Center Stage (Stanley Kwan, 1992)

There was once a painter who one day painted a landscape. [.] The artist was so delighted with his picture that he felt an irresistible urge to walk along the path winding away towards the distant mountains. He entered the picture and followed the path towards the mountains and was never seen again by any man.

– Béla Balázs, Theory of Film

The celebrated Hong Kong action star Maggie Cheung was originally scheduled to play the part of the late-Qing courtesan Scarlet in Hou Hsiao-hsien’s 1998 film Flowers of Shanghai. To prepare for this role, which was unlike any other she had ever played, she reputedly expressed an interest in procuring some opium in order to enter more completely the decadent fin-de-siècle ethos that characterized the late Qing (1). This anecdote foregrounds cinema’s frequent reliance on a process of projective identification, of participants provisionally stepping out of their conventional positions and “entering the picture” of distinctly alien ones.

This process of projection and displacement is not only applicable to the process of cinematic production but is also frequently stressed in discussions of the ways in which cinema is perceived. For instance, scattered throughout a number of modern Western writings on film and visual reproduction, one finds occasional references to the legend (dating back to the Tang dynasty) of the Chinese painter who walked right into the landscape painting he had just completed, thus finding himself absorbed back into the mimetic world of his own creation (2). Béla Balázs (3) and Siegfried Kracauer (4) both cite this legend as an allegory illustrating the ability of cinema to engross and absorb the spectator. Conversely, Walter Benjamin, in his celebrated essay on “The Work of Art in the Age of Mechanical Reproduction,” cites the same account as an illustration of precisely what the cinema is not (5). I begin here, then, from the rather odd fact that these particular modern European theorists should have found it useful to explain the workings of cinematic mimeticism by contrasting it with – or, even more interestingly, comparing it to – a patently fantastic legend which, historically and culturally speaking, is radically removed from their own respective traditions.

This double movement through time and space is mirrored in Hou Hsiao-hsien’s Flowers of Shanghai, which is not only set in a Mainland Chinese city (very unusual for this Taiwan director), but is also an adaptation of a novel by the same title originally written more than a century earlier. Although Maggie Cheung eventually backed out of the project to work on another film instead, her aborted Flowers of Shanghai encounter can nevertheless serve as a prism through which to reexamine a cluster of interrelated themes of specular failure and specular return in two of Cheung’s earlier 1990’s films: Stanley Kwan’s 1992 cinematic retrospective Ruan Lingyu, or Center Stage [also translated as “Actress“], and Olivier Assayas’ 1996 French production, Irma Vep.

Aside from their shared actress/protagonist, both Center Stage and Irma Vep are based on a remarkably similar conceptual premise. In both cases, Maggie Cheung appears in the contemporary “present,” attempting to recreate a series of films from the early decades of the century. In Stanley Kwan’s film, Cheung is charged with reenacting the life and films of the movie star Ruan Lingyu, one of the most influential actresses from Shanghai during the 1920’s and 1930’s. In Assayas’ metatextual comedy four years later, Cheung plays herself, having been called to France by the aging and eccentric director “Henri Vidal” to film a remake of the Louis Feuillade’s 1915 silent classic Les Vampires. In both of these films, the figure of Maggie Cheung provides a figurative locus for the anxieties revolving around the inevitable loss, and potential recovery, of a certain auratic presence. Accordingly, a recurrent concern in both works is whether or not Cheung will be able to preserve successfully the essential spirit of the historical films and film stars which she has been charged with recreating. At the same time, even as she is trying to live up to the legendary status of these early twentieth century film stars, her own nascent star status repeatedly comes back to haunt her.

“Aura,” in Walter Benjamin’s sense, refers to a perceived uniqueness and inherent presence that derives from its imbeddedness within the “fabric” of a specific tradition. He argues, however, that the advent of mechanical reproduction constitutes a profound challenge to this auratic dimension of cultural production, insofar as it makes possible the essentially endless reproduction of art (eg. cinema, photography) – reproductions which are precise simulacra of the “original,” and which are therefore deprived of the auratic presence from which the art work ostensibly derived its significance in the first place. One of the legacies of this development of mechanical reproduction, then, is a sense of specular failure, whereby a superficial visual fidelity is achieved at the expense of an underlying aura of presence. In the following discussion, I will explore how this moment of “specular failure” may be accompanied by a certain recuperation of the “aura,” whereby a certain sense of auratic presence is created from within a techno-cultural regime of mechanical reproduction itself.

Center Stage and Irma Vep

Center Stage

Stanley Kwan’s Center Stage opens with a series of early twentieth century still images relating to the life and works of the Shanghai film star Ruan Lingyu, accompanied by a voice-over providing a capsule summary of her life and career. This is followed by another brief segment in which Maggie Cheung, appearing as “herself,” discusses Ruan’s legacy. Cheung opens her comments by laughing as she asks rhetorically, “Isn’t she [Ruan Lingyu] an exact replica of myself?”

The following sequence, however, proceeds to problematize this initial gesture of identification. As we are presented with more black-and-white photographs and drawings of the historical Ruan Lingyu, the narrative voice-over resumes by observing that Ruan took her own life at the age of 25, but that people nevertheless “still remember her as a movie star.” The narrator then asks Maggie how she would like to be remembered half a century later, to which she replies modestly, “That’s not so important to me. If future people do remember me, it won’t be the same as Ruan Lingyu. As she halted her career at the age of twenty-five when she was at her most glorious; now she is a legend.”

We have here, then, in this brief prefatory sequence, a suggestive moment of specular failure. The preface, that is to say, initially holds up the apparent mirrored relationship between Ruan’s and Cheung’s respective careers (“isn’t she an exact replica of myself?”) but then proceeds to note a necessary inscription of difference between the two. This moment of “difference” lies in Ruan’s “star quality,” which ostensibly colors and problematizes any attempt to establish a strict specular parallel between the women. This “star” quality, moreover, can be productively compared to Benjamin’s concept of the unreproducible “aura,” with the ironic qualification that this “aura” is not only not anterior to technologies of “mechanical reproduction,” but is even itself born of those very same technologies themselves.

The body of Kwan’s film itself depicts Cheung reenacting the figure of the historical Ruan Lingyu. In particular, it reconstructs the behind-the-scenes production of many of Ruan’s major films and, where possible, it juxtaposes these reconstructed scenes with corresponding clips from the original films. The tenuousness and ultimate contingency of this process of cinematic mirroring is further reinforced by the fact that many of the other “reconstructed” scenes correspond to films from which we no longer have any extant footage (as is explicitly noted in Kwan’s film), thus leaving us with no original “model” with which to compare the fictional recreations presented in Center Stage.

Given that Kwan’s film itself is concerned with attempting to come to terms with the unique “auratic quality” of the historical Ruan Lingyu, its success would thought to be predicated on its ability to sublate Cheung’s own charismatic presence in favor of Ruan’s own. At the same time, the film’s energy is arguably the result of the unresolved tension between these two spectral presences. For instance, Cheung’s modest demurral at the opening of the film resonates ironically throughout the film, as her own virtuoso performance partially gives lie to her suggestion that she would not be able to live up to Ruan’s legendary precedent. Furthermore, one of the latter scenes in the movie depicts Cheung-as-Ruan meeting with a small group of directors and producers as they reflect on the impact which the introduction of sound would have on the Chinese film industry, not to mention on Ruan’s own individual career. This question is significant because Ruan built her career in the era of silent films, and she herself does not even speak very standard Mandarin Chinese. What we have here, then, is a gesture to a moment of fractured mirroring, in which Ruan would attempt to capitalize on her own earlier fame even as she must effectively reinvent herself in the face of new technological development.

Irma Vep

In Assayas’ Irma Vep, this theme of specular failure is developed even more explicitly. The film opens with Cheung’s arrival in Paris, where she is scheduled to work with the eccentric director Henri Vidal (played by Jean-Pierre Léaud) in his somewhat quixotic attempt to recreate the early twentieth-century French classic Les Vampires.

In implicit parallel to the short sequence at the beginning of Center Stage in which Cheung appears as “herself,” there is a brief sequence near the beginning of Irma Vep in which the immediacy of the film’s basic premise of having Cheung appearing as “herself” is redoubled by a short series of clips from one of her own earlier action flicks – these ostensibly being examples of the sort of performance which interested Vidal in Cheung in the first place.

Unlike Center Stage, however, Irma Vep is explicitly predicated on an attempt not to sublate Cheung’s own contemporary star status, but rather to capitalize on the productive tension created between Cheung’s distinct charisma and the rather distinct cultural and aesthetic ethos of the masked, silent role she must recreate. Vidal had personally hand-picked Cheung because of his fondness for her Hong Kong action-flicks, but his confidence in the entire project quickly wanes, leaving Cheung struggling in vain to renew his interest. In the interim, even as the production of the Les Vampire remake becomes increasingly confused and chaotic, Cheung ironically finds herself delving deeper and deeper into the Irma Vep character. Aided by a couple of inadvertent encounters with narcotics (an ironic anticipation, perhaps, of her flirtation with opium in Flowers of Shanghai), she finds herself identifying more and more with that masked cat-burglar, such that at one point we find her prowling along catwalks and roof-tops after having stolen an expensive necklace from another room in the luxury hotel she is staying in.

In the end, Vidal abruptly abandons the film altogether, and Cheung herself is unceremoniously dumped from the project when Vidal’s replacement as director deems it ridiculous that they be using an Asian actress to recreate this French cinematic classic. Interestingly enough, this is actually the first non-Asian film which the real-life Cheung had appeared in, and in a subsequent interview she explained that one of the things that attracted her to the project in the first place was precisely that it did not attempt to type-cast her as merely an “Asian woman stereotype.” (6) Underlying Cheung’s scripted failure to mimetically recreate the figure of Irma Vep, therefore, is an explicit concern on the part of Cheung, the actress, that she be allowed to preserve a certain artistic presence and individuality, and not be reduced to a simulacrum of herself and of the characters which she typically performs.

Like Center Stage, Irma Vep develops the themes of specular failure and spectral return through a careful attention to the material plasticity of the cinematic medium itself. Both films intersperse the main narrative with both clips and stills from the early twentieth century silent films they are paying homage to, as well as with their own simulations and recreations of those same early films.

Irma Vep

Furthermore, in Irma Vep, the materiality of the cinematic medium is brought to the fore in a most direct way, in that the movie as a whole ends with an almost painfully long sequence in which we see much of the footage that was filmed over the course of the movie (that is to say, the footage of Vidal’s and Cheung’s contemporary recreation of Les Vampires) scandalously vandalized with extraneous sounds and crude line drawings (this vandalism, it is strongly implied, being the work of the troubled and wayward director Vidal himself).

This somewhat shocking conclusion to the film brings us back to the celebrated, and oft-reproduced, Benjamin article on the mechanical reproducibility of art with which I opened the present essay, where he links the rise of various technologies of mechanical reproduction (e.g., photography, cinema, etc.) to the gradual loss of a sense of “auratic” presence which is retrospectively projected back onto an earlier aesthetic tradition. Benjamin explains that the mechanical reproduction has the effect of detaching the art-work from its previous embeddedness in a specific time and space, thus “detach[ing] the reproduced object from the domain of tradition.” (7) Although many discussions of Benjamin note his attention to this evacuation of tradition, Benjamin also goes on to give equal weight to the second step in the process, whereby “in permitting the reproduction to meet the beholder or listener in his own particular situation, [mechanical reproduction] reactivates the object reproduced.” (8) For Benjamin, it is precisely this ability of the post-auratic simulacrum to forge new affective bonds with the viewer/observer which helps to account for the aesthetic and political potential inherent in reproductive technologies such as cinema.


In his short story “Pierre Menard, author of the Quixote,” Borges writes of a contemporary author who was so fascinated by Cervantes’ classic novel that he decided to devote himself to recreating it word for word (9). After immersing himself in an exhaustive story of the Cervantes’ life and writings, Menard finally sat down to attempt to rewrite the timeless masterpiece. Oddly enough, he found that he had indeed succeeded in “rewriting,” verbatim, the exact text of the beginning of Cervantes’ original novel, but with the catch that each word and phrase had a rather distinct significance for him than it did Cervantes himself. That is to say, at one level his project was a success in that he had succeeded in creating an exact simulacrum of portions of the original novel, but at the same time was a failure insofar as this surface continuity between the two pieces occluded a more significant divergence at the level of the actual connotations and contextual resonances of the resulting text, that is to say, at the level of its original auratic presence.

In a way, however, this ultimate specular failure was already, in a sense, explicitly anticipated by the way in which Menard conceptualized his Quixotic project to begin with: “Initially, Menard’s method was relatively simple: [.] – be Miguel de Cervantes. [.] Being, somehow, Cervantes, and arriving thereby at the Quixote – that looked to Menard less challenging (and therefore less interesting) than continuing to be Pierre Menard and coming to the Quixote through the experiences of Pierre Menard.”(10) The goal of Menard’s project, therefore, was never simply to produce an exact reproduction of the Cervantes’ novel, but rather to re-produce it in a way that productively drew on a diaologic interaction between the respective contextual specificities of Menard and Cervantes himself. The result, therefore, is not merely a literary simulacrum which puts into the question the unique “aura” of the original work, but also, and more interestingly, a “palimpsest” of the auratic specificities associated with both Cervantes’s and Menard’s respective productions.

I would claim that both Kwan’s and Assayas’ films similarly find themselves located precariously at the limits of the auratic. They both present a mediated reflection on the star culture of the early twentieth century film industry (both Chinese and French), while at the same time having rather ambiguous implications for the reproducibility of that same “star aura.” This ambivalence of the auratic can be seen most clearly during the moments in both films in which brief clips from the original productions are explicitly juxtaposed with Cheung’s own precise recreations of the same scenes. These moments both pay homage to the original films and foregound Cheung’s own talent and unique charisma. Like Pierre Menard, Maggie Cheung seeks to immerse herself directly into the historical context of the films she is charged with recreating, but in the end she succeeds only in producing a surface recreation of the target films. In their respective juxtapositions of selected scenes from the original films and Maggie Cheung’s own precise recreations of the same scenes, both Kwan’s and Assayas’ films suggest that even this sort of direct, mimetic simulacrum will necessarily fail to recapture the inherent “aura” of the original. At the same time, however, both works point to the way in which new forms of auratic presence can be carved out within the system of correspondences established by technologies of mechanical reproduction themselves.

I will conclude, then, by returning briefly to the Hou Hsiao-hsien film with which I began. Unlike Kwan’s and Assayas’ respective films, Hou’s Flowers of Shanghai seeks not to recover a turn of the century cinematic legacy, but rather to bring to life an influential turn-of-the-century literary landmark. At the same time, however, it can also be seen as a moment in which contemporary film technology effectively peers back into its own technological and cultural origins. That is to say, Hou’s 1998 film was not only released shortly after the centennial of the first publication (1894) of the Han Bangqing novel on which it is explicitly based, (11) but also marks the almost precise anniversary of the introduction of cinema into China (in 1896, by the suggestively surnamed Lumiére Brothers).

The ambivalences of this moment of doubled retrospective projection (of looking back at the implicitly forward-looking implications of a crucial historical watershed) are foregrounded by a moment of optical anxiety at the very beginning of the film. As several of the characters sit around a courtesan-quarters banquet table playing drinking games and telling stories, one character recounts the story of two lovers – a courtesan and her patron – who are so infatuated with each other that they do nothing else than merely sit around all day staring into each others’ eyes. At one point it is recommended that they try to get out more, but they have objections to each of the proposed activities (e.g., riding in a carriage, etc.). Their objection to having their photograph taken is particularly revealing, as they feel that it will have the effect of stealing away their “eye-beams.” After this suggestion, the story continues, the same character repeatedly observed the couple in bed in the morning, with the courtesan, Crystal, diligently licking Tao Yufu’s closed eyes – implicitly reacting to photography’s perceived threat to visual presence by repairing to a somatic fetishization of the optical faculty itself.

Complementing the above discussion of the way in which two films with Maggie Cheung struggle to come to terms with the relationship between cinematic technologies and the potential loss of a perceived auratic presence, I then conclude with the preceding scene from a film in which Cheung is herself deliberately absent, and which looks back to a cluster of ambivalences about the nature of photographic reproduction and spectral recuperation on the very eve of cinema’s invention and introduction into China.


  1. Hou Hsiao-hsien [Hou Xiaoxian] and Zhu Tianwen, Jishang zhi meng: Haishang dianying quanji lu (Taipei: Yuanliu, 1998), p.125
  2. For comparison, the text of the original Chinese version of this same anecdote reads as follows:Just as his contemporary the poet Li Bo [Li Bai] had drowned in a river trying to catch hold of the reflection of the moon, about whose beauty he had so often sung, legend recounts that Wu Daozi disappeared into the mist of a landscape he had just painted (cited and translated by François Cheng, Empty and Full: The Language of Chinese Painting, [Boston: Shambhala, 1994], pp. 28-29).
  3. Béla Balázs, Theory of the Film: Character and Growth of a New Art (New York: Dover, 1970), p. 48
  4. Siegfried Kracauer, Theory of Film: The Redemption of Physical Reality (New York: Oxford University Press, 1960), pp. 164-165
  5. Walter Benjamin, “The Work of Art in the Age of Mechanical Reproduction,” in Hannah Arendt, ed., Illuminations, pp. 238-9. Here, Benjamin is drawing a distinction between art forms which invite “contemplation” (like the Chinese painting), and those which invite “distraction” (like film).
  6. Interview with Shelly Kraicer, 1996; seen “Irma Vep: Review by Shelly Kraicer,” http://www.chinesecinemas.org/irmavep.html
  7. Walter Benjamin, p. 221
  8. Walter Benjamin, p. 221
  9. Jorge Luis Borges, “Pierre Menard, Author of the Quixote,” in Jorge Luis Borges: Collected Fictions, Andrew Hurley, trans., (New York: Penguin Books, 1998), pp. 88-95.
  10. Borges, p. 91
  11. The original novel was first serialized in 1892, but was not published as a single, unified work until 1894.

About The Author

Carlos Rojas is a postdoctoral fellow and adjunct assistant professor at Columbia University. His dissertation is entitled "Flowers in the Mirror: Visuality, Gender, and Reflections on Chinese Modernity."

Related Posts