In its narratological sense, metalepsis, first identified by Genette, is a paradoxical contamination between the world of the telling and the world of the told: “any intrusion by the extradiegetic narrator or narratee into the diegetic universe (or by diegetic characters into a metadiegetic universe, etc.), or the inverse […]” ( 1980: 234–35). Described as “taking hold of (telling) by changing level” (235 n. 51) and thus combining the principle of narrative levels (Pier → Narrative Levels) with the rhetorical figure of metalepsis originating in ancient legal discourse, narrative metalepsis is a “deliberate transgression of the threshold of embedding” resulting in “intrusions [that] disturb, to say the least, the distinction between levels.” It produces an effect of “humor” or of “the fantastic” or “some mixture of the two […], unless it functions as a figure of the creative imagination […]” (Genette  1988: 88). Genette (2004) also argues that not only is metalepsis a violation of the separation between syntactically defined levels, but also a deviant referential operation, a violation of semantic thresholds of representation that involves the beholder in an ontological transgression of universes and points toward a theory of fiction (Schaeffer → Fictional vs. Factual Narration).
More is at issue, then, than localized rhetorical or stylistic devices, for metalepsis has been characterized as “undermining the separation between narration and story” (Rimmon-Kenan  2002: 93), as a “strange loop” (Hofstadter 1979) in the structure of narrative levels or a “short circuit” between the “fictional world and the ontological level occupied by the author” (McHale 1987: 119, 213), as a “narrative short circuit” causing “a sudden collapse of the narrative system” (Wolf 1993: 356–58), as producing a “disruptive effect on the fabric of narrative” (Malina 2002: 1), etc. Unlike factual narrative, moreover, fictional narrative betrays “at least the potential for narrative metalepsis” (Nelles 1997: 152). Such considerations raise not only the question of the metatextual status of metalepsis (Neumann & Nünning → Metanarration and Metafiction) and that of rhetorical as opposed to ontological metalepsis together with an array of topics bearing on transmediality (Ryan → Narration in Various Media) and transdisciplinarity (Meuter → Narration in Various Disciplines), but they also suggest that fictional narrative is by nature metaleptic, bound to the paradox of “a current presentation of the past” (Bessière 2005), or that “[a]ll fictions are woven through with metalepses” (Genette 2004: 131).
Narrative metalepsis as a concept results from the convergence of rhetoric (placing it alongside metaphor and metonymy as tropes of transformation, substitution and succession) and the principle of narrative levels. Genette ( 1980: 232–34) explains that metadiegetic (or second-degree) narrative bears either an explanatory, a thematic or an enunciative (rather than content-based) relation to the primary narrative, and it is under the latter that his comments on metalepsis are included, emphasizing “a shifting but sacred frontier between two worlds, the world in which one tells, the world of which one tells” (236). Essentially, metalepsis functions with varying dosages of three parameters: (a) illusion of contemporaneousness between the time of the telling and the time of the told (Wolf → Illusion (Aesthetic)); (b) transgressive merging of two or more levels; (c) doubling of the narrator/narratee axis with the author/reader axis. These features are illustrated by Balzac’s “While the venerable churchman climbs the ramps of Angoulême, it is not useless to explain…”—a “minimal” metalepsis (cf. Pier 2005: 249–50) which, being incipiently transgressive, leaps the boundary between narrator (Margolin → Narrator) and extradiegetic narratee on the communicative plane and puts story time on hold while the narrator intervenes with a metanarrative comment, demonstrating the latent metaleptic quality of narrative embedding in general.
Genette’s remarks, though concise, stake out the key features of metalepsis, one of the least debated of his theoretical innovations for many years. It is with subsequent and more differentiated developments that the scope and import of this narrative practice that goes against the grain of codified narratological categories has come to be more fully appreciated. Following a proposal by Ryan (2005, 2006: 204–30, 246–48), it is now widely acknowledged that metalepsis breaks down into a rhetorical (Genette) and an ontological variety (McHale), parallel to the distinction between illocutionary boundary at discourse level and ontological boundary at story level. “Rhetorical metalepsis,” Ryan claims, “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” (2006: 207). It has been shown by Fludernik, however, that Genette’s narrative metalepsis is in effect an umbrella term containing an implicit typology that integrates Ryan’s distinction (Fludernik 2003: 382–89): (a) authorial metalepsis (Virgil “has Dido die”): a metafictional strategy that undermines mimetic illusion, foregrounding the inventedness of the story; (b) narratorial or type 1 ontological metalepsis (in Eliot’s Adam Bede, the narrator invites the narratee to accompany him to Reverend Irwine’s study): transgression from the extradiegetic to the intradiegetic level is illusionary, drawing a fine line between the reader’s immersion and lifting of the mimetic illusion; (c) lectorial or type 2 ontological metalepsis (in a story by Cortázar, the reader of a novel is [almost] killed by a character in that novel): implication of the narratee on the story level or passage of a character from an embedded to an embedding level (also occurs in second-person narration); (d) rhetorical or discourse metalepsis (the Balzac example above).
Given the fluid transitions between these types, it can be seen that the more pronounced forms of metalepsis are contained embryonically in the fourth variety, suggesting that rhetorical metalepsis covers all four—whence the present author’s proposal to rename Fludernik’s and Ryan’s rhetorical metalepsis “minimal” metalepsis. Rather than two distinct types of metalepsis—one rhetorical, the other ontological—what is at stake are the forms and degrees of violation of the boundary between the telling and the told, two aspects of the effects of narrative discourse and, more generally, the role such violations play in artistic representation (cf. Häsner 2001: 40–3 on the “accentuation” of metaleptic relations).
Genette’s rhetorical theory of metalepsis highlights the relation between “figural” and “fictional” metalepsis. Both “figure” and “fiction” derive from the Latin fingere (to fashion, represent, feign, invent), such that a figure of substitution (i.e. a trope such as metaphor, metonymy, litote, etc.) forms the “embryo” or “outline” (esquisse) of a fiction (Genette 2004: 16–8). With emphasis on authorial metalepsis as a particular type of metonymy in which cause is expressed for effect or effect for cause and on the figural and fictional transgressions this entails, a fiction, taking form in the passage between figure as a formal but semantically weak verbal schema and figure as a transfer of meaning, is defined as “a figure taken literally and treated as an actual event” (20). In contrast to narrative considered as the “expansion” of a verb (cf. Genette  1980: 30), fiction can be regarded as a figure taken à la lettre, and in the case of metalepsis “fictively literalized,” it introduces into narratology the problem of ontological transgression in representation. The focus falls no longer on metalepsis as a narrative category forming a system with other describable categories (prolepsis, analepsis, etc.), but on the functioning of representation and the intersection of narrative and fiction. Called into question is the Coleridgean “willing suspension of disbelief,” triggering “a playful simulation of belief,” as in the fantastic or the marvelous mode (Genette 2004: 23, 25): with metalepsis, it is the reader’s belief, not disbelief, that is suspended, setting up a reading contract based not on verisimilitude, but on “a shared knowledge of illusion” (Baron 2005: 298; cf. Macé 2007).
In effect, the rhetorical and the ontological conceptions may represent not so much two types of metalepsis as they point to the two main approaches to the phenomenon, the one based primarily in the (rhetorical) effects produced by representation through discourse or other semiotic means, the other in the problems of logical paradox encountered by modern science. This can in fact be seen in the partially overlapping concerns of the two orientations. Ryan (2005: 205 n. 3) notes that Genette’s discussion bears on the two types without differentiating them, and also that “figural” and “fictional” metalepsis correspond roughly to “rhetorical” and “ontological” metalepsis (2006: 247 n. 3). It is useful to bear in mind, however, that for Genette fiction is addressed in rhetorical and pragmatic terms, while the ontological approach takes the transdisciplinary ramifications of scientific logic as its reference point.
Originating in rhetoric, later to be integrated into narrative theory, metalepsis is now seen as a more widespread phenomenon than initially thought and also to have affinities that vary according to different factors. Thus metalepsis, being paradoxical, is more likely to be cultivated by the baroque, by romanticism or by certain types of modernism than by mimetically inclined classicism or realism, much as it shows a greater propensity for the comic and the ironic than it does for the tragic or the lyric (cf. Pier & Schaeffer 2005: 10–1; Grabe et al. eds. 2006). Furthermore, being restricted by neither genre nor media, metalepsis is manifested in various ways and to different degrees: the theater arts, thanks to the possibilities of audience participation, are metalepsis-friendly; the cinema, with its technical capacity for hypotyposis (what is presented is depicted as though it were before one’s very eyes), can be highly metaleptic, contrary to music, suggesting that metalepsis is bound to the question of representation; the pictorial arts, as demonstrated by the works of Escher and Magritte, possess considerable metaleptic potential, but this is not the case of sculpture, where boundaries between levels are more difficult to define; digital media, with their capacity for generating virtual realities, are fertile terrain for ontological transgressions. And finally, metalepsis is not restricted to high culture, since it is freely resorted to in popular culture, as witnessed by reality TV or by unscripted spectator interventions at sporting events.
It is important to bear in mind that although narrative metalepsis is a recent concept in the history of poetics, the practice itself, under different denominations or none at all, extends back to antiquity. The fact that as a concept it can now be theorized and applied according to definable criteria casts a new light on the theory and analysis of narrative and, more generally, on representation as a cultural phenomenon.
The etymology of metalepsis is disputed, but its sense can readily be grasped from the word’s Latin equivalent—transumptio: “assuming one thing for another.” Metalepsis has a complex history in that it has been regarded either as a variety of metonymy, a particular form of synonymy, or both. As metonymy, it has been identified: (a) in simple form, or expression of the consequent understood as the precedent or vice versa and; (b) as a chain of associations (“a few ears of corn” for “a few years,” the transfer of sense implying “a few harvests” and “a few summers”). Metalepsis can also be understood in Quintilian’s sense as the intermediate step or transition between a term which is transferred and the thing to which it is transferred, resulting in an inappropriate synonym (Morier 1961; Burkhardt 2001; Meyer-Minnemann 2005: 140–43; Roussin 2005: 40–4; on metalepsis and evidentia, see Häsner 2001: 20–7; Cornils 2005).
From the perspective of narrative theory, two positions derive from the rhetoric of metalepsis. Firstly, Genette (2004: 7–16), drawing on the first two types above, notes that metalepsis shares with metaphor and metonymy the principle of transfer of sense and considers it (following Dumarsais) a metonymy of the simple type; he then expands it (with Fontanier) beyond the single word to include an entire proposition. Metalepsis, he argues, combines cause for effect or effect for cause with substitution of an indirect for a direct expression. He points out the importance, in narrative, of authorial metalepsis, by which an author “is represented or represents himself as producing what, in the final analysis, he only relates” (Fontanier). He draws attention to the proximity for the two rhetoricians of metalepsis and hypotyposis (a figure in which the copy is treated, illusorily, as though it were the original), but particularly to the fact that with metalepsis, the author pretends to intervene in a story which is in fact a representation, so that transgression of the threshold of embedding merges with that of the threshold of representation, affirming the existence of the very boundaries that are effaced.
There have also been proposals to refer narrative metalepsis back to metalepsis as use of an inappropriate synonym, notably by Meyer-Minnemann (2005) and Schlickers (2005) (see also Nelles 1997: 152–57). The emphasis here is not on authorial metalepsis as a type of metonymy, but on the paradoxical transgression of boundaries, of which there are two main types: one at discourse level with breaching of the “me-here-now” of enunciation (in verbis transgression), the other at story level with violation of the coordinates of the enunciate (in corpore transgression). Taking a cue from Genette, this model provides for metalepsis of enunciation and metalepsis of the enunciate in which each functions either vertically (bottom-up or top-down) or horizontally, i.e. without change of level (dubbed “perilepsis” by Prince 2006: 628). To take only a few illustrations: (a) a vertical metalepsis of enunciation (top-down) would be the Balzac example cited above; (b) a horizontal metalepsis of enunciation occurs with the juxtaposition of two communicative situations at the same level; (c) with transgression of the diegetic, ontological, spatial or temporal order, there occurs a vertical metalepsis of the enunciate; (d) a horizontal metalepsis of the enunciate is produced when e.g. Woody Allen enters the world of Madame Bovary. In this system, metalepsis is seen as producing an effect of strangeness, either comical or fantastic, but it is not regarded as a figure of fictionality in Genette’s sense (on the fictionality of paradoxical narration, however, see Meyer-Minnemann 2006).
For narrative metalepsis in an ontological perspective, paradox is central, as it involves the logically inconsistent passage between two separate domains through suspension of the excluded middle. At issue is the problem, originating in logic and mathematics, of maintaining distinct levels through avoidance of self-reference by elaborating meta- levels, an endeavor that requires the addition of recursive meta-levels ad infinitum. The inevitable paradox is captured by Gödel’s theorem, although it has long plagued scientific thought in the form of the liar’s paradox (Epimenides is a Cretan and says “All Cretans are liars”); it is also conveyed visually by the Möbius strip, Klein’s bottle and Escher’s drawings. Hofstadter (1979) has examined various manifestations of this paradox in his important transdisciplinary study, even providing a recursive dialogue (103–26) that illustrates the problem of metalepsis, although the term appears nowhere in the book.
McHale has integrated these paradoxes into the poetics of postmodernist fiction, a type of writing that “foregrounds ontological issues of text and world” (1987: 27). Adopting an ontology taken from possible worlds theory (33–6), he recasts Genette’s narrative levels in terms of ontological levels so that a metalepsis produced by violation of levels raises ontological considerations resulting from recursive embedding (120). In particular, McHale identifies metalepsis with the “Strange Loop,” a phenomenon that occurs “whenever, by moving upwards (or downwards) through the levels of some hierarchical system, we unexpectedly find ourselves right back where we started.” Strange Loops occur within a “Tangled Hierarchy”: “when what you presume are clean hierarchical levels take you by surprise and fold back in a hierarchy-violating way” (Hofstadter 1979: 10, 691; qtd. in McHale 1987: 119). He also draws attention to the metaleptic function of the second-person pronoun (223–25), as does Genette (2004: 96–9; cf. Fludernik’s 2003: 389 lectorial or type 2 ontological metalepsis); but he does not distinguish between rhetorical and ontological metalepsis, nor does Wolf (2005b), whose definition of metalepsis combines ontology with possible worlds theory (93).
A particular capacity for generating feedback loops and hierarchies of levels is demonstrated by the computer, dubbed “metaleptic machine” by Ryan. A case in point is the “Metalepticon,” a computer algorithm designed by Meister (2005) to reproduce the recursive structures of Escher’s Drawing Hands: here, however, computational powers are quickly exhausted and “Program Space Full/looping error” is displayed. Meister concludes from this unrealizable abstract formal model that metalepsis annuls the “contract of representation” required for the cognitive and hermeneutic processing of esthetically incarnated metalepses (245–46). Also related to issues of communication is metalepsis as a virtual reality (possibility for the beholder to physically enter the imagined world) and as an existential phenomenon (Emma Bovary modeling her life after the heroine of a sentimental novel) (Ryan 2006: 227). On the other hand, the recursive chain is broken when it is recognized, for example, that the creator of Drawing Hands occupies a space outside the representation in question, even though that creator can in turn be portrayed in a (meta-) representation (cf. “Authorship Triangle”; Hofstadter 1979: 94–95, 688–89)—a situation not unlike that of authorial metalepsis.
As seen in 2.1 and 2.2 above, Genette’s original conception of narrative metalepsis hinted at a typology without actually proposing one. Since then, a number of typologies have been elaborated, a survey of which reveals that to varying degrees theories of metalepsis discriminate between minimally and conspicuously transgressive changes of level. Ontological approaches tend to focus on the latter while rhetorical approaches also take account of the metaleptic potential of e.g. the apostrophic “gentle reader.”
Nelles (1997: 152–57), referring narrative metalepsis back to metonymy as trope (Quintilian), differentiates “unmarked” (at discourse level) from “distinctly marked” (at story level) metalepsis and, for the latter, “intrametalepsis” (movement from the embedding to the embedded level) from “extrametalepsis” (movement in the opposite direction), subdividing each type into analeptic and proleptic forms on the temporal plane (on “inward” vs. “outward” metalepses, see Malina 2002: 46–50). The degree of transgression—knowledge of the other world as opposed to physically penetrating it—is characterized as either epistemological (verbal) or ontological (modal). According to Pier (2005: 253), there is a tendency in intrametaleptic movements to favor the narrator/narratee relation, and in extrametaleptic movements the character/narrator relation.
Wagner (2002: 243–48), for whom the metatextual nature of metalepsis signals the constructedness of narrative along the lines of the Russian formalist notion of “defamiliarization,” emphasizes the reversibility of metaleptic displacements between extra-, intra- and metadiegetic levels. He also draws attention to circulation, at a given level, between collateral fictive universes, not unlike the “horizontal” metalepses included in the Meyer-Minnemann/Schlickers model. And finally, Wagner takes up the question, largely neglected, of the compositional distribution of metalepses: their location, amplitude and frequency can have a significant impact on the strategy and readability of a narrative (cf. Häsner 2001: 40–3).
Defamiliarization and composition point to the Russian formalists’ use of metalepsis, although the term was not employed by them. At issue was neither a rhetorical figure nor an ontological paradox or a typology of its use, but “laying bear the device”: the deliberate distortion of form aimed at highlighting the artificial relations between “form” and “materials,” between sujet and fabula, and the fact the art is “made” of devices. As shown in particular by Šklovskij (1921) in his essay on Tristram Shandy, the numerous digressions, etc. “lay bare” the relations between the time of the telling and the time of the told, thus conflating narration and action in a seemingly unmotivated way and drawing attention to the idea that form, not the world, is the content of the novel (cf. Schmid 2005).
Herman (1997: 133–36) analyzes metalepsis firstly by identifying the textual markers that, in the formal sense, signal “illicit movements up or down the hierarchy of diegetic levels structuring narrative discourse” and, from the functional perspective, “transgression of the ontological boundaries.” In terms of possible worlds theory, metalepsis solicits temporary entry of the reader into a re-centered modal system. Since, in this account, metalepsis abolishes the distinction between storyworld and the world(s) from which addressees relocate, Herman adopts Goffman’s concept of frame analysis as a set of expectations about narrative universes in place of diegetic level.
Wolf, considering the forms of disturbance of mimetic illusion caused by the failure to observe ontological boundaries, sets the mixing of extra-fictional reality with textually produced fiction off from violations of levels in inner-fictional boundaries (1993: 349–72). The latter, metalepses, are a metafictional technique characterized as a “narrative short circuit” and are likened to Hofstadter’s (1979: 134 passim) “heterarchy,” a structure distinct from hierarchy in that it possesses no single “highest level” (cf. McHale 1987: 120). Metalepsis occurs (a) between the extradiegetic and the intradiegetic levels or (b) between the intradiegetic and one or more hypodiegetic (metadiegetic) levels (on “exterior” vs. ”interior” metalepsis, see Cohn 2005). Both (a), marked by punctual violations of levels by characters and/or their words, and (b), punctual short circuits between intradiegetic “reality” and “fiction,” are found in minimal and conspicuous forms and can take place either bottom-up or top-down. In its complex form, metalepsis combines the previous two types, setting in motion a recurrent contamination of levels, as in the Möbius strip, and whose effects are distributed throughout a given narrative.
Where the above typologies can be grouped under the heading “meta-,” situating metalepsis on the same conceptual plane as metanarrative, metadiegesis, etc., another typology, the most elaborate to date, is built up around the suffix “-lepsis” in the sense of “action of taking” (Lang 2006; cf. Meyer-Minnemann 2006). Narration is paradoxical when, in violation of the principle “either one or the other” (cf. the liar’s paradox), coherence (Toolan → Coherence) is undermined. On this basis, Lang provides a typology of paradoxical narration divided into devices that cancel out boundaries (syllepsis, epanalepsis, the latter term designating specular devices including the mise en abyme) and those that transgress boundaries (metalepsis, hyperlepsis, the latter equivalent to Genette’s pseudodiegesis: a metadiegetic narrative presented as though it were diegetic). As in the Meyer-Minnemann/Schlickers typology, each of these devices is analyzed into vertical (bottom-up or top-down) and horizontal relations of discourse and story, respectively. In contrast to the other models presented, this typology includes metalepsis among other forms of paradoxical narration.
As shown, inter alia, by Lang’s typology, the scope of paradox-producing devices is not restricted to metalepsis. Thus the effects of pseudodiegesis (or hyperlepsis), also studied under the term trompe l’œil by McHale (1987: 115–19), can produce “variable realities” as destabilizing as those of metalepsis. As for mise en abyme, it shares with metalepsis the feature of embedding, but it additionally includes resemblance between levels (e.g. a story within a story) and reduplication and is characterized by reflexivity rather than by transgression of levels. Only in the case of “aporistic reduplication” (“fragment supposedly including the work in which it is included”; Dällenbach 1977: 51) does mise en abyme coincide with metalepsis (called “pure” mise en abyme by Cohn 2005: the reader has the impression of belonging to an infinite series of fictions; cf. McHale 1987: 124–28). And finally, while metalepsis is generally found within a given text, violating that text’s system of diegetic levels, infringement of boundaries can also take place across texts. Such is the case in horizontal metalepsis of enunciation, studied by Rabau (2005) under the term “heterometalepsis,” but it also occurs in horizontal metalepsis of the enunciate, a phenomenon that coincides with “transfictionality” as when, say, Sherlock Holmes appears in the fictional universe of Madame Bovary (cf. Lavocat 2007 on metaleptic and intrametaleptic transfictionality). This dimension of metalepsis opens up issues of transtextual relations (cf. Genette 1982), but it also touches on the numerous implications of metalepsis for fictionality and metafictionality.
The violation of levels and boundaries is not limited to narrative, and while metalepsis in its narrative form was originally studied in verbal narratives, it is not a media-specific phenomenon. This is confirmed by a number of contributions in Pier & Schaeffer (eds. 2005) as well as by Genette (2004), much of which is devoted to metalepsis in theater, film, television, painting and photography (see also Genette 2009: 176–80), and Wolf (2005b) which, additionally, looks at comic strips. It would seem, then, that metalepsis has a significant role to play in transmedial narratology (e.g. Ryan 2006: 3–30, ed. 2004) and in intermediality (e.g. Wolf 2005a), although to date this connection remains largely unexplored.
More than a rhetorical flourish, metalepsis raises the question of the porosity of levels and borders in cultural representations, but not their dissolution. Originating in structuralist narratology, it calls for re-examination of the theoretical basis of established models and thus merits serious consideration in charting out transdisciplinary approaches to narrative theory. Among topics requiring further study are: (a) relative weight of local vs. global effects of metalepsis; (b) metalepsis and fictionality (breaking/intensification of mimetic illusion, immersion, etc.); (c) metalepsis and related practices in historical poetics going back to biblical narrative as well as a historical inventory of artistic movements and corpuses employing these devices; (d) the role of metalepsis in trans-/intermediality with regard in particular to multimedia and popular culture.