Narratology and translation

I will begin with a translation. Nearly 50 years ago, Claude Bremond (1964: 4–5) famously argued that, in any narrative message, there is

a layer of autonomous significance endowed with a structure that can be isolated from the whole of the message: the story [le récit] … The structure of the latter is independent of the techniques that support it. It can be transposed from one to another without losing anything of its essential properties: the subject of a tale can serve as argument for a ballet, that of a novel can be brought to stage or screen, one can recount a movie to those who have not seen it. These are words we read, images we see, gestures we decipher, but through them it is a story that we follow; and it can be the same story. The narrated [le raconté] has its distinctive significant elements, its racontants: these are not words, images, or gestures but the events, situations, and behaviors signified by words, by images, by gestures.

Bremond’s argument about the transposability of the narrated also applies, in some measure, to the narrating, which has its distinctive significant elements too: the order in which situations and events are presented, for instance, the speed at which they are depicted, the frequency with which they are described. Now, certain aspects of the narrating are not transposable from one medium to another. As Sol Worth (1975) emphasized, “pictures can’t say ain’t,” for example; they can’t say “I” or “you”; and it is not clear what a first-person narrative painting would look like, as opposed to a second- or third-person one. But certain aspects of the narrated are not (readily) transposable either. How would a picture represent a character who is maybe 145 or maybe 150 pounds or one who is 36 years old and not 35? What about character traits like beauty, diffidence, or ambition? Or what about explicit causal links between events? Besides, Bremond made sure to speak of essential properties. Perhaps some, most, or all untransposable aspects are not essential (which does not mean that they are not important). Whatever the case may be, in the domain that most concerns me here—that of cross-linguistic grammatical comparison and, more specifically, that of translation from (written) language to (written) language rather than that of semiotic transposition in general—there seems to be relatively little narrated or narrating that is not translatable, granted that translation always involves at least some nonequivalence.

In what follows, I will not focus on this or that nonequivalence in this or that translation; I will not point to mistakes and infelicities; I will not assess the benefits and drawbacks of fidelity or transparency. Rather, using mainly English and French as source and target languages, I will discuss some of the effects that translation’s inevitable nonequivalences, variances, or paraphrases have on narrating and narrated features. In other words, I will use translation to revisit these features, to reconsider their basicness, centrality, or indispensability, and to reassess the narratological models they bring about.

A good starting point is the narratological category perhaps inappropriately called “voice,” which subsumes narrator, narratee, and narration (Genette, 1980: 212–262). Perhaps because of its importance in literature, criticism, and narratology, this is probably the most studied category from the angle of translation and its troubles, as work by Susan Lanser (1999), Ina Schabert (2010), and Dennis Schofield (1998) attests. If, in any language, a narrator (whether homodiegetic or heterodiegetic) is always a first person, in many languages, different distinctions are made regarding the grammatical category person. Some languages mark first, second, and/or third persons not only for the singular and the plural but also for the dual (or even the trial and the paucal); some distinguish between masculine, feminine, and neuter; some use classifying systems that discriminate between inclusive and exclusive first person plural or between proximate and obviative third person or between two or (many) more degrees of formality and informality; and some are even said to divide the category into more than three persons. Apart from indicating that taxonomies restricted to classes like first-, second-, and third-person narratives are at best imprecise, these distinctions also more than suggest how certain narrative identities, differences, and effects can get lost (or gained) in translation. After all, it may be easier to feature a grammatically genderless narrator in English than in French, where many more adjectives have to be explicitly marked masculine or feminine, just as it may be easier in French to point to a narrator’s gender through grammatical (as distinct from merely connotative) means. As Lanser (1999: 175–176) points out:

[t]here is a moment in the first chapter of George Sand’s La Mare au Diable [The Devil’s Pool] when the masculine form of an adjective provides the sign that the narrator is male; in the English translation no such signal would be manifest. Conversely, translators of George Eliot’s Adam Bede into French would surely have found it difficult to avoid sexing the narrator of Chapter 17, who undertakes many acts of self-characterization, in such constructions as ‘I am content,’ that English does not, but French would inflect.

Similarly, Ina Schabert (2010: 78–80) shows how French renditions of Charlotte Brontë’s (or Currer Bell’s) Shirley gender the narrator masculine as soon as adjectives and past participles referring to it are used whereas, in the English version, the narrator’s masculine gender is not made explicit before the very end. She likewise points out that the French translations of The Mill on the Floss mark its narrator as male from the beginning while, on the contrary, the English original features an ambi-gendered narrative voice.

Probably more striking for the context of narratorial gendering is a text like Anne Garréta’s Sphinx (Garréta, 1986). In this French novel of desire, mourning, and loneliness, where the autodiegetic narrator and a character named A … live through an extraordinary love story, the (biologically male or female) sex and (culturally masculine or feminine) gender of both figures remain indeterminate. The point here is not that Sphinx is untranslatable into English. It very much is translatable though, to my knowledge, it has not been translated yet. The point is that what constitutes a tour de force in French may require less linguistic virtuosity in English and may thus prove less striking and less efficient in suggesting, say, that love transcends sexual and gender differences (cf. Schabert, 2010: 88).

What is true of narrators is, of course, also true of narratees (and characters or even inanimate entities). It has been said, for example, that Julio Cortázar’s story Usted se tendió a tu lado (You Lay by Your Side) was not translated into English “because of the difficulty of translating the systematic use of both plural/formal and singular/informal ‘second-person’ pronouns” (Schofield, 1998: 15). It has also been argued that the famous second person of Italo Calvino’s Se una notte d’inverno un viaggiatore (If on a Winter’s Night a Traveler) affects the reader differently in the opening passages of the original Italian version, where it is marked as masculine, and in the English version, where no such marking obtains (Kacandes, 1990: 224; Schofield, 1998: 15). It could further be pointed out that, although both Michel Butor’s La Modification (A Change of Heart) and Georges Perec’s Un Homme qui dort (A Man Asleep) are second-person narratives, one important difference between them concerning the narratee is obvious in French but not in English since Perec adopts the informal tu instead of the formal vous used by Butor. In fact, many similar grammatical variances can obtain. Imagine, for instance, a French-language narrative featuring two distinct narratees, one of whom is addressed as tu and the other as vous; or imagine a narrative in which an individual narratee and a group of other narratees are both designated by vous but are distinguished by the adjectival epithets accompanying them.

As for narration, that is, the very act of producing a narrative as signified in the text, the very act of recounting (and commenting on) a series of situations and events, it too can be affected by grammatical differences between source language and target language. Albert Camus’s L’Etranger (The Stranger, The Outsider, or perhaps The Foreigner) constitutes a well-known example. The novel features a posterior narration but, instead of using the passé simple (or preterit) as was expected for French fiction until recently, it uses the passé composé, for which the English present perfect is not quite a good tense equivalent. Instead of yielding “Today, mother has died,” the first words of its famous opening—”Aujourd’hui, maman est morte. Ou peut-être hier”—yield, among many possibilities, “Mother died today,” “Maman died today,” “Mom [or Mommy] died today,” or “Today, Maman died.” (Bloom, 2012). Once again, I do not mean to emphasize untranslatability. Granted, Camus’s choice of tense is fraught with meaning. Rather (or even more) than stressing the straightforward, informal, everyday quality of the protagonist’s narration, which paradoxically also resorts, at times, to highly formal imperfect subjunctives, the passé composé used throughout L’Etranger foregrounds the consequential character of the events narrated and their relevance for the normally detached protagonist as well as the series of similar presents in which he exists, his passivism, his indifference. But, in spite of the preterit they had to adopt, translators were able to convey (many of) these significations through various means. What was more difficult to communicate was the exceptional nature of Camus’s choice. Ironically, now that much (“literary’) French fiction is written in the passé composé, English translations of L’Etranger may seem closer than they did to what the French text has today become.

If the translating of temporal and aspectual properties of a narration can be complicated, so can the translating of modal properties. For certain contexts, the subjunctive is used in French to signify presuppositional uncertainty regarding what is taken to be the case whereas the indicative is used to signify presuppositional certainty. This particular distinction has no exact equivalent in English where both something like “Il ne pense pas que la terre soit ronde” (subjunctive) and something like “Il ne pense pas que la terre est ronde” (indicative) could translate as “He doesn’t think that the earth is round”

Likewise, the translating of metalinguistic elements can be thorny. I am thinking in particular of passages that narrators translate from one language or linguistic code to another one for the benefit of their narratees. When the heterodiegetic narrator of La Chartreuse de Parme, for example, translates his protagonist’s Italian phrases, as in “Fabrice était tellement troublé qu’il répondit en italien: —L’ho comprato poco fa (Je viens de l’acheter à l’instant),” (Stendhal, 1961: 41–42) or when the homodiegetic narrator of The Sun Also Rises says “I had taken six seats for all the fights. Three of them were barreras, the first row at the ring-side, and three were sobrepuertos, seats with wooden backs, half-way up the amphitheatre” (Hemingway, 1954: 161), the translation of these two passages into English and French respectively is not especially problematic. But things might well prove different with an Italian translation of Stendhal or a Spanish translation of Hemingway; and things would also prove different for an English translation if Fabrice answered in English and for a French translation if Jake Barnes glossed French terms.

Metalinguistic glosses sometimes precede the terms they explain, as in “Weirdo, I said, that’s what hurluberlu means” or in “il lui parla de traitrise, de treachery.” Usually, however, they follow them more or less immediately and they can occur in a footnote or in the main body of the text. Often, no metalinguistic interpretation is provided even though linguistic code switching occurs. Toward the end of Lawrence Durrell’s Justine, for instance, the text goes from English to French—“A pity you are not in the habit of shedding tears. Ça aurait été un soulagement”—with italics signaling the switch in languages (Durrell, 1961: 241). Italics are also used in the French version (“Quel dommage que vous n’ayez pas l’habitude de verser des larmes! Ç’aurait été un soulagement”) but a footnote affecting the reader’s immersion is added for clarification: “En français dans le texte.” (Durrell, 1968: 401).

Like metalinguistic discourse, attributive discourse—which accompanies a character’s discourse and specifies the act of the speaker or thinker, identifies the latter, and indicates various dimensions or features of the act, the character, the setting in which they appear, and so on—attributive discourse can present translation difficulties when the source text is English and the target text French. Whereas intercalated or incidental tag clauses in English can follow the Subject-Verb order (“she said,” “John answered”) or the Verb-Subject one (“said she,” “answered John”) and can thus give different emphases to act and agent, they are restricted to the latter order in (standard) French.

Some fundamental narrative features other than voice, such as the order of presentation of events and the achronies or anachronies, the prolepses or analepses it can involve, are hardly affected by translation, if at all. However, some features, such as that of narrative speed, are more sensitive to linguistic peculiarities or to changes caused by them in other features. Since speed is a function of the relationship between the duration of the narrated—the (approximate) time the situations and events presented go on or are thought to go on—and the length of the text narrating these situations and events (in words, lines, or pages, for example), it is affected by most, not to say all, translations. But the changes are usually not very significant narratively. Still, should metalinguistic specifications, say, appear in one version but not in another (where they would be superfluous), alterations in speed might prove more significant and, consequently, so might alterations in rhythm. Obviously, the presence or absence of such specifications can also have an effect on the distance between narrator and narratee (or narrator and character, or narratee and character) and, like other linguistic elements, can modify the tone of the narrative, making it more or less plodding, for example, pedantic, or recherché.

Compared to categories involving the narrating, categories strictly pertaining to the narrated are perhaps less likely to be impacted by translation. However linguistically intricate the representation of an event may be, for instance, it is not clear that its translation would have specifically narrative repercussions. Certainly, the connotations associated with any number of representational elements do vary from language to language and, along with them, so do themes and thematic links. But these are far from constituting distinctively narrative (or narratological) features. The situation may be a little different in the case of existents where, as mentioned earlier, differences in grammatical gendering can lead to translational predicaments and where other linguistic features too can prove translationally troublesome. For example, an inanimate object gendered masculine (or feminine) in a French text and representing an actant, playing an actantial role, constituting an actor, would be neuter in English and this would modify gender configurations. Similarly, since connotations attached to characters’ utterances, say, are part of characterization, the latter can be modified by translation.

Some of the translation troubles I noted confirm what has long been known, namely that story (the narrated, the “what”) is not only more transposable but also more translatable than discourse (the narrating, the “way”). Indeed, they clarify the opposition between a view of narrative as primarily a verbal mode, a narrator’s recounting of a series of events in words, and a view of it as a verbal or nonverbal representational arrangement of events, a configuration, a frame, a schema. The former view underlines linguistic specificity and artfulness. It gestures toward untranslatability. The latter view emphasizes cognitive (and affective) capacity. It espouses translatability. A narratological model mindful of both views and aiming to account for semiotic, modal, and configurational commonalities and differences should therefore make it clear that the depiction of events cohering into a story (configured into a story) can be inherently without narrator (and narratee) or else can be told, recounted, narrated by a narrator that can go from maximal overtness—when the narrating supplants the narrated—to maximal covertness (and disappearance). Such a model would capture the difference between the essential inapplicability and the merely possible absence of a certain feature or category. Such a model may also encourage a reconsideration of other domain-pertinent problems, such as the narrative status of dreams, musings, memories, or various apprehensions of reality. More generally, it may encourage a revisiting of the difference between symbolization and representation as well as emphasize the (possible) necessity for a narrative to be a reproducible deployment of meaning. A dream may be narratable but this does not mean that it is a narrative. The same is true of a baseball game or any number of activities effectively occurring in the real word: they can be represented (in words, say, or on television) but they are not necessarily representations. Furthermore, such a model would promote the demimetization of narratology no doubt required by the ever larger number of narratives which multiply ontological contradictions, temporal inconsistencies, or psychological impossibilities (cf. Richardson, 2006). Last but by no means least, such a model would not only entail changes in the status and characterization of other narratological features and categories (like various aspects of voice, for instance, various indications of causal relationships, or, to mention an aspect of narrative that narratology has neglected and that translation calls attention to, various marks of narrational focus); it would stimulate the further exploration of the links between the narratorial act and narrativity as well as those between the latter and anthropomorphization. To what extent does the materialization of a narrator (and narratee) in a narrative affect the latter’s narrativeness, for example, and to what extent is narrativeness a function of experientiality (cf. Fludernik, 1996; Prince, 1999, 2008b)?

As I have noted in passing, along with confirming or clarifying basic narratological assumptions and positions, translation and the linguistic reflections, comparisons, or transfers evoked by it promote a richer and more precise view of possible narratological properties and narrative kinds: first-, second-, or third-person narrative, of course, but also male, female, ungendered, or ambigendered narrator, singular, dual, or trial narratee, proximate or obviative narration, and so on and so forth. By the same token, such reflections and comparisons nourish the rhetorical force of narratology. Narratological models aim to characterize narratively all and only possible narratives, however improbable or uninteresting these may be. They thus constitute a rich reservoir of (old and new) narrative potentialities for restless narratologists or aspiring storytellers. Translation, as suggested by my brief discussion of narratees, for instance, can help theory invent practice by uncovering, foregrounding, or adding to these potentialities (cf. Genette, 1988: 157).

Besides contributing to the enrichment and characterization of the set of narratological properties or the set of narratives these properties inform, translation can shed light on the essentialness or importance of the latter. I have already noted that, when it comes to narrative painting and however important the distinction may be for verbal narrative, it does not make much sense to speak of and distinguish between first-, second-, and third-person narrations. In other words, these narratological categories refer to features that do not seem necessary for the constitution of narratives. Similarly, the fact that certain aspectual inflections, say, or certain facets of attributive discourse do not quite survive translation might mean that, whatever their significance, they are not as narratively essential as other features. Now, I have argued in various texts and contexts that narratologists of all persuasions should attempt to found their discipline empirically and experimentally (Prince, 2003, 2008a). We have all too often taken for universally true statements conventional or locally (very locally) suggestive and persuasive claims about narrative functioning and, more particularly, about the centrality or importance of this or that narrative feature. We should instead ground narratology through empirical or experimental (cross-cultural and cross-media) study if we want to account for what is the case. Of course, that type of study itself presents a number of difficulties. It is not easy, for instance, to find or devise (laboratory) specimens free from the crippling disease of clumsiness or weirdness. Perhaps versions of the same texts in different languages can serve as (or constitute the basis for) specimens and can thus help to shed light on the adequacy of various claims and the weight or significance of various features.

I have also frequently argued that narratologists should make room for pragmatics in their investigations (Prince, 1983, 1995). In addition to a morpho-syntactic component describing the macro- and microstructures of all and only sequences of narrated situations and events, a semantic component interpreting these structures, and a narrating component operating on the interpreted structures and characterizing so-called narrative discourse, narratological models should include a pragmatic component specifying (some of the) cognitive, affective, and communicative factors which impact the processing and narrativity of the output of the first three parts. For example, this last component might indicate how the same text can be read as including certain narrative elements or, on the contrary, as not including them.

Consider that many narratives are narratively ambiguous. In fact, the confusion between post hoc and propter hoc, which Roland Barthes (1975: 248) regarded as “the mainspring of the narrative activity,” is not the only confusion to operate in or arise from narrative texts. For example, free indirect discourse and narratized discourse are sometimes difficult to distinguish, as in “Against the light, the resemblance was clear to him. She was really the spitting image of Mary.” Free indirect discourse can also sometimes be indistinguishable from interior monologue just as internal focalization can be indistinguishable from zero focalization (as in “They went to the party together. She was quickly overcome by the noise, the lights, the number of people”) or just as iterative narrative can be indistinguishable from singulative narrative (as in “He looked for her under the rubble without listening to his daughter who cried to him: ‘Be careful!’”) or else just as scene can be indistinguishable from stretch or perhaps even from summary, heterodiegetic from homodiegetic, and so on. Some of these ambiguities are no doubt partly due to the fact that narrative features have not always been defined with sufficient precision or to the fact that they have inherently ‘fuzzy’ or ‘relative’ rather than ‘exact’ or ‘absolute’ values. For instance, even when not calculated in terms of our experience and view of the real world, certain tempos in a narrative—summary, say, or stretch—must still be measured in terms of and in relation to other tempos in the narrative itself. But some ambiguities may result from certain other factors that translation can help us to pinpoint. Given my example of the “daughter who cried to him: ‘Be careful!’,” say, a translation into French would yield “sa fille qui lui cria: ‘Fais attention!’” (singulative preterit) or the equally valid “sa fille qui lui criait: ‘Fais attention!’” (iterative imperfect). Narratologists must therefore not only or not so much refine certain features in order to make them more precise; they must also make it plain that, in some cases, a single (set of) feature(s) unambiguously accounts for a particular configuration but that, in other cases, one of two (or more sets of) features can properly be invoked. We should empirically determine the features whose use would yield the same forms. Moreover, we should allow for that phenomenon in our narratological models. We should make explicit room in them for what I call the ‘voice of the reader’ (the switches that can orient a text one way or another, the possible interpretive decisions that can determine the particular direction a text will take), for example by specifying the potential of alternative operations on given structural arrangements to produce identical outputs.

Translation also points to another significant pragmatic factor affecting narrative texts, their meaningfulness, their tellability or reportability. The latter is partly a function of their unusualness (or that of one of their elements), which is why news stories quickly stop being reportable, why—as translation underlines—the force of the passé composé in Camus’s L’Etranger can vary, and why Anne Garréta’s Sphinx is probably more arresting in French than it would be in English. Perhaps an unusualness factor, to be activated or deactivated depending on circumstance, could be attached to every constituent of our narrative models.

There are undoubtedly other ways in which translation and the textual or contextual factors it grapples with or evokes—grammatical particularities, narratological obscurities, verbal ambiguities, historical processes at the individual or collective level—there are other ways in which translation can spotlight various aspects of narrative and its functioning. Indeed, translation may well constitute one of the most powerful tools at our disposal as narratologists. If every reader and every writer is a translator, if transactions always involve translations, if we are all beings in translation, as has become increasingly clear (cf. Brodzki, 2007), it is only normal that our narratologies too are in translation.