Jacques Rivette (1974)
Interviewed by Carlos Clarens, Edgardo Cozarinsky

With your three most recent films -- L'Amour fou, Out 1: and Celine et Julie vont en bateau -- one feels that you are becoming progressively more and more interested in the actors.

JACQUES RIVETTE: When I began making films my point of view was that of a cinephile, so my ideas about what I wanted to do were abstract. Then, after the experience of my first two films, I realized I had taken the wrong direction as regards methods of shooting. The cinema ofmise en scene, where everything is carefully preplanned and where you try to ensure that what is seen on the screen corresponds as closely as possible to your original plan, was not a method in which I felt at ease or worked well. What bothered me from the outset, after I had finally managed to finish Paris Nous Appartient with all its tribulations, was what the characters said, the words they used. I had written the dialogue beforehand with my co-writer Jean Gruault (though I was ninety per cent responsible) and then it was reworked and pruned during shooting, as the film otherwise would have run four-and-a-half hours. The actors sometimes changed a word here and there, as always happens in films, but basically the dialogue was what I had written -- and I found it a source of intense embarrassment. So much so that when I began work on La Religieuse, which was a project that took quite a while to get off the ground, I determined this time to use what was basically a pre-existing text.

La Religieuse is your only film adapted from a literary source -- and a classic into the bargain.

What attracted us to the Diderot novel was that the text is highly wrought and very theatrical. Diderot wrote it two years after his two major plays, Le Pere de Famille and Le Fils Naturel, and it's completely bound up with those plays. In its sensibility, its use of language, it reveals the same theatrical perspective. In his letters, Diderot speaks of scenes, views, images in connection with La Religieuse, all of which encouraged us to make the film. In Diderot's mind it was something very theatrical, very visual, and in effect it is composed of big scenes. We only added some five little scenes to serve as transitions; our main task was pruning, because if we had retained all the material in the novel, the film would have run five hours. Actually, our first adaptation of La Religieuse would have made a three-hour film. There were several successive stages. Jean Gruault had been an actor, though without any real vocation for it. He wanted to write plays, and during the editing of Paris Nous Appartient he brought me the first draft of a stage adaptation of La Religieuse. 'La Religieuse is really a play,' he told me. And, 'while I was writing the play, I was thinking it would be even better as a film.' We staged a production at the Studio des Champs Elysees, with Anna Karina. It wasn't so much a play as a sort of compromise, a bastard enterprise based on the script which the two of us had written in the meantime.

In retrospect, La Religieuse looks like a transitional work. You are still using post-synchronization in it, for instance.

Direct sound was impossible for that film. I'm very fond of accents, but with Francisco Rabal you couldn't understand a word. La ReIigieuse again dissatisfied me when I'd finished it. The project had been so long in preparation that even during shooting, when we finally got that far, I found myself irritated as well as disillusioned by the whole thing. The text had been so written and rewritten and revised, and I knew it so thoroughly by heart, that I realized I was no longer really listening to the actors. Changing anything during shooting was out of the question: we had done all the condensing possible during the process of adaptation, and the text had become the Bible! And there were material problems, because this was a pseudo-expensive film, a pseudo-big production: once the actors and technicians had been taken care of, there wasn't very much money left to make the picture. Since I found myself inattentive to the production and not listening to the actors, I realized this was not a method of filming that interested me.

After La Religieuse, I made a programme for the television series 'Cineastes de Notre Temps'. Jean Renoir, Le Patron was very simply shot, because at the time Renoir was already beginning to have difficulty in getting about. After two or three days shooting, I rea1ized that the best thing to do was to sit down facing him and ask questions, and then show films. So the filming was super-simple and the fifteen day schedule was one long conversation -- not just the part we filmed, but our talks in between times during the days we spent in Provence, in the village where his father is buried, and in the place where he filmed La Regle du Jeu. Afterwards came three months of editing with Jean Eustache, during which we looked at all the material again and at quite a few films, trying to find an approach which would be -- I'm not talking about results but the method -- roughly speaking Renoir's own or Jean Rouch's.

That was when I began to want to try the experiment of L'Amour Fou. The film began as a sort of gamble by Georges de Beauregard. He needed to make films to keep his production company going, and one day he asked if I had anything in mind that wouldn't cost too much. I suggested L'Amour Fou, and there was no argument because it was something that could be shot in five weeks with a small cast and a minimum crew. Afterwards, he wasn't too pleased to be presented with a film two or three times the length he wanted. So we didn't do him much good, but I was delighted to have been able to make it.

From L' Amour Fou onwards, the script seems to be just a formality for you, something that can be shown to a producer or to the Centre du Cinema.

Not only for me; a lot of directors are moving in the same direction. This has become more evident over the last ten years or so, because there are now cameras enabling you to do things which were possible in the silent days but were practically impossible, unless you had the energy of a Renoir, during the first thirty years of the sound film. There is a persistent idea of a cinema partitioned off in tiers: first you look for a subject, then you write as detailed a script as possible, on the basis of which you find someone to put up the money, for which purpose you pencil in the names of certain actors opposite fully defined characters. Once you have got all the elements together, often compromising some of your original ideas in the process, comes another stage: the actual shooting. You shoot little bits here and there, as meticulously as possible, and then you stick them together, and you're pleased if you end up with something that corresponds to what was described more or less in your two hundred type-written pages. Personally I find all this a dreadful bore. I don't say that you can't make great films like this: Alain Resnais does and he seems perfectly at ease within this formula. But the system stifles me; in my idea of cinema, all the stages should be totally interacting.

I want to return, though with quite different methods, aims and end-products, to the old Dziga Vertov idea: that the montage should be conceived with the project and not merely with the exposed film. This may sound like a conceit, but you could say that the script is written in the montage, and that the montage is established before shooting. I have been going in this direction for three films now, in L'Amour Fou more than in Celine et Julie, and in Out 1 most of all. I hate to have the feeling, either during the shooting or the editing, that everything is fixed and nothing can be changed. I reject the word 'script' entirely -- at any rate in the usual sense. I prefer the old usage -- usually scenario -- which it had in the Commedia dell'Arte, meaning an outline or scheme: it implies a dynamism, a number of ideas and principles from which one can set out to find the best possible approach to the filming. I now prefer my shooting schedules to be as short as possible, and the editing to last as long as possible.

How did you happen on a structure so rigorous and yet so free as that of Out 1?

Originally the idea was to do four parallel feuilleton stories, linked at the beginning of each episode by still shots connecting with the other episodes, rather like the old serials. We later abandoned this idea, but in the four-hour version we made use of black-and-white stills, either as a recall or an ellipse, a connection or a pause; in some cases the stills match with moments already seen, in others with scenes to come, occasionally with sequences that have been removed from this version entirely. In the thirteen hour version, Jean-Pierre Leaud received his first message after four hours; in the four-hour version, he receives it after fifteen or twenty minutes. In the thirteen-hour version the plot got under way after four hours of documentary, both true and false, on the two theatrical groups: it was something of a documentary about the modern experimental theatre, a little Peter Brookish or Living Theatre. And there were bits about Juliet Berto and Jean-Pierre Leaud, who were the characters outside all this; you saw him distributing his envelopes, and her doing her bit of hustling. Then came the idea of connecting the four parallel stories. It was at this point that Juliet and Jean-Pierre began to function as go-betweens for the audience, each in his own way, by trying to discover the significance of scenes or sequences of events which are as yet meaningless.

The four-hour version is almost an abstract of the thirteen-hour one. There was, for instance, a second death -- that of Juliet Berto. In the shorter version this has gone, allowing the open but none the less disturbing ending, where one feels that Bulle Ogier is going towards her doom.

Out 1 really took shape only in the month before shooting began, although we had discussed it in rather a vague way for some months with the four actors (Michel Lonsdale, Juliet Berto, Bulle Ogier, Michele Moretti), and also rather more spasmodically with Leaud. We traced the path the film was to take a week before shooting began, because we needed some sort of outline to be able to plan the work and shoot the maximum footage in the minimum time. That was how we managed to shoot a thirteen-hour film in six weeks. Afterwards, for the first cut, we joined the material more or less as a first assembly; basically the thirteen hours comprise a montage, deliberately rather loose, largely retaining the improvisational aspect, sometimes even including hesitations and repetitions, which hardly survives at all in the four-hour version.

In the thirteen-hour version there were very long sequences of pure reportage about the two groups of actors, and also moments of 'letting go', particularly where the camera was concerned: ten-minute takes of the actors left entirely to themselves and cracking up rather spectacularly. It became something of a psychodrama. Of course the thirteen-hour version was edited to some extent, but always trying to retain this 'first assembly' feeling. The main impulse behind Out 1 was provided by a screening of the similar kind of montage Jean Rouch had made for his film Petit a Petit. It ran for eight hours and was totally different from the four-hour version Rouch made, theoretically for television, and the ninety-minute version shown in cinemas. I was so impressed by the original that I refused to see the shorter versions. Out of that -- with disastrous consequences, as you know -- came the thirteen-hour version of Out 1. We still don't know whether there will ever be enough money to make a print, although the negative must be in pretty good shape because it wasn't tampered with too much.

Why did you bring Balzac and L'Histoire du Treize into the film?

I hadn't even read L'Histoire du Treize at the time, but I thought it was something everybody would know, or that people would at least get the reference. Actually there is nothing sinister about the conspiracy; it's really more of a Utopian plot, the key to which is revealed by only one or two people: by Lonsdale, in the long scene by the river with Doniol-Valcroze, and to a certain extent by Francoise Fabian and Jean Bouise in their very long scene.

Leaud conjures these phantoms, rather as in Celine et Julie vont en bateau: through the incantation of letters, of words which he dissects, in analyzing the texts of the messages.

Leaud is an actor who goes all the way once you have suggested a direction to take. It was he who launched out down this track, somewhere between Balzac and Lewis Carroll. Bulle Ogier was busy with another film and just turned up to shoot her scenes. Lonsdale, on the other hand, was in attendance throughout and was interested in what was happening to the other characters when his own wasn't involved, so he had the key to the whole thing. But perhaps it was also because he is a director at heart, even though he hasn't directed anything yet.

The disparity between a rigorous overall structure and improvisation in detail, between documentary and vast conspiracy, rather suggests an unexpected reconciliation of Lang with Rossellini.

They are both film-makers I admire. In point of fact I think that Phenix, the script I wrote with Eduardo de Gregorio after Out 1 and haven't been able to film yet, is an attempt to do something halfway between Murnau and Renoir. Actually, Rossellini had a hand in the genesis of Paris Nous Appartient. Around 1955-56 he was in Paris and wanted to produce a series of films by young directors. He asked some of us from the Cahiers group to put up projects. I remember there was Godard, who didn't do anything at all; Truffaut, who presented a rough draft of what was to become Les Quatre Cents Coups; Chabrol, who was already working on the script for Le Beau Serge, which he shot in 1957; and myself, with a theme somewhat derivative of The Blackboard Jungle, about racism and students of various nationalities within the setting of the Cite Universitaire. It was an extremely phoney script, because I projected a purely external view of the Cite, never having lived there myself, and also because I was rather under the influence of a particular area of American cinema, including Richard Brooks. Rossellini tore it to pieces pitilessly but quite justifiably, and I set to work with Jean Gruault on another script which luckily wasn't filmed either but out of which, more or less, came Paris Nous Appartient, which was also concerned with students and other outsiders. After that Rossellini left for India, and nothing came of these projects; but Paris Nous Appartient may perhaps be said to derive from the experience, if only as a reaction.

In Out 1, as in Paris Nous Appartient, there is a suggestion of obsession about characters who are marginal to the action: Betty Schneider in Paris Nous Appartient, Leaud and Berto in Out 1. In both cases the setting is Paris, and both films end with an escape to nature, a nature by no means reassuring, with isolated house and water prominently featured -- the Seine in Paris Nous Appartient, the sea in Out 1.

When we arrived at the house on the beach towards the end of shooting on Out 1, we had the feeling of stepping into a horror film complete with locked room, missing key and all the rest. Suzanne Schiffman, (1) Michel Lonsdale and I arrived the evening before shooting was due to begin to consider the possibilities, and I found myself faced with this house which was so much more interesting than I had expected. It was there, too, that we found the room with two mirrors opposite each other which Bulle enters near the end. We used these mirrors in the first place because they were there; but also because by then we had realized the considerable role played by mirrors in the material we had shot. We hadn't thought about this before starting: it was one of those things that happen as you go along. Actually, the element of obsession in Paris Nous Appartient only became apparent once the film had been edited. The ironical thing is that Out 1, which is about conspiracy and obsession, was shot in an atmosphere of complete relaxation, whereas Celine et Julie, a much sunnier and more amusing film, was beset by tensions and was very hard going -- not only because it was the cameraman's first film (he did an excellent job), but because it had to be completed within a four-week schedule, twenty working days.

How did you envisage the montage of Out 1?

The first cut, resulting in the thirteen-hour version, was done with Nicole Lubtchansky. Afterwards I felt I was no longer seeing what I was watching on the editing table, and I called in Denise de Casabianca, who shut herself up alone for a fortnight with the thirteen hours to get to know them. After that I began work with her on shaping the shorter version, which actually runs four hours and twenty minutes. Of course we didn't try to make a resume of the thirteen hours, but to find material there for another film with its own rhythm and its own inner design.

While watching Celine et Julie, I was also thinking of Comedie Policiere and the work of the TSE group. (2)

Yes, of course. There was a first stage, mostly involving Juliet Berto and Dominique Labourier -- Eduardo de Gregorio too, but much less so -- in which we tried to define the characters, then the whole beginning of the film, the way they meet each other. After that came a second stage in which we tried to find the shape the action would take. Quite early on I felt I wanted to have a film-within-the-film. It's something I have often wanted to do, and I have one in Phenix too. Even before I knew what the second film would be about, I planned to use Bulle Ogier and Marie-France Pisier, so as to have a second couple not only acting this story but relating to the other pair. And there we got stuck for quite a time. At that point Eduardo began working more closely with us, innumerable ideas were thrown around, and gradually we narrowed it down to the Henry James story The Other House, which Eduardo knew and whose basic theme we used. I still haven't read it; it hasn't been translated into French. But we also borrowed from other James stories, in particular an early short story called A Romance of Certain Old Clothes. Some lines in the film are lifted direct from this story, notably in Bulle Ogier's speech about the trunk and the clothes kept stored in it during her first scene with Barbet Schroeder. Only four lines, but it amused Bulle and she wanted to say them.

During the credits for the film, 'Phantom Ladies over Paris' appears immediately after 'Celine et Julie vont en bateau', one in black on white, the other white on black, exactly like the names of the two 'pairs' of actresses. Does this imply that it is the negative of the 'positive' main title, or is it simply the name of the film-within-the-film?

When we discussed how the second film should be integrated with the first, we considered various possibilities. At one stage the idea was to go much further in fragmenting the second film, particularly in dispersing the various elements, letting the montage range freely, thus permitting a variety of different meanings. At this point, naturally, we thought of Comedie Policiere and about a writer who has been much in view since L' Annee Derniere a Marienbad (and even before): Adolfo Bioy Casares and his novel The Invention of Morel. Of course we knew all this existed -- Marienbad and the TSE and The Invention of Morel -- but we were trying to find a motif for ourselves which would be both similar to theirs and at the same time different; and we finally hit on the motif of the candy. After that everything progressed very rapidly. The candy-making and due allowances -- played the role taken by L' Histoire des Trieze in Out 1. We suddenly found the gimmick which enabled us to link all the elements -- to provide on the one hand a mechanism for pulling the film together, and on the other something to hold the spectator's attention throughout.

For the beginning of the film, particularly everything to do with Montmartre in the first quarter of an hour, we had thought of presenting a sort of illusory Paris, something like the city you saw in American films of the Fifties like An American in Paris. But we would have needed more time -- not more money, just time to look around while shooting -- so it remained more an idea than an actuality in what we were able to shoot. We still had no final title. Someone said: 'If it was an American film, it would have been called Phantom Ladies Over Paris'; it became a sort of joke with us. But when the editing was finished, I wanted it to appear on the credits too. The way it is done makes it look as though it were the title of the film-within-the-film. But if the film is shown in America, I'd like them to use the title Celine et Julie vont en bateau, although it's difficult to translate literally. Celine and Julie take a Trip suggests something to do with LSD. Take a Boat might be better, although this misses the significance in French of vont en bateau. Monter en bateau is to tell someone a rather involved story and have them believe you; aller en bateau is to be caught up in a story you're being told -- which is relevant because Celine and Julie tell each other stories and believe these stories. They embark on their own fiction: hence the shot of the boat at the end.

How did you decide which actress should play which role?

We were a little undecided at first. We felt they could swap roles if they liked, but actually it was better that they should play the pairs they did. But that's why I wanted to keep for the end the rather facile effect Hitchcock has used many times, notably in Strangers on a Train, which suggests they could have played the other two pairs. It's a conventional effect to end what is after all a conventional film. Julie is a wholly sensible character. In fact the film is something which happens to Julie: she's a character with her own inner psychology, possessing a past, a present and even a future, as is suggested by implication in her meeting with Gilou, her girlhood fiance; everything that happens to her is linked to her own imagination. Celine, on the other hand, is purely exterior, a behaviorist character, who is really seen from Julie's point of view. We are really with Julie, receiving flashes about Celine; there are some when Julie isn't present, but they are all flashes of comedy, of exhibition -- the scene with Gilou, the conversation with her friends from the theatre.

Where everything is fantasy with Celine it is memory with Juile. For instance, the photograph food in the trunk which heralds the appearance of the mysterious house. This was the failure of the film version of The Turn of the Screw, where Deborah Kerr was given a sort of get out, in that it is after finding a photograph of Quint and Miss Jessel that she begins to 'see' them.

All I remember about The Innocents is how silly I found it, psychologising everything to death, turning the governess so blatantly into a neurotic. Of course you can read James that way, but who is to say that ghosts don't exist? The Nightcomers, although extremely badly made, had a much more subtle theme. Nevertheless I think that Henry James, like Bioy Casares, is unfilmable. They are authors who can be filmed diagonally, taking up their themes, but never literally. A few days ago I saw the Italian adaptation of The Invention of Morel; an unimaginably obvious, drearily literal illustration of the book. It's incredible how anyone could make such a dull film out of that marvelous novel, where the things that work in the book as words on the page simply won't transfer to the screen, even if, as the director claims, he tried to transpose from first person narrative to third. It simply doesn't work at all. The adaptation should have been much more ruthless, even a betrayal in the sense that Hitchcock betrayed Boileau and Narcejac for Vertigo -- a film I thought about while watching The Invention of Morel, especially when the hero began to spy on his wife. Everything should have been reconstructed from there; above all, the mechanism shouldn't have been revealed.

Celine et Julie vont en bateau suggests another Bioy Casares story even more: than The Invention of Morel, EI Perjurio de la Nieve (3), in which the hero comes to an isolated house where the same actions and the same words are repeated day after day to arrest time, and it is his presence that breaks the spell.

I haven't read much of Bioy Casares. Contrary to what people think, I'm not particularly well read. But Eduardo knew that story and told me a bit about it while we were working. With this type of fiction -- semi Anglo-Saxon, semi Argentinian -- there are only two things you can do. You can borrow the mechanism to do something different, or you can take a kernel from the plot and start rebuilding from there, jettisoning the entire original mechanism. I was struck by this while watching that Italian film the other evening: the folly of trying to make a literal transposition of this sort of fiction.

When Paris Nous Appartient came out, several people found comparisons with Borges. Admittedly you see a book by Borges on the heroine's table in the opening shot. But this was simply because Suzanne Schiffman, who was the assistant on the film, happened to be reading it. We needed a book on the table, we saw the title Enquetes (Otras Inquisiciones) and thought it suited the film quite well, so on to the table it went. Anyway, two or three critics referred to Borges, no doubt because of that shot. So I then began to read Borges, and of course found him magnificent. For quite awhile I wanted to do something with Theme of the Traitor and the Hero -- this was well before Bertolucci's Spider's Sratagem, at a time when I couldn't get La Religieuse off the ground and was trying to set up something else -- and for six months I struggled to construct a script from this story. I tried to fit all the levels of the Borges story into my adaptation, until I got lost in my own labyrinth and finally said 'It's unfilmable.' Later I saw Bertolucci's film, which I like very much, and as I see it he too abandoned any attempt to film Borges. The Spider's Stratagem tells another story, in the middle of which is told the 'Theme of the Traitor and the Hero'. What I wanted above all was to retain the last sentence, (4) which -- as always with Borges -- hints at the possibility of quite a different story.

  1. Suzanne Schiffman. Artistic adviser to Truffaut, Godard, Rivette, scriptwriter on Out 1 and La Nuit Americaine among others, a major influence on French films since 1959.

  2. TSE. A basically Argentinian theatre troupe. Among their best-known productions: Dracula, Goddess, L'Histoire du Theatre, Comedie Policiere, Luxe.

  3. El Perjurio de la Nieve was the basis for Leopoldo Torre Nilsson's first feature El Crimen de Oribe, 1950, co-directed with his father Leopoldo Torres Rios.

  4. Translation, by James E. Irby, of the last paragraph of Borges' story: "In Nolan's work, the passages imitated from Shakespeare are the least dramatic; Ryan suspects that the author interpolated them so that in the future someone might hit upon the truth. He understands that he too forms part of Nolan's plot... After a series of tenacious hesitations, he resolves to keep his discovery silent. He publishes a book dedicated to the hero's glory; this too, perhaps, was foreseen." (Labyrinths, Penguin)

Originally appeared in Sight & Sound Vol. XLIII, No. 4, Autumn 1974; p. 195-8.