Tense in Fiction: Narrative Mode vs. Speech Mode

The twentieth century began with a growing resentment of the author’s guiding presence in every sentence, the enveloping of every fact with comment. Authorial guidance was given in narrative sentences, as opposed to speech or thought by characters: narrative mode versus speech mode, each with a different grammar. We tend to forget that epic and romance and the early modern novel consisted overwhelmingly of narrative sentences and little dialogue, a preference for telling over showing. Open any book by Defoe or Swift, and you get pages of uninterrupted narrative sentences that would make a novel unreadable today. With Sterne and Fielding we get more speech forms, not only from characters but in playful author addresses to the reader. Then with Jane Austen dialogue elegantly enters, but it develops slowly. Her characters are fixed psychologically and socially through speech. With Henry James comes the claimed preference for showing over telling, but in fact the narrative sentence, more and more complex, still largely dominates.

The novel has been making a huge effort to shake off the authoritarianism of the traditional narrative mode. At first, one solution was to take even the narrative sentences right inside the character as free indirect discourse; another was to get away from that and allow far more direct speech mode, either inside the character (interior monologue, as in Joyce) or as dialogue with outside viewpoint but no comment (as in the ‘anti-Modernists’ Hemingway or Waugh, as opposed to James or Woolf). An extreme version of that was the novel wholly or almost wholly in dialogue (as in examples by Henry Green or Ivy Compton-Burnett), but this, as Nathalie Sarraute later commented, was pushing the novel into the theater, where it is bound to be inferior. However, some indication of who was speaking, and how, was thought minimally necessary, and continued late. For example, in 1963 in Iris Murdoch’s The Unicorn: ‘She thought, her heart sinking a little’. These parentheticals, as linguists call them, and their desperate variants, already parodied by Joyce in Ulysses, were later attacked by Sarraute as ‘symbols of the old régime’.

There had been many attempts to get free of this past tense narrative sentence. In nineteenth-century fiction, brief passages in the historic present are used for vivid descriptive non-action scenes before safe return to the past; and the present tense is favored for universalizing moral or social comments from the author. Or, much easier, the handing over of narrative to a character inside the story who, unlike Robinson Crusoe and other early I-narrators, are allowed to pass from the classic narrative sentence to the present and other tenses of speech mode. Already Dostoyevsky’s Notes from Underground (1864) is a non-stop monologue by the hero. So are diary forms, and, earlier, epistolary forms (still predominantly in narrative mode, however). And, of course, the varyingly rhythmed (chopped, telegraphic, unpunctuated) interior monologues of Stephen, Bloom, and Molly in Ulysses (1922). As a narrative tense, however, the present was first experimented in 1874 by Flaubert in La Tentation de Saint Antoine and in 1887 by Edouard Dujardin in Les lauriers sont coupés, then by Joyce, first very occasionally as historic presents for sudden vigor in Ulysses, then coming into its own in Finnegans Wake (but here Joyce purposely blurs the very distinction I am talking about).

In 1966, a structuralist linguist, Emile Benveniste, clarified the distinction between histoire and discours (‘history’ and ‘speech’). So as not to create confusion with Genette’s later use of discours for ‘treatment’ and récit for ‘story,’ from the Russian Formalist fabula and syuzhet (the what and the how, the raw material of a story and the way a story is organized), I have here preferred ‘speech’ versus ‘narrative’. Benveniste’s distinction is entirely about the how, two types of telling, one for history, which uses a particular type of impersonal sentence to tell of events, and the other for the speech system (addressing the reader or, in the novel as in life, characters and people addressing each other). Here is Benveniste on the narrative sentence used for history: ‘Nobody speaks here, events seem to narrate themselves. The fundamental tense is the aorist, which is the time of events outside the person of a narrator’. If the aorist really expresses ‘the time of events outside the person of a narrator,’ this would explain the death of the once imposed passé simple in speech—and a possible difference of attitude in English and other languages, where the past tense can and does also belong to the speech system. There is no dead aorist.

The main point here, however, is that the ‘historical’ past tense is speakerless, and—as historical narrative also ignores audience—it is not only speakerless but addressee-less. In other words, the narrative sentence of historical narrative is outside the personal speech system and cannot use its tenses. Nor can it use its deictics: it has to say ‘then’ for ‘now,’ ‘there’ for ‘here,’ ‘the next day’ for ‘tomorrow,’ ‘the day before’ for ‘yesterday,’ ‘the following year’ for ‘next year,’ and so on; it should even avoid deictic verbs, which situate the speaker in space (e.g., ‘came’ versus ‘went’). The tenses of speech mode are the present, the present perfect, and the future. The French imperfect, which in English is the past progressive (was going) or iterative (would [often] go), seems to belong to both historical narrative and the personal speech system. Hence Proust could use the imperfect for the dead aorist throughout his long narrative without causing much shock or even notice. In English, as the past tense is also part of the speech system, there is no dead aorist, as already noted.

By the 1930s the author’s control was felt as omniscient and godlike. It was felt in the use of the distancing past tense and its distancing deictics, in the distancing third person even when plunged inside a consciousness (James or Woolf); it was felt even in the outside viewpoint with minimum comment and more dialogue, not only with the parentheticals that went with dialogue, but also in indirect discourse, which is a summary of what a character actually said, as well as in free indirect discourse, which I’ll come to in a moment. The last two devices (indirect discourse and free indirect discourse) vanished, more or less, in the nouveau roman, with its brand-new, startling present tense and its equally startling variations. It is not the use of the present tense as such that is original, but how it is used. Let us take four examples from the fifties (in order of ‘newness’ rather than date):

1. BECKETT, MOLLOY (1950/55) OPENING:

I am in my mother’s room. It’s I who live there now; don’t know how I got there.

2. DURAS, LE SQUARE (1955) OPENING:

L’enfant arriva tranquillement du fond du square et se planta devant la jeune fille.
‘J’ai faim,’ déclara-t-il.
Ce fut pour l’homme l’occasion d’engager la conversation.
‘C’est vrai que c’est l’heure de goûter,’ dit-il.
La jeune fille ne se formalisa pas. Au contraire, elle lui adressa un sourire de sympathie.

3. SARRAUTE, LE PLANÉTARIUM (1959) OPENING:

Non vraiment, on aurait beau chercher, on ne pourrait rien trouver à redire, c’est parfait… une vraie surprise, une chance… Une merveille contre ce mur beige aux reflets dorés…

4. ROBBE-GRILLET, LES GOMMES (1953) OPENING:

Dans la pénombre de la salle de café le patron dispose les tables et les chaises, les cendriers, les siphons d’eau gazeuse; il est six heures du matin.
Il n’a pas besoin de voir clair, il ne sait même pas ce qu’il fait. […] Bientôt malheureusement le temps ne sera plus le maître. Enveloppés de leur cerne d’erreur et de doute, les événements de cette journée, si minimes qu’ils puissent être, vont dans quelques instants commencer leur besogne. […] Mais il est encore trop tôt. […] l’unique personnage présent en scène n’a pas encore recouvré son existence propre.

Beckett, in this particular matter of tense, has dropped the past narrative sentence for the conventional ‘pronouned’ speech form, as in, say, Dostoyevsky’s Notes from Underground or Céline’s Voyage au bout de la nuit. This is the only ‘legitimate’ use of the term ‘narrator’. Duras is still using the traditional past tense narrative sentence to introduce the man and girl, but once the child is gone, taking with him the quotation marks and the parentheticals, the rest is in dialogue (designated with mere dashes), therefore in conventional speech forms. Duras, in other words, still uses the convention of past tense for narrative. But she will soon adopt the present tense, if not always dropping the parentheticals but emphasizing them, for example, in La Maladie de la mort (1982): Et puis elle demande: Vous voulez quoi? Vous dites que vous voulez essayer [indirect speech for several lines]. Elle demande: Essayer quoi? Vous dites: D’aimer. And in most of her later novels the present tense is used vividly, dramatically, to increase the emotion, and sometimes omnisciently.

With Sarraute, we are plunged into speech forms (in various tenses) but inside the consciousness of someone. As in Robbe-Grillet we do not know whose mind we are in, there is no je (in the opening), but (as in James or Woolf) that mind is represented by the third person (elle fait un grand effort, elle capte, elle tire), getting more and more excited, gushing internally, but not narrating. Robbe-Grillet, however, uses the present tense of speech form throughout the novel, paradoxically to replace its antagonist, the authoritative but speakerless omniscient past narrative sentence. In this he has been much imitated, but he will add two further paradoxes in subsequent novels. Here, however, he is giving the supposedly speakerless speaker the same omniscience as the traditional author, in both present and future. These differences, then, in these early examples, may be gropings against the authorial authority, difficult to jettison, still there. But each will develop in various and specific ways. Here I am only concerned with Robbe-Grillet and his paradoxical use of the present tense to replace the past tense narrative sentence. His second novel, Le Voyeur, makes an interesting alternation of past tenses (passé simple and imperfect) for ‘real’ events and the present tense for ‘unreal’ ones, namely fantasies (as had Flaubert for St. Anthony’s temptations).

With La Jalousie he will go further than Camus or his own contemporaries in two respects. First, and paradoxically, by using the present tense of the speech system and its deictics, but with all the impersonal speakerless tone of the past-tense narrative sentence. That is, not just for momentary vividness as in the historic present; nor in direct address to the reader, like Sterne or Fielding; nor as a playful ‘look-at-me,’ like the Postmoderns; nor as inner speech or sous-conversations, like Sarraute; nor even with occasional omniscience, as here. Rather, he uses the present tense as a ‘scientific’ present tense. This is clearly derived from cinema. Robbe-Grillet was originally an engineer and must have felt the similarity between the speakerless past-tense narrative sentence and the scientific use of the present tense. Since the traditional narrative sentence had become less and less impersonal and speakerless, he must have decided to replace it with a speakerless, ‘scientific’ present. He simply puts down objectively and in detail all that hits his central consciousness (‘protagonist’).

In the second paradox, Robbe-Grillet limits this new present-tense narrative sentence to one conscience, unlike the traditional past-tense narrative sentence, which is far freer to go where it pleases. It is like a scientific sentence limited to one result at the time of utterance, yet—and this is his great subtlety—constantly varied with other possibilities, even at times, corrections. The regard evokes a detective rather than a scientist: To the right, a simple shape, more blurred, already covered by several days’ deposit, is nevertheless still discernible; at a certain angle it acquires sufficient clarity for its outline to be followed without too much difficulty. It is a kind of cross: an elongated main shape, like a table knife but wider (La Jalousie, tr. Richard Howard).

The third paradox is that he never evokes an act of seeing or a consciousness, that is, there is no seer, only the seen, no énonciation (‘I was convinced,’ ‘surely,’ ‘it seemed,’ etc.), only énoncé. This is why I used the phrase ‘all that hits his central consciousness’ above, rather than ‘all that his central consciousness observes.’ There is no act of observation. Moreover, he dispenses with the first or third or indeed any personal pronoun normally attached to the present tense, just as science can. For instance, in La Jalousie, we never know who the jealous person is whose objectivized registrations we follow: ‘She takes a few steps into the room, goes over to the heavy chest and opens its top drawer. She shifts the papers in the rear of the drawer, leans over and, in order to see the rear of the drawer better, pulls it a little further out of the chest.’ Such a text can give two readings of the jealous character: as an observer of detached, scientific objectivity, or as an obsessed character, the latter chiefly because of the detail.

This startling change of narrative technique may well have been foreshadowed in French by Proust, Céline, Camus, and Beckett, but not in that paradoxical mix (the present tense used in narrative sentences in which no one speaks). All four use the present tense with its conventional first-person pronoun, so that we’re wholly in speech mode. Flaubert comes the closest to using the present tense as narrative sentence, yet he reserves it for hallucinations. Sarraute uses it as a ‘silent’ speech form rather than as narrative sentence, expanding her sous-conversations through more and more emotion and ending up (as she had said of novels in dialogue) in the theater. Duras too, who also made films and wrote scenarios, uses it as a narrative sentence, as does Robbe-Grillet, but a narrative sentence emotively as highly charged as Sarraute’s inner dialogue. Indeed, in Sarraute’s later novels we hear our old friend the omniscient author telling us things the character can’t know (such as ‘she can’t hear’), but they are told in the present tense, which thus seems used for greater immediacy and emotivity, rather than for ‘scientific’ objectivity. Only Robbe-Grillett, in La Jalousie, has fully achieved a form that I shall call the scientific present tense.

This classic question of who speaks led to a strange dispute, during the seventies and eighties, between certain linguists, notably Ann Banfield, and a few literary critics who held on to the traditional view that there are two voices in free indirect discourse and were quite unable to give up author intrusion, so merely changed the critical word ‘author’ to ‘narrator’. Free indirect speech, the paradoxical intrusion of expressive elements from direct discourse (speech) into a past narrative sentence, is unique to modern fiction. It was invented in the eighteenth century [Goethe, Wilhelm Meister’s Apprenticeship, 1795; a proto-free indirect speech], developed brilliantly by Jane Austen, and culminated in E. M. Forster, Flaubert, Woolf. It can even be part of a dialogue. Here are two typical examples from E. M. Forster’s A Room with a View:

Example 1: A conversation then ensued, not on unfamiliar lines. Miss Bartlett was, after all, a wee bit tired, and thought they had better spend the morning settling in; unless Lucy would rather like to go out? Lucy would rather like to go out, as it was her first day in Florence, but, of course, she could go alone. Miss Bartlett would not allow this. Of course she would accompany Lucy everywhere. Oh, certainly not; Lucy would stop with her cousin. Oh no! that would never do! Oh yes!

Here is my own favorite example, from Jane Austen’s Pride and Prejudice. Mrs. Bennett is bidding farewell to Mr. Bingley.

‘Next time you call,’ said she, ‘I hope we shall be more lucky.’ [direct discourse]
He should be particularly happy at any time, &c &c, and if she would give him leave, would take an early opportunity of waiting on them. [free indirect discourse]
‘Can you come tomorrow?’ [direct discourse]
Yes, he had no engagement at all for tomorrow [free indirect discourse]; and her invitation was accepted with alacrity. [narrativized discourse]

Such swift changes of register are frequent in Austen. But who says ‘&c &c’? The two-voiced theory supporters would hear narrator irony about polite formulas, but in Banfield’s theory (she would use linguistic argumentation), it would represent the character’s own awareness, at a non-verbalized, semi-conscious level (formulas he could add but does not, or formulas he is adding and still uttering, while thinking ‘&c &c’), but which we are not given.

Much more often, however, there are no such deictics, and the sentence is formally just like a straight narrative sentence. Here is a famously ambiguous example from Jane Austen’s Emma:

He [Frank Churchill] stopped and rose again, and seemed quite embarrassed [narrative sentence]. He was more in love with her than Emma had supposed.

That last sentence reads as a narrative sentence like the first, coming from the author (or narrator in modern parlance) and therefore, by convention, is true within the fiction. But on second reading, when we know that Emma was gravely mistaken, we can only read it as Emma’s thought, in free indirect discourse.

Ann Banfield has rigorously analyzed this kind of free indirect discourse, which she prefers to call ‘represented speech and thought’. I’ll do the same in this section, since the term is more accurate, not only because it can represent both speech and thought, but also because this type of sentence does not ‘imitate’ speech or thought but ‘represents’—re-presents—it. This is an important distinction. Although Banfield starts from a slightly revised version of Benveniste’s distinction between histoire and discours, her analysis is not structuralist, for she bases her argument on generative grammar. Her book is called Unspeakable Sentences (1982), because both narrative sentences (where no-one speaks) and represented-speech-and-thought sentences are specific to the narrative mode, whereas direct and indirect discourse can occur in the personal speech system. These two types of sentence, the narrative sentence and the sentence of represented speech and thought, are both speakerless, literally unspeakable, in the speech mode of the personal system.

Banfield’s book is a rigorous attempt to answer the traditional view of represented speech and thought as representing two voices, that of the character and that of a so-called narrator. This view has to attribute different bits of the sentence to one or the other. She cites many literary examples given by holders of that traditional view, demolishing them one by one. For instance, we read, in Emma Bovary’s consciousness: ‘She loved the sick lamb, the sacred heart pierced with sharp thorns, or poor Jesus falling as he carries the cross.’ The traditional view has always read this kind of sentence as author/narrator irony about pious sentimentality. But for Banfield, we are inside Emma’s mind, but she is not consciously ‘saying these words’ to herself; they ‘re-present’ her vague thoughts and feelings. Banfield blames the modern silencing of the author for these various double-voiced formulations and for introducing ‘the notion of a narrator as a created persona distinct from the author alongside the notion of point of view,’ which includes evaluation. The author, she says, is ‘expunged’ but ‘reintroduced as narrator who is responsible for all the sentences in the text, as the speaker is in discourse.’ She defends an alternative theory, which sees author and narrator as distinct constructs of literary theory where ‘narrator’ is restricted to cases when the author does create a narrator, namely the first-person narrative.

Banfield insists that even as a construct, the author is not represented in the text but can only be inferred through interpretation. Most of Banfield’s book is devoted to syntax, but she also deals with her opponents’ semantic arguments, which posit either an ‘empathy’ of the narrator or an ironic commentary. On empathy, which is linked to ambiguous represented speech and thought sentences, she invokes a consistency criterion: the facts of the fiction must be consistent with one another. For example, the sentence about Frank Churchill being in love with Emma can’t come from an undramatized narrator, nor can the character’s value judgments. As for irony, the classic case for the dual voice theory, she insists that ‘one can argue that represented speech and thought is the main vehicle for irony without thereby implying that this is an argument for the dual-voice position. For the ironic reading of a sentence cannot be treated as part of its meaning, in any accepted sense of linguistic meaning’. The two meanings of an ambiguous sentence have exactly the same syntactic and semantic structure. With irony we are beyond grammatical or even semantic rules. The irony is provided by the reader (which is why it can also be missed). Irony is a way of reading. ‘There is nothing to which the notion “narrator’s point of view” can be semantically attached that is not equivalent to the character’s point of view’.

The present tense is also difficult to use for more generic reasons: the present is the predominant tense of lyric poetry. It is apt for lyrical description, less so for fictional action, which has for so long depended on the ‘guarantee’ of the past narrative sentence, so that the use of the present for action recalls those brief historic presents for vivid, non-action scenes in the classical novel. The present tense is also the tense of general statements and universal questions, the novelistic equivalent of ‘scientific statements,’ such as authors loved and still love to make in their intrusions, which can sound either dogmatic or banal. This is true at least in English, which, unlike French, also has the continuous present for simultaneous action and so tends to reserve the normal present tense for generalization. Consequently, young writers who think, a little behind the times, ‘Ah, the present, that’s the latest thing,’ and then merely replace the past in the traditional narrative sentence by the present, achieve, to my mind, little more than the critics who have merely replaced the word ‘author’ by the word ‘narrator’. Indeed, the present tense nowadays often reads like a mechanical replacement of the past narrative sentence, regardless of true effect. It is often reduced to mere vividness, like the historic present, and so still coddles the reader, wrapping every fact in the safety of its ‘cool’ reassurance.