Keywords

1 Introduction

In this chapter, I attempt to test the applicability of Lars Elleström’s model of intermedial relations (2020) to analyses of the device of narrative metalepsis in a number of different media products stemming from two different media types: literature and film. At the same time, the chapter has the more ambitious aim to illustrate certain aspects of the model itself using the concept of metalepsis. The overarching aim is to highlight the transmediality of the device of metalepsis, that is, its potential to “be employed in various media” (Rajewsky 2002: 13). As such, the chapter is an attempt to break with what Lars Elleström sees as one problematic tendency in today’s intermedial studies, that is, that “[m]edia in general are studied through concepts developed for language analysis” (Elleström 2020: 6). Metalepsis is a good example in such an endeavour, since its definition was coined within the frame of a linguistic medium, namely literature. This is, I argue, no reason to persist in studying metalepsis from a language-based perspective, which is the tendency even in the numerous cases of approaching metalepsis in other media types than linguistic ones. It is, I argue, not sufficient to acknowledge the transmedial potential of metalepsis. The device should also be seen completely beyond its initial connection to the language-based context, which is best done by using Elleström’s model of intermedial relations.

The initial definition of metalepsis is, as mentioned, clearly linked to language, especially writtenlanguage, in the form of the qualified medium of literature. Indeed, French narratologist Gérard Genette gave the first definition of metalepsis as “any intrusion by the extradiegetic narrator or narratee into the diegetic universe (or by the diegetic characters into a metadiegetic universe, etc.), or the inverse” (Genette 1980 [1972]: 234–235). The all-importance of the narrator is something that all subsequent studies have had a hard time avoiding, even in the case of media types where the existence of a narrator is not necessary, or even impossible. Indeed, theorists explain metalepses in narratorless media, such as painting or music by analogies to literature, finding that certain devices are ways to imitate a narrativevoice where such a thing cannot exist. Even if this is true in many individual cases, it is my contention that it should not be transformed into a rule. Indeed, narrative levels can be both constructed and transgressed without the use of a narrativevoice, as I will try to show in the chapter.

Interestingly enough, the transmediality of the concept of metalepsis was alluded to already in Genette’s first theorizations in 1972. First, he explicitly illustrated metalepsis with examples from other media types than literature—such as theatre, paintings (Genette 1980 [1972]: 235) and cinema (1980 [1972]: 237). Then, implicitly, when he used the term ‘world’ in his other definition of the concept as a transgression of “the sacred frontier between two worlds, the world in which one tells, the world of which one tells” (1980 [1972]: 236), he suggested the extension of the concept of narration beyond its linguistic frame and its inclusion into the broader spectrum of representation. A decade after the first definition, Genette emphasized the importance of representation for metalepsis even more strongly, avoiding the reference to a narrator, and defining it as the “deliberate transgression of the threshold of embedding” (Genette 1988 [1983]: 88). As Jean-Marie Schaeffer pointed out, this redefinition permitted the transfer of the concept to the field of representation in general (2005: 325, 327).

This expansion of metalepsis beyond only language-based media to the field of representation in general should have led to a liberation of the concept of its linguistic straitjacket. Unfortunately, this has not really happened. As in Genette’s early definition, where the act of telling is still central even when a certain move towards representation is made, the linguistic basis persisted both in definitions and in analyses. This is, for example, very clearly the case for Sabine Lang’s detailed typology of the so-called paradoxical narrative devices, which includes metalepsis as the most conspicuous case, where the main criterion is the difference between story and discourse (discurso and historia in the Spanish original, Lang 2006: 32). This is, to an even higher degree, the case for Sonja Klimek’s approach, when she claims that narrative levels only concern “written and not visual fiction” (Klimek and Kukkonen 2011: 25). This is also the case, admittedly to a lesser degree, even in a number of studies, from Brian McHale’s Postmodernist Fiction (1987) to more recent ones such as Werner Wolf’s “Metalepsis as a transgeneric and transmedial phenomenon” (2005), the anthology Metalepsis in Popular Culture edited by Sonja Klimek and Karin Kukkonen (2011), Jan Alber’s and Alice Bell’s “Ontological metalepsis and unnatural narratology” (2012) or Jeff Thoss’s When Storyworlds Collide: Metalepsis in Popular Fiction, Film and Comics (2015). These studies all have in common the aim to expand the application field for metalepsis to representational media in general. However, despite the opening to other media, the studies are still dependent on the concept’s initial connection to language, and have the tendency, criticized by Elleström, to compare only two media types, out of which literature is always the basis (2020: 5–6).

It might seem that even the current study does the same thing. However, even if only two media types are studied, the point is to consider the media beyond their essence, and concentrate on their modalities. This is, I argue, a more productive way of approaching transmediality than studying many different media, because no matter how many media one might study, there would always remain cases that could contradict the results. Looking at a medium’s basic ingredients, which is the aim of Elleström’s model of intermedial relations, is a way out from this dilemma. It also helps in expanding metalepsis beyond language to representation in general, since representation is the core of what Elleström calls the semiotic modality.

According to Elleström’s model, each medium consists of four elementary modalities: the material, the sensorial, the spatiotemporal and the semiotic. These modalities form a transmedial basis on which more complex features can be constructed. These features, that Elleström first called “compound media characteristics” (2014) and now changed to “represented media characteristics” or simply “media characteristics” (2020: 80), cannot be transmedial in the same way as the elementary modalities, since indeed they are more specific, and also more or less closely linked to a specific medium. This can be explained in a simple way by the device of the subjective camera, which is a media characteristic linked to a medium where a camera is necessary. Still, at a level below, the level of elementary modality, a subjective camera has to do with the sensorial modality, more exactly with the visual mode, which belongs to the transmedial basis. It is exactly by acknowledging such properties of lower complexity that one can discover certain similarities between devices from different media which seem completely incompatible at a first look.

Paradoxical narrative devices such as metalepsis are examples of represented or compound media characteristics. As such, they could hardly be considered as fully transmedial. Not even narration in general can be considered to be a trait that can be found in all media, despite the so called narrative turn, according to which everything is narrative (see, for instance, Richardson 2000 and Hyvärinen 2006). But this is not a good reason to reduce the applicability of the paradoxical narrative devices to only one qualified medium: literature. The main reason is the fact that narration can hardly be seen as only a literary trait. The other reason is, as I already mentioned, that metalepsis is linked to representation in general rather than to narration through language only.

Still, as already mentioned, there is a tendency even today to concentrate on two media-specific aspects of metalepsis, which both have reductive effects. First of all, there is the language-based aspect, especially highlighted in the insistence on the need of a narratorial voice for the metalepsis to occur. Secondly, there is the fictional aspect, since several theorists find that metalepsis can only occur in fiction. Some even find metalepsis to be “a popular candidate for a fiction specific category” (Bareis 2008: 164).

I argue that both these aspects can be approached more productively by using Elleström’s model of intermedial relations (2020). Indeed, both narratorial voice and fictionality are media characteristics that can be studied in relation to the basic modalities of the media types in which metalepsis occurs, rather than being considered as necessary ingredients for the concept of metalepsis. I will approach both these aspects below, on the basis of a number of examples from two media types.

2 Metalepsis in Literature

I start with an example of metalepsis in the qualified medium where its existence is undisputed: literature. The example is, moreover, one of the best known cases of metalepsis, and has been often analysed as such by theorists as Dorrit Cohn (2005: 123), Sabine Schlickers (2005: 161), Sabine Lang (2006) and Bénédicte Vauthier (1999). It comes from Spanish author Miguel de Unamuno’s novel Mist (Niebla in original), written in 1907 but published only in 1914 (2013 [1914]).

At first view, the novel seems to be realistic, even though there is a certain play with the narrative instance at the beginning. Indeed, the first pages are narrated by a certain Victor Goti, in the first person. However, as it appears after these pages, this Victor Goti is not the main narrator. Goti only supposedly writes a prologue to the novel we are reading, and contradicts some of the facts therein, pretending to know the characters better than Unamuno. Unamuno himself responds to this prologue, in a post-prologue, defending himself against Victor Goti’s accusations, before letting a heterodiegetic narrator (who will later prove to be himself) take over the narration. This narrator tells the story of a certain Augusto Pérez, a rich young man without a clear goal in life. After his mother’s death, Augusto falls in love, or rather thinks that he falls in love, with Eugenia, a young pianist of lower social status. Eugenia, however, does not love Augusto, but runs away from him with a young man of even lower social status. Having even lost this last goal in life, Augusto decides to commit suicide, but not before going to Salamanca to discuss the matter with a certain Miguel de Unamuno, who had written an essay on the question of suicide from a philosophical point of view.

It is this dialogue with Unamuno that constitutes the famous metalepsis in this novel. Indeed, during their talk, which takes place in chapter 31 out of 33, Augusto finds out that he is only a character of fiction created by his interlocutor. Unamuno reveals to Augusto that he is destined to die, not to commit suicide, as he was planning. Augusto is shocked at first, but then starts to contradict Unamuno, questioning whether he too exists in the real world. Something that we, as readers, could ponder upon too, since the Unamuno of the story is hardly the same Unamuno as the one in the real world. Augusto leaves Salamanca after the dialogue in a very hasty manner, arrives at home and discusses his fictional status with his servants, who do not understand the matter. Eventually, Augusto dies, but it is not stated clearly if it is through suicide by eating compulsively or if it is by Unamuno’s narratorial decision.

In chapter 31, the narrator abandons his heterodiegetic status and becomes homodiegetic, meaning that he narrates at the first person, as a character in the story. This is how this transition is done, at the beginning of the chapter: “The storm in Augusto’s soul ended in a terrible calm. He had resolved to kill himself, to put an end to that self which had been the cause of all his misery. But before carrying out his plan it occurred to him, like a drowning sailor who grasps at a near plank, to come and talk it over with me, the author of this whole story” (Unamuno 2013 [1914]: 147, my emphasis). Or rather, I would argue, the narrator of the story, since the character of Unamuno is fictional too, as I already suggested.

Admittedly, the metalepsis here respects Genette’s definition, since the frontier between the world in which one tells and the world of which one tells is clearly broken. However, it is interesting to notice that it is not until this chapter that the narrator appears as a character in the storyworld. Indeed, not only does he use the first person when speaking, but he is represented as a character, who physically meets Augusto. It is what theoreticians, such as Marie-Laure Ryan (2006: 207) and Monika Fludernik (2005: 79), classify as an ontological metalepsis, the most conspicuous and shocking example of the different classes of metalepsis. What I consider noteworthy, however, is that it seems that even in the case of such a typically literary device, there is a need for a kind of representation that goes beyond the strictly literary context. What I mean is that the narrative levels are less conspicuously transgressed through the literary device of the voice of the narrator. They are more clearly transgressed by a represented physical meeting between two characters that are not supposed to exist in the same world. It is true that the meeting becomes all the more shocking thanks to the narrator’s using the first person, and referring to the story as his own creation, but such a meeting could be described without the use of a first person narrator, just by letting the different worlds collide. The only requirement is that the worlds should be at different, incompatible narrative levels.

It is probably the requirement of ‘narrative levels’ that links metalepsis so closely to language. But, as I already mentioned, narration can be realized through other means than just language. And indeed, Genette’s definition of the two worlds on the basis of ‘telling’ seems to be an unfortunate slip of the tongue. What should be at stake are the levels of representation, and metalepsis should be the violation of the frontier between different levels of representation. Neither the definition based on the notion of a narrator (I recall: “intrusion by the extradiegetic narrator or narratee into the diegetic universe”) is particularly relevant, since the violation can be achieved by other means than the narrator. Just to mention one example, an object that moves from one narrative level to another narrative level would be an undoubted case of metalepsis too.

What is particularly interesting is that the language-based aspect of metalepsis seems to be hard to get rid of even in other media. One example is the cinematic adaptation of Unamuno’s novel, made by Spanish director Fernando Méndez-Leite in 1975. The meeting between Unamuno and Augusto is represented in the film too, but Unamuno is heard as a narrator in voice-over, addressing the viewers and explaining that the meeting is between him, as author, and a character created by him. The words are very close to the text of the novel. However, two additional devices typical of the cinematographic medium are used. The first one is the editing technique, used here in such a way as to give the viewer the impression that Augusto appears and then disappears from Unamuno’s room out of nowhere, almost in a supernatural way. It is as if Unamuno created him right there and then in front of us, the viewers. This is a very good example of how other means than strictly linguistic, in this case visual, can be used in order to create a metaleptical effect. The question is, however, whether the effect would have been metaleptical if the viewer had not been aware of the different narrative levels established previously, in a language-based way, by the voice-over. I argue that even if it is not done in the film, such an establishment could have been done by other means, and choosing to repeat the narrator’s words from the novel is probably the director’s choice to show his debt to the original media product, by incorporating the part that can be incorporated in the new medium. Admittedly, to use Elleström’s words, the text has been transmediated (2014), from written to oral, but it is still a reference to the original work. The second cinematographic device is the use of sound effects. There is a certain change in the quality of the sound directly before Augusto’s appearance and until his disappearance, clearly perceptible for the viewer, who gets the impression that the narrativevoice changes its status from being intradiegetic to being extradiegetic—that is, addressing the viewer directly, as a voice-over. It is still the same voice, but it is as if it came from another place and sounding as if reading out loud from a novel. Here, I argue, the change is solely done by devices belonging to the mode of sound. A narrator’s voice is imitated in its aural characteristic, not claimed by words. It is something that cannot be done in a novel, where the sound as mode is simply absent.

This whole discussion around the ways the metaleptical encounter between the narrator and the characters created by him is represented can seem rather academic. As a matter of fact, even if we can conclude that there is only one narrative level in the scene, since the narrator is actually just a represented character, this does not mean that there is not a metalepsis there. Actually, the metalepsis had already occurred before the encounter, when the author fictionalized himself, moving from the extradiegetic level to the intradiegetic level. This is shown more clearly by the cinematographic media product, where it is much more important to expose visually the moment of this diegetic change so that everything becomes more comprehensible.

Beyond this rather academic analysis, the question is why metalepsis is used. The question is relevant at a general level, since metalepsis is a device which points to the artificiality of a media product, so it should rather be avoided in films aiming at realism. In the specific case of Niebla, one answer is suggested in the novel itself: it could be that becoming a character in a literary work is a way to immortalize oneself. And indeed, it is a strategy that seems to work. A little more than a century after its publication, Unamuno is, in a certain way, paradoxically present in the midst of his readers. But is that a result of his transformation into a character in his own book? Is it not rather the result of the success of the books? Beyond these questions, there is something particularly interesting in the way in which reality and fiction meet in such a paradoxical way. What happens is that a confusion is created at different levels. For instance, is it the author who moves towards the inferior narrative level—metalepsis—or is it the character, Augusto Pérez, who moves towards the superior level—antimetalepsis according to Genette’s definition? Indeed, Augusto is the one who decides to visit Unamuno. It is true that this decision is ultimately orchestrated by Unamuno, as it is underlined by his lack of surprise and enigmatic smile when Augusto appears in his office (2013 [1914]: 147). However, the development of the situation eventually surprises Unamuno, first when Augusto tells him that “you are nothing more than a pretext for bringing my history into the world” (2013 [1914]: 149), then when he acknowledges his reaction: “These sallies of Augusto were beginning to make me uneasy” (2013 [1914]: 151). The gradual loss of authority over his own creation complicates things, putting into question whether the details are planned by the author or not.

Before this dialogue, the narrator’s authority had not been put into question at all, with the exception of the already mentioned prologue. Even then, Victor’s words had not really questioned Unamuno’s authority in general; they only questioned his credibility concerning the version of the facts. On the contrary, the prologue establishes the extradiegetic status of the narrator, Unamuno, which is reinforced in chapter 31, when the narrator refers to himself as “me, the author of this whole story” (Unamuno 2013 [1914]: 147). Thanks to this clear distinction between the different narrative levels, the metalepsis is very effective in Niebla. The fusion of the extradiegetic and the intradiegetic levels in chapter 31 has indeed a strong effect on a reader who until then was certain of being able to keep the levels apart. This gives also the opportunity to double entendres in the novel, such as the one when Augusto tells his servant: “I come from another world, Liduvina, and I am going to another world” (2013 [1914]: 155). This other world could, indeed, be the world of the dead, but also the world of the author, the extradiegetic, real world. This double entendre, besides being a game, can be a way to underline the difficulty to distinguish between fiction and life, between dream and reality.

In conclusion, one of the main elements of Niebla is the fusion between different narrative levels, especially between the extradiegetic level of the author and the diegetic level of the characters. Moreover, Niebla anticipates what Genette would define as one of the possible effects of metalepsis: the existential doubt. Indeed, Augusto comments upon the meeting with his creator by making an analogy between the act of writing and God’s act of creation. The result could be that the real Unamuno could in his turn be a fictional creature, the product of God’s dreams. This is actually suggested by Augusto: “Very well, then, my lord creator Don Miguel”, he says, using an ambiguous “lord creator” that could refer to God, “you too are to die, you too! And you will return to that nothing of which you came! God will stop dreaming you!” (2013 [1914]: 154). Genette’s words about the existential doubt that can be created by metalepsis are close to Augusto’s: “The most troubling thing about metalepsis indeed lies in this unacceptable and insistent hypothesis that the extradiegetic is perhaps always diegetic and that the narrator and his narratees—you and I—perhaps belong to some narrative” (Genette 1980 [1972]: 236). In the following parts, I will investigate whether metalepsis can be achieved in similar ways in a media type where language is less central, that is, film, to see whether the same effects can be created using other media characteristics.

3 Metalepsis in Film

One of the reasons why I choose to analyse metalepsis in film is that film consists of complex modes which can be combined in creative ways. The material modality of film is, for instance, flexible enough, as we shall see, to incorporate other media types, or to give the impression of incorporating them. For instance, since the basic medium of film consists of moving images, it can easily incorporate still images. And since film has a sound track, it can easily incorporate any media type that is based on sound, such as music, radio, and so forth. Thus, I can analyse the effects that other modes, which are less important in literature, can have on metalepsis.

Another important reason why film is a suitable study object in this context is that it has a very special relationship to literature. This is not only because film, as a qualified medium, has a strong tradition of adapting literary works (which is a form of transmediation, to use Elleström’s words [2014]), but also because film, as I just mentioned, can incorporate writtentext. For instance, since the visual mode is of great importance in film, a writtentext can be shown on the screen, (which is a form of media representation, according to Elleström’s model [2014]). Moreover, film can imitate more complex literary media characteristics in very interesting ways. The use of a narrator in voice-over, as we saw in the adaptation of Unamuno’s novel Niebla, is just one such example, even if not the most interesting one, since it is quite obvious. In intermedial theory, this is what Werner Wolf would call “intermedial imitation” and Irina Rajewsky “intermedial reference”. It is, as Wolf puts it, when “the signifiers of the work and/or its structure are affected by the non-dominant medium, since they appear to imitate its quality or structure” (2002: 25), or, as Rajewsky puts it, when “the given media-product thematizes, evokes, or imitates elements or structures of another, conventionally distinct medium” (2005: 53).

The adaptation of Niebla is illustrative as an example of the difficulty of completely breaking with the literary model, in the case of metalepsis. As I already mentioned, I argue that a narrator’s voice-over was not actually needed in order to prepare the field for the metalepsis to occur and to be understood as such by the public. Meanwhile, it is understandable that the director imitates certain details from a specific source media product (the specific novel by Miguel de Unamuno) when adapting it to another media type. It is also understandable that metalepsis, being a very complex media characteristic, cannot be imitated without imitating its smaller units, such as the narrator’s voice. What is less understandable meanwhile is that many films imitate a narrator’s voice as a general feature (that is not from a specific novel) in order to achieve metalepsis. I will present here two such cases, Stranger than Fiction (2006) and Le tableau (The Painting) (2011), but I will also go beyond the narrator’s voice and analyse whether other filmic media characteristics are used when staging the metalepses. By doing so, I aim at showing that metalepsis is not necessarily dependent of a narrator’s voice. The third case is an example of a film, Woman at War (2018), where the narrator’s voice is not used at all in the creation of metalepsis.

3.1 Metalepsis Through Narrative Voice in Film

The first film I choose to analyse is Marc Forster’sStranger than Fiction (2006). As the title suggested, the play with the relationship between fiction and reality is the main theme of this film, and is achieved through a metalepsis which actually reminds of the one in Niebla. Indeed, the main character, Harold Crick, played by Will Ferrell, hears the voice of a narrator, played by Emma Thompson, who seems to create the events he is experiencing, and that the viewers are viewing on the screen. The viewers hear the narrator’s voice-over as well, and have no reason to consider this strange at all. Indeed, this is a case of an intermedial imitation of a narrator’s voice, common for many films. The first time the viewer reacts is when the character himself reacts, very early in the film, when starting to hear the voice, which describes him and his ongoing actions.

Interestingly enough, what the metalepsis actually does is to draw the attention to the artificiality of the device of a narratorial voice-over. What in literature is unavoidable, that is, the use of a narrator narrating the events, is unnatural in film, and has become an accepted device only because of its frequent use. The existence of an extradiegetic voice at the same level as the storyworld is actually a logical impossibility, since the two worlds should not be able to coexist. The ability to keep the worlds apart, even if they are represented as coexisting, is only achieved through training, and a viewer who experiences such a case for the first time would probably react with the same surprise as does Harold.

The similarities with Niebla do not finish at this point. Indeed, not only does the character have a physical contact with his creator, but he also hears that the creator plans his death, as Unamuno did for Augusto. Harold reacts at first as does Augusto, trying to change the course of the events. For instance, he spends all his time at home trying to control his destiny by doing nothing and thus not taking any risks, but even so, he is almost killed by a wrecking crew which demolishes his apartment, because of having mistaken the building for an abandoned one. Having thus realized that he cannot control the plot, Harold accepts his fate and tries to enjoy the time he has left. One day, by coincidence, Harold hears the voice he had in his head when watching a television interview with author Karen Eiffel. Indeed, the voice is the same as the narrator’s, but as a character. When confronted by Harold, who did his best to find her, Karen Eiffel is not aware of what she had caused. She is just a character in the story, but in a strange loop, it is the story she writes. Therefore, she is shocked too.

It is certainly not by chance that a British actor has been chosen for the role of Karen Eiffel. First of all, the British accent of the narrator’s voice contrasts in an obvious way with the American accents of the other characters, who are supposed to live in Chicago. This contrast contributes to creating the impression that the voice is heard at another narrative level, the extradiegetic one, that is, the level that should not be heard by the characters from the diegesis. The fact that the voice had been disembodied until the author first appeared on the TV set also contributed to this impression, as did the sound effects. Besides, also by effect of contrast, British accent can give an elevated impression, which can be considered as a reference to literature. All these devices are filmic media characteristics, relying on the mode of sound, and could thus not be used in writtenliterature, where the mode of sound is absent. It is therefore an interesting example of how intermedial imitation can be used in innovative ways, thus casting new light on concepts that previously were considered to belong to literature.

When Karen Eiffel, the author, appears as a represented character in the film, her accent is still British, and her voice is the same, but it loses its distinctiveness. This is, I argue, the result of the viewers being used to the filmic conventions. A voice belonging to a character who is visually represented is less distinctive than a disembodied voice-over. Even if the voice-over is known by the viewers to belong to one character, not showing the character and using certain sound effects can easily give the impression that the utterance comes from another level at the moment it is made. That is why the first sign of metalepsis, when Harold actually hears this strange voice, is so shocking for both Harold and the viewers.

The second example of metalepsis in film that I choose to analyse is Jean-François Laguionie’sLe tableau (The Painting) from 2011. It is an animated film which can be considered as a metalepsis from the start until the end, since its first scene is a metaleptic entrance into the world of film, where the viewer stays until the last scene. Besides, the film is a very good example of different types of metalepses, according to classifications established by theoreticians in the field, showing how film as basic medium can contribute to a better understanding of these classes and of metalepsis in general.

The most important distinction the film addresses is the one between vertical metalepsis and horizontal metalepsis. This is a distinction first made by Klaus Meyer-Minnemann and Sabine Schlickers (2005) and by Sabine Lang (2006). Their definition of vertical metalepsis as a transgression of the border between worlds situated at different narrative levels (the world of the narrator and the world of the narrated events) is close to Genette’s initial definition, that is, “any intrusion by the extradiegetic narrator or narratee into the diegetic universe” (Genette 1988 [1983]: 234–235). Horizontal metalepsis, defined as the transgression of the borders between worlds situated at the same narrative level (Meyer-Minnemann and Schlickers 2010: 140), is an expansion, rather than a contradiction, of Genette’s definition. Still, the concept of the horizontal metalepsis has raised certain controversy among theoreticians. Sonja Klimek considers, for instance, that we should “respect Genette’s initial definition” and restrict the use of metalepsis only to cases where there is a vertical infringement, that is, between “the world of the creator” and “the world that is created within the artefact” (Klimek 2011: 26). My analysis of the metalepses in Le tableau will try to show how cinematic medium can help to reconcile these views.

When I claimed above that the whole film is a metalepsis, it was because it is framed by two vertical metalepses, which form the entrance into and the exit from the world of the film. The whole story thus becomes the result of the initial metalepsis, and its fictional status is regularly reminded by metaleptical comments or by metaleptical scenes. The first scene is not only worth studying as an interesting case of a vertical metalepsis, but also because it occurs both visually and verbally. Indeed, while the camera is zooming in, giving the viewer the impression that the frame is left behind and that the limit established by the canvas is transgressed, a character in the painting, a girl called Lola, comes to life and addresses the viewers with the words: “Voilà! Vous venez de pénétrer dans le tableau” (“Here you are! You just entered the painting”). Such a comment could be interpreted as a metalepsis of the discursive kind according to Dorrit Cohn (“métalepse discursive” in the French original), who found it more “harmless” (“inoffensive” in the French original) than the “diegetic metalepsis” (“métalepse de l’histoire” in the French original) (Cohn 2005: 122). John Pier would consider this example a “minimal metalepsis” (“métalepse minimale” in the French original) (Pier 2005: 249), considering that the transgression of the frontiers is only suggested in such a case, not real. Monika Fludernik would classify it as “rhetorical metalepsis” (2005: 79–81) as opposed to the “ontological metalepsis”, according to a classification made previously by Marie-Laure Ryan. According to Ryan, ontological metalepsis “opens a passage between levels that result in their interpenetration, or mutual contamination,” while rhetorical metalepsis only “opens a small window that allows a quick glance across levels, but the window closes after a few sentences, and the operation ends up reasserting the existence of the boundaries” (Ryan 2006: 207). Indeed, the comment is quite close to the example these theorists use to illustrate these kinds of harmless metalepses, namely an extract from Honoré de Balzac’s Les illusions perdues where the narrator addresses the narratee with the words: “While the venerable churchman climbs the ramps of the Angouleme, it is not useless to explain the network of interests into which he was going to set foot” (quoted, for instance, by Fludernik [2005]). In the French original: “Pendant que le vénérable ecclésiastique monte les rampes d’Angoulême, il n’est pas inutile d’expliquer le lacis d’intérêts dans lequel il allait mettre le pied.”). An even better example, also mentioned by theorists, comes from another well-known French nineteenth-century novel, Alexandre Dumas’s The Count of Monte Cristo:

We will leave Villefort on the road to Paris, travelling—thanks to trebled fees—with all speed, and passing through two or three apartments, enter at the Tuileries the little room with the arched window, so well known as having been the favorite closet of Napoleon and Louis XVIII., and now of Louis Philippe. (Dumas 2019 [1844]: chapter 10)

What complicates things in the film, however, is the visually represented ‘penetration’ of the painting. It could be argued that this is only a metaphor of the act of immersion into the painting, since we as readers or viewers penetrate in some way the world represented by the media product. But this interpretation is a part of the semiotic modality. At a lower level, the spatiotemporal modality, the fact remains that the zooming in of the camera gives the viewer the impression of entering a space, which is only a virtual space, represented on the screen. Indeed, neither the painting nor the space represented by the painting exists in reality. Still, by using visual devices offered by the cinematic medium, the viewer gets the impression of transgressing the two borders: first the border between reality and the world of Lola as narrator, and then the border between that world and the world of the painting. This second border is even more clearly marked by the frame of the painting.

My point is that Lola’s verbal comment is not really needed to give the viewer the impression of the metaleptical transgression. What the comment does is that it reminds the language-based original definition of metalepsis as the transgression of the frontier between “the world in which one tells, the world of which one tells” (Genette 1980 [1972]: 236, see even above). The fact that it is an intrusion into an ontologically different world appears more clearly in a movie, thanks to the visual dimension, but this does not mean that there are no narrative levels involved in this transgression. The world in which one tells corresponds to the world of the camera before it penetrates through the frame into the world of the painting, after which the character of Lola can enter in direct dialogue with the viewer. To be more precise, this example corresponds to Genette’s first definition of metalepsis, “any intrusion by the extradiegetic narrator or narratee into the diegetic universe” (Genette 1980 [1972]: 234–235), since it is the narratee, meaning the recipient of the narration, who is entering the diegetic universe of the painting. Following John Pier’s dichotomy, this would be a descendant metalepsis, of the kind that Pier exemplifies with an extract from George Eliot’s Adam Bede (1859), when the narratee and the narrator together transgress the demarcation line between the narrative levels (Pier 2005: 250).

However, there are even more interesting examples of metalepses in this movie, especially of the kind that could be analysed as horizontal (that is, I recall, metalepses which occur between worlds at the same narrative level). Indeed, after the metaleptical start, the film represents the events in the world of the painting in a traditional way, without several levels or worlds involved, making the viewer forget the initial metaleptical intrusion. All of a sudden, Lola, who has abandoned her ability to communicate with an extradiegetic narratee and has become a character inside the world of the painting, arrives at the frontier of her world, the canvas of the painting, and jumps out of it, followed immediately afterwards by two other characters. They enter an extradiegetic world, an ontologically different world, the world of the creator, the painter, a world which could be the same as the initial world of the narratee, the viewers’ world. This is of course a hierarchically different world, so the transgression is undoubtedly vertical. According to John Pier’s typology, this would be the case of an ascending metalepsis, which he exemplifies with an extract from Mark Twain’s The Adventures of Huckleberry Finn (1883), more precisely when Huck goes up one level and enters Mark Twain’s world (Pier 2005: 252). This transgression is so important in the movie Le tableau that exactly the moment of the jump is used on the DVD cover.

What happens next in the film, though, is very interesting from a theoretical point of view. When Lola jumps into her creator’s world, she lands on the frame of another painting. Shortly after that, the second painting falls down from the wall because of Lola’s weight, and Lola falls through the canvas into the world of this second painting, representing a war. She becomes a character inside that world, able to interact with the soldiers there.

This example of metalepsis could be analysed as an ontological metalepsis, as opposed to the more ‘rhetorical’ or ‘discursive’ cases studied above. Indeed, Lola literally travels between different worlds in a way that could be analysed through the lenses of the possible world theories. According to Alber’s and Bell’s attempt to apply these theories to metalepses (Alber and Bell 2012), such a travel between different ontological worlds would correspond, at least partly, to Daniel Lewis’s concretist philosophical position. According to the Concretists, as opposed to Abstractionists, such as Nicholas Rescher, Saul Kripke, Jaakko Hintikka or Alvin Plantinga, an individual cannot exist simultaneously in the same world. Thus, when Lola penetrates the world of the painter, she disappears from the world of the painting.

What is more important is that this is a horizontal metalepsis, since Lola actually crosses the frontier between two worlds situated at the same level: the worlds of two paintings hanging on the same wall. It is this kind of metalepsis that Sonja Klimek finds incompatible with Genette’s original definition, since the transgression is between worlds, not between narrative levels. But what this movie example shows is that in order to move from one painting to the other, Lola must necessarily pass through her creator’s world, the world ‘in which one tells’.

One could argue that this example cannot be applied to literature, where the movement between two parallel worlds could occur without a necessary transition through a higher narrative level. But does the fact that such a transition is not represented imply that it does not occur? Is it not rather that the transition is implicit, and does not have to be spelled out explicitly? My answer would be that this depends. In some cases, such as when a character from a fictional work appears in a later fictional work, it is probably better to talk about what Marie-Laure Ryan calls “transfictionality”, meaning “the migration of fictional entities across different texts” (Ryan 2013). But in the cases when such a migration is actually represented, when a character is represented when he or she literally leaves the storyworld, the paradoxical dimension is more obvious. However, even if I cannot provide a definitive answer as to the nature of horizontal metalepsis in the context of this article, it has appeared as obvious that a transmedial perspective is necessary in order to better understand this narrative device.

Before going over to the following film, I want to mention succinctly the metalepsis at the end of the film. What happens there is that Lola finally meets her creator, a painter, and talks to him, the way Augusto does with Unamuno. Her comment to him is, “But who painted you?” A question which is relevant even from a theoretical point of view since, indeed, the painter from the film is nothing else than represented, even though he is not an animated, but a live action character. The question also underlines that metalepsis has the same effect in film as in literature, as we saw in Niebla: making one doubt about his or her own existence.

3.2 Metalepsis Without Narrative Voice in Film

The third film to be analysed is also the one in which the metalepses are created without any implication of a narrator. Thus, the language-based features are overshadowed by other features, both visual and aural ones. What is particularly interesting is that the aural features are not related to language, but to music, and the so-called world in which one tells has no relevance, since there is no narrator speaking. Therefore, the film is an illustrative argument against the language-based definitions of metalepsis, and a good example of the transmediality of the concept.

The film in question is an Icelandic-Ukrainian coproduction from 2018 entitled Woman at War and directed by Benedikt Erlingsson. The plot is completely centred on the main character, Halla, a choir conductor and eco-activist, played by Halldóra Geirharðsdóttir, and on her attempts to sabotage an aluminum plant in the Icelandic highlands. From this point of view, the film would fit Jørgen Bruhn’s definition of ecomedia, that is, “representations of environmental issues in any specific media product” (Bruhn 2020: 121). However, at the same time, as Halla sabotages this plant in ingenious but dangerous ways, an application she made years earlier to adopt an orphan child from Ukraine is approved. The film follows the two different lines of plot in parallel, thus showing that illegal activism can very well be compatible with a humanistic act.

Even if the plot is thrilling, the merits of the film lie rather in its cinematic language, as critics have noticed. Just to mention one reaction, Peter Bradshaw, from The Guardian, finds the film “confidently and rather stylishly made” (Bradshaw 2018). However, Bradshaw finds this stylishness a bit exaggerated and unmotivated, especially when it comes to the use of music:

There is also a lot of Icelandic folk music from a tuba, accordion and drums trio, and a rather beautiful singing group, whose performances turn out to be diegetic. That is, the camera pans around to show the musicians themselves, standing weirdly, incongruously in the background of the shot, variously trilling or parping away. It is a comic effect that is a bit distracting, subject to diminishing returns, and which ironises and undermines the action and obstructs your natural tendency to invest emotionally in Halla’s dilemma. (Bradshaw 2018)

It is not my aim to contradict Bradshaw’s opinion, but as a specific feature, the use of music is one of the most interesting details of Woman at War. Besides, it is through music that metalepsis is constructed in this film, and that is done in innovative ways. This is a good example of what Werner Wolf in an article calls ‘musical metalepsis’, which he contrasts with verbal or pictorial metalepses, that is, metalepses constructed through verbal or pictorial means (Wolf 2019). As it appears clearly in Wolf’s article, instrumental music alone, as a media product, cannot achieve metalepses. It does so only when it is incorporated in film: “what in ‘absolute’ instrumental music would be impossible but which the plurimedial combination of music, narrative and the moving image in the sound film can produce, namely musical metalepsis” (Wolf 2019: 29).

I follow Wolf’s definition of the concept of musical metalepsis, and analyse how music can be used in order to create metalepsis in other media where it is used, since Woman at War is particularly interesting from this point of view. Indeed, Bradshaw is right to point out that what is important in the use of music in this film is the performance of the musicians. Admittedly, the music itself is important in the film, from different points of views. Already the fact that Halla is a choir-leader, with several scenes presenting her while leading the choir, is a sign that music should be given particular attention. The score itself, specially composed for the film by Davíð Þór Jónsson, is original in its minimalism. Generally, it seems to be used in conventional ways, meaning the ways it has been established in classical Hollywood film. Here is a short list of the possible functions of music in such films:

It can establish setting, […] it can fashion a mood and create atmosphere; it can call attention to elements onscreen or offscreen, thus clarifying matters of plot and narrative progression; it can reinforce or foreshadow narrative developments and contribute to the way we respond to them; it can elucidate characters’ motivations and help us to know what they are thinking; it can contribute to the creation of emotions […] [it] can unify a series of images […] [and] encourages our absorption into the film by distracting us from its technological basis […]. (Kalinak 2010: 1)

All these functions apply to the use of music in Woman at War. However, there is much more to it. To start with, still in the traditional way, music highlights the double action of the film. The Icelandic music trio, who play the drums, the piano and the accordion, generally play music that reinforces the atmosphere of danger that has to do with Halla’s sabotaging of the aluminium plant. The other group, a Ukrainian female choir singing folk songs, is used for the other plot line, that is, the adoption of the Ukrainian orphan. The fact that the two groups only play together at the end, when Halla meets the girl she adopts, could be seen as a metaphorical use of music, to “clarify” the plot, as Kathryn Kalinak mentions in the quotation above.

As Bradshaw notices, music is also used in unexpected ways in Woman at War. One of these ways enters indeed in complete contradiction with what Kalinak listed as a function of film music, namely that it is “distracting us from its technological basis”. It is exactly when music is used metaleptically that this happens. Indeed, already in the first scene, when the credits still run, the music suddenly seems to lose its extradiegetic status, since the Icelandic trio appears in the frame, when the camera pans the beautiful Icelandic landscape. They stay there, “weirdly, incongruously” in the field, as Bradshaw rightfully notices. What Bradshaw misinterprets, however, is that music thus turns out to be diegetic. Indeed, just because the musicians are visually represented does not mean that they become characters in the diegesis. It is a misconception, only caused by the conventions from traditional films, that everything that is seen on the screen exists at the same level. It is enough to reconsider this by applying the mode of hearing, which should apply in the same way: everything that is heard in the film exists at the same level. Which obviously is not true, since extradiegetic music, or extradiegetic narrators in form of voice-over, are used frequently. So why could not extradiegetic elements be visually represented on the screen without “distracting” the viewers, as Bradshaw claims?

There are, indeed, clear arguments against considering the music in the film as diegetic. Actually, the music is not heard by the characters. Only two characters hear the music the same way as Harold Crick hears the narrator’s voice-over: Halla (a couple of times) and a Spanish backpacker who happens to be close by when Halla is hunted in the highlands (once). When these two characters show that they hear the extradiegetic music, it is when metalepsis actually occurs. Because, I argue, until the two different levels, the extradiegetic and the diegetic, interact, there is no real transgression. Or rather, I should say, the transgression is covered the same way as the use of extradiegetic music is covered by its frequent use. Because, as I already mentioned above in the analysis of Stanger than Fiction, all coexistence of diegetic and extradiegetic elements is transgressive, and thus metaleptical. So, what these examples show is that metalepsis actually loses its raison d’être when it is used in conventional ways.

The extradiegetic and the diegetic collide in another way too in Woman at War. This is also through music, or rather through the visual representation of the performance of the musicians. Indeed, the Icelandic trio actually interferes with the events in the diegesis at least three times. Two of these instants involve the use of music. Indeed, by changing the rhythm, the musicians warn the two characters that are able to hear them (Halla and the Spanish backpacker) that something dangerous is going to happen. In both cases, the characters react, and the course of action is changed: rather slightly in the case of the Spanish backpacker, who turns his head and discovers Halla running in the fields, and more importantly in the case of Halla, who realizes that the police are stopping people at the airport, and decides not to take that flight. The third case is even more clearly metaleptical, since the musicians physically interfere with the events in the plot. It is when, while playing the extradiegetic music as usual in Halla’s apartment, they decide to put on the TV set so that she can find out that the police have started the hunt. The TV set thus functions metaleptically as well, since it is at the same time both a diegetic screen and a metadiegeticscreen, according to the distinction made by Andrea Virginás in this volume (2020).

In conclusion, the musical metalepses in Woman at War are particularly interesting to study in a transmedial perspective, and by concentrating on the modalities of the media product, according to Elleström’s model of intermedial relations (2020). Indeed, what appears clearly is that when the mode of seeing is applied to extradiegetic music, that is, when the musicians are visually represented, what was hitherto a conventional use of extradiegetic music can be experienced as metaleptical. More importantly, when the effects that the extradiegetic world has on the events in the diegesis are visually shown, the metalepsis becomes clearer for the viewer.

4 Conclusion

This chapter did not aim at changing the definition of metalepsis. Its first definition, formulated by Genette in 1972, is still valid as a general starting point: “any intrusion by the extradiegetic narrator or narratee into the diegetic universe (or by the diegetic characters into a metadiegetic universe, etc.), or the inverse” (Genette 1980 [1972]: 234–235). I tried to show that metalepsis is a transmedial device, which can appear even in media products that are not language-based. What seems to be mandatory is the capacity to represent. And I thereby do not mean to represent in general, because all media represent in some ways, but to represent in some detail such features that are necessary for perceiving, for instance, ontological clashes and narrative levels.

What I particularly wanted to show is that the concept of narrator can obstruct the expansion of the applicability of metalepsis to other media. An intrusion, or rather a transgression, is needed, between two worlds, but these worlds do not necessarily have to be created by a narrator’s voice. Consequently, the initial definition can be improved, but it already possesses a flexibility, since it mentions “diegetic universes”. Admittedly, it lacks even a specific reference to horizontal transgressions, that is, between parallel universes, but I argue that that kind of transgression can be studied within the same framework, since the parallel universes are necessarily created from a superior level.

In order to prove my point, I started with the media type in which metalepsis was first defined, that is, literature, and studied one of the best known examples of metalepsis: the meeting between Unamuno, the author and the narrator of the novel Niebla, and a character created by him, Augusto Pérez. A closer look showed that even if the narrator is important in that scene, the most shocking is the representation of the whole meeting. Indeed, it is first when Unamuno, the narrator, becomes a character who physically meets Augusto that the metalepsis becomes really obvious and astonishing, both for Augusto (and eventually even for Unamuno) and for us, readers. What is interesting is that the shocking effect can be expanded to existence in general, since ultimately, even in reality, human beings can have the impression of being the creation of a God. This effect is obviously possible even in the case of other media types, as is explicitly mentioned in the filmLe tableau.

After the literary example, I analysed metalepses in three films, chosen not only because they actually contain metalepses, but also because they create it in different ways. The first example, Stranger than Fiction, is also the one which is the closest to metalepsis in literature, since the metalepses are occurring between a narrator, whose voice imitates a narrativevoice in literature, and a character created by this narrator. However, the ways in which the metalepses are achieved rely especially on the mode of sound, which is absent in literature. So, even if what we have in this film is a so-called intermedial imitation of a narrator, the ways in which that is done highlights the transmedial potential of metalepsis. A potential that is even clearer in the other two films that I analyse, since they take gradually greater distance from the use of a narrator.

The second film that I analysed, Le tableau, starts by representing a narrator who performs a metaleptical transgression, but this initial imitation of a literary narrator is not repeated, and the metalepses later on are constructed with the help of filmic media characteristics. One of these metalepses, constructed by applying to the visual mode, can be used as an argument for the possibility of horizontal metalepses.

The last film, Woman at War, is the one in which no reference is made to literature when it comes to its metalepses. Here too, it is through the mode of sound that the metalepses are principally created, but this time by using the qualified medium of music. What the analysis shows is that these ‘musical metalepses’ are all the more effective when the visual mode is used too, showing music as a performance act.

All in all, the analyses have shown that a transmedial perspective on metalepsis is beneficial, even in the case of literature. This is in line with Elleström’s view on media not as essential and hermetically closed entities, but as open and flexible concepts (2014, 2020). It has also proved productive to concentrate on the modalities of media, and to analyse how different modes can affect metalepses. What the analyses have shown too is that the complex device of metalepsis can be used to illustrate aspects of Elleström’s model of intermedial relations, especially those related to modalities (2020).