evander’s mother. A similar reading can be found in two of the so-called
Vatican mythographers: in the first, fabula 69, and in the second, fabula
176, both dealing with King Evander (and not King Latinus). Isidor fol-
lowed this interpretation as well.

45

In his introduction to Veraldar saga,

46

Jakob Benediktsson, who seems not to have been aware of its presence in
the Breta sögur, attributes its origin to Isidore’s Etymologies, I, IV, 1:

Latinas litteras Carmentis nympha prima Italis tradidit. Carmentis autem

dicta, quia carminibus futura canebat. Ceterum proprie vocata [est] Nicos-

trate.

47

(The nymph called Carmentis was the first to teach the Latin alphabet to the

Italians. And she is called Carmen because she prophesied the future in verse

[carmen]. She is, otherwise and more properly, called Nicostrate.)

But the presence of this feature in the Breta sögur, connected as we know
them to be to the Veraldar saga, confirms a direct transmission through
the Historia Romana, though it does not preclude knowledge of Isidore.
Indeed, the Historia has the passage at the exact place where King Latinus’s
genealogy is given, as in the Old Icelandic texts:

48

Post hunc [Picum] eius filius Faunus, qui fuit pater Latini, cuius mater Car-

mentis Nicostrata creditur Latinas litteras repperisse.. . . Regnante tamen

44. Servii Grammatici qui fervntvr in vergilii carmina commentarii, ed. Georgius Thilo and

Hermannus Hagen (Leipzig: Teubner, 1884), II, fasc. II; In Aen. VIII, 51, p. 206.

45. Mythographi Vaticani I et II, ed. Péter Kulcsár, Corpus Christianorum, series latina, 91C

(Turnhout: Brepols, 1987), pp. 30 and 234; Isidori Hispalensis Etymologarivm sive Originvm

Libri XX, ed. W. M. Lindsay (Oxford: Clarendon Press, 1911; repr., 1985), I, lib. I, p. iv. The

tradition that Evander’s mother Carmenta was a prophetess (from the Latin “carmina”) is

mentioned as well in Dionysius of Halicarnassus’s Roman Antiquities, I, p. xxxi. Interestingly

enough, we find a reference to “Evander’s mother Carmentis” in the Gesta Regvm Anglo-

rum, Book II, p. 206. A link between this passage of the Gesta, dedicated to Pallas’s funeral

monument, and the Breta sögur is not to be ruled out (William of Malmesbury, Gesta Regum

Anglorum, The History of the English Kings, ed. and trans. R. A. B. Mynors, completed by R. M.

Thomson and M. Winterbottom [Oxford: Clarendon Press, 1998], I, p. 384).

46. Veraldar Saga, inledning, p. xlviii.

47. Etymologiarum sive Originvm Libri, lib. I, p. iv.

48. The genealogy given by the Historia Romana corresponds to the sequence already at

hand in classical literature (see Metamorphoses, XIV) and widely transmitted through Eusebius-

Hieronymus’s Chronicon: Saturnus—Picus—Faunus—Latinus. This genealogy is familiar to

the medieval historiographers dealing with Roman history. The beginning of Trójumanna

saga in Hauksbók indirectly draws on it as well, though there is ambiguity about the figure

of “Saturnus.”

Trojan Origins

505

Latino, qui Latinam linguam correxit et Latinos de suo nomine appellavit,

Troia a Grecis capta est. . . .

49

(After [Picus] [reigned] his son Faunus, who was the father of Latinus, whose

mother Carmentis Nicostrata is supposed to have invented the Latin alphabet.

. . . Under the reign of Latinus, who improved the Latin language and called

the Latins after himself, Troy was taken by the Greeks. . . .)

In the corresponding part of its narrative, AM 573 4to does not include
the passage and offers a “better” text if we compare it to Virgil. Indeed,
its author calls the king by his right name and only refers to his daughter,
without naming her. The name of Lavinia must have appeared, in the
Virgilian form, later in the text, but this part of the text is now lost.

50

But

if we take a closer look at this variant, we note that the longer version of
Breta sögur can be reconciled with the etymological explanation, which
attributes the invention of the Latin alphabet to Latinus’s mother and has
Lavinia called by her regular name. It is therefore not inconsistent with the
version found in Veraldar saga; on the other hand, it contradicts Haukr’s
version, which attributes the Latin alphabet to Latinus’s daughter. Given all
these observations, I would argue that Hauksbók changed Lavinia’s name
into “Latina,” cancelled the name “Nicostrata” to remove the reference
to the source, and did not use any model for this modification. If the
author of Veraldar saga and Haukr Erlendsson shared the same version
of Latinus’s genealogy, the former most probably had a more accurate
rendering, and Haukr produced an inconsistent one. To Haukr’s credit,
however, we should keep in mind that this story did indeed have several
variants, as noted earlier. Haukr might thus have known several traditions
and hesitated between drawing on any of them. Whatever the reason for
the variant, from the two examples studied here, I am inclined to conclude
that the Historia Romana is the text where the interpolation originated.
These observations, which need to be tested by a more thorough future
examination of the textual evidence, seem to indicate that a text drawing
on the Historia Romana came to Iceland and was used together with the
source for the Old Icelandic Brut. We do not know, however, how and
when the connection was made between the “Roman” and the Galfridian
material. There is a possibility that the texts correspond to two different
times of transmission.
The second group of interpolations includes variants from other sources
which are derived from Ovid or involve Virgilian and Ovidian provenance.
In many cases, Servius has been used as well. In one sense, it is artificial to
separate these examples from the previous ones if we consider the prob-
ability that all of them eventually merged in the same compilation. However,

49. Pauli Diaconi Historia Romana, p. 6.

50. It can be read in Stockholm Papp. 58, fol. 152r.

506

Tétrel

by doing so, I wish to emphasize the fact that these interpolations belong
to the grammatical tradition and could as well have been introduced inde-
pendently through this medium. If this were the case, these readings could
betray the use of an accessus to Virgil (with Ovidian/Servian glosses on the
Aeneid) together with the principal model (Virgil himself) and alongside
the Roman History, which was used to enhance the historical background.
These passages are the ones about which the shorter and the longer versions
of the Breta sögur agree, roughly speaking, and do not have any counterpart
in Veraldar saga or in AM 764 4to. They deal with the mythical raising of
Alba Longa and the meeting between Dido and Aeneas.

51

As will be shown,

the Old Icelandic texts sometimes show connections with the other Bruts
in our corpus regarding these interpolations, a fact which proves that they
belong to a tradition known in Europe.
As has already been observed, the longer version of Breta sögur has a more
developed and interpolated Aeneid section. Its author had immediate access
to a summary of the Aeneid, already interpolated. On the other hand, the
shorter version, as found in Hauksbók, made use of exactly the same inter-
polated Aeneid. Indeed, Hauksbók shares the same Virgilian readings as the
longer version, though it shortens them, and sometimes has details which
are not found in the longer text but can be demonstrated as belonging to
the same tradition. This is indeed the case with the passage where Hauksbók
contains a summary of an episode related to the founding of Alba Longa.
The beginning of Aeneid, Book VIII, tells us how Aeneas, before his meeting
with king Evander, is gratified with a sign of a divine nature: a white sow
and thirty piglets appear to confirm a prophecy made by Tiberinus, god of
the River Tiber, that Alba Longa shall be built there by his son Ascanius.
Hauksbók and the longer version have a different reading at this point, but
they draw on the same model. Indeed, it appears that Hauksbók made use
of the Virgilian text but modified it:

Sva sem þv ferr veg þin þa mantv finna vndir einv tre þvi er vlex heitir hvita

gyllti með .xxx. grisvm alhvitvm ok i þeim stað skalltv borg reisa ok skalltv

kalla Albans borg.

52

51. Concerning the story of Dido and Aeneas (and the raising of Carthago by Dido) the

tradition is notoriously problematic in medieval chronicles. According to Servius’s tradition

(see below), the encounter could not have taken place. Many chroniclers who have knowl-

edge of the Virgilian epic resigned themselves to mentioning several alternate traditions,

but chose mainly to follow the Servian theory. Making this choice, they simply reproduce

the pattern at hand in the Chronicon by Eusebius-Hieronymus, which was very often their

ultimate model. That Veraldar saga and AM 764 4to should not even mention Queen Dido,

even though they do mention the fight against Turnus, is therefore understandable. For the

Chronicon, see Patrologiae Cursus Completus, Ser. Graeca, XIX, accurante J.-P. Migne, Eusebius

Pamphili Caesariensis episcopus, a. 421 (Paris: 1857).

52. Hauksbók, p. 233; Aeneid, VIII, ll. 26–85. Ulex (‘ilex’, ‘oak’) is copied untranslated; it

is found in Virgil, l. 43.

Trojan Origins

507

(On your way, under a tree called ulex, you shall encounter a white sow with

thirty white piglets. And in the same place you shall raise a city and name

it Albansborg.)

This passage is amplified with non-Virgilian materials for which I have not
found any model: Hauksbók tells us in the chapter immediately following
how Aeneas himself founded the city, “as big and powerful as Karthago.”
Later on, the passage corresponding to the statement in the Historia regum
Britanniae that Ascanius raised the city is, of course, emended: the text
explains how Ascanius “had the city extended” (lét auka Albansborg).

53

AM 573 4to, on the other hand, has a nonsensical reading of the above-
mentioned passage but respected the Virgilian text (and/or the Historia
regum Britanniae) when the story again mentions the name of Alba: indeed,
according to the longer version, Aeneas built a city named “Karthago,”
and later on, Alba was raised by Ascanius. This variant shows that the
shorter and the longer versions have a similar origin; that is to say, a text
where the comparison with “Karthago” appears, as opposed to Aeneid,
Book VIII. The Icelandic authors, then, differently but equally incorrectly
interpreted a unique lectio. So far, the Old Icelandic Brut is the only text
I know which offers this reading.
Another striking difference between the Brut tradition and the Breta
sögur is the presence of a developed tale of Dido and Aeneas influenced
by Ovid. In keeping with the Historia regum Britanniae, Wace and Layamon
have skipped the episode in their Virgilian summary, whereas the Old Ice-
landic text pays a good deal of attention to it. This episode is usually absent
from the Aeneas sections in early chronicles, but its use is well-attested in
Latin school poetry.

54

Therefore, even if one admits the possibility that

the model for the Aeneas section of the Breta sögur was a combination of
sources not confined to Christian histories, one can nevertheless not rule
out the hypothesis that the episode draws on early sources, since a tradi-

53. Hauksbók, p. 238.

54. Christopher Baswell, Virgil in Medieval England, p. 20, reminds us that Servius’s in-

fluence on the early historical compilations was strong. Another tradition preventing the

expansion of the episode was articulated around the idea that Dido remained chaste and

pure (See J. Monfrin, “Les premières compilations françaises,” p. 207); the romance tradi-

tion was possibly influenced by the long rendering of Dido’s misfortune as narrated in the

Roman d’Eneas in the early twelfth century. This heterogeneous reception certainly did not

help to maintain clarity for the episode at the time when the Bruts started to be composed

in Europe. The early Latin poems on Dido have been studied and edited by Peter Dronke,

“Dido’s Lament: From Medieval Latin Lyric to Chaucer,” in Kontinuität und Wandel. Lateini-

sche Poesie von Naevius bis Baudelaire, Franco Munari zum 65. Geburtstag, ed. U. J. Stache, W.

Maaz, F. Wagner (Hildesheim: Weidmann, 1986), pp. 364–90, and Otto Schumann, “Eine

mittelalterliche Klage der Dido,” in Liber Floridus, Mittellateinische Studien, Paul Lehmann

zum 65. Geburtstag, ed. B. Bischoff and S. Brechter (St Ottilien: Eos Verlag, 1950), pp.

319–28.

508

Tétrel

tion of Dido’s planctus was already to be found in twelfth-century poems.
But it definitely belongs to a different background from that of Geoffrey’s
chronicle and, more likely, to a later Ovidian tradition.

55

In any case, if

the interpolation goes back to the first rendering of the “saga of Aeneas”
attached to the Brut, one cannot be sure whether this episode was or was
not added by the Icelandic author from personal knowledge.
Among several versions of Dido’s episode, the Breta sögur (or their mod-
el) have chosen a variant based on the seventh of the Heroides, ll. 133–38.
In the Ovidian epistle, indeed, Dido is pregnant, but not in Aeneid, Book
IV, where she deplores the absence of a “little Aeneas” who could soothe
her sorrow. A different tradition, that of Aeneas and Dido’s “infertile”
union, can be found in the Roman d’Enéas

56

and in texts based on the

Eneas from the late twelfth and early thirteenth century onwards.
Though rendered in two different ways, both Old Icelandic versions go
back to a model rearranged according to the Ovidian reading.

Nv hefir þv svikit mik ok sva gvðin eð sama. Nv skyt ek ollv minv mali til þeira

at þav hefni min a þer. gef gavm at að eigi virðiz sva þersi þinni gerð at þv fyri

farir bœðe mer ok þeim þinum syni er (með) mer leyniz af þinni hervist ok

þegar ek veit at þv villt eigi til vars rikis koma þa skalltv raðinn vita min bana

ok sva barnsins þers er ek fer með. (Hauksbók, 232, 25–29)

(And now you have betrayed me and also the gods. I pray to them with all my

heart that they should avenge me on you. And consider that your deed shall

appear as a murder of both me and your son, who came into my womb from

your presence. And since I am certain that you do not wish to come back to

our land, you can expect me and the child within my womb to be dead.)

AM 573 4to presents a similar reading. The beginning of the passage has
a slightly different wording:

ok svá mikinn ósóma hefir þú til tekit við mik at eigi má ek svá búit bera.

Hygg at því at eigi [virðisk]. . . .

(And you have brought so much shame on me that I cannot bear it. Consider

that it not [appear]. . . .)

57

But the changes do not offer evidence of a diverging source.
The tale of Dido and Aeneas became popular in French tradition mostly
in the later Middle Ages; one of the best-known poems responsible for its
transmission was the lengthy historical-allegorical Ovide Moralisé, based on

55. “Malgré la popularité dont Ovide semble avoir joui au Moyen âge, nous avons conservé

assez peu de manuscrits de ses œuvres qui aient été copiés avant le 13e siècle.” Birger Munk-

Olsen, La Reception de la Littérature Classique au Moyen âge (IX–XII) (Copenhagen, 1995).

56. Le Roman d’Enéas, ed. J.-J. Salverda de Grave (Paris: Champion, 1891; repr., 1985), p.

54, ll. 1739–44.

57. AM 573 4to, last line of fol. 25r.

Trojan Origins

509

the Metamorphoses and enriched with classical and late-antique materials.

58

The Ovide Moralisé refers to Dido’s unborn child in Book XIV and is one
of the first Old French texts to do so:

Mes seule ne morrai je mie:

Il me laist grosse et empreignie

D’un enfant qu’il a engendré.

59

(But I shall not die alone:

He leaves me with a child

That he begot.)

This historical-allegorical poem is an interesting parallel for us, partly be-
cause of its widespread influence (late medieval writers such as Guillaume
de Machaut and Christine de Pisan drew on it), but mainly because it ex-
hibits parallels with the Old Icelandic Trojan part of the cycle, especially
the euhemeristic preamble of the Trójumanna saga found in Hauksbók.
This tale about Saturnus and his lineage was composed or, most probably,
copied by Haukr as a preamble to the Pseudo Dares and is to some degree
inspired by Ovid’s Metamorphoses (Book I, end of Book II, beginning of Book
III, and possibly part of Book VII); it has been enriched with other materi-
als and is found only in Hauksbók. A very brief comparison between this
preamble, Ovid’s Metamorphoses, and the Ovide Moralisé shows interesting
links between the two vernacular versions. If they derive from a common
source for the prehistory of Troy, then that source probably contained the
detail of Dido’s pregnancy as well.
However, a short sentence found in Hauksbók and the longer version
at the end of the passage proves, at least, that both versions of Breta sögur
draw on the same source for Dido’s planctus. Following Dido’s letter is a
comment which does not come from Ovid. In the Aeneid,

60

Dido conjures

up a future “avenger” of her people. The Old Icelandic texts identify him
and add a few comments on Hannibal’s vengeance on the Romans. If the
source from which the Old Icelandic text derives was not responsible for
the identification, it is possible that it goes back to Servius’s commentary
on Aeneid, Book IV, identifying Dido’s avenger (et Hannibalem ostendit).

61

58. Pascale Bourgain notes that from the end of the thirteenth century and early into the

fourteenth, the legends developed very soon after the Ovide Moralisé, which first promoted

it in the French vernacular literature: “En fait, la légende de Didon s’est pratiquement

détachée, rendue autonome de son promoteur” (“Virgile et la poésie latine du bas moyen

âge,” in Enée et Didon, naissance, fonctionnement et survie d’un mythe, Actes du colloque inter-

national organisé à la Sorbonne Nouvelle, 6–9 décembre 1988, Paris, ed. R. Martin [Paris:

CNRS, 1990]). See also Eberhard Leube, Fortuna in Karthago. Der Aeneas-Dido Mythos Vergils

in den romanischen Literaturen vom 14. bis zum 16. Jahrhundert (Heidelberg: 1969).

59. Ovide Moralisé, ed. C. De Boer (Wiesbaden: 1915; repr., 1968), V, Book XIV, p. 23, ll.

471–73.

60. Aeneid, IV, 624–29.

61. Servii Grammatici Qvi feruntur in Virgilii Aeneidos Libros, Aen. IV, 625, p. 573.

510

Tétrel

But we may observe that both the shorter and the longer versions of Breta
sögur share this interpolation, which must therefore have been available
in the (same) source.
The last item on which I shall focus is the story of Aeneas’s death, which
has been retained in the Breta sögur as well as in AM 764 4to. In his above-
mentioned article on the Histoire Ancienne, Jacques Monfrin emphasized
the fact that this motif had a rather intricate tradition in the Middle Ages:

62

the Histoire Ancienne, following the Servian method, does not favor any
single version but identifies several possible scenarios concerning Aeneas’s
death. I transcribe the reading from manuscript BNF f. fr 20125:

quant eneas ot la terre auques en pais mise et delivree. la mort que nul [ . . . ]

nen [ . . . ] espargne li coru sore et en tel maniere que petit seit on coment il

perdi lame. quar li un dient quil fu ocis dun effondre. li autre dient que li deu

len ravirent et li autre dient que li cors de lui fu troues en une aigue iouste le

toivre ausi com un estanc. nunicum lapelent cil de la contree et si ne vesqui

eneas que .iij. ans puis quil ot lauine espousee.

63

(After Eneas had delivered and pacified the land, death, that spares nobody,

caught him in such a way that it is difficult to know how he died. Some say

he was killed by a thunderbolt, some say he was taken to heaven by the gods,

and others say that his body was found in a pond next to the river Tiber,

which the inhabitants call “Nunicus.” Eneas did not live more than three

years from the time he had married Lavine.)

The other medieval translations are silent on the subject of Aeneas’s death.
The only extended rendering I know of, aside from the version given by
the Histoire Ancienne, is the chapter in the prose Croniques des Bretons (§
VIII, ch. 2). This chronicle, most probably copied from the Histoire Anci-
enne, mentions the three versions but is more precise than the manuscript
20125. I quote the passage from G. Veysseyre’s edition:

Quant Eneas ot mise la terre et tout le paÿs en paix et qu’il n’avoit mais

nulle guerre fors seulement a Merencius, la mort, qui n’espargne nullui, lui

courut sus en telle maniere que nul ne sceut comment il morut: les ungs

dient qu’il fut foudroiez si comme Ovide, les autres dient qu’il fu ravi des

dieux si comme Virgile, les autres dient qu’il fu noiez en un estant qui est

prés du Tibre et que la fu trouvé son corps. Mais Titus Livius dit autrement,

car il dit que la derreniere des batailles Eneas fit en Lombardie, laquelle fu

moult grande: touteffoiz la prouesce de lui et de sa gent, il ot victoire de ses

ennemis, mais il y mourut. Cestui qui fist l’Istoire de Brutus ne dit riens de

la mort Eneas.

64

(When Eneas pacified the whole land and the whole country, when all the

wars were over, except for the war against Merencius, death, that does not

spare anyone, fell on him in such a way that nobody knew how he died: some,

62. Monfrin, “Les premières compilations,” p. 208.

63. BNF, f. fr. 20125, fol. 176, col. b.

64. Veysseyre, Translater Geoffroy de Monmouth, III, p. 74.

Trojan Origins

511

like Ovid, say he was killed by a thunderbolt; others, like Virgil, say he was

ravished by the gods; others say he drowned in a pond close to the Tiber

river, and his body was found there. But Livy puts it differently; indeed, he

says that the last battle Eneas had to fight was in Lombardy, and it was a hard

fight. Despite his prowess and that of his men, he vanquished his foes but

lost his life there. The author of Brutus’s story does not say anything about

Eneas’s death.)

According to G. Veysseyre, the last sentence in the paragraph mentioning
“celui qui fist l’Istoire de Brutus” (the author of Brutus’s story) probably
stands for the author of the Historia regum Britanniae. Besides, the French
author tries to make a precise statement about the diverging sources he is
citing, and I believe he is to be trusted in this statement. That is to say, he
did not find any detail about Aeneas’s death in his Historia regum Britan-
niae text, and therefore copied the corresponding chapter of the Histoire
Ancienne with additional materials: we note in this paragraph, for instance,
that he attributes the death by thunderbolt to “Ovid.”
Three of our Icelandic texts choose precisely this option, without
showing knowledge of the other two, and without attributing it: Elding
laust hann til bana (Breta sögur, longer version/ Stockh papp 58); en elld-
ing slo hann til bana (Breta sögur, Hauksbók); hann drap elldinng til bana
(AM 764 4to) (He was killed by lightning). Veraldar saga does not say
how Eneas died.
Neither the late-antique renderings nor the Historia Romana show
knowledge of this version. Ultimately, the Historia regum Britanniae (ed-
ited variant and vulgate versions) do not speak of Aeneas’s death. The
antique tradition apart from Virgil is unclear, but none of the Latin clas-
sical authors, to the best of my knowledge, uses lightning as an explana-
tion for Eneas’s death.

65

If we take a closer look at the classical tradition,

we notice that Virgil himself mentions the hero’s apotheosis in Books
I, 259, and XII, 794. That Aeneas became one of the national gods (an
indiges) is also implied in these two references. But Virgil mentions nei-
ther the cause of the death nor the whole process of Aeneas being taken
to heaven. A more thorough account of the motif, however, is given in
Metamorphoses, Book XIV: Venus searched for Aeneas’s body and finally
found it on the banks of the river Numicius (not yet called Tiber) where
it had been lying. Venus then cleaned the body and kept the “melior
pars” to be raised to Heaven. From this moment on, Aeneas was made
one of the divine protectors (indiges) of Rome. Livy has more or less

65. Cf Livy, I, ii. 6, and Dionysius of Halicarnassus, Roman Antiquities, I, p. 64. Being

struck by a thunderbolt suggests a punishment for hubris, and therefore seems irrelevant

for Aeneas. But if we keep in mind that the figure of Aeneas was used both as a positive and

a negative model in the medieval reception of the Aeneid (a founder of empire and/or a

traitor), the motif of a thunderbolt might seem more relevant.

512

Tétrel

the same rendering in Book I, ii, 6, except that in this account Aeneas
was killed in battle against the Rutulians. Dyonisius of Halicarnassus has
both traditions: Aeneas disappeared in a fight against Mezencius, “and
when the body of Aeneas was nowhere to be seen, some concluded that
it had been translated to the gods and others that he had perished in
the river beside which the battle was fought.”

66

These combined expla-

nations account for all the variants offered by the French Brut and the
Aeneas section of the Histoire Ancienne except one: the death by light-
ning, which is exactly the one the author has in common with the Old
Icelandic texts. I have not found any better solution than following the
French author himself, when he attributes this version to Ovid, but as-
sume that a transfer occurred in the source: in the Metamorphoses, Book
XIV, the section immediately following that on Aeneas indiges is a short
account about the first kings of Alba starting with Ascanius. One of the
latter’s offspring, Remulus, brother of Acrota, is killed by a thunderbolt
(as a punishment). The transfer was possibly made to Aeneas’s episode
from this tale, which occurs just a few lines later. Besides, the choice of
Ovid as an authority most probably betrays Servius’s direct or indirect
influence. A closer match with the French Brut, both in content and
style, can be found in Servius’s comment on Aeneid, Book I, 259:

Aeneas. . . secundum quosdam in Numicum cecidit fluvium, secundum Ovi-

dum in caelum raptus est, appellatus tamen est Iuppiter Indiges.

67

(Aeneas, according to them, fell into the river Numicus; according to Ovid,

he was taken to heaven, and he is called Iuppiter Indiges.)

According to Bömer, it is understandable that Servius should quote Ovid
rather than Virgil, because the former refers to a fact while the latter only
utters a prophecy.

68

Thus I allow myself to reconstruct the story of Aeneas’s death, as it stands
in the Breta sögur (shorter and longer versions) and in AM 764 4to, as the
following combination, which must have been available for the author of
the Histoire Ancienne and the author of the French Croniques: the episode of
Aeneas’s death derived from Ovid’s Metamorphoses, transmitted through a
Servian commentary, and in the course of the transmission, a transfer may

66. Dionysius of Halicarnassus, I, p. 64 (Translation from: Roman Antiquities, Books I-II,

trans. by Earnest Cary [Cambridge, MA: Harvard Univ. Press, 1937]).

67. Franz Bömer, P. Ovidius Naso, Metamorphosen, Buch XIV–XV, Kommentar (Heidelberg:

Carl Winter Unversitätsverlag, 1986), pp. 146–97. Pp. 154–55 are dedicated to “Aeneas

Indiges” (Servius here quoted from Bömer). Servius was used by the First Mythographer,

fabula 199, on Aeneas’s journey. The fabula mentions two versions (Aeneas taken to heaven

or killed in a battle against Mezencius), but does not know of a death by lightning.

68. “Servius nennt hier wohl deswegen Ovid und nicht Vergil, weil dieser die Tatsache

(raptus est), Vergil nur die Prophezeihung berichtet” (Bömer, Kommentar, p. 155).

Trojan Origins

513

have been made from Remulus’s death to Aeneas’s.

69

Whether the precise

source is extant somewhere, either in Latin or in French, is a question I
cannot yet answer. But we can be sure that the variant involving Aeneas
being killed by lightning was not made up by the Old Icelandic author or
copied by the scribe of AM 764 4to, Haukr, or the author of the longer
version, nor was it invented by the French author.
These observations, which may be completed or modified by others
in the future, lead to the following conclusions concerning the sources
of “Aeneas’s saga”: The first conclusion, which seems firmly established,
is that both versions of the Breta sögur rely on the same source for this
part of the text. The fact that this “saga” was obviously intended to serve
as a preamble, an independent introductory chapter, does not mean, of
course, that the conclusion is valid for the entirety of the Breta sögur. To be
confirmed, this theory has to be verified on the basis of other parts of the
text. The second conclusion concerns the community of sources of our
Icelandic corpus. As I have tried to show in this study, there is evidence
that both versions of the Breta sögur, Veraldar saga, and the universal history
contained in AM 764 4to made use of similar materials. I have further
tried to show that there are connections with the kind of work represented
by the French Histoire Ancienne jusqu’à César and the French prose Bruts,
which are to be found in similar contexts. In all these texts, the connection
originated in a common pattern uniting the story of the fall of Troy, the
story of the origins of Rome, and the story of the founding of Britain. The
easiest way to explain the connections seems to be to postulate a common
use of materials. In Iceland, these materials could have been found either
in a universal history devoted to the secular “empires,” the Roman empire
being one of them, with a large extension of the Roman sections drawing
on Paulus Diaconus, or it could have been a copy of the Historia Romana
itself, in one of its rewritings at least. It could have also been transmitted
either alone, or together with the History of the Langobards. The common
use of the Historia Romana, both in a “British” inflected history and in a
similar “Roman” version, is easily explained, since the materials are often
contiguous in universal compilations. Furthermore, we note a certain
influence from Servius’s commentary in relation to Virgilian and Ovidian
interpolations, which might suggest the existence of a secondary source,
possibly a source imported at a later date.

69. If there is no tradition of Aeneas being killed by a thunderbolt, such a tradition exists

about Anchises, Aeneas’s father. Servius, In Aen. I, 617, explains why Anchises lost an eye:

Venus asked Jupiter to strike him with lightning as punishment for his arrogance and lack

of discretion. But when the thunderbolt fell, she showed mercy and tried to intercept it.

She was only partially successful, and Anchises lost an eye. A similar reading can be found

in the Second Mythographer, fabula 222.

514

Tétrel