The Hollywood Dance-In: Abstract and Material Relations of Corporeal Reproduction

Abstract

This essay asks what the figure of the Hollywood dance-in—a dancer who performed in place of a star prior to filming and who assisted the choreographer in the creation of dance numbers—can reveal about the reproduction of corporeality as an operation that is both abstract and material. Focusing on the white film star Gene Kelly and his Mexican-born dance-in Alex Romero, the essay shows how the men functioned as literal and virtual doubles for one another in the rehearsal process and argues for an understanding of their relations of reproduction as queer and racially charged.

1. Introduction

A 1953 article in Newsweek on the making of the 1953 film musical The Band Wagon, directed by Vincente Minnelli, choreographed by Michael Kidd, and starring Fred Astaire and Cyd Charisse, gives readers a glimpse of “the literally exhausting work [that] goes into the shaping of a dance production” (Finklea & Austerlitz 1953, pp. 49–50). Citing Charisse’s twelve-hour days and the group effort required to convert ideas into concrete dance steps, the article also points to the “well nigh indispensable help” that stars receive from their “dance-ins.” These supporting figures, the article explains, “are the dancers’ counterpart to the movie actor’s ‘stand-ins’—people who accommodate the director, cameraman, costume and scenic designers, etc., when the actors themselves are not on the set.” Like stand-ins, dance-ins were “Hollywood unknowns” (Slide 2012), part of a largely anonymous and uncredited labor force who were nonetheless critical to the film industry. But the “inside help” provided by dance-ins surpassed the work of stand-ins:

For while the stand-ins are not expected to act, but merely to correspond to the physical presence of those they represent, the dance-ins are supposed to help the dance director formulate his ideas. While the stars are off the set, the dance-ins work out parts of routines which they will later demonstrate to the leading dancers, thus saving the latter a good deal of time and energy.

(Finklea & Austerlitz 1953, p. 50)

Dance-ins thus served multiple roles simultaneously: they filled in for absent stars and functioned as choreographers’ assistants, “work[ing] out parts of routines” and helping to teach those routines to the stars. As such, the bodies of dance-ins not only accommodated the filmic apparatus; they also mediated the conversion of choreographic ideas into concrete steps, as well as the acquisition of choreography by film stars.1

This essay makes the dance-in work double time in another sense, for I treat the dance-in as both a material and a conceptual figure. Materially, the dance-in is important as a ubiquitous flesh-and-bone presence in rehearsal on the sets of film musicals, a skilled mover whose behind-the-scenes labor was essential to the realization of dance on screen.2 At the same time, because dance-ins hold such a unique intermediary position—they serve as the connective tissue between idea and instantiation, choreographer and star, live performance and screen image—they also perform important conceptual work. To be sure, the theoretical implications of the dance-in are manifold, touching on issues from labor to surrogation to inter-mediality. My aim here, though, is to ask how the dance-in can advance our thinking about the often invisible relations that govern the reproduction of corporeality.

On the most fundamental level, dance-ins force us to attend to the relationship between the bodies we see on screen and the bodies that help shape those screen images. As stars’ counterparts, they call attention to the “correspondences” between bodies and raise questions about how physicality, especially as a site where race, gender, and sexuality intersect, is and is not shared across bodies. As choreographers’ assistants, dance-ins highlight the creation of choreography itself as a relational, reproductive operation. And as bodies that occupy the space between choreographers and stars, they suggest that the transmission of movement may be something other than a “straight” affair. My contention in this essay, then, is that dance-ins prompt us on the one hand to reckon with the fleshy materialities that lay obscured behind virtual projections of bodies on screen and, on the other, to project outward from the materiality of the body and see it as an index of abstract relational forces.3

In a rare exception to the dance-in’s customary anonymity, the Newsweek article cited above identifies Pat Denise and Alex Romero as Charisse’s and Astaire’s respective dance-ins. Denise (1922–2018) was a white, Canadian-born ballet-trained dancer who danced with the Ballet Russes, appeared in the choruses of a number of Hollywood films, and assisted choreographers Hermes Pan, Jack Cole, Eugene Loring, and Jack Baker (Billman 1997, pp. 294–96). Alex Romero (1913–2007) was a Mexican-born dancer, proficient in Spanish dance, ballet, tap, and jazz, who served as a contract dancer for MGM before assisting choreographers such as Jack Cole, Robert Alton, Hermes Pan, Michael Kidd, and Gene Kelly. He went on to become a full-fledged choreographer for stars like Elvis Presley and is perhaps best known, as the title of Mark Knowles’s biography submits, as The Man Who Made the Jailhouse Rock (Knowles 2013).

I have analyzed the dance-in through the lens of white women’s corporeality elsewhere (Kraut 2019). In this essay, I continue to interrogate the gendered and racialized relations between dance-in and star, but this time with a focus on the Mexican/American Alex Romero and his work in white Hollywood. Knowles’s book, based on extensive interviews with Romero and many of his collaborators, is an invaluable source on Romero’s life and career, and I rely heavily on it, as well as on other published interviews and various archival and photographic sources. My emphasis, though, is not on Romero’s accomplishments as a choreographer but on the significance of his labor as a dance-in and assistant for the dancer/choreographer/director Gene Kelly. According to Knowles, Kelly and Romero first met when Kelly was shooting the 1944 non-musical film Christmas Holiday (Knowles 2013, p. 171, n. 5) and first worked together on the 1948 musical Words and Music, on which Romero was hired as a staff assistant to dance director Robert Alton (Knowles 2013, p. 50). Romero went on to assist Kelly in films like On the Town (1949) and An American in Paris (1951), both of which appear on the American Film Institute’s list of the twenty-five “Greatest Movie Musicals of All-Time” (Dirks n.d.).4 Together and separately, Kelly and Romero contributed to the popularization of a distinctive but eclectic style of jazz dance on film, one that borrowed from and combined the Africanist rhythms of tap dance, the Latin-, Africanist-, and South Asian-inflected technique developed by Jack Cole, and the European aesthetics of ballet.

As two men who shared certain physical characteristics but not others, who copied, exchanged places with, and assisted each other but within an uneven field of power, Romero and Kelly are a particularly useful pair for exploring the reproductive relations of racialized masculinity as it was corporealized in mid-twentieth century jazz dance and captured in mid-twentieth century film musicals. Making sense of those relations requires drawing together dance studies, performance studies, and film studies with queer theory and critical race theory. In what follows, I first approach the dance-in as a double that operates across material and virtual registers and then theorize its function as a mode of corporeal reproduction. Turning to the homosocial aspects of Romero’s and Kelly’s relationship, I consider the implications of being “caught” in the act of destabilizing the original/copy binary and speculate about the potentially queer structures of identification and desire between the men. Finally, I ask how Romero’s racially ambiguous status bears on the “relational schema” (Muñoz 2013, p. 113) that existed between white star and brown dance-in.

2. The Dance-In as Material and Virtual Double

One way of thinking about the relationality of the dance-in is in terms of a double. The figure of the double, of course, is a recurrent trope across disciplinary formations and has been theorized by scholars in a variety of fields. Given the dance-in’s liminal position between film and live performance, it is useful to start by comparing discussions of the double in the fields of film and dance.

In a 2000 article titled “Missing Persons and Bodies of Evidence,” Ann Chisholm historicizes and theorizes the paradoxical functions of body doubles in the US film industry. As Chisholm shows, the practice of body doubling, in which “bodies or body parts are filmed for the purpose of replacing the bodies or body parts of motion picture actors,” sometimes known as photodoubling, became commonplace in the late 1960s but can be traced back to the 1920s and the emergence of the stand-in, which was part of the film “industry’s move toward efficiency and replication” (Chisholm 2000, pp. 123, 127). Increasingly expected to look and act like the stars they were replacing, stand-ins developed long working relationships with their more famous counterparts, becoming their “virtual doubles.” Yet despite the crucial function they played in the production of the film star, doubles “inherited the negative valences associated with being a second-rate physical copy” and were often concealed or disavowed by industry publicity (Chisholm 2000, pp. 127–28). As a result, Chisholm, invoking Jacques Derrida, argues that doubles were “positioned as a dangerous supplement within the economies of cinematic production and representation.” These supplements were also predicated on absence, insofar as either the star’s body or the double’s body had to disappear for the star’s image to assume coherence on screen (Chisholm 2000, pp. 125, 142–43).

Supplements and virtual doubles are arguably just as vital to the production of dancing bodies. In a 2006 essay entitled “Paradoxical Body” (originally published in Portugese in 2001), the philosopher José Gil goes so far as to assert that “To dance is to produce dancing doubles” (Gil 2006, p. 25). As dancers sense their own movement, Gil argues, they form “virtual images … according to the map [they have] created from the choreography.” For Gil, who is not writing about dance in relationship to film, the virtual is the conceptual realm in which dancers construct images of themselves in/as space. The effect of these conceptual images is a multiplication of bodies, regardless of how many dancers are materially present:

Paradoxically, the narcissistic position of the dancer does not demand an “I.” Rather, it demands (at least) one other body that can detach itself from the visible body and dance with it. Thanks to the space of the body, the dancer, while dancing, creates virtual doubles or multiples of his or her body who guarantee a stable point of view over movement ….

(Gil 2006, p. 24)

In the case of duos, the physical second dancer (what Gil calls the “actual body”) “realizes the virtual double of the dancer” (Gil 2006, p. 25). Even in the absence of a screen, the movement of dance depends on and produces an interplay between virtual and physical bodies.5

There are echoes here of the way Susan Leigh Foster has theorized the construction of the dancing body via various training practices. In her 1997 essay “Dancing Bodies,” a foundational text in the field of dance studies, Foster writes that “Training … creates two bodies: one, perceived and tangible; the other aesthetically ideal” (Foster 1997, p. 237). For Foster, the perceived body is the one derived from the “sensory information” a dancer takes in—visual, aural, haptic, olfactory, and perhaps most important, kinaesthetic”—while the ideal body combines aspects of the dancer’s perceived body with “fantasized visual or kinesthetic images of a body, images of other dancers’ bodies, and cinematic or video images of dancing bodies.” Foster also points to a “third kind of body,” which she terms the “demonstrative body;” this third body “mediates the acquisition of…skills by exemplifying correct or incorrect movement” (Foster 1997, pp. 237–38). Taking both virtual and material form, the demonstrative body “displays itself in the body of the teacher, and sometimes in one’s own image in the mirror and in the bodies of other students in the class and their mirror images” (Foster 1997, p. 238).

Existing at the intersection of cinematic and performance practices, the dance-in shares features with both the missing bodies theorized by Chisholm and the virtual doubles and supplemental bodies theorized by Gil and Foster. But because dance-ins, as the Newsweek article establishes, also double as choreographers’ assistants, and because they blur the lines between the construction of filmic images and the construction of corporealities, the full extent of their doubling exceeds any one of these theoretical models.

In a 2017 online post on Facebook, Gene Kelly’s widow Patricia Ward Kelly framed the dance-in in labor-saving terms. “You always have a dance-in,” she recalls Kelly saying. “For instance, if you want to set it up for the camera you have a dance-in go in and hit the marks and dance for you while they’re lighting. Otherwise you were pooped out before you went in and did the take. Then you rehearse a couple times. But they rehearse it for you. They hit the marks.” The memory accompanies a photo of Kelly rehearsing for That’s Entertainment Part II (1976), flanked by Alex Romero and a white woman named Robin Hoctor (Kelly 2017).6 The dance-in depicted here aligns with Chisholm’s description of the stand-in as a figure of efficiency and “an ancillary physical presence that filled the space left absent by the star prior to filming” (Chisholm 2000, p. 126). A substitute body that can help the filmic apparatus anticipate the choreographic marks it needs to hit to create its projectable images, the dance-in rehearses in place of the star, who is spared from “pooping out” too soon.

But in a 1975 interview with Marilyn Hunt, housed at the New York Public Library for the Performing Arts, Kelly made clear that his reliance on Romero extended beyond these parameters. Describing the rehearsal process for the ten-minute “Slaughter on Tenth Avenue,” the slinky, jazz-ballet number featuring Kelly and Vera-Ellen in Words and Music, Kelly explained that Romero “was very strong and my size. And we both had enough muscles to lift the girl all day and to try these various things” (Kelly 1975). Though Kelly frequently used Carol Haney and Jeannie Coyne as assistants, because he “couldn’t have girls doing lifts and things like that,” and because his frequent collaborator Stanley Donen “wasn’t that complete of a dancer,” Romero became indispensable to him.

The gender dynamics of Kelly’s and Romero’s relationship merit their own discussion, and I will take them up below. For now, it is worth dwelling on the functions and effects of the physical resemblance between Romero and Kelly. Kelly elaborates on their rehearsal process:

We’d stand, we’d look at each other and try these different lifts …. At times like that, you want to see how a lift will look. If he does something whether accidentally or tries something, or I’ll say, “Try another twist. We know we have her up there.” And he does something, I’ll see it. Even if he doesn’t quite know what he did because so much of that, as you know, is trial and error, when you get a variation on a lift.

(Kelly 1975)

Here, Romero, who was officially employed as an assistant dance director to Robert Alton on Words and Music, does much more than hit Kelly’s predetermined marks and much more than save the star from fatigue; he engages in a steady back-and-forth exchange with Kelly, who juggles his roles as dancer in and choreographer of “Slaughter,” to help determine what and where those marks should be. Unlike a stand-in, Romero does this not by temporarily becoming the object of the camera’s gaze (helping the camera choreograph moves before filming commences) but by inhabiting Kelly’s physicality and projecting it back to Kelly. Lifting Vera-Ellen in Kelly’s place, Romero helps Kelly visualize his own body in the partnering choreography and plot the next moves accordingly. In these moments, Romero is both the object of Kelly’s gaze and a reflection of Kelly’s corporeality. Or, considering Kelly’s comment that the two matched each other in strength and size and that “we’d look at each other,” it is more accurate to say that the two men served as mirrors for one another as they tested out different movement possibilities with Vera-Ellen.

Photos of Romero, Kelly, and Vera-Ellen rehearsing the “A Day in New York” number from On the Town, taken by Jon Brenneis, provide a visual illustration of this testing and mirroring. While lighter in tone and more balletic than “Slaughter,” “A Day in New York” also features several lifts, and both dances contain a section where Kelly and Vera-Ellen make use of a barre. (Unlike “Slaughter,” Romero appears in parts of “A Day in New York,” where he partners Carol Haney.) Figure 1 shows Romero and Kelly helping Vera-Ellen position her body on top of a ballet barre, finding the proper “mark.” Figure 2 shows Romero and Vera-Ellen positioned below and above the barre respectively, while Kelly looks on, taking in the choreography. Here Romero tries out a mark that Kelly will later hit. Figure 3 is a screenshot of the film version of the “A Day in New York” number, showing that Kelly has taken Romero’s place underneath the barre.

These images make clear how Romero functioned as the physical incarnation of the virtual body theorized by Gil: Romero is the “one other body that can detach itself from the visible body and dance with it,” making it possible for Kelly (and Romero) to “contemplate…himself from the place of the other” (Gil 2006, pp. 24–25). Romero also oscillates between, or combines, all three of the bodies that Foster identifies as crucial to the construction of a dancing body: the perceived, the ideal, and the demonstrative. Romero, that is, not only generates tangible, sensory information; he is also the projection of an image that Kelly seeks to embody, and, at the same time, he mediates Kelly’s assimilation of choreography by pre-figuring it materially and virtually.

3. The Dance-In as Queer Mode of Reproduction

As much as Romero facilitated the conversion of physical materiality into filmic image, there are ways in which he also operated like the cinematic apparatus itself. Asked what a choreographer’s assistant does in an interview with Rose Eichenbaum, published in her 2004 book Masters of Movement, Romero explained:

The way it normally works is you stand behind the choreographer, and every time he makes a move, you learn it. You have to be plenty sharp at picking things up because he doesn’t know what he’s going to leave in or take out. He throws out so much stuff that he doesn’t always remember what he’s done. He’ll come back to you and say, “What did I just do?” And you have to repeat whatever he demonstrated.

(Eichenbaum 2004, p. 166)

In response, Eichenbaum comments, “You’re like a human video recorder,” to which Romero gives a simple “Yes.” In this telling, Romero casts himself as both the choreographer’s shadow and as much more than shadow.7 As an assistant, he stands behind the choreographer’s body while standing in for the choreographer’s bodily memory. Required to pick up new choreography nearly instantaneously, he mirrors back whatever movement has just been demonstrated.

The analogizing of Romero to a video recorder not only adds to the list of the kind of doubling the dance-in/assistant performs; it also shifts the focus from the dance-in as double to the dance-in as figure of reproduction. In functioning “like” a camera, Romero illustrates dance scholar Harmony Bench’s claim that “bodies are not antithetical to machines, but … are themselves already machines and come into relationships with technologies as machines” (Bench 2004). Yet Romero’s “likeness” rather than equivalence to a video recorder acknowledges a gap between technological and body-to-body modes of reproduction. Indeed, the slight divergences in Romero’s and Kelly’s physicalities in Figure 2 and Figure 3 above—note the men’s different hand positions and the different angles of their heads—are evidence of the non-exact nature of human mimesis. And, as should be clear by now, the dance-in does not just replicate movement (if never identically) but also helps materialize the choreography through “trial-and-error” embodiment. This raises the question: what kind of reproduction produces both originals and copies? In this section, I explore the possibility that there is something queer about the reproductive relations that the dance-in ushers into existence. To do so, we need attend to the anxieties and the desires that accompany the material exchanges between Kelly and Romero.

Elsewhere in their interview, Eichenbaum asks Romero if “choreographers ever pick[ed] your brain for dance moves,” to which Romero responds, “Oh, yes.” Explaining that he “would make suggestions, but only when…asked to,” he offers the following story:

One day I’m in a rehearsal room with Gene Kelly, and he’s doing this number, you see, when suddenly he gets stuck. “Alex,” he says, “what can I do next?” I think to myself, “Well, he never went in that direction and maybe we could add a turn over here,” and suddenly this thing is coming together. He says, “That’s great. Do a little more.” So I choreograph at least a quarter of the number, and he’s applauding me, when suddenly the door opens and in walks the producer, Arthur Freed. Gene pushes me aside and says, “Cool it, Alex.”

For clarification, Eichenbaum asks if Romero means “that Gene Kelly didn’t want anyone to see that you were helping him.” “He didn’t want anyone to know that he might need help,” Romero replies (Eichenbaum 2004, p. 166).

Romero’s assertion that he was responsible for “at least a quarter” of the choreography for an unnamed movie number that was attributed to Kelly is of course significant. But I am not interested here in ascertaining which number, or in re-assigning credit for it. Instead, I want to home in on the moves Kelly makes when producer Freed enters the rehearsal space, moves that seek to conceal the nature of Kelly and Romero’s relationship. Like the industry publicity releases that studios issued to “shroud” the use of body doubles in film (Chisholm 2000, p. 128), Kelly’s push and instruction to “cool it” betray a nervousness that deserves closer scrutiny.

Performance studies scholar Rebecca Schneider’s 2001 article “Hello Dolly Well Hello Dolly: The Double and its Theatre” is helpful toward this end. Schneider connects the anxiety around the 1997 cloning of Dolly the sheep to “a rather ancient Western cultural distrust of mimesis,” which she explains as:

a fear of indiscreet origins …. The fear is that the copy will not only tamper with the original, but will author the original—or, perhaps most fearful, that the copy (the rib, the second) will come to be acknowledged as author, father, First.

(Schneider 2001, p. 96, emphasis in original)

Gene Kelly’s attempt to disguise his reliance on Romero seems driven by just this fear. To expose Romero, the assistant and dance-in, as the creator of a chunk of film choreography might threaten Kelly’s status as “author, father, First.” “You have to understand,” Romero tells Eischenbaum, “that he had a huge reputation to maintain” (Eichenbaum 2004, p. 166).8

Schneider’s terms (indebted to the insights of Freudian psychoanalysis and Derridian deconstruction) are a nod to the gendered subtext to threats to originality—or, rather, gendered subtexts, plural. First, the inversion of the original/copy hierarchy unsettles the patriarchal order and the “proper” lines of heteronormative, genealogical transmission. Second, Schneider contends, the reproductive practices of cloning and mimesis have differently gendered associations: while “‘cloning’ has long been synonymous with high tech, and high tech has signified masculinity,” the “no tech” practice of mimesis is “primitivized, feminized, and debased” (Schneider 2001, p. 98).

Returning to the comparison between Romero and a video recorder, the gendering of his reproductive function seems anything but stable. Machine-like, Romero blurs the lines between masculinized technology and feminized “no tech,” just as his alternation between replicating and generating choreography blurs the lines between the masculinized author and the feminized copy. The dance-in, like the clone, thus disrupts “the heteronormative tenet that femininity and masculinity are fully distinct from each other” (Schneider 2001, p. 98). Like cloning, too, the dance-in represents a mode of reproduction “unmoored from mechanics of sperm and egg,” just as it represents a mode of transmission unmoored from patrilineal genealogy (Schneider 2001, p. 98). Accordingly, the reproductive relations that the dance-in instantiates might also be said to pose a threat to the “heterosexual prerogative” (Schneider 2001, p. 101).

Under this light, the queer dimensions of the relationship between dance-in and star become more salient. This, of course, is not the same as saying that Kelly or Romero were queer (both were married to women, although that precludes nothing). It is, however, to ask how we might apply dance scholar Clare Croft’s understanding of queer as a force that disrupts the gender binary to the reproductive relations between Romero and Kelly (Croft 2017, pp. 2, 6). Kelly, as is well known, made it his mission to promote dancing as a masculine activity.9 But popular culture is more than capable of upholding heteronormative texts and queer subtexts simultaneously.10 Steven Cohan, moreover, has argued that “what Kelly signified as a male dancer during his years at MGM was more complicated, less stable and coherent” than the image of male virility he worked so hard to project (Cohan 2005, p. 151). Especially given the “queer reputation” of producer Arthur Freed’s MGM unit—it was nicknamed “Freed’s fairies” (Cohan 2005, p. 47)—Kelly’s pushing aside of Romero and his directive to “cool it” are not hard to read as attempts to cover over something queer. At the very least, that the public airing of Kelly’s private relationship with Romero imperiled Kelly’s “huge reputation” indicates the extent to which Romero’s mode of reproduction both buttressed and destabilized Kelly’s masculinity. Just as Cohan proposes that the Freed Unit was “a liminal space…where, regardless of the sexual partnering, queer and straight outlooks intermingled, influencing each other’s work” (Cohan 2005, p. 48), the dance-in’s liminal position created the conditions for heteronormative and queer modes of relationality to coexist.

4. The Dance-In as Homosocial Buddy

While most queer readings of musicals are focused on reception practices and queer identifications with performances that cut against the grain of heterosexual plot lines,11 the work of the dance-in shifts the focus to queer circuits of exchange that are not visible to spectators. Still, scholars’ insights about on-stage and on-screen performances can help illuminate those that take place behind the scenes. Film scholar Steven Cohan’s and theatre scholar Stacy Wolf’s analyses of “buddy pairings” in musicals are particularly instructive in this regard. In his reading of Kelly’s dancing as a “camp performance of masculinity which dialectically crosses the categories of gender and sexuality,” Cohan (2005, p. 166) notes the frequency with which Kelly is paired with an effeminate sidekick in his film musicals. Citing Gerald Mast’s comment that “dancing is something Kelly does with his buddies” (quoted in Cohan 2005, p. 165), Cohan shows that, in films like Anchors Aweigh (1945) and Singin’ in the Rain (1952), a “tacit homoerotic admiration underlies the buddy bond.” Even when “the narrative closes with the successful formation of a heterosexual couple,” the close bond between Kelly and his buddy, especially in dance numbers, remains a source of “erotic affect” that produces a lingering “disturbance to heteronormality” (Cohan 2005, pp. 166, 189). Stacy Wolf, meanwhile, observes that “[w]hile the heterosexual couple’s meeting, greeting, and mating provide the raison d’être and architecture of midcentury Broadway musicals, virtually all of them also feature a duet between women.” More than diversions, these duets are “representation[s] of intense homosociality” that, however subtly, undermine the musical’s heterosexual union (Wolf 2006, pp. 352, 354).

How, then, might we construe the homosocial relations between Romero and Kelly? Paired together in the rehearsal studio, serving as each other’s virtual and material double, matching in size and strength but asymmetrical in terms of “reputation” and power, the two men danced a peculiar kind of “queer collaborative duet” (Wolf 2006, p. 354). Like the duets that Cohan and Wolf describe, Romero’s and Kelly’s collaborations were often in service of heterosexual partnerings, as the cases of Words and Music and On the Town attest. But the representational ends do not cancel out the significance of the reproductive means. And the presence of a woman like Vera-Ellen arguably only increases the possibility of an erotic charge between the two men, as Eve Kosofsky Sedgwick’s influential argument about “male homosocial desire within the structural context of triangular, heterosexual desire” demonstrates (Sedgwick 2016, p. 16). To be sure, the assertion Cohan makes about Singin’ in the Rain—that “the romantic duo … is not achieved by heterosexually combining man and woman but by subtracting the extra male from the initial equation” (Cohan 2005, p. 188)—could apply equally to the “Slaughter” and “A Day in New York” choreography, as described above. The difference is that the dance-in’s subtraction took place prior to filming rather than within the diegesis.

What I have been trying to suggest, however, is that the dance-in/assistant and star/choreographer create a complicated material and virtual circuit of homosocial exchange, even in the absence of a third dancer, and that we cannot ignore the possibility of queer desire within this circuit. As Susan Foster notes in her account of dance training practices, “when I look at another student in class, I see her or his body not as that of a friend or an acquaintance, but as the bodily instantiation of desired or undesired, correct or incorrect, values” (Foster 1997, p. 238). To the extent that the demonstrative body is a figure with and through whom a dancer fashions their self, it exemplifies what queer theorists like Sedgwick have called the “slippery relation…between desire and identification” (Sedgwick 2016, p. 24). Diana Fuss’s discussion of the “vampiric structure of the look in women’s fashion photography” (Fuss 1992, p. 729) provides another instance of this slipperiness. As she writes, the ways that fashion magazines position female spectators as consumers of their own images “demands both separation and identification, both a having and a becoming—indeed, a having through a becoming.” She explains:

Vampirism works more like an inverted form of identification—identification pulled inside out—where the subject, in the act of interiorizing the other, simultaneously reproduces itself externally in that other. Vampirism is both other-incorporating and self-reproducing; it delimits a more ambiguous space where desire and identification appear less opposed than coterminous, where the desire to be the other (identification) draws its very sustenance from the desire to have the other.12

(Fuss 1992, p. 730)

Much of this could also describe the relational structure between Romero and Kelly, who alternatingly observe and incorporate one another, who reproduce themselves in the image of one another, and who, in so doing, (re)produce themselves. Recalling that one of Romero’s fundamental roles was to have his “brain picked” by choreographers, the metaphor of vampirism to describe his queer relationality with Kelly seems remarkably apropos.

5. The Dance-In as Dark Copy

In a 2005 follow-up to her examination of originality, doubling, and repetition, Rebecca Schneider quotes artist Antoinette LaFarge’s statement that, with the primacy of the copy, art has become “virtually identical with its dark twin, forgery” (qtd in Schneider 2005, p. 26). In response, Schneider asks, “why is the twin marked as dark? Why is it color that draws a line between one twin and its criminalized other?” Her point is that within Western constructions of the original/copy binary, the copy’s charge is not only feminized but also racialized.13 In the case of the Mexican-born Romero and the Irish American Kelly, this racial charge must be considered in more than metaphorical terms.

Knowles’s biography of Romero does not directly address the significance of race to Romero’s career, but it does tell us that “Alex was … proud to be Mexican, and his Latino heritage defined a large part of who he was” (Knowles 2013, pp. 10–11). It also provides details that help situate Romero in the US racial landscape at mid-century. Born Alejandro Bernardo Quiroga in Monterrey, Mexico to a politically elite family, Romero emigrated to the US after his father and a number of his brothers were killed during the Mexican Revolution. In the 1920s, he settled with his mother and remaining siblings in a “poor Mexican neighborhood” in Los Angeles, where his older brother Carlos had come to pursue film work (Eichenbaum 2004, pp. 164–65; Knowles 2013, pp. 9–15).14 As a teenager, Alex accompanied his surviving brothers as they toured vaudeville in a Spanish dance act. Having changed their name to “The Romeros” to avoid mis-pronunciations of Quiroga, the troupe was a competitor to the Cansino family, the famous Spanish dance ensemble headed by Rita Hayworth’s uncle (Knowles 2013, p. 16; Clark 1962).

Romero appears to have lived some of the contradictions that Anthony Macías (2008) describes in his book Mexican American Mojo: Popular Music, Dance, and Urban Culture in Los Angeles, 1935–1968. This period, Macías tells us, was “one of emerging middle-class formations and burgeoning civil rights expectations” that was nonetheless “beset by police brutality, vigilante violence, racial discrimination, underfunded schools, Anglo stereotyping,” and other forces of urban racism (Macías 2008, p. 11). In school, where ‘[t]he American kids hated the Mexicans. I mean, hated them” (Knowles 2013, p. 15), Romero’s teacher Anglicized his name from Alejandro to Alex. This did not prevent him from being teased by his white peers, although his darker-skinned brother Mario, who “looked the most Mexican,” bore the “brunt of prejudice.” If, relative to Mario, Romero “looked ‘like an Anglo,’” his light skin no doubt aided his ability to break into Hollywood (Knowles 2013, pp. 168, n.9; 173, n.14).

Like the Mexican American musicians about whom Macías writes, Romero absorbed Euro American as well as African American cultural influences, even while preserving elements of his Mexican heritage. As Romero told Eichenbaum, while touring vaudeville, he was so taken by the hoofing of the African American trio King, King, and King that he began teaching himself tap and in turn taught his brothers, who incorporated it into their act (Eichenbaum 2004, pp. 165–66; Clark 1962). And as a member of Jack Cole’s troupe of studio dancers on contract at Columbia Pictures, Romero absorbed all kinds of racial influences—African American, Afro Cuban, South Asian, European.15 Romero’s dance fluency and relatively light skin allowed him to “pass” in all-white (or seemingly all-white) dance ensembles.16 But he was not always granted the privileges of whiteness outside the studio lot. Knowles’s biography describes an incident in which Romero was stopped by two police officers as he was on his way to deposit a paycheck during a rehearsal break. Admittedly, as Romero told Knowles, he was busy practicing a Jack Cole routine and oblivious to his surroundings. Still, the fact that the police asked Romero for identification and questioned the legitimacy of his paycheck suggests that he did not enjoy unfettered mobility as he moved through Los Angeles (Knowles 2013, pp. 95–96).17

In his performance history of Latina/o performers on stage and screen in the twentieth century, Brian Herrera shows that Latinx performers’ “ethnic or racial legibility…varied widely depending on the audience, the role, or both” (Herrera 2015, p. 16). If, on the whole, Romero’s on-screen performances as a chorus dancer were racially “unmarked,”18 did this also extend to his off-screen work? Kelly, recall, described Romero as matching him in size and strength. But in photographs of the two together, like Figure 4, in which the two rehearse another segment of “A Day in New York” with Carol Haney, Romero’s skin appears slightly darker than the Irish American’s. When Romero and Kelly reproduced themselves in the image of one another, were they projecting whiteness or brownness, or something in excess of either?

Scholars like Carol Clover and Pamela Krayenbuhl have written persuasively about how Kelly’s staging of masculinity was mediated by blackness, typically absented on screen but invoked via his dance vocabulary and movement style. Krayenbuhl coins the term “blackbodying” to describe Kelly’s “post-blackface” mode of adopting an Africanist bodily comportment, which injected his dances with a “racial indeterminacy” that served to “hedge against [their] gender ambiguity” (Krayenbuhl 2017, pp. 124, 126, 296). Kelly’s work with Romero raises the possibility that Kelly was also engaged in “brownbodying,” refracting his filmic image through the body of the “not quite/not white” Romero.19 Or perhaps it is more accurate to say that Kelly’s blackbodying was at times mediated by the multiracial-bodying Romero, who had physically assimilated black, brown, white, and Asian dance styles and whose racial legibility was contingent on context. Romero’s encounter with police officers while dancing in the streets of Los Angeles, it should be noted, is uncannily similar to Kelly’s run-in with a policeman during his iconic “Singin’ in the Rain” performance, which Clover maintains “hints at some other story behind this one, some other dancer and some other policeman.” However coincidental the parallels between Romero’s off-screen and Kelly’s on-screen confrontations, the “racial resonances” that haunt Kelly’s dancing must include brownness as well as blackness (Clover 1995, p. 738).

We should take care, however, not to collapse the distinctions between Kelly’s appropriation of blackness and his inter-corporeal exchanges with Romero. The labor-saving nature of Romero’s work as a dance-in for Kelly and Kelly’s reliance on Romero to co-create choreography while preserving all of the credit for himself do reproduce patterns of white exploitation, “love and theft,” and “invisibilization” (Lott 1993; Gottschild 1996). And the power imbalances between Kelly, the celebrated white choreographer-director and star, and Romero, the less-known Mexican/American dance-in and assistant, are undeniable. But the complexity of the doubling and mirroring they performed for one another, their dependence on one another materially and virtually, the queerness of their reproductive relations, and the “racial fluidity” (Herrera 2015, p. 59) of Romero’s position make “appropriation” an insufficient term for characterizing their relationship. Instead, we might think of the ambiguous space of their exchange as a “rich, complicated, and sometimes troubling collaborative scene,” to borrow Jose Muñoz’s depiction of the literary relationship between the white Eve Sedgwick and the African American Gary Fisher (Muñoz 2013, p. 104). Muñoz (2013, pp. 108, 110–12) proposes the concept of “the communism of the incommensurate” as an alternative to relations of either equivalence or “pure submission and domination.” As a “relational schema,” the communism of the incommensurate allows for the “experience of being-in-common-in-difference” and of thinking of subjecthood as “imbricated in a larger circuit of belonging” (Muñoz 2013, pp. 112–13).20

Romero himself certainly grasped the incommensurability that structured his relationship with Kelly. “I understood what was at stake,” he told Eichenbaum after sharing the anecdote about how Kelly pushed him aside when Arthur Freed entered their rehearsal space. “And I didn’t care about getting any credit,” he added. “I always felt that Gene Kelly respected me and appreciated what I did for him. I loved the man” (Eichenbaum 2004, p. 166). Aware of the uneven field of power and accepting of private expressions of respect and indebtedness over public acknowledgments of credit, Romero pronounces his love for Kelly. This love—an articulation of connectedness to the man whose need for assistance helped Romero carve a space for himself in the overwhelmingly white Hollywood, and whose dancing body Romero replicated and helped shape, stood in for and stood aside for—speaks to the intimacy of their queer working relationship, a relationship where power was fluid but never equalized.21 And yet, the very fact that Romero chose to recount the story about Kelly suggests a lingering ambivalence about the incommensurability of their relations.22 Outing Kelly’s anxiety even as he declares his love for the man, Romero both undercuts and excuses the white star.

And it should not go unnoticed that once Romero found employment as a lead choreographer and dance director, he “rarely used assistants, preferring to work alone” (Knowles 2013, p. 187 n.6), with one major exception: after casting a dancer named Alex Ruiz in the 1957 Elvis Presley film Jailhouse Rock, Romero hired Ruiz as an assistant, and the two worked closely together for the next decade or so. According to Knowles, Ruiz was a “powerful, athletic dancer who understood Romero’s style immediately.” To avoid confusion on film sets, Ruiz was known as “Little Alex” to Romero’s “Big Alex” (Knowles 2013, p. 188 n.6). Larry Billman’s Film Choreographers and Dance Directors, which contains a short entry on Ruiz, reports that he was born in Arizona and raised in San Francisco, trained in tap and ballet, and danced in both Hollywood films and in the San Francisco ballet (Billman 1997, p. 476). A series of photographs in a January 1962 Dance Magazine article on Romero show Romero and Ruiz rehearsing for MGM’s The Wonderful World of the Brother Grimm (1962) with the actress Yvette Mimieux (Clark 1962; see Figure 5 and Figure 6).23 Like Romero and Kelly, Romero and Ruiz appear to be similar in stature and muscularity, and Ruiz appears to have an ever so slightly darker skin tone than Romero. Where once Kelly and Romero took turns developing choreography with Vera-Ellen, now Romero and Ruiz do the same with Mimieux. Repetition—but with a difference. Now in a position to select his own virtual and material double and to orchestrate his own “communism of the incommensurate” (Muñoz 2013, p. 112), Romero does so by generating what Deborah Paredez might call a “circuit of Latinidad” (Paredez 2009, p. 131).

Finally, then, Romero’s hiring of Ruiz reinforces the importance of tending to the dance-in’s material and conceptual registers. The co-presence of “Little” and “Big” Alexes on the film set is another example of the literal doubling, swapping, supplementing, and subtracting of bodies on which the creation of dance for film rested. It also points again to the queer mode of reproduction that the dance-in enacts: the blurring of distinctions between original and copy, the spawning of heterosexual representations through homosocial pairings and Sedgwickian triangles, the construction of a corporeal self through the desire/identification/embodiment for/with/of the other. And, in the end, Romero’s subtle act of centering brownness in the white-dominated studio spaces of Hollywood is a reminder that bodily materialities always matter to the relations of reproduction, even when those materialities are concealed behind projections of whiteness.