The three films which Laura Mulvey and I have made together (Penthesilea,1974; Riddles of the Sphinx, 1977; AMY!, 1980) all attack a single set of problems which are at the same time political, psychoanalytic, and semiotic. According to the Lacanian schema, each of us is subjected to the symbolic order, the superimposition of social law and regulation on the needs and instincts of nature. This symbolic order, this law, is authorized in the name of the father, the third term which intervenes to break up the dyad of mother and child at the joint moment of entry into the Oedipus complex and acquisition of language. Thus the symbolic shatters the imaginary plenitude of the mother-child relationship and restructures the course of our identifications. We must find our places, as men or women, within that symbolic order, an order which is constituted through verbal language and on the alienation of sign from object.

The patriarchal, character of this symbolic order necessarily makes it problematic for women and a fortiori for feminists. In our films we have tried to investigate the limits of a validation of the imaginary (the myth of the Amazons, the dyadic mother-child relationship, the exemplary heroine) as a form of resistance to patriarchy and the possibility and implications of transforming the symbolic order to one which is nonpatriarchal. Since the symbolic order is tied to the acquisition of language, this has involved paying particular attention to the place of verbal language within our films, which in many respects are interrogations of language itself, symbolic quests in search of the place from which women could utter the repressed counter-meanings of patriarchal discourse.

One possible strategy, which we have always rejected, would be to avoid verbal language altogether and produce a kind of film which, lacking the explicit presence of words, was made only of icons and indices, images and traces. From the beginning of cinema, again and again, we can find an impulse on the part of filmmakers to banish language and reduce film to the status of a "pure" visual art. Not only was the advent of sound resisted (seen as tributary to verbal language) but there was a campaign, within silent cinema, against the written intertitle, both in commercial and avant-garde film. Up to this day, many avant-garde filmmakers persist in refusing verbal language or reducing it to a minimal and epiphenomenal role. Another possibility, the inverse of this, would be to restrict ourselves to writing and not make films at all.

In fact, however, it is precisely the interface between image and word which concerns us. It is here that sexual difference, the subject of our films, takes shape. To restrict ourselves to the imaginary would be to restrict ourselves to the (prolonged) pre-Oedipal phase which, even if it could be interpreted as a form of resistance, would involve, in psychoanalytic terms, both a repression and a regression, and, in political terms, a flight from the society in which we are in fact living and from history. To restrict ourselves to the symbolic, on the other hand, would involve a denial (or, at least, an undervaluation) of the persistence of the pre-Oedipal and of the imaginary within and alongside the symbolic, ceaselessly structured by it yet escaping it in the form of desire. This would be to deny history in another way, by revoking wish and memory in their full force.

Verbal language is a crucial component of film, both as signifier and as signified, as crucial as the image. Each is deprived of a dimension of its sense in the absence of the other. This is not to argue, of course, for a mutual reinforcement of the two-standard practice in commercial cinema, where word and image are used to add more and more reciprocal redundancy. On the contrary, it is in the dialectic of fit and misfit that the value of working with both word and image lies, as well as in the heterogeneity of the registers of each. Language is the component of film which both threatens to regulate the spectator, assigned a place within the symbolic order, and also offers the hope of liberation from the closed world of identification and the lure of the image. Language, therefore, is both a friend and a foe, against which we must be on our guard, whose help we need but whose claims we must combat. Hence the fractured and dislodged body of language in our films.

Penthesilea is constructed around two theoretical axes- the myth of the Amazons, shown as a palimpsest, whose genealogy reveals in every repetition its status as an inscription of male fears and fantasies, and the symbolic blockage of woman's voice, unable to find a place from which to speak against the order of patriarchy. The myth can only be reinscribed as a myth of and for women when the historic place of women is changed. The film then sets up a problem, contrasting the graphic inscriptions on the picture track with the voice on the sound track, unable to speak its desire. This exclusion from the symbolic leads to the demand for a change in the field of language, but without any indication of how this might come about.

In all three films we have made, woman's voice predominates rather than man's. In Penthesilea, there are four modes in which we hear this voice. In the first section of the film, we hear women's cries and groans-the absence of language is underlined by the fact that we are seeing a version of Kleist's play, Penthesilea, renowned for its language and its status as literature, reduced to a mime. In the third section, we hear a woman's voice, with music, striving to articulate sounds into a language, but not yet succeeding, poised still on the brink. In the fourth section, we hear the words of an American suffragist, Jessie Ashley, who ends by lamenting her inability to speak as she desires and proclaiming her return to silence, contrasting her sense of failure with the delusion of success among those women who have learned to speak from within the patriarchal order, to accept the place of servitude, and its language, in asking for women's rights from their masters. Finally, at the very end of the film, the actress who had played the part of Penthesilea in the mime removes her make-up and calls for a new language, direct to camera, though hesitatingly, as if unused speaking.

The emphasis throughout is on exclusion from language (the only man's voice-mine-is the only articulate one) or rather on exclusion from the possibility of speaking against the symbolic order because of the insistent subjection of women within it. Thus to speak, to quote the words of Lacan which form the epigraph to the third section, is to use those "seals of phobia," symptomatic emblems of male fears, which thereby become "blazons of self-punishment" when adopted by women. Riddles of the Sphinx goes a stage further than Penthesilea. It sets out to capture not only the predicament of women, but the outlines of a prospect of change. It takes as its starting point not the Amazons, the site of imaginary identification, but the sphinx, interpreted metaphorically as the repressed signifier of the mother in the female unconscious. To be heard, woman's voice must come from this place, the place of the sphinx.

The concept of the "female unconscious" does not imply any essential or natural "femaleness" of the unconscious. It gives a name to what was described by Freud in his paper. "A child is being beaten"-because the trajectory of the girl through the Oedipus complex is different from that of the boy, so the pattern of repression is different for her and the genealogy of her fantasies and unconscious desires. Freud uses the beating fantasies of men and women to trace quite different and distinct histories of repression and regression, based on the different attitudes taken up by boys and girls within the Oedipal constellation, quite distinct contents of the male and female unconscious, in effect. Here, as elsewhere, he openly argues against "essentialist" explanations: the contents of the unconscious, together with sexual dierence itself, are historically formed.

In Riddles of the Sphinx, we were concerned with the fact that whereas within the patriarchal order the image of the mother was offered to women for identification, her symbolic place was occluded by that of the father, and this involved a repression-perhaps a primal repression. We conceived of this place of the maternal instance, metaphorically like that of the sphinx, as "outside the gates of the city" and returning, for Oedipus, as the repressed always must, in the fateful prize of marriage to his actual mother, Jocasta, awarded to him for his destruction of the power of the sphinx by answering her riddle. By producing order in the symbolic (answering the riddle), Oedipus produced disorder in the real (incestuous marriage). Thus the mother, whose symbolic power he has destroyed, returns to destroy him as he assumes the mantle of patriarchy. The myth, like that of the Amazons, is sealed by male phobia.

The sphinx, however, unlike Penthesilea, exists differentially within language. Penthesilea is locked in the imaginary symmetry of struggle and love with Achilles, but the sphinx speaks from a distinct place with a distinct form of language. The riddle is metaphoric, interrogative, and incomplete; it involves wordplay, enigma, and disguise. The language of Oedipus, by contrast, is transparent, assertive, and closed. It is important to stress that the sphinx represents an alternative form of language-she is not outside language as she is outside the city of Thebes, the realm of patriarchy, but is able to offer a different discourse, potentially the nucleus of a nonpatriarchal symbolic, based on a different Oedipal structure-or, perhaps it would be better to say, a different mode of entry into language, kinship, and history.

The voices in Riddles of the Sphinx, like those in Penthesilea, are predominantly those of women. The only man's voice heard is that of Louise's husband, Chris, who speaks a few lines, redolent of the goodwill and sound sense of patriarchy in its full banality. The first woman's voice is that of Laura Mulvey, speaking, as I did in Penthesilea, from the position of filmmaker. Thus the symbolic role of author is opened for female identification. Moreover, Laura not only speaks to introduce the voice of the sphinx, but also listens finally to the voice of the sphinx herself, thus repositioning herself as author. As she listens, Laura writes: her inscription is a function of the voice to which she is listening. She is no longer origin of the discourse but an instance of its intertextuality. The tape recorder to which she listens, like the video monitors at the end of Penthesilea, represents an access to memory which produces a new discourse.

The second voice is that of the sphinx. The sphinx is introduced first on the picture track, with lips sealed: she is to be voice-off, her symbolic function separated from her image, which must be dissolved before she can be heard. Even then, she can only be heard with difficulty, "buried" beneath the music. The voice of the sphinx has three modes: poetic (structured around allusion to the role and experience of motherhood); questioning (about politics and the unconscious); and remembering. This third mode is the most complex because in it the voice of the sphinx uses the first person pronun with an unanchored reference, shifting between herself, Louise, and Louise's daughter (perhaps) in an ambiguous play of signification of the shifter. It is no longer possible to "fill" the symbolic place of the sphinx with a single fixed identification. There is a dispersal of the subject, dissociated from a single ego or consciousness.

The last words of the sphinx appear on the tape to which Laura listens. They run as follows:

I was looking at an island in the glass. It was an island of comfort in a sea of blood. It was lonely on the island. I held tight. It was night and, in the night, I felt the past. Each drop was red. Blood flows thicker than milk, doesn't it? Blood shows on silk, doesn't it? It goes quicker. Spilt. No use trying. No use replying. Spilt. It goes stickier. The wind blew- along, the surface of the sea. It bled and bled. The island was an echo of the past. It was an island of comfort, which faded as it glinted in the glass.

Like a riddle, the sphinx's words are based on repetition (of words, syntactic forms, and rhymes), metaphor, wordplay, and subtexts-especially of the two proverbs, "Blood is thicker than water," with its reference to the order of kinship, condensed with other senses of blood, having to do with menstruation and injury; and "No use crying over spilt milk," with its similar reference to mother's milk and its displacements, via "water," to the mother-child dyad (subordinated to the order of kinship) and, via "spilt," to bloodshed. Finally, blood too is linked to water, to the sea, and hence to the womb. It is a riddle without an answer, whose meaning is dispersed.

The third set of voices are those of women placed within the diegesis and the symbolic order. We do not hear the voice of Louise, the protagonist, until the moment we see her face, the moment when she is parted from her daughter and enters the public realm. Embedded within the dialogue are three monologues: one, a political voice, speaks about the attitudes of trade unions to women's demands for day-care nurseries; the second, a theoretical voice, speaks both about her own child and, in Lacanian terms, about "weaning from the dyad"; and the third, the voice of the protagonist, Louise, reads the narrative of a dream from her friend's diary. This series of voices signals a movement from the politics of consciousness through theory to the problems of the unconscious, leading finally to Louise's attempts to read the language of hieroglyphs, a metaphor for the enigmatic voice of the sphinx.

The theoretical voice, that of Mary Kelly, which supplies some elements of a metalanguage, is distanced from the mainstream of the diegesis in order to mark the movement to a different level of language. This speech could not be spoken by the sphinx because the sphinx uses not a metalanguage, but a counter-language. Access to a metalanguage, like access to memory, is, however, a condition of possibility for a counter-language, a necessary phase of Louise's journey through the word; and hence it is crucial that she (and the spectator of the film) should encounter this difficult theoretical text and, of course, that it should be spoken by a woman.

The third film, AMY!, recapitulates and extends the registers of language and voice used in the first two. The key speech, placed in symbolic contrast to the male voices of newspaper headlines and popular love song, is a montage of fragments from a number of texts. It is impossible to identify a continuous source for the writing, which is heterogeneous and disharmonic, but different in form from the experiments in counter-language attempted in Riddles of the Sphinx. It is placed in a metaphoric relationship to the perversity of Amy's flight, the thrill of which is suppressed as it is rewritten into the form of legend, based once again on male fantasy, within the patriarchal order. This speech is placed between her refusal of the name-of-the-father and the identity assigned her by the patriarchy and her transcription into a fetishized emblem within the museum-morgue of patriarchal legend (the plane in which she made her flight suspended motionless from the roof of the Science Museum).

The project of a counter-language necessarily involves more than the ways in which verbal language is used on the sound track. Words also appear on the picture track, in the form of titles, epigraphs, intertitles, and superimpositions. In AMY! the flight around which the film pivots is presented as a process of reading, as Amy's journey is followed across the map. The text is both a litany and an itinerary through a series of languages, as the morphology of the place-names changes, until the original English is reencountered finally in Australia and the flight is stopped, to be transcriled into the language of the Empire in legend. The map, too, is contrasted with the letters that Amy burns in an earlier scene, which in turn are transformed into the language of the engineering manual and the daily drill of the mechanic.

There is another sense, however, in which the concept of language relates to the picture track, apart from the presence of written words as "graphic enclaves" within the sequence of images. Freud himself described how the mise-en-scène of the dream produced a form of displaced writing, a pictograph, rebus, or emblem. In our three films the images themselves are designed to have this kind of pictographic or emblematic force. The films have a module construction, in which individual shots are treated as tableaux, as text rather than representation. Thus, in Penthesilea, each section is transcribed through a different cultural form: mime and gesture, writing, painting and sculpture, cinema, and, finally, video. The governing image of the last section is that of memory as a form of writing (Freud's "Mystic Writing-Pad") by which the previous sections of the film are interwoven, as in Griffith's Intolerance, through retranscription.

The camera, too, is treated as a camera-stylo (camera-pen), in Astruc's phrase. Camera position and camera movement are elements of design (disegno) in the historic sense: frame determines the geometrical limits of the picture-space and movement inscribes the trajectory of the frame within a geometrically interpretable space. Thus 360-degree pans, which make up most of Riddles of the Sphinx, inscribe the form of a circle; tracks inscribe straight lines; the complex camera movement of the second section of Penthesilea inscribes an elaborate arabesque, and so on. Cinematography itself is a form of inscription, rather than a capture and reproduction of the world.

In Penthesilea an epigraph drawn from Mallarmé suggests that the face of a mime and, by extension, the screen itself, is like "a not-yet-written page." I was always struck by the Mallarméan resonances of Hitchcock's dictum that "there is a rectangle up there-a white rectangle in a theater-and it has to be filled." A film is like a book whose pages are extended in time, to be inscribed with graphic signs. The project of a counter-language, then, is one which implies a transformation of the symbolic order of these graphic inscriptions, neither a refusal of "writing" and a relapse into a realm of the image as such, nor an acceptance of the canonical codes of the dominant narrative cinema, the cinematic law, so to speak. Here, of course, the same attention must be paid to the place of the look as is paid to the place of the voice on the sound track.