At first glance, autofiction seems to be at odds with theories of fiction and fictionality. Serge Doubrovsky’s inaugural definition of autofiction, on the back cover of Fils (1977), arguably capitalizes on a broadly postmodern or poststructuralist consensus around the fictional status of self-narration: even if the events and facts recounted are “strictly real,” the “adventure” of language produces a fiction.Footnote 1 By contrast, both semantic and pragmatictheories of fiction and fictionality, especially as they have developed since the 1990s, have tended to reaffirm the fundamental distinction between fictional and nonfictional narratives, aiming to specify the borders, the autonomy—the “distinction” as Dorrit Cohn (1999) puts it—of fiction. Although this ostensible opposition between autofiction and theories of fiction requires some qualification (as we shall see), it no doubt accounts for the misunderstandings that arise in debates around autofiction, autobiography, and autobiographical fiction. For those who wish to maintain a clear distinction between fact and fiction, autofiction must appear defective: a narcissistic failure of imagination, in the eyes of the partisans of fiction (e.g., Petit 1999); an excuse for carelessness with the truth, for those who defend the specific referential claims of factual literature (e.g., Lejeune 2007, 3). At first glance, the same characteristics that have allowed autofiction to become a broad, fluid, almost infinitely extensible term also make it the object of an impasse—amenable to sociological study, perhaps, as Jean-Louis Jeannelle has argued, but resistant to literary-theoretical investigation (2013, 226).

Attempting to move beyond this theoretical deadlock, I will posit here that theories of fiction and fictionality can indeed shed light on autofiction, and vice versa. I will first examine how accounts of autofiction engage with theoretical approaches to both autobiography and fiction, before asking whether autofiction can be reconciled with existing definitions of fictionality. Drawing on pragmatic, narratological, and rhetorical theories of fictionality, I will then aim to locate factual and fictional modes at work within texts, showing how they operate at the level of formal devices or narrative frames to foreground either referential force or the work of fictionalization. This chapter thus aims to bring some precision to our understanding of the fictional in autofiction, while also accounting for ambiguities in reception. Ultimately, we will see that autofictional texts allow for a range of configurations of the fact/fiction relationship, while theory can help us locate sites and signposts of fictionality or factuality within works. Conversely, due to the very ambiguity and hybridity of autofictional texts, they can serve as a useful empirical testing ground for theories of fiction, which have traditionally based their arguments on narratives and entities already established as generically fictional (the exception here are rhetorical theories of fictionality, discussed later, which identify local uses of fictionality within nonfictional discourse [Walsh 2007; Nielsen, Phelan, and Walsh 2014]). This dialectical approach takes theories of fiction out of their comfort zone in the novel, while also attempting to bring some clarity to the debates about autofiction.

The Theoretical Adventures of Autofiction

Autofictional texts bring to light disjunctions between theory and practice, as well as divergent understandings of the fact/fiction divide. Doubrovsky’s initial coining of the genre descriptor responds in large part to renewed scholarly interest in the genre of autobiography, in particular Philippe Lejeune’s influential work—as Doubrovsky himself confirms in a letter to Lejeune, mentioning his desire to fill the “empty square” in the latter’s analysis (cited in Lejeune 1986, 63). In 1975, Lejeune had posited the identity of author, narrator, and character as the fundamental condition for autobiography. This identity may or may not be backed up by an explicit “autobiographical pact” that affirms it within the text (Lejeune [1975] 1996, 26). Aside from the classic case where we find both a proper name and an explicit pact (the canonical model of Rousseau’sConfessions, Lejeune’s case 3b), it is possible, within Lejeune’sschema, to find “indeterminate” cases where no proper name or pact allows genericidentification (case 2b); to find an autobiographical pact without mention of a proper name (case 2c); or to observe an identity of proper names without a direct autobiographical pact (case 3a) (Lejeune [1975] 1996, 28–30). What does not seem possible, however, is an explicit divergence of identity and pact, hence the two empty squares in Lejeune’s chart of autodiegetic narratives (28). The difficulty is not just empirical (even if Lejeune initially presents it as such, mentioning a lack of examples) but also logical, for it is precisely the identity of author, narrator, and character that grounds the autobiographical pact for Lejeune. Yet, Doubrovsky insists on this divergence in Fils. On the one hand, the narrator is clearly named “Serge Doubrovsky” and is a university professor in the United States; on the other, in addition to the famous sentence on “autofiction” on the back cover, the front cover bears the designation “novel” (“roman”). The notion of autofiction thus originates as a curious and contradictory theoretical experiment, one that simultaneously occupies one of the empty squares in Philippe Lejeune’s chart and undermines the pragmatic basis of Lejeune’s distinctions.

At the same time, Doubrovsky’s aim is not to define a genre but to describe a particular literary practice—his own, which leans toward the auto rather than the fictional. As Arnaud Schmitt notes (2010, 126), Doubrovsky’s term is flawed “since autofiction as a substantive lays stress on the non-referential part of the personal discourse, whereas Doubrovsky’s textual practice went rather in the opposite direction.” Frank Zipfel observes that Doubrovsky seems to use “fiction” merely to designate a specific form of non-chronological and associative construction, which, strictly speaking, does not make his narrative fictional (2009, 299). Doubrovsky’s own conception of autofiction as a “fiction based on strictly real events and facts” thus remains close to autobiography, while incorporating (at least in its initial form) a psychoanalytic dimension and a focus on linguistic “adventure” (see Gasparini 2008, 19–31). In practice, of course, readers may not know in which respects exactly the author “strictly” adheres to the facts. Other theorists’ definitions, taking seriously the fiction component of the term, demand an overt “self-fictionalization process” or “autofabulation” (Colonna 2004, 75), which combines onomasticidentity with avowed fictionality. But autofiction has also been seen as a category of “undecidable” texts (Bersani, Lecarme, and Vercier 1982, 150–165) or even as a way of erasing the distinction between reality and fiction altogether (Vilain 2005, 124–125).

Autofiction owes much of its success, but also its ambiguity, to the semantic fuzziness of the term fiction, which (in both English and French) is sometimes conflated with narrative emplotment in general or else with literariness as such. This ambiguity may also be cultivated to express suspicion of referential discourse. Doubrovsky acknowledges as much in a 2005 interview, where he describes “autofiction” as a postmodern variant of autobiography, suited to a moment that no longer believes in the literal truth of historical narrative (2005, 212). Other practitioners of autofiction make similar claims, sometimes in rather contradictory terms. For instance, Catherine Cusset distinguishes between strict factual accuracy and the writer’s quest for truth, while claiming that “the only fiction in autofiction is the work on the language” (2012). While the recognition of the gap between life and stories might seem innocuous enough, it can be identified with a broader skepticism that rejects the notions of unitary selfhood and transparent self-discovery (see Zipfel 2009, 308), and even the possibility of referential language. This skeptical position risks extending the notion of fiction to the point of meaninglessness, subsuming all utterances into this category in line with what Marie-Laure Ryan (1997) calls the “doctrine of panfictionality.”

Fictionality and Hybridity

Theories of fiction, by contrast, have aimed to develop precise accounts of fiction and fictionality, even as they offer different approaches to the relationship between fictional and referential narrative. For semantictheories of fiction, for instance, fiction depends on structures of reference and the ontological status of named entities, not on stylistic factors or elements of narrative construction. In philosophical debates on nonexistent entities, the proper names of fictional characters serve as paradigmatic cases—notably Odysseus (Frege [1892] 1948), Mr. Pickwick (Ryle 1933), and Sherlock Holmes (Kripke 1972; Lewis 1978)—whether it is to conclude that these names have sense but no reference (Frege), to allow reference to pertain in some possible world (Lewis), or to argue for the “discursive unity” of fictional worlds beyond the propositional content of their components (Pavel 1986, 16).

While the fictional status of Sherlock Holmes or Mr. Pickwick is not in doubt, the semantic or ontological issue at stake in autofictional writing is the referential force (or lack thereof) of the autodiegetic “I” and the associated proper name. We should note here that several definitions of fiction rely precisely on the status of the speaker or subject. This is the case for Käte Hamburger’s distinction in The Logic of Literature between reality statements that originate with a real “I-Origo”—that is, a genuine statement subject at the center of a “system of temporal and spatial coordinates” (1973, 67)—and statements that refer what is narrated to fictive “I-Origines” (73). On this basis, Hamburger goes so far as to exclude all first-person narratives from the category of fiction proper: because of their autobiographical origins in the statement system (213), first-person narratives can only be “feigned” reality statements (rather than fictive figurations), while their “degree of feint is subject to variation” (328).Footnote 2 It is also from the point of view of the speaker’s relationship to his/her speech acts that later philosophers and narratologists will distinguish fact from fiction, albeit without following Hamburger’s distinction between I-Origo and I-Origines. John Searle separates fictional utterances from truth claims via an account of “pretended illocutions” (1975, 326)—operating according to conventions that suspend the speaker’s normal commitments to the truth of propositions (the question, then, is not who speaks, but whether the speaker is pretending to make assertions). In Fiction and Diction, Gérard Genette refers to both Lejeune and Searle in basing his definition of fiction on the relationship between author and narrator: while their “rigorous identification […] defines factual narrative,” “their dissociation […] defines fiction, that is, a type of narrative whose veracity is not seriously assumed by the author” (1993, 70). The question hinges, as with Lejeune, on establishing the identity or nonidentity of author and narrator. Autofiction, in this perspective, can only be a logical and pragmatic contradiction, amounting to the statement “It is I and it is not I” (Genette 1993, 77).

Dorrit Cohn, who defines fiction as “literary nonreferential narrative” (1999, 12), differs from Genette in positing the existence of formal “signposts of fictionality” that mark the unique domain of fiction: the “synchronic bi-level” model of narrative (that is, the division into story and discourse), certain narrative modes free of referential constraints (for instance, techniques for the presentation of consciousness), and the “doubling of the narrative instance into author and narrator” (130). This last criterion coincides with Genette’s distinction and serves to define nonfiction in terms of the identity of author and narrator. Cohn groups autofiction, mentioned only in passing, with other “crossbreeds” that “adopt the contradictory practice of naming their fictional self-narrators after their authors, thereby effectively ambiguating the distinction between fiction and nonfiction for self-narrated lives” (1999, 94). Cohn’s account, like Genette’s, is consistent with Lejeune’s view of the autobiographical pact—or what he will later call the “truth pact”—as an all or nothing proposition, with respect to which the term “autofiction” can only generate perplexity (Lejeune 2005, 25–26).

We fall back, then, on the impasse already outlined above, which seems to make the notion of autofiction incompatible with any serious account of fictionality. The same seems to hold for Kendall Walton’s definition of fiction as “make-believe” (1990), or Jean-Marie Schaeffer’s “shared ludic feint” (2010, 138–139). In its Aristotelian argument for the distinction between reality and its mimetic representation, and its insistence on the pragmatic framing that creates the conditions of make-believe or ludic feint, Schaeffer’s account can be partially aligned with Searle’s account of “pretense” and, by extension, with Lejeune’sreading pact; thus, Schaeffer speaks of a “pragmatic contract,” by which a fiction announces itself as such (2010, 137). As for Olivier Caïra’s extension of the category of fiction to incorporate non-mimetic, “axiomatic” fictions (such as rule-based games), it still rests on a basic distinction between “documentary” and “fictional” communication, established via pragmatic “framing operations” (2011, 75).

To compound the problem, these approaches to fiction often have difficulty accounting for literary nonfiction. Genette is most explicit about this issue, asserting that fictions are “constitutively” literary, while the status of factual narratives depends on the “conditional” criterion of “diction”—that is, on formal features or a subjective judgment of aesthetic value (1993, 138). Of course, this exclusion itself helps explain the success of autofiction as a strategy for legitimizing autobiographical writing, as the novelist Marie Darrieussecq ironically quips: “Since autobiography is questionable and conditional, and since all fiction is literary, let’s bring autobiography into the field of fiction” (1996, 372–373). To put the point differently, this strategy mobilizes a postmodern, panfictionalist view of narrative (“all narrative is fictional”) in order to claim a place within a literary regime that privileges fiction.

Some theorists, nevertheless, have successfully brought the fact/fiction distinctions of fiction theory into dialogue with approaches to autofiction. Arnaud Schmitt returns to Hamburger’s (rather than Schaeffer’s) conception of “feint” in order to define “self-narration” (autonarration), which he prefers as a term over “autofiction,” as a “loosely referential literary genre” (2010, 129) that admits of degrees of proximity to reality (2007, 22), and draws on the formal resources of the novel to prioritize self-exploration and self-expression over precise factual accuracy. Self-narration amounts to a “more sophisticated” (2010, 133) mode of autobiography, specific to twentieth-century conceptions of the self. Frank Zipfel prefaces his reflection on autofiction with a definition of fiction as a rule-governed speech act, which produces a narrative with a “non-real” story that is to be received with an attitude of “make-believe” (2009, 289). He then distinguishes between three conceptions of autofiction: autofiction as a particular form of autobiography that (debatably) conflates fiction with narrative construction (Doubrovsky’s definition) (290); autofiction as a particular kind of fictional telling (corresponding to Colonna’s“autofabulation”), where a fictional figure has the name of the author (302–303); and finally, autofiction as a combination of fictional and autobiographical pacts (304).

The recent revival of “non-communicational” theories of fiction, such as Sylvie Patron’s “optional-narrator” approach, which relates Hamburger’s positions to those of Ann Banfield and S.-Y. Kuroda (Patron 2009), invites us to rethink the question of the speaker in both factual and fictional narratives. In this light, we may perhaps move beyond the simple binary of the identity/nonidentity of narrator and author that leaves no space for autofiction as a third term. In the domain of autofiction studies, Schmitt’s return to Hamburger’s conception of “feint” points to some intriguing possibilities, while reminding us of the ambiguous proximity of first-person fiction to autobiography. Without pursuing this last issue fully, I will focus on the question of the combined reading pact, following Schmitt’s argument against the possibility of a “simultaneous approach” (2010, 128), as well as Zipfel’s claim that readers in fact switch between modes of reading (2009, 306). In this volume, Martina Wagner-Egelhaaf (Chap. 2) and Alexandra Effe and Alison Gibbons (Chap. 4) also make the case for an oscillation in the reader’s attitude to autofictional texts. Such oscillations, I would argue, require a fine-tuned attention to the shifting voice(s) of the work. Speaking of non-Francophone autofiction, Karen Ferreira-Meyers points to a general consensus that readers remain “very much able to see and keep a line of demarcation between fact and fiction” (2018, 42). Françoise Lavocat (2016, 522) has argued for the broader necessity of just such a demarcation, noting that apprehending the modes of hybridization between the fictional and the factual need not entail erasing the contours of either sphere.

To apprehend these contours, it is helpful to turn to Richard Walsh’s account of fictionality as “a distinctive rhetorical resource, functioning directly as part of the pragmatics of serious conversation” (2007, 1). Such “rhetorical” approaches, which distinguish between fiction as a genre and “fictionality as a quality or fictive discourse as a mode” (Nielsen, Phelan, and Walsh 2014, 62), allow for a distinction between “local” and “global” fictionality, such that nonfictional texts can contain “passages of fictionality” (67). In practice, as we shall see, the local and global are not entirely separable: for instance, particular formal devices that function as “signposts of fictionality” in a fictional context can produce different effects in a factual one. Still, when applied to autodiegetic narratives, the distinction between local and global fictionality allows us to theorize forms of hybridity that do not erase the border between fact and fiction or require the simultaneous adoption of contradictory modes of reading. Autofictional texts themselves help us draw out some of these distinctions and theorize kinds and degrees of fictionality. To demonstrate this point, I now turn to a few specific cases. Needless to say, these cases do not cover all configurations of the fact/fiction relationship or the great variety of autofictional practices. They do, however, serve to exemplify the operations and effects of fictionality and factuality at the level of narrative voice and readerly contracts, and moreover showcase a variety of autofictional texts across Anglophone, Francophone, and Scandinavian literatures.

I/Not I

The assertion, “It is I and it is not I,” which constitutes for Genette the fundamental contradiction of autofiction (1993, 77), is in fact a common claim in both fictional and factual autodiegetic narratives. Generally speaking, it does not open a breach at the level of the text’s pragmatic contract, but rather functions on the thematic level to expresses a non-unitary conception of the self. However, narrative voice and conceptions of selfhood are not always easily separable, especially when what is at stake is not only the relationship between a narrated and narrating I but also experimentation with new uses of the first person. The French writer Annie Ernaux, for instance, describes her use of the “transpersonal I” as an attempt to grasp social and familial reality: “The I that I use seems to me an impersonal form, barely gendered, sometimes more an utterance by the ‘other’ than by ‘me’: a transpersonal form, in short. It is not a means of constructing an identity through a text, of self-fictionalizing myself [de m’autofictionner]” (Ernaux 1993, 221; my translation).

Ernaux rejects the term autofiction, repeatedly insisting on her scrupulous adherence to the factual. Her works since La Place (1983; translated as A Man’s Place [1996]) can certainly not be considered fiction in the sense of involving “shared ludic feint,” in Schaeffer’s terms. Still, the quotation above illustrates the enunciative gap that she opens up within the I itself. In the phrase, “The I that I use,” are both instances of the pronoun transpersonal? Or does a personal I (that of the author) authorize and instrumentalize the transpersonal form (of the narrator)? This kind of metadiscursive commentary that introduces a split within the first person is also present in Ernaux’s literary works. If there is any destabilizing of the referential ground of Ernaux’s texts, it surely lies in the simultaneous impossibility and necessity of this “I-Origo” (to use Hamburger’s term). Ernaux’s experiments with the I invent an ambiguousspace of projection, positioned between the individual and the collective. In Les Années (2008; translated as The Years [2017]), she extends the experiment to other pronouns: the impersonal on (one), the third-person elle (she), and the first-person plural nous (we):

We [on] changed [changeait] plates for dessert, quite mortified [mortifiée] that our fondue bourguignonne had not been greeted with the expected congratulations, but with curiosity and comments that were disappointing at best, considering the trouble we’d gone to [qu’on s’était donné] with the sauces, and even a touch condescending. (Ernaux 2017, 91; French original: 2008, 97)

Globally, in its narrative framing and paratextual apparatus, Les Années remains a work of nonfiction. Still, the “I-Origo” that anchors the deixis in a given time and place is troubled by Ernaux’s play with pronouns. In the passage above, for instance, the impersonal pronounon is paired with the feminine singular gender agreement of mortifiée, while the imperfect tense suggests an iterative action somewhat incompatible with the specific scene of the failed fondue bourguignonne. What we are invited to participate in is not ludic pretense, but rather an interpretative operation that transforms a specific scene and personal feeling into a general scenario and a broader social symptom, while still maintaining a tension between the singular and the collective.

Many other writers also experiment with the first person without necessarily troubling the foundations of the autobiographical pact. Édouard Louis’sEn finir avec Eddy Bellegueule (2014; translated as The End of Eddy [2017]) stages in its title the desire for a violent break between the child Eddy Bellegueule (and his social milieu) and the narrating subject, the author Édouard Louis. This onomastic split, along with the designation “novel” (roman) on the book’s cover, may help to account for the frequent characterization of this work (and Louis’s other works) as autofiction, but in terms of global factuality, the autobiographical pact does not seem to be in question. Like Ernaux’s works, those of Louis rule out overt fictionalization, as they rely on the referential force of self-exposure as the basis for the contagious sense of shame that confronts the reader with the violence of social and sexual norms.

We might contrast these cases with some globally fictional first-person narratives that are sometimes characterized as autofiction, and which share Ernaux’s concern with intersubjectivity and the transpersonal. In Rachel Cusk’s “Outline” trilogy (Outline [2014], Transit [2017], and Kudos [2018]), the narrator appears only in her interactions with others.Footnote 3 The basic principle of the work is articulated in the first volume, Outline, in a momentary miseen abyme—voiced not by the main narrator but by another character, Anne, who describes her conversation with a man sitting next to her on a plane:

in everything he said about himself, she found in her own nature a corresponding negative. This anti-description, for want of a better way of putting it, had made something clear to her by a reverse kind of exposition: while he talked she began to see herself as a shape, an outline, with all the detail filled in around it while the shape itself remained blank. Yet this shape, even while its content remained unknown, gave her for the first time since the incident a sense of who she now was. (Cusk 2014, 239–240)

The narrator, a writer named Faye (although she is only named once in each book of the trilogy), records Anne’s theory of fragile selfhood with the same passivity as she receives other characters’ impressions and ideas. Although Faye clearly shares several characteristics with Cusk (biographical features that serve as “identification operators,” as in the cases analyzed by Arnaud Schmitt in this volume), our desire to identify them stems above all from the attentive, receptive, and porous point of view that the narrative invites us to inhabit, by recording less Faye’s own interiority than the voluble self-disclosures of those she encounters. The first-person subject at the center of the text is concerned less with direct self-narration than with a curious form of exteriorization, as those around the narrator offer reflections on everyday life, human relationships, and the human condition in extended monologues.

The trilogy produces what we might call an “autofictional effect,” despite the lack of onomasticidentity between author and character. The effect goes beyond the possibility of reading these texts as fictionalized autobiography (seeing Faye as a version of Cusk), although that is certainly one interpretative option. It has to do with the ambiguities and discomfort that Cusk produces by reversing the standard fictional strategies for representing consciousness. We may recall that Dorrit Cohn’s “signposts of fictionality” involve, above all, techniques for the representation of consciousness. Cusk turns this novelistic interiority inside out, as the narrator simply receives the externalized thoughts of others. A metanarrative moment toward the end of the third volume, Kudos, acknowledges and exploits the artifice of this technique. Over dinner at a writers’ conference, a woman named Sophia delivers an extended monologue on marriage, motherhood, divorce, her relationship to men, and her own sense of self, as her audience gradually becomes “visibly uncomfortable” with such self-exhibition (Cusk 2018, 162). Another novelist later remarks that “things had got pretty intense back there” (164). The scene foregrounds Cusk’s own writing technique in the form of a fictional writer’s inappropriate confession, raising the discomforting question of who is actually speaking in the novel’s monologues. Throughout her trilogy, Cusk sustains the intensity of a multicentered subjectivity that is not performed as much as it is compulsively exuded. It emanates from individuals while also constituting an impersonal, diffracted, or projected version of the narrator’s own consciousness.

We might compare this reading experience with another case where a displaced I governs an ambiguously fictional account. Olivia Laing’sCrudo is presented as a novel in both its subtitle and the publisher’s disclaimer at the beginning of the book: “Crudo is a work of fiction. Incidents, dialogue, and characters, with the exception of certain public and historical figures, are products of the author’s imagination or are used fictitiously.” The frequency of such disclaimers in contemporary literature is itself symptomatic, indicating the presence not just of fiction but of fictionalization. A person used fictitiously is not quite a fictional character, but a figure that draws us into the realm of the counterfactual (see Gallagher 2018; Prendergast 2019). The novel’s opening establishes a very particular kind of novelistic frame:

Kathy, by which I mean I, was getting married. Kathy, by which I mean I, had just got off a plane from New York. It was 19:45 on 13 May 2017. She’d been upgraded to business, she was feeling fancy, she bought two bottles of duty-free champagne in orange boxes, that was the kind of person she was going to be from now on. Kathy was met at the airport by the man she was living with, soon to become the man she was going to marry, soon, presumably, to become the man she had married and so on till death. (Laing 2018a, 1)

What displacement of meaning and self does the first sentence enact, exactly? Who does “I” refer to? Does Kathy stand in for Laing, the extradiegetic author? Does “she” refer to Kathy or Laing? Or does “she” stand in for “I,” in an enunciative distancing of the intradiegetic narrator, complete with its own form of self-ironizing free indirect discourse (“that was the kind of person she was going to be from now on”)?

The book’s following pages complicate things further as we learn that Kathy’s lover will break up with her because he does not think two writers should be together: “Kathy had written several books—Great Expectations, Blood and Guts in High School, I expect you’ve heard of them. The man with whom she was sleeping had not written any books. Kathy was angry. I mean I. I was angry” (1–2). Serving as a kind of referential “punctum” (Barthes 2002, 5:809) the book titles pierce through the fiction to link the character-narrator with Kathy Acker, the late author of Great Expectations (1982) and Blood and Guts in High School (1984). For the hesitant reader, the phrase, “I expect you’ve heard of them,” serves as a signpost of factuality (to adapt Cohn’s term). Or more precisely, it is a signpost of counterfactuality, which throws a possible version of Kathy Acker, still alive, into a post-Brexit Britain where she lives a life that in some way resembles that of Olivia Laing. None of this play with the authorial/narratorial persona or with referential elements entirely dissolves the novelistic frame: the I retains the identity of a fictional Kathy or counterfactualKathy Acker. Still, it is not surprising that Crudo has been read as autofiction, despite Laing’s own dislike of the term (Laing 2018b). Thisreception is conditioned by a number of factors, not least the opening’s dramatic staging of the “deictic displacement” often associated with fiction (Hamburger 1973, 127–128). When foregrounded in this way, the signpost of fictionality becomes a signpost of fictionalization, only accentuating the referential uncertainty surrounding the first-person narrator and thereby encouraging an autofictional reading, regardless of the author’s stated intention.

Autofiction: A Novel

Cusk’s and Laing’s books both proclaim themselves to be novels, but so does Édouard Louis’s primarily autobiographical narrative. The genre designation on its own—the paratextual indication “a novel” on the book’s cover or title page, often appearing as a subtitle—has become an extremely fragile signpost of fictionality, even as it still functions as a marker of literary prestige. Detachment from direct autobiographical enunciation, in the cases of Cusk’s trilogy and Laing’sCrudo, still hinges on the proper name, whether it is discreet or insistent in its presence (Cusk’s self-effacing “Faye” versus Laing’s “Kathy”). In French-language literature, in particular, the indication roman very frequently coexists with a broadly referential pact and produces an ambivalent reception—including when these works cross into other languages, and especially into an Anglophone literary market characterized by a clearer-cut fiction/nonfiction distinction.

Another of these designated “novels,” Christophe Boltanski’sLa Cache (2015; translated as The Safe House [2017]) is an autobiographical story of the Boltanski family, told via a description of spaces in the family home, including the “in-between”—the tiny storage room where the author’s grandfather Étienne Boltanski hid out during the war (2017, 159). Boltanski’s tale of his grandparents’ home is organized around this secret place, the tiny safe house within the house. It is also a story about names: proper names that have been altered, rejected, or reinvented, and are no longer entirely “proper” (122) bear the problematic trace of an obscured origin, of the path that led the “Bolts” from Russia to France. We should note here that “Boltanski” is now a household name in French intellectual and artistic life: Christophe’s uncle is the artist Christian Boltanski, his father the sociologist Luc Boltanski. But Boltanski (Christophe) explores what is hidden behind the public life of his eccentric family. In any case, this kind of investigation of names and identities (tied in this case to the condition of the Jewishdiaspora in Europe) does not in itself undermine the autobiographical contract founded on the proper name.

Where, then, are the sites of fictionalization in Boltanski’s account? The book’s structure is non-linear, organized around spaces (rooms, but also the family car) treated as extensions of characters; it delves into family mythologies. This approach might be contrasted with the novels of the author’s grandmother, who, publishing under the penname “Annie Lauran,” “advocated a ‘tape-recorder literature,’ which depended on strictly cataloguing real life” (9). Boltanski, for his part, adopts a mode of writing that does not appear strictly documentary but adapts techniques sometimes associated with fiction. Consider the following use of “penetrative” narrative devices that Cohn (1999, 16) claims are “unavailable to narrators who aim for referential (nonfictional) presentation”:

She didn’t so much want to rediscover her youth (Rediscover what? The abandonment she’d been a victim of? Her godmother? Polio? The war?) or thwart old age. She wanted to escape time altogether. No beginning, no end. No path sewn with obstacles to avoid. She wanted to be ageless. A state that was neither tender nor ungrateful nor green nor ripe. Not canonical, but undetermined or absent. She would have liked to float in a vague space. Eternally in between. (Boltanski 2017, 146)

The reader faced with this passage has different options. One is to read it as the site of a local fictionalization within a globally factual narrative, attributing the thoughts of the grandmother (Marie-Élise/Myriam/Annie Lauran) to the imaginative speculation of the author. Another option is to infer that the account is based on family archives, conversation, and oral tradition, but that this documentary-testimonial basis remains implicit in the text. The focalization remains ambiguous, between inner and outer. The presentation of the grandmother’s mental states serves as an “immersion vector,” to borrow Jean-Marie Schaeffer’s term for “ludic feints, ludic beginnings, that the creators utilize to give birth to a fictional universe and that permit the receivers to reactivate this universe mimetically” (2010, 218). In this case, however, the “ludic feint of mental acts” (Schaeffer 2010, 219) does not serve as a gateway to a fictional universe; we do not read the book as fiction, exactly, but maintain a distinction between the fictional and the fictionalized that determines our reading of the formal indices. In any case, it is precisely the book’s referential ground that makes visible the work of fictionalization—in this instance, Boltanski’s transformation of a set of intensely private individuals and an eccentric family unit into characters whose lives we can follow.

What this example illustrates is also the difficulty of moving from local effects of fictionalization to a definition of autofiction as a genre. It may be more useful, as we have also seen with our other cases, to consider a range of strategies and textual features that can produce autofictional effects. The forms of hybridity and the relationship between global and local factuality and fictionality vary widely. Karl Ove Knausgaard’s minute descriptions of attending a children’s birthday party (2013, 21–59), eating cornflakes (56–57), or making a cup of tea (2014, 372), for instance, produce a different kind of fictional effect from Boltanski’s use of free indirect discourse. Knausgaard’sMin kamp (My Struggle) transports into self-narration the apparently superfluous details that Roland Barthes (2002, 3:25) associates with the “reality effect” in nineteenth-century literature. Used in a nonfictional, autobiographical context, however, they serve to capture the texture of daily experience and to cultivate a deliberate form of egalitarian, undifferentiated attention (see Lerner 2014). This hyperrealism appears as both acutely phenomenological and highly artificial, in that such minute recording evidently exceeds the capacities of memory. This is fictionalized reconstruction, deployed in the service of an overall autobiographical project.

Camille Laurens employs fictionalization to quite different effect in Dans ces bras-là (2000; translated as In His Arms [2004]), which integrates self-exploration into a fictional frame by staging a set of confessional dialogues with a psychoanalyst. The direct expression of the fictional pact takes the ambiguousform of a hypothetical account of a future book, narrated by someone who resembles the author but is not identical to her: “I wouldn’t be the woman in the book” (2004, 7). Is it the author or the narrator who imagines the (this?) book? This paradoxical assertion perhaps brings us as close as we can come to an overtly autofictional pact. It also creates an enunciative distance that is in tension with the intimate thematic content of the book. Indeed, Laurens’s strategies of fictionalization are in part a form of self-protection: Laurens herself describes them as attempts to bypass censorship in its various forms (from self-censorship to lawsuits for invasion of privacyFootnote 4), as well as a means of allowing the reader a margin of manoeuver (2007, 224–225). Fictionalization appears here as a strategy of subterfuge, allowing Laurens to explore the limits of the sayable while evading some of the risks and commitments of truth-telling.

The above analyses can only scratch the surface of the range of practices and effects that shape the territory of the autofictional. Still, I hope to have shown where we can begin to locate the fictional in autobiographical writing, or, conversely, the factual in fictional first-person narratives. Literary theory, as Gasparini notes (2008, 246), has historically had difficulty tackling generichybridity, but it can nevertheless supply us with tools for taking on this task. Theories of fiction and fictionality can help us avoid the pitfalls of a panfictionalist position that would ultimately deny any specificity to the autofictional (by identifying all narrative with fiction), and they give us tools for analyzing the textual strategies that produce particular autofictional effects. Autofictional texts, in turn, can shed light on key debates in the theory of fiction, for instance, in their complex play with forms of deictic displacement and other signposts of fictionality. By presenting us with a wide range of configurations of the fact/fiction relationship, these works make the nature and affordances of fictionality visible, and they illustrate the multiple forms that literary hybridity can take. In this sense, it is perhaps not possible to define a single kind of hybrid that would constitute the genre of autofiction. Broadly speaking, however, autofictional texts present us not with autonomous fictional worlds, but with sites of fictionalization where the referential ground of the I is maintained to a greater or lesser extent. As a corollary, different forms and degrees of fictionality are present at different moments or within particular aspects of the work, and produce specific effects. The autofictional is also a complex phenomenon of reception. Autofictional texts offer a salutary challenge to literary theory by highlighting the divide between ontological theories and pragmatic approaches, or between communicational and non-communicational theories of fictional discourse—reopening, for instance, the question of the distinction between fiction and feint. Finally, far from erasing all boundaries, these works bring new attention to the interactions of the factual and the fictional. While autofictional writing sometimes provokes epistemic anxiety and even moral condemnation in the so-called post-truth era, its contemporary forms can also be read as a response to this moment, revealing authors’ heightened awareness of the stakes of both fiction and truth-telling.