3. Female Authorship
Reconsidered

All of the films discussed in the previous chapter take as their central premise and plot the relationship between women and image-making, and may thus be read as explorations of female authorship in the cinema. The importance of female authorship is accentuated by the fact that, with the exception of Julie Dash, the filmmakers themselves appear in their films, from Helke Sander’s role as protagonist of Redupers, to the more cameolike appearances of Yvonne Rainer (as a voice and briefly on screen) in The Man Who Envied Women and Patricia Rozema (in a window as Polly climbs up a building during one of her fantasies) in I’ve Heard the Mermaids Singing. In more general ways, of course, all of the films trace a relationship between women and cinematic production.

In this chapter, I will examine how female authorship has been theorized in feminist film studies, and I will focus in particular on the example of Dorothy Arzner, one of the few women to have been successful as a director in Hollywood in a career that spanned from the late 1920s to the early 1940s. Arzner was one of the early “rediscoveries” of feminist film theory, and she and her work remain to this day the most important case study of female authorship in the cinema. While the most significant work on Arzner’s career was done in feminist film studies of the early to mid-1970s, I will suggest that important dimensions of her status as a female author have yet to be explored.

In contemporary feminist literary criticism, inquiries into the nature of female authorship have been shaped by responses to two somewhat obvious assumptions: first, that no matter how tenuous, fractured, or complicated, there is a connection between the writer’s gender, her personhood, and her texts; and second, that there exists a female tradition in literature, whether defined in terms of models of mutual influence, shared themes, or common distances from the dominant culture. A wide range of critical practice is held within these assumptions. But insofar as a self-evident category of womanhood may be implicit in the female author defined as the source of a text and as a moment in a female-specific tradition, these seemingly obvious assumptions evoke what has become in contemporary theory a dreaded epithet: essentialism.

A decade or so ago, a friend of mine remarked sarcastically upon the prevalence of “oedipus detectors” at a Modern Language Association meeting, that is, critics eager to sniff out any remnants of oedipal scenarios in work that was ostensibly “progressive,” “feminist,” or “postoedipal.” Contemporary feminist criticism—and feminist film studies in particular—is marked by a similar presence of “essentialist detectors.” For virtually any mention of “real women” (especially insofar as authors are concerned) tends to inspire a by-now-familiar recitation of the “dangers” of essentialism—an affirmation of the difference between men and women as given, and an attendant belief in the positive value of female identity which, repressed by patriarchy, will be given its true voice by feminism. While there is obviously much to be said about the risks of essentialism, the contemporary practice of essentialism detection has avoided the complex relationship between “woman” and “women,” usually by bracketing the category of “women” altogether.

Even though discussions of the works of women filmmakers have been central to the development of feminist film studies, theoretical discussions of female authorship in the cinema have been surprisingly sparse. While virtually all feminist critics would agree that the works of Germaine Dulac, Maya Deren, and Dorothy Arzner (to name the most frequently invoked “historical figures”) are important, there has been considerable reluctance to use any of them as privileged examples to theorize female authorship in the cinema, unless, that is, such theorizing affirms the difficulty of women’s relationship to the cinematic apparatus. This reluctance reflects the current association of “theory” with “antiessentialism.” In the realm of feminist literary theory and criticism, however, antiessentialism has not had quite the same widespread effects of negation. In the works of critics such as Margaret Homans and Nancy K. Miller, for instance, female authorship is analyzed not in terms of simple categories of agency and authority, but rather in terms of complex textual and cultural processes which dramatize and foreground women’s relationships to language, plot, and the institutions of literature.¹ My point is not that feminist film critics have the proverbial “much to learn” from feminist literary critics, but rather that the paradigm of female authorship in literature may provide a useful point of departure to examine the status of female authorship in the cinema.

For such a point of departure, I turn to two anecdotes, one “literary” and one “theoretical,” both of which stage an encounter between women’s writing and the cinema in similar ways. My first anecdote, the more “literary” one, concerns two contemporary novels by women concerned with the vicissitudes of female writing. In both novels, cinema becomes a persuasive metaphor for the difficult and sometimes impenetrable obstacles that confront the woman writer. Doris Lessing’s novel The Golden Notebook explores the relations between female identity and artistic production, and a formulation of that relation is represented through the cinema. Woman is the viewer, man the projectionist, and the whole viewing process a form of control and domination. Writer Anna Wulf describes her vision of events from her own past as films shown to her by an invisible male projectionist. The films represent what Anna calls the “burden of recreating order out of the chaos that my life had become.” Yet Anna is horrified by this vision of cinematic order:

They were all, so I saw now, conventionally, well-made films, as if they had been done in a studio; then I saw the titles: these films, which were everything I hated most, had been directed by me. The projectionist kept running these films very fast, and then pausing on the credits, and I could hear his jeering laugh at Directed by Anna Wulf. Then he would run another few scenes, every scene glossy with untruth, false and stupid.²

Lessing’s cinematic metaphor is informed by a relationship between viewer and image, and between projectionist and screen, that is profoundly patriarchal in the sense that separation, hierarchy, and power are here synonymous with sexual division. The conventionality and gloss of untruth of the films are complicit with Anna Wulf’s alienation from her name that appears on them. If, for Lessing, the conditions of film viewing suggest patriarchal domination, then the most immediate terms of that metaphor are the simultaneous evocation and denial of female authorship. Cinema embodies distance from the self—or at least, distance from the female self, a distance produced by the mockery of female authorship in the titles of the film. As evoked in The Golden Notebook within the context of the female narrator’s relationship to language and to experience, cinema functions as a particularly and peculiarly negative inflection of the female authorial signature.

In her novel The Quest for Christa T., Christa Wolf evokes the cinema as a form of illusory presence, as a fantasy control of the past. The female narrator of The Quest for Christa T. describes her search for Christa, as well as for the very possibility of memory: “I even name her name, and now I’m quite certain of her. But all the time I know that it’s a film of shadows being run off the reel, a film that was once projected in the real light of cities, landscapes, living rooms.”³ Film may create images of the past, but the images are contained by a reified memory. The cinema thus suggests a past that has been categorized, hierarchized, and neatly tucked away.⁴ Like Lessing’s Anna Wulf, the narrator of The Quest of Christa T. searches for the connections between female identity and language. While less explicit in its patriarchal configuration, Wolf’s metaphor nonetheless posits cinema as resistant to the process of active searching generally, and female self-expression specifically. The female narrator in The Quest for Christa T. is engaged, in Wolf’s words, in a search for “the secret of the third person, who is there without being tangible and who, when circumstances favor her, can bring down more reality upon herself than the first person: I.” As in The Golden Notebook, cinema obstructs the writing of female self-representation, thus embodying what Wolf calls “the difficulty of saying ‘I.’ ”⁵

If we are to take Lessing’s and Wolf’s metaphoric representations of the cinema at their word, then the difficulty of saying “I” for the woman filmmaker is far greater than for the woman writer. Feminist interrogations of the cinema have supported Lessing’s and Wolf’s metaphors, for the narrative and visual staging of cinematic desire relies, as most theoretical accounts would have it, on the massive disavowal of sexual difference and the subsequent alignment of cinematic representation with male-centered scenarios. To be sure, one could argue—with more than a touch of defensiveness—that such metaphoric renderings of the cinema suggest the strategic importance of the works of women filmmakers. For if the cinema is symptomatic of alienation (Lessing) and reification (Wolf), then the attempts by women directors to redefine, appropriate, or otherwise reinvent the cinema are crucial demonstrations that the boundaries of that supremely patriarchal form are more permeable, more open to feminist and female influence, than these film-inspired metaphors would suggest. At the same time, it could be argued that the works of women filmmakers offer reformulations of cinematic identification and desire, reformulations that posit cinematic metaphors quite different from those in the passages from Lessing and Wolf cited above. In other words, the “difficulty of saying ‘I’ ” does not necessarily mean that female authorship is impossible in the cinema, but rather that it functions differently than in literature.

If Lessing’s and Wolf’s formulations reflect the spirit of much feminist writing about film, suggesting that the cinema is peculiarly and forcefully resistant to the female creator, yet another obstacle to the theorizing of female authorship in the cinema emerges when the literary comparison is pursued in another direction. My second, more properly “theoretical” anecdote of the relationship between female authorship in its literary and cinematic forms is drawn from the introduction to Revision, a collection of essays on feminist film theory and criticism. The editors of the volume note that feminist film critics have “reason to be envious” of those feminist critics working in literature who “were able to turn to a comparatively substantial canon of works by women writers.” Unlike literature, the cinema has no such evidence of a female-authored cinema to which feminist critics might logically turn to begin to elaborate the components of women’s cinema or of a feminist film aesthetic. “For where in the classic cinema,” the editors ask, “do we encounter anything like an ‘autonomous tradition,’ with ‘distinctive features’ and ‘lines of influence’? And if, with some difficulty, we can conceive of Lois Weber and Dorothy Arzner as the Jane Austen and George Eliot of Hollywood, to whom do they trace their own influence?”⁶

While feminist literary critics have their own disagreements about the validity of the concept of a “female tradition” (autonomous or not) or a female “canon” (substantial or not), it is true that feminist film critics simply do not have the body of evidence to suggest how and in what ways female-authored cinema would be substantially different from cinema directed and created by men. The absence of this body of evidence notwithstanding, however, it seems to me that the reluctance of many feminist critics to speak, as feminist literary critics do, of a “female tradition” in cinema had to do with a number of other factors, ranging from theoretical frameworks in which any discussions of “personhood” are suspect, to the peculiar status of authorship in the cinema. Particularly insofar as the classical Hollywood cinema is concerned, the conventional equation of authorship with the role of the film director can repress or negate the significant ways in which female signatures do appear on film. For instance, consideration of the role of the often-forgotten, often-female screenwriter might suggest more of a female imprint on the film text; and the role of the actress does not always conform to common feminist wisdom about the controlling male gaze located in the persona of the male director—witness Bette Davis as a case in point.⁷

The reluctance to speak of a “female tradition” has perhaps been most influenced, however, by the fear of essentialism—the fear, that is, that any discussion of “female texts” presumes the uniqueness and autonomy of female representation, thus validating rather than challenging the dualism of patriarchal hierarchy. However, the act of discarding the concept of female authorship and of an attendant female tradition in the cinema as necessarily compromised by essentialist definitions of woman can be equally dualistic, in assuming that the only models of connection and influence are unquestionably essentialist ones. Sometimes it is assumed that any discussion of authorship is a throwback to the era of biographical criticism, to the text as transparent and simple reflection of the author’s life. While the limitations of such an approach are obvious, purely textual models of cinematic representation have their own limitations insofar as the narrative strategies of many contemporary women’s films are concerned, for these strategies frequently involve an inscription of authorship in literal terms, with the director herself a performer in her film.

Any discussion of female authorship in the cinema must take into account the curious history of definitions of cinematic authorship in general.⁸ It was not really until the 1950s that “auteurism” became a fixture of film theory and criticism. The French term did not connote then, as French terms have in the past two decades of film studies, a particularly complex entity. For auteurism refers to the view that the film director is the single force responsible for the final film, and that throughout the films of a given auteur a body of themes and preoccupations will be discernible.⁹ The obviousness of these claims is complicated, rather, by the fact that the object of inquiry for auteurist critics was primarily the Hollywood cinema. To speak of a “Hitchcock” or a “John Ford” or a “Nicholas Ray” film as opposed to an “MGM” or a “John Wayne” film was, if not a necessarily radical enterprise, then at least a historically significant one, in that a shift was marked in the very ways in which one speaks of film. For the corporate industrial model of film production was being challenged by a liberal humanist one, and “Hitchcock” does not carry quite the same capitalist, industrial, or corporate baggage as “MGM.”

Despite the opposition between the industry and the creative individual from which auteurism emerges, however, the terms do not differ all that much in their partiarchal connotations; “MGM” and “Hitchcock” may be patriarchal in different ways, but they share a common ground. The cinematic auteur was identified as a transcendental figure resistant to the leveling forces of the Hollywood industry; to use Roland Barthes’s words, the auteur theory in cinema reinstated the “formidable paternity” of the individual creator threatened by the institutions of mass culture of which the cinema is a paradigmatic and even privileged example.¹⁰ Thus it does not require too much imagination to see Alexandre Astruc’s famous equation of the camera with a writer’s pen, in his phrase “caméro-stylo,” as informed by the same kind of metaphorical equivalence between pen and penis that has defined both the Western literary tradition (symptomatically) and feminist literary history (critically).¹¹ The phallic denominator can be read several ways, most obviously as a denial of the possibility of any female agency. Conversely, it can be argued that the privileging of female authorship risks appropriating, for women, an extremely patriarchal notion of cinematic creation. At stake, then, is whether the adjective female in female authorship inflects the noun authorship in a way significant enough to challenge or displace its patriarchal and proprietary implications.

Whether authorship constitutes a patriarchal and/or phallocentric notion in its own right raises the specter of the “Franco-American Dis-connection” (to use Domna Stanton’s phrase) that has been the source of much critical debate, or confusion, depending upon your point of view, in contemporary feminist theory.¹² The position usually described as “American”—and therefore empirical and historical—would claim female authorship as basic to the goals of a feminist appropriation of (cinematic) culture, and the position described as “French”—theoretical and deconstructive—would find “authorship” and “appropriation” equally complicitous in their mimicry of patriarchal definitions of self, expression, and representation.¹³ Although it is a commonly held assumption that contemporary film studies, especially as they developed in England, are virtually synonymous with “French theory,” the fate of auteurism, particularly in relationship to feminist film theory, has not followed such an easily charted or one-directional path. In a famous 1973 essay, for example, Claire Johnston argued against the dismissal of the auteur theory. While acknowledging that “some developments of the auteur theory have led to a tendency to deify the personality of the (male) director,” Johnston argues nonetheless for the importance of auteurism for feminism. She notes that “the development of the auteur theory marked an important intervention in film criticism: its polemics challenged the entrenched view of Hollywood as monolithic, and stripped of its normative aspects the classification of films by director has proved an extremely productive way of ordering our experience of the cinema.”¹⁴

Johnston’s argument recalls Peter Wollen’s writings on auteur theory, where the cinematic auteur is defined less as a creative individual and more as a figure whose imprint on a film is measured by the repetition of sets of oppositions and the network of preoccupations, including unconscious ones.¹⁵ Her analysis needs to be seen in the context of a certain moment in feminist criticism, when notions of “good roles” for women (and therefore “positive” versus “negative” images) had much critical currency. Johnston turns that critical currency on its head in a comparison of Howard Hawks and John Ford. She argues that the apparently more “positive” and “liberated” heroines of Hawks’s films are pure functions of male desire. For John Ford, women function in more ambivalent ways. Whereas in Hawks’s films woman is “a traumatic presence which must be negated,” in Ford’s films woman “becomes a cipher onto which Ford projects his profoundly ambivalent attitude to the concepts of civilisation and psychological ‘wholeness.’ ”¹⁶ Defined as a narrative and visual system associated with a given director, Johnston’s auteurism allows for a kind of analysis which goes beyond the categories of “good” and “bad” (images or roles) and into the far more productive critical territory of symptom and contradiction.

While Johnston’s analysis seems to stress equally the importance of auteurism and of “symptomatic readings,” her work is read today far more in the context of the latter. As with Peter Wollen’s work on authorship, one senses that perhaps the auteurist part is a backdrop upon which more significant critical and theoretical assumptions are projected—those of structuralism and semiotics in the case of Wollen, and those of Althusserian-based critical readings in the case of Johnston. The kind of analysis for which Johnston argues—analysis of the position of “woman” within the narrative and visual structures of the cinema—has by and large been pursued without much direct consideration for the auteur theory, or for auteurs. ¹⁷ Despite the importance of auteurism in staking out what Johnston would call progressive claims for film criticism, the analysis of the kinds of structures to which Johnston alludes in the films of Hawks or Ford has been pursued within the framework of textual and ideological analyses of that ubiquitous entity, the classical Hollywood cinema, rather than within the scope of authorship.

By and large, the preferred mode of textual analysis, given its particular attention to unconscious resonances within narrative and visual structures, has had little room for an exploration of auteurism. One notable exception is Raymond Bellour’s analyses of Hitchcock’s role as “enunciator” in his films, which nonetheless define authorship in explicity literal and narrow textual terms—i.e., the fact that Hitchcock’s famous cameo appearances in his films occur at crucial moments of the exposition and/or resolution of cinematic desire.¹⁸ More frequently in contemporary film studies, one speaks of a “Hawks” film or a “Ford” film in the same way one would speak of a “horror” film or a “film noir”—as a convenient categorization of films with similar preoccupations and similar stylistic and narrative features. Such a demystification of authorship might well be more progressive than Johnston’s defense of authorship. Conversely, authorship itself may have assumed a symptomatic status, in which case it has not been demystified so much as concealed within and displaced onto other concerns, evoking a process similar to what Nancy K. Miller has observed in the field of literary studies, where the concept of authorship has been not so much revised as it has been repressed “in favor of the (new) monolith of anonymous textuality.”¹⁹

In film theory and criticism of the last decade, auteurism is rarely invoked, and when it is, it is more as a curiosity, as a historical development surely influential, but even more surely surpassed. In this context, Kaja Silverman has suggested that a curious slippage occurs in feminist discussions of the avantgarde works of women filmmakers, for the concept of authorship—largely bracketed in textual analysis—reappears, but in an extratextual way.

The author often emerges. . . as a largely untheorized category, placed definitively “outside” the text, and assumed to be the punctual source of its sounds and images. A certain nostalgia for an unproblematic agency permeates much of the writing to which I refer. There is no sense in which the feminist author, like her phallic counterpart, might be constructed in and through discourse—that she might be inseparable from the desire that circulates within her texts, investing itself not only in their formal articulation, but in recurring diegetic elements.²⁰

Silverman recommends a theorization of female authorship that would account for a diversity of authorial inscriptions, ranging from thematic preoccupations, to the designation of a character or group of characters as a stand-in for the author, to the various enunciative strategies (sonoric as well as visual) whereby the film auteur’s presence is marked (whether explicitly or implicitly), to the “fantasmatic scene” that structures an author’s work.²¹

The concept of female authorship in the cinema may well have a currency similar to categories of genre or of style. But can female authorship be so easily assimilated to the existing taxonomy of the cinema? Present categories of authorship are undoubtedly much more useful in analyzing the configurations of “woman” on screen than in coming to terms with the ways in which women directors inflect cinematic practice in new and challenging ways. The analysis of female authorship in the cinema raises somewhat different questions than does the analysis of male authorship, not only for the obvious reason that women have not had the same relationship to the institutions of the cinema as men have, but also because the articulation of female authorship threatens to upset the erasure of “women” which is central to the articulation of “woman” in the cinema. Virtually all feminist critics who argue in defense of female authorship as a useful and necessary category assume the political necessity for doing so. Hence, Kaja Silverman urges that the gendered positions of libidinal desire within the text be read “in relation to the biological gender of the biographical author, since it is clearly not the same thing, socially or politically, for a woman to speak with a female voice as it is for a man to do so, and vice versa.”²² The notion of female authorship is not simply a useful political strategy; it is crucial to the reinvention of the cinema that has been undertaken by women filmmakers and feminist spectators.

One of the most productive ironies of feminist theory may be that, if “woman” and “women” do not coincide (to borrow Teresa de Lauretis’s formulation), they also connect in tenuous and often complex ways. It is customary in much feminist film theory to read “subject” to “object” as “male” is to “female.” But a more productive exploration of female authorship insofar as “woman” and “women” are concerned may result when subject-object relationships are considered within and among women. Visions of “woman” that appear on screen may be largely the projections of patriarchal fantasies, and the “women” who make films and who see them may have problematic relations at best with those visions. While it is tempting to use de Lauretis’s distinction as an opposition between traditional cinematic representations of “woman” and those “women” filmmakers who challenge and reinvent them, the gap, the noncoincidence are better defined by exploring the tensions within both “woman” and “women.”²³

One such strategy has been directed toward the “reading against the grain” of traditional cinematic representations of women, demonstrating how they can be read in ways that contradict or otherwise problematize their function within male-centered discourse.²⁴ Surprisingly little comparable attention has been paid, however, to the function and position of the woman director. Central to a theorizing of female authorship in the cinema is an expanded definition of textuality attentive to the complex network of intersections, distances, and resistances of “woman” to “women.” The challenge of female authorship in the cinema for feminist theory is in the demonstration of how the divisions, overlaps, and distances between “woman” and “women” connect with the contradictory status of cinema as the embodiment of both omnipotent control and individual fantasy.

The feminist rediscovery of Dorothy Arzner in the 1970s remains the most important attempt to theorize female authorship in the cinema. Arzner may not be feminist film theory’s answer to George Eliot, but her career as a woman director in Hollywood with a significant body of work (and in whose work—true to the most rudimentary definitions of film authorship—a number of preoccupations reappear) has posed issues most central to a feminist theory of female cinematic authorship. As one of the very small handful of women directors who were successful in Hollywood, particularly during the studio years, Arzner has served as an important example of a woman director working within the Hollywood system who managed, in however limited ways, to make films that disturb the conventions of Hollywood narrative.

The significance of this argument, advanced primarily by Claire Johnston and Pam Cook, in which Arzner is defined as a director “critical” of the Hollywood cinema, needs to be seen in the context of the development of the notion of the film auteur. Arzner was very definitely not one of the directors for whom auteurist claims were made in the heyday of auteurist criticism. For despite the core themes and preoccupations visible across her work, Arzner does not satisfy any of the specific requirements of cinematic authorship as they were advanced on either side of the Atlantic—there is little of the flourish of mise-en-scène that auteurists attributed to other directors, for instance, and the preoccupations visible from film to film that might identify a particular signature do not reflect the life-and-death, civilization-versus-the-wilderness struggles that tended to define the range of more “properly” auteurist themes.²⁵

Given the extent to which feminist analysis of the cinema has relied on the distinction between dominant and alternative film, the claims that can be made for an alternative vision that exists within and alongside the dominant cinema will be crucial in gauging the specific ways in which women directors engage with “women’s cinema” as divided between representations that perpetuate patriarchal definitions of femininity, and representations that challenge them and offer other modes of identification and pleasure. One can read in responses to Arzner’s work reflections of larger assumptions concerning the Hollywood cinema. At one extreme is Andrew Britton’s assessment of Arzner, in his study of Katharine Hepburn, as the auteur of Christopher Strong (1933), the film in which Hepburn appears as an aviatrix who falls in love with an older, married man.

That Christopher Strong functions as a “critique of the effect of patriarchal heterosexual relations on relations between women” suggests that the classical cinema lends itself quite readily to heterogeneity and conflicting ideological allegiances, whether the “critique” is the effect of the woman director or the female star.²⁶ At the opposite extreme, Jacquelyn Suter’s analysis of Christopher Strong evolves from the assumption that whatever “female discourse” there is in the film is subsumed and neutralized by the patriarchal discourse on monogamy.²⁷ If the classical cinema described by Britton seems remarkably open to effects of subversion and criticism, the classical cinema described by Suter is just as remarkably closed to any meanings but patriarchal ones, and one is left to assume that female authorship is either a simple affirmation of agency, or virtually an impossibility as far as Hollywood cinema is concerned.

In contrast, Claire Johnston’s analyses of Arzner are reminiscent of Roland Barthes’s description of Balzac as representative of a “limited plurality” within classical discourse.²⁸ For Johnston suggests that the strategies of her films open up limited criticisms of the Hollywood cinema. Johnston’s claims for female authorship in Arzner’s films rely on notions of defamiliarization and dislocation, and more precisely on the assumption popularized within film studies, primarily by Jean-Louis Comolli and Jean Narboni, that there exists within the classical Hollywood cinema a category of films in which realist conventions are criticized from within, generating a kind of internal critique. Claims for this “progressive” text have been made from a variety of vantage points, virtually all of them concerned with ideological value—with, that is, the possibility of a Hollywood film that critiques the very values that are ostensibly promoted, from the literal dark underside of bourgeois ideology “exposed” in Young Mr. Lincoln to the impossibility of familial ties for women in Mildred Pierce.²⁹ For Johnston, female desire is the auteurist preoccupation that generates a critique of patriarchal ideology in Arzner’s films.

Initially, Johnston’s analysis of Arzner appears to rely on a definition of the classical cinema that allows for more heterogeneity and more articulation of contradiction than is the case in those analyses that posit a rigid distinction between the classical cinema and its alternatives. However, Arzner’s films can be identified as “progressive” only in relationship to a norm that allows for no divergences from purely classical filmmaking. More problematical within the present context, there is nothing in this kind of analysis to suggest what these marks of dislocation and critique have to do with distinctly female authorship. Many “woman’s films” are motivated by the representation of female desire, and feminist critics have shown how these films might also be read as driven by such an internal—if often unconscious—critique.³⁰ It is not clear, in other words, to what extent the fact of female authorship gives a particular or distinct inflection to the representation of female desire.

The “political” reasons for insisting on the relevance of the author’s gender are not adequate in and of themselves, for they can easily harden into an idealized abstraction, and the name “Dorothy Arzner” would thus become just one more signature to add to the pantheon of (male) directors who critique the conventions of Hollywood cinema from within. While the importance of Arzner’s signature in extratextual terms is undeniable, stressing that importance should not be a substitute for an examination of the textual ramifications of female authorship. Yet Johnston’s approach to those textual ramifications in Arzner’s work seems torn between female authorship understood (“politically”) as agency and self-representation, on the one hand, and as a negative inflection of the norms of classical cinema, on the other.³¹ This ambivalence—which could be read in terms of the conflicting claims of the so-called American and French positions—is not particularly productive, for the agency thus affirmed dissolves into negation and the impossibility of a female position, evoking Julia Kristéva’s extremely limited hypothesis that “women’s practice can only be negative, in opposition to that which exists, to say that ‘this is not it’ and ‘it is not yet.’ ”³²

Noting that structural coherence in Arzner’s films comes from the discourse of the woman, Johnston relies on the notion of defamiliarization, derived from the Russian formalists’ priem ostranenie, the “device of making strange,” to assess the effects of the woman’s discourse on patriarchal meaning: “the work of the woman’s discourse renders the narrative strange, subverting and dislocating it at the level of meaning.”³³ Johnston discusses in this context what has become the single most famous scene from any of Arzner’s films, when Judy (Maureen O’Hara), who has played ballet stooge to the vaudeville performer Bubbles (Lucille Ball) in Dance, Girl, Dance (1940), confronts her audience and tells them how she sees them. This is, Johnston argues, the only real break between dominant discourse and the discourse of the woman in Arzner’s work, and it is a break that is quickly recuperated within the film, for the audience applauds Judy, and she and Bubbles are quickly dispatched to center stage, where they engage in a catfight, to the delight of the audience. The moment in Dance, Girl, Dance when Judy faces her audience is a privileged moment in feminist film theory and criticism, foregrounding as it does the sexual hierarchy of the gaze, with female agency defined as the return of the look, thus “citing” the objectification of woman.³⁴

The celebrity accorded this particular scene in Arzner’s film needs to be evaluated in the context of feminist film theory in the mid-1970s. Confronted with the persuasive psychoanalytically based theoretical model according to which women either did not or could not exist on screen, the discovery of Arzner, and especially of Judy’s “return of the gaze,” offered some glimmer of historical hope as to the possibility of a female intervention in the cinema. To be sure, the scope of the intervention is limited, for as Johnston herself stresses, Judy’s radical act is quickly recuperated within the film when the audience gets up to cheer her on, and she and Bubbles begin to fight on stage. But the need to revise Johnston’s model of authorship is most apparent in this reading of recuperation, for it is informed by the assumption that such a “break” can be only a brief eruption, and can occur in classical cinema only if it is then immediately contained within the laws of male spectatorial desire.

Only one kind of look (Judy’s return of the look to her audience) and one kind of spectacle (where men are the agents of the look and women its objects) have received attention in Dance, Girl, Dance. In other words, the disruptive force of female desire central to Arzner’s work exists primarily within the symmetry of masculinity and femininity.³⁵ However, I would suggest that female authorship acquires its most significant contours in Arzner’s work through relations between and among women. The female gaze is defined early on in the film as central to the aspirations of women as they are shaped within a community of women. Madame Basilova, the older woman who is in charge of the dancing troupe of which Judy and Bubbles are a part, is seen gazing through the rails of a stairway as Judy practices her ballet, and the gaze of Judy herself is isolated as she looks longingly at a rehearsal of the ballet company which she wishes to join. Even Judy’s famous scolding of the audience is identified primarily as a communication, not between a female performer and a male audience (the audience is not, in any case, exclusively male) but between the performer and the female member of the audience (secretary to Steven Adams, the man who will eventually become Judy’s love interest) who stands up to applaud her.³⁶ And the catfight that erupts between Judy and Bubbles on stage is less a recuperative move—i.e., transforming the potential threat of Judy’s confrontation into an even more tantalizing spectacle—than the claiming by the two women of the stage as an extension of their conflicted friendship, rather than as the alienated site of performance.

To be sure, the men—promoters as well as onlookers—eagerly consume the spectacle of Judy and Bubbles in a catfight. But I see this response less as a sign of pure recuperation by the male-centered system of looks and spectacles, and more as the dramatization of the tension between performance and self-expression which the film attempts to resolve. Although Johnston is more concerned with the devices that give Arzner’s films “structural coherence,” it is tempting to conclude from her analysis that Judy functions as a metaphoric rendering of the woman filmmaker herself, thus establishing something of a homology between Arzner’s position vis-à-vis the classical Hollywood cinema and Judy’s position on stage.³⁷ The stage is, in other words, both the site of the objectification of the female body and the site for the theatricalizing of female friendship. This “both/and”—the stage (and, by metaphoric implication, the cinema itself) is an arena simultaneously of patriarchal exploitation and of female self-representation—stands in contrast to the more limited view of Arzner’s films in Johnston’s work, where more of a “neither/nor” logic is operative—neither patriarchal discourse nor the “discourse of the woman” allows women a vantage point from which to speak, represent, or imagine themselves.

Reading Arzner’s films in terms of the “both/and” suggests an irony more far-reaching than that described by Johnston. Johnston’s reading of Arzner is suggestive of Shoshana Felman’s definition of irony as “dragging authority as such into a scene which it cannot master, of which it is not aware and which, for that very reason, is the scene of its own self-destruction. . . . ”³⁸ The irony in Dance, Girl, Dance, however, does not just demonstrate how the patriarchal discourse of the cinema excludes women, but rather how the cinema functions in two radically different ways, both of which are “true,” as it were, and totally incompatible. I am borrowing here from Donna Haraway’s definition of irony: “Irony is about contradictions that do not resolve into larger wholes, even dialectically, about the tension of holding incompatible things together because both or all are necessary and true.”³⁹ This insistence on two equally compelling and incompatible truths constitutes a form of irony far more complex than Johnston’s analysis of defamiliarization.

Johnston’s notion of Arzner’s irony assumes a patriarchal form of representation which may have its gaps and its weak links, but which remains dominant in every sense of the word. For Johnston, Arzner’s irony can be only the irony of negativity, of puncturing holes in patriarchal assumptions. Such a view of irony has less to do, I would argue, with the limitations of Arzner’s career (e.g., as a woman director working within the inevitable limitations of the Hollywood system) than with the limitations of the film theory from which it grows. If the cinema is understood as a one-dimensional system of male subjects and female objects, then it is not difficult to understand how the irony in Arzner’s films is limited, or at least would be read as limited. While rigid hierarchies of sexual difference are indeed characteristic of dominant cinema, they are not absolute, and Arzner’s films represent other kinds of cinematic pleasure and desire.

An assessment of Arzner’s importance within the framework of female authorship needs to account not only for how Arzner problematizes the pleasures of the cinematic institution as we understand it—e.g., in terms of the voyeurism and fetishism reenacted through the power of the male gaze and the objectification of the female body—but also for how, in her films, those pleasures are identified in ways that are not reducible to the theoretical clichés of the omnipotence of the male gaze. The irony of Dance, Girl, Dance emerges from the conflicting demands of performance and self-expression, which are linked in their turn to heterosexual romance and female friendship. Female friendship acquires a resistant function in the way that it exerts a pressure against the supposed “natural” laws of heterosexual romance. Relations between women and communities of women have a privileged status in Arzner’s films. To be sure, Arzner’s films offer plots—particularly insofar as resolutions are concerned—that are compatible with the romantic expectations of the classical Hollywood cinema; communities of women may be central, but boy still meets girl.

Claire Johnston claims that the conclusion of Dance, Girl, Dance, where Judy is embraced by Steven Adams, destined for a fusion of professional mentoring and romance, is marked by Judy’s defeat. This strikes me more as wishful feminist thinking than as a convincing reading of the film’s conclusion, which “works” within the conventions of Hollywood romance. Noting Judy’s final comment as she is in Steven’s arms—“when I think how simple things could have been, I just have to laugh”—Johnston says that “this irony marks her defeat and final engulfment, but at the same time it is the final mark of subversion of the discourse of the male.”⁴⁰ If the “discourse of the male” is subverted in Dance, Girl, Dance, it has less to do with the resolution of the film and more to do with the process of heterosexual initiation which the film has traced. Judy’s attractions to men are shaped by substitutions for women and female rivalry—Steven Adams is a professional mentor to substitute for Basilova, and Jimmie Harris is an infantile man who is desirable mainly because Bubbles wants him too.⁴¹ Therefore, the heterosexual romance provides the conclusion of the film, but only after it has been mediated by relationships between women.

A controversial area in feminist theory and criticism has been the connection between lesbianism and female friendship in those fictional worlds which, like Arzner’s, take communities of women as their inspiration. Barbara Smith’s suggestion that the relationship between Nel and Sula in Toni Morrison’s novel Sula can be read in lesbian terms has been provocative to say the least, particularly given Toni Morrison’s own assertion that “there is no homosexuality in Sula,”⁴² But the case of Arzner is somewhat different. What is known about Arzner implies that she herself was a lesbian.⁴³ But this assertion raises as many questions as it presumably answers, concerning both the responsibility of a critic vis-à-vis an individual who was presumably in the closet, and the compulsion to define lesbianism as something in need of proof.⁴⁴ Bonnie Zimmerman has suggested that “if a text lends itself to a lesbian reading, then no amount of biographical ‘proof’ ought to be necessary to establish it as a lesbian text.”⁴⁵ The point is well taken, but in Arzner’s case another “text” mediates the relationship between director, her films, and their reception. For Arzner’s films are virtually no longer read independently of her persona—an issue to which I will return momentarily. Nonetheless, if relationships between and among women account for much more narrative and visual momentum than do the relations between men and women in Arzner’s work, then one begins to wonder about the perspective that informs these preoccupations.

For all of the attention that has been given to Arzner’s work, one striking aspect of her persona—and of her films—has been largely ignored. Although the photographs of Arzner that have accompanied feminist analyses of her work depict a woman who favored a look and a style connoting lesbian identity, discussions of her work always stop short of any recognition that sexual identity might have something to do with how her films function, particularly concerning the “discourse of the woman” and female communities, or that the contours of female authorship in her films might be defined in lesbian terms. This marginalization is all the more notable, given how visible Arzner has been as an image in feminist film theory. With the possible exception of Maya Deren, Arzner is more frequently represented visually than any other woman director central to contemporary feminist discussions of film. And unlike Deren, who appeared extensively in her own films, Arzner does not have the reputation of being a particularly self-promoting, visible, or “out” (in several senses of the term) woman director.

Sarah Halprin has suggested that the reason for this omission is, in part, the suspicion of any kind of biographical information in analysis of female authorship:

most discussions of Dorothy Arzner’s films, especially those by the English school, carefully avoid any mention of Arzner’s appearance in relation to some of the images in her films. Lengthy analyses of Dance, Girl, Dance ignore the fact that while the “main” characters, Judy and Bubbles, are recurrently placed as immature within the context of the film, there are two “minor” characters who both dress and look remarkably similar to Arzner herself (i.e., tailored, “mannish,” in the manner of Radclyffe Hall and other famous lesbians of the time) and are placed as mature, single, independent women who are crucial to the career of young Judy and who are clearly seen as oppressed by social stereotyping, of which they are contemptuous. Such a reading provides a whole new way of relating to the film and to other Arzner films, encouraging a discussion of lesbian stereotypes, relations between lesbians and heterosexual women as presented in various films and as perceived by any specific contemporary audience.⁴⁶

Indeed, one of the most critical aspects of Arzner’s work is the way in which heterosexuality is assumed equivocally, without necessarily violating many of the conventions of the Hollywood film.

In his book on gay sexuality and film, Vito Russo quotes another Hollywood director on Arzner: “an obviously lesbian director like Dorothy Arzner got away with her lifestyle because she was officially closeted and because ‘it made her one of the boys.’ ”⁴⁷ An interview with Arzner by Karyn Kay and Gerald Peary gives some evidence of her status as “one of the boys,” at least insofar as identification is concerned, for in discussing both Christopher Strong and Craig’s Wife, Arzner insists that her sympathies lie with the male characters.⁴⁸ However, one has only to look at the photographs of Arzner that have accompanied essays about her work in recent years to see that this is not a director so easily assimilated to the boys’ club of Hollywood. Arzner preferred masculine attire, in the manner, as Halprin says, of Radclyffe Hall. Two dominant tropes shape the photographic mise-en-scène of the Arzner persona. She is portrayed against the backdrop of the large-scale apparatus of the Hollywood cinema, or she is shown with other women, usually actresses, most of whom are most emphatically “feminine,” creating a striking contrast indeed.

Both tropes appear in the photograph on the cover of the collection edited by Claire Johnston, The Work of Dorothy Arzner: Towards a Feminist Cinema. We see Arzner in profile, slouching directorially on a perch next to a very large camera; seated next to her is a man. They both look toward what initially appears to be the unidentified field of vision (Figure 14). When the pamphlet is opened, however, the photograph continues on the back cover: two young women, one holding packages, look at each other, their positions reflecting symmetrically those of Arzner and her male companion (Figure 15). It is difficult to read precisely the tenor of the scene (from Working Girls [1930]) between the two women: some hostility perhaps, or desperation. The camera occupies the center of the photograph, as a large, looming—and predictably phallic—presence (Figure 16). The look on the man’s face suggests quite strongly the clichés of the male gaze that have been central to feminist film theory; from his perspective the two women exist as objects of voyeuristic pleasure. Arzner’s look has quite another function, however, and one that has received very little critical attention, and that is to decenter the man’s look, and to eroticize the exchange of looks between the two women.⁴⁹

Figure 12. Dorothy Arzner (Museum of Modern Art)

Figure 13. Dorothy Arzner and Billie Burke (Museum of Modern Art)

While virtually none of the feminist critics who analyze Arzner’s work have discussed lesbianism, a curious syndrome is suggested by this use of “accompanying illustrations.” The photograph on the covers of the pamphlet edited by Claire Johnston speaks rather literally what is unspoken in the written text, in a teasing kind of way. Johnston’s and Cook’s essays are reprinted in a recent collection of essays on feminist film theory, not one of which discusses erotic connections between women.⁵⁰ Yet on the cover of the book is a photograph of Arzner and Rosalind Russell exchanging a meaningful look with more than a hint of female homoeroticism. One begins to suspect that the simultaneous evocation and dispelling of an erotic bond between women in Arzner’s work is a structuring absence in feminist film theory. Like any good symptom it rather obsessively draws attention to itself. Arzner’s lesbian persona may not be theorized in relationship to her films, but her visibility in feminist film criticism suggests that one of the mechanisms cited most frequently as central to male spectatorial desire informs feminist film theory, too. I am referring, of course, to fetishism.

To be sure, any parallel between a classical, male-centered trajectory such as fetishism and the dynamics of feminist theory can be made only tentatively.⁵¹ But there is nonetheless a telling fit between Octave Mannoni’s formula for disavowal (“I know very well, but all the same . . .”), adapted by Christian Metz to analyze cinematic fetishism, and the consistent and simultaneous evocation and disavowal of Arzner’s lesbian persona.⁵² A heterosexual master code, where any and all combinations of “masculinity,” from the male gaze to Arzner’s clothing, and “femininity,” from conventional objectification of the female body to the female objects of Arzner’s gaze, has shaped discussions of Arzner’s work. The narrative and visual structures of her films are praised for their “critique” of the Hollywood system, but the critique is so limited as to only affirm the dominance of the object in question.

Figure 14

Figure 15

Figure 16

The photographs of Arzner are interesting not only in the biographical terms suggested by Sarah Halprin, but also in textual terms. For one of the most distinctive ways in which Arzner’s authorial presence is felt in her films is in the emphasis placed on communities of women, to be sure, but also in the erotic charge identified within those communities. If heterosexual initiation is central to Arzner’s films, it is precisely in its function as rite of passage (rather than natural destiny) that a marginal presence is felt—an authorial presence that is lesbian, as well as female. Consider, for instance, Christopher Strong. Katharine Hepburn first appears in the film as a prize-winning object in a scavenger hunt, for she can claim that she is over twenty-one and has never had a love affair. Christopher Strong, the man with whom she will eventually become involved, is the male version of this prize-winning object, for he has been married for more than five years and has always been faithful to his wife. As Cynthia Darrington, Hepburn dresses in decidedly unfeminine clothing and walks with a swagger that is masculine, or athletic, depending upon your point of view. Hepburn’s jodhpurs and boots may well be, as Beverle Houston puts it, “that upper-class costume for a woman performing men’s activities.”⁵³ But this is also clothing that strongly denotes lesbian identity, and which (to stress again Sarah Halprin’s point) is evocative of the way Arzner herself, and other lesbians of the time, dressed.

Figure 17. Katharine Hepburn as aviatrix Cynthia Darrington in Dorothy Arzner’s Christopher Strong (Museum of Modern Art)

Figure 18. Katharine Hepburn and Colin Clive in Christopher Strong (Museum of Modern Art)

Cynthia’s “virginity” becomes a euphemistic catchall for a variety of margins in which she is situated, both as a woman devoted to her career and as a woman without a sexual identity. The film traces the acquisition of heterosexual identity, with some peculiar representations of femininity along the way, including Hepburn dressed as a moth. I am not arguing that Christopher Strong, like the dream which says one thing but ostensibly “really” means its mirror opposite, can be decoded as a coherent “lesbian film,” or that the real subject of the film is the tension between gay and straight identities. The critical attitude toward heterosexuality takes the form of inflections, of bits and pieces of tone and gesture and emphasis, as a result of which the conventions of heterosexual behavior become loosened up, shaken free of some of their identifications with the patriarchal status quo.

Most important, perhaps, the acquisition of heterosexuality becomes the downfall of Cynthia Darrington. Suter has described Christopher Strong in terms of how the feminine discourse, represented by the various female characters in the film, is submerged by patriarchal discourse, the central term of which is monogamy. The proof offered for such a claim is, as is often the case in textual analysis, convincing on one level but quite tentative on another, for it is a proof which begins from and ends with the assumption of a patriarchal master code. Even the “feminine discourse” described by Suter is nothing but a pale reflection of that master code, with nonmonogamy its most radical expression. The possibility that “feminine discourse” in Christopher Strong might exceed heterosexual boundaries is not taken into account.⁵⁴ As should be obvious by now, I am arguing that it is precisely in its ironic inflection of heterosexual norms, whether by the mirroring gesture that suggests a reflection of Arzner herself or by the definition of the female community as resistant to, rather than complicitous with, heterosexual relations, that Arzner’s signature is written on her films.

These two components central to female authorship in Arzner’s work—female communities and the mirroring of Arzner herself—are not identical. The one, stressing the importance of female communities and friendship among women, may function as a pressure exerted against the rituals of heterosexual initiation, but is not necessarily opposed to them. This foregrounding of relationships among women disturbs the fit between female friendship and heterosexual romance, but the fit is still there, the compatibility with the conventions of the classical Hollywood cinema is still possible. The representation of lesbian codes, as in the mirroring of Arzner’s—and other lesbians’—dress, constitutes the second strategy, which is more marginal and not integrated into narrative flow. These two authorial inscriptions—the emphasis on female communities, the citing of marginal lesbian gestures—are not situated on a “continuum,” that model of continuity from female friendship to explicit lesbianism so favored in much contemporary lesbian-feminist writing.⁵⁵ Rather, these two strategies exist in tension with each other, constituting yet another level of irony in Arzner’s work. Female communities are compatible with the classical Hollywood narrative; the lesbian gesture occupies no such position of compatibility, it does not mesh easily with narrative continuity in Arzner’s film.

Thus, in Dance, Girl, Dance, Arzner accentuates not only the woman’s desire as embodied in Judy and her relationships with other women, but also secondary female figures who never really become central, but who do not evaporate into the margins, either—such as the secretary (who leads the applause during Judy’s “return of the gaze” number) and Basilova (the dance teacher and director of the troupe). That these figures do not simply “disappear” suggests even more strongly their impossible relationship to the Hollywood plot, a relationship that is possible insofar as Judy is concerned. In Craig’s Wife, however, there is more of an immediate relationship between marginality and female communities, although in this case, the marginality has less of a lesbian inflection, both dress-and gesturewise. Julia Lesage has noted that in Craig’s Wife, Arzner rereads George Kelley’s play, the source of the film, so that the secondary women characters are treated much more fully than in the play.⁵⁶

Craig’s Wife—preoccupied with heterosexual demise rather than initiation—shows us a woman so obsessively concerned with her house that nothing else is of interest to her. Harriet Craig (Rosalind Russell) married as “a way towards emancipation. . . . I married to be independent.” If marriage is a business contract, then Harriet Craig’s capital is her house. Indeed, Harriet’s sense of economy is pursued with a vengeance. And the men in the film are the victims, explicit or not, of her obsession. It is Harriet’s husband who married for love, not money, and in a subplot of the film, a friend of Walter Craig is so obsessed by his wife’s unfaithfulness that he kills her and then himself.

At the conclusion of the film, virtually everyone has cleared out of Harriet’s house. Her niece has left with her fiance, her servants have either quit or been fired, and Walter has finally packed up and left in disgust. Harriet seems pathetically neurotic and alone. The widow next door (Billie Burke) brings Harriet some roses. In Kelley’s play, Harriet has become a mirror image of her neighbor, for both are portrayed as women alone, to be pitied. But in Arzner’s film, the neighbor represents Harriet’s one last chance for connection with another human being. Thus the figure who in Kelley’s play is a pale echo of Harriet, becomes in the film the suggestion of another identity and of the possibility of a female community. The resolution of Arzner’s Craig’s Wife has little to do with the loss of a husband, and more to do with situating Harriet Craig’s fantasy come horribly true alongside the possibility of connection with another woman. And while Billie Burke is hardly evocative of lesbianism (as Basilova is in Dance, Girl, Dance), she and Rosalind Russell make for a play of contrasts visually similar to those visible in photographs of Arzner with more “feminine” women.⁵⁷

Figure 19. Billie Burke and Rosalind Russell in Dorothy Arzner’s Craig’s Wife (Museum of Modern Art)

To be sure, Arzner’s authorship extends to an ironic perspective on patriarchal institutions in general, and in this sense her films do not require or assume a lesbian audience, as if this was or is likely to happen within the institutions of the Hollywood cinema. At the same time that the irony of Arzner’s films appeals to a wide range of female experiences, and is thus readable across a wide spectrum, ranging from lesbian to heterosexual and from female to feminist, the marks of female authorship in her work do not constitute a universal category of female authorship in the cinema. The female signature in Arzner’s work is marked by that irony of equally compelling and incompatible discourses to which I have referred, and the lesbian inflection articulates the division between female communities which do function within a heterosexual universe, and the eruptions of lesbian marginality which do not. This lesbian irony taps differing and competing views of lesbianism within contemporary feminist and lesbian theory. Lesbianism has been defined as the most intense form of female and feminist bonding, on the one hand; and as distinctly opposed to heterosexuality (whether practiced by women or men), on the other. In Arzner’s own time, these competing definitions were read as the conflict between a desexualized nineteenth-century ideal of romantic friendship among women, and the “mannish lesbian” (exemplified by Radclyffe Hall), defined by herself and her critics as a sexual being.⁵⁸ Arzner’s continued “visibility” suggests not only that the tension is far from being resolved, but also that debates about lesbian identity inform, even (and especially!) in unconscious ways, the thinking of feminists who do not identify as lesbians.

I see, then, several points to draw from the example of Dorothy Arzner as far as female authorship in the cinema is concerned. The preoccupations with female communities and heterosexual initiation are visible and readable only if we are attentive to how the cinema, traditionally and historically, has offered pleasures other than those that have received the most sustained critical and theoretical attention in recent years. Female authorship finds an inadequate metaphor in the female gaze as it returns the ostensibly central and overriding force of the male gaze. Other forms of the female gaze—such as the exchange of looks between and among women—open up other possibilities for cinematic meaning and pleasure and identification. In addition, a female signature can take other forms besides the gaze—costume and gesture, and the strategies of reading “marginality” in the case of Arzner. Textually, the most pervasive sign of female authorship in Arzner’s film is irony, and that irony is most appropriately described as the confrontation between two equally compelling, and incompatible, discourses.

I am suggesting, of course, that lesbian irony constitutes one of the pleasures in Arzner’s films, and that irony is a desirable aim in women’s cinema. Irony can misfire, however.⁵⁹ It has been argued that in Jackie Raynal’s film Deux Fois, for instance, the ironic elaboration of woman-as-object-of-spectacle is rendered decidedly problematic by the fact that it is only in offering herself as an object of spectacle that the category of woman-as-object-of-spectacle can be criticized; it is only by affirming the validity of patriarchal representation that any critique is possible.⁶⁰ I have in mind here another kind of misfiring, when the ironic reading of patriarchal conventions collides with other coded forms of representation which may serve, quite disturbingly, as a support for that irony.

In Dance, Girl, Dance, for instance, racial stereotypes emerge at three key moments in the narrative of the film. In the opening scenes, at a nightclub in Akron, Ohio, the camera moves over the heads of the members of the audience as it approaches the stage where the female dance troupe is performing. Intercut is an image of the black members of the band, who are smiling like the proverbially happy musicians. While an equivalence seems to be established between women and blacks as objects of spectacle, I see little basis for any “critical” use of the racist stereotype. Later in the film, when Judy watches longingly the rehearsal of the ballet company, another stereotype emerges. The performance number portrays the encounter between ballet and other forms of dance and body language within the context of the city. At one point in the performance, the music switches suddenly to imitate a jazzy tune, and a white couple in blackface struts across the stage. During one of the concluding scenes of the film, when Judy and Bubbles resolve their friendship in a court of law, the ostensibly “amusing” conclusion to the scene occurs when the clerk announces the arrival of a black couple whose names are “Abraham Lincoln Johnson” and “Martha Washington Johnson.”

However disparate these stereotypes, they do emerge at crucial moments in the deployment of irony and performance. In each case, the racial stereotype appears when the sexual hierarchy of the look is deflected or otherwise put into question. The black performers at the beginning of the film are defined securely within the parameters of objectification when it is apparent—much to Bubbles’ irritation and eventual attendant desire—that Jimmie Harris, one of the spectators in the audience, is totally unengaged in the spectacle on stage. The appearance of the white couple in blackface occurs when the centrality of Judy’s desire, as defined by her longing gaze at the performance, is affirmed. In the courtroom scene, Judy assumes aggressively and enthusiastically the court as her stage, and the racist stereotype of “Mr. and Mrs. Johnson” appears once the rivalry between the two women is on the verge of resolution.

In each of these instances, the stereotype affirms the distinction between white subject and black object just when the distinction between male subject and female object is being put into question. While there is nothing in Dance, Girl, Dance that approximates a sustained discourse on race, these brief allusions to racist stereotypes are eruptions that cannot be dismissed or disregarded as mere background or as unconscious reflections of a dominant cinematic practice that was racist. The marks of authorship in Dance, Girl, Dance include these racist clichés as well as the ironic inflection of the heterosexual contract. I want to stress that female irony is not just a function of sexual hierarchy, but that virtually all forms of narrative and visual opposition are potentially significant. To ignore, in Arzner’s case, the intertwining of sexual and racial codes of performance is to claim female authorship as a white preserve. The racist stereotypes which serve as an anchor of distinct otherness in Dance, Girl, Dance speak to a more general problem in female authorship. While Arzner’s films suggest other forms of cinematic pleasure that have been relatively untheorized within film studies, these forms cannot be posited in any kind of simple way as “alternatives.” I think it is a mistake to assume that the racist clichés are symptomatic of the compromises that inevitably occur with any attempt to create different visions within the classical Hollywood cinema. Such clichés are possible within virtually any kind of film practice.

Dorothy Arzner has come to represent both a textual practice (consciously) and an image (less consciously) in feminist film theory, and the relationship between the textual practice and the image suggests an area of fascination, if not love, that dares not speak its (her) name. The preferred term sexual difference in feminist film theory can slide from the tension between masculinity and femininity into a crude determinism whereby there is no representation without heterosexuality. Challenging that implicit homophobia would be reason enough to read the marks of lesbian authorship in Arzner’s work. There are two other issues which the designation of lesbian authorship crystallizes in particularly important ways. First, female authorship cannot be a useful concept if it perpetuates the notion of a monolithic essentialist identity, with a feminist inflection, perhaps, but no less problematic for that. Feminists have said frequently enough that when unchallenged, the notion of the “human subject” refers inevitably to a subject that is white, male, and heterosexual.

Similarly, the unexamined female subject may not be male, but is usually assumed to be nonetheless white and heterosexual. To be sure, there is much “female bonding”—to use the preferred phrase whereby lesbianism is usually repressed—in Arzner’s films, but that female bonding takes many forms, one of which is lesbian; and it is the lesbian inflection where Arzner’s authorial signature is most in evidence. Second, lesbianism raises some crucial questions concerning identification and desire in the cinema, questions with particular relevance to female cinematic authorship. Cinema offers simultaneous affirmation and dissolution of the binary oppositions upon which our most fundamental notions of self and other are based. In feminist film theory, one of the most basic working assumptions has been that in the classical cinema, at least, there is a fit between the hierarchies of masculinity and femininity on the one hand, and activity and passivity on the other. If disrupting and disturbing that fit is a major task for filmmakers and theorists, then lesbianism would seem to have a strategically important function. For one of the “problems” that lesbianism poses, insofar as representation is concerned, is precisely the fit between the paradigms of sex and agency, the alignment of masculinity with activity and femininity with passivity.

It is perhaps “no coincidence” that one discourse in which the “problem” of lesbianism is thus posed most acutely is psychoanalysis. For reasons both historical and theoretical, the most persuasive as well as controversial accounts of cinematic identification and desire have been influenced by psychoanalysis. Laura Mulvey’s classic account of sexual hierarchy in narrative cinema established the by-now-familiar refrain that the ideal spectator of the classical cinema, whatever his/her biological sex or cultural gender, is male. Many critics have challenged or extended the implications of Mulvey’s account, most frequently arguing that for women (and sometimes for men as well), cinematic identification occurs at the very least across gender lines, whether in transvestite or bisexual terms.⁶¹ However complex such accounts, they tend to leave unexamined another basic assumption common both to Mulvey’s account and to contemporary psychoanalytic accounts of identification, and that is that cinematic identification not only functions to affirm heterosexual norms but finds its most basic condition of possibility in the heterosexual division of the universe.

While feminist film theory and criticism have devoted extensive attention to the function of the male gaze in film, the accompanying heterosexual scenario has not received much attention, except for the occasional nod to what seems to be more the realm of the obvious than the explorable or questionable. Even David Bordwell, Kristin Thompson, and Janet Staiger’s The Classical Hollywood Cinema—a model of historical and formal precision—characterizes heterosexual love as a theme that links cinema with other historical forms in a very simple way. “Almost invariably,” writes David Bordwell, one of the main lines of action in the classical Hollywood cinema “involves heterosexual romantic love. This is, of course, not startling news.” Bordwell goes on to specify that in the sample of 100 films used in the study, “ninety-five involved romance in at least one line of action, while eighty-five made that the principal line of action.”

Thus, Bordwell concludes, “in this emphasis upon heterosexual love, Hollywood continues traditions stemming from the chivalric romance, the bourgeois novel, and the American melodrama.”⁶² The unbroken narrative connecting chivalric romance and Hollywood plots suggests, quite accurately of course, that heterosexual love is common to both forms. But this is a bit like saying that the Iliad and Citizen Kane are alike in that they both explore the relationship between the individual and society—true, perhaps, but only in the most a-historical way. When feminists criticize heterosexual scenarios, or, to use Monique Wittig’s phrase, the “heterosexual contract,” it is rarely heterosexuality as a simple “attraction to the opposite sex” that is under scrutiny, but rather the absolute equation between one kind of heterosexuality, drawn as the norm against which all differences are measured as “perversions,” and cultural meaning. Or as Wittig puts it, the heterosexual contract “produces the difference between the sexes as a political and philosophical dogma.”⁶³

An impressive body of feminist writing has been devoted to the exploration of how—following Luce Irigaray—heterosexuality functions as a ruse, a decoy relation to mask male homosocial and homosexual bonds. “Reigning everywhere, although prohibited in practice,” Irigaray writes, “hom(m)o-sexuality is played out through the bodies of women, matter, or sign, and heterosexuality has been up to now just an alibi for the smooth workings of man’s relations with himself.”⁶⁴ Comparatively little attention has been paid to how heterosexual economies work to assure that any exchange between women remains firmly ensconced within that “hom(m)osexual” economy. To be sure, male and female homosexualities occupy quite different positions, and given the logic of the masculine “same” that dominates the patriarchal order, female homosexuality cannot be ascribed functions that are similar to male homosexuality. However, the two homosexualities share the potential to disrupt, in however different ways, the reign of the “hom(m)osexual.” Irigaray speaks of the “fault, the infraction, the misconduct, and the challenge that female homosexuality entails.” For lesbianism threatens to upset the alignment between masculinity and activity, and femininity and passivity. Hence, writes Irigaray, “the problem can be minimized if female homosexuality is regarded merely as an imitation of male behavior.”⁶⁵

Irigaray’s discussion of the disruptive potential of female homosexuality emerges from her symptomatic reading and rewriting of Freud. In the Freudian text that occasions Irigaray’s remarks on the “problem” of female homosexuality within psychoanalysis, “The Psychogenesis of a Case of Homosexuality in a Woman” (1920), questions of narration and identification, masculinity and femininity, and dominant and alternative practice are posed in ways that are particularly relevant to lesbian authorship in the cinema. Jacqueline Rose has said of the case history that here, Freud “is in a way at his most radical, rejecting the concept of cure, insisting that the most psychoanalysis can do is restore the original bisexual disposition of the patient, defining homosexuality as nonneurotic.”⁶⁶

In the case history, Freud describes the brief analysis of a young woman who was brought to him by her parents after her unsuccessful suicide attempt. This “beautiful and clever girl of eighteen” pursued with great enthusiasm her attraction to a woman ten years older than she, and her parents (and her father in particular) were particularly distressed by her simultaneous brazenness (“she did not scruple to appear in the most frequented streets in the company of her questionable friend”) and deception (“she disdained no means of deception, no excuses and no lies that would make meetings with her possible and cover them”).⁶⁷ The suicide attempt occurred when these two factors that so distressed her parents coincided in full view of her father. After the young woman and her female companion were greeted by the woman’s father with extreme displeasure as his path crossed theirs on the street one day (as Freud notes, the scene had all the elements of a mise-en-scène planned by the young woman), the young woman threw herself in desperation over a railway fence.

Despite the apparent gravity of the suicide attempt, Freud saw little hope for successful analysis, for the woman—while not necessarily hostile to analysis, as was Dora, to whom this patient has frequently been compared—was nonetheless brought to analysis of a will other than her own.⁶⁸ In addition, Freud saw little actual illness in the young woman, at least as far as her sexuality was concerned; rather than resolving a neurotic conflict, Freud was being asked to assist in “converting one variety of the genital organization of sexuality into the other” (p. 137). As Freud proceeds to untangle the various threads of the young woman’s lesbian attachment, a somewhat confusing and often contradictory portrait of homosexuality emerges.⁶⁹

The woman’s sexuality is read through a variety of oppositions which form the territory of psychoanalysis—body and mind (“in both sexes the degree of physical hermaphroditism is to a great extent independent of the psychical hermaphroditism” [p. 140]); masculine and feminine desire (“She had thus not only chosen a feminine love-object, but had also developed a masculine attitude towards this object” [p. 141]); maternal and paternal identification (written before Freud hypothesized more extensively about the importance of the preoedipal phase for women, the case history nonetheless acknowledges the maternal object as, if not on the same level of importance as the oedipal scenario, then at the very least constitutive of her sexual identity). The case history is written within the field of these opposing terms, but there are shades of a breakdown of opposition, and the subsequent interdependence of the opposing terms. Hence, Freud speculates that the woman to whom the analysand was so intensely attracted evoked two love objects, her mother and her brother.

Her latest choice corresponded, therefore, not only with her feminine but also with her masculine ideal; it combined gratification of the homosexual tendency with that of the heterosexual one. It is well known that analysis of male homosexuals has in numerous cases revealed the same combination, which should warn us not to form too simple a conception of the nature and genesis of inversion, and to keep in mind the extensive influence of the bisexuality of mankind. (p. 143)

Indeed, this case history occasions some of Freud’s most famous pronouncements on the importance of bisexuality. Speculating that rage toward her father caused the young woman to turn away from men altogether, Freud notes that “in all of us, throughout life, the libido normally oscillates between male and female objects; the bachelor gives up his men friends when he marries and returns to club-life when married life has lost its savour” (p. 144).

But the “bisexuality of mankind” posited in the case history takes two distinctly different forms. On the one hand, it is posited as an originary force, a kind of biological given from which a variety of factors—Freud sometimes privileges predisposition, and sometimes environment—will determine one’s choice of sexual aim and sexual object. On the other, bisexuality emerges in a much more challenging and disturbing way as the violent play of warring forces, as evidenced most particularly in the young woman’s suicide attempt. For the desperate jump over the railroad wall is no quivering oscillation, and it is far from the kind of serial bisexuality alluded to in the above quotation about bachelors, marriage, and club-life. Rather, in the suicide attempt the battle of maternal and paternal objects attains crisis proportions and provokes a parallel crisis in representation. There are two divergent conceptions of bisexuality in the case history—one that assures that the young woman is either really like a man (in her choice of role) or really like a heterosexual (in her choice of love object), and the other which suggests, rather, a deeper tension between the desire to be seen by the father and the desire to construct an alternative scenario of desire altogether.

Despite its reputation as a more successful exploration of questions posed in the case history of Dora, “The Psychogenesis of a Case of Homosexuality in a Woman” does not read as a particularly convincing narrative in its own right. The “problem” of the case history centers on the woman’s self-representation, on her desire, not simply for the loved object but for a certain staging of that desire. In the event of the attempted suicide, it is not clear to what extent it was an unconscious attempt to put an end to parental—and particularly paternal—disapproval by literal self-annihilation, or rather an equally unconscious attempt to dramatize her conflicting allegiances by creating a scene where she is at once active subject and passive object (Freud notes frequently that the young woman’s amorous feelings took a “masculine” form). The suicide attempt is best described as both of these simultaneously—one a desire for resolution, the other a desire for another language altogether whereby to represent her conflicted desires.⁷⁰ Put another way, the suicide attempt crystallizes the position of “homosexuality in a woman” as a problem of representation and of narrative.

Freud discusses the young woman’s case in ways that suggest quite strongly the pressure of lesbianism against a system of explanation and representation. Throughout the case history, the young woman’s “masculinity” is the inevitable frame of reference. Masculinity acquires a variety of definitions in the course of the essay, at times associated with the biological characteristics of men (the young woman favored her father in appearance), and at others equated with the mere fact of agency or activity (she displayed a preference for being “lover rather than beloved” [p. 141]). But “masculinity” never really “takes” as an explanation, since throughout the case history the woman remains an embodiment of conflicting desires. The suicide attempt turns upon what has become, in the cinema, a classic account of the activation of desire, the folding of spectacle into narrative. However, in the standard account, woman leans more toward the spectacle with man defined as the active agent. Here, the woman’s desires to narrate and to be seen collide, leading her to make quite a spectacle of herself, but without a narrative of her own to contextualize that spectacle. As the young woman recounted the scene, the disapproving gaze from her father led her to tell her female companion of his disapproval, and her companion then adopted the opinion of the father and said that they should not see each other again. The sudden collapse, the identity between lover and father, the erasure of tension, seem to precipitate the woman’s quite literal fall. The woman’s desire for self-annihilation occurs, in other words, when her desire becomes fully representable within conventional terms.

What I am suggesting, then, is that the conditions of the representability of the lesbian scenario in this case history are simultaneously those of a tension, a conflict (which is “readable” in other than homosexual terms), and those of a pressure exerted against the overwhelmingly heterosexual assumptions of the language of psychoanalysis, a desire for another representation of desire. Or as Wittig puts it, “Homosexuality is the desire for one’s own sex. But it is also the desire for something else that is not connoted. This desire is resistance to the norm.”⁷¹ Expanding on Irigaray, de Lauretis writes: “Lesbian representation, or rather, its condition of possibility, depends on separating out the two contrary undertows that constitute the paradox of sexual (in)difference, on isolating but maintaining the two senses of homosexuality and hommo-sexuality.”⁷² The irony in Arzner’s signature suggests the division to which de Lauretis refers, between a representation of female communities and an inscription of marginality. That irony stands in (ironic) contrast to feminist film theory’s division of Arzner into a textual hommo-sexual (in print) and a visible homosexual (in pictures).

Given Arzner’s career in Hollywood, and the realist plots central to her films, her influence would seem to be most apparent among those filmmakers who have appropriated the forms of Hollywood cinema to feminist or even lesbian ends—Susan Seidelman (Desperately Seeking Susan), Donna Deitch (Desert Hearts). A more notable connection, however, exists with those contemporary women filmmakers whose films extend the possibilities of lesbian irony while revising the components of the classical cinema and inventing new cinematic forms simultaneously. I turn now to two films which are remarkable explorations of the desire, so succinctly expressed in the case history, to see and to be seen, to detach and to fuse, to narrate one’s own desire and to exceed or otherwise complicate the very terms of that narration.