Alain Resnais and Alain Robbe-Grillet: The implementation of a modern project
Alain Resnais and Dephine Seyrig 1961
To appreciate the modernity of Last Year in Marienbad, we need to look first at the two “authors” who collaborated on its making: Alain Robbe-Grillet, the screenwriter, and Alain Resnais, the director. The film is effectively the setting for a dialogue between literary modernity and cinematic modernity. When Alain Robbe-Grillet wrote the screenplay for Marienbad, he had already experimented with a highly innovative style in his novels Les Gommes (1953), Le Voyeur (1955) and La Jalousie (1957) and was firmly associated with the nouveau roman (“New Novel”), a term applied to the works of a small group of French writers who sought to overturn the literary foundations of the traditional novel in the 1950s. The nouveau roman authors rejected plot and character (in the sense of an entity determined in threefold fashion by physical, psychological and social factors) in order to focus on the formal dimension of writing. This aesthetic break was to be the means of translating a new vision of the world and saw the stream of consciousness of an often anonymous “I”, plunged into stupor in the face of a reality deprived of meaning, replaced by sequences of extremely precise descriptions.
It should be noted that Marguerite Duras and Jean Cayrol, like Robbe-Grillet, are also writers linked with the nouveau roman. We know that Duras wrote the screenplay for Hiroshima mon amour and Cayrol that for Muriel ou le temps d’un retour: the modernity of Resnais’ first feature films was thus closely bound up with literary modernity of their epoch. This point is important in order to understand what Resnais invents with Marienbad: part of the film adheres to the “literary contraption” constructed by Robbe-Grillet, but the director’s peculiar genius consists of drawing sustenance from the text in order better to detach himself from it and of pursuing a highly original investigation into the powers of cinematic representation.
The “literary contraption”
Let us turn briefly to the premise of the film. It is, as has often been said, the “story of a persuading”. We are in a spa resort, in a grand, luxury hotel peopled by lifeless characters. A man attempts to convince a woman that they met last year, in a very similar hotel, perhaps in Marienbad. She denies it and seems to remember nothing. The man therefore starts telling her their story. He relates snatches and fragments, which become more elaborate as the film proceeds. The man checks and corrects himself; different, sometimes contradictory versions succeed each other. Gradually the young woman’s recollection seems to improve and she starts to accept the story she is being told and to introduce her own modifications. But now that the story is being fleshed out, we discover that it corresponds to the events that are in the process of unfolding before our eyes, with the result that, at the end of the film, we are back at “last year” and the entire story seems doomed to repeat itself.
The plot is based on a technique of literary description used by Robbe-Grillet, which he presents as follows: it is “a description that starts from nothing; it does not start by offering a general overview, but may arise out of a fragment of no importance – something that most resembles a point – from which it invents lines, planes, an architecture. Our impression that it is inventing them is only reinforced when it suddenly contradicts, repeats or checks itself, bifurcates etc. Yet we start to glimpse something and we believe this something is going to become clear. But the lines of the drawing accumulate, overload and negate each other and change place, with the result that the image is cast into doubt even as it is taking shape. A few more paragraphs and, when the description comes to an end, we realise that it has left nothing behind it: it has taken place in a double movement of creation and erasure – a movement we find, moreover, in the book at all levels and in particular in its global structure.” We can see how much the composition of the Marienbad screenplay as a whole owes to the movement of Robbe-Grillet’s writing. It was naturally necessary to find an equivalent for this form of literary description when it came to the film adaption. The writer imagined widely varying solutions, ranging from the time-lag between the visual narration and the soundtrack (the pictures we see contradict what the man is saying off-screen) to the progressive and imperceptible modification of the décor (the young woman’s bedroom is bare at the start of the film, but gradually fills up with increasingly luxurious and ornate furnishings).
In a fascinating essay, Mireille Latil-Le Dantec seeks to identify the points of alignment and divergences that exist between the work of Resnais and that of Robbe-Grillet. She notes that if “Marienbad unquestionably takes up the labyrinthine and repetitive ‘double circuit’, with its return to the point of departure, found in Les Gommes, Le Voyeur and Dans le labyrinthe”, the film can also be seen as “close to the litany of forgetting of Hiroshima mon amour, to the disorderly path towards a confession dragged from a certain resistance, to the progressive structuring through words of a vague, ambiguous event that cannot be verified other than through a vision”. In this sense, Marienbad seems a continuation both of Robbe-Grillet’s areas of enquiry and those of Resnais. Toute la mémoire du monde (1956), Resnais’ documentary about the Bibliothèque Nationale, may thus be considered a draft for Marienbad: here we already find the long tracking shots that follow labyrinthine corridors, in which moving through space becomes a journey through the meanderings of memory and time.
It should be stressed that Resnais was also involved in the task of writing the Marienbad screenplay, which began with a long discussion between the two men. As Robbe-Grillet has said: “We began by speaking of things like glances that are directed off-screen, what linkages are needed when relations of causality are altogether uncertain, ambiguity, the obscure nature of the slightest amorous adventure. And we found we agreed on all that. The question of defining an anecdote was something for later: the important thing was in the telling. As long as the kinds of form were agreed on, we’d be able to think up the subject.” On the basis of this first conversation, Robbe-Grillet produced not a traditional scenario but what was effectively a shooting script, in which the dialogue was accompanied by a detailed, shot-by-shot description accompanied by indications regarding camera movement and soundtrack. The scriptwriter thus sought to think the work directly in cinematic terms. This choice is logical given that he assigned greater importance to its form (“the telling”) than to its storyline. We should also remember that Robbe-Grillet was at that time preparing to make his own films (he was working on L’Immortelle, which came out in 1963). After Robbe-Grillet presented the finished script to Resnais, the two men discussed it at length. Resnais then proceeded to work alone: he shot and edited the film by himself and only showed it to Robbe-Grillet in its final cut.
When the film was released, Resnais and Robbe-Grillet repeatedly insisted that each agreed wholeheartedly with what the other had done. In fact, as one might imagine, the reality was a little different, and the two men subsequently spoke of several areas of dissension. According to Resnais, Robbe-Grillet accused him of having added “psychology” and of not having followed all his instructions regarding the music and the use of sound in general. As his ciné-novel makes clear, the screenwriter had conceived a universe of sound careful constructed around the noises of the hotel (and their distortion, their time-lag vis-à-vis the picture, etc.), an avenue that was ultimately little explored by the film. Robbe-Grillet was not at all happy with the music chosen by Resnais, which he found too sentimental. The writer had pictured something much more strident, aggressive and discontinuous. We may imagine, too, that Robbe-Grillet – who at one point had suggested shooting the script in the Paris metro –strongly disliked the sumptuous and almost languorous aspect of the final film. Although, on the surface, Marienbad is a very cold work, certain scenes possess a poignant dimension, notably thanks to the acting of Delphine Seyrig. The misty, pathos-laden atmosphere in which Marienbad is steeped is completely absent from the novels of Robbe-Grillet and from his films.
To understand the difference between Robbe-Grillet’s artistic project and that of Resnais, it is necessary to look at the question of time. Latil-Le Dantec links the repetitive “double circuit” of Robbe-Grillet’s first films with the pattern, recurrent in Resnais, of a “repetition that fails”. “Muriel”, she notes, “is ‘the time of a return’, i.e. the vain attempt to go back, with the same characters and in the same settings, to an earlier situation. La Guerre est finie  is the crossing of a border ‘one more time’... [in] Je t’aime je t’aime , the one opportunity for an exact repetition (one minute of… last year) offered by science fiction nevertheless results – due to the failure of the experiment – in [the protagonist’s] restless wandering”. From this point of view, in Resnais as in Robbe-Grillet, “the present moment denies continuity”: we are on the crest of a perpetual here-and-now, which abolishes all possibility of inscribing ourselves in permanence. For Robbe-Grillet, there is something in this that opens the door to the “exhilaration of invention”: repetition is “failure of the search for meaning”, but also “force of becoming and dynamism of meaning”. By contrast, repetition in Resnais corresponds to a vain and desperate attempt to avoid erasure by time; his characters are painfully aware that the present devours and swallows everything.
The repercussions of these divergent visions of Marienbad’s mise en scène can be demonstrated by comparing the end of the film and that of Robbe-Grillet’s ciné-novel. The differences are many; we shall only mention those we consider the most significant. First of all, the location: in the ciné-novel, when the lovers (X and A) meet in order to leave for good, they are in one of the hotel lounges. Resnais’ choice is more dramatic: A is waiting at the foot of a curving staircase and X descends the steps to join her. Another difference: in the ciné-novel, A is described as having a travel bag which she picks up when the lovers get up and leave. In the film, the bag has disappeared – A’s departure is thus swathed in unreality. Robbe-Grillet had envisaged a cut with a fade to black before the characters exited the shot. The film, by contrast, shows them walking away towards the far end of a vast hall and disappearing behind one of the elements of the décor.
The script also envisaged M, upon his arrival after the lovers have left, performing various innocuous actions before returning to his room. Resnais, on the contrary, shows M starting to descend the stairs (like X before him), but stopping to survey the empty hall from the top of the steps. This elevated position gives him the appearance of a master of ceremonies. M does not move, but an axial cut carries us forward and frames him in close-up. This choice is decisive: the close-up confers upon this moment an emotional charge which is absent from the ciné-novel (because it invites us to share the character’s sense of loss). It means, above all, that the sequence closes with the persistence of a gaze. When we “exit” the hotel (via an outdoor shot found in the ciné-novel as in the film), we are not faced with an empty, abandoned scene; someone is watching. In his ciné-novel, Robbe-Grillet imagines the last shot being filmed with the dolly rolling out. The viewer was to have the impression of retreating backwards, because the story was over. In the film, by contrast, the hotel is filmed in a stationary shot and forms an imposing black mass that stands out against a sky striped with clouds. In this very expressive, pathos-laden shot, the hotel resembles a haunted castle (a few isolated lights glow in the darkness). This shot confirms our feeling that the characters have never really left; that they are prisoners in an enchanted setting, and that they are phantoms condemned to re-enact the same story over and over again under the enigmatic gaze of M, the holder of a secret.
The divergent endings of the ciné-novel and the film show that the two authors each held a different artistic perspective. In Resnais “the fiction doesn’t end with itself”, whereas it is the opposite for Robbe-Grillet, who explains: “The universe in which the whole film unfolds is […] a world without a past, which is wholly self-contained and which gradually fades. This man and this woman begin to exist only when they appear on the screen for the first time. Before that they are nothing, and – once the film is over – they are once again nothing. Their existence only lasts as long as the film lasts.” Seen in this light, we might think that the modernity of the film is less radical than that of Robbe-Grillet’s text.
The film and the French “New Wave”
To understand Marienbad’s reception by the public and the critics, it is necessary to situate the work in the very specific historical context of the explosion of the French nouvelle vague (“New Wave”). The film came out in 1961, in the aftermath of the first feature films by Claude Chabrol (Les Cousins and Le Beau Serge, both 1959), François Truffaut (Les 400 coups, 1959) and Jean-Luc Godard (À bout de souffle, 1960). These three young film-makers were part of a close-knit group: they had all been critics for the magazine Cahiers du cinéma, where they made no secret of their desire to introduce a genuine aesthetic break into French cinema. From the start, their artistic ambitions went hand in hand with a genuine rethinking of the means of production. They wanted to encourage the freedom and spontaneity of filming by shooting with small teams and on a low budget. The quest of the nouvelle vague was in the first place a quest for authenticity, in which the burning sincerity of the author was matched by a type of realism greatly inspired by the documentary approach (anchored in contemporary life, shot outdoors and in natural settings, using hand-held cameras, etc.) The renewal of forms only took place at a second stage. This characterisation of the nouvelle vague is evidenced in particular by the first films by Éric Rohmer (Le Signe du Lion, 1962), Jacques Rozier (Adieu Philippine, 1962) and Agnès Varda (Cléo de 5 à 7, 1962). From this point of view, Resnais was a little different right from the start.
Hiroshima mon amour was made with a specific budget and its extremely impressive mise en scène plainly departs from the assumed amateurism that hallmarked the early works of Rohmer, Godard and Truffaut. But the film came out in June 1959, that is to say, just after the Cannes triumph of 400 coups, and this historical coincidence proved crucial. The modernity of the film, both in its treatment of time and in its editing, aroused the unbounded admiration of the Cahiers du cinéma’s contributors, who linked the film to their common goal of artistic revolution. It is interesting to review briefly what the Cahiers critics said about Hiroshima. At a round-table discussion, Jacques Rivette offered his highly perceptive thoughts on the film’s editing, which in his view translated “the sense of the fragmentation of the basic unit: the world has shattered, it is fragmented into a series of tiny pieces, and it is a matter of putting the jigsaw back together again.” Rohmer declared Resnais to be a “Cubist” and also compared Hiroshima to the American novel (as represented by William Faulkner and John Dos Passos). In order to describe (and legitimise) the cinematic modernity of the film, in other words, the Cahiers team of writers inscribed it within the lineage of literary modernity and pictorial modernity.
The reception enjoyed by Hiroshima allows us to understand why Marienbad was considered a major work of contemporary cinema. In Marienbad we find the interweaving of different periods in time already present in Hiroshima mon amour, along with the same use of repetition and fragmentation. We also find Resnais again employed his celebrated “leopard cuts”: these consist of a match on action between two shots of the same actor taken in different settings. Sylvie Baudrot, the continuity girl on Marienbad, thus revealed that a long scene of dialogue, in which Delphine Seyrig and Giorgio Albertazzi are walking side by side down one of the hotel corridors, was in fact shot in three different locations (even though the sequence is intended to give an effect of continuity in terms of dialogue and rhythm). Baudrot explained that the dolly out was a montage of three takes, one filmed at Nymphenburg palace, the second at Schleissheim palace, and the last in the studio. She added that the montage cuts had only been crudely masked, because Resnais wanted viewers to see that there were three different corridors. From a formal point of view, the director here ventures something very bold and modern that makes absolutely no appearance in the script written by Robbe-Grillet.
Suzanne Liandrat-Guigues and Jean-Louis Leutrat ask, however, whether the critical reception given to Marienbad, which focused on the modernity of the film, has not given rise to a misunderstanding. They underline the fact that the mise en scène imagined by Resnais possesses a very timeless dimension. The film-maker has stated that the film has the air of taking place in the 1930s – adding, by way of emphasis, “I say ‘has the air’.” The entire visual style of Marienbad was designed to evoke a certain golden age of classic cinema. Sacha Vierny, the director of photography, explains in particular that Seyrig’s face, her silhouette and the way in which she was lit, made reference to the movies of the end of the silent era. According to the first assistant director, Jean Léon, “Resnais wanted to rediscover the style of the grand productions of the old days, of those fabulous costumes covered in feathers, boas, the final era of the silents and the beginning of the talkies. He’d shown Bernard Evein photos of L’Herbier films like L’Inhumaine and L’Argent, the costumes for which were created by the great fashion designers.”
Robbe-Grillet’s text perhaps took Resnais back to his earliest influences – Surrealist literature, Mandrake the Magician and the hypnotic dimension of silent cinema. What was “the story of a persuading” became the story of a bewitching. It is also possible that Resnais wished to temper the extremely radical modernity of Robbe-Grillet’s project via the classicism of his own mise en scène. At the end of the day, the difference existing between the vision of the writer and that of the director proved highly productive, since it fuelled the indecipherable dimension of Marienbad. Do the characters have an existence outside the double movement of creation and erasure in which the film’s “story” unfolds? Should we keep our distance and consider the work as a cold, cerebral game, or allow ourselves to be captivated by the light, the reflections, the trembling looks and the murmuring voices?
The articulation of space and time
When all is said and done, the true point of convergence between Resnais and Robbe-Grillet lies in the way in which Marienbad overturns the articulation of time and space that had, until that point, formed the basis of cinematic narration. In order to characterise this work of subversion in detail, it is interesting to consider the analyses proposed by Gilles Deleuze in The Time-Image. The philosopher writes that, in Resnais, “the image no longer has space and movement as its primary characters but topology and time.” For Deleuze, in classic films action and sensory-motor links (situation-action, action-reaction, excitation-response) prevail and are served by an editing that connects movements and thus restores an indirect image of time. With the crisis of the action-image, the time-image appears: the situation no longer prolongs itself in action but becomes a purely optical and auditory situation, which “shows” (the characters have become “seers”). This time-image “has subordinated movement. It is this reversal which means that time is no longer the measure of movement but movement is the perspective of time: it constitutes a whole new cinema of time, with a new conception and new forms of montage”.
In Hiroshima, Marienbad and Muriel, we are present at the coexistence of different sheets of time: it becomes impossible to distinguish between the past and the present because they are in constant interaction. To describe this type of structure, Deleuze draws upon models supplied by topology. This branch of mathematics is sometime called rubber-sheet geometry, because it studies invariant properties under the effect of biunivocal continuous transformations, that is to say the continuous deformation of figures, among other things. Topology, because it is bound up with notions of continuity and limits, with qualitative properties and with the relative positions of geometric entities, allows us to rethink the phenomena of transformation, generation, contamination, dilation and contraction, capture and liberation. Following Deleuze’s intuition, topology enables us describe the form of Marienbad and analyse the way in which the film subverts the coordinates of cinematic narration.
In the film, a certain number of elements (the statuary group, the photograph of A, the game of Nim, the smashed glass, the broken heel…) function as organising centres. They contain potential virtualities which are going to become real in the space-time created by the unfolding of the story. The work develops by means of successive slidings, with each repeated element progressing towards other zones, from neighbouring locality to neighbouring locality, by contamination. In topological terms, we may speak of the formation of continuums. The different elements aggregate and give rise to narrative segments, shot and reshot in unceasing transformations. The topological model demonstrates the dynamic, shifting dimension of Marienbad’s structure: it allows us to picture the organisation of the story in terms of relocations, to view the relationships between the different parts of the film in a spatial perspective. The sheets of time coexist and interact with each other, serving to subvert the logic of succession imposed by the running of the film.
This movement, in Marienbad, has its origins in the “literary contraption” put in place by Robbe-Grillet’s shooting script, but it is subtly extended by the match cuts imagined by Resnais. The connections which establish themselves between one picture and the next constantly bring back into play the articulation between continuity and discontinuity on which traditional editing is based. Let us thus return to the start of the film and a shot in which the camera shows us A, plunged into shadow, while off-screen we hear the voice of X saying “Vous êtes toujours la même…” (“You’re still the same”). In the next shot, the dialogue is continued by a couple in front of a lift, but the woman is no longer A and the man has a different voice to that heard off-screen in the preceding shot. At the end of their conversation, the couple passes in front of A. The camera, which is following them on a dolly, stops in front of her, still in the shadows. There follows a match cut in which A is dressed differently and is this time seen in full light. The off-screen voice of Albertazzi repeats: “Vous êtes toujours la même…” Later, A and X are leaning on a balustrade outside in the park and looking at each other, smiling. The match cut with the following shot shows X standing in front of the balustrade. But he has changed his suit (he is now in a dinner jacket) and we wonder if we are still in the same location (is it the same balustrade, or another somewhere else in the park?) These mismatches allow “switch-overs” not only between spatial settings but also between times. We pass from one location to another not via movement and action, but because two takes present similarities that allow them to be joined into the same continuum.
The move from one place to another is made not by following the characters’ journey but according to a pure editing logic: there is no more trajectory, no more action, but simply links and connections. This is also true of relocation in time: the length of time which could have built up out of the accumulation of movements is annulled by cuts, which make us jump from one sheet of time to another.
According to Robert Benayoun, at the end of their first meeting Resnais and Robbe-Grillet had agreed “upon the principle of an examination of non-narrative ‘cinematic forms’ based on a certain indecipherable quality”. Can we speak, in the case of Marienbad, of non-narrative forms? It seems to me that the significance of Robbe-Grillet and Resnais’ film lies in the fact that there is indeed narration, but a narration that is subverted, prevented and unbalanced by the powers of editing, opening onto a new thinking on relationships of space and time.
All three works were published by Les Éditions de Minuit (Paris). They have been translated into English as The Erasers (1964), The Voyeur (1958) and Jealousy (1959).
The core members of this group included (in addition to Alain Robbe-Grillet) Jean Ricardou, Michel Butor, Nathalie Sarraute, Claude Ollier, Robert Pinget and Claude Simon.
These two films, directed by Resnais, were released in 1959 and 1963 respectively and fall chronologically speaking either side of Last Year in Marienbad.
“…l’histoire d’une persuasion”. As stated by Robbe-Grillet in the introduction to L’Année dernière à Marienbad: ciné-roman (Paris 1961). Cited here from the English translation, Last Year in Marienbad: a Ciné-Novel, translated by Richard Howard, London 1962, p. 9.
Alain Robbe-Grillet, Pour un nouveau roman, Paris 2013 (first published 1963), p. 160.
Mireille Latil-Le Dantec, “Notes sur la fiction et l’imaginaire chez Resnais et Robbe-Grillet”, in Études cinématographiques, nos. 100–103, Alain Resnais et Alain Robbe-Grillet, évolution d’une écriture, Michel Estève (ed.), Paris 1974, pp. 120–121.
Interview with Nicole Zand for France-Observateur, cited by Jean-Louis Leutrat in L’Année dernière à Marienbad, London 2000, p. 18.
Jean-Daniel Roob, Alain Resnais: qui êtes-vous?, Lyon 1986, p. 116.
It should be noted that in Marienbad this atmosphere is due in large part to the lighting, an aspect which Robbe-Grillet did not take into consideration at all when he wrote his shooting script.
Latil-Le Dantec, op. cit., p. 122.
Robbe-Grillet, Pour un nouveau roman, op. cit., p. 168.
Latil-Le Dantec, op. cit., pp. 130–131.
Ibid., p. 123.
In the ciné-novel the woman is called A, the man who is pursuing her, X, and the man from whom she is separating, M.
Resnais, it should be recalled, has always professed a taste for the supernatural.
Latil-Le Dantec, op. cit., p. 128.
Robbe-Grillet, Pour un nouveau roman, op. cit., p. 165.
À bout de souffle, for example, goes much further than Les 400 coups, by introducing bold editing, exploiting the discontinuity of the soundtrack and instigating a very loose and scattered temporality.
Jean Dormachi, Jacques Doniol-Valcroze, Jean-Luc Godard, Pierre Kast, Jacques Rivette, Éric Rohmer, “Hiroshima, notre amour”, Cahiers du cinéma, no. 97, July 1959, p. 4.
These cuts were invented with Henri Colpi; cf. “Le raccord léopard”, extract from a letter by Henri Colpi published in Cahiers du cinéma, no. 699, April 2014, p. 39.
Anecdote related by François Thomas in L’Atelier d’Alain Resnais, Paris 1989, p. 157.
Suzanne Liandrat-Guigues and Jean-Louis Leutrat, Alain Resnais, Liaisons secrètes, accords vagabonds, Paris 2006, pp. 16–17.
Ibid., p. 57.
Emma Wilson, Alain Resnais, Manchester 2006, p. 74.
Leutrat, op. cit., p. 62. L’Inhumaine and L’Argent date from 1924 and 1928 respectively.
The hero of a comic strip created by Phil Davies, Mandrake the Magician triumphs over his adversaries by means of hypnosis.
Gilles Deleuze, The Time-Image, translated by Hugh Tomlinson and Robert Galeta, 1989, p. 125 (first published as L’Image-temps, Paris 1985).
These analyses form the conclusion to L’Image-mouvement (devoted to films made before the Second World War) and the point of departure of L’Image-temps. Cf. Gilles Deleuze, L’Image-mouvement, Paris 1983.
Deleuze, The Time-Image, op. cit., p. 22.
Ibid., p. 119.
bid., p. 169.
This argument is developed by Claude-Paul Bruter in his interdisciplinary work Topologie et Perception, vol. 1, Bases philosophiques et mathématiques, Paris 1985.
I pursue this topological reading in greater depth in my book Alain Resnais, une lecture topologique, Paris 2000.
Bruter, op. cit., p. 40.
Robert Benayoun, Alain Resnais, arpenteur de l’imaginaire, Paris 2008 (first published by Éditions Stock 1980), pp. 85–86.