This article examines some of the complex interactions between salsa music and dance by focusing on physical interpretations of specific types of metric ambiguities and disruptions. It explores both the fairly frequent displacement dissonances that arise when the established clave pattern is flipped, paused, or broken and the grouping dissonances that are somewhat rare occurrences in salsa music, showing how dancers' responses to these metric disruptions depend heavily on the unique features of each musical context. Annotated videos break down salsa's fundamental dance and musical structures, encouraging readers to contemplate the artful interpretations presented by experienced dance practitioners and to engage with these interesting musical passages more intimately by trying out the dance steps for themselves.
IN HIS METICULOUS chronicle and critique of the work of salsa musicians primarily in the salsa dura (hard salsa) era of the 1970s, César Miguel Rondón (2008: 224) writes:
From the Caribbean perspective, the quality of the music did not come from sacrificing its support of dancing; rather, it grew out of that very role. This does not suggest a mediocre category of “music to dance to,” either; instead it emphasizes that the music, in addition to its own quality, innovations, and variations as music, also carries within itself the condition of the dance. In our part of the world, making music that cannot be danced to simply makes no sense. This approach to the music is quite different from the very exclusionist thinking in the United States. There the assumed conflict between quality music and dance music is so pronounced that one ends up denying the other. Here, on the contrary, both categories of music are fused, mutually supportive, and inevitably complementary. Obviously, in the New York salsa scene, the U.S. view of popular music never eradicated the Caribbean idea, which continued to guide the development of salsa, despite the “majestic” construction of North American culture. Even if they seldom used any visual enhancements, New York Latin orchestras always performed music that inevitably encouraged their audiences to dance.1
Among other things, this statement articulates well the fundamentally intertwined nature of music and dance in salsa; in short, salsa has always been and continues to be both a music that inspires dance and a dance that inspires music. However, dance scholar Priscilla Renta (2004: 140–41) notes that initial academic salsa research tended to focus on music and not on dance, reflecting the divide along disciplinary lines that Rondón observes more generally in North American culture. Indeed, over more than four decades, music scholars have had a lot to say about salsa, defining it as a unique genre; addressing its somewhat complex historical evolution and relationships with other dance-music forms; exploring how salsa has come to be embraced as part of a pan-Latinx sense of identity, in addition to being incorporated into the many personal identities of participating individuals; tracing the important contributions made by various (Latin and non-Latin) communities worldwide as they drew inspiration from and sometimes fused with other (often geographically differentiated) dance-music traditions; and scrutinizing the rhythmic/metric and melodic/harmonic features of salsa's instrumental and vocal layers within a formal structure derived primarily from Cuban predecessors.2 However, beyond acknowledging that dance is indeed an essential component of this collaborative art form, salsa music scholarship has paid remarkably little attention to who is doing what on the dance floor, frequently glossing over the important contributions dancers make (individually and collectively) and neglecting to discuss with specificity the influential role that dancers' gestures play in the creation and interpretation of the music.3
In the time since Renta made her observation, dance-focused salsa scholarship has developed along somewhat parallel lines, providing important studies of historical evolution, culture, community, politics, identity, race, and gender, not to mention the physical mechanics and motivations that help define the different geographically associated styles of salsa dance.4 While not usually employing music notation or specific music-theoretical terminology, dance-focused studies often take care to represent fundamental dance-music interactions in ways that demonstrate a deeper understanding of and engagement with musical structures than music-focused studies typically provide for dance structures. However, analytically rigorous research that integrates salsa music and dance on a detailed technical level using the discipline-specific terminology common to each field has, for the most part, yet to be done.5
There are many practical reasons that such integrated projects have not yet been tackled by salsa scholars. If we consider the nature of academia through the lens that Rondón provides, we note that dance and music (particularly as studied in the Western tradition) have evolved into highly specialized fields, resulting in practitioners, scholars, and audiences/readers who, despite possibly seeing the need, often do not have equivalent expertise in both areas to fully engage in such interdisciplinary work.6 We might simply conclude, then, that given its multifaceted nature, salsa needed to be broken down into these separate projects, with researchers each providing one piece of a particularly rich puzzle within the scope of their own investigation and expertise—and in many ways, this is a plausible hypothesis. However, to the extent that we take seriously Rondón's comments about the nature of salsa as a fundamentally integrated art form, we need to examine in some detail how it came to be explored and presented in this divided way in order to make informed decisions about how to put the puzzle back together again—to bring music and dance into the same kind of engaging dialogue (and on equal footing, so to speak) as happens between musicians and dancers in the salsa club. Indeed, this historical context influenced in significant ways the positional listening perspective I adopt in the analyses presented in this article.
Rondón's comments allude to the fact that dance music, by its very nature, has long been considered to be of lesser quality within a Western art music tradition that came to privilege absolute music as the pinnacle of artistic achievement.7 Renta suggests that salsa and other Latin dances have been especially marginalized both because of their origins in and continued practice as social dance traditions (compared with theatrical/concert dances such as ballet that have been more readily embraced as “art”) and because of European colonizers' need to suppress the bodies of those they wished to conquer and control. She writes:
The power of dance and its potential for military mobilization made the dancing body of color a particular threat for Europeans in their efforts to colonize the Caribbean and the Americas. . . . Consequently, colonialism brought with it a physical oppression that included the suppression of Afro-Latin(o/a) dance forms. Historically, Afro-Latin(o/a) dance communities have resisted this form of oppression in part by keeping their dance forms alive, a practice that persists in Latin America, the Caribbean, and the U.S. among Latinos/as. Salsa dance inherits this legacy of resistance against colonial subjugation from its Afro-Caribbean ancestors. The marginalization of Afro-Latin(o/a) dance practices and its scholarship stems from such history. (Renta 2004: 141)
While some dancers have indeed sought to “elevate” salsa dance to the status of art by bringing it out of the clubs and onto the choreographed stage (see McMains 2015: 315–17), others have endeavored to retain a strong connection to salsa's roots as an improvisation-based street dance that grows organically within communities, and still others have tried to balance the tension between these practices by claiming to do both simultaneously. However, Frances R. Aparicio (1998: 97) notes that attempts to discourage spontaneous, communal, physical responses to salsa music within the concert/stage environment are often met with resistance as well:
Despite articulatory efforts by producers and the entertainment industry to recontextualize salsa into concert music, to contain it into music to be only listened to, unidirectionally consumed as another commodity, salsa concertgoers will notice that whenever a Latina/o audience is present, there is dancing in the aisles and in any available open space within the confines of the theater or auditorium. This practice, consistently repressed by security guards for supposed safety codes, reveals, on the one hand, the profound relationship between salsa music, the cultural collectivity of dancing, and our bodies as social sites. On the other, it attests to Latinas/os' resistance to salsa's commodification through its strategic, physical insertion and cultural containment within the dominant spaces of concert halls and auditoriums.
Given the crucial importance of the dancing body itself in both past and present salsa practice, its absence—even in early salsa dance scholarship—is all the more conspicuous. By way of explanation, Renta (2004: 140–41) looks back to the religious and philosophical notions of disembodiment that began with thinkers such as Plato and Descartes, noting that in addition to contributing to the oppression of specific populations along lines of gender, race, and sexual orientation, this mind/body split rendered embodied knowledge subjective and not worthy of scholarly concern. While it seems that she is building a case for a reversal of these antiquated attitudes toward the body, Renta ends up justifying the use of lived experience (her own and that of fellow salsa dancers, performers, and instructors) as part of her research method by citing the dearth of scholarly literature on salsa dance rather than as a valid way of knowing in and of itself (142). Writing more than a decade later, dance scholar Juliet McMains (2015: 17) is unapologetic about the use of her own body in observation and analysis. In support of this scholarly practice, she writes:
Just because I used my own body as a vehicle through which to compare different dance styles does not mean that I privileged my experience over that of others. On the contrary, my physical dialogues with informants focus on understanding their embodied knowledge. I consider dances with informants to be interviews in another language, one in which my whole body listens. Such physical interviews are not any more or less biased because they are filtered through my own body than are verbal interviews, which is why neither is used in isolation.
In this more recent vein of salsa dance research, my analytic approach here is fundamentally shaped by experiential knowledge—my training as a musician, my experiences as a dancer and dance instructor, and my observations of the practices of other dancers and musicians over nearly twenty years of participation in salsa communities, primarily in Canada and the United States. I present the dance perspective as an effective lens through which to analyze and interpret music, encouraging readers to engage with the analyses both as observers and as active participants—indeed, it is not the same experience to watch someone dancing as it is to dance oneself.
In this article, I investigate some of the complex interactions that occur between salsa music and dance by focusing on dancers' physical interpretations of specific types of metric ambiguities and disruptions found in salsa music.8 The four songs I transcribe and analyze—Joe Arroyo's “La rebelión,” Víctor Manuelle's “He tratado,” El Gran Combo de Puerto Rico's “Me liberé,” and Grupo Niche's “Cali pachanguero”—were selected both because they provide a variety of interesting metric irregularities for analysis and because they are very well-known songs, frequently played in salsa clubs around the world. Through these musical examples, I illustrate both the fairly frequent displacement dissonances that arise when the established clave pattern (the “key” to the salsa groove) is flipped, broken, or paused and the grouping dissonances that are somewhat rare occurrences in salsa music, showing how the dancers' responses to these metric disruptions depend heavily on the musical context. I focus on these events because they affect dancing bodies of all experiential levels; for dancers just learning the basic salsa steps while reading this article, these events are tangible and readily apparent after only a little background information and training, yet they continue to provide interesting challenges even for advanced dancers, who engage with these disruptions in a variety of artistically innovative ways. In this study I represent salsa practice by analyzing what salsa dancers actually do with these particular musical scenarios (without any knowledge of the investigation at hand), rather than suggesting what some imagined dancer might do with such situations. Thus the videos I have created myself primarily illustrate the basic elements of the dance, whereas videos of other dancers were selected from YouTube because they represent fairly typical dance responses and/or showcase exceptionally engaging interpretations of these musical passages. Collectively, these videos depict a range of circumstances in which salsa dancing occurs—social dance floors, staged performances, studio classes, and instructional videos—but with an overall emphasis on improvised, rather than choreographed, dance.9
Of the physical configurations (solo, couple, team/group) in which salsa dance may be practiced, I am interested primarily in the improvised negotiations that occur between a “lead” and a “follow” as they navigate their own unique way through the music as a dancing couple on the social dance floor. While the role of the lead usually involves being the decision maker in the partnership and thus is often privileged in the pedagogy, performance, and promotion of salsa dance, McMains's comment above implies that it is actually the follow who is particularly well suited to making analytic observations about the range of possible dance responses, because a defining requirement of this role is the ability to “listen” carefully to each partner in order to interpret the (often subtle) physical cues that allow complex movement patterns to be executed with style and grace.10 The physical responses that I suggest in my analyses, then, are represented primarily from the follow's perspective and are modeled on my own intuitions as a dancer (who most often dances the role of follow) and on my observations of the behaviors of other skilled salsa dancers—those who have deeply internalized both dance and musical structures and thus can make informed, in-the-moment, interpretive decisions based on a wealth of bodily knowledge and experience.11
An overview of dance and musical structures
To lay the groundwork for my analyses of these four songs, some essential dance and musical structures must be outlined. In our 2019 article Chris Stover and I define some of the basic music and dance structures in salsa, providing a framework within which to make more detailed analytic observations about dance-music interactions. In this section I review some of this prior work and weave in additional details from other scholars' contributions to provide the necessary background for the analyses that follow.
Figure 1 and Video 1 show the basic salsa dance step used throughout this article, commonly referred to as the “on-1 basic” because the changes of direction (known within the dance community as “breaks”) that motivate each half of the eight-count pattern fall on the downbeat of each 4/4 measure of music (each dance count corresponds to one beat of music in this context).12 It should be noted that this is not the only version of the salsa basic in regular circulation; indeed, many versions have been practiced throughout salsa's history and continue to be in widespread use on modern dance floors.13 I employ the on-1 basic here because it is generally accepted as easier for beginning dancers to learn, allowing the reader to incorporate the salsa steps more quickly and comfortably into their body's own repertoire of movements in order to focus attention on the analytic observations that are of primary interest in this investigation.14 In particular, on-1 allows dancers to begin their movements and place their biggest gestures (the breaks) on downbeats, which are usually easier for listeners (especially those not well versed in this musical tradition) to identify within the complex musical texture.15
To return to my description of the on-1 dance basic, then, three steps are taken in each half of the eight-count pattern (on counts 1-2-3 and 5-6-7), resulting in a “quick-quick-slow” dance rhythm as motion is stretched out (or ceases entirely) on counts 4 and 8. All beats on which steps occur receive some degree of emphasis or accent; however, breaks receive particular physical accentuation, as they articulate the points of farthest remove from the dancer's center position, change the direction of the body's motion, and provide the biggest, most consistently visible gestures of the pattern (Simpson-Litke and Stover 2019: 79). Inspired by a notational system developed by Kara Yoo Leaman to facilitate choreomusical analysis of ballet and other dance forms, Figure 2 converts the footwork diagram of Figure 1 into staff notation, using colored footprints as the noteheads of quarter-note durations within measures of 4/4 while maintaining the bird's-eye perspective of the dancer's feet via notehead placement on staff lines.16 These examples highlight two types of physical accent that occur in the on-1 basic—the “slow” steps that precede pauses (shown with blue footprints) receive (interonset) durational (or D) accents, while breaking steps (shown with red footprints) are analogous to contour (or C) accents because they mark the outer spatial boundaries of the basic.
Because salsa is primarily a partner dance, it is helpful to consider the contour of the moving couple in addition to the contour of each individual's footwork pattern. Figure 3a shows the dancing couple's footwork as viewed on the dance floor from above, and Figure 3b converts these patterns into staff notation from the visual perspective of each individual dancer (again, to encourage readers to embody each role for themselves). The uppercase Cs of Figure 2 and the red footprints of Figure 3 show that backward breaks receive more physical emphasis than forward breaks because they articulate the points of farthest remove in the dancing couple's overall motion and are the most visually trackable gestures in the context of more complicated movement patterns. While the basic footwork is the foundation for all dance structures in salsa, it is usually not the aspect of the dance to which viewers are most instinctively drawn, especially those not well versed in the salsa style. Indeed, when we turn to the video analyses, the viewer may find it difficult not to be dazzled by other body movements (fast-moving spins, rippling body rolls, decorative arm flourishes, crisp head flicks, sensual hair combs, etc.); it will be important to remember that, in the midst of all this complex physical activity, the follow's backward break on the right foot is usually a clear indication of where the hypermetric downbeat falls and provides an important visual distinction between the “one” and the “five” in the dancers' eight-count pattern.
For the sake of visual simplicity, then, my analyses usually track only the follow's footwork, with the understanding that it is representative of the dancing couple.17 However, in one analytic example both dancers' footwork patterns need to be tracked independently and simultaneously, in which case, I represent the couple's footwork on two staves and from each dancer's individual perspective, as in Figure 3b. As you dance along with Video 2 (an illustration of Figure 3), note how the direction of the lead's physical motion is exactly opposite that of the follow. Once a couple has established a particular metric entrainment, they usually align their steps with the music consistently for the duration of the song; while steps are sometimes added, omitted, syncopated, and most certainly reoriented to execute more complex movement patterns, an underlying basic alternation of feet (R-L-R/L-R-L for the follow, and vice versa for the lead) is maintained throughout. However, many salsa songs feature interesting (hyper)metric conflicts, ambiguities, or changes that dancers must navigate in their physical interpretation of the music. Before investigating these events, some of salsa's musical features need to be defined.
Example 1 contains information about the essential elements of a typical salsa arrangement and is an amalgamation of Christopher Washburne's (2008) figures 6.1, 6.2, and 6.3 (with additional annotations); the summary below draws heavily from his description. In the top half of the example, we see an outline of salsa's standard formal organization, derived from the Cuban son(-montuno) form and consisting of a precomposed first section featuring the main theme (cuerpo), often employing a standard song form like AABA or verse-refrain, and an open-ended second section featuring a number of subsections, often improvisatory in nature. In the montuno subsection, a call-and-response (coro-pregón) between the singers is supported by the instrumentalists with a repetitive harmonic and rhythmic vamp. The call of the lead singer (sonero) is usually improvised—indeed, the ability to spin creative melodic lines and lyrics (las inspiraciones) on the spot has traditionally been considered the mark of a truly skilled sonero—and followed by a precomposed response (coro) from the chorus (coristas). The montuno usually returns (in full or abbreviation) a number of times in the open-ended second half of the musical form and is often interspersed with instrumental subsections like the mambo (a precomposed subsection featuring a heightened level of intensity created by virtuosic horn writing, new musical material, and harmonic shifts) and moñas (which may be precomposed or improvised by the horn players, often in the form of short riff-like vamps that are layered on top of each other to build energy and excitement). More extended instrumental solos can also occur within moñas or in the montuno itself (168–69).18 In addition, “rhythmic breaks” (henceforth referred to as cierres so as not to create confusion with dance breaks), in which all members of the rhythm section (piano, bass, and percussion) play together in rhythmic unison, are often used to demarcate the ending and beginning of the various sections within the arrangement (172; Gerard 1998: 112–13).19
The bottom half of Example 1 shows the interlocking percussion complexes that are typical of the theme and montuno sections. The dotted boxes highlight that some percussionists play rhythmic patterns that repeat every measure (maracas, güiro, bongó, timbale drum),20 while the dashed boxes show that others project the two-measure hypermetric unit (clave, timbales, congas, mambo bell, cencerro bell) meant to coincide with the eight-count dance basic. While some of these rhythmic patterns remain more or less constant throughout,21 the roles of two percussionists should be examined in more detail due to their importance in articulating the main two-part formal division. (1) In the theme, the timbales player (timbalero) plays the cáscara (a.k.a. paila) figure. During the montuno, however, this player moves to the mambo bell (playing the “mambo montuno ride pattern”) and the timbale drum (Washburne 2008: 171). (2) In the theme, the bongó player (bongocero) plays the martillo (literally, the “hammer”) figure and then shifts in the montuno to the cencerro bell (cowbell) to play the campana, although if the arrangement calls for a piano solo during this section, the bongocero will usually return to the bongós (Gerard 1998: 71). In large part, it is these instrumental changes, combined with the (re)entrance of the piano's repeated guajeo, that make the switch from the first half of the song to the second half audible, exciting, and energizing for musicians and dancers alike.22 As Washburne (2008: 171–72) observes: “The result of the two bells sounding concurrently is dynamically louder and, thus, used for the more energetic sections of the arrangement. Often a simultaneous slight hastening of the tempo—what percussionist Johnny Rodríguez calls ‘agitating the ride’—occurs with the introduction of the bell patterns, enhancing the effects of the increased energy level.” These formal sections and their unique characteristics are important to keep in mind as the following analyses consider the effects of musical energy and momentum on the dancers.23
Example 1 places the son clave rhythm at the top of each system to make the important point that “all musical and dance components in salsa performance are governed, to varying degrees, by the clave rhythm” (Washburne 2008: 178). In the words of Cuban trumpeter Jesús Alémany, “Clave is the vertebra of my musical feeling, the crucial way that the bass line, percussion and chorus, and of course the dance itself, link together” (quoted in Steward 1999: 12). In traditional salsa practice, this means four main things:
The clave rhythm is often performed on the instrument of the same name by a lead singer (or, alternatively, on the woodblock by the timbales player); however, clave may not be played in some sections or possibly at all in some songs but is understood to remain in effect nonetheless. In other words, clave is more than just another rhythmic pattern within the percussion complex; it is a timeline, a point of reference, a coordinating force that governs the behavior of the other musical elements—rhythmic, melodic, and harmonic—and is thought to continue, uninterrupted, from the first note to the last.
Clave can occur in two equally acceptable orientations: 2-3, where each hypermetric unit begins with the rhythmically straighter “2-side” measure (in which two clave strikes occur) and is followed by the more syncopated “3-side” measure (in which three clave strikes occur), as shown in Example 1; or 3-2, where the order of these two clave measures is reversed (along with the corresponding measures in the other percussionists' parts).24
To play en clave (as opposed to cruzao or “crossed”), musicians are required to keep track of which side of clave is in play at any given moment so that they structure their individual musical decisions accordingly. This becomes an especially important and challenging task for soloists who are improvising musical material on the spot (i.e., not playing one of the predetermined patterns identified in Example 1), but it can also pose challenges for salsa arrangers (as they craft the precomposed sections of the song like the theme and mambo, sometimes adapting tunes from other sources to the salsa style).25
In the context of a salsa song, melodic material played most frequently by piano, bass, and horns often features units that are multiples of two measures in length. It is the relationship between these melodic groupings and clave (and, by extension, all the other repetitive percussion patterns) that determines whether the music is in 2-3 or 3-2 orientation (or “flips” from one orientation to another in the midst of a song, usually via odd-measured phrase expansions or contractions). In my analyses I explore interesting treatments of this hypermetric arrangement—unexpected, unusual, or ambiguous interactions between clave, melody, and dance.
As the final concept needed to set up this analytic discussion, Figure 4 presents two theoretical situations in which established relationships between these three elements are disrupted; the brackets above the staves indicate how the notated clave is grouped by some imaginary melodic material. In both situations, we see that a 3-2 clave is clear for the first two measures and that a second such grouping begins in m. 3. In m. 4, however, the examples diverge in important ways. Figure 4a presents a somewhat musically scandalous situation whereby the clave pattern is broken, either by the inclusion of an “extra” measure in the phrase or by the “deletion” of an expected measure that does not materialize, resulting in an abrupt hypermetric disruption. According to salsa traditionalists, “once a song begins, the clave does not change its measure order (e.g., a 3-2-2-3 clave sequence [or a 2-3-3-2 sequence, as we have here] is rare and considered inappropriate by today's salsa performers)” (Washburne 1998: 163; 2002: 120). While this kind of situation is considered quite transgressive in some salsa circles, many music and dance scholars have noted that throughout salsa's history adherence to clave as an unbreakable rule has not always been strictly maintained at different times, in different places, and by different artists.26 Nonetheless, this situation creates a dilemma for the dancers, causing them either to find themselves suddenly dancing “on the wrong counts” if no adjustment is made or, as the suggested footwork in the example shows, to adopt the new hypermetric downbeat by abruptly changing their footwork. Because the direction of physical motion on count 1 is completely opposite that on count 5, dancers may respond negatively to the sensation of moving forward when they should be moving backward, and vice versa.
Figure 4b shows a subtler hypermetric disruption where clave remains consistent throughout but a change in melodic grouping structure in m. 4 flips the clave from 3-2 to 2-3 orientation. This type of disruption is commonplace in even the most traditional salsa music (and its Cuban predecessors) as it does not constitute a violation of clave, and indeed, some arrangers view it as a necessary occurrence to make certain tunes “fit” the clave rhythm. Quoting well-known salsa arranger Marty Sheller, Charley Gerard (1998: 35) writes:
Usually somewhere in the course of the song there is a phrase that makes the form of the clave obvious. Then you work back from there to see how the other phrases fit. And if there's a phrase that goes against the clave, then you have to make some kind of alteration such as adding or subtracting one bar here or there. It can usually be done in a subtle way so that it sounds hip. I'll try to put something in there that makes the melody fit into the clave in a musical way so that people who are dancing are not going to realize that it “turned around” there—even though a musician will realize it. It just flows along.
While I disagree with Sheller that dancers will fail to recognize that this type of event has occurred, I do agree that they may or may not respond to this event with a change of footwork. In fact, McMains (2015: 126) suggests that such footwork changes have not always been the norm—in the mambo era of the 1950s, for example, there was no consensus about how dancers needed to line up with the two-bar musical phrase—but coincided with the development of studio teaching in New York: “When Eddie Torres's teaching system began to dominate the city in the 1990s, studio-trained dancers began adhering to a strict rhythmic structure. . . . The counts were so strictly adhered to that when a song included an orphan bar such that the phrase began anew after only four counts, dancers would reset their dancing as well to maintain the sanctity of the counts.”27 My point is that in both situations shown in Figure 4, a choice must be made by the dancing couple: when confronted with the disruption, do they change established footwork patterns to align with new hypermetric downbeats, or do they tolerate the potentially uncomfortable feeling of dancing “on the wrong counts” of the music? I believe the answer to this question and the nature of the physical response depend heavily on the unique features of each musical context, so I now turn to some specific musical examples.28
Dance-music analysis no. 1: Joe Arroyo's “La rebelión”
Beloved for his dedication to raising awareness of Afro-Latinx political histories, Colombian sonero Joe Arroyo represents the “tropical” salsa style, for which he earned popular acclaim with his band La Verdad (The Truth) from 1981 until his death in 2011. Recorded in 1986 on the album Musa Original, “La rebelión” tells the story of an African couple brought in chains to Cartagena by Spanish slave traders in the 1600s and, despite being somewhat formally unusual, is one of Arroyo's most famous and best-loved songs. Example 2 provides my transcription (in collaboration with Chris Stover) of the song's opening sections.29 Eight measures of instrumental material (four measures repeated) form a kind of preintroduction that establishes a clear 2-3 clave. This is followed by a sixteen-measure instrumental introduction, with four additional measures (mm. 25–28) providing a “holding pattern” until the beginning of the first vocal verse in m. 29. The projected length of this section of the verse (what I call verse A) is another sixteen-measure unit, featuring what we might expect to be four complete iterations of a two-measure vocal call alternating with a two-measure instrumental response. However, in m. 44, a hypermetric disruption of the type shown in Figure 4a occurs—the second half of the verse (verse B) sounds like it enters one measure early, cutting the last horn response short and suddenly breaking the clave (note the two 2-side measures in immediate succession), which remains in 2-3 orientation despite the “missing” measure.30
While not yet into the improvisatory second half of the form, many montuno-like features begin in m. 44—for the ten measures of vocal material that follow (of which only the first few measures are transcribed in Example 2), the timbalero switches to the montuno ride pattern, the bongocero's cowbell is quite audible, and the piano plays a repeated guajeo, energizing this second section of the verse; the coristas then enter with repeated coro responses to horn calls (eight measures), which leads into a mambo-like section (another eight measures). However, this coordinated horn material soon turns back around to the opening introduction (m. 9), leading into a second verse that provides new lyrical content over returning musical structures, including a second clave break. In the discussion of Figure 4a above, I suggested two possible physical responses to this type of dancers' dilemma, as demonstrated simultaneously (for visual comparison) in Video 3. I encourage the reader to dance along with this video (at least) twice—once with the feet on the left (interpretation 1), which proceed with an unchanged basic throughout, and once with the feet on the right (interpretation 2), which make accommodations for the clave breaks by resetting the basic once in m. 44 (as depicted in the footwork notation of Example 2 and demonstrated at time stamp 0:58) and a second time with the return of this material (not transcribed in Example 2 but demonstrated at 2:10).31
As felt/seen in interpretation 1, if we begin dancing to this song by breaking backward on count 1 and forward on count 5 (i.e., using the follow's footwork), we suddenly find ourselves interpreting the material after m. 44 in the opposite physical orientation, breaking forward on count 1 and backward on count 5 (i.e., using the lead's footwork) for the second half of the verse and return to the opening material. As a result, a displacement dissonance is created between the hypermetric downbeats of the music and those of the basic step during that section.32 Indeed, this is likely what would happen if a dancer was experiencing the song for the first time. Dancers who are familiar with the song may choose interpretation 2, adjusting their footwork with the clave breaks so that the alignment between dance counts and musical phrase structure is maintained throughout (i.e., no direct displacement dissonance is perceived). However, given that (a) the clave break in this example is abrupt and (b) the upcoming second verse involves a second clave break and “omitted” measure, it is also possible that an experienced dancer would tolerate or perhaps even embrace the temporary dissonance created here instead of attempting a sudden and potentially awkward change of footwork.33 Indeed, this interpretation adds a layer of metric complexity to the dance-music experience that would not otherwise exist.34 I encourage the reader to dance through interpretation 1 again, noting both the feeling of aural-physical tension that begins at 0:58 and the feeling of resolution that occurs when the dance-music alignment is “righted” at 2:10.
In my hearing of the overall form of this song, the true montuno section begins after a more significant cierre and features the typical vocal call-and-response format interspersed with extended trumpet and piano solos. The song finishes with a coda that returns to the instrumental material of the preintroduction. Table 1 summarizes this large-scale formal organization and provides corresponding time stamps for Video 4, an improvised performance of the full song by dancers Andria Andreou and Alexandros Iacovides. Video 4 presents the dance at full speed with general annotations about formal sections and locations of hypermetric disruptions, while Videos 5 and 6 break down two specific passages of interest at half speed and with more detailed annotations, as described below.35
As noted in Table 1, many of the song's important formal divisions are articulated in this dance interpretation by shifts between open work (where the dancers improvise individually, focusing on more elaborate footwork and body isolations) and partner work (where the dancers hold hands and coordinate their movements via the principles of leading and following).36 In this dance interpretation (and in many others I surveyed for this song), the dancers tend to use open work for sections that are sparser in texture, where repeated melodic/harmonic vamps are absent, percussion complexes are not (yet) in full swing, and the groove is downplayed to allow specific rhythmic/melodic gestures played by individual instruments to be highlighted. In open work, the dancers do not have to coordinate their physical movements with each other, so they do not have to tie their footwork as precisely to the basic step or metric grid; this allows each dancer more freedom to accentuate individual musical gestures with different body parts.37
Let us examine in detail how these dancers navigate this song, focusing especially on the clave breaks (shown with a dashed line in Table 1) in the first half and an unexpected hypermetric dissonance created solely by the dancers in the second half. As shown at half speed in Video 5, the follow uses chest pops, hip swivels, and shoulder shimmies to highlight the clave rhythm in the instrumental preintroduction, while the lead's shoulder movements, head flicks, and footwork are often coordinated with the melodic rhythm of piano and bass. When the full texture (horn melody, piano guajeo, bass tumbao, full percussion) enters for the song's real introduction (0:21), the dancers switch to energetic partner work, aligning their basic footwork with the metric grid reinforced by the groove; as expected, the follow's backward break provides a reasonably consistent indication of where hypermetric downbeats fall (highlighted with “1” annotations in Video 5). The punchy unison horn articulations in m. 23 are then punctuated by the dancers' sharp arm movements (0:53), and they release again into open work (0:56), matching the smoothness of the vocal call with the fluidity of rippling body rolls, hip swivels, wrist curls, and spins and then providing contrast by aligning sharp arm, chest, shoulder, and hip movements, jumps, and spins with the unison rhythms of the horn responses (at 1:13, 1:22, 1:32, and 1:41). Because the dancers have, for the most part, stopped dancing to the metric grid through this entire section, the hypermetric disruption in the middle of the verse poses no aural-physical dissonance or gestural awkwardness. The dancers simply restart their basic step with a resumption of partner work when the full musical texture returns in m. 44 (1:43), seemingly unaffected by the “missing” measure and providing yet another possibility for the physical navigation of this clave break.
To return to the full performance in Video 4, when analogous musical material returns in the second verse, these dancers embody this parallelism by providing an analogous physical interpretation: partner work for the return of the introduction (1:28), freezing side by side and body rolling during the holding pattern (1:47), open work in call-and-response fashion for the first part of the verse (1:52), and then back to partner work with the second hypermetric disruption (2:10). For the big cierre prior to the entrance of the montuno (2:30), the lead wraps up the follow beside him, stopping her momentum and adding hip gyrations until the musical groove returns. The couple stays in partner work for the montuno (2:33) and trumpet solo (2:57) and then releases into open work for the start of the piano solo (3:31), which contains a second passage of hypermetric interest.
As broken down at half speed in Video 6, the lead's focus on matching fancy footwork and body isolations (including hip-hop-style popping, locking, and ticking) to piano rhythms causes him to temporarily lose track of the metric grid during this solo. As a result, he attempts to reenter partner work on the “wrong” counts at 0:42, causing his partner (who had been accurately tracking the hypermetric downbeats prior to this moment) to make a split-second correction in order to go with him. This creates a very brief hypermetric displacement dissonance between dance and music—one that does not seem to be prompted by specific musical cues (like the previous clave breaks) but that demonstrates the challenge of tracking these hypermetric units, particularly while improvising creatively (alongside a pianist who is also improvising creatively) and switching frequently between open and partner work. The lead realizes his “mistake” almost immediately and covers it by realigning with piano rhythms until he has located the music's hypermetric downbeat again at 0:57.
A final return to Video 4 shows that the couple stays in partner work for the remainder of the piano solo and return of the montuno (4:11) before releasing into open work one last time for the coda (4:28). In short, these dancers embody many musical elements simultaneously in their improvisatory interpretation of this song, from the rhythmic patterns of individual instruments that ride atop the metric grid on the small scale, to the textural changes and motivic/thematic contrasts and returns that articulate the formal structure on the large scale. Interestingly, while this couple chose to navigate around the clave breaks that could have led to a hypermetric displacement dissonance in the first half of the song, they created a new and unexpected dissonance in the second half of the song—one that could not have been anticipated by analysis of the music alone.
Dance-music analysis no. 2: Víctor Manuelle's “He tratado”
Produced by Sergio George and released on Nuyorican sonero Víctor Manuelle's album A Pesar de Todo in 1997, “He tratado” is representative of the smooth salsa romántica style and provides another example of a broken clave but with some important distinctions. The refrain transcribed in Example 3 is repeated twice in the precomposed first half of the song. In this excerpt, two-measure vocal groupings are punctuated by horn responses and combine with a four-measure unit to form an eight-measure phrase. The second phrase in this excerpt begins with the anacrusis to m. 9 but is quickly disrupted by a measure of rest following the piano's dramatic descending octave glissando in m. 10, creating an irregular nine-measure phrase. If we were to imagine the clave pattern continuing through m. 11 into m. 12, the same side would be heard twice in succession, as in “La rebelión.” Dancers who are inexperienced or unfamiliar with this song will frequently dance through the measure of rest and find themselves “on the wrong counts” beginning in m. 12.
Even more quickly than in the previous song, this uncomfortable feeling of hypermetric displacement dissonance is righted when an analogous measure is “added” in the following phrase. Thus, while the dancer's dilemma in this excerpt is short-lived, perhaps making a footwork change unnecessary throughout mm. 12–19, the more common and perhaps more musically sensitive interpretation of this passage is for dancers to halt their footwork in mm. 11 and 20 and to align with the new hypermetric downbeat after the disruption in each phrase. The fact that there is complete silence in all instrumental layers of the music suggests that these moments should be interpreted less as instances of broken clave and more as “paused” clave, not to be included in the hypermetric or footwork count.38 The reader is encouraged to dance along with Video 7, noting how Manuelle's live performances of this song feature several obvious visual clues (bright flashes of light, cutoff gestures with Manuelle's hands) that alert dancers to these pauses in the musical texture even if experiencing the song for the first time. Video 8 shows an improvised performance of this song in which dancers Giana Montoya and Nery Garcia put these clave pauses to expressive use with dramatic and sensual poses.
Dance-music analysis no. 3: El Gran Combo's “Me liberé”
“La rebelión” and “He tratado” provide examples where the clave pattern itself is altered but the correspondence between clave orientation and melodic groupings remains consistent throughout; in “La rebelión” the melodic groupings always align with a 2-3 clave orientation, whereas in “He tratado” the melodic groupings always align with a 3-2 clave orientation, despite the “addition” or “subtraction” of measures that disrupt the regularity of the overall hypermetric structure and formal proportions. In contrast, “Me liberé” from El Gran Combo de Puerto Rico's Grammy-winning 2001 album, Nuevo Milenio, El Mismo Sabor, provides an example of the subtler type of disruption described abstractly in Figure 4b, where the clave pattern remains constant throughout but the correspondence between clave and melodic groupings changes—again, often via the “insertion” or “deletion” of musical material. Example 4 provides my transcription of the second montuno section (following the mambo), in which three repeated subsections (labeled A, B, and C) occur. The second page of the example (mm. 9–18) is shown twice to accommodate two possible dance interpretations, with both follow and lead footwork provided throughout Example 4. One-measure cierres (mm. 9 and 18) are added to transition from the A subsection to the B subsection, and then from the C subsection back to the A subsection, flipping the clave orientation from 2-3 to 3-2 and back again. Dance interpretation no. 1 shows the footwork changes that would need to take place if dancers wanted to align their steps with the melodic groupings throughout. However, because there is neither a real pause in the musical momentum during these transitional measures (the cierres both flow from the previous subsection and seem to propel the music forward into the next subsection) nor a breaking of the clave pattern, the hypermetric disruptions in this example are less obvious, and footwork changes may actually take a little longer to execute than dance interpretation no. 1 suggests.39
Thus far I have described only situations where each member of the dancing couple navigates the hypermetric disruption in the same way—as one dancing unit: either both dancers adjust to a new footwork pattern, or both maintain the old footwork pattern. However, we have already seen that dance couples sometimes separate into open work, the independence of which allows one dancing partner to adjust to a new footwork pattern while the other partner maintains the old pattern. This situation occurs quite frequently in choreographed salsa performances, where the open work is designed specifically for the stage; that is, sequences of flashy footwork are executed individually but in synchrony, and with the couple(s) facing the audience rather than each other. This type of open work provides an effective change in dance texture and can be used to highlight changes in musical texture, the synchrony of dance choreography possibly coinciding with passages that feature unison instrumental playing or other types of musical synchrony.
To illustrate, dance interpretation no. 2 shows how dance performers might take advantage of the added measure at the end of the A subsection to transition into coordinated open work for the B and C subsections, later transitioning back into partner work via the second added measure in the return to the A subsection. There are many ways to accomplish this, but one possibility might be for the follow to add an extra step on the count of 4 in m. 9 so that she can begin another consecutive backward break, aligning with her partner's steps beginning on the downbeat of m. 10. Because of the clave flip and his lack of footwork adjustment, the lead finds himself dancing on the “follow's counts” (backward on 1, forward on 5) for the B and C subsections but then arrives back on his own counts (forward on 1, backward on 5) for the return of the A subsection. To realign with the A material, the follow adds another extra step in m. 18. Video 9 provides this precise footwork interpretation (with a few choreographed embellishments added), and Video 10 shows this idea on the stage, although using a slightly different section of the song.40 We see in this choreographed stage performance how couples from Dansschool Ritmo Latino transition from partner work to open work via an added step by the follows, with the leads maintaining their basic footwork throughout. Following a full-speed run-through featuring the whole dance team, Video 10 provides a slowed-down version that zooms in on just one dancing couple; downbeat footwork annotations (follow in blue, lead in red) show how the couple starts out dancing on opposite feet and ends up dancing on the same feet via the follow's added step at 1:21.
Dance-music analysis no. 4: Grupo Niche's “Cali pachanguero”
As my final example, Colombian Grupo Niche's “Cali pachanguero” features several types of metric disruption for analysis and interpretation. There are many recorded versions of this song, ranging in length from the very short radio version (clocking in at 3:19) to more complete versions running closer to six minutes, as shown in Table 2. Figure 5 provides an overview of the most extensive studio-recorded version of the song I could find (see Table 2, top entry), with measure numbers running along the bottom row of the array, color coding to indicate repeated motivic material, syllables of text to identify phrase beginnings, and asterisks to show the hypermetric alternation of strong and weak measures, primarily as a method of keeping track of formal proportions and deviations.41 As can be seen from this representation, most of the song features groupings that are multiples of two measures. However, in three instances this otherwise regular hypermeter is disrupted—near the beginning of the song (in the instrumental introduction), in the middle (in a vocal interlude prior to the entrance of the montuno), and near the end (in a mambo-like section featuring the tres, bass, finger snaps, and horns).42 With the exception of the radio version (which preserves the first two disruptions but cuts out the third) and the remixed version (which changes material more radically in a number of places), the versions I surveyed maintained at least one iteration of each disruption, suggesting they are integral components of the song; let us examine each of these passages in some detail.
Example 5 provides my transcription of the song's introduction, which articulates four-measure phrases that subdivide into two-measure calls from the upper brass, followed by two-measure responses from the lower brass. The trumpet material strongly establishes the downbeats of the 4/4 meter with contour and durational accents, while the trombone material pits dotted quarter-note durations against it. Indeed, these dotted quarters seem to be motivated by and provide an echo of the dotted-quarter durations on the 3-side of clave (displaced by an eighth note in each appearance), and they form important motivic connections on multiple metric levels as we move through the song.43
In the fourth call-and-response phrase that begins in m. 13, the trombones enter one measure early, and their response is cut off one measure early by the melodic material that begins in m. 15. Although the compression of the fourth phrase by two measures does not create a footwork problem, it does signal a more significant disruption that is soon to come. In mm. 15–20, all melodic instrumental layers switch from articulating the accents and grouping patterns of the prevailing 4/4 meter to those of an implied 3/4 meter, creating a striking grouping dissonance.44 The placement of melodic repetition within this 3/4 section is interesting on a slightly higher metric-organizational level as well. Four iterations of a three-note sequential pattern are heard over three complete notated measures of 4/4, at which point the pattern repeats. After these six measures, a two-measure tag ending reconciles the melody with the 4/4 meter in preparation for the entrance of the coro in m. 23. Thus this 3+3+2 grouping of measures plays out the 3-side of clave (a.k.a. the tresillo rhythm) on a higher level of rhythmic-metric structure, much in the same way as the trombone material played with the clave's threes on the more immediate lower level.45
I have observed at least four solutions to the particular dancer's dilemma presented in mm. 15–20. As indicated by follow footwork option 1 in Example 5, dancers may try to maintain their original eight-count footwork pattern, despite the distinct lack of support for this 4/4 meter in the music. Indeed, the passage is very disorienting if approached in this way, as this option physically reinforces the notated meter and thus the grouping dissonance that occurs throughout these measures, preventing the dancer (or dance observer) from settling comfortably into the melodically prominent 3/4 meter. However, if the dancer can block out the melodic structure, focusing instead on the percussion elements, which maintain the 4/4 meter, they will eventually find themselves comfortably aligned with the melody again in m. 21. A second option indicated in this transcription is to change the footwork pattern to conform to the 3/4 accents and melodic grouping of the music in some way. This option requires modification of the structure of the basic step in order to dance to the rhythm of the music, rather than to the previously established (and soon to return) 4/4 metric grid. For instance, one could retain all the original steps but take out the pauses to squeeze the original eight-count basic into six counts, thus eliminating the durational accents in the footwork pattern and increasing the rate of steps. As with the first option, this second option also works out evenly, allowing the dancer to resume the normal basic in m. 21.46 As a third but related option, one could insert additional pauses, stepping or posing only on the accented beats of the implied 3/4 meter (i.e., leaving out the steps in parentheses in option 2). A fourth option reflects the fact that all the material prior to the entrance of the coro is introductory and serves a real practical function in the salsa club—time to have a drink, find a new dance partner, make your way onto the dance floor, and otherwise prepare to start dancing in m. 23. (Indeed, I have yet to find a social dancing video of this section, most likely for this reason.) Video 11 demonstrates footwork options 1–3; the reader is encouraged to observe how each option feels in relation to the music by dancing along.
The coro refrain that begins in m. 23 alternates with the sonero's verses until a new call and response between the vocalists begins in m. 94 (which establishes the material for and leads into the montuno proper in m. 110). However, the vocal interlude that bridges these two call-and-response sections provides the second dancing dilemma found in this song. Indeed, it is the melodic material beginning in m. 91 that Washburne (2008) transcribes and points to on numerous occasions as an example of a particularly cruzao passage, using it to support his assertion that this song was composed without much regard for clave. Washburne writes: “The reasons for the crossing of clave in ‘Cali Pachanguero’ are unclear. Coming from Colombia, the arranger may not have understood the use of clave as well as [other arrangers]. Perhaps the arranger viewed other musical factors as taking precedence over adherence to the clave and made a conscious decision to depart from conventional practice” (183). He suggests that the problems in the song are so egregious that many older dancers and musicians will not dance to it, while also acknowledging that it became a huge hit despite these potential issues. However, having transcribed the passage for myself and later discovering and comparing it with Washburne's, I noticed some discrepancies that suggest an alternative interpretation for these musical “errors.”
Washburne's transcription of the passage is reproduced in Example 6, while mine is provided in Example 7 (note that my transcription begins eight measures earlier than Washburne's, in anticipation of points about dance interpretations I make below). Washburne (2008: 182) transcribes only the vocal line, placing clave on the system beneath and Xs above to show where melody and clave align, and suggesting that “three of the twelve measures notated [the first, third, and last] do not have any correlation to the clave strokes.”47 However, on comparing my transcription with Washburne's, I realized that we were working from different versions of the song, since two of his “problem” measures (the third and last) do not even exist in my transcription, thus significantly minimizing (or possibly eliminating) the cruzao effect of the passage. Compare the two audio versions (beginning just before the passage in question) provided in Video 12.48
I have not yet fully uncovered the origins of and reasons for these contrasting versions; however, regardless of the discrepancies, my interpretation of this passage also differs from Washburne's in another way. As mm. 91–93 of my transcription show, I do not consider clave to be in operation at all during the measures that Washburne finds most problematic. Instead, the instruments drop out completely so that the sonero can accent every third eighth note (a reference back to the rhythmic figures of the introduction) to play once again with the listener's perception of the 4/4 meter. Indeed, in most versions of the song, if the dancing couple were to keep their footwork going throughout this pause in the musical texture, they would find themselves dancing “on the wrong counts” when the new section enters in m. 94, as shown in footwork option 1 of Example 7. Option 2 and Video 13 show how dancers Melissa Mitro and Oscar Naranjo manage to solve this “problem” section by executing a flashy neck drop in place of a regular footwork pattern.
Finally, the hypermetric disruptions that occur at the end of this song are perhaps the most interesting of all, featuring a clave flip from the 3-2 orientation that has been heard consistently throughout the song to 2-3 orientation via a transitional section that involves a displacement dissonance between the melodic groupings of different musical layers. Let us investigate how the first of these flips is accomplished by examining the transcription provided in Example 8. With the anacrusis to m. 182, the tres and bass play a repeated four-measure melody that coincides with the established 3-2 clave. This mambo-like material continues, but one measure after it begins again for a third repetition, the horns enter with accents that conflict with the hypermetric downbeats previously established by the tres and bass. The material that begins in m. 191, then, forms a transitional section where conflicting hypermetric downbeats are articulated by different instrumental layers. When the coda section begins in m. 201, its melodic groupings coincide clearly with a 2-3 clave orientation for the first time in the song.49 Thus, to align with the new hypermetric downbeats that arrive unambiguously in m. 201, the dancer must transition from listening to the tres and bass material to listening to the horns. In Video 14 this transition is embodied quite clearly through the dancers' movements: after executing an eight-measure spin, they dance as a couple to the metric grid during the tres/bass material; the couple then separates, dancing to the horn rhythms in solo open-work fashion during the transitional mm. 191 to 200; finally, the dancers move back into partner work once the new hypermetric orientation has been clarified in m. 201. More specifically, note how the follow punctuates these sharp horn figures with head flicks, arm gestures, and hip swivels before resuming her newly aligned basic step. (Video 14 plays the entire excerpt with annotated score and hypermetric footwork counts at full speed, and then replays the excerpt starting from the mambo at half speed.)
These analyses from four of salsa's most popular songs begin to show how dancers contribute to this collaborative art form in a number of important and varied ways. From the outside, we might observe both how the physical gestures of dancers highlight and interpret events that already occur in the music itself, and how they create new layers that add to the richness and complexity of the collective endeavor. From the inside, we might reflect on how these hypermetric disruptions affect our own bodies, taking pleasure both in the presence of disruption and in its absence. On one foot, disruptions and ambiguities present challenge and excitement, causing us to pay close attention to musical elements in order to coordinate our movements in a convincing way—these events literally keep us on our toes. On the other foot, a straightforward metric groove is relaxing and reassuring; when we do not have to concentrate on where the hypermetric downbeat is, we are free to explore other aspects of our dancing—like musicians, dancers can improvise more creatively when they know that the metric grid is stable beneath them. While my analytic discussion in this article has focused primarily on how the hypermetric grid of the music plays with or against the hypermetric grid of the basic footwork pattern, and while this initial project has proven to be reasonably complex, this is just the first step in understanding the depth of rhythmic interactions that occur between salsa musicians and dancers in the heat of live performance and improvisation.
Trombonist and ethnomusicologist Christopher Washburne (1998: 180) provides a somewhat contrasting perspective of the practices of New York musicians in the salsa romántica era of the 1980s and 1990s when he suggests that synchronized body movements do play an important role in helping horn sections achieve a “hard-driving swing”; however, he stops short of making claims for “visual enhancements” in the form of real dance moves as part of these groups' regular performance practices.
See Aparicio 1998; Aparicio and Jáquez 2002; Aparicio and Valentín-Escobar 1994; Berríos-Miranda and Dudley 2008; Boggs 1992; Cortez, Falcón, and Flores 1976; Den Tandt 1997; Duany 1984; Gerard 1998; Glasser 1995; Manuel 1994; Padilla 1990; Roberts 1999; Román-Velázquez 1999a, 1999b; Rondón 2008; Shain 2009; Singer 1982; Steward 1999; Washburne 1998, 2002, 2008; and Waxer 1994, 2002a, 2002b.
For example, even though Rondón (2008: 33, 62–63, 157) argues for the essential role of dance in salsa throughout his book, he neglects to mention any salsa dancers by name, despite the close working relationships that musicians and dancers are known to have had throughout this era. See, for instance, the connection between dance icon Eddie Torres, who was musically trained as a child, and the “King of Mambo/Salsa” Tito Puente, who was also an accomplished dancer before an injury turned him toward his illustrious career in music (Hutchinson 2004: 118, 123–24; McMains 2015: 174–75).
See Borland 2009; Bosse 2008; Delgado and Muñoz 1997; Díaz 1994; García 2013; Hutchinson 2004, 2014; McMains 2015; Pietrobruno 2006; Renta 2004; Skinner 2008; Sloat 2002; Urquía 2005; and Wieschiolek 2003.
To the best of my knowledge, no researcher had undertaken a study of the complex interactions between salsa music and dance prior to Simpson-Litke and Stover 2019. Dance scholar Juliet McMains and musician Ben Thomas (2013) attempt a similar collaboration, bringing dance and music into conversation, and while their article features examples taken from specific dance-music forms, including salsa (210), their analytic approach is more descriptive than technical in nature.
Two examples are drawn from the salsa music literature as illustrations: (1) Following his helpful breakdown of salsa's instrumental layers and arranging techniques, Charley Gerard (1998: 108) opens his brief and unnuanced chapter on dance in the second edition of his book with this telling comment: “A lifetime of staying on the sidelines of the dance floor and an obsession with musical details kept me from investigating Latin dancing when I was writing the first edition of Salsa.” (2) After describing the importance of physical movement in locking into the salsa groove, Washburne (1998: 180) makes a similar comment: “As a performing salsero and music scholar, I lack the proper qualifications to explore these connections in more depth, however, they are deserving of greater attention. An analysis of kinetic processes and their relation to sound phenomenon could prove beneficial.”
With its origins in this tradition, the discipline of music theory has played its own role in the marginalization and suppression of particular voices within and outside the field, with its established scholars often working as the gatekeepers of “quality,” determining which repertoires are worthy of study and which analytic approaches are valid. While some scholars in the field have made attempts to embrace new positional listening perspectives and to promote study of more diverse repertoires, Yayoi Uno Everett (2019), Philip A. Ewell (2019, 2020), Ellie M. Hisama (2019), and Joseph Straus (2019) have provided recent reminders that we still have far to go in exposing and rooting out the systemic racism, sexism, homophobia, ableism, and other forms of discrimination and oppression that are deeply embedded in our discipline.
A “music-first” analytic approach is validated by dancers' frequent engagement with prerecorded musical tracks. While true conversations do take place in the spontaneous interactions between dancers and live music in salsa clubs (I explore this aspect of salsa practice in other projects), I believe that the kinds of metric events I investigate here are usually planned out in musical arrangements ahead of time, rather than improvised on the spot, so dancers are less likely to influence the behaviors of musicians in these moments. However, I also suggest that my analytic approach in this article is more reciprocal than it might initially appear. While musical gestures might elicit or inspire particular physical responses, these dance gestures in turn highlight and thus affect how the observer of the collective art form perceives musical elements.
In addition to choreographed stage routines, videos of social dancing in clubs or house parties have emerged on the internet in recent years, further demonstrating an acknowledgment of and appreciation for improvised dancing in the global salsa community. Indeed, famous stage performers and instructors earn (additional) credibility and respect within the community of viewers if they demonstrate an ability to adapt their unique styles to spontaneous interactions with a partner and with the music on the social dance floor (McMains 2015: 281).
In simple terms, the lead's main role is to determine the sequence of moves for the partnership and to communicate clearly to the follow what move is coming up next through a commonly understood language of physical cues. The follow's main role, then, is to read these cues and to respond appropriately by executing the requested move. Several scholars (e.g., Aparicio 1998; Aparicio and Valentín-Escobar 1994; Borland 2009; Díaz 1994; García 2013; McMains 2015; Skinner 2008; Washburne 2008; Wieschiolek 2003) have addressed the problematic gender dynamics in salsa: misogynist lyrics, suggestive album covers, sexual harassment in night clubs, old-boy's-club mentalities in salsa bands and dance companies, and the dance structure itself (the man as lead, the woman as follow). While the follow's role may be viewed as passive and submissive, many experienced female dancers throughout salsa's history have chosen not to dance it this way. Instead, they have pushed the boundaries of the dance structure in innovative ways, challenging machismo attitudes and expressing their own individuality and style. At times, I find it convenient to refer to follows and leads with gendered pronouns, but I also wish to stress that these dance roles need not continue to be tied to gender (for discussions of same-sex salsa and women taking the lead, see McMains 2015: 137–45 and 284–92, respectively).
It is important to acknowledge that expertise in salsa may be obtained and recognized in various ways: while some salsa musicians and dancers are classically or studio trained, others have learned and honed their exceptional skills via mentorship within rich community-based traditions of dance and music making, and still others have drawn on some combination of both.
As is common in ballroom dance footwork diagrams (and in the historical tradition of the Beauchamp-Feuillet track-drawing system for baroque court dances), Figure 1 provides a bird's-eye view of the footprints a dancer would make on the floor, while Video 1 is shot from two perspectives—above and behind the dancer's feet—to encourage readers to embody these steps by putting themselves in the dancer's shoes. As explained by Simpson-Litke and Stover (2019: 80), many dance teachers simplify this salsa basic by instructing students to bring the feet together and back to the middle centerline at the end of each half (on counts 3 and 7). In practice, there is some variability in the placement of each half's third step in relation to the second step, and my footprint and video representations attempt to show what tends to happen in the flow of dancing, when momentum often causes the dancer to travel across the dashed centerline rather than land right on it.
McMains (2015) provides a detailed discussion of the development of the salsa basics. The most popular alternative to breaking on-1 is “on-2,” originating in New York with two prominent variants (see Hutchinson 2004; Renta 2004; Simpson-Litke and Stover 2019), but traditions of breaking “on-3” and “on-4” have occurred in different places and at different times through salsa history. In my work I have remained neutral in the debate over which of these basics is superior while still acknowledging the important, sometimes identity-defining role this debate plays in certain dancing communities. I focus instead on the richness contributed by this variety of options, as each basic encourages the dancer to place physical accents in different temporal locations and thus to interact with and highlight different musical events, extending the range of interpretive possibilities within the art form. For diagrams and video demonstrations of the on-2 versions of the basic, see Simpson-Litke and Stover 2019: 80–87, example 7, and videos 2–3.
The phenomena I explore in my analyses here are primarily hypermetric. In this respect, the differences between the dance basics are inconsequential, as all feature the same length of cycle. Two-measure dance units are not unique to salsa but are found regularly in other Latin partner dances (cha-cha, bachata, cumbia, etc.), non-Latin partner dances (minuet, swing, etc.), and other dance forms. For instance, see hypermetric investigations of the minuet in Leaman 2016, McKee 2012, and Sánchez-Kisielewska 2015.
In support of their style of dancing, on-1 dancers often point to the “millions of dancers around the world who instinctively respond to salsa music by dancing on-1” and suggest that the musical elements that on-2 dancers refer to in support of their style are overstated, take specific training and experience to perceive, and, depending on the style of salsa music, are not always prominent in the musical texture—at all, or at least compared with the musical elements that emphasize downbeats (McMains 2015: 154). Along the same lines, on-2 dancers often take pride in this general perception that their style requires greater musical sophistication, with many dancers switching from on-1 to on-2 as their level of dancing advances.
I first became aware of this type of notation for dance in Leaman 2013 and adapted it for my own salsa purposes in Simpson-Litke 2014. Like a number of other dance-notation systems (Benesh, Conté, Saint-Léon, Stepanov, Sutton; for more information, see Guest 1998), Leaman's choreomusical scores for ballet use the horizontal dimension of the staff to represent series of events in time, while the vertical dimension (lines and sometimes spaces) indicates the height of gesture in physical space. However, like me, Leaman (2015, 2016) also maps physical space onto staff lines differently in her representation of other dance forms; for instance, she uses the top line to indicate forward steps (rather than a high position or upward motion) and the bottom line to indicate steps to the rear (rather than a low position or downward motion) in the running man and Charleston moves found in the Melbourne shuffle of rave dancing. (All these representations contrast with something like Labanotation, which presents time vertically and arranges body parts horizontally on the page.)
The lead's essential footwork can be obtained easily from the follow's by switching the two halves of the basic; that is, the lead executes the forward break on the count of 1 and the backward break on the count of 5.
Gerard (1998: 112) outlines a simplified version of this form, saying that there are four essential sections: “a verse, a montuno section based on a short series of repeated chord changes, an instrumental section called mambo with greater harmonic movement than in the montuno, and a return to the montuno.” Both Washburne and Gerard agree that the intro, moña, and coda are optional sections. While Washburne (1998: 66) indicates that the nightclub environment supports the extension of tunes by eight to fifteen minutes, allowing time for musicians to take solos and for dancers to find new partners on a crowded dance floor, McMains (2015: 107) suggests that the increased athleticism of studio-taught and competition-influenced salsa dancing has led many of today's dancers to prefer shorter songs of six to seven minutes, requiring live musicians to make some concessions if they want to be hired for these types of dance events.
I suspect that metric disruptions are more likely to occur in the precomposed sections of the salsa form than in the improvisatory sections, where the maintenance of a consistent metric structure is what allows individual soloists to create material spontaneously while keeping the ensemble together. In keeping with David Temperley's (2008) observations of hypermetric shifts used to emphasize formal divisions between sections in common-practice music and Nicole Biamonte's (2014: 7.1) linking of metric dissonance to formal function in rock music, specifically “as a truncated link at the end of a formal section that provides momentum toward the end of a phrase,” hypermetric disruptions in salsa music seem to occur frequently in the cierres between formal sections.
Several important sonic events provide audible and consistent cues alerting listeners to the beginnings of these one-measure cycles, even within this dense texture; in particular, in Example 1 see the change from high to low drum in the bongó's martillo pattern and open tones of the congas' tumbao pattern on beat 4. The güiro's accents every fourth eighth note articulate a consistent half-measure pulse stream as well.
Conga players (congueros) often play variations on this slap/open tone pattern, incorporating rhythms from other dance forms like the Puerto Rican bomba and plena and the Cuban rumba, as well as improvised solo flourishes (Washburne 2008: 170).
The guajeo, also known as the montuno because it characterizes this section of the form, is a repeated melodic/harmonic pattern played by the pianist. While the guajeo is usually uniquely composed for each song, many feature similar rhythmic patterns, melodic contours, and harmonic progressions. See, e.g., Gerard 1998: 54–58. Using the Salsa Rhythm application (salsabeatmachine.org), one can hear each component of the salsa rhythm section individually and in various combinations.
In a similar vein but on a larger scale, Edgardo Díaz Díaz (1994, discussed in Aparicio 1998: 96) provides an interesting study of how musicians might tailor the order of songs within each set, inserting other dance forms (merengue, bolero, etc.) between salsa songs to produce the right balance of tempi and energy for the physical requirements of the dancers during an evening.
We could also think of 2-3 and 3-2 as rotations of each other, beginning four beats apart in a continuous eight-beat clave cycle. See Simpson-Litke and Stover 2019: 75–78.
Gerard (1998: 31–50) and Washburne (1998: 162–63; 2008: 180) provide general guidelines to determine if a musical phrase is in clave and to which orientation of clave it corresponds. Both authors note that developing the “correct” clave sense (i.e., the sense acceptable to traditional salseros) takes years, and even the most experienced and talented salsa musicians have, from time to time, found themselves on the “wrong” side of it. Washburne 2008: 194–97 provides fascinating discussions of “clave as identity” and “clave as cultural weapon,” used to create boundaries between “insiders” and “outsiders” within the genre and as a potential point of debate even within otherwise cohesive circles.
Many scholars have explored how, in the origins of salsa, clave traditions from older Cuban dance-music forms (son and rumba) were used and developed primarily by Puerto Ricans living in New York in the 1960s and 1970s. Changing attitudes toward clave emerged as salsa was embraced in other Latin countries (that had looser connections to Cuba and that often incorporated their own regional dance-music forms) and as younger generations of musicians and producers (with a greater desire for innovation and record sales than for preserving a strict clave-based tradition) emerged in the 1980s and 1990s (see McMains 2015: 85; Washburne 2008: 184). Even within the tradition, however, salsa has always had its renegades, who have gone against clave to achieve other musical effects. Sue Steward (1999: 12) writes: “The biggest sin for such purists is to play out of clave, though that has never bothered mavericks like Willie Colón, who started breaking the rules in the 1970s, or today's leading salsa producer, Sergio George, who delights in reversing clave and changing its accent. Colombia's salsa musicians have their own sense of clave which stops and starts and changes mid-song, and the dancers match it.”
Just as it takes time for musicians to develop a good “clave sense,” it takes time for dancers to develop a sensitivity to hypermetric units, should this be important to them (it is not of concern to all dancing communities). A number of YouTube videos are designed to help dancers refine their listening in these ways and demonstrate how footwork changes may be executed smoothly to reflect musical changes when needed. For examples, see videos by Anna LEV 2016, Black Belt Salsa 2017, and Joel Salsa NYC 2009.
Temperley (2008) entertains various listener responses to hypermetric shifts in different musical passages, suggesting that a change from one hypermetric organization to another might be perceived to occur (a) suddenly at a particular moment in time, (b) through a hypermetrically ambiguous transitional passage, or (c) via some more complex cognitive process—for instance, where the hypermetric organization of a later passage retrospectively influences perception of the hypermetric organization of previous musical material. I suggest that, if salsa dancers choose to physically represent a hypermetric change, it will most often be suddenly, as in Temperley's first listening option, because a clear and unambiguous basic step is necessary for the physical coordination of dancers in partner work. However, Temperley's second option might be physically engaged through a move to open work; that is, if dancers are not bound to each other, they are also not bound to the (hyper)metric grid through an obligation to maintain basic footwork and thus have more freedom to physically represent musically ambiguous passages in other ways. Finally, in consideration of Temperley's third option, there is no way to redance a passage once it has passed; however, the perception of later musical material might influence how earlier musical material is physically interpreted in a subsequent dance performance of the song.
This example is also used in Simpson-Litke and Stover 2019: example 13, which focuses primarily on other structural aspects of the song.
This is an example of a sudden hypermetrical shift, as discussed in Temperley 2008.
The audio for Video 3 is taken from a YouTube channel by the Salsa Beat Machine (creator of the Salsa Rhythm app), whose goals are to help relatively inexperienced salsa dancers “develop better timing, learn about different key instruments involved in Salsa compositions, and improve [their] ability to execute [their] dance patterns . . . while developing concepts of musicality” (play.google.com/store/apps/details?id=com.salsarhythm&hl=en_US&gl=US). The creator of the audio track has added his own voice-over to clarify the musical count of “one” for the dancer throughout the song. Indeed, the preponderance of such videos, classes, workshops, and apps designed to address the subjects of salsa timing and musicality supports my assertion that dancers are acutely aware that these metric “problems” exist in many salsa songs and that to ignore them might render one's dancing “unmusical,” at least in this respect. For the full video, which contains animations and other entertaining information about the song, see Salsa Beat Machine 2010.
Using terminology developed by Harald Krebs (1999), this would be a D(8+4) dissonance, where 1 = the quarter note; that is, both music and dance cycles are eight quarter-note beats long, but their beginnings are displaced in relation to each other by four beats.
When faced with metric change, Andrew Imbrie (1973) proposes that a “conservative” listener (embodied in interpretation 1) will maintain the old metric structure for as long as possible, while a “radical” listener (embodied in interpretation 2) will quickly adopt the new metric structure. Similar ideas are engaged in a number of related writings in music psychology and music theory; see, e.g., W. Tecumseh Fitch and Andrew J. Rosenfeld's (2007) “coexistence,” “assimilation,” and “resetting” cognitive strategies for dealing with syncopated rhythms and Temperley's (2008) discussion of listeners' perceptions of sudden versus gradual hypermetrical transitions.
In an analogous discussion of a particular Cuban rumba practice, McMains (2015) notes that some dancers align their steps with the 3-side of clave but not with the 2-side. Instead of suggesting that the 2-side be simply ignored by dancers in this practice, however, she proposes that they “enjoy the tension produced by stepping against it” (186), which aligns closely with my interpretive suggestions here. Several dance-music scholars have observed these types of strategies for embracing metric dissonance, particularly in choreographed dance forms. For instance, Stephanie Jordan (1986) began explorations of rhythmic counterpoint and conflict between music and dance (including at the hypermetric level), noting choreographer Doris Humphrey's opposition to dance structures that follow musical structures too closely (176–77), and has continued to examine such relations in her later books (Jordan 2000, 2015). See also Bell 2017, Horlacher 2018, and Leaman 2016 on ballet; Stevens 2018 on eighteenth-century contredanse; Short 2019 on musical theater; Goldberg 2019 on Bulgarian folk dance; and Bilidas 2019 on tap dance.
When watching this dance interpretation, the viewer may note that, rather than a “quick-quick-slow” rhythm, the follow's steps are often (nearly) equal in duration. I believe this is happening for three main reasons: (1) At faster tempi it is harder to maintain a strict distinction between the “quick” and the “slow” durations. This phenomenon is documented in Richard Jankowsky's (2013) work on the microtiming of gradually accelerating time cycles in Tunisian music, where nonisochrony collapses into near isochrony as tempi increase. (2) The follow is often preparing for moves (like the cross-body lead) that cause her to travel across the dance floor on the count of 5; it is likely that her partner is already beginning to pull her forward after the count of 3, causing her to begin travel a bit earlier than count 5 and evening out the durations of the steps as a result. (3) In the most commonly danced variant of on-2, dancers actually step consistently on the anacrusis to each downbeat, meaning that the lengths of steps (as measured in quarter notes) are 〈1.5, 1, 1.5〉 rather than 〈1, 1, 2〉 as with on-1. It is quite possible that familiarity with this on-2 timing is an influence here as well (the “New York” sign in the background of this video is a clear reference to the birthplace of on-2 salsa).
Open-work passages are also called “shines,” but as Renta (2004: 150) notes, this is a derogatory term that “harkens back to the days when African-American shoe-shine boys [which is itself a derogatory term, as it was often used in reference to adult African Americans] would offer to dance for change.” While the phrase is still in common usage in the salsa dance community, I will avoid it.
Sydney Hutchinson (2004), Renta (2004), and McMains (2015) describe in some detail how polyrhythmic elements in salsa and other Latin music are often matched by the dancers' polycentric movements (i.e., initiated with or emanating from multiple body parts simultaneously), noting how these aspects of Latin dance-music practice are derived from African traditions of music/movement and are somewhat incongruent with European dance aesthetic principles.
William Rothstein (1989: 56–57) discusses such events in common-practice music as “elongated upbeats,” not counted as part of any hypermeasure and effectively suspending the sense of hypermeter momentarily. He generally groups this “extra” measure with the hypermetric downbeat of the phrase that follows, and Manuelle's anacrustic vocal entries in mm. 11 and 20 of Example 3 support this interpretation.
That is, instead of adjusting footwork abruptly in m. 10 as depicted here, the lead may need the first couple of measures (or perhaps the whole first four-measure statement) of the B subsection to realize that the music has changed and to maneuver the couple's footwork into the new orientation in a way that is (ideally) not too jarring or ungraceful for the follow (see footwork-changing videos by Anna LEV, Black Belt Salsa, and Joel Salsa NYC cited in the bibliography for tips on how to physically accomplish this). Similarly, the return back to the A subsection might be executed in a more gradual way than my transcription suggests.
Video 10 begins just prior to the montuno section (not transcribed in Example 4); when the montuno does enter at 0:35, the listener will recognize the A subsection coro material but will also notice some musical deviations because this video shows the first montuno (rather than the second montuno transcribed in Example 4). In the first montuno, the four-measure coro alternates with a four-measure inspiración (the sonero actually overlaps a bit with the coro each time, but these are the essential proportions). In the second montuno, the same coro returns but is then followed by a truncated inspiración and a shorter call-and-response between coro (“Me liberé”) and horns. The B and C subsections and the important cierres that create clave flips are newly added, which is why I focused on this second montuno in my analysis. Unfortunately, a suitable video was not available for this later section of the song.
To receive an asterisk, the musical material in that measure has to serve as some sort of beginning. One asterisk marks the beginning of a(n anticipated) two-measure unit (motive), two asterisks mark the beginning of a(n anticipated) four-measure unit (phrase), and three asterisks indicate the beginning of an (anticipated) eight-measure unit (subsection). Temperley (2008: 306) provides a helpful summary of the metrical preference rules from Lerdahl and Jackendoff 1983 that are especially important with regard to hypermetric analysis.
A survey of four other versions listed in Table 2 shows that cuts tend to be made in the final cadential sections of the song, usually omitting repetitions of the instrumental mambo and/or coda sections, with two exceptions. (1) The radio version leaves out some surprisingly important material in the first half of the song as well, presumably to fit within the constraints of airtime. (2) The remixed version features more significant structural changes, including different melodic and harmonic content, additional measures that change the formal proportions within sections, reordering of musical material, and so on.
For more on three-generated rhythms superimposed on duple metric frames in popular music, see Cohn 2016.
In Krebs's (1999) terminology, this is a G(4/3) dissonance, where 1 = the quarter note.
Thanks to Richard Cohn for pointing out the connection between the grouping of measures in this section and tresillo.
By changing the step in this way, the dancer is replacing six quarter-note rests with what amounts to one full six-step basic.
In the first three measures, there are actually more articulations in the melody that align with clave than Washburne notes, so his comments are a little unclear. I suspect that he is considering as correspondences only places where melodic accents align and is disregarding the unaccented notes altogether.
Washburne was working from a version that was released in the United States on a compilation album (Grandes Exitos) in 1989; while I could not obtain this exact recording, I believe that it is similar to the remix referred to as version 5 in Table 2 and provided in Video 12, the only version I found that featured those additional measures. I had initially wondered if this was an early version that had later been modified to delete the cruzao measures and “fix” the passage, as sometimes happens, but on listening to several other versions (including what I believe to be the original 1984 version), I think that this was not the case.
This is an example of a gradual hypermetrical shift, as discussed by Temperley (2008).