Mode of Reception and Function of Medieval Texts: A Comparative Study of Elye de Saint-Gilles and Elis saga ok Rósamundu

Journal of English and Germanic Philology—January© 2013 by the Board of Trustees of the University of Illinois

Mode of Reception and Function of Medieval Texts: A Comparative Study

of Elye de Saint-Gilles and Elis saga ok Rósamundu

Stefka G. Eriksen, University of Oslo

The aim of this article is to analyze the function and mode of reception1 of variant versions of a medieval text2 as they appear in different codico-logical, cultural, and social contexts. The text in question is the story of Elye, which was translated from Old French to Old Norse in the thirteenth century. The Old French chanson de geste Elye de Saint-Gilles originated in the twelfth century but is preserved in only one medieval manuscript, Paris, Bibliothèque nationale de France, fr. 25516 (fols. 76r–95r), from North East France, dated ca. 1280. The manuscript is associated with the House of Flanders. The Old Norse translation Elis saga ok Rósamundu was commissioned by King Hákon Hákonarson of Norway (r. 1217–63), according to the testimony of the oldest manuscript, and is preserved in one Norwegian manuscript and a number of Icelandic vellum and paper manuscripts and fragments.3 The following is a comparison of the Old French version with the Old Norse version found in the Norwegian manuscript De la Gardie 4–7 folio (DG 4–7),4 dated to ca. 1270.5 This is the oldest version of the saga, and is considered to be a copy of a copy of the actual translation. The manuscript was written in southwest Norway, possibly around the town of Bergen, either in the royal chancellery or in the nearby Lyse abbey, both of which were scribal milieus of sufficient

1. Mode of reception refers to the manner in which a text was received, that is, visually, aurally, or a combination of these. 2. To distinguish between various levels of a text, I use “text” to signify an abstract entity. “Version” or “variant” refers to each manifestation of a “text” in a manuscript. “Manuscript” signifies the materiality of a book. 3. See Ordbog over det Norrøne Prosasprog. Registre (Copenhagen: Den arnamagnæanske kommission, 1989), p. 236; and Foster Blaisdell, “Elis saga ok Rósamundu: Holm 7 – AM 119 – Holm 17 – Holm 47,” Opuscula, VIII (1985), 153–57, on the relationship between some of the extant manuscripts. 4. Uppsala, University Library in Uppsala, De la Gardie collection 4–7 folio, fols. 6r–17v. 5. For a discussion of Elis saga in the manuscripts Holm Perg 6 4to and Holm Perg 7 fol., see Stefka Eriksen,“The Materiality of Medieval Texts: A Comparison between Elie de Saint-Gille and two versions of Elis saga,” Arkiv för Nordisk Filologi, 124 (2009), 69–88; and Eriksen, “Writing and Reading in Medieval Manuscript Culture. The Transmission of the Story of Elye in Old French and Old Norse Literary Contexts” (PhD diss., Univ. of Oslo, 2010).

stature to produce a manuscript like DG 4–7.6 The French and Norwegian manuscripts are almost contemporary, yet there is no direct relationship between them; that is, the Norwegian version is certainly not based on the extant French version, since the latter is somewhat younger. The two manuscripts are linked to two great political and cultural milieus and it is the mode of reception of the two versions and their function in their respective contexts that will be discussed here.

THEORETICAL INTRODUCTION

A text’s mode of reception may be seen as an aspect of the relationship between orality and literacy in medieval culture. According to the so-called oral-formulaic tradition7 and great-divide tradition,8 orality and literacy are seen as the opposing ends in a dichotomy. In recent research, it has become common to focus on the dynamic and intertwined relation-ship between orality and literacy. In addition, the distinction between the modes of composition and the transmission and reception of a text has been emphasized, since these may relate to orality and literacy in vary-ing degrees.9 Whereas a written text may originally have been composed orally, it may subsequently have been written down with the intention of being publicly read or orally performed. The mode of reception of a text may be approached in various ways. One method is to study the prologue, epilogue, and other relevant parts of a text for explicit information concerning the mode of reception in such expressions as: “As you will hear,” “As you have heard,” “. . . those who

6. See Mattias Tveitane, introduction to Elis saga, Strengleikar and Other Texts. Uppsala University Library Delagardieska samlingen Nos. 4–7 folio and AM 666b quarto, Corpus Codicum Norvegicorum Medii Ævi. Quarto Serie 4 (Oslo: Selskapet til utgivelse af gamle norske hånd-skrifter, 1972), p. 13; and Ludvig Holm Olsen, Den gammelnorske oversettelsen av Pamphilus, med en undersøkelse av paleografi og lydverk (Oslo: Det Norske Videnskaps-Akademi i Oslo, 1940), p. 83. 7. See Albert Lord, The Singer of Tales (Cambridge, MA: Harvard Univ. Press, 1960), and Milman Parry, The Making of Homeric Verse (Oxford: Clarendon, 1971); and John Miles Foley, Theory of Oral Composition. History and Methodology (Bloomington: Indiana Univ. Press, 1988). 8. In his Orality and Literacy. The Technologizing of the Word (London: Methuen, 1982), Walter Ong addresses the orality-literacy debate through the “Great Divide” theory. See also Jack Goody, The Domestication of the Savage Mind (Cambridge: Cambridge Univ. Press, 1984); and Erik Havelock, The Muse Learns to Write. Reflections on Orality and Literacy from Antiquity to the Present (New Haven: Yale Univ. Press, 1986). 9. See Michael Clanchy, From Memory to Written Record. England 1066–1307, 3rd ed. (Oxford: Blackwell, 1997); D. H. Green, Medieval Listening and Reading. The Primary Reception of German Literature 800–1300 (Cambridge: Cambridge Univ. Press, 1994); and Joyce Coleman, Public Reading and the Reading Public in Late Medieval England and France (Cambridge: Cambridge Univ. Press, 1996).

2 Eriksen

read and hear,” etc.10 Such expressions have been interpreted as residual formulae, without directly signifying the mode of reception intended by the author/scribe. Their formulaic character does not, however, contradict an interpretation with regard to the mode of reception, since formulas may have been used consciously by an author/scribe to facilitate the flow of the text and make it appropriate for an oral performance. Studying rhetorical features such as formulae, addresses to the audience, a narrator’s comments, the use of direct speech, and the use of temporal vs. spatial ad-verbs is another approach to the discussion of a text’s mode of reception.11 The use of such rhetorical devices has also been interpreted either as an indicator of the oral composition of a text, especially by the advocates of the oral-formulaic theory, or as a tool used by a consciously writing author who intended to create a certain rhetorical flow appropriate for a specific mode of reception. It is the latter approach that will be pursued here. Yet another method of elucidating the mode of reception is to study various grammatical aspects of the language in a text. For example, using various colloquial grammatical constructions in a written text may have been a way of adapting it to an intended listening audience.12 Further, the style and the rhythm of a text, as expressed through the extensive use of alliteration, end rhyme, assonance, and other metrical patterns, may also indicate that the text was intended to be listened to. The narrative structure of a text is also a relevant aspect for establishing how the dramaturgical line of the plot may have functioned, depending on whether the text was heard or read privately. A final method, distinct from the others, is to investigate

10. See studies of Old French sources by Ruth Crosby, “Oral Delivery in the Middle Ages,” Speculum, 11(1936), 88–110; Coleman, Public Reading and the Reading Public; and Evelyn Birge Vitz, Orality and Performance in Early French Romances (Woodbridge: D. S. Brewer, 1999). Clanchy, From Memory to Written Record, discusses the significance of such terms in Latin, Middle English, and Old French. Green, Medieval Listening and Reading, pp. 36–53, focuses on Middle High German language and texts. See also Terje Spurkland, who has studied the respective Old Norse terms in various sources: “Måtte Herren hjelpe den mann som ristet disse runer og likså han som leser dem!,” Norsk Lingvistisk Tidskrift, 12 (1994), 3–15, and “Þeir báru fram bréf ok segja ørendi þau sem fylgðu. Om brevveksling i middelalderen,” in Den Nordiske Renessansen i Høymiddelalderen, ed. Jón Viðar Sigurðsson and Preben Meulengracht Sørensen (Oslo: Historisk institutt, Universitetet i Oslo, 2000), pp. 47–63. 11. Various rhetorical devices have been discussed from the perspective of orality and literacy in respect to composition, transmission, and reception. See, for instance, Green, Medieval Listening and Reading; Suzanne Fleischmann, “A Linguistic Perspective on the Laisses Similaires: Orality and the Pragmatics of Narrative Discourse,” Romance Philology 42(1989), 70–89; Fleischmann, “Philology, Linguistics, and the Discourse of the Medieval Text,” Specu-lum, 65 (1990), 19–37; and Keith Busby, Codex and Context. Reading Old French Verse Narrative in Manuscript, 2 vols. (Amsterdam: Rodopi, 2002). 12. See Fleischman, “Philology, Linguistics, and the Discourse of the Medieval Text”; and Fleischman, Tense and Narrativity: From Medieval Performance to Modern Fiction (Austin: Univ. of Texas Press, 1990).

Mode of Reception and Function of Medieval Texts 3

the way the text appears in the manuscript, that is, the way it is written (e.g., abbreviated and punctuated)13 and the way it is arranged and struc-tured on the manuscript page (by means of illustrations, rubrics, initials, etc.).14 Visual and graphical highlights on the manuscript page may have served to guide a reader during a vocal performance or to facilitate the comprehension of the text during a private silent reading. The manifold methods of analyzing a text’s reception may sometimes point in different directions. Seemingly conflicting evidence may indi-cate that one and the same text was intended to have been both publicly and privately read, both listened to and looked upon, in various contexts and settings. The various aspects of a text are, however, seldom taken into consideration collectively when the mode of reception is discussed. How does the materiality of a manuscript, for example, relate and correspond to the rhetorical, lexical, grammatical, or narrative aspects of the text? In this article an eclectic approach will be pursued through an analysis of the correspondence between one graphical aspect of the manuscript—the occurrence of initials—and various textual aspects.15 The initials may have provided guidance during the reading of the manuscript, both public and private. Sometimes initials may correspond to textual aspects, thereby indicating that the text was to be listened to. It is plausible to argue that the graphical emphasis was intended to be vocalized by means of textual emphasis. At other times, initials may correspond to textual clues suggest-ing an intended visual reception, thus making plausible that private read-ing is intended. Studying the correspondence between these two aspects (graphical and textual) offers a means of synthesizing existing research on the mode of reception of Old French and Old Norse literature. At the same time, it provides a new means of commenting on the issue and more interestingly it elucidates how a manuscript’s layout and presentation may reflect a certain mode of reception. The method, even though here applied on limited material, is thus potentially relevant for and applicable to any medieval manuscript material, regardless of such traditional categories as genre, language, and culture.16

13. See M. B. Parkes, Pause and Effect. An Introduction to the History of Punctuation in the West (Berkeley: Univ. of California Press, 1993); and Albert Derolez, The Palaeography of Gothic Manuscript Books. From the Twelfth to the Early Sixteenth Century (Cambridge: Cambridge Univ. Press, 2003). 14. See Paul Saenger, “Silent Reading: Its Impact on Late Medieval Script and Society,” Viator, 13 (1982), 367–414; Saenger, Space Between Words: The Origin of Silent Reading (Stan-ford: Stanford Univ. Press, 1997); and Busby, Codex and Context. 15. Note that other graphical aspects, such as titles, rubrics, illustrations, majuscules, and also abbreviations, word-clusters, and corrections, may be studied from the same perspective as well. For a theoretical discussion and a survey of these graphical aspects in three versions of Elis saga, see Eriksen, “Writing and Reading in Medieval Manuscript Culture.” 16. The function of initials in other texts and genres has been commented on by Richard

4 Eriksen

THE STORY OF ELYE TRANSLATED

Elis saga ok Rósamundu is traditionally regarded as belonging to the genre of the riddarasögur, which includes translations of Old French romances, chansons de geste, lais, and fabliaux. Sometimes translations from Middle High German, such as Þiðreks saga, and Latin, such as Alexanders saga, Breta sögur, Trójumanna saga, are also included in the genre.17 King Hákon Hákonarson is traditionally considered as one of the main promoters of translation of Latin and vernacular literature. In 1226, he commissioned the translation of Thomas’s Tristan. Four other texts mention the king as having commissioned a translation, one of which is Elis saga, but other texts may have been translated under similar circumstances. The com-mission of translations may be seen as part of an extensive political and cultural program initiated by the king, which entailed peace within the kingdom, revision of the provincial laws leading to the development of a national law, external expansion, development of strong personal and political contacts with various European kings, to mention just some of King Hákon’s many achievements.18 From a literary point of view, the aim of the translation program may have been to contribute to the position-ing of Old Norse language and literature within the European Latin and vernacular polysystem. With a new type of literature, there may also have been a new awareness of writing19 and possibly new attitudes to the recep-tion and function of books and literature in general.

THE CONTENT OF THE TWO VERSIONS

The Old French and Old Norse versions of the story of Elye tell substan-tially the same tale, but there are also some differences. The beginning

Hartmann, “Initials and Laisse Division in Two Later Epics: Aiol and Parise la Duchesse,” Olifant, 12 (1987), 5−27; Hans Runte, “Initial Readers of Chrétien de Troyes,” in Continuations. Essays on Medieval French Literature and Language. In Honor of John L. Grigsby, ed. Norris J. Lacy and Gloria Torrini-Roblin (Birmingham, AL: Summa Publications 1989), pp. 121−32; Roger Middleton, “Coloured Capitals in the Manuscripts of Erec et Enide,” in Les Manuscrits de Chrétien de Troyes: The Manuscripts of Chrétien de Troyes, ed. Keith Busby, Terry Nixon, Alison Stones, and Lorri Walters (Amsterdam: Rodopi, 1993), pp. 149−94. 17. For a survey of the corpus of the translated riddarasögur, see Jürg Glauser, “Romances (Translated Riddarasögur),” in A Companion to Old Norse-Icelandic Literature and Culture, ed. Rory Mc Turk (Malden, MA: Blackwell Publishing, 2005), pp. 272–387. 18. Sverre Bagge, “The Norwegian Monarchy in the Thirteenth Century,” in Kings and Kingship in Medieval Europe, ed. Anne J. Duggan (London: King’s College London Centre for Late Antique and Medieval Studies, 1993), pp. 159–77. 19. See Jürg Glauser, “Staging the Text. On the Development of a Consciousness of Writ-ing in the Norwegian and Icelandic Literature of the Middle Ages,” in Along the Oral-Written


of the Old French poem20 tells how Elye, after being challenged by his father, proves himself worthy of being a knight, is dubbed, and leaves home to seek adventure. He undergoes a series of troublesome episodes on his journey: he fights against Saracens to help a group of Christian warriors, and is captured and abducted on a Saracen ship, from which he later escapes. This leads to his meeting his fellow comrade, Galopin, a former thief, with whom Elye once more fights against Saracens. Un-fortunately, Elye is wounded, and Galopin helps him hide in the gardens of a castle, which turns out to be the home of his enemies, but also of Rosamunde, a famously beautiful Saracen princess. She not only heals his wounds and saves his life but eventually she is also baptized because of her love for Elye. Elye fights several more battles against various Saracens, but at the end, he is compelled to wait for military assistance. A great Christian army, consisting of the king and his well-known vassals, comes to his help. They take over the Saracen castle, baptize all the people, and turn it into a Christian stronghold. Because Elye has sponsored Rosamunde’s christening, the two cannot be married. Galopin and Rosa-munde are married instead and become the rulers of the new Christian stronghold, while Elye, the king, and the army return to France, where Elye is married to the king’s sister. Elis saga in DG 4–7 differs somewhat from the French version, chiefly inasmuch as the text is not extant in its entirety. There is a lacuna, which must have contained the account of Elis’s capture, abduction on the Sara-cen ship, and later escape.21 The other major difference concerns the conclusion of the Old Norse saga. The saga ends prior to the grand battle between Christians and Saracens, which is only barely suggested as a future possibility. A possible future marriage between Elis and Rósamunda is also only suggested.22 The lacking final scene in the Norwegian translation is thus a major difference between the two versions.

Continuum: Types of Texts, Relations and Their Implications, ed. S. Rancović, L. Melve, and E. Mundal (Turnhout: Brepols, 2010), pp. 14–52. 20. Gaston Raynaud, Elie de Saint Gille: chanson de geste, publié avec introduction, glossaire et index. Accompagnée de la rédaction norvégienne, trad. par Eugene Koelbing (Paris: Firmin Didot, 1879), p. xiii; Patrick de Winter, La bibliothèque de Philippe le Hardi, duc de Bourgogne (1364–1404): étude sur les manuscrits à peintures d’une collection princière à l’époque du style gothique international (Paris: Editions du Centre national de la recherche scientifique, 1985), p. 234; Busby, Codex and Context, pp. 534–35. 21. The lacuna in DG 4–7 is filled by the version in AM 533 4to in Kölbing’s edition. 22. Note that the conclusion of the version in the Icelandic manuscripts differs from that in DG 4–7, since the final grand battle is accounted for. However, this conclusion differs from that of the Old French version as well.

6 Eriksen

INITIALS: SIZE AND STYLE

Keeping in mind the cultural background and the differences between the two versions, the following is devoted to the presentation and analysis of their graphical appearance and its relationship to the text. The Old French poem is structured by initials of various sizes, which oc-cur at the beginning of the laisses (see fig.1). The first initial, a historiated initial, spans eleven lines and the width of one column of the text. The majority of the initials in Elye are two-line capitals. They are simple with no extra decoration, appear alternatively in red and blue, and structure the poem visually.23 In addition, there are a few three-line pen-flourished initials that appear in connection with illustrations. In all there are six miniatures (one in an initial) in the poem, which structure the text on yet another level. All the initials in the Old Norse version span two lines, except for the intended initial at the beginning of the saga, which would have spanned five lines (see fig. 2), but which is missing, as are some of the two-line initials.24 The initials are relatively simple: either green with red details, or red with green details. In the following, the placement of the initials in the two versions will be considered in relation to four different textual aspects: the rhythm of the text, various rhetorical devices, lexical evidence concerning the mode of reception, and the narrative structure of the texts.

INITIALS AND RHYTHM

The initials in the Old French version correspond to the rhythmical pat-tern of the poem, since a change in assonance occurs with every initial but one.25 A basic feature of chansons de geste is that they are structured in laisses, or a series of verses, varying from two to three lines to several hundred, which have the same final assonance.26 The beginning of a new

23. The style of the initials may be a guide to the dating of a manuscript and may point to the place of origin. This is not a primary aim here, but for a detailed study on the issue, see Sonia Scott-Fleming, The Analysis of Pen Flourishing in Thirteenth-Century Manuscripts (Leiden: E. J. Brill, 1989). 24. For possible explanations, see Tveitane, Elis saga, Strengleikar and Other Texts, p. 16; and Holm Olsen, Den gammelnorske oversettelsen av Pamphilus, p. 14. 25. For a presentation of the various types of assonance, see Raynaud, Elie de Saint Gille, pp. x–xi. 26. For a more detailed description of the rhythmical peculiarities characteristic of chansons de geste, see Gérard Gros and Marie-Madeleine Fragonard, Les Formes poétiques du Moyen Âge à la Renaissance (Paris: Nathan, 1995).


Fig. 1. Front page of Elye de Saint-Gilles, fol. 76, BnF fr. 25516, ca. 1280.

Fig. 2. Front page of Elis saga, fol. 6r, DG 4–7 fol., ca. 1270.

laisse is most often marked by an initial in the manuscript, even though there are exceptions. In other words, the beginning of a new laisse, which is graphically marked by an initial in the manuscript, would also have been audible to a listening audience because of the change of assonance from the preceding laisse. The three lines just above the initial in the second column in figure 1, which initiates the second laisse, and the three lines from this new laisse are an example. The assonance, in this case, changes from -i- to -a-:27

Les huis ont desfremés et les cambres ovrirent;Elye i ont trové et sa seror Olive.Droit de devant lor pere les menerent et guient. (vv. 29–31)

(They unlocked the doors and opened the rooms;There they found Elye and his sister Olive.Straight before their father they ushered and brought them.)

Julïens se sëoit ens el palais de marbre,Tout entor lui sa gent et son barnage.Il les a apelés conme preudon et sage. (vv. 32–34)

(Julien was seated in the marble palace,All around him gathered his people and his noblemen;He spoke to them as a worthy and wise man.)

In addition to assonance, rhythm may also be linked to the initials. Al-though most of the poem Elye is written in dodecasyllabic form, with the caesura dividing the verse into 6+6 syllables, some of the verses connected to initials are decasyllabic (6+4 or 4+6). In the notes to their edition of Elye de Saint-Gilles, A. Richard Hartman and Sandra Malicote suggest that the decasyllabic lines emphasize dramatic moments in the narrative.28 Thus, the graphical emphasis of the initials would have been expressed audibly through the shift in assonance and the shortness of the verse. The Old Norse saga is in prose, and while it is generally less rhythmical than a chanson de geste, rhythm is produced through alliteration and cursus. Alliteration occurs at the beginning of Chapters 1 and 31:29

HÆYRIT horskir menn. æina fagra saugu. dyrlegs drengscaps. um raustan riddera scap & lofsæla atgerð. (6ra15; p. 1)30

(LISTEN, wise men, to a beautiful story of glorious manhood, of valiant chivalry, of glorious deeds.)

27. Quotations, English translations, and laisse numbers are from A. Richard Hartman and Sandra Malicote, Elye of Saint-Gilles. A Chanson de Geste (New York: Italica Press, 2011). 28. Hartmann and Malicote, Elye, notes to ll. 2026, 2039, 2049 (p. 220); 2377 (p. 226). 29. I use the term “chapter” to designate the section between two initials, for lack of a specific term corresponding to Old French laisse. 30. The references indicate folio number, recto or verso, column (a or b), and line num-ber and are based on Tveitane’s facsimile edition, Elis saga, Strengleikar and Other Texts. For a critical edition of Elis saga, see Kölbing, Elis saga ok Rosamundu, p. 1. Subsequent references to pagination are to this edition. The English translation is mine.

10 Eriksen

ROSamunda hin kurtæisa hin friða & hin frægia hin liosa & hin lofsæla (12vb31; p. 76)

(Rósamunda, the courteous, the beautiful and famous, the bright and the glorious)

Alliterating word pairs are used regularly throughout the text as well, and not only in relation to initials.31 Traditionally, alliteration is regarded as a main feature of the “courtly style” of the translated riddarasögur,32 such as in the Strengleikar33 and in Ívens saga.34 It appears also in texts with learned and native stylistic features, such as in the Old Norse Homily Book,35 Barlaams saga,36 and even the Gulathing Law in Codex Rantzovianus from ca. 1250, Bergen.37

Scholars have debated whether cursus, or a prose rhythm influenced by cursus, can be detected in Old Norse prose texts, such as Elis saga among others.38 Cursus is a rhythmical pattern commonly used in Latin, which is formed when a set number of syllables appear between two stressed syl-lables at the end of a sentence. There are various types of cursus, that is, planus, velox, and tardus, and these are defined by the length of the last word. The most significant determinant for cursus is, however, the number of syllables between the two last accents and the number of syllables after the last accent, and not where a new word begins.39

31. Some examples are: gæfu goðlæikr (6ra23; p.1), hirðlegs haversclæiks (6ra25; p. 1), miclar oc milðar and miscunnsamar (6ra35; p. 2), valkum oc vanðræðum (7ra34; p. 12), særðan and suivirt (7va36; p. 17), virðuliger vinur (8ra31; p. 22), rikustu ridderar (8ra32; p. 22), feginn af fagnaðe (10va12; p. 42), herra kuað hinn hundhæiðni hunndr (13rb39; p. 81), oð and œr (14ra28; p. 88), festa honom fullan frið (14rb13; p. 89), ne angr ne aræðe (14rb15; p. 89), er þæím hað and hegomi. snæyping & suivirðing (14rb32; p. 90), vaskliga vaxinn and vel mannaðr (6rb20; p. 3), felauss and fulgðar laus (6va8; p. 5). 32. For a discussion of the different stylistic features of Old Norse literature, see Reidar Astås, “Lærd stil, Høvisk stil, Florissant stil i norrøn prosa,” Maal og Minne (1987), 24–38. 33. See Ingvild Budal, “Strengleikar og Lais. Høviske noveller i omsetjing frå gammalfransk til gammalnorsk. I. Tekstanalyse. II. Synoptisk utgåve” (PhD diss., Univ. of Bergen, 2009), chap. 7. 34. Marianne Kalinke, “Alliteration in Ívens saga,” Modern Language Review, 74 (1979), 871–83. 35. See Trygve Knudsen, “Innledning,” in Gammelnorsk Homiliebok etter AM619 qv. (Oslo: Selskapet til utgivelse av gamle norske håndskrifter, 1952), p.7. 36. Magnus Rindal, Barlaams ok Josaphats Saga. Manuscript no. 6 fol. in the Royal Library, Stockholm and the Norwegian Fragments, Corpus Codicum Norvegicorum Medii Aevi. Quarto Series (Oslo: Selskapet til utgivelse av gamle norske håndskrifter, 1980), pp. vii–xx. 37. Finn Hødnebø, “Tinget og Loven,” in Den Eldre Gulatingsloven, E Donatione Variorum 137 4to (Codex Rantzovianus) i Det Kongelige Bibliotek, København oc AM 309 fol. (93r−100v), AM 315 E fol., AM 315 F fol., AM 468 C 12 (Bindet), NRA 1B, ed. Finn Hødnebø and Magnus Rindal (Oslo: Selskapet til utgivelse av gamle norske håndskrifter, 1995), pp. 9–20. 38. For a discussion of cursus, see the following essays by Jakob Benediktsson, in Lærdómslistir. Afmælisrit 20. júlí 1987, ed. Halldór Guðmundsson, Sverrir Tómasson, Örnólfur Thorsson (Reykjavík: Mál og menning, 1987): “Cursus in Old Norse Literature,” pp. 175–82; “Cursus hos Bergr Sokkason,” pp. 262–69; “Traces of Latin Prose-Rhythm in Old Norse Literature,” pp. 153–60. See also Kirsten Berg, “Cursus i norrønt? – en metodediskusjon med eksempler,” Maal og Minne (1999), 165–86. 39. Berg, “Cursus i norrønt?” p.166.


Jakob Benediktsson has argued that cursus was adapted from Latin into Old Norse, especially in translations of religious works from the latter half of the twelfth century and in later literature featuring a courtly style. Kirst-en Berg has noted the difficulty of defining cursus, since it is influenced by the punctuation in a manuscript, the merging of words, and the style of a text. These aspects are not always consistent in a manuscript, which is problematic if one is to argue for a conscious use of cursus. Berg claims to find such use of cursus or other forms of prose rhythm in several translated sagas from the second half of the thirteenth century, namely, Barlaams ok Josaphats saga, Konungs skuggsjá, and Elis saga. According to Berg, it is plausible that cursus was used as a conscious strategy in the version of Elis saga in DG 4–7, one resembling that found in Niðrstigningarsaga in the manuscript AM 645 4to from the first half of the thirteenth century. In a sample study of the first six chapters of Elis saga, Berg discovers that the most common type of cursus is planus.40 Berg does not give any specific examples from Elis saga, but claims that planus, tardus, or velox appears at the end of 39.6 percent of the sentences in the first part of the saga. Such a statement is not only difficult to cross-check but also debatable, since Berg uses Kölbing’s edition as a base for her study. In the edition, the punctuation, which is one of the main determinators for cursus use, is regularly emended vis-à-vis that occurring in DG 4–7. The rare use of alliteration in relation to initials and the debatable occur-rence of cursus make it difficult to argue for the conscious use of rhythmical patterns at the end of sentences and chapters in Elis saga in DG 4–7. The correspondence between rhythmical and graphical patterns is much more consistent in the Old French poem than in the Old Norse saga.

INITIALS AND RHETORICAL DEVICES

The initials in the Old French text correspond to various rhetorical de-vices. Some of these are narrator’s comments, which appear either at the end or at the beginning of a laisse. The comments vary in character: some include emotional exclamations and appeals to God; others refer backward or forward in the story; still others are of a proverbial nature. The following are examples of the various kinds of comments:

1. Narrator’s comment that refers to the further development of the plot, end of laisse 6:

Il troverent Elye a l’issir d’un boscage;Mais ançois qu’il s’en partent, lor fera tel damage,K’escus i avra frais et armes de cors traits. (vv. 277–79)

40. Berg, “Cursus i norrønt?” p. 175, p. 177.

12 Eriksen

(They found Elye, as he was leaving the woods,But before they’ll take leave of him, he’ll do them such harmThat he will have fresh shields and arms taken from their bodies.)

2. Comments that reveal the narrator’s personal opinion in an emo-tional fashion, from the beginning of laisse 23:

Mout fu grant la bataille et li estor campés;Dieus! Con i fiert Guillame, li marcis au cort nés! (vv. 710–11)

(The fighting was very heavy, and the battle raged,God! How William struck blows, the short-nosed marquis!)

3. Narrator’s comment that links real-world knowledge to the narra-tive plot, from the beginning of laisse 22:

Ausi con li faucons fait les oiseus füir,Fait Guillame d’Orenge paien et Sarrasin. (vv. 659–60)

(Just as the falcon sends the other birds fleeing,So did William of Orange pursue the pagans.)

Narratorial comments also occur in relation to initials in the Old Norse translation, often at the end of a chapter and before an initial. Thence, the initials marking the narrator’s comments would have helped the reader vocalize the discourse appropriately during a public reading of both ver-sions of the text. The narrator’s comments are, however, generally less emotional in the saga than in the Old French poem. Some of them are intertextual in character and refer to the Christian faith of the narrator, for example, one that appears at the end of Chapter 6:

Nu mœta þæír elisi sem hann ræið orskoginum. & er nu vón at hann bríoti spiotscapt sitt. fyrr en hann braút komiz. nu gæti guð hans her & huervetna. er huervetna gætír í sinni miskunn. (8ra22; p. 21)

(Now they meet Elis who was riding out of the wood. And it is now to be ex-pected that he break the shaft of his spear before he gets away. Let now God, who watches over everywhere in his mercy, protect him here and everywhere.)

Sometimes the narrator expresses his personal evaluation of the situation, for example, at the end of Chapter 33:

& er æigi nu kynlict hann ferr læiðar sinnar. sem skiotazt ma hann. (13rb18; p. 21)

(And now it is not strange that he goes on his way as fast as he can.)

The whole of Chapter 14 is an address to the audience by the narrator, which has a proverbial character:

Nu lyðit goðgæfliga. betra er fogr froðe en kuiðar fylli. þo scal við saugu súpa. en æí ofmikit drecka sœmð. er saugu at segia ef hæyrenðr til lyða. en tapat starfi at hafna at hæyra. (9va1; p. 33)


(Now listen carefully! Fine learning is better than a filled stomach; one is to drink when stories are told, but not too much; it is an honor to tell a story, if people listen, but it is a wasted effort, if nobody listens.)

A second rhetorical device that appears in relation to initials is repeti-tion. There are various types of repetition in the Old French text. The first type occurs when important information from one laisse is given at the beginning of the next (see the beginnings of laisses 7, 8, 16, and 36). Sometimes the repetition also involves lexical repetition with variation. The opening of laisse 11 reads:

Il hurte le destrier et li lasque le resne, (v. 382)he spurred on his warhorse and gave it rein.

and is almost identical to the last line of the previous laisse:

Il hurte le destrier et le resne li lasque. (v. 381)he spurred his warhorse on, giving it full rein.

It should be noted that some phrases are repeated often enough to be con-sidered formulaic, and these appear not only in conjunction with initials. Another type of repetition that occurs in the Old French text is of an episodic nature, when a similar event occurs two or three times. Some new details or a tiny twist of plot take place, and this distinguishes the new episodes from previous ones, driving the narration forward. It is the frame or the structure of the episodes that is repeated. It has been observed that the length in the number of verses of such repeated episodes may vary, and it has been suggested that amplifying the episode may be a means of emphasizing it.41 An example of such a repetition may be the beginning of laisse 23 (vv. 710–13), which is relatively short, and the beginning of laisse 24 (vv. 732–42), which is three times longer. In both excerpts the Christian knights, who are caught up among the Saracens, address Elye and urge him to draw closer to them; then, when he is about to do so in laisse 23, a pagan comes forward, and Elye kills him. This leads to a repetition of the episode in laisse 24; the second pagan tries to convince him to give up his Christian faith and to believe in Mahomet, which Elye naturally refuses, and so on. The wording is not exactly the same, but still quite similar. In addition, both times, after the Christians ask for help, the poet adds a verse:

Et il si feïst senpre, bien s’en fu apensés (v. 721)(And he would have done so quickly, he’d already thought of it)

Et il si feïst senpre, bien s’en est percheüs (v. 741)(And he’d have done it immediately, he’d perceived as much himself)

41. Edward A. Heinemann, “Measuring Units of Poetic Discourse: Analogies between Laisse and Verse in the Chansons de Geste,” in Romance Epic. Essays on a Medieval Literary Genre, ed. Hans-Erich Keller (Kalamazoo: Medieval Institute Publications, Western Michigan Univ., 1987), p. 24.

14 Eriksen

Repetition occurs at the beginning of chapters in the Old Norse version as well, but the nature of the repetition is different from that in the Old French version. Many of the repetitions are much shorter, no longer than a simple clause in the first sentence of a chapter, in passages like “Sem elis hafði skilt . . .” (When Elis had understood . . .) (see chaps. 12, 13, 16, 27, and 35). On several occasions there are also lexical repetitions, with variation:

“Goðe faðer kuað mærin” (14ra27; p. 88)(My good father, said the maiden)

“Goðe faðir kuað mærin” (14ra31; p. 88)(My good father, said the maiden)

“Goða dottir kuað konungr” (14rb16; p. 89)(My good daughter, said the king)

“Sem konungrinn skilðe at sonr hans lætk vera síukr” (13rb34; p. 81)(When the king understood that his son was claiming to be sick)

“Konungrinn skilðe orð hans. & at iosi lezk vera siukr” (13va6; p. 82)(The king heard his words that also Josi claimed to be sick)

“þat væít tru mín iungfru kuað herra elis” (14va10; p. 91)(By my faith I know, my lady, said Elis)

“þat væít maghun herra iarl quað mærin” (14va41; p. 92)(By Maghun I know, my sir, said the maiden)

Chapters 35 and 36 are an example of episodic repetition, that is, a simi-lar event occurs two times. But not all episodic repetitions from the Old French are preserved in the Old Norse version, as, for example, the rep-etition of the episode mentioned above (vv. 721 and 741). Another rhetorical device that appears very often in conjunction with initials is the use of temporal adverbs and connectives. As already men-tioned, the extensive use of temporal adverbs, narrator’s comments, and repetitions has been interpreted independently of the physical aspects of the manuscript in which these appear, as tools which render a text suit-able for a listening audience. In the Old French version, the two temporal adverbs are quant and or.

Quant l’entent l’amiraus, por poi que il ne derve (v. 2068)(when the emir heard him, he nearly went out of his mind) (last line of laisse 58)

Or en penst Dameldé, qui tout a a jugier (v. 1950)(Now may the Lord God grant His blessing, Who has dominion over all) (at the end of laisse 55)

The initials in the Old Norse version also occur in conjunction with tem-poral adverbs. In Old Norse, however, the temporal adverbs are more varied, and include nu, sem, nu sem, þa, en nu siðan, etc.


Nu ríða haufðingiar. fim firer þæím er herteknir ero (8ra27; p. 21)(Now the chieftains ride, five in front of those who had been captured)

SEM kursot or tabarie læít rodoan daudan avellinnom (8va6; p. 25)(When Kursot from Tabarie saw Rodoan dead on the plain)

Þa er Josiar sa malatret systur son sinn. liggia (9rb1; p. 31)(When Josiar saw Malatret, his nephew, lie )

EN nu siðan er þiofrinn hafði comit augum a hestinn (15ra23; p. 96)(And once the thief had seen the horse)

Thus, the same types of rhetorical devices are used in conjunction with the initials in both versions, but these are used in different ways. In the Old Norse version, the narrator’s comments are fewer and less emotional; the repetitions are also fewer and shorter, while more types of temporal adverbs are used.

INITIALS AND LEXICAL EVIDENCE

At the beginning and end of the Old French poem, as well as in other places (possibly for breaks, but not always in relation to initials), there is explicit lexical evidence that the text was meant to be listened to. Here are some examples:

Or faites pais, Signor, que Dieus vous benëie (v. 1, laisse 1)(Now, be silent lords; may God bless you)

Or m’escoutés, Signor, que Dieus grant bien vous doinst (v. 490, laisse 15)(Now listen, my lords, may God give you highest good)

Or escoutés, Signor, que Dieus grant bien vous don (v. 572, laisse 19)(Now listen, my lords, and may God provide well for you)

Or vous dirons noveles du bacheler Elies42 (v. 2083, laisse 60)(Now we’ll give you the news of the young knight Elye)

The very end of the poem is rich in lexical evidence relevant for the in-tended mode of reception.

Menestrel s’en loërent, quant vint al dessevrerEnsi dona li rois sa seror al vis clerA Elye le preu, fil Julïen le ber,Qui fu dus de Saint Gille, si con oï avés.D’Elyë vint Ayous, si con avant orés,Ichi faut li romans de Julïen le ber,Et d’Elyë, son fil, qui tant pot endurer.

42. Hartmann and Malicote, Elye, p. 221. Note that Raynaud changes the whole line to Du bacheler Elie or vous dirons noveles to achieve assonance on “-e.” Hartmann and Malicote keep the reading of the manuscript and argue that the name may have been pronounced alternatively /el jƐs/ for the assonance.

16 Eriksen

Cil engenra Ayoul, qui tant fist a loer,Si con vous m’orés dire, sel volés escouter.Explichit li romans d’Elye (vv. 2753–61)

(The minstrels praised the festivities, when it came time to partThus the king gave in marriage Avisse, his bright-faced sister,To Elye the noble son of worthy Julien,Who was duke of Saint-Gilles, as you have already heard.From Elye came Ayoul, as you will hear presently.Here ends the story of Julien the nobleAnd of Elye his son, who had to endure so much!Elye will beget Ayoul, who did so many praiseworthy deeds,Just as you’ll hear me tell, if you want to listen.Here ends the story of Elye.)

Note that the occurrence of dire, for example, in vv. 1919 and 2083, does not have to indicate oral communication but may be used neutrally in respect to the intended mode of reception and the writing process. How-ever, given the evidence of the examples as a group, the intended aural reception of the text is plausible. Moreover, such comments also appear in the middle of various laisses and do not always correspond to initials. The Old Norse saga contains fewer examples of the type. Only Chap-ters 1 and 14 begin with constructions of the type hæyrit / lydit:

HÆYRIT horskir menn. æina fagra saugu. dyrlegs drengscaps. um raustan riddera scap & lofsæla atgerð. æins uirðulegs hertoga. (6ra15; p. 1)

(LISTEN, wise men, to a beautiful saga of glorious manhood, of valiant chivalry, of the glorious and praiseworthy deeds of a virtuous count.)

NU lydit goðgæfliga . . . . sœmð er saugu at segia ef hæyrende til lyða. en tapat starfi at hafna at hæyra (9va1; p. 33)

(Now listen carefully! . . . it is honorable to tell stories, if people listen, but it is a wasted effort, if nobody listens.)

Thus, even though lexical evidence for the mode of reception is less abun-dant in the Old Norse translation than in the Old French version, such evidence does appear and corresponds to initials in both versions.

INITIALS, ILLUSTRATIONS, AND NARRATIVE STRUCTURE

Initials divide the Old French poem into laisses of various lengths, the length of a laisse varying between 5 and 376 verses. It has been suggested that chansons de geste sometimes may be structured by means of long and short laisses in order to provide rhythm and variation in the poem’s flow. Edward Heinemann, for example, argues that in chansons de geste the individual laisses (which are almost always marked with an initial) do not


always correspond to narrative episodes.43 The correspondence and lack of correspondence between episodes and laisses creates variation in the rhythm of the poem. The alternation of short and long laisses also contrib-utes to creating rhythm. The table below illustrates the correspondence between the length of a laisse and the content of the longest laisses in the poem Elye. The sections consisting of laisses that are of similar and relatively short length, sometimes retell events that may be considered a single scene (laisses 6–14: the dialogue between Elye and the pagans who have captured the Christian knights); at other times, these consist of episodic repetitions that have similar narrative significance (laisses 43–45), present different voices in one dialogue (laisses 47–50), or include a series of changes of scene that drive the narrative forward at a constant speed (laisses 60–68). The regularity in the length of the laisses in these sections corresponds thus to their repetitive, steady, and typical content. The content of the longest laisses, however, seems to be central to the narratological structure of the poem, as illustrated in the table below:

Table 1

Nr of laisse Graphic(length) Narrative Content Illustration

5 (106 vv.) 1. Elye is off on his journey and has his first X (at the very adventure (meets the messenger) beginning of the text)24 (69 vv.) 2. Elye is taken prisoner on the ship27 (122 vv.) 3. Elye escapes from the Saracens’ ship X29 (113 vv.) 4. Elye meets the four thieves, one of whom X is Galopin37 (58 vv.) 5. Elye and Rosamunde meet for the first time X and she cures him47 (138 vv.) 6. Rosamunde’s father decides to marry her to Lubien 56 (71 vv.) 7. Galopin steals a special horse for Elye X69 (376 vv.) 8. The great fight between Christians and X (toward the end Saracens. The Christians win. Sobrie is of the laisse) made a Christian city. Rosamunde is baptized and marries Galopin; Elye is married to the sister of the emperor in Paris.

The plot of the poem, if described in terms of these peaks, is a story about Elye who leaves his father’s house to find adventure. He is taken prisoner on a ship, from which he escapes; he meets his comrade Galo-pin; they end up at Sobrieborg, where they meet Rosamunde; the Sara-

43. Edward A. Heinemann, “Measuring Units of Poetic Discourse,” pp. 21−34; Heine-mann, L’art métrique de la chanson de geste. Essai sur la musicalité du récit (Geneva: Librairie Droz, 1993).

18 Eriksen

cen princess saves Elye’s life and is later about to be married to another man. The poem ends with a great fight and victory for the Christian army, but a somewhat unhappy ending for the couple. In addition, the illustration program of the poem seems tightly related to the narrative structure of the poem, as defined by the length of the laisses, since the subject of three of the miniatures (for narrative sections 3, 4, and 8 in the table above) corresponds directly to the theme of the laisses.44 The first miniature appears in the historiated initial at the very beginning of the narrative and not at the beginning of laisse 5. However, the subject of the miniature and laisse 5 may be seen as connected, since the illustra-tion portrays a courtly counsel scene between Elye’s parents and their knights and ladies in which Elye’s fate is discussed. This court scene is the starting point that leads to Elye’s first adventure. The miniatures for narrative sections 5 and 7 above correspond to two secondary peaks in the plot. Laisse 37, corresponding to the fourth miniature, is the final long laisse in a series of peaks, while laisse 56, accompanied by the fifth miniature, is longer than its surrounding laisses. There exists a credible correspondence between the initials and the narrative structure of the story, since the laisses of similar length contain episodes with similar function, providing a regular pace in the development of the plot, while the longer laisses correspond to the most significant episodes, in which there is a break in the regularity of plot development, and thus they provide a skeleton for the narrative. The Old Norse text is also divided into chapters of various length that vary from 4 to 105 lines. Note that while the Old French poem has 69 laisses, the Old Norse text has only 54 chapters, that is, the textual divisions in the two versions are not the same. The difference may be explained, to some degree, by the lacuna in the manuscript. In addition, as already mentioned, the Old Norse version is shorter than the Old French. Still, it seems that the translator(s)/scribe(s) did not always reproduce the laisse structure of their exemplar in the chapter structure of the Old Norse text. If defined by the length of the chapters, the narrative structure of the Old Norse saga differs considerably from that of the Old French poem. The correspondence between the placement of the initials and the narra-tive structure of the saga is not as remarkable as in the Old French poem. This was possibly a less common strategy in an Old Norse prose text than

44. The function of illustration in the reception of a text is discussed more extensively in Eriksen, “Writing and Reading in Medieval Manuscript Culture,” pp. 151–61. The function of miniatures and illustrations in manuscripts, their relationship to the text, and the impli-cation they have for our knowledge of the cultural context are discussed by Alison Stones, Lori Waters, and Patricia Stirnemann, in Les Manuscrits de Chrétien de Troyes, ed. Busby, Nixon, Stones, and Walters.


in a chanson de geste. The content of the clearest narrative peaks in Elis saga are given below:

Table 2

Chapters Episodes

2–3 Julien challenges Elis to prove his manhood, or otherwise lose his inheritance. Elis is insulted and rejects his father. Elis’s mother defends her son’s position. 5 Elis meets the messenger on the road24 Elis meets the thieves and Galopin30 Elis and Galopin meet Rósamunda and she heals Elis37 The king decides to marry Rósamunda to another man 49 Elis gets out on the field to fight and the others go to watch; Kaifas hits Rósamunda

If the tables for Elye de Saint-Gilles and Elis saga are compared, it turns out that some of the summits are the same, but not all. There is absolute cor-respondence in three of these summits: when Elis meets the messenger on the road (the fifth narrative section in both versions); when he meets the thieves and Galopin (laisse 29 in Old French, chap. 24 in Old Norse); and when Rosamunde’s father decides to marry her to another man (laisse 47 in Elye and chap. 37 in the saga). Otherwise, the description of the starting point for Elis’s search for adventure is important in the Old Norse text but is not emphasized equally in the Old French. The same scene is depicted, however, in the first illuminated initial of the poem. The episode in which Elis and Galopin meet Rósamunda and she heals Elis appears as a summit in the Old Norse version, but only as a secondary summit in the Old French. The episode in which Galopin steals the horse is a secondary summit in the Old French poem and does not appear as a major episode in the Old Norse version. Further, the episode in which Elis goes out to fight a duel with a Saracen in order to protect Rósamunda from being married is cen-tral in the Old Norse version but is not at all important in the Old French version. The Old French version ends with a description of a grand battle, a newly Christianized land, and two marriages; none of these episodes oc-curs in the Old Norse version, even though the battle is suggested as a pos-sible continuation of the plot. Thus, despite the similarities, the narrative structures of the two versions, as defined by means of initials, are different. The use of initials structures the two versions in such a way that a different arrangement of episodes is highlighted as most significant. While the Old French text appears as a description of a series of adventures of the epony-mous protagonist, culminating in a grand battle between Christians and Saracens, the Old Norse version appears more as a story about a hero and his relations to his father, his mother, his friend, his lady, and his enemies.

20 Eriksen

SUMMARY AND CONCLUSIONS

The preceding is an analysis and comparison of the correspondence between one graphical aspect (the initials) and four textual aspects (rhyth-mical, rhetorical, lexical, and structural) in two versions of one nar-rative, the Old French chanson de geste Elye de Saint-Gille and the Old Norse Elis saga. The main result of this investigation is that both versions evince a certain correspondence of graphical and textual aspects, but the type and degree of correspondence differ. The rhythm in the Old French poem is defined by assonance whereas in the Old Norse saga this is achieved, but rarely, through alliteration and possibly a certain prose rhythm reminiscent of Latin cursus. In consequence, the prose version is considerably less rhythmical than the verse text. The narrator’s com-ments in the Old French text are more dramatic and emotional than in the Old Norse text. There are a greater number and variety of repeti-tions in the Old French poem than in the Old Norse saga, which is not surprising, since repetition is one of the major characteristics of chansons de geste. The Old Norse translation, however, contains a greater variety of temporal adverbs than the Old French version. The lexical evidence for the expected reception is more abundant in the Old French than in the Old Norse version. The narrative structure of the Old French text is more clearly and convincingly defined by the length of the laisses than the narrative structure, suggested by chapter division, of the Old Norse version. In addition, the thematic focus seems to be slightly different in the two versions. Some of these discrepancies may be explained by the different form and style of the two versions, compelling the two scribes to employ dif-ferent means. But there are several other possible implications for the intended mode of reception of the texts and their function, the manner of translation, the status of translation, and the relative status of the two languages and literary traditions. In respect to the intended mode of reception, the very presence of illuminations in the Old French manuscript may be an indication that the manuscript was meant to be looked at and enjoyed visually. Private reading, therefore, springs to mind. The consistent correspondence be-tween graphical and textual features that suggests intended listening also suggests intended vocalization. Thus, the poem may have been designed for public and private reading in various contexts. Public reading in small groups, with the manuscript visually accessible to all listeners, is one pos-sible setting. In addition, the manuscript may have been available for visual inspection at a court, and the listening audience may have been familiar with its images already.


Initials correspond to textual emphasis in the Old Norse saga as well, but the degree of correspondence is clearly lower. Even though the dif-ferences may possibly be explained by an existing writing norm, they may also suggest a greater tendency toward private reading than public reading. The tentative nature of this conclusion has to be emphasized, since both versions contain evidence pointing in both directions, but to a different extent. Thus it is necessary not only to consider various aspects of a text but also to indicate which manuscript version one studies when discussing mode of reception. When it comes to the possible function of the two versions, the disparity in their structure renders their core message rather different. The Old French poem is a story about the crusades, about the taking and Chris-tianization of new land, about strong Christian norms and a powerful king. The Old Norse saga, however, is a story of personal adventure, inner conflicts, friendship, and a love relationship. These themes also appear in the Old French poem, but they remain secondary, since the final scene in the chanson de geste is entirely missing in the Old Norse saga. The saga may in fact be based on an Old French version with a different ending, which would also explain the differences in narrative structure. Nonethe-less, the scribe who copied Elis saga in DG 4–7 must have found the text worthy of transmission, suitable for transcription, and possibly fulfilling an intended function. This hypothesis is tenable if one bears in mind the content of the two manuscripts and the political involvement in the crusades of the two respective cultures. The Old French manuscript BnF 25516 contains four texts, including two in addition to Elye de Saint-Gilles that deal with a Christian-Saracen conflict: Boeve de Hamptome and Aiol, the latter forming a cycle together with Elye de Saint-Gilles. In addition, there existed other chansons de geste concerned with the same topic, for example, Le Couronnement de Louis, La Prise d’Orange, La Charroi de Nîmes, and Aymeri de Narbonne, which circulated in the same period as Elye de Saint-Gilles and to which the poem alludes.45 This interest in the Christian-Saracen conflict may be explained by the fact that the House of Flanders, with which the manuscript is associated, had actively participated in the crusades; one of the counts, Baldwin of Flanders, had even been declared emperor of the Latin Empire of Constantinople in 1204.46

For its part, DG 4–7 contains fragments of three other texts besides Elis saga, namely Pamphilus, a Dialogue between Courage and Fear, and the

45. For a more detailed discussion of the subject, see Sandra C. O. Malicote, “Visual and Verbal Allusion. Disputatio and the Poetics of Elie de Saint Gille and Aiol,” Romania, 124 (2006), pp. 92–96. 46. Robert Fossier, ed., The Cambridge Illustrated History of the Middle Ages, vol. 2: 900–1250, trans. Stuart Airlie and Robyn Marsack (Cambridge: Cambridge Univ. Press, 1997), p. 275.

22 Eriksen

Strengleikar collection. All of these, to varying degrees and in varying ways, comment on the nature of love relationships and inner moral conflicts. Although Norway was not an active participant in the crusades, it is claimed that on one of his visits to Norway, Matthew Paris presented a letter from King Louis IX of France to King Hákon Hákonarson, asking him to lead the Seventh Crusade. King Hákon seems to have committed himself to take the cross three times (based on various documents), but he seems not to have kept any of these promises.47 Thus, it seems that the narrative structure of the two texts accords well with the respective overall theme of the manuscripts in which they are transmitted and also with the historical background of the milieus in which they were written down. The divergences in the two versions may also obtain from differences in the mode and status of copying/translating as a text-generating activ-ity and the competence of the scribe(s)/translator(s). Although Elye and Elis saga are versions of one and the same work, there is no direct link between the texts in the two manuscripts; the differences between them indicate that changes were made by scribes and translator(s) during the transmission process from the no longer extant original poem. Since we do not know who made the changes, the following may apply to both a scribe copying the Old French or the Old Norse text and a translator. The Old Norse text differs from the Old French and seems not to have been translated or revised word for word, but rather, changes were made as seemed appropriate. One obvious reason for this is the difference in form (verse vs. prose), the change from one to the other having neces-sitated some adjustments. This may, however, not be an explanation for omissions, additions, and differences in narrative structure. For example, since more types of temporal adverbs occur in the Old Norse text, it may be that the translator did not aim for standardization but rather for varia-tion. At other times he changed the nature of the narrator’s comments, rendering them less emotional, and thus testifying to the use of different rhetorical devices. This indicates that translating and copying (we do not know who made the changes, the translator or the copyist) were activities similar to composing, which required competence in available rhetori-cal tools and their appropriate use. The similarities in the processes of composing and translating in the Middle Ages have been argued for by such scholars as Rita Copeland.48

The lack of absolute equivalence in translation may also indicate inter-

47. Henry G. Leach, Angevin Britain and Scandinavia (Cambridge, MA: Harvard Univ. Press, 1921), p. 104; Peter Jackson, The Seventh Crusade 1244–1254: Sources and Documents (Aldershot: Ashgate, 2007), p. 23. 48. Rita Copeland, Rhetoric, Hermeneutics and Translation in the Middle Ages. Academic Tradi-tions and Vernacular Texts (Cambridge: Cambridge Univ. Press, 1991).


ference from the writing norm of the target culture; the translation was in a style that complied with local writing norms. The use of various tem-poral adverbs, and not only those corresponding to quant and or, occurs throughout Elis saga. The above discussed temporal adverbs and several others (nu með þvi, meðan, siðan, i þvi, jafnana, þegar, ok, en þo, helldr, etc.) occur regularly at the beginning of new sentences that are initiated by a capital. Thus they are very often visually emphasized. Many of these, with some variation, also appear often at the beginning of sentences found in other genres, such as Konungs skuggsjá, Barlaams saga, The Legendary Olavs saga, and Þiðreks saga, all of which are found in manuscripts more or less contemporary with DG 4–7.49 Prose rhythm and certain proverbial expres-sions occur also in other Old Norse texts written in the same period as Elis saga, in those mentioned above. The differences between the French and Norse versions may thus indicate that the style of narration in the Old Norse version complied with local standards for storytelling; the scribe chose not to be faithful to the style of the Old French original but rather to his own language and literary practices, and this may suggest that for the scribe the Old Norse language and literary tradition had greater status than that of the Old French source. Based on the differences in the dramaturgical structure of the two texts, it seems that the scribe was familiar with the entire text, which he rewrote, adjusting the narrative summits somewhat. Such an approach is characteristic of professional and competent translators, according to Gideon Toury’s Law of Interference.50 Postulating a professional or at least competent translator may be supported by the information given in the epilogue of Elis saga in DG 4–7, that it was translated by an Abbot Rob-ert, who may be the same as the Brother Robert who translated Tristrams saga some years earlier. The translation is also said to have been commis-sioned by King Hákon Hákonarson. From this one can presumably also infer that the scribe of DG 4–7 was active either in the royal Norwegian

49. Konungs skuggsjá is preserved in the Norwegian manuscript AM 243 bα fol., from ca. 1275. For a detailed description, see D. A. Seip and Ludvig Holm-Olsen, Konungs Skuggsiá. Speculum Regale. De Norske Håndskrifter i Faksimile. Festgave fra Universitetet i Oslo til H.M. Kong Haakon VII på hans 75-årsdag 3. august 1947 (Oslo: Oslo Univ. Press, 1947). Barlaams saga is preserved in the Norwegian manuscript Holm Perg 6 fol., from ca. 1250–1300. See Mag-nus Rindal, introduction to Barlaams ok Josaphats Saga, pp. vii–xx. The Legendary Olavs saga is preserved in the Norwegian manuscript DG8 II, from ca. 1250. See Anne Holtsmark, “Innledning,” in Olav Tryggvasons Saga etter AM 310 4to, med en Innledning av Anne Holtsmark, Corpus Codicum Norvegicorum Medii Aevi. Quarto Series, 5 (Oslo: Selskapet til utgivelse av gamle norske håndskrifter, 1974), pp. 9–20. Þiðreks saga is preserved in the Norwegian manuscript Holm Perg 4 fol., from ca.1275–1300. See Þiðriks saga af Bern, ed. Henrik Ber-telsen (Copenhagen: Samfund til Udgivelse af gammel nordisk Litteratur, 1905–11). 50. Gideon Toury, Descriptive Translation Studies and Beyond (Amsterdam: J. Benjamins, 1995), p. 279.

24 Eriksen

chancellery or in a monastic institution,51 and it is thus plausible that the scribe was a member of the top cultural class in Norway, possibly an erudite religious figure with substantial experience in writing. To summarize, based on the correspondence between the materiality of the two manuscripts BNF 25516 and DG 4–7 fol. and the textuality of Elye de Saint-Gille and Elis saga, one can conclude that more or less the same story was to be received in different ways and possibly fulfilled different functions. The Old French poem may have been meant to be received both visually and aurally and to familiarize the courtly audience of the House of Flanders with their own historical past. The manuscript itself testifies to a powerful and highly cultured provenance. The Old Norse saga, however, may have been intended primarily for private reading, even though a listen-ing audience cannot be excluded. The saga may have functioned as com-mentary on personal moral issues and appropriate solutions of personal conflicts of various types. Given the hermeneutical nature of translating and copying in the Middle Ages and the great competence of the transla-tor and scribe(s) of Elis saga, the tale changed in character from a political and religious narrative to a story of personal conflicts and ideals.

51. Mattias Tveitane, Elis saga, Strengleikar and Other Texts, p. 26.
