India Song

(Marguerite Duras, France, 1975)


Durassic Park


For about 60 years – the first 60 years of cinema – the relation of a narration (whether in written inter-titles for silent films, or spoken voice-over for sound films) to a narrative was relatively simple and clear. We watched an event unfold on screen; and, laid on top of it, a commentary that did not pre-empt the visual event, but tactfully prepared it, set up its context (“So I came to the river, unaware of what would happen that day …”) – and then, just as tactfully, vanished from the film until its next brief, necessary, functional appearance at the start or end of another scene. Many fine classical films – from How Green Was My Valley (John Ford, 1940) to Carlito’s Way (Brian De Palma, 1993) – have streamlined and perfected, for artistically expressive as well as functional storytelling purposes, this technique.


Only a few special, exceptional works throughout the greater part of cinema history contradict this convention. In the Chinese film Spring in a Small City (Fei Mu, 1948)  – the daring original version, not its comfortable and slick 2002 remake – the voiceover of the wife in this tangled but understated marital melodrama poetically reiterates what is plainly visible on screen, covers incidents she has in fact not witnessed, and puts sad, unspoken realities into brutal words. In the Japanese production of a supremely eccentric Austrian-American, Josef von Sternberg’s The Saga of Anatahan (1953), the almost constant narration, spoken by the auteur himself, regularly destabilises its own presumed omniscience over the story. And in Marcel Hanoun’s remarkable Une simple histoire (1959), the idea of spelling out or doubling the patiently descriptive image-track (devoted to a single mother’s everyday actions) with an equally pedantic voice-over account (belonging, this time, to the central character) reaches a rigorously systematic yet subtly varied point of extremity.


In his classic essay on Sternberg (originally titled “Aquarium”), the modernist French novelist Claude Ollier (1922-2014) suggested of the voice-over in The Saga of Anatahan:


The text which he reads is in fact more than just an element linking one scene to another, and more than just a commentary. It is a musical and signifying part of a poetic fresco for a screen in black and white. […] The voice seems to summon a scene into being as often as to hold forth on its execution, seeming to solicit the mise en scène more than to criticise it. (1)


The various New Waves of the 1960s – French, Brazilian, Czech, Japanese, Polish, Italian, Indian – scrambled the customary functions of voice-over in many, freewheeling ways: fragments of spoken text inserted as quotation, lyrical interlude, reflection, jazzy punctuation, complication or punctuation of narrative point-of-view … the sorts of modern voice-over techniques mined, ever since, by Martin Scorsese or Leos Carax. Yet, in an odd phenomenon of global synchronicity, it was during the 1970s – and essentially there alone – that the intermittent pre-‘60s radicalism of experimental narration asserted itself in force across a number of extraordinary, singular works, including Badlands (Terrence Malick, 1973), Fontane Effi Briest (Rainer Werner Fassbinder, 1974), Doomed Love (Manoel de Oliveira, 1978), Barry Lyndon (Stanley Kubrick, 1975) and, last but far from least, India Song. Which is a curious bunch of titles indeed, crossing serial television production, genre cinema, classic literary adaptations and the avant-garde.


In these films, the conventional relation of visual event to verbal narration is thoroughly, utterly inverted. The images may show or propose little – a couple driving, a character sitting at a table, a tableau vivant of an awkwardly static grouping of characters – but the narration gives us an enormous amount of information about and reflection upon this visual event: what’s happening before, during and after it; its context and significance; and sometimes a poetic weaving of different voices around it and spreading outward from it.


There is always a significant shift (in relation to convention) involved in the way these texts are delivered, the imagined place from where they are spoken: Fassbinder, like Sternberg, reads the voice-over himself; Sissy Spacek’s Holly in Badlands swings wildly from impossible omniscience to unacknowledged avoidance and repression; the novels used by Kubrick and de Oliveira (a constant in the latter’s career, as Abraham Valley [1993] shows) are retained in huge, verbatim chunks.


These spoken texts can seem “novelistic” – even when they are not actually adapted from prior novels – in their density and detail, their highly written structures of literary artifice. And the resulting films may seem minimalistic – even as they are, to those open and sensitive to their charge, textually dense and deeply affecting. Both novelistic and minimalistic won’t get us terribly far, as terms or tool, into these films.


Duras’ India Song is a key movie in the small tradition of works that are simultaneously minimalist and rich. But it is – and always has been, since its premiere screening at the Cannes Film Festival in 1975 (2)  – divisive. There are viewers for whom Duras’ filmic minimalism will always be too much: desiccated, overly formal (if not formalist) and abstract, literally posturing, flat, too heavily leaning upon the sombre effect of the literary word. Many poor imitations subsequent decades (in short and feature films, as well as art-videos and gallery installation pieces) have done nothing to improve this bad reputation for those ill disposed towards Duras as an artist-celebrity.


Emilie Bickerton, for instance, concludes that the “distillation” of Duras’ oeuvre could run thus: “A detached, half-conscious protagonist looks out on a world that she is ultimately powerless to affect … a world dominated by ennui”. (3) But to those who love the film – beginning with the testimony of the never-very-modest auteur herself in the informative making-of documentary La Couleur des Mots (1984) (4) – its substance is wholly and instantly emotional. I will come back to the question of the kinds of spectator emotion it releases or puts into play.


Duras frequently presented India Song – like much of her cinematic, theatrical and literary work – as autobiographical. Not issuing directly from her own experience – as is her most celebrated novel, The Lover (1984, poorly adapted to film in 1992) – but based on her observation of people around her when she was young. Duras grew up in French Indo-China (known today as South Vietnam); in working up India Song (firstly, as a theatrical piece commissioned for the British stage), she decided to “widen the frame” by altering the geography, and hence what she called (in La Couleur des Mots) the “destinies” of her characters: “I moved everything to Calcutta”.


In so doing, she abstracted two different situations of colonialism, in order to equalise them as material for her modernist/minimalist melodrama of the passions. Even specific socio– or geo-political associations translate into an allegory of passion for Duras: “She is Calcutta” and “He alone is Lahore”, she remarks of the main characters in her “Notes on India Song”, (5) irresistibly replaying the memory of gender and national division structuring Hiroshima mon amour (1959), which she wrote for Alain Resnais. I am interested, in this piece, less in a political critique of Duras’ decision (rarely undertaken by her champions), than in taking aesthetic measure of the displacement it involves. (6)


For the displacement does not stop there. The author’s experience of a childhood in Indochina is made over into an imagining of a decadent, white colonial culture in decline in India; and that imaginary is filmed, very obviously, in contemporary France – specifically locations including the decaying Palais Rothschild – without any attempt to disguise the raw artifice of this fact. Another displacement, minor but glaring, that is surprisingly easy to overlook or forget: that this extremely French film has an insistently English title, and that – in a way that evokes, in a surreal détournement, the unfussy logic of Hollywood’s historic spectaculars set in non-American lands – its presumably British characters speak almost solely in the author’s native tongue. From Indo-China to India to France: this dizzying succession of moves, both geographically and temporally, leads us finally to an unusual space-time that seems to belong to cinema – or more exactly to the unfolding of this one film – alone.


As savvy spectators of today, we need to remind ourselves of the strangeness of India Song in this regard – and how radical a gesture it must have been in 1975. This is a film in which the act of representing the past – recreating it, evoking and conjuring it, building it on screen – undergoes a massive process of dematerialisation. There are costumes and sets that more or less evoke that past, and music participating in the historical mood … but very little else.


Is the costume-drama flimsiness of India Song a matter of economy, production expedience? On the contrary, it was, from the outset, a fundamental aesthetic decision. The past is gone, cannot be represented, is lost in the mists of its re-mediation and narrativisation; personal experience can only be reconstructed accordingly, at the risk of what Bickerton calls an extreme (and, for her, crippling) “sense of otherness from the world”. (7) This is the type of writing or art-making, from the cliff-edge of oblivion, deliberately riddled with absences and impossibilities, that is familiar from the work of Duras’ literary comrade, Maurice Blanchot. (8)


Little wonder that so many reviews and commentaries on India Song refer to it, casually, not as a historical re-creation at all, but as the prime example of another, amorphous kind of cinematic genre: the memory film (closely linked, as it happens, to the equally amorphous brand-label of the dream film). Critics including David Bordwell rank India Song, in this regard, alongside Gertrud (Carl Dreyer, 1964) or Last Year in Marienbad (Resnais, 1961) as particularly haunting examples of this type; (9) the collected works of Terence Davies (such as Distant Voices, Still Lives, 1988) and Bird (Clint Eastwood, 1988) have also been nominated as prime memory films, alongside the entire “haunted cinema” of Víctor Erice (see this audiovisual essay).


What is this special category of the memory film? Decidedly not a film with elaborate and lengthy flashbacks, a guiding voice-over (“I recall when it all began …”), and a plot construction founded on the fictional act of remembrance – such as (to take a magisterial example) Once Upon a Time in America (Sergio Leone, 1984) or, in a more fractured, modernist, multi-layered vein, The Ceremony (aka Ceremonies, Nagisa Oshima, 1971) and Time Regained (Raúl Ruiz, 1999). (10) Rather, it is a film that, in totality, presents itself as filtered through memory, the end-result of a process of memory, or a mosaic of memories, unflagged as such except through clues offered by style and mood (usually melancholic or elegiac).


However, designating India Song as a memory film runs the risk of cohering it, papering over its abundant cracks, in a way that the work itself militantly resists. In this regard, one does well to make the effort of trying to re-experience the film in its time, that initial moment of the “India Song event” (as Joël Farges and François Barat dubbed it). (11) For, just as it witnesses the intriguing flowering of extreme and unconventional voice-over narrations in 1970s cinema, India Song sums up a brief, bold era in art and thought when gaps, fissures, ruptures and contradictions were prized almost above all else.


A profoundly anti-classical moment: whereas normal films are daily praised for appearing seamless, India Song is a film full of gaping seams – indeed, it is built on them. My own memory of seeing the film as a young cinephile in the ‘70s  (12) corroborates this: what was striking then (to adapt a phrase of Duras’ own explication) was the fact that the image and sound proceeded on two separate paths, merely “touching” each other now and again, creating a frisson of narrative (and of the narrative’s world), but never fully constituting it (Duras described the film as “the bringing to ruin of any kind of reconstitution”) (13); that certain notable images – such as, immortally, the vision of Delphine Seyrig’s exposed breast, standing up in the frame like a mountain – seemed to emanate from no discernible point-of-view belonging to any visible fictional character; that Carlos D’Alessio’s extraordinary old-style, dance-band score, as a recording, is simply engineered but not produced, hence retaining the rough charm of an unmixed, untouched live performance (bum notes and all). (14)


What is the upshot of all this? The story – this sad story of Anne-Marie Stretter, relayed in novelistic form in The Ravishing of Lol V. Stein (1964) and The Vice-Consul (1965) – is told neither in the images nor the sounds; rather it is alluded to, echoed, replayed in various fragmentary ways. The specific time and place of the story, the time and place of its unfolding, is thus neither seen nor heard. This is another constitutive aspect of the film’s strangeness: from where, exactly, do these images and sounds (separately) issue? From what point in space and time? Cinema – even the most conventional, mainstream cinema – has accustomed us to routine ambiguity, a strategic vagueness, in the placement of voice-over narration: it usually comes from somewhere further ahead in time, when the narrating-perceiving-remembering subject is older and wiser, having reached the point of being able to tell his or her tale. But to take the assumed present tense of the image and subject it to the same wandering vagueness or ambiguity is a truly radical move, and Duras’ gesture – her conjuring of a filmic fiction as “this place where doubt is cast” (15) – resonates through many of the most decisive experiments to follow in cinema history.


Many in-depth discussions of India Song dwell on aspects of its highly original visual style – its lighting scheme, wall-to-floor mirrors reminiscent of certain Russian silent films, rhythms of camera and bodily movement – and, in broad terms, the system of its image-sound relation (taken even further by Duras in the film’s avant-garde sequel of sorts, Son nom de Venise dans Calcutta désert [1976], which recycles exactly the same soundtrack, but over completely depopulated shots seemingly taken in and around the same locations). (16) But what of the film’s emotional impact, so determining for those who adore it, so lacking for those who do not? The peculiar emotional register of India Song is caught in Ollier’s discussion of The Saga of Anatahan in “Aquarium”, written four years before Duras’ shoot in 1974:


Emotion no longer arises from the events themselves, but from the alternation of drawing near and moving back at the suggestion of the “voice off”, a coming and going based on the alternation of the struggle between violence (shown and lived) and calm (recited, earned by age and reflection), between madness and wisdom, between self-destruction and renewal and recovery. In its strict and distant course, this voice communicates over and over again the impression of a destined and inevitable development. (17)


This description corresponds to the passion, intermingled with contemplative distance, that Duras herself frequently evoked: “To speak about the reception in India Song is to inhibit the rapture I feel when I speak about it”. (18) But there is another, quite different but no less intense emotion I have often observed (and even participated in) at several public screenings of the film: a kind of delirious, hysterical, mad laughter. And this is not necessarily a sign of disrespect, nor entirely explicable as a nervous, defensive reaction – for quite a number of the film’s fans have testified to this very same experience, in many different periods and countries since 1975.


Why does India Song – and particularly, in my own viewing history, the prolongation of Michael Lonsdale’s wild screaming as the Vice-Consul on the soundtrack – provoke such an emotional reaction? The answer, I believe, lies in a condition of the cinematic apparatus that Thomas Elsaesser described well in 1969: the viewing situation generates what he calls a psychic matrix from the agitational effect of the motion-picture upon the immobile spectator. (19) Managing this essentially negative condition – one that easily gives rises to violent emotional reactions – and turning it into a positive experience is no small part of the everyday art and craft of narrative movies. But when the immobility of audience members is met by the immobility of events and bodies in the image-track of a film such as India Song, something goes fascinatingly askew, and other kinds of sensation – like the kind of jouissance which Roland Barthes once promised exists on the nether side of boredom (20) – come into play, exactly the kinds of sensation elaborated by a century of avant-garde cinema. What usually counts as psychic compensation for the spectator during (brief) moments of immobility in funky mainstream cinema (think of the tableau or vignette effects in Wes Anderson’s films) is the soundtrack – usually a beloved or cultish pop song that freezes the action more definitively than the numbers in a traditional musical comedy do.


But the veritable shuttle-parade of displacements on all levels in India Song takes us way beyond the facile pleasure of such compensations, and into a hitherto unknown terrain of cinematic thought and sensation – a terrain that we are still exploring (whether in practice or critical theory) today, long after the death in 1996 of Marguerite Duras.





1. Claude Ollier, “Josef von Sternberg”, in Richard Roud (ed.), Cinema: A Critical Dictionary (London: Secker and Warburg, 1980), pp. 958-959. The original French text, “Aquarium”, now appears in Ollier, Ce soir à Marienbad (Paris: Les Impressions Nouvelles, 2020); the English translation for Roud by Michael Graham is superb and exact. back


2. See Marguerite Duras, Duras by Duras (San Francisco: City Lights Books, 1987), for many passionate testimonies dating from the first year of the film’s public life. back


3. Emilie Bickerton, “The Timeless Marguerite Duras”, The Times Literary Supplement (25 July 2007). back


4. La Couleur des Mots (The Colour of Words), directed by Jérôme Beaujour and Jean Mascolo, with interviews conducted by Dominique Noguez, is available as a Postface on the India Song DVD (Benoît Jacob Video, France). back


5. Duras by Duras, pp. 14-15. back


6. It can be argued that the task of politicising Duras falls precisely to other, later filmmakers who admire and have been influenced by her: in this sense, if H-Story (Nobuhiro Suwa, 2000) fills in the geo-political outline of Hiroshima mon amour, then A Song of Ceylon (1985) by the Australian-Sri Lankan experimental filmmaker Laleen Jayamanne offers a pointed rewriting or rewiring (along another geo-cultural displacement) of India Song. back


7. Bickerton, “The Timeless Marguerite Duras”. back


8. See, for an example of Blanchot’s commentary on Duras, The Unavowable Community (New York: Station Hill Press, 1988). back


9. See David Bordwell, The Films of Carl-Theodor Dreyer (Berkeley: University of California Press, 1981). The notion of the “haunted memory” film is also explored in my book Mysteries of Cinema (Perth: University of Western Australia Publishing, 2020). back


10. For a discussion of this temporal narrative structure, see Adrian Martin, Once Upon a Time in America (London: British Film Institute, 1998). back


11. Duras by Duras, p. 1. back


12. See Adrian Martin, “Scenes” (1982), reprinted in Mysteries of Cinema. back


13. Duras by Duras, p. 16. back


14. The music – in exactly the recorded form that the film uses it – appears on the invaluable CD (credited to D’Alessio and Duras), India Song et autres musiques de films (Le Chante du monde, 1991, LDX 274818). D’Alessio, born in Argentina 1935, worked extensively with Duras on film and theatre projects; he became best known, near the end of his life in 1992, for his collaboration with the directorial team of Jean-Pierre Jeunet and Marc Caro (Delicatessen, 1991). He can be glimpsed performing the central India Song theme on piano in La Couleur des Mots. back


15. Duras by Duras, p. 13. back


16. See, for examples, Elizabeth Lyon, “The Cinema of Lol V. Stein”, in Constance Penley (ed.), Feminism and Film Theory (London: British Film Institute & Routledge, 1988), pp. 244-271; Manny Farber, Negative Space (New York: Da Capo, 1998). back


17. Ollier, “Josef von Sternberg”, p. 959. back


18. Duras by Duras, p. 12. back


19. Thomas Elsaesser, “Narrative Cinema and Audience-Oriented Aesthetics”, in Tony Bennett, Susan Boyd-Bowman, Colin Mercer & Janet Wollacott (eds.), Popular Television and Film (London: British Film Institute, 1981), pp. 270-282. A subsequent 1974 revision of this late 1960s text, again retouched in 2001 and retitled “Narrative Cinema and Audience Aesthetics: The Mise en Scène of the Spectator”, appears in Elsaesser, The Persistence of Hollywood (London: Routledge, 2012), pp. 95-104. back


20. See Roland Barthes, The Pleasure of the Text (New York: Hill and Wang, 1975). back

© Adrian Martin June 2008

Film Critic: Adrian Martin
home    reviews    essays    search