Literature, we know, does not progress – it merely changes; but literary criticism can, and does, progress, through the development of new theory and methodology. Nothing illustrates this truth more strikingly than the criticism of narrative literature, historically considered. Until comparatively recently most of the theory in this area of literary research derived from a treatise primarily concerned with dramatic representation – Aristotle’s Poetics. Patristic and medieval commentators on the Bible developed a sophisticated machinery for interpreting narrative, but not for analysing it. Neoclassical criticism was obsessively concerned with trying to extrapolate from Aristotle and other classical sources a set of rules for composing epic poems. The novel arose in a complete vacuum of critical terminology appropriate to it. Not until the end of the nineteenth century did some self-conscious and experimental novelists – pre-eminently Henry James in the English literary world – stimulate by their practice and precept a serious and systematic study of realistic narrative. The poetics of the emergent New Criticism of England and America, however, was heavily biased towards lyric poetry and verse drama; and the major modern breakthrough in the theory of narrative remained virtually unknown to English-speaking students of the novel until comparatively recently.

This breakthrough was the distinction established by the Russian Formalists in the early decades of this century, between fabula and sjuzet: between the ‘what’ and the ‘how’ of narrative. It sounds like the familiar distinction between content and form; but isn’t quite the same, and avoids most of the aesthetic traps and fallacies of that opposition. It is more like the distinction between deep and surface structure in generative grammar. Fabula is the story as it might have been enacted in real time and space, the story conceptualised in the most neutral, objective possible way; sjuzet is the transformation of that hypothetical ‘kernel’ story in a particular discourse, with all the distortions, gaps and emphases that are inevitable in such a process, whether they are consciously willed or not. By examining and classifying the differences between fabula and sjuzet, and the range of options open to the narrative artist in this respect, a true ‘poetics of fiction’ becomes possible.

English and American critics were alerted to the importance of Russian Formalist work on narrative primarily through the mediation of Parisian structuralism in the 1960s. A key figure here was Tzvetan Todorov, an emigre Bulgarian poetician who translated the Russian Formalists into French, and made his own original contributions to narratology in such books as Grammaire du Décaméron (1969), Introduction à la littérature Fantastique (1970), and the essays collected in Poetique de la Prose (1971). Another key – figure was of course Roland Barthes, who for a time enthusiastically pursued a semiological approach to narrative, culminating in that remarkable critique of the ‘classic realist text’, S/Z (1970).

Gérard Genette, another star of the Parisian nouvelle critique, has more kinship with Todorov than with Barthes. Like Todorov, he is essentially a formalist, and a rhetorician; he aspires to the cool intelligibility and objectivity of the science of linguistics, and eschews the ideological passion and the exciting, intimidating, aphoristic style of Barthes. But on the other hand, Genette is less abstract, less clinical, and more empirical than Todorov. He likes to work with texts and is a brilliant explicator of them. He is not interested in theory for its own sake, but as a tool to understanding and appreciation. One would not wish to set Genette against either Todorov or Barthes, for his work needs theirs to complement it (especially with regard to the narrative codes of suspense, enigma and connotation); but of all the major figures of the nouvelle critique, he is closest in style and spirit to the Anglo-American new critical tradition (which he knows well) and the most accessible to English and American readers – though the last to be translated.

Narrative Discourse was originally published in 1972, constituting the major part of Genette’s third book of criticism, Figures III. In the intervening years it has achieved classic status among critics interested in the theory of narrative, and has left its mark already on several important books in English (e.g. Seymour Chatman’s Story and Discourse, 1978 and Frank Kermode’s The Genesis of Secrecy, 1979). It expounds a method of identifying and classifying the ways in which experience is organized and represented in narrative, combined with and illustrated by a critical commentary on Proust’s A la Recherche du Temps Perdu. Any student of Proust will find the book fascinating and rewarding as a work of descriptive criticism, but I will focus attention here on Genette’s contribution to theory.

What Genette has done is to study the ways in which the sjuzet may differ from the fabula under the general headings of time, mood and voice , and then to break down these broad categories into more delicate ones. Genette does not adopt the Russian Formalist’s terms, or Todorov’s translations of them (histoire/discours) exactly, but proposes a new, tripartite scheme: histoire (corresponding to Todorov’s histoire), récit (corresponding to Todorov’s discours) and narration (the act of narrating itself). These three terms are translated here as story, narrative and narrating, respectively.

The general category of time breaks down into the aspects of order, duration and frequency. The rearrangement of the chronological order of the histoire in the récit by what Genetre calls analepsis (retrospect, flashback) and prolepsis (anticipation, flashforward) is, of course, one of the most venerable marks of literary narrative, going back at least as far as Homer, though modern fiction (e.g. Conrad, Virginia Woolf, Muriel Spark) has made particularly elaborate play with it. At any rate, it is well-known to criticism. Less well-known, or less often remarked, is the variable relationship between the putative duration of the action of the story and the duration of the reading of the narrative. This controls the pace of the narrative and (in co-operation with other aspects discussed below) the tone and thematic emphasis of the narrative. A ‘fast-moving’ story, for instance, is one in which a great many events which affect the situation of the protagonist are narrated in a short space of text. ‘Frequency’ refers to the number of times a given event in the story is narrated in the narrative. One may narrate once what happens once (e.g. Moll Flanders’ first theft); narrate n times what happened once (e.g. Lord Jim’s leap into the lifeboat); narrate n times what happened n times (e.g. sexual intercourse between Lady Chatterley and Mellors); or narrate once what happened n times (this, if combined with condensed duration, can summarise unimportant repetitions, e.g. it rained every day that week, or , if combined with stretched duration, can represent a habitual but important action, what Genette calls ‘iterative narration’, like the famous opening of the Recherche: ‘For a long time I used to go to bed early’).

Under the headings of Mood and Voice (terms which derive from classical grammar, and complement Time, or Tense) Genette treats what is known to Anglo-American criticism as ‘point of view’; but he makes a huge advance in methodology by pointing out that this phrase conceals and confuses two very different aspects of narrative: the perception of the action and the narrating of that action: the question, who sees? and the question, who speaks? These are never one and the same person except in present-tense interior monologue such as Molly Bloom’s in Ulysses. Even in ‘first-person’ pseudo-autobiographical novels like Great Expectations, the mature narrator who speaks is not identical with the naive hero who perceives, though the writer is free to open or close the gap between the narrative and the narrating instance according to his purposes. In the same way, in the classic ‘third-person’ authorial narrative the narrator may make himself more or less prominent by focalising the narrative in his own person (e.g. Tom Jones) or in the consciousness of a created character (e.g. Strether in The Ambassadors). (Genette, incidentally, quite rightly protests against the illogicality of this customary ‘first-person’/’third-person’ terminology, since all narration must be the utterance of an ‘I’, whether implied or explicit.) This opening and closing of the gap between the narrating and the narrated instance Genette calls ‘distance’, and he groups it with ‘perspective’ (the question, who sees?) under the general heading of Mood. Here his fondness for organizing his theory in tidy, quasi-metaphorical categories derived from classical philology has perhaps misled him into making a false distinction, since distance seems to be just as much a matter of voice as of perspective.

This particular question, as Genette notes, was first raised by Plato in his distinction between diagesis (pure narrative, in which the poet speaks ‘in his own voice’) and mimesis (in which the poet pretends to be speaking in the voices of a character); and one of the principal ways in which novelistic narrative varies and controls ‘distance’ is by the use of free indirect speech, a device that Roy Pascal has called, in his monograph on the subject, ‘the Dual Voice’ – a subtly equivocal mode of narrative discourse that mixes mimesis (the voice of the character) and diagesis (the voice of the narrator) sometimes inextricably .

Genette’s discussion of the Platonic distinction is among the most fascinating and thought-provoking of the theoretical passages in his book, but it must be said that his own use of the term ‘diagesis’ (and ‘diagetic’) is potentially confusing. Early in the book he explains in a note that he is going to use the term with the same meaning as ‘histoire’ or ‘story’ – ‘a use which comes to us from the theoreticians of cinematographic narrative’. In this sense diagesis is logically prior to and quite distinct from the narrative text, which is then a ‘mimesis’ of it. But later in the book, in the chapter on Mood, he distinguishes between diagesis and mimesis as two kinds of representation within the narrative text (Plato’s original distinction), and indeed acknowledges that there is no such thing as pure mimesis in written narrative (since even direct speech is transcribed rather than performed there) – only ‘degrees of diagesis’.

It is easy to see how this double application of the term, diagesis, arose: we cannot conceptualise the ‘story’ (the fabula or histoire) except as a bald summary of events in chronological order, which looks very much like Plato’s diagesis. But of course, such summaries are never truly objective or transparently referential accounts of the raw story-stuff, as Genette himself demonstrates from Plato’s diagetic rewriting of the scene between Chryses and the Achaeans in the Iliad. Any summary is always a selection and an interpretation of ‘the facts’. In cinematic theory, identification of the ‘story’ with diagesis makes more sense, because film is an inherently mimetic medium – in cinematic narrative one might say that there can be no pure diagesis, only degrees of mimesis – and a diagetic synopsis of a film narrative is clearly of a wholly different order from the film itself. To sum up, although there are valid reasons for applying the diagesis/mimesis distinction to both story v. narrative and to two modes of representation within narrative (scene v. summary, showing v. telling) it seems to me that it would cause less confusion if it were restricted to the latter use in literary criticism.

There are other minor details of Genette’s classification with which one might quibble, and with which other theorists have already quibbled, but this does not affect the essential validity and usefulness of his basic categories of order, duration, frequency, perspective, and voice. Indeed, Genette seems excessively self-deprecatory when he says in his Afterword, ‘I do not expect “posterity” to retain too large a part of these propositions.’ In the same place he observes dryly that, ‘In an area we regularly grant to intuition and empiricism, the proliferation of concepts and terms will doubtless have angered more than one reader,’ and in Britain particularly he may expect to meet such a response. It would, however, be very regrettable if British readers allowed themselves to be deterred from reading Genette by the unfamiliarity, and at times pedantic preciousness, of his technical jargon. The categories and distinctions denoted by this jargon are real and important. Their value is not merely descriptive but heuristic: that is to say, they do not merely enable us to name formal features of texts that we are aware of – they make us aware of formal features that we are likely to overlook in a simply intuitive, empirical reading. The value of such theory for literary education can hardly be exaggerated. Ask an undergraduate to write an interpretative essay on, say Heart of Darkness, and you are likely to get back either a naive misreading, or a digest of what more mature minds have made of that text. Equip the same student with Genette’s theory, and ask him to report on the relations between story, narrative and narrating in Heart of Darkness (looking at such things as the use of a frame narrative, the points at which Marlow’s narrating deviates from chronological order, the variations in duration in his narrative, the difference between his naive perspective in the narrated instance, and mature perspective in the narrating instance, etc. etc.), and to ask himself why Conrad chose to realise the story in that way, and you are much more likely to get a piece of work that is both genuinely illuminating and genuinely independent.