On Ogier le Danois’s legend and the document of Oulx 83 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 On Ogier le Danois’s legend and the document of Oulx 1 Giorgio Barachini (Merano) https: / / orcid.org/ 0009-0009-1928-3526 Riassunto: Questo articolo si occupa delle fonti della leggenda di Ogier le Danois, a partire dal documento di Oulx scoperto dal Prof. G. A. Beckmann nel 2004. Il primo scopo di questo saggio è quello di riesaminare il documento, nel quale tra i testimoni leggiamo il nome (o i nomi) di Otgerius spata. g. curta , e di fornirne un’analisi cronologica, storica e filologica al fine di valutarne l’affidabilità nella discussione sulle origini della leggenda di Ogier. Il secondo intento dell’articolo riguarda l’uso di tale documento (o più in generale, di documenti tardi) per dimostrare l’esistenza di una tradizione orale nella poesia epica romanza, in particolare nella leggenda di Ogier. In molti casi tali documenti sono dubbi intrinsecamente ed estrinsecamente. In questo modo, concentrandosi sul/ sui racconto/ i riguardanti Ogier, l’articolo riflette sulla tradizione epica romanza e sulle sue interpretazioni divergenti. Keywords: Ogier le Danois, Archival research, French epic poetry, Chanson de geste, Traditionalism, Individualism 1. Ogier’s legend between Traditionalism and Individualism This article deals with the sources of Ogier le Danois’ legend. Because of its supposed antiquity, this legend has been at the centre of the dispute between Traditionalists and Individualists 2 . My purpose is to face this matter from a specific, even narrow perspective: the scripta of a medieval archival document. From this starting point, inferences will be made exploring both tendencies, thereby aiming to be as neutral as possible in regard to them. Indeed, a close interconnection between orality and literacy was a peculiar feature of the Middle Ages. It is undeniable that the legend existed much before the preserved chanson de geste which concerns Ogier’s feats. Nevertheless, it is more difficult to say how old this legend actually is, to trace back how many sources were put together to obtain the extant hero and what they 1 I want to express my gratitude to Prof. Linda Paterson and Claudia Pellegrini for their courteous help, careful reading and useful suggestions. 2 Traditionalist ante litteram was Voretzsch (1891). The individualist perspective started of course with Bédier (1914-1929), followed by Becker (1942). The traditionalist standpoint was expressed by Lot (1940-1941) and Lejeune (1948). Recently it has been re-proposed by Beckmann (2004). On the topic, see also Le Gentil (1957), Togeby (1969), Aebischer (1975a), Aebischer (1975b). 84 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 were. It is equally difficult to ascertain which part of the epic tale is the oldest and how the elements were arranged together. And it is even hard to know if the shifts of perspective we can observe in the preserved tale were intended by a remanieur or are due to oral tradition 3 . We are only sure of this: Ogier is a complex character resulting from the combination of different, hitherto not fully disclosed elements. The most transparent one is the exploitation of the memory of the Frankish duke Autchar connected to the Lombard War, but this is not enough to explain other puzzling data such as the hero’s Danish provenance or his duplicity (loyal vassal vs. rebel). A great number of details in which Autchar was not historically involved go back to the age of Charlemagne, as well as to the first centuries of the Holy Roman Empire (Saxon Wars, Viking invasions, Arabic incursions). This is the reason we find it difficult to understand the role of Autchar’s legend as the founding core of the tale. In many cases, the sources that scholars have used to trace back all details to Autchar or, more generally, to history, are both extrinsically and intrinsically suspect. Extrinsically, because scholars make connections between the historical Autchar and the epic Ogier, basing them on unverified interpretations which in turn often rely on other unconfirmed suppositions; they risk tautologically claiming the truth of their own premises. Intrinsically, because errors could have been introduced into the sources themselves during their manuscript transmission, and these should be verified before they can be relied on as historically accurate. An example of extrinsically doubtful arguments is the name of Ogier’s sword, Courtain . In the Nota Emilianensis , this name was already interpreted as referring to the shortness of the hero’s weapon. It has been explained in two ways. On the one hand, it has been said that weapons in the 8 th -9 th centuries were shorter than those used in the 11 th ; however, to make such comparison, people in the 11 th century would have had to be conscious of the sword length three hundred years earlier, and this is not very likely. On the other hand, Beckmann (2004: 424-26) has recognised the weapon typology as the so-called sax , a dagger which could be as long as a short sword, up to 70cm. After the 8 th century, the use of such a weapon declined and disappeared in continental Europe. This feature of Ogier’s figure is considered a heritage of his origin from the Frankish Autchar: Ogier descends from Autchar, because the first uses a short sword in the epic poems and the latter lived in the 8 th century when short swords were in use. Nevertheless, this theory cannot explain why, among all epic figures based on people who lived in the 8 th century, Ogier is the only one who preserved his alleged original weapon. Moreover, we do not notice that Ogier fights differently from the other knights in the poems, so the name of his weapon would be a simple crystallization of an old feature that the singers of tales could not handle. So, something does not fit in this explanation. Hence, the sword 3 I am thinking of the motif of the war between Charlemagne and Ogier, which was originally historic - Ogier escaping to Italy with Carloman’s sons, as referred in l. 4402-11 of Eusebi’s edition (Eusebi 1963). This motif was romanticised later and referred to the killing of Ogier’s son, Baudouinet, by Karlot, Charlemagne’s son, during a chess game. On Ogier le Danois’s legend and the document of Oulx 85 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 name should be separated from its chronological origin. Beckmann makes an interesting observation, when he sets aside the Frankish Autchar and connects the sword name with the epithet Danois , Dane, which accompanies the hero from the first mentions of his name. In fact, among Norsemen, short swords ( sax ) were not only in use in early centuries, but also in the following ones up to the 12 th . We therefore should not take this element as a proof of the antiquity of the legend, but as a feature of the hero’s epic provenance. We have to assume it is a sign of an archaic, savage, barbaric way of fighting, differing from the advanced practices of Carolingian Frankish knights. A similar meaning must be seen for the epithet «Dane» as well (Lejeune 1948: 92, Le Gentil 1957: 206-07, 226-27, Togeby 1969: 18, 68) 4 . Instead of being evidence for the origin of Ogier’s legend, the sword name concerns the contemporary way in which Frankish (or French) rulers and intellectuals perceived the northern Germanic people (especially Norsemen and Angles) 5 . 2. The document of Oulx In 2004, Gustav Beckmann (2004: 421-24) found a document connected to Ogier’s legend in the cartulary of the former Monastery of St. Lawrence of Oulx (in the Alps in western Piedmont). The document, a donation made by count Guigues of Albon to the above-mentioned monastery, is connected to the chanson de geste and its antiquity, since the name of the last recorded witness is «Otgerius spata. g. curta». Beckmann interprets this presence as the definitive proof of an ancient oral tradition concerning Autchar. As this argument is intended to have a decisive impact on the 4 Beckmann (2004: 426 N47) rejects this interpretation, because only the Vikings, not the Danes, could be seen as barbarians in these ages, but Charlemagne had no relationship with the Vikings, i.e., in his interpretation, with Norwegian and Swedish regions. Yet, the Danes of Charlemagne’s age were not clearly distinguished from other Norsemen and were often equalized to the Vikings (cf. the toponym «Danelaw» in England, where Vikings settled in the 9 th century), and it is not unlikely that the French could use the ethnonym of the closer people (Dane) to name the other Norsemen. In a more recent contribution Prof. Beckmann (2017: 677-79) deems Dane to be an expression for «Ostler» (‘man from the East’), and, according to this, we find the Danish hero as the commander of Bavarians in the Chanson de Roland . 5 I cannot agree with Beckmann’s arguments when they become evolutionistic. For him, the idea of the short sword would have led to that of the Germanic people, then the idea of the Germanic people would have led to that of the Dane. I do not see how we can scientifically take the short sword as the primordial element and the hero’s provenance as the inferred one (it could be easily viceversa or both could have arisen together as the hero’s peculiar characterisation), without refreshing the theory that the short sword relates to Autchar, but we have set this option aside earlier. Yet, if we accept that Autchar is concerned, we go back to our first question: why do other epic heroes originating from the 8 th century not use any short weapon? 86 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 dating of the legend and has not been discussed by the research community so far, I decided to re-examine the document 6 . The original document of the donation is lost, and we have to manage with one medieval and three modern copies. I inspected all of them, thanks to the courteous help of the archival institutions. Three of the copies are preserved in the Archivio di Stato di Torino (A, B 1 , C 2 ), the fourth in the Bibliothèque municipale de Grenoble (C 1 ). In addition, I have also checked the modern printed editions of the text by Guichenon, Rivautella-Berta and Collino (Guichenon 1660: 197, Rivautella/ Berta 1753: 135, Collino 1908: 18). From the latter, which is the only available critical edition, I have taken the manuscript acronyms, with the exception of C 2 (see below). The witnesses have the following classification: A: Archivio di Stato di Torino. Sezione corte. Materie ecclesiastiche. Benefizi di qua dai monti. Prevostura di San Lorenzo di Oulx . Mazzo I di prima addizione. Fascicolo 4. The parchment codex has the title: «Cartario della Prevostura di Oulx». It was copied in the 13 th century in Oulx. The document resides at f. 139rv and is numbered 152 (rubric: do(n)acio d(omi)ni Guigonis comitis. ) 7 . B 1 : Archivio di Stato di Torino. Sezione corte. Materie ecclesiastiche. Benefizi di qua dai monti. Prevostura di San Lorenzo di Oulx . Mazzo I di prima addizione. Fascicolo 2. The paper codex has the title: «Estratto dei privilegi a favore della Prevostura di San Lorenzo di Oulx». The codex, which is not in fact a simple excerpt but a complete copy, is also called Codice Peralda , named after the general vicar of the prevostship, Ugo Peralda, who copied the manuscript at the end of the 16 th century in Oulx (between 1582 and 1599). The document resides at f. 72r and is numbered 152 (title: Dona(ti)o d(omi)nj guigonis comitis./ ). C 1 : Bibliothèque municipale de Grenoble, ms. U.5221. The paper manuscript has the title: «Cartulaire du monastère d’Oulx en Briançonnais». The manuscript was copied in the 18 th century in Grenoble. The document resides at pp. 166-167 and gives the correct reference to f. 139 of ms. A in the margin (title: Donatio d(omi)ni Guigonis Comitis ). 6 I must stress once more: no doubt an oral tradition produced the extant epic French texts, but it can be misleading to look over-zealously for evidence of this by trying to fill the gap between deeds and texts with many suppositions and venturing to use later documents to prove the antiquity of the epic legends. All archival documents connected with the chansons de geste date from the mid-11 th century, and this recent dating cannot prove the existence of oral songs and their connection to the more ancient deeds which inspired them. It proves only that at that age the epic legends matured and became suitable for French society. For the centuries of état latent we should rely on other arguments. 7 The folio numbering dates back to the 15 th and 16 th centuries, that of the documents to the 19 th century (see Collino 1904: 172). On Ogier le Danois’s legend and the document of Oulx 87 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 C 2 : Archivio di Stato di Torino. Paesi. Saluzzo. Marchesato di Saluzzo. Documenti ceduti dalla Francia nel 1949 . Mazzo 1. The paper quire has the title: «Recueil de copies d’actes destinés à justifier que le marquisat de Saluces devait dépendre du Dauphiné». This copy, to which I ascribe a new abbreviation (C 2 ), was not used in Collino’s edition (1908: 18): «Copia, di cui non si tien conto, in Archives Departementales [sic] de l’Isère ». In the latter archive, it originally had the shelfmark B 3853, but it was moved to Turin in 1951 due to an international agreement of 1949. It is an uncomplete copy made in the 18 th century and has most likely exploited the same exemplar as C 1 (both have the same text, except for a clause lacking in C 2 ). The document resides at f. 3 (title: Donatio Guigonis Comitis Ecclesiae S(anc)ti Petri de Ultio de quibusdam mansis, decimis et offerendis etc. ). In the margin, we read the reference: «Cartul. Vltiense. f. 139.», i.e. ms. A. As already stated, the document is a donation, in which «Guigo il vecchio d’Albon, e suo figlio Guigo il grasso danno alla chiesa dei santi Lorenzo e Pietro d’Oulx un manso in Cesana» (Collino 1908: 18). Collino’s transcription, quoted by Beckmann, is as follows (the numbers referring to the apparatus are omitted). Jn nomine sancte et indiuidue trinitatis. Anno ab incarnacione domini. m . l [ x ]. iii . Jndiccione prima. Ego Guigo qui nomine uocor senex. atque filius meus Guigo pinguis dono et confirmo pro anime mee mercede et pro anima patris mei et matris mee et parentum meorum ecclesie beati petri cum ceteris apostolis et ecclesie sancti laurencii martiris in loco qui dicitur ple[ b ]s martirum mansum unum cum omnibus rebus ad se pertinentibus iacet in loco qui dicitur sesana et canonicis in supranominatis ecclesiis regulariter uiuentibus et omnibus eorum successoribus. ut pro animabus nostris ipsi apud omnipotentem intercessores existant. quod totum factum est consilio domini ade castellani briençonis existentis. Jnterfuerunt testes quidam canonici. Girardus ualençole. Vldricus prepositus et uuarnerius. et martinus. Galterius: et quidam laici. Aurucius. Armannus presbiteralis. letardus crossus. Girardus garembo. Otgerius spata. (g.) curta. Cat(b)aldus diaconus. 3. The date The date of the document is 1053 but the editor’s conjectural correction shifts it to 1063. The problem is that 1053 does not fit with the next indication of the first indiction year, because 1053 was a sixth indiction year. Beckmann (2004: 423) reports two different proposals for the correction. The first was discussed by Terrebasse (1875: 55-59, part. 57-58). In his research, which goes back to the 1860s, he adhered to the abrupt correction of Gaspard Moyse de Fontanieu (in his handwritten, so far unpublished notes) and L’art de verifier les dates (Clément 1784: 454). Terrebasse considered that the year 1053 was incorrect 88 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 and the first indiction was correct. He argued that it was more likely that the error had involved the complex Roman numeral expressed in thousands, rather than the simple numbering of the indiction. To such an argument we can add that the indiction is generally considered reliable as an alternative year counting system because of its reduced cyclical period. Hence, on such bases, the document date has been shifted to 1063, which is in fact a first indiction year (after March 25 th , as the Incarnation style is used in the document). Collino also accepted this correction and emended « m . l [ x ]. iii .». In contrast, Georges de Manteyer (1908: 183-84; 1925: 36, 81 N29) has assumed the year 1053 to be the correct date, implying an error in the indiction year 8 . Beckmann has used Manteyer as his main historical source, being biassed towards an earlier date of the document, and has explained his choice as follows: «[Manteyer] vertritt also offenbar die Meinung, daß im Original die Indiktion VI stand, schon bei der Niederschrift des Archetyps aber das V übersehen wurde» (2004: 423). Indeed, 1053 was a sixth indiction year. Moreover, Beckmann has expressed the opinion that such errors with Roman numbering were usual, even daily («alltäglich») in medieval documents. This naive argument is unacceptable, because dating errors are nevertheless fewer than dates correctly written, and they are definitely not qualifiable as daily. If we now compare the actual texts of the witnesses, the manner in which Beckmann has attempted to prove an earlier date for the document (that is, dating it back in 1053 and emending the indiction) raises serious doubts about this choice of date. In the manuscripts we find the following readings for the year: A: m .l.iii B 1 : m o .liii o . C 1 : m. lxiii. C 2 : mlxiii For the indiction we read: A: p i ma B 1 : j a . C 1 , C 2 : i . Regarding the indiction, it is patently wrong to argue that copyists have made an error while transcribing the ordinal numeral on the spurious grounds that we should presume that in the original document the ordinal numeral was written with Roman 8 Manteyer did not discuss his choice and did not provide any explanation for it. He simply deems the year as correct, without mentioning the indiction. On Ogier le Danois’s legend and the document of Oulx 89 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 numerals, not with full letters. The manuscripts in fact attest both scriptiones (the most ancient ms. A reading «p i ma»). The date is most clearly indicated by the readings of C 1 and C 2 , whose texts report the year 1063. This variant is not registered in Collino’s apparatus, hence it is unknown to Beckmann. With this testimony, we are led to take 1063 as the most likely date for the donation, though all doubts about a potential manipulation cannot be dispelled. In fact, as we have already seen, C 1 and C 2 are copies made in the 18 th century in Grenoble, using a codex parallel to B 1 . Their copyists (or, more likely, their exemplar’s) collated their text with ms. A, the oldest copy of the document: they have precisely annotated the folio numeration of A in the margins of C 1 and C 2 . Now, the first person who noted the inconsistency between date and indiction was the above-mentioned Gaspard Moyse de Fontanieu, if we trust Terrebasse’s testimony. As Fontanieu was the Intendant of the Dauphiné between 1724 and 1740, it is possible that both C 1 and C 2 are copies transcribed on behalf of him or on the impulse of his entourage. So we cannot exclude the possibility that the texts of C 1 and C 2 have been emended on the base of the Intendant’s conjecture. Nevertheless, to attenuate this assertion, it is more plausible that the conjecture and the emendation were already in the lost exemplar which both manuscripts exploited. This exemplar was a manuscript written between the end of the 16 th and the beginning of the 18 th century; most probably close to the first date, as at the end of the 16 th century a manuscript loaned by the monastery of Oulx was already in Grenoble and according to Collino (1904: 180-81) it is likely that it was the exemplar of C 1 and C 2 . Moreover, in the reconstruction of the textual tradition made by the editor, the witness C 1 (and of course C 2 , which Collino did not take into account) descends via a lost intermediate from the same hyparchetype (called B) from which the manuscript B 1 also descends, but such a tradition makes us uncertain about the reading that was attested in this branch of the stemma (because of the conflict between B 1 and C 1 / C 2 ) and about the time when the date would have been emended (Collino 1904: 181-90) 9 . To verify if 9 Collino’s reconstruction is the following: This stemma is somewhat tendentious, because it is mainly based on external elements, such as the introductory or working notes the copyists have left. C is considered a parallel copy of B 1 by Collino (1904: 188-89), because in C 1 we find «una nota evidentemente trascritta in C 1 da C a proposito della collocazione del doc. LXIV che è duplicato a fol. 169 del cartolario archetipo. La nota 90 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 Collino’s reconstruction is correct, we should re-examine the whole codex C 1 and Fontanieu’s handwritten notes. I can only underscore that the unscrupulousness of such emendation directly within the copy would be as surprising as the idea that a copyist, while transcribing, would have remembered Fontanieu’s conjecture and would have inserted it into the text. Both events seem quite unlikely to me. Furthermore, Terrebasse (1875: 58) strengthened his correction to the manuscripts A and B 1 (C 1 and C 2 being unknown at that time) with a historical argument. In the cartulary of Oulx, the congregation of St. Lawrence - and hence its existence - is not confirmed before 1057, when Otto of Savoy and his wife Adelaide of Turin granted all their rights in Oulx territory to it (see Collino 1908: 7-10, doc. VII, May 1057). Thus, Terrebasse held it to be impossible that Guigues’ donation, in which we already find a «prepositus», could date back to 1053. By contrast, Manteyer’s and Beckmann’s conjecture has not such historical support. Moreover, we must note that, between the document of 1057 and the one we are dealing with, there is a great similarity when we consider the names of the clerics, but the affinities disappear when we consider the religious titles granted to them. Such titles are to be found only in Guigues’ donation. This points to a more recent date for the document, in any case subsequent to Otto’s and Adelaide’s grant 10 . afferma che la carta stessa era collocata in prima redazione a ‹fol. 65 du cartulaire original et a folio 37 v de la copie›». Collino interprets that the «cartulaire original» is ms. A and the «copie» is the alleged reconstructed ms. C. This is surely wrong, since the «folio 37 v de la copie» does not refer to the folio numbering of C, but to that of B 1 , where the relevant document is transcribed at f. 37v. Philologically, we cannot be sure that both B 1 and the exemplar of C 1 were descendants of A in some way. Indeed, the exemplar of C 1 , whose copyist could compare his own exemplar with both A and B 1 (of course, after Peralda’s copy and after ms. A had moved from Oulx to Susa), particularly raises doubts as to Collino’s alleged position in the stemma and his historical reconstruction. 10 In Guigues of Albon’s donation we find five clerics as «testes» («Girardus ualençole. Vldricus prepositus et uuarnerius. et martinus. Galterius»). They are more or less the same «clericos qui in eodem loco regulariter uiuunt» in Otto’s and Adelaide’s document of 1057 («Giraldus et uldricus. et aicardus. et martinus. et lantelmus. et Vualterius et armanus»; even the relative order is the same). Giraldus, Uldricus, Martinus, Vualterius are common to both. The last witness of 1057, «armanus», should be the same person mentioned as «Armannus presbiteralis» in the donation of Guigues, where however Armannus appears among the laical witnesses. The final position of this witness in 1057 is certainly hierarchical, according to the prevailing use in medieval documents, and the status of «presbiteralis» in Guigues’ donation explains the reason of such position. A presbyteralis was a man who, halfway between the consecrated and the laical status, lived and worked within the religious institution, studying and waiting to be formally admitted (hence he lived there «regulariter», according to the regula ; a presbyteralis is in fact the medieval evolution of the deacon). Depending on cases and on the inclination of the writer, he could be registered among the clerics, strictly in the final position, or among still-laical people. On Ogier le Danois’s legend and the document of Oulx 91 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 4. The witness The crucial point of the document is the mention of Otgerius spata. g. curta . In relation to this, Collino makes a relevant choice in his edited text, which oriented the scholarly perspective thereafter, but it seems to have been somewhat cursory according to palaeographical and philological data 11 . In fact, both manuscripts and previous editors have interpreted the relevant textual string differently from Collino. Hereafter the readings 12 : A: Otgerius spata. g. curta. B 1 : otgeri us spata/ g. curta/ C 1 : Otgerius Spata. G. Curta. C 2 : Otto Spata, G. Curta, Guichenon: Otgerius Spata, G. Carta, Rivautella/ Berta: Otgerius Spata, G· Curta, This gives rise to further doubts. Until 1908, no scribe or editor has thought about the presence of a single witness named Otgerius spata curta here. Manuscripts B 1 , C 1 , C 2 and editors Guichenon and Rivautella-Berta have clearly understood that two different witnesses were named, one called Otgerius Spata , the other G. Curta . In ms. B 1 this issue is certain, since the copyist Ugo Peralda (see above) always distinguished the name of a witness from the following one with a forward slash and marked every abbreviation with a dot after the letter. Ms. C 2 used a similar system with a comma in the place of the forward slash and capital letters for name initials (even abbreviated names). Ms. C 1 is less consistent (it used dots both for separating names and for abbreviations), but its copyist undoubtedly understood the names in the same way as his colleagues. The most ancient manuscript A is unclear. By copying onomastic references to witnesses, donors and procurators throughout the codex, the scribe always opts for using initials in capital letters for the first name, when this is preceded by a separation dot, as in the case of our document. Even when the name is abbreviated with a single letter, this is written with a capital. In many plac- 11 By way of example, the integration of a letter in the word «ple[b]s» is philologically unnecessary, since the reading ples is an oversight of the only scribe of ms. A, whereas the texts of B 1 and C 1 have plebs , which is rightly the acceptable reading without square brackets (it is also to be noted that ms. C 1 indisputably reads plebs , not «plebe», as we find in the scant apparatus). Further, the intervention in the name «Cat(b)aldus» is similarly inappropriate, because either the letter b should be emended to h (a strong, not recommended intervention) or it is in fact suitable and correct, since the name Catbaldus has given rise to the modern French surnames Chabaut, Chabaud. 12 For modern editors cf. Guichenon (1660: 197), Rivautella/ Berta (1753: 135). Guichenon’s edition is clearly inaccurate in many places, among them our reading, which contains an evident oversight concerning the second name. 92 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 es of the manuscript, the single capital letter «G.» stands for many name options such as Guigo, Guillelmus, Gandulphus, Girardus, Gonterius, Gaufredus. If the scribe of A had meant to point unequivocally to a second witness in the concerned textual string, he would have opted for capital G. Yet, we cannot be sure of what he meant: a lowercase in the place of an uppercase is indeed a poor clue. If we admit that the lowercase was not the abbreviation of a name and was not in the original text, we still have to answer lots of questions. What does it represent? How did it originate? Why did the scribe put it into the text during the copying? And what does it actually mean? In the face of difficulties, Collino gave no justification for his choice of putting the letter in brackets. Beckmann (2004: 423) however provided the following explanation: Der belanglose Einzelbuchstabe g zwischen spata und curta schließlich, den Collino zu Recht getilgt hat, ist wohl darauf zurückzuführen, daß der Schreiber zunächst ein falsches Adjektiv (oder eine falsche Graphie von curta ) niederschreiben wollte, dann zwar seinen Fehler sofort erkannte, die Verbesserung jedoch bis zum Trocknen der Tinte verschob und anschließend vergaß. Such an explanation, which from the outset brands the isolated letter g as «belanglose» (‘irrelevant’), does not seem thoughtful to me. Firstly, «g.» is a handwritten primary matter of fact and cannot be deemed irrelevant, particularly because ms. A was copied in the 13 th century, that is at an age in which the epic poem Chevalerie Ogier was well enough known to stimulate textual interpolations elsewhere (even in late witnesses of Carolingian annals, as I will discuss in a forthcoming article). It is difficult to say whether the letter g placed here, between the two words which form the alleged witness’s surname, is a trace of an interpolation (I will go back to this point later), but deeming it irrelevant is a convenient solution for a traditionalistic standpoint such as that found in Beckmann’s essay. Secondly, if the copyist had begun to write «ein falsches Adjektiv», as Beckmann argues, then we are led to make two deductions, assuming that at that time there was an epic hero called Ogier, armed with a short sword or dagger: a) if the adjective beginning with g was the correct one, the other transcribed later, i.e. «curta», has been interpolated by the copyist who was influenced by the epic hero (such an interpolation probably dates back to the 13 th century); b) if «curta» was already in the exemplar of A, its copyist (and not that of A) was so careless that he could not recognize the name of a celebrated epic hero, thereby starting to transcribe another adjective. But this would undermine the postulate itself, namely the alleged fame of the epic figure. So the theory of the «falsches Adjektiv» would result in contradictions about Ogier’s notoriety, and the letter g is shown to be relevant. Thirdly, the possibility of an exchange between minuscule g and c in the gothic script of A or of its exemplar is very unlikely, and this is even less possible at the time of B 1 and C 1 . For the rest, all manuscripts fully agree on transmitting the letter g . On Ogier le Danois’s legend and the document of Oulx 93 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 Finally, Beckmann’s explanation that the copyist would have noted the error immediately but would have waited for ink to dry shows two weak elements. Firstly, Beckmann does not make clear what intervention of the copyist he is thinking about, when he talks about a «Verbesserung» (‘correction’). In these circumstances, the copyist would have implemented (in order of probability) 1) an expunction, without waiting for the ink to dry, and we would have found a dot under, and not after the letter (of course, if this intervention had been in the exemplar, we would not have found anything in the copy); 2) a deletion, but in this case the copyist would not have needed to wait for the ink to dry either; 3) a knife erasure, but we would have found extra spacing between words, what was normally avoided. Moreover, a scraping erasure conflicts with the second element preserved in all manuscripts, even in the oldest A, with a clearness which cannot be easily put aside: the letter g is not isolated at all, but it is clearly followed by a dot. This dot shows that the letter was not written by the copyist randomly or mistakenly. A letter followed by a dot expresses an abbreviation, so with this the scribe had something in mind or was reading something in his exemplar. Such an abbreviation has several expansions. These all argue against the existence of an Otgerius spata curta . For similar abbreviations Cappelli’s dictionary (1929: 147-48) registers the expansions «igitur» and «ergo» (we should assume a small retouching during transcription). For a simple g followed by a dot, it even registers the expansion «gesta» in the 14 th century (Cappelli 1929: 147). Therefore, if we could make this expansion rather more ancient, we would be justified in thinking that the textual string «g. curta» was a marginal note which the copyist of A had conscientiously transferred to the main text. Maybe gesta , or even better gestus, indicated an archival section of the prevostship (cf. Du Cange 1885: 63c), where the full name of the witness, who was simply mentioned as Otgerius spata here, could be read in extenso . Maybe gesta was to be interpreted as the past participle of gero , with the meaning of ‘completed, to be completed’, that is, the name Otgerius spata had to be completed with curta . Maybe gesta meant ‘feats, great deeds’ and directly referred to the epic poem. By finding an anonymous Otgerius spata in the document he was copying, a curious cleric would have left the note «g. curta», as to say: «This Otgerius spata is the one who carries a short ( curta ) sword in the gesta , the epic poems» ( spata curta as instrumental ablative is the likeliest interpretation in any case, with the meaning ‘the one with a short sword’). In all the above-mentioned possibilities the abbreviation leaves us without any certainty that the name Otgerius spata curta is genuine. The name of the Oulx document could be interpolated. The same conclusions would be drawn if we assumed the expansions «igitur» or «ergo» for «g.». Methodologically, both for the dating and for the interposed letter g , it must be stressed that it is improper to use archival information to argue a point without discussing its involved set of problems in depth, for conclusions would be controversial and, in the worst case, wrong, if they had been drawn with the aim of demonstrating a preconceived theory without first verifying and counter-verifying the 94 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 primary data. Nowadays, it is essential to go back directly to primary sources without trusting old editions. We do not yet exactly know what the letter g means, but by inspecting all witnesses we are now sure we cannot easily dismiss its importance. 5. Consequences for Ogier’s legend Since the original document is not preserved, I am not able to propose an uncontroversial explanation for the manuscript issues. The copying procedure, which has produced the extant copies, could be interpreted in various ways, therefore we could really be dealing with a person whose name was openly referring to the epic hero and his weapon, as Beckmann argues. So, it is worth taking this possibility into consideration. In the next paragraphs, I will hypothetically reflect on this, as if the onomastic data had been confirmed. A preliminary set of questions involves the issue that the onomastic feature does not only concern the Christian name, but also the surname. This aspect requires discussion. When we deal with onomastic references to epic poems, e.g., a pair of brothers and cousins named Rotlandus and Olivarius , the two important elements are the first names and their combined occurrence, which certifies an intentional choice 13 . In this perspective the surname, if any, is not important, as it could not be implicated in the naming choice. If we admit that the witness of Oulx was called Otgerius spata curta , we must also ask how he has obtained his full onomastic features. Was the surname spata curta already extant and did parents and godfather choose the first name because of the epic figure? In this case it is still possible that the combination of elements happened independently from epic poems. Or was spata curta an individual sobriquet which would become a surname? In this case we should ascertain if the sobriquet originated from the epic poems and not from any other unrelated reason. In any case, while the surname Spadacorta and its possible variants are not attested in the Italian onomastics, Courtépée is to be found in the French, and this counts as an indirect confirmation of such a surname in the Middle Ages. What could this Otgerius spata curta teach us about Ogier’s legend, if the name were confirmed? Firstly, we would be certain that a hero with such a name and such a renowned weapon existed some time before 1050. If the donation dates back to 1063, the birth year of the witness must have fallen at least in the Forties of the 11 th century. It is useless to point out that Ogier’s weapon is not named in Turoldus’ Chanson de Roland . Omissions like this could be done on purpose or, more likely, happened by chance, and they are unusable for critical aims, since a positive argument cannot 13 On epic onomastics we have now the worthy, complete work by Beckmann (2017), in which the couples of brothers/ cousins are discussed at p. 820-44. On Ogier le Danois’s legend and the document of Oulx 95 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 arise from a lack of information 14 . Since the epithet spata curta is in the Nota Emilianensis , but Courtain lacks in Turoldus, reappearing later in the Chevalerie Ogier and elsewhere, it is legitimate to conclude that this heroic feature already existed around 1050 (this means that Turoldus forgot it), but it is not legitimate to deduce on this basis that Ogier’s legend is as ancient as the hero’s sword and dates back to the 8 th century. We cannot affirm that Ogier was accompanied by the epithet Courte-Espee in the epic poems, because there is no trace of such an epithet in vernacular texts and we are compelled to suspect that spata curta was originally an instrumental ablative with an explicative value - as already pointed out above - in Latin, not in vernacular texts (Ogier, i.e. that figure whose main feature is to combat ‘with a short sword’). However, we can easily admit that the detail of the short sword was a specific feature of this character, which for unknown reasons is not preserved in the Roland , where Ogier plays a subordinate role. Secondly, around 1050, the hero and his legend had some reputation. Otherwise, we would not find a person with the hero’s name in Oulx. If we compare Ogier’s sporadic case with that of several brothers and cousins named Rolandus and Olivarius , we can outline a general scenario, in which Ogier was a known figure, but his fame could not compete with that of more illustrious heroes. Moreover, the image of the epic hero had to be perceived with a positive connotation to serve for onomastic reasons. We have already observed that Ogier is openly faithful and loyal to the emperor in the Chanson de Roland , whereas Ogier as the protagonist of the eponymous poem is a double-edged figure, with a positive and a negative side 15 . The decision to name a real person after the hero shows either that 14 On this element see Beckmann (2004: 421-22, 2017: 678). 15 Beckmann (2004: 443-44) tries to obliterate the dualism of Ogier’s figure. He asserts that there has never been a purely loyal vassal, inasmuch even in the Roland Ogier pronounces outrageous words against his king. I dissent from this interpretation. Ogier’s words quoted by Beckmann are not so outrageous to justify the elimination of Ogier’s duplicity admitted by all other scholars. If at all, they show an impetuous, cocky behaviour on Ogier’s part. Moreover, I disagree that in the Conversio Othgerii militis , written in the abbey of Saint Faron of Meaux, where we find a representation of Ogier’s Moniage and the notice of an abbatial donation near Vercelli in Italy, «deutet eben das Vercelli-Motiv auf die zeitweilige Rebellenexistenz». Vercelli’s quotation would refer to the epic battle narrated in the poems, where the rebellious Ogier fights with the Lombards against the Franks. But the existence to such an abbey near Vercelli («in suburbio Vercellinsi», MGH SS Rer. Merov. 5: 206) is highly disputed by historians and Othgerius is not represented at all as a rebel in the Conversio . Beckmann’s argumentation in these pages, as it was in Lot’s and Lejeune’s, is built on a number of hypothetical elements, in which the previous one should corroborate the next one in a chain of scientific uncertainty (as Aebischer, quoted by Beckmann, already pointed out). Let us take, for example, the above-mentioned abbey near Vercelli. To prove the existence of an ancient Ogier’s legend related to Autchar since the 8 th century, scholars have to presume that such an abbey was connected to Autchar, perhaps granted to him by Desiderius and then ceded by Autchar/ Othgerius to St. Faron with Charlemagne’s permission. This is a sequence of supposed actions that is quite unlikely, especially because Autchar was a prisoner. In any case the whole process is not demonstrable (neither which abbey we are dealing with, nor that it was a possession of St. Faron in whatever period). Our best option is to suppose that «suburbium» indicates a 96 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 there was no intention to refer to his negative side (hereby including the possibility that there was no negative side yet), or that the negative side was neutralised by the positive one, or yet that even in the Middle Ages there were cases of the «attraction of evil», i.e. a negatively depicted character attained such a fame in his excess to be nevertheless perceived as heroically impressive and suitable for conventional onomastic reasons. So was Ogier’s version, to which Otgerius spata curta of Oulx is referring, the same as the one appearing in the Chanson de Roland ? No one doubts anymore that one or more versions of this poem circulated before Turoldus’ composition. I see no effective argument against the idea that such lost versions could date back at least to the fourth decade of the 11 th century. I do not see any difficulties either in supposing that in those lost versions, Ogier had a role of somewhat moderate importance, as he has in Turoldus’ poem. The person in Oulx would confirm this perspective. Such a hero armed with a shorter sword than the other knights for some reason disappeared in Turoldus, and this would explain the cognominatio of Oulx. He could be the leader of the vanguard or a brave knight without specific assignment, and this would explain his not-overwhelming fame. He was a loyal baron of Charlemagne, and this explains the decision to re-use his name, if we exclude the improbable idea of the attraction of evil. On this basis, we could guess at the existence of an eponymous poem, but the witness in Oulx could not be evidence of it. If the attestation of Oulx could be confirmed, it would enhance the dossier on Roland ’s prehistory, maybe more than Ogier’s. On the questions we see as the most important and essential concerning the figure of the Danish hero, namely his double-sided characterisation and the genetic priority of the rebel or loyal vassal, the find of Oulx leaves us unsatisfied. On this subject, the only important element remains the document of circa 1071 found by Rita Lejeune (unexploited by Beckmann), when in the diocese of Valence we find a couple of brothers named Otgerius and Desiderius , who could show the wish to celebrate the rebellious, bloodthirsty (and at the end defeated) vassal and the last Lombard king. Both brothers also reappear in a donation of circa 1074 (Lejeune 1948: 155, huge territory around Vercelli. So the assertion of the Conversio could refer to St. Albin of Mortara (30km to the east on the border of the diocese), where epic legends placed the graves of Ami and Amile, died during the aforementioned fictive battle. In the Chevalerie d’Ogier they are killed by Ogier (l. 5856-918). But this element being fictive, it implies a circulation of epic legends about the Lombard War, that would be significant to Ogier’s prehistory only after the determination of the composition age of the Conversio , which is also disputed. It falls around 1070-1080 according to Lejeune herself, making it unusable to prove the reliability of the indication in the Conversio and at the same time the antiquity of Ogier’s legend. Such links in the chain are very weak and unreliable. If one of them were proved wrong or manipulated, as I intend to show in a forthcoming article for the quotations of the alleged Chronique de Moissac (i.e. the first link of the chain), the whole building would appear very frail. On Ogier le Danois’s legend and the document of Oulx 97 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 referring to Chevalier 1913: 357, doc. 2077) 16 . Nonetheless, a certain prudence is needed here, because the name of the second brother is widely used in the cartulary, Saint Didier being broadly venerated in South-eastern France (Rhone region) with many church dedications, and even Otgerius is frequent in the same region and in the same documents. Therefore, their co-occurrence could be fortuitous. In any case, these documents could indicate that at that time a poem was already dedicated to the rebellious vassal. The document of Oulx, like the Chevalerie d’Ogier de Danemarche , could add that the rebel was already connected to the positive figure we find in the Roland (and in the Enfances and in the epilogue of the Chevalerie and in all other epic poems). Instead, none of these documents helps us by researching the obscure origin of the epithet of the Romance tradition, le Danois/ de Danemarche . So, Otgerius spata curta of Oulx tells us that a character existed, but it does not complete the framework 17 . Ogier still remains a composite figure, unclear in its origins. At least two traditions have produced it, the one connected to the Frank Autchar representing him as a rebel, the other, most likely of northern origin, that connects him with the Saxon Wars («Dane» in this sense), where he perhaps played the role of an ally of Charlemagne 18 . As I said at the beginning, what we still have to investigate, by looking for new documents, is why these traditions, which do not seem to have originally had 16 In the donation of ca. 1071: «Signum Otgerii donatoris. Signum Desiderii fratris ejus» (Chevalier 1898: 181, doc. 158). In the document of ca. 1074 (Chevalier 1898: 187-88, doc. 166): «Signum Otgerii presbiteri. Signum Desiderii fratris sui». See also Aebischer (1954: 1223-38). 17 We find the same limitation in dealing with the false document of Saint-Yrieix-de-la-Perche. As well known, here Charlemagne would have granted the confirmation of all privileges held by the abbey, the document mentioning as witnesses «Turpione, Otgerio palatino ac Guillelmo Curbinaso, Bertranno validissimo, Rotgerio Cornualto» ( MGH DD Karol. I : 357, doc. 251). The forgery was written around 1090 and it does not prove, as Lejeune (1948: 61) argued, the existence of a legend starting with Autchar and pointing to Ogier, but rather the increasing fame of the Chanson de Roland and other written poems in those decades. In the Roncevaux poem Ogier the Dane had a subordinate, but not negligible role. Yet, Autchar’s story found no space in it and could not be compatible with the highly positive characterisation of the Dane. 18 Supporting his idea that Ogier originally had to be only a rebel, Beckmann (2004: 435-38) attempts to trace back the «northern» element to the Viking Ansger, who ravaged northern France around 841. Remaining closer to the epic narratives, the historical figure, if any, was perhaps to be found among Charlemagne’s allies against the Saxons. Other traditions blended into the extant character, like the Saracenic incursions in central Italy, specifically against Rome, between 846 and 849. These events are reflected in the Enfances , when amiral Corsuble lays siege to Rome and is defeated by Ogier and the Franks (this epic episode is known by Metellus of Tegernsee, who transfers it to his monastery’s founder). Even half-erudite traditions could have blended in the legend, like the one concerning Ogier’s father, Gaufrois. He shares with the historical Danish leader Gudhfridh (attested since 804, dead in 810) nothing but the name. Ancestry and deeds are completely invented and Gaufrois appears only in late chansons which pretend to be the «prequel» of Ogier’s tales, in a common cyclisation process. It is difficult to see here a very marginal oral tradition that has come to light very late by chance, rather than the exploitation of a name found in Carolingian annals, to whom the writer’s imagination supplies every piece of information. DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 98 Giorgio Barachini any contact, have blended into each other, how they mixed and which one of them prevailed to forge the main features of the character. Bibliography a eBiScher , P. 1954: «Anthroponymie et toponymie dans la ‹Chevalerie Ogier›. À propos d’un ouvrage récent», RBPH 32/ 4: 1223-38. a eBiScher , P. 1975a: «Le concept d’‹état latent› dans la préhistoire des chansons de geste», in: i D ., Des annales carolingiennes à Doon de Mayence , Genève, Droz: 159-97 (firstly appeared RBPH 47, 1969: 789-839). a eBiScher , P. 1975b: «La mesnie de Doon de Mayence et son plus illustre représentant Ogier le Danois», in: i D ., Des annales carolingiennes à Doon de Mayence , Genève, Droz: 201-22 (firstly appeared in: Mélanges de langue et de littérature médiévales offerts à Pierre le Gentil par ses collègues, ses élèves et ses amis , Paris, SEDES, 1973: 13-32). B ecKer , P. A. 1942: «Ogier von Dänemark», ZFSL 64: 67-88. B ecKmann , G. A. 2004: «Oggero Spatacurta und Ogier le Danois. Zur Komplexität einer epischen Tradition», ZrP 120: 421-56. B ecKmann , G. A. 2017: Onomastik des Rolandsliedes: Namen als Schlüssel zu Strukturen, Welthaltigkeit und Vorgeschichte des Liedes , Boston/ Berlin, De Gruyter. B éDier , J. 1914-1929 [ 1 1908-1913]: Les légendes épiques. Recherches sur la formation des chansons de geste , Paris, Champion. c aPPelli , A. 1929 [ 1 1899]: Lexikon abbreviaturarum. Dizionario di abbreviature latine ed italiane usate nelle carte e codici specialmente del medio evo , Milano, Hoepli. c heValier , U. 1898: Cartulaire de l’Abbaye de Saint-Barnard de Romans. Première partie (817-1093) , Romans, [s.ed]. c heValier , U. 1913: Regeste dauphinois ou répertoire chronologique et analytique des documents imprimés et manuscrits , vol. 1, Valence, Imprimerie Valentinoise. c lément , F. (ed.) 1784: L’art de verifier les dates , vol. 2, Paris, Jombert. c ollino , G. 1908: Le carte della Prevostura d’Oulx raccolte e riordinate cronologicamente fino al 1300 , Pinerolo, Chiantore/ Mascarelli. c ollino , G. 1904: «Sui cartolari della prevostura d’Oulx», Miscellanea di storia italiana 9/ 3: 169- 204. Du c ange , c. Du f reSne 1885: Glossarium mediae et infimae latinitatis , cum supplementis integris D. P. Carpenterii. Editio nova aucta […] a Léopold Favre, vol. 4, Niort, Typographie de L. Favre. e uSeBi , M. (ed.) 1963: La Chevalerie d’Ogier de Danemarche , Milano/ Varese, Istituto Editoriale Cisalpino. g uichenon , S. 1660: Bibliotheca Sebusiana . Miscellae Centuriae II , Lugduni, Apud Guilelmum Barbier. l e g entil , P. 1957: «Ogier le Danois, héros épique», R 78: 199-233. l eJeune , R. 1948: Recherche sur le thème: les chansons de geste et l’histoire , Liège, Faculté de Philosophie et Lettres. l ot , F. 1940-1941: «À quelle époque remonte la connaissance de la légende d’Ogier le Danois? », R 66: 238-53. m anteYer , g. De 1908: La Provence du I er au XII e siècle , Paris, Picard. m anteYer , g. De 1925: Les origines du Dauphiné de Viennois. La première race des comtes d’Albon (843-1228) , Gap, Jean et Peyrot. MGH DD Karol. I = Monumenta Germaniae Historica. Diplomatum Karolinorum tomus I , Hannoverae, Impensis Bibliopolii Hahniani, 1906. On Ogier le Danois’s legend and the document of Oulx 99 DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 MGH SS Rer. Merov . = Monumenta Germaniae Historica. Scriptorum rerum Merovingicarum tomus V , ed. B. K ruSch / W. l eVinSon , Hannoverae et Lipsiae, Impensis Bibliopolii Hahniani, 1910. r iVautella , a./ B erta , F. 1753: Uliciensis ecclesiae chartarium , Augustae Taurinorum, Typis Regiis. t erreBaSSe , a. De 1875: Œuvres posthumes de Alfred de Terrebasse. Notice sur les Dauphins de Vienne , Vienne, Savigné. t ogeBY , K. 1969: Ogier le Danois dans les littératures Européennes , Munksgaard, DSL. V oretzSch , c. 1891: Über die Sage von Ogier dem Dänen und die Entstehung der Chevalerie Ogier , Halle, Niemeyer. 100 Giorgio Barachini DOI 10.24053/ VOX-2024-003 Vox Romanica 83 (2024): 83-100 On Ogier le Danois’s legend and the document of Oulx Abstract: This article deals with the sources of Ogier le Danois’ legend from the perspective of the document of Oulx, discovered in 2004 by Prof. G. A. Beckmann. Its first purpose is to examine once again the aforementioned document, in which among the witnesses we read the name (or names) Otgerius spata. g. curta . The document is analysed from a chronological, historical and philological perspective to test its reliability in the discussion about the origins of Ogier le Danois’ legend. The second purpose of this article concerns the use of such a document (or, more generally, of later documents) to give evidence for the existence of an oral tradition in Romance epic poetry, specifically Ogier’s legend. In many cases the documents are intrinsically and extrinsically doubtful. So, focusing on the tale(s) about Ogier, the article reflects on the Romance epic tradition and its differing interpretations. Keywords: Ogier le Danois, Archival research, French epic poetry, Chanson de geste, Traditionalism, Individualism 100 Giorgio Barachini