When Conventions Travel

By the 1930s, the classical Hollywood mode of narration had been more or less adopted by Chinese filmmakers in terms of the basic principles of continuity, such as scene construction using varying shot scales determined by narrative motivation and punctuation cuts (wipes, fades, dissolves) between scenes. Many examples of borrowing from specific Hollywood films also can be found. The earliest still extant Chinese film, Laborer’s Love (Laogong zhi aiqing 劳工之爱情; Zhang Shichuan 张石川, 1922), for example, drew some of its comedic gags from Hollywood films of just a year earlier, including Buster Keaton’s The Haunted House and Harold Lloyd’s Never Weaken.8

One type of intertextual citation of Hollywood found in Chinese films is the incidental—or even accidental—mimicking of an isolated image or motif. Take, for instance, an iconic image from Wu Yonggang’s silent classic The Goddess, from 1934 (Figure 5a): Ruan Lingyu plays a victimized prostitute who tries to raise and educate her young son despite being preyed upon by the gambler/pimp whose legs we see in the foreground. Was this striking visual composition inspired by King Vidor’s 1932 film Bird of Paradise (Figure 5b)? It certainly is possible. Bird of Paradise had run at the opulent Nanjing Theater in Shanghai beginning on January 1, 1933, and its flamboyant style and racy, pre-Code content was sure to have drawn the attention of many Shanghai film artists. On the other hand, the same composition had been used at least once earlier, in the 1931 Hollywood film Gentleman’s Fate (Mervyn LeRoy).9 In fact, the composition is such a compelling one—for both its symmetry and its ability to convey the dynamics of power, including an implicit sadomasochistic eroticism, of its narrative content—that one can imagine various filmmakers, whether Western or Asian, either independently deploying it or readily copying it.

If such parallels are relatively inconsequential, given that the rampant transnational pilfering of cinematic ideas has been the rule since the birth of the medium, a more complicated case arises when we go beyond the borrowing of isolated images or technical devices and look instead at the colonially inflected spread of more extensive genre and plot elements from Hollywood to China. One example would be the status of The Goddess itself as a maternal melodrama that drew upon not only traditional Confucian values regarding family sacrifice and education but also Hollywood precedents like Stella Dallas (Henry King, 1925).10 Indeed, the broader availability of melodrama as both a quotable Hollywood film genre and a transcultural and transhistorical narrative mode calls attention to the ways in which Western cinematic and narrative conventions were creatively combined with indigenous storytelling traditions.11

A fascinating case of such combinations (though far less well-studied than The Goddess) is the 1931 Chinese film A Spray of Plum Blossoms (Yi jian mei 一剪梅; Bu Wancang 卜万苍), which features a tangle of cross-cultural signifiers, all within a loose adaptation of William Shakespeare’s Two Gentlemen of Verona. In its details, the film draws simultaneously on Chinese cultural reference points and Western cinematic precedents. The “Silvia” character, for example, has an affinity for the titular plum blossoms (a symbol of endurance in China), leading to flower-themed set designs of her personal quarters in her warlord father’s mansion that recall the camellia icons of the 1921 Hollywood film Camille (Ray C. Smallwood), an adaptation of The Lady of the Camellias (La Dame aux Camélias); original novel by Alexandre Dumas fils) starring Alla Nazimova and Rudolph Valentino.12

By the latter third of A Spray of Plum Blossoms, even though the setting in modern times has been made clear by, for example, the arrival of the “Proteus” character in a biplane, the “Valentine” character’s adopted outlaw band after his unjust banishment appears to be drawn partly from the Chinese cultural imagination of the jianghu 江湖—literally “rivers and lakes,” the half-real, half-mythical traditional setting of martial arts fiction (somewhat akin to the Wild West in the American imagination). More confusingly, however, after he becomes an outlaw leader, Valentine’s costume in A Spray of Plum Blossoms, featuring a jerkin-style vest, immediately recalls not Chinese martial arts heroes so much as the Hollywood version of Robin Hood exemplified by Douglas Fairbanks, who was as wildly popular in China as he was in the West (Figure 6).

Here we see not just an isolated citation, nor only the adoption of the basic principles of classical continuity, but a more thorough interpenetration of themes, motifs, and imagery, suggesting a substantially new cultural imaginary in which Hollywood has been thoroughly absorbed into the Chinese cinematic and narrative consciousness. Much of the critical discussion of 1930s Shanghai cinema, both at the time and in more recent scholarship, has in one way or another touched on the question of the relationship, whether actual or desired, between the ubiquity of Hollywood and the Left-Wing Film Movement itself, in which progressive and Communist-sympathizing directors, screenwriters, and other film artists made a number of films which—despite censorship by the Nationalist government—turned a critical mirror on contemporary Chinese society and at least implicitly promoted revolutionary change as well as defense against the growing economic and territorial encroachments of imperialism. In Chinese Communist film historiography, long represented by the two-volume History of the Development of Chinese Film, edited by Cheng Jihua 程季华, these films are appreciated for their left-leaning sentiments and their social realism, which is seen as a precursor to the more fully developed Chinese socialist realism that would follow the establishment of the People’s Republic in 1949.13 However, most scholars more recently have rejected this official teleological narrative. As early as 1989, for example, Chris Berry placed Cheng Jihua’s account in the context in which it was originally published, in 1963, and suggested that the connection of these filmmakers to the Communist Party was not as seamless as Cheng’s account had made it seem.14 Later, Paul Pickowicz would argue that because many of these films are melodramas, they are in fact not revolutionary at all but rather reactionary—at least insofar as they actually adhered to Hollywood genre formulas.15 Still later, in Zhang Zhen’s indispensable history of Chinese cinema up to 1937, the relationship between Chinese films of the 1930s and classical Hollywood is treated with an unprecedented degree of detail and nuance, though, on balance, Zhang, too, deemphasizes these films’ positions in a Communist genealogy, examining them instead as reflections of and influences on the daily lives of Chinese people living under the new conditions of urban modernity.16 Even more recently, Zhiwei Xiao has questioned the actual coherence of any singular leftist film “movement” and pointed out that historical narratives of it often have been self-serving and intended to celebrate specific figures and downplay others according to later political developments.17

Rather than being mutually exclusive, these films’ competing priorities of Hollywood-style entertainment and leftist critique are better viewed as yielding what Siegfried Kracauer has called the “strange constructs” that result not only from cinema’s capacity for realist recording of reality but from its ability to combine disparate things in new ways.18 Zhang Zhen cites these “strange constructs” as helping to constitute “the revolutionary potential of the cinema.”19 Clearly any account of 1930s Shanghai leftist cinema that depicts it solely as politicized critical realism will fail to account for a great deal of what made this cinema so engaging both to its original audiences and to students of Chinese cinema today. At the same time, reducing these films simply to Chinese variations on Hollywood entertainment fails to capture the unique combination of classical entertainment film formulas with genuine revolutionary impulses that they embody. Rather than seeing certain scenes in an ostensibly leftist film as representing the filmmakers’ more politically correct (“hard”) critical realist impulses and others as caving to Hollywood conventions in a concession to the audience’s desire for (“soft”) entertainment and the studio’s desire for profits, here I will explore how the very invocation of Hollywood conventions was sometimes done in a subversively critical manner that generated irrepressible tensions within the classical form.

Mimicry as Critique: Crossroads

Despite being considered a classic of the Left-Wing Film Movement and indulging in moments of bleak critical realism—such as an attempted and then actual suicide by drowning of one of the four college graduates introduced at the beginning of the film—Crossroads is primarily, for most of its running time, a romantic comedy. One would not know that from reading the History of the Development of Chinese Film, which includes a nearly four-page discussion of Crossroads that never once describes it as a comedy of any sort (and romance is mentioned only in a brief aside stating that two of the characters fall in love in the course of the film).20 Nonetheless, Crossroads not only is clearly identifiable as a romantic comedy but also follows a particular plot pattern within that genre that I will call the romantic comedy of misrecognition—where much of the comedy and suspense (the delaying of full romantic coupling typical of rom-coms) is derived from a plot twist in which two potential lovers are unaware that they already share some life connection without recognizing the other’s identity.

A paradigmatic Hollywood rom-com of misrecognition is the 1940 Ernst Lubitsch film The Shop around the Corner, in which a shop clerk (James Stewart) feuds with a coworker (Margaret Sullavan) while falling in long-distance love with a pen pal, without realizing that they are one and the same person. In a late-century variation, You’ve Got Mail (Nora Ephron, 1998), Tom Hanks and Meg Ryan enact much the same scenario of open professional rivalry and dislike combined with a more private anonymous epistolary romance (in this case, through email). Such a repeated narrative device qualifies as what Carl Plantinga labels a “paradigm scenario”—or conventionalized and learned “types and sequences of events that are associated with certain emotions.”21 In the case of the rom-com of misrecognition, the plot activates the audience’s strong desire for the misrecognition to be overcome and the couple to be permanently united.

Like so many Hollywood paradigm scenarios, the rom-com of misrecognition goes back to the silent era. In Crazy Like a Fox (Leo McCarey, 1926), for example, a young woman runs away from home, outraged because her father has engaged her to a friend’s son whom she has never met. Meanwhile, the groom-to-be also hopes to get out of the marriage and hatches a plan to pretend to be mentally unstable when he meets the bride’s family. The resolve of the man and woman to elude the arranged marriage is amplified when they meet as strangers at the train station and fall in love at first sight, not realizing the other’s identity. Needless to say, all ends well when the misrecognition is resolved and the arranged marriage partner turns out to have conveniently coincided with the actual love interest.22

Precedents for the cinematic rom-coms of misrecognition can be found in the sorts of ironic twists of fate featured in the popular American literature of O. Henry (William Sydney Porter) in the late nineteenth and early twentieth centuries. One of his most reliable formulas, in stories such as “The Gift of the Magi” and “The Unfurnished Room,” was an improbable coincidence and consequent misunderstanding that served to demonstrate the depth of a young couple’s love. Paul Fejos confessed to adding a corny “O. Henry twist” to his masterpiece Lonesome, a 1928 rom-com of misrecognition that straddled the transition from silent to sound cinema.23 The plot is simple: a factory worker and a telephone operator meet at Coney Island and quickly fall in love. When they become unwittingly separated in the crowd of thousands, it appears that they will never find each other again, but upon returning home, it turns out that they live in adjacent rooms in the same tenement building. In the final scene, the man and woman have independently returned home, exhausted and brokenhearted. The film crosscuts between their separate rooms, occasionally combining them into one shot to show visually how the sound of the phonograph the man plays to console himself reaches the woman’s space. She becomes angry at the noise coming from her anonymous neighbor and goes to complain to him, only to find that he is in fact her lost lover, so that they are now joyfully reunited and the film ends.

Similarly, in the Chinese Crossroads, one among the group of four struggling male college graduates carries on a battle of the sexes with his unseen neighbor, who, unbeknownst to him, is the same young woman with whom he flirts when they cross paths every day in their respective streetcar commutes. Miriam Hansen speculates that Crossroads borrows from Lonesome its basic misrecognition scenario—two people feud as unseen next-door neighbors in a tenement house while falling in love as anonymous others who randomly encounter each other outside in the city.24 It may well be true that the makers of Crossroads remembered Lonesome from less than a decade earlier, but the director’s account of the plot’s inception does not make open acknowledgment of the influence of that or any other film. In an article published just before Crossroads was released, director Shen Xiling wrote that the inspiration for the film came partly from a friend of his, who suggested the following plotline, as quoted by Shen:

I want to write about a contemporary money-craving couple. They don’t know they are living together because of their different work schedules. At home they fight every day, yet when they run into each other, they pretend they are rich and deceive each other—at last the whole thing is discovered and it becomes a joke.25

It is unclear whether Shen himself knew this, but his friend’s suggested plot synopsis in fact exactly describes the 1932 German film I by Day and You by Night (Ich bei tag und du bei nacht; Ludwig Berger), a masterful romantic comedy of misrecognition that also was in part an extended meditation on how the movies structure our fantasy lives in ways that can be both pleasurable and problematic in the context of socioeconomic class divides—much as Crossroads would do in a key sequence to be discussed shortly.26

In I by Day and You by Night, in an arrangement typical of the economic hardship of the time, a man and a woman separately rent the same bed in a furnished apartment room. The man works all night as a waiter in a local nightclub, while the woman works by day as a manicurist. Both are young and poor, and each sleeps in the rental bed while the other is at work. They never cross paths in the apartment, but each often becomes angry at the absent other for his or her personal habits in the space they must share. Naturally, the two happen to meet outside of the domestic space and quickly fall in love, while, through a series of misunderstandings, both also come to believe that the other is rich and put on airs accordingly to deceive the other about his or her actual poverty. The climactic scene comes near the end of the film, when the man returns home while the woman is still in bed, leading to their comical discovery of each other’s true identities and a happy ending. A metacinematic element is introduced by the fact that the man’s friend is a movie theater projectionist, and the role of popular cinema in modeling happiness and structuring our fantasy lives is a running joke in the film.

The arrangement in Crossroads shares even more with the situation in I by Day and You by Night than Lonesome in that it returns frequently to the domestic proximity gag rather than just revealing the coincidence at the end. In Crossroads, however, the eventual revelation is a surprise only to the man, Lao Zhao (played by Zhao Dan 赵丹), because the woman, Yang Zhiying (Bai Yang 白杨), had figured it out relatively early in the film but kept him in the dark just to play a prank on him. Lao Zhao, cheerful because he is in love, continues his usual practice of singing loudly and throwing trash over the wall into the room next door to annoy his unseen nemesis, who he is convinced is actually his cranky landlady, unaware that it is in fact his new girlfriend. Adding to the comic misunderstanding is Lao Zhao’s pose, instigated by a friend who gave false information to Zhiying, that he lives in a wealthy neighborhood when in fact he is unemployed at the beginning of the film and, by the end, has been laid off again from his night job as a proofreader for a newspaper. Much of the film’s humor, but also its politics, derives from the precarious class status of both Lao Zhao and Zhiying as “petty urbanites” (xiao shimin 小市民)—a class of literate city dwellers in modestly paying salaried jobs (or, increasingly, unemployed)—in depression-era Shanghai. (Zhiying is a women’s vocational school graduate who serves as a trainer at a cotton mill.) The clash between the aspirations represented by Lao Zhao’s pose as a higher-class romantic prospect and the actual precarity of his and his unemployed college friends’ real lives constitutes a central theme of the film.

With a crucial dream sequence that occurs approximately midway through the film, Crossroads shrewdly relates those aspirations to the cultural imaginary offered by Hollywood and subtly critiques the inadequacy of Hollywood’s fantasy world for capturing the realities of life in China. In this sequence, the leisurely reading of a novel leads Zhiying to doze off and have an erotic dream featuring a romantic encounter between her and Lao Zhao. The novel that inspires her libidinal fantasy is the Chinese translation of La dame aux camélias. By this time, Lady of the Camellias had enjoyed decades of wide popularity in China and prominence in the modern Chinese cultural imagination, epitomizing the imported modern Western romantic ideology of heterosexual love as providing the ultimate meaning of life for the individual.27 It had inspired several Chinese rewrites and sequels as well as countless stage adaptations beginning in the late Qing dynasty and continuing well into the Republican period—including one that had starred Bai Yang, the actor playing Zhiying, during her stage career before she began to act in films.28

Literary and stage precedents aside, with the Crossroads dream sequence, we have not only the sleeping character Zhiying imitating Lady of the Camellias protagonist Marguerite Gautier in her dream but the director Shen Xiling, in his filming of the dream sequence, imitating Western directors like George Cukor, whose own Hollywood version of Camille had been released the previous year, or Fred Niblo, who had directed another Hollywood version in 1926, or even Ray Smallwood, who had directed the 1921 Camille mentioned earlier. These are just a few of the dozens of film adaptations of La dame aux camélias that had appeared by the time of Crossroads—including at least three Chinese film versions in addition to many from the United States and Europe.29 The dream sequence in Crossroads thus serves as a metacinematic commentary on the very influence of Western film that is visible elsewhere in Crossroads as well as in other films of its time. The effect of Western films on the petty urbanites of semicolonial Shanghai is parodied precisely by the casting, within the imagination of Zhiying’s erotic dream, of her as Marguerite and Lao Zhao as her lover Armand, so that her erotic desire for Lao Zhao essentially takes the form of a Western movie scene based on the French romance.

Though awkward and fumbling in real life, in Zhiying’s dream, Lao Zhao is as charming, smooth, and confident as his dashing, aristocratic European costume would suggest, whereas Zhiying appears as a beautiful and refined young woman of the world. After he sways her around on a Western-style swing (including a wonderful tracking shot in which the camera moves as if attached to the swing), the two retire to an interior space, where Lao Zhao boldly kisses Zhiying’s bare shoulder and then her lips in a close-up (Figure 7). Kisses on the lips are extremely rare in Chinese films of this time, and I am aware of no other instance in which the lips touch any part of the lover’s body below the neck. In fact, given the much more modest standards of physical intimacy found in Chinese films compared to Hollywood in the 1930s, I would go so far as to speculate that this shot could only happen within a dream sequence, and particularly one that clearly is imitating Hollywood—and perhaps especially one that parodically invokes Greta Garbo, who had starred in the most recent Hollywood version of Camille. The kisses in the Crossroads dream sequence appear to consciously mimic amorous affections paid to Garbo not only in Camille but also in other of her films, such as Mata Hari (George Fitzmaurice, 1931)—precisely the sort of “fleshly” (rougan 肉感) erotic displays that Chinese audiences and critics found both excessive and transgressively titillating in Hollywood movies but generally did not see in Chinese productions.30

Lao Zhao maintains his masculine confidence throughout the sequence, but the fantasy performance of Westernized romance is comically undermined by the scene that immediately follows the dream sequence, in which Lao Zhao at last has a perfect opportunity to go beyond the nervous smiles he has shared in previous coincidental streetcar encounters with Zhiying (enabled by the fact that she takes the same streetcar to work during the day that he takes home after his night shift). This time, instead of passing her quickly on the streetcar steps, he remains on the car with her, helps her retrieve something she has dropped, and then finds himself standing right in front of her during the ride, yet he cannot muster the confidence to say a word to her despite her obviously flirtatious smiles and glances. The scene thus humorously negates Zhiying’s previous dream fantasy of Lao Zhao as a highly sophisticated, confidently amorous Europeanized lover.

In the Lady of the Camellias dream sequence in Crossroads, the object of satire is not so much a Western “original,” because so many adaptations already had proliferated in both the West and China. Nor does the humor come simply from Zhiying’s desire for Lao Zhao, though that is certainly the denotative narrative content of the scene in terms of its plot motivation. Instead, by articulating Zhiying’s desire in this particular way, the scene appears to satirize the mimicking of Western looks and behaviors, learned in large part through movies, as the way to be modern, as well as the imperialist context from which such mimicry arises (see chapter 1). The real object of gentle satire is the imagined selves of Zhiying and Lao Zhao in the scene.31 These alternative selves are both real, in the sense of being new modes of existence made conceivable in part through the co-opted imagination of Western literature and film, and impossible, in that the material conditions of poverty and colonial exploitation ensure that the fantasies shaped by Hollywood’s dream factory will remain only fantasies, even though the desires expressed through or activated by them are real.

The blunt failure of the fantasized world in Zhiying’s dream to bear much resemblance to her real life is made clear even within the dream during its most comic moment. When Lao-Zhao-as-Armand makes his initial appearance with a triumphant smile, his visage is suddenly obscured by a superimposed image of a tag reading “unemployed no. B.” This label is one that Lao Zhao and his three college friends had taped to their graduation photos in a darkly satirical gag to make light of the fact that they all had suffered from unemployment after finishing their degrees. It was when Zhiying discovered Lao Zhao’s portrait and tore off the label covering his face that she first realized her despised neighbor was the handsome young man she had been seeing in the streetcar. The irruption of the “unemployed no. B” label into Zhiying’s dream, once again obscuring Lao Zhao’s face, thus serves as a comic reminder of who this debonair Armand character really is, in the process ironically setting off the opulent world of the fantasy from the much more desperate conditions of the couple’s real life and thus serving as a moment of social critique, however lighthearted, within the dream sequence.

The Crossroads dream sequence’s apparent critique of Hollywood-based fantasy as incompatible with contemporary Chinese reality has a precedent in the 1931 classic film Love and Duty (Lian’ai yu aiqing 恋爱与义务; Bu Wancang 卜万苍). Starring two of the biggest Chinese movie icons of the day, Ruan Lingyu and Jin Yan 金焰, the film as a whole is partly a romantic story of illicit love and partly a maternal melodrama.32 In fact, the film appears to have borrowed some elements from Tolstoy’s Anna Karenina, including the theme of a woman in a loveless marriage having a passionate affair with her true soul mate as well as the intense internal conflict stirred by her attachment to her offspring, from whom she must part when she finally leaves her husband for her lover. In the years before Love and Duty was released, a popular adaptation of Anna Karenina starring Greta Garbo and John Gilbert (Love, Edmund Goulding, 1927) had played in Shanghai. In Love and Duty, some aspects of Ruan Lingyu’s character, Yang Naifan, seem to borrow from Garbo’s Anna. There is, for instance, a moment soon after she has left her husband and taken flight with her lover when the sight of some anonymous small children drives her to tearful despair because they remind her of her own children, whom she has abandoned. Also in both films, the forsaken husband falsely tells the children that their mother is dead.

Most interesting in the present context is a fantasy sequence in Love and Duty that features its male protagonist in a heroic sword-fighting role clearly drawn from Western cinema precedents, particularly Douglas Fairbanks films like The Three Musketeers (Fred Niblo, 1921), Robin Hood (Allan Dwan, 1922), and The Black Pirate (Albert Parker, 1926). Read against the film as a whole, the sequence serves not only to reveal the extent of Hollywood’s colonization of the horizons of the public imagination in Shanghai at the time but also, as would be the case with Crossroads, to critique that same set of romantic conventions as ultimately empty in the face of the reality in China. In the period of taboo romance before Yang Naifan leaves her husband, her lover, Li Zuyi, picks up a novel while he is at home alone, yearning wistfully for her. A close-up shows the book cover with the title A Hero in English and Qinghai yingxiong 情海英雄 (Hero of boundless passion) in Chinese. Spurred by reading the text, Li has a daydream fantasy that casts himself in the hero’s role. In the fantasy sequence, his dress and demeanor strongly resemble those of Douglas Fairbanks in The Black Pirate. While strumming a Spanish mandocello on a veranda, he hears the cry of his lover, cast in his reverie as a damsel in distress being molested by a bearded pirate. In physical Douglas Fairbanks fashion, the hero zip-lines down from the porch on a rope, leaps off a ledge to reach the woman, wrests her away from the scoundrel, and challenges the latter to a duel with drawn sabers as she looks on in alarm. During their fight, the hero manages to flip his rival’s saber into the air and catch it, thus disarming his opponent, but he only laughs cavalierly (just as Fairbanks would smile through sword fights in The Black Pirate) while chivalrously tossing the saber back to the villain (as Fairbanks does in The Three Musketeers) to continue the fight evenly. After fatally stabbing the pirate, the hero is, of course, rewarded with the embrace of the rescued maiden. The film then cuts to an intertitle of the novel’s text, reading, in part, “In the end he won the final victory in affairs of the heart.” Later, after his real-life lover has agreed to leave her husband and run away with him, Li grabs the same novel, kisses it, raises his fists and his gaze to the air in a dramatic gesture, and announces, “I’ve won!” As in Crossroads, the sequence shows the extent to which the character’s fantasy life—and, to some extent, his actual behavior—has been determined by the narrative tropes and iconography of Western popular romance as promulgated by Hollywood (the novel clearly prompting the cinematic imaginary of Fairbanks films). However, the entire sequence is rendered ridiculous in retrospect when, in the film’s “real” world, Li and Yang soon live in hopeless poverty after she leaves her husband, Li literally works himself to death for almost no money (apologizing for ruining Yang’s life before dying), and then Yang labors as a seamstress to raise their daughter and can only watch her other children in anonymity from afar, à la Stella Dallas, eventually committing suicide to spare her now-grown daughter the shame of her existence. In short, the film ends in such unrelenting melodramatic pathos that the Hollywood-fueled fantasy of romantic heroism is exposed as a dangerous lie—even if it is exposed as such partly through the deployment of other Hollywood conventions, particularly those of maternal melodrama.

Insofar as they serve as parodies of Hollywood and implicitly as critiques of the inadequacy of the Hollywood imaginary for molding the sorts of citizens needed by China in times of national crisis, the dream sequences in Love and Duty and Crossroads were cinematic counterparts to critiques already abundant in Chinese critical discourse by the early 1930s. Lu Xun 鲁迅, for example—widely considered the “father of modern Chinese literature” and China’s most prominent cultural commentator until his death in 1936—aimed his satirical ridicule both at Hollywood and at Chinese intellectuals naive enough to think they could change Hollywood by criticizing it. In 1930, a year before Love and Duty was released, Lu Xun mocked Chinese audiences for indulging in Hollywood-induced fantasies, imagining they could enjoy pleasures that were in fact only available to their colonizers (bearing in mind that Shanghai itself was a colonized city at that time, with large sections of it outside Chinese sovereignty in the form of foreign concessions)—making such audience pleasure essentially a form of masochism.33 In an introduction to his translation of the Japanese essay “Modern Film and the Bourgeoisie,” Lu Xun derided Chinese fans’ worship of “valorous knight-errant” figures in Hollywood movies, and he specifically criticized the Chinese reception of Douglas Fairbanks.34 Leading leftist screenwriters and critics Zheng Boqi 郑伯奇 and Xia Yan 夏衍 also wrote scathing critiques of the influence of Hollywood on the Chinese imagination. In a 1932 essay, Zheng Boqi (under the pseudonym Xi Naifang 席耐芳) described Hollywood films as “poison” and criticized Chinese filmmakers for imitating the bourgeois content and ideology of Western films (though he acknowledged that those films should be studied for their methods and techniques).35 During the hard film versus soft film debate, Xia Yan denounced Frank Capra’s Lady for a Day (1933) as “merely a kind of fantasy” (using the English word fantasy in the original).36 In the context of Japan’s invasion of Manchuria in 1931 and bombing of Shanghai in 1932, Zheng Boqi wrote the following:

Right now in China the situation intensifies by the day. The entire old society is rapidly crumbling. Imperialism invades ever more powerfully, carving up the country like a melon. The petty urbanites, habituated to daydreaming, cannot help but gradually face these grim realities. At such a time, we can expose the stark social realities for them to see, no longer needing to gingerly pander to them with saccharine fantasies.37

The dream sequences of Love and Duty and Crossroads can thus be seen as cinematic representations of the “daydreaming” and “fantasies” that Zheng Boqi and Xia Yan insist the petty urbanite audience needs to go beyond, presumably with the help of more critically realist leftist Chinese films.

Indeed, in Crossroads, particularly in the beginning and ending, the lives of Lao Zhao and Zhiying ultimately are framed not in terms of their potential romantic coupling as the ultimate means of personal fulfillment but rather in terms of generational crisis due to economic depression and imperialist aggression, thus earning the film its status as a social realist work of the Left-Wing Film Movement. In his review of the film, Xia Yan specifically praised as “utterly scrupulous” the melancholy opening segment, in which Lao Zhao prevents his despondent classmate from committing suicide and then they together lament their situation in Shanghai’s back alleys—though Xia also commented that the screenplay becomes “completely careless” later, presumably corresponding to the switch in tone from critical realism to romantic comedy.38 At the film’s end, the romantic couple Zhiying and Lao Zhao both have lost their jobs and apparently each other. But then, running into each other by chance along with each of their best friends at the same spot on the river where the other friend—one of the original four college graduates—had earlier tried to drown himself, they learn from a newspaper report that he in fact has now committed suicide, and they consider their own lives as facing a choice represented on one side by the suicide victim’s loss of hope and on the other by another old college friend who had returned to his homeland in northeast China to fight against the invading Japanese. The remaining friends quickly vow to bravely go on living and struggling, and they march back into the Shanghai cityscape arm-in-arm in an on-location extreme long shot. I will return to the questions raised by this ending and those of other left-wing films, as well as the sometimes jolting contrasts in tone between the comic main part of the film and these more serious segments, but first let us look in detail at another leftist classic of 1937 that borrowed even more explicitly from Hollywood than did Crossroads but that also subverted Hollywood conventions yet more starkly.

Undermining Elevation: Street Angel

Yuan Muzhi’s Street Angel took its title from Frank Borzage’s 1928 silent classic of the same name and borrowed many of its motifs from that film as well as from Borzage’s international blockbuster 7th Heaven from a year earlier. In Yuan’s film, however, Borzage’s themes of poverty and redemptive love get reworked in a way that both borrows from the master of Hollywood melodrama and also fundamentally undermines his type of storytelling, thus serving as another prime example of the “Shanghaiing of Hollywood” in the 1930s.

In analyzing the various elements of adaptation from one film text to another, we can usefully employ Jörg Schweinitz’s study of film “stereotypes,” which he defines broadly as “conventionalized representational patterns shared and accepted by artists and viewers within a given period.”39 Schweinitz takes stereotypes as “differentiated functional forms that ensure the activation of affective responses (as intended in aesthetic fictions) in the addressee.”40 That is, insofar as they activate codes already understood by the viewer, stereotypes create affect as part of the spectator’s activity, as we already saw in the case of the “paradigm scenarios” presented by the “rom-com of misrecognition” in Crossroads. In other words, viewers’ emotional responses—such as the pleasurable (because temporary) anxiety evoked by misunderstandings between lovers in romantic comedies—are cued not only by the film they are currently watching but by the previous films from which the spectator learned the same conventions.

One type of conventionalized pattern or stereotype that helps to constitute popular film narratives is the repeated (and therefore immediately recognizable) character type. Yuan Muzhi’s Street Angel shares several character stereotypes with its Borzage antecedents. The male protagonist is a handsome, talented, and good-hearted young man, played by Charles Farrell in both of Borzage’s films and by a youthful Zhao Dan (the same lead actor as in Crossroads) in Yuan’s film (Figure 8). Opposite him is a virtuous young beauty, played by Janet Gaynor in both of Borzage’s films and by the popular singer Zhou Xuan 周璇, in her first film role, in Yuan’s film (Figure 9). In 7th Heaven, the heroine has an evil older sister, on which the heroine’s sister in Yuan’s Street Angel initially appears to be modeled—even down to such details as the penciling of her eyebrows (Figure 10). In Borzage’s Street Angel, a prostitute neighbor tries to come between the main couple in a manner that appears to be quoted with the older sister (also a prostitute) in Yuan’s film (Figure 11). The male protagonists in both 7th Heaven and Yuan’s Street Angel have a ragtag band of misfit friends, simultaneously illustrating that they are respected and that they identify with society’s lower stratum (Figure 12). Several other details of the Chinese Street Angel also are borrowed from Borzage, from the visual trope of a makeshift skywalk between two facing apartments (Figure 13) to the plot scenario of the female protagonists being protected from menacing policemen by a false claim of residency (both of which had appeared in 7th Heaven).

According to Schweinitz, “classical film genres are nothing more than open-structured repertoires of narrative stereotypes,” including “plot patterns” and “narrative formulas” in addition to character types. The basic plot pattern that Yuan’s Street Angel shares with Borzage’s films is a love story that includes the clichéd “boy loses girl” part of the stereotypical romantic progression. That is, just as some sort of misrecognition was key to the rom-com paradigm scenario in Crossroads, these films use the familiar narrative formula in which a romance is threatened either by a falling out between the lovers based on a misunderstanding or by external forces that tear them apart. In both of Borzage’s films, Gaynor plays a young woman who, having initially been forced into prostitution by desperate circumstances, finds redemption through a romance with the handsome and caring young man played by Farrell. In 7th Heaven, made in Hollywood but set in Paris, just as the couple have committed to marriage, the man, Chico, is torn away by conscription to fight in World War I. He eventually is reported to have died in battle, but the couple is happily reunited in the end when it turns out he survived. In the case of Borzage’s Street Angel, made in Hollywood but set in Naples, the woman is thrown in jail by an unforgiving cop—unbeknownst to the man, who believes she has simply abandoned him, leaving him heartbroken and furious with her.

A paradigm scenario of a lovers’ quarrel based on a misunderstanding not only temporarily pits the lovers against each other but also in a sense pits the spectator against the protagonists, insofar as the former still wants the romance to succeed while the latter, at least ostensibly, have given up on it. That is, in terms of how the film is, as Carl Plantinga puts it, “affectively prefocused” for the spectator, the lovers’ quarrel section of a romantic comedy introduces a “benign incongruence” between spectator and character desires: whereas one or more of the lovers has decided that the romance is no longer possible or even desirable—often owing to a mistaken construal of the other’s intentions or character based on some kind of misinformation or misinterpretation—the spectator construes the situation quite differently and strongly desires that the misunderstanding be cleared up and the couple reunited.41 In Borzage’s Street Angel, the male protagonist’s false belief that his lover has intentionally left him, when in fact she has only been unwillingly separated by circumstances, is cleared up in the film’s climax, when the lovers are reunited and all is forgiven after her release from jail. In Yuan’s Street Angel, the scenario of misunderstanding and breakup occurs when the male protagonist, Xiao Chen, mistakenly believes that his girlfriend, Xiao Hong, is falling for a man who is in fact preparing to purchase her as his concubine from her guardians, the owners of a teahouse. Rather than talking things through, the lovers have an angry quarrel and are separated during the following scenes, causing the same sort of temporary benign incongruence between the desires of the audience and the behavior of the protagonists. The incongruity is resolved and the lovers reunited only after Xiao Hong learns of her guardians’ plan to sell her and, in desperation, goes to Xiao Chen for help.

Yet, it is after this reconciliation—the final “boy gets girl back” stage of the clichéd rom-com formula—that the Chinese Street Angel departs significantly from the narrative formulas of the Borzage films. The latter feature happily-ever-after endings at the very moment that the artificial separations are ended and the couples reunited. In Yuan Muzhi’s film, however, this romantic plot pattern is resolved with fully half of the film still remaining, and the plot will now take several twists that do not echo anything from Borzage’s films. Taken together, these developments fundamentally shift the central theme of Borzage’s films—the transcendent spiritual power of heterosexual romantic love—to a social realist exploration of class, in particular, the plight of the urban poor and especially women, who are exploited and trafficked.

One clear way in which Yuan’s Street Angel reworks its Hollywood antecedents is through the metaphorical use of verticality in urban architecture. In the most celebrated shot in 7th Heaven (actually two, though the spectator is not supposed to notice the cut), when Chico first takes Diane to his place to allow her to hide from the police by posing as his wife, the couple climb a spiral staircase all the way to the seventh floor while the camera tracks up in sync with them as if looking through a transparent fourth wall. (In fact, the shot was achieved by a camera rising in an elevator on a set that lacked that fourth wall.) The ascent of the camera to the seventh floor emphasizes the symbolism of the couple’s rise to the utopian space of Chico’s living quarters—the “seventh heaven” of the film’s title—an implausibly spacious apartment atop a high rise in Paris (implausible because Chico is only a city sewer worker), where the couple will fall in love, safe from the dire social conditions that previously had threatened them. In Borzage’s slightly later film Street Angel, the theme is repeated when Farrell’s character tells his lover, “This is just the beginning! We are going to climb . . . higher and higher!”—which amounts to an intertextual reference to the very popular previous film. Love in Borzage’s films has its sensuous elements, but it is above all a spiritual condition that offers redemption and an almost mystical, telepathic bond between two souls. Indeed, in terms of affective impact on the spectator, the love on the seventh floor activates what Plantinga calls the emotion of elevation. Following psychologist Jonathan Haidt, Plantinga defines elevation as “the opposite of social disgust, triggered by the witnessing of acts of human beauty or virtue,” and drawing its “power from a desire for moral betterment.”42 In Plantinga’s view, “this emotion of elevation quite aptly describes the sort of emotional response generated by many narratives of the melodramatic variety,” a genre that Borzage’s films mix with romantic comedy.43

In Borzage’s Hollywood films, elevation is both the effect of love and the implied purpose of art. In Street Angel, for example, the male protagonist, Gino, is a painter who falls in love with a young woman, Angela, who has run away with a circus to escape a criminal justice system that tried to jail her for robbery and solicitation for prostitution when she was desperately trying to get money to buy medicine for her dying mother. During the couple’s courtship, Gino paints a portrait that depicts Angela in a golden light, looking divine and innocent. She finds it beautiful but protests that she is not really like the figure in the painting, to which Gino answers that it is how he sees her. The painting eventually is sold for a pittance to a dishonest art dealer, who adds a halo to her head to make her literally angelic and sells it to a church as a lost masterpiece. At the film’s climax, after Gino attacks Angela and chases her into the very same church in retaliation for leaving him and breaking his heart, it is the occasion of unexpectedly seeing the painting again that makes him see her once more as angelic rather than evil and find it in his heart to forgive her, even without having cleared up the misunderstanding between lovers (that she in fact did not leave him willingly but was rediscovered by the Naples police and arrested). The painting thus serves the same function for Gino that Borzage hopes his films will serve for the viewer: it elevates him to a more spiritual plane of existence, where he can see the good and beauty in life. If it is a realist representation, it is in a prescriptive rather than only descriptive or verisimilar mode, depicting in art an ideal that can then change the viewer’s perceptions and behaviors.

In contrast to Borzage’s affective elevation, associated with literal architectural elevation in the case of 7th Heaven, some Shanghai films of the 1930s would put quite a different twist on the theme. The 1933 film Daybreak (Tianming 天明; Sun Yu) imitates the upward tracking shot of 7th Heaven when its protagonists, a young couple from the countryside, first climb the stairs to the Shanghai apartment that they would now share with friends.44 The idea of climbing to a refuge from the social ills of the city is cruelly mocked in a later shot, however, when we see the heroine, played by Li Lili 黎莉莉, make the same climb after having been raped by the boss of her factory—the first of a sequence of events that would have her tricked and sold into prostitution.

Yuan’s Street Angel systematically employs the trope of elevation as a metaphor for class. The film introduces its fictional world with a shot of a tall skyscraper (clearly a model). The camera tilts down from a view of the top of the building to the bottom, even going beneath ground level to metaphorically reveal what a subtitle calls “Shanghai’s underground stratum.” The shot thus reverses the direction of Borzage’s elevation in 7th Heaven, instead frankly taking the viewer into the dregs of urban humanity rather than offering the fantasy of any heaven on high accessible to the poor. Even though, later, there are endearing scenes of Xiao Chen and Xiao Hong falling in love while interacting through the space between their two apartments—both by singing and playing music to each other and by crossing on their makeshift skyway—their rooms are only on the second floor of the Shanghai tenements, and in any case, they abandon those quarters to go into hiding after Xiao Hong’s guardians try to sell her. The theme associating architectural elevation with class is explicitly returned to when Xiao Chen and his friend attempt to find protection for Xiao Hong by visiting a lawyer whose ad they had seen in an old newspaper. The lawyer’s office is high up in a skyscraper—“This truly is heaven!” remarks the friend—but the poor, shabbily dressed young men are depicted as comically out of place in such an upper-class location, from which they indeed are ejected as soon as it becomes apparent that they cannot pay the lawyer’s fees.

Yuan’s Street Angel evokes but then departs from the paradigm scenarios of Borzage’s films in other significant ways. The character type of the evil sister, for example, seems to follow its 7th Heaven precedent during the first part of the film, when Xiao Hong’s “sister” (most likely just another orphan purchased by the teahouse owners, who then forced her into prostitution while making Xiao Hong entertain guests with her singing) mildly disciplines Xiao Hong and blocks her from spending more time with Xiao Chen, for whom the sister, Xiao Yun, harbors a crush of her own. Later in the film, however, Xiao Yun facilitates Xiao Hong’s escape, develops a romance with one of Xiao Chen’s friends, and occupies a hideout together with the lead couple and their companions. Following this affective rehabilitation of the evil sister figure, the character is sacrificed in the end when her former master finds her and fatally wounds her with a knife.

The incredibly dark final scene of Yuan’s Street Angel, in which the sister dies slowly for lack of money for medical care and the young protagonists seem to have nowhere left to hide from the social evils that victimize them, could not be more different from the elevated endings of Borzage’s films, in which love conquers all and the couple is happily reunited. Near death, after her love interest, Lao Wang, is mentioned, Xiao Yun babbles her final words: “He’s a good man . . . he helps poor people . . . like ant . . . ants.” What starts as a comment on her boyfriend seemingly turns incoherent, but this apparent non sequitur in fact refers to a broader discourse of social commentary in Shanghai in the mid-1930s. As Wen-hsin Yeh has detailed, the widely read leftist journal Dushu shenghuo 读书生活 (Reading and livelihood) had popularized the use of “ants” as a metaphor for China’s swelling numbers of petty urbanites. In a social commentary section in May 1935, contributor Li Gongpu 李公朴 had written that while petty urbanites might think of themselves as nothing but “ants,” they should consider themselves as potentially forming “an army of ants in defense of the nation” in the face of Japanese imperialism.45 Soon an Ants Club (yishe 蚁社) formed and grew to a membership of more than ten thousand, some of whom eventually joined the Communist base in Yan’an during the war against Japan.46 Xiao Yun’s dying reference to “ants” thus not only evokes despair at the helplessness of her and her friends’ current situation of oppression but implicitly hails a potential future of collective revolutionary resistance.

Irresolutions, Fissures, and Transformations

The ending of Yuan’s Street Angel obviously departs fundamentally from Hollywood’s style of storytelling. Hollywood’s classical mode of narration, as David Bordwell has detailed, will as much as possible reach a satisfying conclusion for its main plot threads, providing, if not total closure—because inevitably some details will remain unresolved—then at least a “closure effect,” which, more often than not, features the lead romantic couple in a stable state of happiness.47 As David R. Shumway puts it, in classical Hollywood screwball romantic comedies, romance is “a complex and tenacious ideology” that “holds ‘the bliss of genitality’ to be the end of desire. When the right man or woman is found and returns one’s love, the subject will be satisfied, will lack no more.”48 The story is over. In contrast, by the end of Yuan Muzhi’s Street Angel, the film has shifted into a mode of unsparing critical realism, in which nothing is happily resolved and the love story itself has become secondary to the larger story of social oppression and the strongly implied need for class struggle and revolutionary change. The final shot reverses the first shot after the opening credits, tilting from “Shanghai’s lower stratum” back up the towering skyscraper, but by now such a metaphorical elevation has taken on a wholly ironic tone. Far from leaving the movie theater with a sense of moral elevation, satisfaction at the plot’s resolution, and happiness for the attractive fictional couple, the audience would leave with a deep sense of uneasiness, incompleteness, and perhaps something more like moral outrage.

The irresolution with which Yuan Muzhi’s Street Angel ends, so uncharacteristic of classical Hollywood, was in fact common among Shanghai’s leftist films of the 1930s.49 Yuan’s previous film, City Scenes (Dushi fengguang 都市风光, also translated as Metropolitan Scenes or Scenes of City Life; 1935), which was both a comedy and an avant-garde experiment, ends with all the main characters literally pulling in different directions, unsure of which way to go, until the shot gives way to a giant animated question mark that fills the frame and ends the film.50 The 1933 film adaptation of Mao Dun’s famous May Fourth–era novel Spring Silkworms (Chuncan 春蚕; Cheng Bugao 程步高) shows the devastation of the Chinese craft silk industry by imperialist economic competition and military aggression without offering any sort of resolution or solution. Another 1933 leftist classic, Sun Yu’s silent Little Toys (Xiao wanyi 小玩意), ends with its heroine losing her sanity and shouting in the streets, warning of imminent destruction by imperialist invaders, while the same director’s Big Road (Da lu 大路; 1934) concludes with most of its main characters being killed by strafing Japanese fighter planes—until, through double exposure, the spirits of the slain appear to rise up from their own corpses, as if to struggle on.51 The ending of Crossroads, described earlier, also provides little closure: the remaining protagonists have vowed to struggle rather than commit suicide, but what will actually happen to them? We have no idea.

Refusal of any reassuring closure not only goes against the tenets of classical Hollywood; it even defies the conventions of much critical realism. In his study of British social problem and working-class realist films, John Hill argues that the nature of narrative itself dictates that fictional resolutions be provided to the problems presented in social realist films. Citing Tzvetan Todorov’s definition of a “minimal complete plot” in narrative—an initial equilibrium, a disturbance of that stable state by some force, a struggle against that force, and finally the reestablishment of a new equilibrium—he concludes that “there is a presumption, built into the very structure of conventional narrative, that ‘problems’ can be overcome, can, indeed, be resolved.”52 Hill quotes Thomas Elsaesser as also finding a link between an ideology of “affirmation” and the conventions of narrativity: “As he explains it, there is ‘a kind of a priori optimism located in the very structure of the narrative . . . whatever the problem one can do something about it.’”53 Therefore, argues Hill, “it is in the nature of the conventions of narrative that these ‘problems’ be overcome”; “the articulation of the film’s ‘social problem’ into the problem-solving structure of narrative necessarily implies that it too is capable of resolution.”54 This often leads to socially progressive critical realist films having incongruously reassuring or traditional conclusions: “it is this need for some sort of narrative resolution which tends to encourage the adoptions of socially conservative endings.”55 An intractable problem like poverty, for example, the real resolution of which may only be accomplished by transformation of the entire social order, may appear in a film to attain a piecemeal resolution that leaves the film’s characters in a more positive equilibrium at the film’s end. This is partly due to the convention of classical cinema to have individual characters act as the principal agents of causality in the film’s plot; the problems they deal with might be as broad as society, but their resolutions are artificially reduced to the fortunes of individuals.56

This kind of resolution is evident in countless Hollywood films that might raise difficult issues, such as class disparity, but then solve them through fanciful and purely personal means. A typical example is Sunny Side Up (David Butler, 1929), which played in Shanghai and was the first “talkie” starring the already famous screen couple Janet Gaynor and Charles Farrell—the same pair that had starred in Borzage’s 7th Heaven and Street Angel. Intertitles at the film’s beginning establish the stark divide between the poor masses (the “four million” of New York City) and the ruling upper class (the “400” of a rich Long Island suburb). A shop clerk played by Gaynor goes with her three friends to the posh suburb, where a rich young man (Farrell) wants to use her to make his girlfriend jealous. Class conflict is resolved through a happy ending when the rich fellow realizes he has fallen in love with the poor shop girl and wants to marry her instead. The protagonists as causal agents thus appear to overcome intractable class divisions through individual romantic coupling and random twists of fate.

Some 1930s Shanghai films considered to be progressive, if not fully “left wing,” also resorted to such measures, providing an audience-pleasing “bright tail” (guangming de weiba 光明的尾巴) ending that implausibly “solves” the problems of poverty and injustice for their protagonists.57 The leftist films on which we have focused in this chapter, in contrast, radically refuse recourse to any pat resolution, whether an appeal to conservative Confucian family values, a reversion to Hollywood-style genre formulas, or a resort to the tendency even of social realist films to provisionally “solve” their problems at the end, at least at the level of individual characters. Instead, the problems are presented as unresolved or left to the films’ protagonists—and, more to the point, implicitly to its spectators—to collectively solve after the films’ endings. In defying classical cinema’s convention of resolution, leftist filmmakers like Shen Xiling and Yuan Muzhi (who wrote as well as directed Crossroads and Street Angel, respectively) were responding in part to the calls of prominent leftist critics. Victor Fan has detailed, for example, how Zheng Boqi called on screenwriters to “reconfigure the classical Hollywood notion of plot development and resolution, in order to avoid providing imaginary solutions to real social problems that required political actions.”58 He instead wanted spectators to experience the social contradictions represented without being reassured, so that the pain of the injustices depicted was not washed away by the pleasure of a hopeful ending.

Aside from the refusal of resolution, other challenges to classical film narration in Shanghai leftist films include sometimes abrupt shifts in tone of narration and in characterization. Audiences today, certainly including those who continue to be raised mainly on Hollywood cinema, often perceive these sudden shifts of tone and character as weaknesses or even mistakes in the films’ construction. As early as 1916, Hugo Münsterberg, in one of the first theoretical discussions of the cinema, insisted that the “chief demand” in the depiction of characters in film “is that the characters remain consistent.”59 More broadly, as Bordwell, Staiger, and Thompson have detailed, the classical Hollywood narration highly values what they call the “homogeneity of the fictional world.”60 The classical film pursues, in addition to closure at the end, a sense of diegetic unity, a clear chain of cause and effect in plot, and characters with consistently demarcated traits as well as, in the case of protagonists, psychological depth—meaning, for example, that any change of heart has obvious motivations through narrative events that may entail personal growth. We have seen how Yuan’s Street Angel shifts from a typical rom-com formula to a darker social realism at about the midway point (though romantic comedy reappears at some points later), finally ending in tragedy. The ending of Crossroads at the waterfront, too, seems not only out of character with the rom-com formula but also quite sudden—the entire scene of reunion, revelation of the friend’s suicide, and collective vow to struggle on lasts less than two minutes. Even at the time of its release, Zheng Boqi and Xia Yan, though generally praising the film, criticized its inconsistencies and choppy changes in tone from comedy to tragedy.61 In her study of 1930s leftist Shanghai films, Laikwan Pang also notes that Crossroads “is a light-hearted romance comedy, but the comic portion is sandwiched by two tragic sections. . . . The tragic and comic sections are juxtaposed to each other without any transition.”62 Such shifts often involve what appear to be sudden, seemingly unmotivated changes in character psychology. One well-known example is the Li Lili vehicle Daybreak, in which the lead character transforms from being a simple country girl to an undercover revolutionary with very little of the process of political education that would have been depicted, at least in a quick montage sequence, in a standard socialist realist film from the Soviet Union or later in China. Another Li Lili film directed by Sun Yu, Blood of Passion on the Volcano (Huoshan qing xue 火山清血; 1932), appears to be a work of melodramatic social realism about class exploitation in China for the first thirty-seven minutes but then, for the last hour or so, becomes a romantic adventure/revenge story set on an unnamed tropical island in the Pacific Ocean.

Pang notes that the “sudden emotional shifts” and “‘jumpy’ structure” of many Chinese films of this time can be traced in part to the conventions of traditional Chinese novels and plays, conventions that not only survived but thrived in the burgeoning popular literature of the early twentieth century.63 In the Chinese narrative tradition, as Andrew Plaks has shown, there is an “essential ambivalence in the apprehension of human character,” the upshot of which is “the impression of inconsistency: wavering, backtracking, side-switching, mood-changing, in the portrayal of the heroes of the narrative tradition.”64 He goes on to suggest that while some examples of this might be explainable in terms of psychological realism, “it may be more accurate to understand the phenomenon in terms of an aesthetic system that simply does not demand consistency in the representation of human behavior. Far from being a critical fault, this sort of flexibility, or fluidity, in character portrayal stands as a clear index of the greatness of a given work.”65 According to Plaks, what is valued in the traditional Chinese hero is not necessarily a consistent nobility of character but rather “the capacity of the hero to respond to the needs of the moment” or “to move with the times,” rather than just to have “singleness of purpose as a heroic quality.”66 Given such narrative conventions, it is reasonable to assume that the majority of Chinese film spectators in the 1930s—not the intellectual elite, whose tastes had been formed by the May Fourth Movement, but the “petty urbanites,” who were likely also to be readers of both the classic Chinese vernacular novels and the so-called Mandarin Ducks and Butterflies fiction that had become popular in the Republican era—would have more readily accepted sudden shifts in characterization and narrative direction.67 In fact, for the purposes of leftist film viewing, fictional protagonists’ capacity for personal metamorphosis to respond to the needs of the moment would only reinforce the filmmakers’ implicit agenda of transforming petty urbanite spectators into revolutionary subjects.

Despite the possible influence of the vernacular storytelling tradition in China, in the context of the Hollywood-dominated box office in 1930s Shanghai, we must view the left-wing films examined here as radical in their departures from the Hollywood norm in terms of form as well as content. Leftist filmmakers like Yuan Muzhi were very cognizant of the recent Soviet experiments in film form, as is evident in part from the spectacular urban montage sequence that he used both in City Scenes and as the opening title sequence of Street Angel, one of the finest examples of an avant-garde montage aesthetic in Chinese cinema of the time. For over two minutes, beneath the opening credits of Street Angel is a dazzling and disorienting montage of Shanghai city life that has precedents in “city symphony” films like Walter Ruttmann’s Berlin: Symphony of a Great City (1927) and Dziga Vertov’s Man with a Movie Camera (1929) as well as the city montages in some Hollywood films, such as Murnau’s Sunrise or even the self-conscious opening montage of the Jean Harlow vehicle Bombshell (Victor Fleming, 1933). The Street Angel title sequence frames the entire film to follow with a rapid series of urban scenes and rhetorical tropes that serve to summarize semicolonial Shanghai. The montage lasts 130 seconds and consists of eighty-six shots (including some repetitions of the same shots). The resulting average shot length of one and a half seconds still does not convey the disorientation of the montage’s speed of juxtaposition, as two sections of the montage consist of a series of superimposed shots—all of this occurring underneath the further superimposition of the film’s title credits themselves, so that, counting the titles, there are two to three image layers virtually throughout the title sequence. The dazzling sequence is packed with disparate images: shaky travelling shots propelling the viewer down brightly lit Shanghai streets; closer shots of neon signs, in both Chinese and English, advertising restaurants, inns, places of entertainment, Western products like whisky and Palmolive soap, and international corporations like Frigidaire; shots of city parks with strolling, well-to-do people juxtaposed with quicker shots of city streets and bridges packed with ordinary people, cars, motorcycles, buses, rickshaws, bicycles, and streetcars; Western statues superimposed in contrasting canted angles; a cascade of coins suggesting riches (very similar to that in the Bombshell title-sequence montage); four shots of a pair of bronze lions, cast in England, which for decades guarded the entrance to the Hong Kong and Shanghai Bank; shots that tilt up the facades of various Western-style buildings, several of them identifiable as Christian churches, offering not only a survey of Shanghai’s famous architecture but an exposé of the city’s semicolonial status; and a spectacular finale in which twenty-five shots in just twenty-one seconds juxtapose fireworks explosions with shots of ballroom dancing as well as chorus girls, superimposed electric street lamps, and, finally, the return of the beginning phantom ride down a neon-lit Shanghai street at night, metaphorically suggesting that the eighty-four shots in between had represented views from a whirlwind trip through the city.

This opening sequence no doubt is indebted to the Soviet theories of montage, which had entered China in the early 1930s in the form of screenings of landmark Soviet films like Storm over Asia (Vsevolod Pudovkin, 1928) and The Road of Life (Nikolai Ekk, 1931) as well as translations of Russian film theory, particularly Pudovkin’s The Film Director and Film Material (1926), detailing his theory of montage, which had been translated into Chinese in 1933 by Xia Yan and Zheng Boqi. Moments of Soviet-style montage had appeared in landmark films of the Chinese Left-Wing Film Movement by the mid-1930s. In Big Road, for example, the climactic scene in which the films’ protagonists are attacked by Japanese fighter planes is cut in a manner that briefly recalls Eisenstein, while Jessica Ka Yee Chan has detailed how a Pudovkin-style contrast montage sequence is used to demonstrate class divisions in New Women (Xin nüxing 新女性; Cai Chusheng 蔡楚生, 1934).68

What I wish to emphasize, however, is not just the influence of Soviet avant-garde aesthetics on certain montage sequences but more particularly how the overall structure of the films also features jarring juxtapositions that function in a way analogous to Soviet montage. In other words, the principles of montage—as Eisenstein put it, “Collision. Conflict between two neighboring fragments. Conflict. Collision”—are operative through the relationships between scenes and sequences, not just between shots.69 That is, although the lack of unity in left-wing Shanghai films no doubt is due in part to the conflicting demands of political priorities (Communist-leaning writers and directors who wished to use cinema to awaken the revolutionary masses) and commercial priorities (film studios’ need to attract popular audiences weaned largely on Hollywood imports), such clashes also become the very means of the films’ agitation. The conflicting agendas of political art versus entertainment are demonstrated, for example, by the opening three sequences of Street Angel, in which the experimental city montage title sequence, loaded with politically charged rhetorical tropes emphasizing Shanghai’s imperialist oppression and stark class divisions, is followed by a nearly four-minute slapstick comedy sequence in which several of the film’s main characters are introduced in the context of a wedding procession down Shanghai’s narrow streets, a parade that includes a full Western-style marching band in which the main male protagonist, Xiao Chen, plays the cornet. The comedy sequence features eight distinct sight gags, as, for example, when Xiao Chen peeks into the bridal sedan, sees that the bride is cross-eyed (a joke with class connotations, because she must come from a rich family) and goes temporarily cross-eyed himself in response. With the end of the comical parade sequence, the film suddenly shifts gears again, becoming a vehicle for featuring a popular music star after Xiao Hong, who had been looking down on the wedding parade from a second-floor balcony, turns to go back into the teahouse in which she works as a singing hostess.70 Zhou Xuan was already a popular singer playing her first starring role in a film, so that the song she sings for the guests, “Song of Four Seasons,” would have been an attraction in itself for the film’s original audience, who could even sing along with the help of the lyrics provided in Chinese subtitles, complete with a little white ball bouncing from word to word to help spectators sing in time with the song, karaoke-style.

Thus, just within the first ten minutes of Street Angel, there is an avant-garde city montage, a slapstick comedy sequence, and a pop song performed by a popular singer. In addition to all that, however, there is in these first few minutes a strong undercurrent of documentary-style critical realism, and not just through the shots of city street scenes in the opening montage. The wedding procession sequence, as well, uses footage from a real parade and includes no fewer than sixteen shots of groups of random onlookers watching from the sidewalk and from windows and balconies. These are grouped into two sequences of eight quick shots each, lasting less than a second per shot but evidently showing real people rather than actors. Then, during Xiao Hong’s performance of “Song of Four Seasons,” the film repeatedly cuts away from shots of her singing to images illustrating her song of woe—including shots of war and fleeing refugees, as we gather that she herself was orphaned by war in the northeast. In this way—much as in the case of a similar performance by Li Lili in Sun Yu’s Big Road in 1934—the spectacle of the attractive singing star is disrupted by disturbing documentary-style scenes of chaos and destruction.71

In sum, the juxtaposition of these different styles—Soviet-style montage, slapstick comedy, musical, documentary—like the broader collision of light romantic comedy and dark social realism evident in both Crossroads and Street Angel, creates stylistic and narrative fissures within the texts, which, along with the lack of Hollywood-style resolution in the plot, represent a fundamental departure from classical Hollywood narration. On one hand, the promiscuous mixing of genres recalls that of early (preclassical) cinema, in which, as Miriam Hansen has described, “the juxtaposition of generically distinct frames of reference creates a moment of surprise, a glimpse of a type of narrative whose outcome need not necessarily be known,” requiring a viewer capable “of shifting between diverse positions and referential contexts,” thus arguably encouraging agency in the public sphere in a way that would be lessened by Hollywood’s comparatively unified fictional realism.72 At the same time, given that the leftist Shanghai films of the 1930s intervened in a cinematic culture that already was largely defined by Hollywood’s hegemony, their genre mixing and narrative fissures also suggest a more overtly avant-garde practice, specifically in the sense of the modernist “open works” analyzed by Umberto Eco.73 They use affectively prefocused paradigm scenarios and character types based on the precedents of Hollywood and other entertainment cinemas to stir strong emotions and identifications in the audience, but then they quite intentionally leave those attachments suspended rather than being reassuringly resolved.74 The spectator is expected to be agitated by unresolved contradictions and nudged, not only by critical realist content but also by the provocatively open narrative form to take action in some way or another.

Cinema, with its “strange constructs,” is ideally suited to perform such work on the imagination. As Hansen argues in a discussion of Kracauer, constantly transforming images on the cinema screen provoke a “psycho-physiological mimesis [that] affords the viewing subject the sensation of participating in this transformation, evoking the possibility—both threatening and liberating—of liquefying fixed structures of social, critical-intellectual, gendered identity.”75 In this light, the Hollywood-inspired fantasy worlds critiqued by left-wing critics and parodied by the dream sequences in Love and Duty and Crossroads should not be taken merely as instances of false consciousness that must be overcome to create properly revolutionary subjects. On the contrary, the very practice of envisioning new selves and ways of being, metamorphoses enabled even by Hollywood cinema and often required for survival itself in the rapidly modernizing world of semicolonial Shanghai, might have prepared a spectator not only for new modern possibilities like amorous encounters with an attractive neighbor within the anonymity of urban life but for transformations of a more politically revolutionary nature as well. Films like Crossroads and Street Angel encourage the viewer to imagine both potentialities, without necessarily providing the singular ideological point of view presented in official histories of the Left-Wing Film Movement. Moreover, though my argument does not apply to all the films that have been grouped within that movement, neither does it identify a peculiarity of only a couple of classics from 1937. In fact, Victor Fan has made very similar arguments about the early 1930s films of Sun Yu, perhaps the greatest Chinese filmmaker of his generation. Fan argues that Sun’s films were “narrativized with textual openness” that made room for the spectators’ active political interpretations.76

Consequently, we may view the type of realism in these films as not simply social realism, in which the real ills of society find representation on-screen, nor a variation on the fictional realism prized by Hollywood, but rather—or in addition—a practice of what I have called apophatic realism, in which it is hinted that there is a more essential “real” that escapes representation. Within the fissures created by unresolved juxtapositions of narrative modes, beyond the irresolutions with which these films end, represented only indirectly by open-ended questions (literally, in the case of the question mark ending City Scenes), lies an urgent yet elusive real of revolutionary transformation that could not yet be given positive configuration in cinematic representation in the context of the seemingly intractable social, economic, political, and military calamities of China in the 1930s, not to mention government censorship. Beyond showing the social realities of immediate crisis, these films endeavored to open up spaces in which, with the help of the spectator’s imagination and will, a revolutionary real could potentially make an appearance in the near future. Failures of diegetic unity allow for a shift of the affectively charged spectators’ attention toward the extradiegetic world of their real lives but with the opportunity to see that reality anew and the possibility of changing it. These films thus sought to instill a state of revolutionary liminality, in which the future may not yet have taken on any positive form but past imaginary identifications have been challenged, the present has been rendered unstable, and the possibility for reimagining the self through new social solidarities has been created.

The Question of Vernacular Modernism

The foregoing discussion suggests that the understanding of Shanghai leftist cinema of the 1930s as “vernacular modernism” must be clarified, taking into account possible contradictions within the concept of vernacular modernism itself when it is applied to cinemas around the globe. In her initial proposal of the concept, Miriam Hansen was addressing specifically the classical Hollywood film, arguing that we should see it as an instance of modernist art or culture because it not only is “part of the historical formation of modernity” but in fact offered “something like the first global vernacular”—a cultural form mediating the experience of modernity that could have global appeal and intelligibility partly because even its initial American mass audience, a diverse nation of immigrants, already required the negotiation of great differences.77 In this formulation of vernacular modernism, the visual vernacular language of modernity is the cinema produced in Hollywood.

In Hansen’s later discussions of the same concept, however—as well as in the work of several of her former students and others inspired by her work—the focus shifts from the films of Hollywood itself to the popular cinemas of other societies around the world, in particular, China and Japan, and the ways in which those popular cinemas creatively adapted the classical Hollywood narrative mode.78 Such a broad application of the idea of “vernacular modernism” opens it up to criticism that it homogenizes world cinema, ignoring the different historical contexts of each local experience of modernity as well as the vastly different cultural and linguistic forms and heritages that may get articulated and even preserved through local or regional cinematic practices.

In a sympathetic critique of Hansen’s concept of vernacular modernism, Daniel Morgan argues that the theory becomes least viable when it is presented as “a claim about the medium itself”; if any film that engages in some way with the condition of modernity is therefore a case of vernacular modernism, then every film would have to be included, if only because the medium itself is a modern one.79 Instead of taking this approach—which Morgan believes Hansen sometimes mistakenly pursues—vernacular modernism, according to him, has to be “based on an engagement with individual films, including their specific historical contexts.”80 In her own final (and longest) essay on the topic, Hansen herself calls vernacular modernism “a heuristic and analytic framework for comparing individual films and tracing particular currents in film culture.”81 She argues that even if classical Hollywood, because of its actual historical global hegemony, must be “the ground against which the stylistic accomplishments of indigenous filmmakers can be analyzed,” that still “does not make them simply variants of a dominant style.”82 Instead, local cinemas can take “different positions vis-à-vis Hollywood hegemony,” so that “the question is how filmmakers have appropriated Hollywood (along with other foreign cinemas as well as their own cultural pasts) in creative, eclectic, and revisionist ways to forge aesthetic idioms, and to respond to social conflicts and political pressures, closer to home.”83 Vernacular modernism becomes, therefore, a “comparative lens” rather than a homogenizing singular style or mode of filmmaking that spreads across the world.84 In other words, vernacular modernism is more of a research agenda than a unitary concept, in that local articulations of cinema as a global vernacular will always and crucially be inflected by their historically and culturally specific contexts.

This view, however, still fails to resolve, at least in the case of left-wing Shanghai cinema of the 1930s, the extent to which vernacular modernism is necessarily prescribed by the “classical Hollywood” mode (as defined by the work of Bordwell, Staiger, and Thompson, cited by Hansen) through which it was originally delineated as a concept. Given the neglect of defining traits like character consistency, diegetic unity, and closure in the films examined in this chapter, it seems that, despite all their rampant borrowing from Hollywood and other Western sources, they cannot be described as being fundamentally or unproblematically within the classical narrative mode. Instead, their practice of vernacular modernism borrowed as well from global, often leftist avant-garde practices and included fissures and irresolutions that fundamentally undermined the unity of classical Hollywood narration in an effort to subvert Hollywood itself, challenge its hold on the imagination of China’s “petty urbanites,” and spur the latter to reimagine themselves as potential agents in a collective revolutionary movement rather than as aspirants to the romantic fantasies of individual (or coupled) happiness propagated by Hollywood.