Navigation – Plan du site

The Origins of the Narrative Structures in the Chansons de Geste : Methodological Notes on Joël Grisward’s comparativism

Andrea Ghidoni

Résumés

La recherche par Joël Grisward sur les chansons de geste a introduit un élément important dans l'étude de la formation des légendes héroïques médiévales, à savoir l’élément folklorique ; ses études décrivent la réactivation du système trifonctionnel indo-européenne – selon les théories de Georges Dumézil – que Grisward tente d’extrapoler à partir de la structure narrative. Il est nécessaire de comprendre le modèle sous-jacent de l’œuvre de Grisward, puis de réfléchir sur sa méthode. La présente révision du modèle de Grisward utilise deux outils conceptuels et épistémologiques venus d’autres domaines : le premier est emprunté aux notes inédites sur les légendes germaniques rédigés par Saussure ; le second est le concept de catégorie polythétique formulée par Rodney Needham dans un contexte anthropologique, sur le modèle de réflexions similaires dans le domaine des sciences naturelles.

Haut de page

Texte intégral

  • 1 The masterpiece of Grisward’s comparativism is his Archéologie de l’épopée médiévale (Paris, Payot, (...)
  • 2 Given its importance for the incoming discourse, it is useful to outline briefly what we mean when (...)

11. Joël Grisward1 has probably given the most significant and innovative contribution to the debate on the origins of medieval French epics of the latest decades, reshaping our view of the genre of the chansons de geste. The Griswardian approach has introduced an important component in our way of conceiving the chansons de geste, i.e. the folkloric one. Moreover, his study is systematic and is applied to a song, the Narbonnais, which properly shows the reactivation of that trifunctional Indo-European structure that Grisward tries to extract from the narrative construction2. Thus, the role of folkloric elements comes to light within the complexity of the chansons de geste ; moreover, Grisward’s work demands – according to Georges Dumézil’s teachings – the inclusion of wide-ranging comparativism to the study of medieval literature : his results may be arguable, but they are not useless.

  • 3 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», Perspectives (...)
  • 4 J. Grisward, Archéologie de l’épopée médiévale, p. 19.

22. The best explication of the Griswardian model – particularly as opposed to the historicist one (whose view considers the chansons de geste as the output of an uninterrupted tradition starting from 9th century events of the Charlemagne’s era) –, may be found in an article he himself wrote in 19823, rather than in the introduction of Archéologie (especially at page 19). The said article aimed at clarifying some methodological statements appeared on page 19 of his Archéologie, in particular those regarding the historicist method backed among others by Gaston Paris. Grisward criticizes the traditional approach that establishes a hierarchy inside the so-called ʽNarbonnais cycleʼ, which is considered centred only on Guillaume d’Orange, the principal character of the epopee : such approach originates from the assumption that the cycle had been developed in the centuries around Guillaume with additions and aggregations of characters and events. Grisward rather prefers a wider and more structured consideration of the ensemble of the characters – Guillaume and his brothers : in Grisward’s opinion, « [p]oser le problème en terme de chronologie et d’histoire aboutit à une impasse : la théorie évolutionniste et le découpage chronologique n’enseignent rien sur le sept Aymerides et leur père qu’une préhistoire artificielle et illusoire». The historicist method leads us astray when it focuses on individual historical prototypes, but above all is unfit to catch « le groupe en tant que groupe». The search for historical prototypes establishes in a « diachronie fictive des éléments qui lui sont toujours donnés contemporains», it approaches something that « relève de la structure»4 with the lenses of historicity.

  • 5 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 126.

3In his 1982 article, Grisward intends to specify the meaning of these statements. He confirms that the group of Aymeri’s seven sons of and their sister is to be conceived as a block and a synchrony, « comme un ensemble articulé sur les trois fonctions sociales des Indo-européens. Sa genèse est de l’ordre de la structure». Its birth is adequately explained only by a « processus de transposition d’un moule conceptuel, d’une matrice conceptuelle, en personnages épiques, tels les Pāndava du Mahābhārata». On the contrary, the historicist way offers « une explication fragmentaire, diachronique, éclatée» for « un agrégat d’unités élémentaires séparées». Historicism projects the phenomena towards the unsolvable problem of the origins and is deceived by the illusion that the chanson de geste « a été fabriquée à partir d’ingrédients historiques». Grisward’s conclusion, playing with a well-known formula of large usage in epic studies, is that « [a]u commencement était la structure»5.

  • 6 R. Louis, « L’épopée française est carolingienne», in Coloquios de Rencesvalles, Zaragoza, 1956, p. (...)
  • 7 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 127.

4Unlike René Louis6, Grisward does not find in the relationship between the brave Guillaume and the weak king Louis the historically-attested link between Bernard of Septimania and Louis the Pious, finds in the texts of the so-called ʽGuillaume cycleʼ a dramatization of the « thème littéraire intégré à un ensemble de rapports structurés entre un type de souverain et un type de guerrier»7, whose prototype is the Danish hero Starcatherus : in these patterned myths, a bundle of internal connections is functional to a reflection on royalty facing the role of the warrior.

  • 8 Ibid., p. 129 : « ...il y a une structure, des structures, des schémas mythico-épiques indo-europée (...)
  • 9 Ibid, p. 129.

5The primacy of the structure and its long-term persistence are at the base of the Narbonnais geste (but there is a strong doubt that Grisward would extend this statement to the entire corpus of the chansons de geste)8. Therefore, Grisward’s structuralist comparativism rejects the historicist concept of the aggregation and transformation (in other words, the evolution) of a historical memory sinking its roots in a given event. However, the controversy is not intended to exclude these roots, but to « remettre à sa place [l’Histoire]». The metaphor of stratification, applicable to an operation of archéologie (programmatic even from the title of his opus maximum), is useful to explain the overthrow of the approaches that should be attempted when studying the epic phenomenon9 : History is what can be called a mere ʽtemporal hueʼ, i.e. the action of more recent times on the old material.

  • 10 Ibid., p. 130.

« L’Histoire offre avec l’épopée des points de convergence, des points de rencontre. Une manière d’analogie paraît lisible entre un certain réel et un certain imaginaire, et l’on sait que le second n’est pas la simple métamorphose du premier. Quel type de rapport entretiennent-ils ? Une certaine adéquation de la réalité aux structures préexistantes a-t-elle entraîné l’incarnation de ces structures dans cette réalité ?»10.

6Therefore, the reactivation of the structures should be placed in the 9th century (a William and a Bernard of that century were indeed respectively assigned the positions that the poem of Narbonnais assigns them), when a certain convergence between narrations and reality produces récits which are functional to a topical reflection, in this case on the set of problems referring to the division of the Carolingian Empire among the descendants of Charlemagne.

7However, we cannot help noting the extent to which this statement borders on a paradox, namely reality being adjusted to the imaginary. This is not Grisward’s thought, of course, but it is unavoidable that the survival of the structures collides with the convergence of some surface signifiers (the Carolingian location) toward some historical events ; such signifiers go beyond the full placement in the structure and so they are independent of and prior to the structure. Certainly, it is plausible that the signifiers giving ʽhueʼ to the mythic structures can be borrowed, if not suggested, from reality : this is precisely what transfert means. But such a conclusion cannot be extended to the whole corpus of the chansons de geste, unless one wants to affirm the surprising coincidence that all songs (or most of them) are founded on mythic structures reactivated by the convergence with historical events, almost as if the Carolingian history was nothing more than a pale replica or a reflection of very ancient myths : Grisward precisely mentions « thèmes historiques» that would have been a « miroir propre à susciter comme en reflet la reviviscence d’un univers parallèle mais épique». Basically, a series of historical materials would have provided a new signifier to pre-existing mythic structures specifically enlivened by the fact that the historical events already tended to malleability and vaguely followed the mythical universe of departure. While this reasoning is close to the paradox if applied to one song of deeds, it turns paradoxical if extended to the entire corpus. If indeed the epic universe of the chansons de geste (not only those belonging to the Narbonnais cycle) has a certain coherence – which involves characters, events, places ; and it is really coherent – it seems to me more plausible to assume that there has been a rearrangement of materials through an epic history-based universe built from suggestions of historical nature (such as historical legends unrelated to each other), i.e. a mythic structure that is not pre-existing, but more recent than the historical materials themselves ; this epic myth may have used as models also structured tales taken from a cultural heritage preceding the Carolingian times, but changing their functions toward a coherence that is built poem after poem, with a few initial songs as models for the subsequent ones.

8The Griswardian model rests on a fragile basis when it collides with the fact that the chanson de geste is in most cases based (especially the older ones) on historical materials that challenge the primacy of long-term structure, and with the fact that the ʽhistoricalʼ coloration of ethnic and ahistorical myths is not a painless and surface operation for the structures on which they are based, because also History can organise a mythical universe importing external narrative constructions.

93. Both the historical paradigm – the memory of historical events gets deformed up to achieve a tale with fantastic elements (evolutionary process) – and the Griswardian model – long-term mythic structures get historical coloration (historicizing process) – are not sufficient to explain the epic universe of the chansons de geste structured and founded on History (process of mythopoiesis or epogenesis) : the first one because it neglects (as is underlined by Grisward) the structure ; the second one because, even if it recognises the coherent narrative construction of individual songs, cycles and epic genre, it bends the histoire to its needs, failing to identify the possibility that mythopoiesis may take place starting from historical materials and that it may be placed not at the beginning of the cultural tradition but rather at its end. Grisward’s structuralism is weakened by the logic that the structural principle (what gives order and consistency to the narrative constructions) and the genealogical principle (the oldest materials in a cultural tradition) must necessarily coincide : if the chansons de geste are endowed with a symbolic structure, the latter has to be the legacy of a distant past.

  • 11 MGH, SS, II, p. 612. But the edition by Pertz in the MGH reads the passage in a different way and s (...)

10In the same 1982 article, Grisward offers another example of difficulty as related to historical materials that show traces of a structure that is plausibly independent of the Indo-European trifunctional structure. He reports the case of the Limousin Astronomer, author of a 9th century Vita Hludowici, who mentions a Hadhemarus in the retinue of the military expeditions (809-810 A.D.) led by Louis – who had not yet succeeded his father Charles – in the Iberian Peninsula : in particular, this man would use some guerrilla tactics similar to the ones characterising Aïmer le Chetif in the poem of the Narbonnais. This fact creates a bond between the character of the chronicle and the one of the song, but we must add that elsewhere Hadhemarus is said to be the Count of Narbonne and that the Astronomer himself, recounting the 801-803 expedition against Barcelona, mentions paired Hadhemarus and Wilhelmus, primus signifer ; in so doing, he builds once again the couple of second function (concerning war and force) identified by Grisward in the poem on the Narbonnais11.

  • 12 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 132-133.
  • 13 C. Carozzi, « Le dernier des Carolingiens : de l’histoire au mythe», Moyen Age, 82 (1976), p. 453-7 (...)
  • 14 J. Batany, « Renart et les modèles historiques de la duplicité vers l’an mille», in Third Internati (...)
  • 15 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 132.

11Grisward himself offers two solutions to explain such remarkable coincidence12. The first one is taken from considerations – in other fields – by Claude Carozzi13 and Jean Batany14 : some histoires would build around Hadhemarus a series of relationships focused on his name (link with the guerrilla, with Spain, with Louis, with Guillaume, etc.). « Le nom devient porteur d’un système de sens adéquat au système de sens véhiculé par le héros épique anonyme (et en quête d’un nom)» ; he would be given the traits of the legendary character of Hadhemarus : we could say that the aesthetics of the chanson de geste cares less for the historical verisimilitude than the plasticity and the malleability of the historical characters to enter a mythic structure with a Carolingian setting. Grisward continues15 :

L’introduction de l’Histoire dans le récit mythique (et du récit mythique dans l’Histoire) passerait elle-même par la structure, par une structure mythique intermédiaire, plus récente, contemporaine cette fois, peu ou prou, de l’insertion dans le temps de l’ensemble mythique indo-européen. Ainsi que l’écrit C. Carozzi : « La relation entre la Chanson et l’Histoire est indirecte puisqu’elle est médiatisée par le mythe véhiculé dans la tradition orale».

  • 16 The generic expression insertion dans le temps does not help : does it mean historical coloration o (...)

12Here, the model is similar to the one that I sketched above, which sees three components involved, i.e. the historical element, the ethnic myth and a history-based myth that has been recently formed. Indeed, the name of Hadhemarus would be the focus around which an entire system materialized, an historical narrative (or pseudo-historical) construction that would be further developed within another mythic configuration – epic this time – importing historical elements to fill an absence of names (i.e. of characters). The role played in this case by the Indo-European mythic set is not entirely understandable16, but the usage of the words introduction and insertion – respectively referred to récit mythique and ensemble mythique – suggests that the real stand of the chanson is the recent mythic structure that mediates between the historical materials and the ethnic ones of ancient origin ; these two kinds of materials are essentially united by being passively attracted by a pole, a point of convergence, that renews them.

  • 17 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 132.

13But this is not the paradigm supported by Grisward : it would mean accepting a structural dualism and giving up the functional primacy of the Indo-European structures, for the benefit of the recently-formed « structure mythique intermédiaire». On the contrary, the French scholar introduces a second possibility : the chronicler may have drawn the couple of warriors from « une épopée», from « une légende épique déjà constituée» ; in other words, he could have introduced in his narration imaginative aspects already belonging to an epic récit. Consequently, « Hadhemarus n’en [du Chétif] serait que le double, que la doublure rationalisée, ʽhistoriciséeʼ»17.

  • 18 MGH, SS, II, p. 612-614.

14One may argue that a chronicle contaminated by epic elements would contain many more imaginative aspects than the quotations we read in the case of Hadhemarus. In addition, this character in the same Vita Hludowici is also mentioned in other contexts that do not seem to repeat in any way the compact structure of the Narbonnais and the compagnonnage (there is no reference to kinship in the chronicle) with William of Toulouse. Even in the (short) passages where Hadhemarus is the protagonist of guerrilla actions18, he is always added to a list of commanders who have nothing to do with the epic. Therefore, it seems very rather debatable to identify contaminations in passages where a relationship is established with figures of attested historicity (William of Toulouse, for example).

  • 19 These aporias are noticed partly also by Jean Batany, in an article replying precisely to this piec (...)
  • 20 For the latter, see : M. Bonafin, « Fiaba e chanson de geste. Note in margine a una lettura del Voy (...)

15Grisward’s 1982 article does not manage to fully avoid all the difficulties that History creates to his structuralist model19. It is interesting, then, to discuss more in detail Grisward’s attempt to apply the Indo-European trifunctional structure to a chanson de geste among the oldest ones, that is the Chanson de Roland, which apparently do not exhibit any adherence to a mythic-fairy model as in the cases of the Narbonnais or of the Voyage de Charlemagne20. Indeed, the most important and famous French epic poem shows a solid historical system founded on the facts of Roncesvalles. The experiment conducted by Grisward thus provides material to understand the structuralist-comparativist approach where the historical material is abundant.

  • 21 J. Grisward, « L’or corrupteur et le soleil arrêté ou la substructure mythique de la Chanson de Rol (...)

164. Grisward dedicated an article to the Roland in 198121. The starting point is twofold as it moves from two themes-patterns in the Roland. The first is the episode of the stopping of the sun and of the delayed sunset due to Charles’s prayer (ed. Segre : laisses 178-179) : the miracle allows the French army to continue the battle that is going to win and avenge the death of its best knights at Roncesvalles. This ʽmarvellousʼ episode has been framed by critics from both an archeological point of view and a functional one. In the first case, it has been debated whether the statement of the Annales Anianenses (10th century) – which asserts that while Charles was winning at Zaragoza a rare astronomical event happened, « de hora nona factus est sol hora secunda», ʽmorning was transformed into afternoonʼ : probably it is a mistake to report a solar eclipse – is to be considered trace of a tradition prior to the chanson de geste we know. In the second case, the miraculous episode has been interpreted as a recovery of a biblical-folkloric pattern used to highlight the transcendent origin of the royal power. Not only : unlike other signs showing the divine predilection for Charles (dreams, angels sent to him), this is caused at first by Charles’s action. The characterization is not enough for Grisward : « La résonance d’un motif ou d’un thème ne se perçoit pas à travers une appréhension isolée ; un motif tire sa signification du système de rapports qu’à l’intérieur d’un ensemble il tisse avec les autres motifs, de sa place dans un discours, de sa situation dans une suite narrative».

17The second starting point is what would prompt Charles to intervene with this magical action : what drives Ganelon to betray. On this point, the text itself of the Oxonian manuscript seems to lean towards two options : Ganelon betrays in order to avenge the offense committed by Roland against Ganelon himself when he proposed his stepfather as ambassador to Marsilius ; Ganelon betrays simply due to auri sacra fames (l. 3756). In Grisward’s opinion – he adduces evidences from the Pseudo-Turpin and the Carmen de prodicione Guenonis (early 13th century) – the latter motivation is the genuine one (older, maybe) and it explains more efficiently the logic system underlying the narrative construction of the Roland.

18The corruption of Ganelon (linked to the heathens’ wealth) and Charles’s miracle are related respectively as cause and effect, as an action and a reaction within a trifunctional conception of power and society : the balance between the parties involved is broken and compromised by money (third function), but it is restored by a response which is military at the beginning and then magical-sacral (second and first functions).

19Therefore, according to Grisward, there are two myths belonging to the heritage of ancient Indo-European peoples, both placed at the beginning and at the foundation of a society (be it divine or historical) ; two myths showing the same dynamic. Some Scandinavian myths telling of the war between the Aesir and the Vanir show a similar structure : a hostile action by the Vanir (the sending to the Aesir of a being called ʽPower of goldʼ that sows discord among the women) is returned by a magical act by the king of the Aesir (i.e. Odin, who throws his formidable javelin against the enemies). The other myth is Roman and regards the siege of Rome by the Sabines : their king seduces (with jewels) the girl Tarpeia to gain access to the Capitoline with her complicity and only the prayer addressed to Jupiter by Romulus, barricaded with his troops on the Palatine, manages to resolve the battle in favor of Rome.

20The trifunctional structure is less evident in the Roland than in the Narbonnais, because in the latter chanson the schema is repeated through sequences or conform characters with triadic movements : the intention of distributing the signifiers in trifunctional groups is clear. Otherwise, in the Roland the trifunctionality, even if valid, is much less noticeable and involves episodes of the plot distant and extrapolated from their position : not to mention that they are fundamentally two (solar miracle and mercenary betrayal), whereas the second function (on war and force) is spread over the entire epic setting. If anything, the trifunctional logic emerges comparing this song with the above-mentioned Scandinavian and Latin myths, to confirm the existence of a Indo-European narrative structure acting as archetype. But some doubts persist about the fact that the structural identity can be said unchangeable and uninterrupted (this is related to the question about the real existence of the structure).

21The Scandinavian and Latin myths consisting maybe of the same functional framework of the Roland do not share, apart from the deep structure, any surface feature, but a few generic traits. Hence, we should question ourselves on how this property has been maintained to generate new stories in cultural areas that are distant from those of origin and what represents common identity for these stories. Among the myths chosen by Grisward to exemplify the pattern of the corrupting gold in foundational myths (pattern based on the logic of the three Indo-European functions) there is certainly a family resemblance, but this is not sufficient to establish any kinship between the narrations.

  • 22 J. Grisward, « L’or corrupteur et le soleil arrêté ou la substructure mythique de la Chanson de Rol (...)

22If we wanted to outline the common structure for the identities of all three stories (Roland, the myth of the Aesir and the Roman one), we would have a reduction to the backbone of the common traits. There are very few points of convergence between the three narrative structures and these are very generic : if we look at the details and at the complexity of the signifiers, distinctions become more evident. In Grisward’s view, the cornerstones of the scheme are the corruption par l’or by the Riches and the grande magie by the Forts and the Religieux22 : substantially, a binary syntagmatic movement founded on a ternary paradigmatic structure. However, the pattern of the corrupting gold is well differentiated if we consider the feminine role, only present in the Scandinavian myth and in the Roman one ; furthermore the distinctive feature between these two myths is the betrayal, which misses completely in the Scandinavian myth.

  • 23 We suggest reading of : R. Needham, « Polythetic Classification : Convergence and Consequences», Ma (...)

23Therefore, in Grisward’s intent, structural consistency between the three stories should be evident, actually this similarity gets generic and the mobile identity of this structure can be described only by means of a model that can be defined as « polythetic». Polythetic classifications in anthropology have been used by Rodney Needham, partly borrowed from similar reflections in the natural sciences : they are a useful tool for the analysis, for example, of institutions of the social law, present in different cultures, similar but not perfectly overlapping each other ; so, the clarity of a homogeneous classification (impossible) will be replaced by a sort of blurry picture that allows reflections more productive and profitable. The object of study is dissolved in a bundle of features, without hierarchies of any kind between them and without bonds determining the inclusion or not within the limits of the class – rather : strictly speaking, there should be no limits –, features present in many objects but not necessarily in all. For Needham, polythetic classifications would have the advantage of being more realistic than a monolithic order of words and concepts : monothetic categorisation, based on formal identities formal and on the principles of the Aristotelian logic and of the Western tradition, denies the fuzzy nature of the things23.

24The structure of the motif of the corrupting gold is founded on abstraction and it cannot avoid to erase certain qualifying and differentiating elements, which thus remain at the margin, such as the role of the woman and the fact that she may be seduced without necessarily committing betrayal : if we assume as a cornerstone the seduction of women with gold, we strengthen the unity of classification only for the Scandinavian myth and the Latin one ; on the contrary, taking the perspective of the betrayal, the pattern can be inflected into different versions, with the exclusion of the Scandinavian myth. It is basically a matter of point of view. Such a degree of abstraction can be useful when we classify the examined material according to one or more view angles, but it may not be fit for carrying a genetic study.

25Perhaps the only trait shared by the three myths is the role of gold as a source of seduction and corruption. But it might be deemed as methodologically unconvincing to trace the diachronic identity of the structure in its complexity from a single feature as a guideline, because a structural identity is much more than a single or a few features, but a broader combination of elements.

26Rather, a historical reconstruction may better succeed on the syntagmatic plan of the signifiers and it could recognize traditions and innovations not from circumstantial elements that characterize similar narrative situations, but from a recognizable structure, that would make possible to exclude the polygenesis. In order to do this, it is also necessary to find a monothetic definition of the pattern (a definition not as thin as Grisward’s one, though), while the polythetic approach is useful only in the analysis of a single object, a prototype, on which it is possible to cast light to increase the degree of informativeness thanks to the comparison with similar but not identical witnesses. This restriction does not allow us to go too far back to the roots of the motif, but at least it provides reasonably reliable data to build assumptions or more detailed analyses of the inventoried material. The setting of genetic lines of a pattern based on rarefied analogies risks to become just a suggestive hypothesis and nothing more if deprived of precise material evidence and of certifications of the longue durée of the identity of the motif (an identity, of course, specified and not generic).

  • 24 « By this account of the matter, the first consequence of the adoption of polythetic classification (...)

275. This limit of the polythetic classifications applied to the human sciences is emphasized by Rodney Needham in an article written in 1975. Needham believes that the value of the polythetic classifications is to provide « a high content of information» and to better adhere to the object of study, but he repeats that the morphological comparison is however a price to pay24.

28Such a warning must be born in mind working on a genetic hypothesis. On this point there is a strong divergence between natural sciences and human ones, as the former can almost with certainty determine genetic and evolutionary relationships, for example in biological taxonomy, as between members of the polythetic class you can still recognize « the monothetic core of common properties» and to distinguish the elementary parts that compose the individuals of the class (what Needham denies in the case of human facts). Moreover, the genetic-evolutionary scheme requires guidelines and distinctive elements to be set :

  • 25 Ibid., p. 360.

This implies that sociological taxonomy is highly artificial (as in fact it is), and such that the evolution of institutions may be disparately inferred according to the typological features that are more or less arbitrarily decided upon. There is no sign, at any rate, of a correspondence, with quasi-natural monothetic clusters of features. Rather, it is the construction of a sociological typology of cultural particulars by monothetic criteria, and these too of the most slender and unreliable kind, that is the chief and irremediable weakness of this evolutionary scheme.25

29To draw genetic lines, we need « resemblances» inside the subject matter of the research scarcely compatible with a polythetic classification calling for a wider variety of elements :

  • 26 Ibid., p. 360.

as soon as the polythetic character of the resemblances in question is recognised, the grounds of the taxonomy on which the study rests [i.e. evolutionary hypothesis] are quite removed [...]. Where, therefore, [a set of entities] under comparison are analysed by polythetic criteria, instead of being typed by a few monothetic features, the presumed resemblances are reduced or abolished ; the comparison is vitiated, and the attempt to work out evolutionary interconnexions is thereby doomed to failure.26

  • 27 Ibid., p. 362.
  • 28 Ibid., p. 365 ; Needham is quoting an aphorism (not published) by Evans-Pritchard.

30The polythetic approach will be far more useful – Needham argues – only in the method of the "cluster analysis", i.e. in a statistical analysis where the number of variables is very high. Of course, we must put aside any definition-oriented comparison, when asked « whether a definition of a polythetic concept is after all a definition, since it is certainly imprecise»27 : difficulties increase when the compared material increases, as it happens when facing semantically complex human phenomena. In essence, the idea of ​​polythetic class offers new perspectives and ways of thinking but in fact complicates comparative studies confirming that « there’s only one method in social anthropology, the comparative method – and that’s impossible»28.

31In conclusion, returning to the pattern of the corrupting gold, we note that the three narrative structures found by Grisward are not comparable and therefore the identity of any Indo-European scheme is confusing and opaque. The persistence of these structures in a limited number of generic traits is not sufficient to establish a family relationship, and even if they existed it would be difficult to say that the structure remains the same to the change of the signifiers in which it is embodied. The signifiers – by which I mean always what Grisward defines as the latest coat of paint or the top layer of a narrative : historical materials in the case of the chansons de geste –, far from being only the semblance in which a durable structure is embodied, are, on the contrary, the elements around which is to coagulate a new structure, a framework which can also be formed of materials borrowed from more or less ancient models, but that is modified to allow the emergence of the cultural specificities of the new text or the relationships implicit in the signifiers, which are not inert, but affect the structure.

326. In the case of the Narbonnais – poem which, differently from Roland, does not show historical roots and, instead, presents a mechanical, repetitive and fairytale narrative structure –, the triadic placement of the elements, both on the paradigmatic plan (the relations between the characters, for example, or between the cities) and on the syntagmatic one (the distribution of estates and titles by the father), respects the trifunctional logic. It is a narrative scheme replicated constantly in the text as shown Grisward in Archéologie de l’épopée médiévale : be enough the reference to this, without going into further demonstrations.

  • 29 Preparatory to the comprehension of the following paragraphs is the knowledge of the contents of an (...)

33Confine ourselves to the study of the pattern of the father-king who distributes lands and titles (and social functions) among his children, the so-called partage du monde. The narrative scheme of the song is found by Grisward in other texts, some very ancient (in Indo-Iranian and Persian contexts) and one more recent (the Nerbonesi by Andrea da Barberino, text, according to Grisward, depository of a variant of the epic cycle that could conserve some archaic traits)29.

  • 30 J. Grisward, Archéologie de l’épopée médiévale, p. 81.

« Les analyses [...] ont mis en lumière dans Les Narbonnais deux schémas dans lesquels la structure trifonctionnelle et la comparaison avec le témoins indo-iraniens invitent à reconnaître un héritage indo-européen : le vieil Aymeri, fidèle à un type mythique très ancien (indo-européen), accomplit les gestes fondamentaux de tout « premier roi» ; il organise la société dans ses divisions fonctionnelles [...]. Organisateur d’une société tripartie et partageur d’un monde caractérisé trifonctionnellement, Aymeri de Narbonne représente donc face à la Scythie, à l’Iran et à l’Inde, le témoin occidental le plus précieux de ce mythe d’origine indo-européen qui expliquait par un partage trifonctionnel entre frères l’organisation sociale et géographique du monde»30.

  • 31 R. Needham, « Polythetic Classification : Convergence and Consequences», p. 361.

34Even in this case the fundamental scheme of the pattern can be described with a polythetic model, with a list of pertinent features but whose presence (or absence) is not required to include (or exclude) a witness within (or from) the range of the motif of the partage du monde. More precisely, the exchange of single sections between a myth/story and the other allows to strengthen the affinity and the family resemblance between the witnesses, despite the traits present in all occurrences are also in this case very few. An almost complete scheme is not observable in any witness, except perhaps by combining the different variants that cleave individual traditions : on the one hand, the Indo-Iranian myths ; on the other hand, the stories about Aymeri/Amerigo, i.e. the Narbonnais and the prose work of Andrea da Barberino. In this way each of the two traditions would collect the pattern of the dividing up of the inheritance, the trifunctional logic (not always present in all witnesses) and the motif of the trial of the sons by the father. A unitary identity, however, remains non-existent and is only an abstraction, like the smelling panther imagined by Dante in his De vulgari eloquentia, sniffed everywhere but impossible to find. I think at this point useful to recall some words of Rodney Needham about the « abstract stipulation», that is a polythetic model politetico assumed as starting point for genetic derivations31 :

Even in this quite promising line of “evolutionary” investigation, therefore, the parallels with a natural science are superficial. An abstract stipulation occupies the place of a common ancestor ; the similarity of social forms has nothing to do with common descent.

  • 32 J. Grisward, Archéologie de l’épopée médiévale, p. 123-124.

35What I wish to emphasize is the difficulty of establishing the identity of the structure that would be reactivated or reincarnated in every witness of the myth, in all its variations. A statement like the one by Grisward, according to which32

les deux variantes, française et italienne, de la légende d’Aymeri de Narbonne et de ses fils sont bâties sur le même schéma que l’aventure de Yayāti et de ses petit-fils. Mais si la structure est identique, la mise en œuvre et le contexte sont sensiblement différents...

36is incorrect from the point of view of the paradigm underlying. In fact, the identity is not monolithic and it is not credible that a myth presenting a single structure has always been dismembered among his witnesses. Rather, if it is legitimate to speak of long-term traditions, it seems more plausible to think of a group of tales or myths strongly intertwined with each other because of structural similarity, a certain amount of combinations of a set of elements constantly open to rework. But in this case we cannot speak of an identical structure varied depending on the implementation and the context, but it is the action of synthetic and combinatorial implementation to create the structure, from fragmentary connections of the elements and suggestions induced by myths and stories very similar to each other. The structural elements may even be ancient, as well as the relationships that can exist ​​between them, but the set is only a hazy nebula of fragmentary relations, whereas a structure that is unique and compact is the scope of each single recombination and recreation.

37Conceiving the compositional scheme as an itinerant structure that is nearly always identical implies disregarding the unstable nature of the folkloric culture, lacking of bonds and untouchable narrative models, whereas, on the other hand, individual patterns are reproducible, interchangeable between them within structural frames similar but not identical. In this sense, a polythetic model – with which the reconstruction of genetic lines and abstract ancestors becomes risky and incorrect – gains a diachronic function as it provides an idea of the complexity of the narrative material that coagulates in the combination that gives rise to a single text, in which the inherited narrative modules acquire new functions.

  • 33 The individual witness of a pattern is therefore a particular setting of predetermined elements, a (...)

38What distinguishes the composition of the Narbonnais from the variants of the other witnesses of the motif of the partage du monde highlights that at the base of the composition of the song there is not the imposition of an ancient structure reincarnated and actualized, but individual modules exploited to achieve the formation of a new structure. In fact, this poem, compared to other witnesses of the pattern of the partage du monde, is not only based on the trifunctional logic, but it is also the only witness of the motif to replicate the triadic module on multiple levels and with insistence : therefore, the partage involves not only lands, but also social appointments ; likewise the trifunctional logic acts later in the plot, as the characters are coherent to their functional nature and the narrative sequences replicates continuously tripartite modules. The trifunctional logic permeates the entire chanson de geste, as shown by Grisward, so as to give the impression that it is not a passively replicated pattern, but the insisted reuse of a dynamic module that generates new structures, just to give the song a coherent and systematic architecture and a clear mythic frame : what corresponds to the aesthetics of the 13th century chansons de geste, interwoven with fictional and folkloric elements and with a certain tendency to the fairy style33.

  • 34 The pattern of the enfances, in its entirety, provides « l’exhérédation du futur souverain, son exi (...)

39Then, the models may have been one or more ethnic myths known to the culture of the milieu from which the text stems, one or more tales morphologically not distant from the myths analyzed above, whose structure and patterns have been isolated and reused freely – with, so to speak, ludic insistence – on the narrative support of the chanson de geste (and in particular, on that of the subgenre of the enfances)34.

407. Grisward’s genealogical hypotheses could be accepted (not always and not as a whole), what instead is less convincing is precisely the model underlying this type of transfert, where long-term mythic structures would wear a ʽrefreshedʼ dress and where the so-called ʽCarolingianʼ signifiers would be the surface layer compared to the deep structures. What has emerged is essentially the scattering of the traces of descent of the French epic poetry and its mythemes from a remote past of cultural traditions and therefore the impossibility of finding the source that gave rise to the individual songs or to the epic cycles. The genre of the chansons de geste is equipped with a substantial narrative unity missing in the cultural traditions prior to 11th century and the time frame preceding the epogenesis is one that contemplates a breaking of the structures and a crumbling of the constitutive elements of the songs : elements which only in these songs are immersed in a series of functional relationships.

  • 35 If actually this model emerges in Saussure’s notes, is matter of discussion. D’Arco Silvio Avalle i (...)
  • 36 For the notes on Germanic legends, I follow the text of the anthology edited by A. Marinetti e M. M (...)

41The model that seems most appropriate to represent this cultural frame is what emerges – or at least it can be distinguished – from the so-called Notes on Germanic Legends by Ferdinand de Saussure35. The Saussurian reflections are addressed to the theory of the sign-character, but their semiological model is still valid and can also be extended to other types of sign, as a recognizable structure in a narrative text or an epic universe taken as a paradigm of a literary genre. The Saussurian notes36 are kept in two boxes in the Bibliothèque Publique et Universitaire of Ginevra, numbered Ms. fr. 3958 and Ms. fr. 3959 ; they have as main topic a series of reflections on two medieval legendary cores, the legend of the Nibelungs and that of Tristan : the intent of Saussure is the identification of historical prototypes in the first case, of mythical prototypes in the second. What is relevant for us, however, is the material that is open to theoretical perspectives, a series of notes of considerable importance in the semiological field although it must be admitted that they are extremely scarce compared to the amount of notes of practical address.

42In the case of the Nibelungs, Saussure attempts to reconstruct the link between the legend and its variants and the Burgundian historical facts that are the basis of the récits, clearly net of all the changes and all the tranferts that have occurred in the cultural tradition. It is therefore a methodological problem, i.e. what are the correct steps to identify the invariant particles that can allow the recognition and the identification of mythical and historical events.

43However, the historicist conception of departure is already highly innovative. Saussure conceives the transformation as an extremely complex and sophisticated process that requires the scholar to adopt a method describing the phenomenological complexity and locating the invariants. Instead, in the Tristanian corpus the transformation involves not only historical events but also other legends, in this case Greek myths. Therefore, the framework of the cultural tradition is complicated by the interplay of multiple components, whose coupling produces variations that dim the identity of the legend in its witnesses and lends narratives to a componential analysis contrasting with their consideration as unities.

  • 37 « 1. Dans une légende donnée, et sans autre moyen de contrôle, il est faux de croire qu’il soit plu (...)

44Faced with this complex picture, the historical legend can be traced in its origins only by appealing to the traits that remain unchanged over the course of the tradition : the story must contain a minimum basis of elements in which it is possible to recognize the historical material. Thus, the comparison between legends is compromised : it is impossible to formulate a supporting core to compare legends, a crochet on which hang from in order to establish genealogical relationships (see : Note 3959/11)37.

45These Saussure’s statements are connected with what has been said about the comparativist method. Comparison among different legends is possible only on a documentary level, while genetic links cannot be pursued because it is not possible to find a crochet to build a stable identity : the result is subdued to the danger of reconstructing a « légende inconnue» as an ancestor calling to mind Needham’s « abstract stipulation». Also the comparison between variants of the same legend is very likely to be sterile from a genealogical point of view, due to the extreme complexity of the internal combination of the elements ; moreover, it is impossible still another time to assign a primacy to any among such traits – it mirrors what I have just said about using an elastic and polythetic set of tales or patterns with a genealogical function. Thus, only history-based legends, according to Saussure, can be equipped with a somehow fixed identity, because at least the historical events supposed original can work as archetype for certain details of the plot. Of course, leaving the narrowness of the historicism, we cannot deny the possibility to also follow the origins of non-historical legends : but the crochet is always necessary to settle the identity and this is hardly provable for Grisward’s long-term mythic structures. Comparison between « légendes en elles-mêmes» – this expression is used by Saussure immediately below the passage we have just commented – requires a monothetic definition or however a less elastic one, too.

46The discourse on the sign identity in these Saussure’s notes concerns in particular the unities of the legend, i.e. the character-symbol (Note 3959/4 p. 1) :

Il n’y a aucune méthode à supposer que le symbole doive rester fixe, ni qu’il doive varier indéfiniment, il doit probablement varier dans certaines limites. [...] Donc en principe, on devrait purement renoncer à suivre, vu que la somme des modifications n’est pas calculable. En fait, nous voyons qu’on peut relativement espérer suivre, même à de grands intervalles de temps et de distance.

47Saussure confirms that the incalculability of the modifications takes place on a semiological and theoretical plan, although a minimum degree of traceability is basically possible, provided that, as we have seen above, a historical basis is available. An extremely fragile conception of identity emerges, in which even the littlest change of the combination of the features forming a symbol at a given time can move the entire layout of the identity.

  • 38 Cf. Note 3958/8 p. 21v : « Comme on le voit, au fond l’incapacité à maintenir une identité certaine (...)
  • 39 On the illusions of literary scholar following the vicissitudes of narrative sign, Saussure comes b (...)

48The symbol, which is the semiological unit of the legend, is unable to maintain a stable identity and Saussure persists in arguing that any semiological definition is the result of a purely scholarly operation : from a semiological point of view, the transformations of a legend and its components are not related to time or to the accidents of tradition but they are the result of the inability of the sign-symbol to confirm itself identical in each witness of the legend38. Saussure rejects the existence of the unity of narrative sign, which would remain such only for the effect of a scholarly definition or of a formula, while what really can be equipped with existence is the single element, which – assembled with other elements – lets emerge the blurry appearance of a sign39.

49Saussure is uncertain between two points of view. Adopting the lenses of the semiologist, he cannot but recognize the associative and componential nature of signs and legends, properties that prevent the establishment of a diachronic identity more or less recognizable ; on the other hand, he also has confidence in the practical act of reconstructing the history of sign and legend, provided that the latter have a historical basis or otherwise a secure handhold, i.e. a prototypical identity of the sign by which you can calculate the transformations of narrative signs.

  • 40 D. S. Avalle, Ferdinand de Saussure tra strutturalismo e semiologia, p. 106.

50The comment by Avalle is as follows40 :

La semiologia della narratività non è più tanto la scienza dei possibili narrativi in quanto tali, quanto piuttosto della « lettura della narrazione», che equivale a dire dell’osservatore esterno. Solo in questo settore è lecito parlare di simbolo con uno stacco dalla matrice storica che in effetti non è mai stato consumato in nessun momento della evoluzione della leggenda.

  • 41 Ibid., p. 107.
  • 42 See above : Note 3959/4 p.1.

51Applying the ideas of Saussure-Avalle to Griswardian comparativism, in my opinion this is revealed as a method for tracking – in legends that maintain few cultural ties (not to say none) one with the other – structures only recognizable by the external observer who puts in succession and then in perspective the various witnesses of the structure-sign. What is missing is the crochet, which undoubtedly cannot be a functional structure – such as the one built by Grisward – whose signifiers can be considered secondary, but otherwise it must be a combination of signifiers, as it can be the narration – concrete and not abstract – of a historical event ; the « matrice storica» introduced by Avalle, in the case of ahistorical myths, is instead a text (or a pattern, a narrative structure with its symbols) whose identity – certainly not immutable – can be traced in those texts that use it as a source or that are culturally contiguous and later, so that the transformational process that governs the relationship between the witnesses of the legend (or of the pattern) can be rebuilt. Only in this way it is possible to reconstruct « il nesso che lega le differenti versioni di una leggenda le une alle altre, attraverso processi di vario di tipo “di riduzione di proporzioni o di amplificazione degli avvenimenti”». Under these conditions, for Saussure, « sembra che sia ancora possibile « seguire» una leggenda nella sua evoluzione storica»41. The signs, far from remaining fixed, are subject to constant revision, are contingent tabular systems, even if certainly the transformation is limited to the short term (Saussure does not exclude longer periods, but it is always necessary to establish the limits of transformation ; it should be pointed out that the « grands intervalles de temps» of Saussure are certainly not those of Grisward) : due to this, it is at least necessary a temporal and cultural ʽconcrete contactʼ between the variants of the sign to be compared, so that the processes of transmission can be recognised. Saussure’s position is neither absolute skepticism nor the (positivist) optimism of genealogical reconstruction, it covers rather a middle course between reconstruction of transformational processes and recognition of the dispersion of the signs during this same process : « [i]l n’y a aucune méthode à supposer que le symbole doive rester fixe, ni qu’il doive varier indéfiniment, il doit probablement varier dans certaines limites»42.

52The recovery of diachronic identity of the narrative sign is possible only from a close comparison of texts that maintain together historical, textual, cultural, geographical relations rather firm, as well as a community of signifiers, and this method can be called philologically correct, since it considers the literary text in its historical nature. Instead, an approach semiologically correct – providing a detachment from the historical and cultural matrix of the texts – compares the signs of the legends (or their narrative structures) not for the purpose of obtaining genealogical data but to increase knowledge about the « possibili narrativi», to build at most the nebula of elements joining with various combinations the establishment of a cultural heritage – which is, as I said before, perhaps the only way in which a polythetic and flexible set can provide useful validation from a genealogical viewpoint.

  • 43 Note 3959/10 p. 18 : « Ce qui fait la noblesse de la légende comme de la langue, c’est que condamné (...)

53The long-term structural identity cannot be considered valid for genetic reconstruction : assumed as immutable the structure in the course of the comparativist research, the signifiers become symbols of functions and the « fait naturel de transmission sémiologique» – which, according, to Saussure, engenders the narrative – gets lost ; the Saussurian sign of the Notes does not consist of fixed meanings (functional structure) and variable signifiers (symbols), but the structure escapes from the combination of the signifiers in every moment of their varied transmission43.

  • 44 D. S. Avalle, Ferdinand de Saussure tra strutturalismo e semiologia, p. 103-104.
  • 45 Note 3959/7 p. 35v : « Un des éléments de destruction semés d’avance pour l’histoire dans ‹le terra (...)
  • 46 The article published here is an excerpt from my doctoral thesis discussed at the Università di Mac (...)

548. In conclusion, the Griswardian structure – based on functions and always identical to itself regardless of the éléments – stiffens the process of transmission of the legends, i.e. their cultural tradition, whereas meaning and sens change actually with each new narrative construction. What resists time (at least within certain limits) are a few clusters of signifiers, which do not cease to interact with other elements up to lose after a certain time any glimmer of identity. The fixed identity (or at least minimally elastic) is the cynosure of a search method limited to the comparison of systems « a struttura associativa statica o sincronica»44. A genealogical approach instead has to do with systems « a struttura associativa dinamica o diacronica». In the phenomenology of literature, the first systems correspond to « momento unitario di un tutto finito dove le varie parti si tengono in equilibrio (ἔργoν)», while the second system to « varie strategie messe in atto dal locutore al fine di raggiungere tale equilibrio (ἐνέργεια)», to « farsi di un tale organismo, alla cronologia delle varianti». The new equilibrium is a ʽdramatic unityʼ subordinating to itself the relations previously maintained by narrative materials45. In the case of the genre of the chansons de geste, the genealogical approach can identify at best the materials borrowed from history or folklore in each of the chansons, i.e. the masses that establish, variously combined, new connections and new structures between themselves46.

Haut de page

Bibliographie

d’Arco Silvio Avalle, « Dai sistemi di segni alle nebulose di elementi», Strumenti critici, 19 (1972), p. 229-242

d’Arco Silvio Avalle, L’ontologia del segno in Saussure, Torino, Giappichelli, 1973

d’Arco Silvio Avalle, Dal mito alla letteratura e ritorno, Milano, il Saggiatore-Mondadori, 1990

d’Arco Silvio Avalle, Ferdinand de Saussure tra strutturalismo e semiologia, Bologna, Mulino, 1995

Jean Batany, « Renart et les modèles historiques de la duplicité vers l’an mille», in Third International Beast Epic, Fable and Fabliau, Koln, Böhlau, 1981, p. 1-24

Jean Batany, « Mythes indo-européens ou mythe des Indo-européens : le témoignage médiéval», Annales E.S.C., 2 (1985), p. 415-422

Massimo Bonafin, « Fiaba e chanson de geste. Note in margine a una lettura del Voyage de Charlemagne», Medioevo romanzo, 9 (1984), p. 3-16

Massimo Bonafin, « I guerrieri al simposio. Morfologia di un motivo», L’immagine riflessa, 2 (1993), p. 1-37

Massimo Bonafin, « Prove di un’antropologia del personaggio», in Le vie del racconto. Temi antropologici, nuclei mitici e rielaborazione letteraria nella narrazione medievale germanica e romanza, Padova, Unipress, 2008, p. 3-18

Massimo Bonafin, Guerrieri al simposio. Il Voyage de Charlemagne e la tradizione dei vanti, Alessandria, Edizioni dell’Orso, 2010

Giovanni Bottiroli, « Differenze di famiglia», in Problemi del personaggio, Bergamo, Il Sestante, 2001

Dominique Boutet, « Du mythe à la chanson de geste : le problème de 1’ajustement dans les Narbonnais», Six études sur la chanson de geste, Revue des langues romanes, 91-1 (1987), p. 25-35

Jeremy Brough, « The Tripartite Ideology of the Indo-Europeans : An Experiment in Method», Bulletin of the School of Oriental and African Studies, 12 (1959), p. 69-85

Claude Carozzi, « Le dernier des Carolingiens : de l’histoire au mythe», Moyen Age, 82 (1976), p. 453-76

Marcel Detienne, Comparer l’incomparable, Paris, Éd. du Seuil, 2000

Georges, Dumézil, L’idéologie tripartite des Indo-Européens, Bruxelles, Latomus, 1958

Georges Dumézil, Mythe et épopée, Paris, Gallimard, 1968-1973

Georges Dumézil, Heur et malheur du guerrier, aspects de la fonction guerrière chez les Indo-Européens, Paris, Presse Universitaire de France, 1969

Andrea Fassò, « Miti indoeuropei ed epica comparata», in Le forme e la storia, 2 (1990), p. 359-93

Andrea Fassò, « L’ideologia tripartita», in Lo spazio letterario del medioevo. 2. Il medioevo volgare, Roma, Salerno Editrice, 2001, p. 83-104

Andrea Ghidoni, « Il motivo dei vanti tra radici storiche e archetipi : alcune riflessioni di metodo sulle classificazioni politetiche», L’immagine riflessa, 20 (2011), p. 155-180

Joël Grisward, Archéologie de l’épopée médiévale, Paris, Payot, 1981

Joël Grisward, « L’or corrupteur et le soleil arrêté ou la substructure mythique de la Chanson de Roland», in Georges Dumézil. Cahier pour un Temps, Paris, Pandora, 1981, p. 257-70

Joël Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», Perspectives médiévales, 8 (1982), p. 125-133

Cristiano Grottanelli, « Temi duméziliani fuori dal mondo indoeuropeo», Opus, 2 (1983), p. 365-389

Cristiano Grottanelli, Ideologie miti massacri. Indoeuropei di Georges Dumézil, Palermo, Sellerio, 1993

René Louis, « L’épopée française est carolingienne», in Coloquios de Rencesvalles, Zaragoza, 1956, p. 327-460

Rodney Needham, « Polythetic Classification : Convergence and Consequences», Man n.s. 10 (1975), p. 349-369

Rodney Needham, Casi esemplari, Milano, Medusa, 2002

Rodney Needham, Caratteri primordiali, Milano, Medusa, 2006

Aldo Prosdocimi, « Sul Saussure delle leggende germaniche», Cahiers Ferdinand de Saussure, 37 (1983), p. 35-106

Benson Saler, Conceptualizing Religion. Immanent Anthropologists, Trascendent Natives, and Unbounded Categories, New York, Brill, 1993

Ferdinand de Saussure, Le leggende germaniche, Padova, Zielo-Este, 1986

Bernard Sergent, « Observations sur l’origine du cycle des Narbonnais», Romania, 105 (1984), p. 462-491

Somiglianze di famiglia : tipologie e classificazioni fra scienza e letteratura, edited by M. Bonafin, Alessandria, Edizioni dell’Orso, 2011

François Suard, Guide de la chanson de geste et de sa postérité littéraire (XIe-XVe siècle), Paris, Champion, 2011

Haut de page

Notes

1 The masterpiece of Grisward’s comparativism is his Archéologie de l’épopée médiévale (Paris, Payot, 1981).

2 Given its importance for the incoming discourse, it is useful to outline briefly what we mean when we talk of trifunctional Indo-European ideology. According to Georges Dumézil, the majority of the cultural elements (social conception, myths, theology, literature, etc.) of this ancestral people (but it would be rather right to employ plural terms) were marked by an ideology that outlined three spheres of beliefs or functions of human activities, in a hierarchic order : the first one linked to sacred life, to knowledge and to sovereignty ; the second one connected with physical force, military power and weapons ; the third one fixed on economic aspects, on every form of wealth, on agricultural activities, on fecundity and fertility. Among the numerous works by Dumézil we quote here only Mythe et épopée (Paris, Gallimard, 1968-1973).

3 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», Perspectives médiévales, 8 (1982), p. 125-133.

4 J. Grisward, Archéologie de l’épopée médiévale, p. 19.

5 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 126.

6 R. Louis, « L’épopée française est carolingienne», in Coloquios de Rencesvalles, Zaragoza, 1956, p. 327-460.

7 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 127.

8 Ibid., p. 129 : « ...il y a une structure, des structures, des schémas mythico-épiques indo-européens, des “histoires”. L’épopée française ne surgit pas comme une création spontanée du VIIIe ou du IXe siècle. [...] L’histoire carolingienne n’est pas primitive, elle ne consiste pas le point zéro d’où il faut partir. L’analyse structurale met au jour l’existence d’une préhistoire ou d’une ultra-histoire de la chanson de geste [...]. L’épopée française [...] n’est pas engendrée par cette Histoire carolingienne transformée, “embellie”, tripatouillée».

9 Ibid, p. 129.

10 Ibid., p. 130.

11 MGH, SS, II, p. 612. But the edition by Pertz in the MGH reads the passage in a different way and so changes its meaning : « Willelmus primus, signifer Hadhemarus». The passage is under discussion, so we do not go further and consider the text of Grisward (who, however, does not mention his reference edition).

12 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 132-133.

13 C. Carozzi, « Le dernier des Carolingiens : de l’histoire au mythe», Moyen Age, 82 (1976), p. 453-76.

14 J. Batany, « Renart et les modèles historiques de la duplicité vers l’an mille», in Third International Beast Epic, Fable and Fabliau, Koln, Böhlau, 1981, p. 1-24.

15 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 132.

16 The generic expression insertion dans le temps does not help : does it mean historical coloration or simple recovery of actuality for a narrative structure ?

17 J. Grisward, « Épopée indo-européenne et épopée médiévale : histoire ou histoires ?», p. 132.

18 MGH, SS, II, p. 612-614.

19 These aporias are noticed partly also by Jean Batany, in an article replying precisely to this piece by Grisward (J. Batany, « Mythes indo-européens ou mythe des Indo-européens : le témoignage médiéval», Annales E.S.C., 2 (1985), p. 415-422). The scholar, who can be defined a historicist neotraditionalist, underlines the primacy of history as compared with myth, but he concludes that the one does not exclude the other : « Les constructions narratives de l’épopée peuvent être issues à la fois du mythe et de strates de vérité historique dont l’une ou l’autre a pu jouer un rôle primordial (la strate carolingienne dans la chanson de geste, celle de l’an mille dans Renart). Les études de type dumézilien risquent évidemment de négliger l’apport « historique», et d’oublier ce qu’il a pu apporter d’original dans chaque culture» (p. 419).

20 For the latter, see : M. Bonafin, « Fiaba e chanson de geste. Note in margine a una lettura del Voyage de Charlemagne», Medioevo romanzo, 9 (1984), p. 3-16 (p. 10 ; p. 15-16).

21 J. Grisward, « L’or corrupteur et le soleil arrêté ou la substructure mythique de la Chanson de Roland», in Georges Dumézil. Cahier pour un Temps, Paris, Pandora, 1981, p. 257-70.

22 J. Grisward, « L’or corrupteur et le soleil arrêté ou la substructure mythique de la Chanson de Roland», p. 267.

23 We suggest reading of : R. Needham, « Polythetic Classification : Convergence and Consequences», Man n.s. 10 (1975), p. 349-369 ; R. Needham, Casi esemplari, Milano, Medusa, 2002 ; R. Needham, Caratteri primordiali, Milano, Medusa, 2006. Further readings : M. Bonafin, Guerrieri al simposio. Il Voyage de Charlemagne e la tradizione dei vanti, Alessandria, Edizioni dell’Orso, 2010 ; A. Ghidoni, « Il motivo dei vanti tra radici storiche e archetipi : alcune riflessioni di metodo sulle classificazioni politetiche», in Somiglianze di famiglia : tipologie e classificazioni fra scienza e letteratura, edited by M. Bonafin, Alessandria, Edizioni dell’Orso, 2011, p. 155-180.

24 « By this account of the matter, the first consequence of the adoption of polythetic classification in social anthropology is that comparative studies, whether morphological or functional or statistical, are rendered more daunting and perhaps even unfeasible» (R. Needham, « Polythetic Classification : Convergence and Consequences», p. 358).

25 Ibid., p. 360.

26 Ibid., p. 360.

27 Ibid., p. 362.

28 Ibid., p. 365 ; Needham is quoting an aphorism (not published) by Evans-Pritchard.

29 Preparatory to the comprehension of the following paragraphs is the knowledge of the contents of any individual tale. So, the reader is invited to consider the synopses scattered in the Archéologie de l’épopée médiévale.

30 J. Grisward, Archéologie de l’épopée médiévale, p. 81.

31 R. Needham, « Polythetic Classification : Convergence and Consequences», p. 361.

32 J. Grisward, Archéologie de l’épopée médiévale, p. 123-124.

33 The individual witness of a pattern is therefore a particular setting of predetermined elements, a selection and, if we talk about authorial texts, a choice. Interesting in this regard the definition of object comparable by Marcel Detienne : « Les comparables que nous mettions en place par intuitions successives, ce sont donc ces orientations, ces relations en chaîne, ces choix. Des choix parmi des possibles. Quand une société (non pas globale, mais localisée en sa ou ses microconfigurations retenue) adopte un élément de pensée, elle fait une choix parmi d’autres qu’elle aurait pu faire. Le travail de l’analyste [...] consiste à reconnaître les contraintes des configurations mises en perspective. Il s’agit de comprendre comme un microsystème de pensée est poussé à organiser ses éléments constitutif en regard des entrées mises à l’affiche précédemment [...]. Dans ce comparatisme-ci, et constructif on le voit, les ʽcomparablesʼ, ce ne sont pas des types pour établir une typologie [...]. Les comparables, ce sont les plaques d’enchaînements décidée par une choix, une choix initial», they are « des assemblages réalisés à partir de choix logiques, voisins mais différents» (M. Detienne, Comparer l’incomparable, Paris, Éd. du Seuil, 2000, p. 54).

34 The pattern of the enfances, in its entirety, provides « l’exhérédation du futur souverain, son exil dans une terre étrangère, ses exploits auprès d’un prince de ce pays [...], la conquête amoureuse qu’il fait, à cause de sa vaillance, d’une belle étrangère, le plus souvent la fille de son nouveau seigneur, la fuite à laquelle le contraignent de nouvelles trahisons, les nouveaux exploits puis le retour triomphal dans son pays, suivi ou non du mariage avec la belle princesse» (F. Suard, Guide de la chanson de geste et de sa postérité littéraire (XIe-XVe siècle), Paris, Champion, 2011, p. 107). Of course, in the case of the Narbonnais the scheme is without its amorous implications, even if you can see it through the story of Guillaume’s brother Beuve that marries the daughter of Yon de Bordeaux.

35 If actually this model emerges in Saussure’s notes, is matter of discussion. D’Arco Silvio Avalle in several papers (for example in : « Dai sistemi di segni alle nebulose di elementi», Strumenti critici, 19 (1972), p. 229-242 ; Ferdinand de Saussure tra strutturalismo e semiologia, Bologna, Mulino, 1995) supported the extraordinary modernity of this Saussurian reflection not published, while Aldo Prosdocimi (« Sul Saussure delle leggende germaniche», Cahiers Ferdinand de Saussure, 37 (1983), p. 35-106) is much more cautious and believes that the thought of Saussure does not deviate too much from the historicism of his time and from the orthodoxy of his other writings. Although some of the objections by Prosdocimi are reasonable (for example, the fact that a historicist attitude is prevalent in these notes), I consider the interpretation by Avalle functional to my speech.

36 For the notes on Germanic legends, I follow the text of the anthology edited by A. Marinetti e M. Meli : F. de Saussure, Le leggende germaniche, Padova, Zielo-Este, 1986.

37 « 1. Dans une légende donnée, et sans autre moyen de contrôle, il est faux de croire qu’il soit plus ou moins possible de savoir quel sont les traits essentiels, quels sont les traits qui relient la légende présente et visible à la légende inconnue antérieure. Cela est une chose impossible. 2. Dans une légende donnée, et avec des moyens de comparaison provenant de versions différents, il est encore à affirmer que nous ne pouvons nullement juger de la valeur comparative des éléments, des incidents communs, des mots placés dans tel ou tel contexte qui en change le sens. Je n’ai aucune foi non-seulement dans le résultat concret qu’on prétend tirer de la comparaison de deux ou trois légendes, mais même dans le résultat de la comparaison des versions d’une seule légende, lorsqu’on est abandonné à la combinaison interne. Dilemme. Légende non en rapport avec fait historique, en ce cas efforts inutiles, méthode condamnable en elle-même, par impossibilité de savoir quel détail prévaut. Ou bien en rapport avec un fait historique. Dans ce cas, – que nous ne posons point comme attendu, mais comme inattendu –, il y aurait en effet une possibilité, non prévue, de suivre la légende. Il y aurait un crochet extérieur où pendre la légende».

38 Cf. Note 3958/8 p. 21v : « Comme on le voit, au fond l’incapacité à maintenir une identité certaine ne doit pas être mise sur le compte des effets du Temps – c’est là l’erreur remarquable de ceux qui s’occupent des signes – mais * est déposée d’avance dans la constitution même de l’être que l’on choye et observe comme un organisme, alors qu’il n’est que le fantôme obtenu par la combinaison fuyante de deux ou trois idées. Tout [...] est une affaire de définition. Loin de partir de cette unité qui n’existe pas à nul moment, on devrait se rendre compte qu’elle est la formule que nous donnons d’un état momentané d’assemblage, – les éléments seuls existant. Ainsi, Dietrich pris dans son essence vraie n’est pas un personnage historique ou anhistorique ; il est purement la combinaison de trois ou quatre traits qui peuvent se dissocier à tout moment, entraînant la dissolution de l’unité tout entière. * Pas de création plus ou moins fragile : mais une création radicalement dénuée de principe d’unité ; c’est seule la durée relative de certains traits qui donnent illusion là-dessus, et c’est la leçon de tous le jours pour qui étudie, de voir que l’association – que nous chérissons parfois – n’est que une bulle de savon, n’est pas même une bulle de savon, laquelle possède du moins [p. 43] son unité physique et mathématique et non accidentelle et indigne de [...] d’une façon méthodique et non accidentelle».

39 On the illusions of literary scholar following the vicissitudes of narrative sign, Saussure comes back in other notes (3958/4 p. 63v-64r) : « Les symboles ne sont jamais, comme tout [sic] espèce de signe, que le résultat d’une évolution qui a créé un rapport involontaire des choses : ils ne s’inventent, ni ne s’imposent sur le coup. Est admissible un symbole qui s’explique comme n’ayant pas été d’abord un symbole. Par exemple, et en restant sur le genre de sol naturel aux légendes, si dans une bataille [...] alors en effet la geste pourra passer pour symbolique à la fin, – et encore sera-t-il faux de l’appeler symbolique, vu que l’interprétation symbolique n’est que chez le critique qui voit la succession des versions et des expressions de l’événement. [...] et il n’y a là par conséquent aucune symbole à la fin, comme il n’y en avait aucun au commencement non plus. On peut parler de réduction de proportion ou d’amplification des événements à la suite d’un temps écoulé, c’est-à-dire d’un nombre indéfini de récitations transformées, mais non de symbolisation à un moment quelconque».

40 D. S. Avalle, Ferdinand de Saussure tra strutturalismo e semiologia, p. 106.

41 Ibid., p. 107.

42 See above : Note 3959/4 p.1.

43 Note 3959/10 p. 18 : « Ce qui fait la noblesse de la légende comme de la langue, c’est que condamnées l’une et l’autre à ne se servir que d’éléments apportés devant elles et d’un sens quelconque, elles les réunissent et en tirent continuellement un sens nouveau. Une loi grave précise, qu’on ferait bien de méditer avant de conclure à la fausseté de cette conception de la légende : nous ne voyons nulle part fleurir une chose qui ne soit la combinaison d’éléments inertes, et nous voyons nulle part que la matière soit autre chose que l’aliment continuel que la pensée digère, ordonne, commande, mais sans pouvoir s’en passer. Imaginer qu’une légende commence par un sens a eu depuis sa première origine le sens qu’elle a, ou plutôt imaginer qu’elle n’a pas pu avoir un sens absolument quelconque, est une opération qui me dépasse. Elle semble réellement supposer qu’il ne s’est jamais transmis d’éléments matériels sur cette légende à travers les siècles ; car étant donné cinq ou six éléments matériels, le sens changera dans l’espace de quelques minutes si je les donne à combiner à cinq ou six personnes travaillant séparément».

44 D. S. Avalle, Ferdinand de Saussure tra strutturalismo e semiologia, p. 103-104.

45 Note 3959/7 p. 35v : « Un des éléments de destruction semés d’avance pour l’histoire dans ‹le terrain particulier qu’est› la légende, c’est que les détails, avec le temps, s’arrangeront presque tout seuls en une unité dramatique à laquelle seront subordonnés beaucoup de choses, et sacrifiées beauc(ou)p d’autres mais que tel soit depuis le commencement, depuis le premier jet, l’arrangement épique des événements, c’est ce que je ne croirai jamais». Something similar can be read in Note 3958/7, p. 34v-35r : « On peut dire que précisément la donnée d’ensemble est sujette à se modifier par le simple travail d’où sort la véritable épopée et qui introduit un sens et une unité dans les événements».

46 The article published here is an excerpt from my doctoral thesis discussed at the Università di Macerata (Italy) in March 2013 (title : Gormund et Isembart e le origini delle chansons de geste (con una nuova edizione del testo) ; advisor : prof. Massimo Bonafin). My dissertation consists of two sections, one more specifically philological (going to be published in : Gormund et Isembart, Alessandria, Edizioni dell’Orso), the other theoretical, which includes this excerpt.

Haut de page

Pour citer cet article

Référence électronique

Andrea Ghidoni, « The Origins of the Narrative Structures in the Chansons de Geste : Methodological Notes on Joël Grisward’s comparativism », Perspectives médiévales [En ligne], 35 | 2014, mis en ligne le 01 janvier 2014, consulté le 27 mai 2017. URL : http://peme.revues.org/4321 ; DOI : 10.4000/peme.4321

Haut de page

Auteur

Andrea Ghidoni

Università degli Studi di Macerata - Italie

Haut de page

Droits d’auteur

© Perspectives médiévales

Haut de page