AAWM Journal 12/1 (2024)

**Singing in Tune: Berimbau Tuning and Song Mode in Capoeira (1937–1978)**

Juan Diego Díaz and Alex Rossi

Berimbau, Capoeira, Singing modes, Microtones, Tuning, Tonicization

The berimbau musical bow is the signature instrument of the Afro-Brazilian art of capoeira. Alone or in groups of two or three, it leads the capoeira’s ensemble by creating various kinds of grooves that accompany songs and physical games. As pitched instruments, they have the potential to guide singers melodically or even harmonically. In practice, however, practitioners rarely talk about precise bow tuning or singing tonalities and their approaches are widely diverse. This article investigates the tuning of the berimbau and the relationship between berimbau tuning and song mode. Our goal is to infer the most common tuning and singing practices used by practitioners through a systematic study of field recordings by foreign researchers, commercial albums by practitioners, movies, documentary films, TV programs, and transcriptions of capoeira music from the 1937 – 1978 period, a formative period in the development of modern capoeira. With a method combining musical analysis at tonal and microtonal level, we analyze a corpus of 448 songs to propose a five-tier taxonomy of relationships between tuning and mode: (1) Direct tonicization, when singers directly tonicize one of the berimbau pitches; (2) Indirect tonicization, when singers use one of the bow pitches as an indirect reference—the berimbau is not tuned to the song’s fundamental but to a different scale degree such as the third or the fifth; (3) Chordal tonicization, when berimbau pitches form a dyad or triad that sets up the singing mode; (4) Overtone tonicization, when singers tonicize not a bow note but one of its prominent overtones; and (5) Non-tonicization, when singers seem to disregard bow pitches. We propose that these categories of tonicization may correspond to unspoken ideal aesthetics that practitioners developed in the 1937 – 1978 period and persist today.

Juan Diego Díaz is Associate Professor of Ethnomusicology at the University of California, Davis.

Alex Rossi is a PhD candidate at the University of California Davis.

Click for DOI, citation, and PDF version.

Click to read Appendixes.

[1] The berimbau musical bow is the signature instrument of the Afro-Brazilian art of capoeira. Alone or in groups of two or three, it leads the capoeira ensemble by creating various kinds of grooves that accompany songs and physical games where players try to outmaneuver each other using combat and dance movements. As pitched instruments, they have the potential to guide singers melodically and harmonically. In practice, however, practitioners rarely talk about precise bow tuning or singing tonalities and their approaches are widely diverse. This article investigates the tuning of the berimbau and its relationship to song mode. Through close musical analysis of the earliest musical materials on capoeira, our goal is to infer the most common tuning and singing practices used by practitioners. The materials we analyze mainly consist of field and commercial recordings, but also movies, documentary films, transcriptions, and textual sources from 1937 to 1978, a formative period in the development of musical practices in capoeira.

[2] Practitioners’ discourse about the relationship between berimbau tuning and song mode is typically vague and implicit. For instance, most mestres (masters) teach their students to tune the three berimbaus of the capoeira angola ensemble in three relative levels (low, medium, and high) without specifying the intervals between those levels.^[1] Likewise, students are advised to project their voices clearly and loudly, but are rarely given information about the specific mode they should sing in. Common instructions about the relationship between singing and berimbau playing include “when you sing you need to listen to the berimbau,” or “the voice should accompany or complement the berimbau.”^[2] Even when mestres give specific instructions for berimbau tuning, they do not usually describe how a singer should match the berimbau pitches. One of the reasons for these seemingly vague approaches is that, for most mestres, maximizing bow resonance and voice projection respectively take precedence over the specific pitches that can be sounded by the berimbaus and singing in a particular mode. Correspondingly, scholarly discourse has been vague about or unconcerned with these issues (e.g., Querino 1916; Carneiro 1936, 1937; Almeida 1942; Ott 1955–7; Rego 1968; Shaffer 1977). In her seminal 1977 monograph on the berimbau, Kay Shaffer, for instance, focused on the instrument’s history and rhythms, referring only tangentially to tuning and without giving specific details. However, this does not mean practitioners are not interested in these topics or that they consider them unimportant. Due in part to the online availability of historical capoeira recordings (precisely the ones we examine in this article), some current practitioners like mestres Cabello and Negoativo have developed a renewed interest in berimbau tuning and related ways of singing. The most concrete example is the 2020 book, Capoeira Afinada, by Brazilian capoeira practitioners Bárbara Kristensen and C. R. Santos (aka Contramestre Latino), which explores berimbau tuning and song mode in various historical recordings, including a small sample of the ones we analyze here. This book pioneered the study of tuning in capoeira, proving that practitioners do care about these topics even if they tend to talk about them implicitly. We expand the work that these practitioners began by providing a more rigorous musical analysis of a more complete historical corpus.^[3]

[3] In this paper we propose a method of analyzing these historical capoeira recordings using a combination of traditional transcription by ear with digitally generated song melograms that consider both tonal and microtonal activity, the interaction of which we see as crucial to the relationship between berimbau tuning and song mode. In this method, which we call hybrid transcription, high-resolution pitch detection software enhances the human act of musical transcription by identifying the microtonal nuances that human ears are unable to measure. Our analysis of recordings across nearly four decades yielded a large amount of data that we summarize in Appendix A. Moreover, we propose a five-tiered taxonomy of relationships between berimbau tuning and song mode that we believe are integral to capoeira aesthetics, even if they appear to be largely absent from practitioners’ discourse. Thus, our goal in this paper is two-fold: Firstly, we provide a novel methodological approach for the study of microtonal music through hybrid transcription. Secondly, we propose a taxonomy of approaches by which capoeira practitioners match their sung notes to perceived berimbau pitches.

[4] The article opens by providing some background information on musical aesthetics in capoeira, focusing on berimbau construction, playing technique, tuning, and repertoire, as well as singing practice. This is followed by a section outlining our method of analysis and another introducing the corpus of studied materials. Subsequently, we provide an analytical demonstration using samples of the patterns we identified. Finally, we present our proposed five-tiered taxonomy of relations between bow tuning and singing and conclude by discussing the broader implications of this study and suggesting ways to deepen the analysis ethnographically.

Background

Berimbau

[5] The berimbau used by capoeiristas today consists of a wooden staff with an approximate length of between 1.4 and 1.6 meters and a thickness of around 2.5 centimeters. The two ends are braced and bent into a bow shape by a steel wire. A gourd resonator, or cabaça, is fastened (with its cavity facing the player’s body) near one of the ends, thus dividing the string into two unequal parts. Players hit the longer side of the string with a thin wooden stick called baqueta while holding the bow vertically with the other hand, producing a base pitch known as solto. The hand holding the bow also holds a flat stone (pedra) or a coin (dobrão) that touches the string in different ways to produce two additional notes. By hitting the string with the baqueta while firmly pressing the stone or coin against the string, a musician produces a raised note called preso. The preso tone lies anywhere from a minor second to a major second above the solto. Finally, a berimbau player might obtain a buzzing note or chiado by hitting the string with the baqueta while loosely touching it with the stone or coin. Berimbau players also hold a small wicker shaker called caxixi (which is filled with seeds and has a handle) in the hand that holds the baqueta. While playing the berimbau, musicians will move the gourd toward and away from their stomachs to either mute, unmute, or produce a characteristic vibrato or “wah-wah” effect.

[6] While photos from the first half of the twentieth century (Carneiro 1936: 112–13; Calvacanti and Verger 1947: 57; Tavares and Verger 1948: 9–11) confirm that the berimbaus used in capoeira during this period were like the one described above, some authors suggest that they were shorter in length. In the first known academic article on capoeira, Renato Almeida (1942: 157), for instance, wrote that he had seen berimbaus approximately half a meter long at a roda (capoeira circle where performance takes place) in the Bahian town Santo Antônio de Jesús. However, US American anthropologist Ruth Landes wrote in her 1947 book that the berimbaus she saw in a street presentation in Salvador measured about 1.8 meters (Landes 1994: 100). As our analysis will reveal, berimbau tuning during the 1940s was generally lower than in subsequent decades and now, implying that berimbaus were not shorter but rather longer, as Landes reported. At any rate, these dimensions are relevant because they, along with the tension of the wire, the position of the gourd, and the place where the coin stops the string, affect pitch.

[7] In all the recordings we analyzed, the berimbau pitches we heard as fundamentals remain in the E♭2 – F3 range. Significantly, these pitches and the intervals formed between them often escape the 12-tone equal-temperament tuning system (12-TET) so commonly used in Western music theory. Chiados, on the other hand, are unpitched and add a contrasting buzz to solto and preso. Because the caxixi moves with the hand that holds the baqueta, its rattling sound is heard every time the string is hit.

[8] Capoeira practitioners use the three berimbau notes described above to create “melo-rhythmic” patterns (Nzewi 1974) that they generically refer to as toques. Toques are played within a groove logic of ostinato and variations, meaning that they alternate the restatement of a basic pattern with improvised variations. They are a central part of the capoeira groove, providing the backdrop for songs and physical games (jogo). Each group typically has a repertoire of ten to fifteen toques that they use in their performances, musical training, and recordings. Importantly, each of these toques has various extramusical associations, some historical, others related to the nature of the physical game. For instance, practitioners of the capoeira angola style relate toque de Angola, notated in Example 2, to a slow and elegant game close to the floor.

[9] A capoeira bateria, or musical ensemble, typically features one, two, or three berimbaus. In a three-berimbau bateria, the instruments are typically organized by size: the lowest-pitched bow is called gunga, the medium-pitched is médio, and the highest-pitched is viola.[4] In two of the analyzed recordings (Pastinha 1969 and Grupo Bonfim 1994 [1969]), however, four berimbaus could be heard in a few tracks. The berimbaus in the capoeira bateria are completed by a set of percussion instruments that provide mostly unvaried rhythmic accompaniment. Depending on the style, these accompanying instruments may vary. For instance, the emblematic capoeira angola ensemble has three berimbaus accompanied by two pandeiros (tambourines), a conga-like drum called atabaque, a two-pronged bell called agogô, and a scraper called reco-reco. Conversely, the characteristic bateria in the capoeira regional style consists of one berimbau and two pandeiros. In the recordings under discussion, we heard combinations of all these instruments. While played by the full bateria, toques are ultimately determined by berimbau patterns. When these toques are played by more than one berimbau, voicings may change, meaning that berimbaus may play the same pattern homorhythmically or with complementary rhythms. Example 1 shows the full texture of a capoeira angola bateria with three berimbaus playing a toque commonly known as Angola in Bahia. Notice how the berimbau texture moves from homophony in measures 1 and 2 to polyphony when viola improvises in measures 3 and 4, and how gunga and medio play “reversed” versions of the same pattern—that is, when gunga plays an open note, medio plays a raised pitch and vice versa. This manner of voicing is common, but, as we will see in the analysis of historical recordings, there is a great variety of voicing options.

Example 1. Toque de angola in “Paranaê” (GCAP 1996), track 10 (0:00 – 0:14).

Song

[10] The capoeira song repertoire has hundreds, perhaps thousands of songs. Capoeiristas typically group them in four main types: ladainhas, quadras, chulas, and corridos. The ladainha is an introductory solo song sung by the most senior member of a group, often the mestre, which precedes the capoeira jogo. These are typically structured in quatrains or sestets, follow rhythmic-melodic patterns characteristic of a group or singer, and commonly last from half a minute to a couple of minutes. There are, however, exceptionally long ladainhas—in a field recording from 1951, for instance, Mestre Waldemar sings an eight-minute ladainha (Leeds 1951). In rodas of the capoeira regional style, ladainhas are often replaced by a kind of four-line quatrain called quadra. The ladainha or quadra is then followed by formulaic responsorial songs called chulas (or louvacões), in which the choir repeats back the soloist’s verses with the word camará at the end. For example, a mestre may sing “Iê viva meu Deus,” to which the choir will respond “Iê viva meu Deus, camará.” Finally, a sequence of corridos is sung in a non-predetermined order to accompany the physical games. Corridos feature the greatest melodic, rhythmic, and structural variety of the capoeira song repertoire and are thematically connected to the games. Many ladainhas, quadras, chulas, and corridos heard in today’s rodas were canonized in the recordings we analyze in this article.

[11] Singers use a variety of modes for every song, with pitch collections ranging between three and seven notes per octave. Although all songs feature notes from either a major or minor triad that clearly imply a home note or fundamental, it is often difficult to assert that these songs are effectively in what are called major or minor modes in Western music theory because the pitch collection may be incomplete. A song using the pitches C–D–E–G–A, with C as the fundamental (i.e., C major pentatonic), for instance, can theoretically be in C Ionian (major), C Lydian, or C Mixolydian. Likewise, another song using the same five pitches, but with A as the fundamental (i.e., A minor pentatonic), could be in A Aeolian (minor), A Dorian, or A Phrygian. Many related genres from Bahia such as samba de viola or samba de roda use these and other modes and capoeira singers have historically partaken of the melodic richness that these modes make possible.^[5] Because our main concern is to determine whether there is a relationship between berimbau tuning and song mode, rather than to establish what exactly that mode could be, we have simplified the modes into two families: major modes and minor modes. Our analysis consists in comparing berimbau pitches to what we heard as the song’s home note (and sometimes to its third or fifth, too). While this does not imply that singers are necessarily thinking about these intervals, modes, or implied chord progressions, we noticed that some musicians in the recordings sing in parallel thirds or sixths, which could mean that they do follow implied functional harmonies. We therefore use the term “tonicization” liberally to refer to instances when singers use berimbau pitches as a reference in one of the four ways explained at the end of this article, rather than in the strict sense of Western music theory. Ladainhas and quadras are generative for this exploration because they are extended solo songs meant to be delivered in a highly personal style. In them, singers may use conversational styles with micro-rhythmic and microtonal activity. As we will see, in the transition from ladainha or quadra to chulas the choir may follow the mode set up by the soloist or change to its relative major or minor key.

[12] Finally, the range of most songs typically remains within an octave, although some ladainhas can extend beyond this interval. Mestre Traíra, for example, often exceeds an octave in his 1963 ladainhas (Nascimento 1963), reaching up to an eleventh, as in “Tava em Casa.”

Methodology

Hybrid transcription

[13] As we have noted, the interval between the open and raised pitches of a given berimbau can range from a minor second to a major second. Because the gap rarely matches the intervals in 12-TET, measuring, notating, and analyzing these intervals cannot be adequately done using Western-tuned instruments or staff notation alone. Likewise, the pitches of a particular song generally fall in the cracks of standard Western tuning systems (e.g., A440). However, they were easier for us to sing and reproduce because the intervals between sung pitches usually conform to those of Western modes. To analyze these pitches and intervals and notate them, we have implemented Alexander J. Ellis’s cents system, which subdivides the Western semitone into one hundred steps. Each berimbau pitch was then identified and notated according to its closest Western pitch plus or minus a number of cents. For example, we notated the berimbau pitches in Caiçara’s first track of his 1969 album (Moraes 1969) as A2+10 / B2+5. This means that the solto note is an A in the second octave plus 10 cents (that is, one-tenth of a semitone), and the preso note is a B in the second octave plus 5 cents. In this case, the interval between the two notes is 195 cents, just 5 cents shy of a major second. In our analysis of song modes, we provide the central pitch (i.e., the perceived fundamental) using the labels of Western music theory plus or minus the appropriate number of cents. For example, a corrido may be analyzed as being in B+20 major. This means that the home note of that corrido is a B that is sharp by about 20 cents.

[14] However, measuring these pitches and the intervals formed between them cannot be done with our ears alone, nor with an instrument tuned in 12-TET such as the piano. We are fortunate that at the time of doing this research there are various pieces of open-source software that can not only identify pitches with microtonal precision but represent them graphically. Direct automatic transcription of recorded berimbau pitches with existing software is, however, not yet an option because the polyphonic textures are extremely thick (it is nearly impossible for any program to identify individual melodies or the two pitches of a given berimbau) and the quality of the recordings is generally quite poor (there is often too much static and noise) for existing software to isolate a melodic item and transcribe it in any meaningful way. In response to these challenges, we devised a method of hybrid transcription that takes advantage of both our capoeira- and Western-trained ears and the software Singscope, which produces a melogram (i.e., a graphic profile of pitch contour) within a grid that shows time in the horizontal axis and cents in the vertical axis when triggered by melodic sounds. The method consists of the following steps:

Looping and listening to a brief excerpt of the recording in a headset.
Selecting a specific musical item (e.g., a berimbau pitch or a song melody) and focusing our attention on it.
Through repeated listening experiences, internalizing that musical item to the point that we could sing along with it, matching the heard pitch(es).
Singing it back to Singscope for it to produce a melogram with microtonal precision.
Repeating step (4) to produce various melograms of the same musical item to check for consistency and pitch stability in our singing voice.

[15] As we can see, this method is essentially an exercise of human transcription, except that in the last steps (4 and 5) we use specialized software that can discern microtonal nuance. Singscope (or other software of that type) is indispensable in this method because our ears or instruments like the piano alone cannot identify pitches with the required microtonal precision.

[16] For the sake of comparison, consider how we transcribed the capoeira texture in Example 20b, which consists of a ladainha and chulas accompanied by three berimbaus. Here is a summary of the method:

Looping a recording of the excerpt (0:00–0:14) and listening to it repeatedly.
Selecting a musical item (say, the solo vocal melody) and focusing our attention on it.
Internalizing the song melody through repeated listening experiences to the point that we could sing it or play it on the piano, matching the pitches of the recording.
Transcribing the rhythm and pitches of the melody that we had internalized in step 3 in staff paper and then in Finale. To this end, we sang each pitch of the internalized melody out loud and checked them in the piano and then with Finale, which offers the option to sound every pitch that is entered in the score. Because the song melody is modal, our ear training in Western music also helped us identify intervals and relative pitches.
Crosschecking and adjusting the accuracy of our transcription by comparing (note by note) the recorded melody with our own performance of the transcription and ultimately with a MIDI version that Finale generates from the transcribed melody.
Repeating this process for each musical item in the texture (each berimbau, etc.).

[17] Clearly, the transcribed song melody in Example 20b is a visual representation of the transcriber singing or reproducing the melody with a melodic instrument such as the piano or the guitar, not of the data directly from the recording. In human transcription, recorded sounds are therefore inevitably filtered by the human body (i.e., the transcriber’s ears, singing voice, and/or playing fingers). In our method of hybrid transcription, we follow a similar process up to the point of embodiment and externalization of individual pitches and melodies. From this point on, we cannot use an instrument like the piano to identify the pitches that we hear and sing because there are no pianos with 1200 keys between a given pair of octaves. Even if we could match our sung pitches to those on a violin or other unfretted chordophone, our ears alone would still not be able to identify them with the required microtonal precision. Moving forward, we rely on Singscope for the identification and representation of those pitches. Once again, the principle behind our hybrid transcription of capoeira recordings is essentially the same followed in human transcription, in the sense that they both rely on the transcriber’s internalization and externalization of recorded pitches.

[18] We recognize that externalizing our own musical perception of these recordings through our singing voices can be problematic because aural perception is a subjective experience shaped by our individual musical experiences and enculturation. This should not come as a surprise to our readers, as these problems exist with any form of human transcription. We do believe, however, that our extensive formal musical training and our combined three decades of practice in capoeira have made us enculturated listeners and helped us mitigate some of these issues. Because we know and routinely perform the multi-bow patterns heard in the recordings, we can identify the timbral differences between berimbaus well enough to determine which instrument a pitch belongs to. Thus, our biases while listening to berimbau sounds should not be radically different to those of the singers in these recordings.

[19] To temper our perceptual biases, each of us analyzed the recordings independently and we subsequently crosschecked our results. In the few cases when our perceptions differed beyond an acceptable margin of error (more on this below), we repeated the process together until we could reach a consensus. We stress that our perceptual differences rarely varied, and when they did, the margin of difference was minimal. We believe that this system of crosschecking between two long-term capoeira practitioners brings some objectivity to our measurements. Because human-made transcription is inevitably an exercise in musical embodiment, our method of singing back the sounds to an app that measures pitches with high precision is a form of hybrid transcription. The input is filtered through our senses and the output through automated software. This admittedly imperfect solution has allowed us to study a trove of low-quality recordings while we continue to wait for the high-precision automatic transcription machines or software that scholars like Charles Seeger have dreamt about since the early days of ethnomusicology.

[20] While we were highly accurate in maintaining microtonal precision with our singing voices when reproducing single pitches, it was more difficult to do so with sequences of pitches because we tended to instinctively “auto correct” our singing to conform to the interval categories that we are used to. For instance, we noticed that when we sang the two pitches of a given berimbau in succession, our singing voices tended to adjust the interval to either a minor second (100 cents) or a major second (200 cents). However, when we heard and sung each of these pitches separately (i.e., letting five seconds pass between each measurement) with Singscope, the resulting interval we obtained was more variable—the interval was rarely 100 or 200 cents, more commonly falling somewhere in between. By measuring bow pitches separately, we have attempted to represent them more accurately and to address some of the biases we faced as transcribers.

[21] While analyzing song melodies, we had the advantage that capoeira songs conform to Western modes, even if those melodies do not use all scale degrees and even if the pitch collection does not conform to a standard tuning system such as A440. In other words, although the pitches of a given song may fall outside of the 12-TET system, the intervals between those pitches are familiar and easily reproducible for us. Because we are interested in establishing whether there is a relationship between berimbau tuning and song mode rather than determining the exact singing mode, we simplified the singing modes into two categories of major and minor families and focused our measuring efforts on the fundamental. Once we determined the fundamental in each song (almost invariably at phrase endings), we slowed down the excerpt and sang it to back to Singscope sustaining the fundamental for a longer time to make it more easily identifiable in the melogram. To double check that the experience of hearing and singing the surrounding pitches did not affect our perception of the fundamental, we isolated it, looped it and sang it back to Singscope to produce a single-line melogram of it. Later, we highlight the observation that some singers may gradually raise their singing mode. To account for these subtle changes, we measured the fundamental at multiple points in every song, each time isolating the pitch to neutralize possible biases caused by hearing surrounding pitches.

Toques: Emphasized pitches and transcription

[22] We also consider berimbau toques to be important to singers when determining song mode. In the basic form of the various toques we encountered (i.e., when improvisation is not considered) one of the two pitches consistently articulates the downbeat or syncopates it. Because of their central position in the toque’s cycle, we consider these pitches to be emphasized. We hypothesize that these emphasized pitches are particularly present in singers’ musical consciousness and influence their chosen song modes. Example 2 shows an example of a common toque in which the raised pitch, preso, lingers from the second half of beat two, and is tied to the downbeat. In this case preso is the emphasized pitch. We posit that rhythmic placement is more important in determining the emphasized pitch than pitch recurrence or combined duration. In the berimbau toque notated in Example 3, the solto has six attacks with a combined duration of two beats, and the preso has two attacks over one beat. Despite the solto being played more times over a longer combined duration, preso is the emphasized note because it falls on the downbeat, the metric position that singers emphasize.

[23] To make our berimbau transcriptions legible to a wider group of readers, we use two notational systems. The first is a combination of the time unit box system (TUBS) with an onomatopoeic system used by capoeiristas to name berimbau notes. In this system don represents the berimbau’s open note (solto); din is the raised pitch (preso); chi is the buzz note or chiado; and tch is the sound of the caxixi when shaken by itself. The second is inspired by a system used by Kay Shaffer (1977) to notate capoeira toques, which was in turn an adaptation of a notation developed by Luiz D’Anunciação in 1971. The system consists of a single-line staff using the durational figures and time signatures of staff notation. As shown in the legend for notation of Example 2, some noteheads are adapted to represent specific berimbau sounds. Importantly, we understand downbeat placement differently than Shaffer. As can be verified in her transcriptions of the toque de Angola, Shaffer places the downbeat on what we hear as beat two (Shaffer 1977: first page after bibliography). Because song phrasing consistently emphasizes this point, our decision is consistent with the interpretations of most capoeira scholars, including Anunciação (1971, 1990), Lewis (1992), Larrain (2004), Beyer (2004, 2015–16), Murphy (2006), and Galm (2010).^[6]

Example 2. Toque de angola in two notational systems emphasizing the raised pitch.

Example 3. São Bento Repicado toque (Shaffer 1977) emphasizing the raised pitch.

[24] With this method, we study the extent to which capoeira singers attend to berimbau pitches in these recordings and postulate hypotheses about patterns they may have followed. On some occasions the relationships will be obvious; for instance, when a singer consistently tonicizes one of the berimbau pitches. On others, the relationship may be looser or more abstract, such as when a singer switches tonality from one song to the next, or tonicizes pitches different to those heard in the berimbaus.

Tolerance

[25] Through our systematic listening experience of these recordings, which extended across three years, we noticed that when singers use berimbau pitches as a reference, they rarely match bow notes at the microtonal level. Rather, they roam around in close proximity to the pitch in a manner that sounds harmonious, but that is mismatched when measured at high precision. Before discussing how our hybrid transcriptions account for these subtle differences, it is important to understand how the berimbau produces musical tones. In a study of the berimbau’s psychoacoustic properties, physicists Rui Vilão and Santino Melo (2014) demonstrated that the berimbau string vibrates in a highly “inharmonic” way (i.e., its harmonics significantly deviate from the natural harmonic series). This is due to the fact that the frequency spectrum of a note is affected not only by the vibration of the wire, but also by that of the gourd and even of the coin or stone that stops the wire. They suggest that “The gourd is likely to play a fundamental role in the definition of pitch through an amplification of the higher harmonics, in particular the fourth and fifth harmonics” (Vilão and Melo 2014: 1155). More importantly, they show that the loudest overtone in the frequency spectrum of a berimbau tone is not the fundamental, but an upper inharmonic overtone such as the fifth or the eighth harmonic (ibid.). Furthermore, their frequency charts show that, in addition to the loudest, there are other loud harmonics spread throughout the spectrum. The spectrogram in Example 4 shows the frequency spectrum of two berimbau notes, first a solto and then a preso. In both cases we can see that there are multiple loud harmonics accompanying the loudest ones (460Hz or B♭4-23 and 495Hz or B4-4). Ordinarily, the perceived frequency of a tone is determined by the fundamental and reinforced by its harmonic overtones. But as Vilão and Melo (2014) have shown, this is not the case with the berimbau because the instrument’s fundamental is almost entirely absent from its frequency spectrum. The loudest harmonics are rather selected inharmonic overtones which, according to Koulaguina et al. (2015: 922–23), can stand out as “separate distinct sounds” through a process called “harmonic enhancement.” Although during our analysis of the corpus we often heard various “separate distinct” overtones of a given berimbau tone, one of them was always clearly the loudest. We have therefore conceptualized this loudest overtone as a working fundamental because we assume that the recorded musicians heard this same loudest overtone when they were recorded. Because it is possible that the dominant overtones we heard are different to the ones the recorded musicians may have heard due to their relative position during the recording session, limitations of the recording equipment, or hearing loss, we have considered a category called “overtone tonicization” (discussed later).

Example 4. Spectrogram of a berimbau’s two notes (solto on the left and preso on the right). The loudest overtones, perceived as working fundamentals, are circled.

[26] To consider the subtle differences that singers consistently demonstrated when they tonicized a berimbau note, we introduced a degree of tolerance for both our measurement of berimbau and song pitches as well as our analysis of how close the song fundamental needs to be to a bow note to be considered tonicized to that note. When we compared our individual measurements of each bow pitch in the corpus, we noticed that most of the time the widest margin of error was 25 cents in either direction. We therefore decided to set the tolerance for pitch measurements at plus or minus 25 cents—if one of us determined that a bow pitch was A2+40, acceptable pitch assessments for the other would have to fall in the A2+15 and A2+65 range. We also found that most of our independent assessments of song fundamentals were within 25 cents of each other. In the few occasions when the difference between our individual assessments exceeded the 25-cent margin, we repeated the measurement together until we reached consensus—i.e., until our perceptions were less than 25 cents from one another. Since a 50-cent margin is only 4.2% of the 1200 cents of a full octave, the chances that our independent pitch assessments would randomly coincide are quite low. This level of alignment between our perceptions is reassuring considering that studies of pitch-matching accuracy have found that Western trained singers can match a piano note with an average of 28 cents of precision when they are distracted by a chord with distant harmonies in relation to the target note (Estis et al. 2011: 178). In the same study, they found that untrained participants performed that task significantly worse: with an average of 2 semitones (200 cents) of (im)precision (ibid.). There are multiple potentially distracting factors in the analyzed recordings, including the simultaneous and preceding sounds of other berimbaus, each with their own inharmonic spectra; the singing voices, who have different levels of capoeira and formal musical training and therefore sing with different degrees of tonal precision; and the sounds of percussion instruments like the pandeiro and the agogô bell, which multiply inharmonicity. Given these extreme conditions of potential distraction, a 25-cent margin of tolerance for our independent assessments of pitch accuracy is relatively conservative.

[27] Although most studies in experimental psychology place the just noticeable difference (JND) between two pitches at 10 cents, applying this threshold for berimbau recordings poses various problems. First, JND is typically tested using sinusoidal tones, rather than complex tones with inharmonicities such as those produced by the berimbau. Because the inharmonicities of the berimbau can produce various harmonics that can potentially be heard as competing fundamentals, they are more ambiguous and harder to match. Second, the calculation of JND is based on perception rather than performance. It is one thing to listen to two tones and tell how close they are from each other, and another to match a pitch with the singing voice. As Estis et al. (2011) have demonstrated, the latter will produce, on average, wider variations than the former. Compounded, these two factors prompted us to widen the margin of tolerance between our perception, matching, and singing of recorded berimbau pitches beyond a 10-cent threshold. Our method of hybrid transcription not only investigates recorded singers matching bow notes with their voice, but also ours as transcribers. Because the recorded singers tonicized bow pitches under the same distracting conditions explained above, we used the same tolerance of 25 cents in either direction for the measurement of tonicization—if the bow pitch we found was, for instance, A2+50, we considered any singing mode with a fundamental between A2+25 and A2+75 (25 cents either way) as directly tonicized to that pitch. Our system has two advantages over 12-TET. First, it allows for precise location of berimbau pitches and song tonalities—while 12-TET polarizes all pitches and modes into 12 discrete steps within an octave, our system allows for 240 locations, resulting from moving across the octave in steps of 5 cents each. Second, it adds precision to the measurement of tonicization by narrowing the margin of tolerance from 100 to 50 cents. Theoretically, in 12-TET any song mode with a fundamental between A2-50 and A2+49 will have to be polarized to A2; in our system only song modes with fundamentals between A2-25 and A2+25 count as tonicized to A2.

[28] Finally, we assume that pitches one or more octaves apart from each other are musically equivalent because their closely related overtones make them sound essentially the same (Babbit 1965). Although inharmonicity, such as that of the berimbau, produces deviations of octave perception, those deviations rarely exceed 3%, especially between adjacent octaves (Burns and Ward 1978: 456). Therefore, the bow pitch G2, for instance, may facilitate tonicized singing in G major in any octave. Likewise, berimbau pitches F2 and F3 may prompt a singer to sing in F3 minor. Since the berimbau pitches we heard sit in the E♭2 – F3 range, low for our comfortable vocal registers, when transcribing these pitches we sang them one octave above the actual pitches. For this reason, all melograms need to be adjusted by lowering pitches one octave, which is technically the same given said octave equivalence. For our analysis, we produced melograms of each musical item in the body of recordings, including songs and berimbau pitches. We selected a few of these melograms for illustration in our analytical section and supplemented them with staff-notation realizations to make it easier for our readers to interpret this data at a macro level.

[29] Although the transcriptions we present in this article are rarely, if ever, used or discussed by players during rehearsal or performance, this does not mean that practitioners do not care about tuning issues or that they do not have ways to determine and achieve specific tuning and singing results. Most capoeira students can attest that their mestres are meticulous when tuning berimbaus before a performance and fastidious about letting others touch them. Our goal is to investigate how those tuning practices were developed during the seminal period between the 1940s and 1970s, when the first capoeira recordings were created. Because we cannot ask the musicians from this period the questions we would have liked to, in this article we rely on the tools of musical analysis and our own experience as capoeira practitioners. Ethnographic studies with contemporary practitioners may reveal whether the patterns we found in our musical analysis are part of implicit or explicit aesthetics.

Scope and Repertoire

[30] While this article covers sources from 1937 to 1978, our focused analysis begins in 1940, when the first capoeira recordings appeared. Our sources can be grouped into four categories: 1) field recordings made by foreign researchers; 2) commercial albums made by Brazilian capoeira mestres themselves; 3) audiovisual sources, including videos, movies, and documentary films; and 4) printed sources, including textual descriptions and musical transcriptions. Of these, our focus will be on the first three; textual sources will mainly be used to contextualize and supplement our analysis. Up to 1969, the corpus of capoeira recordings is scattered, but from the 1970s onwards the practice of recording and releasing albums became commonplace among capoeira groups in Brazil. This led to an explosion of commercial recordings that reflected and promoted musical diversity among practitioners. For this reason, 1969 marks the point where our analysis shifts from being comprehensive to being selective. To study some of the musical trends in the later period, we selected four recordings from the 1970s, which we take as emblematic of the tuning and singing practices and experimentations of the period. Lastly, in some of these recordings capoeira singers also recorded songs from other Afro-Brazilian genres accompanied by the berimbau. These genres include samba de roda, maculelê, and the music that accompanies Candomblé and Caboclô religious ceremonies.^[7] Because these songs also provide information about capoeiristas’ tonicization practices, we have included them in our analysis.

Table 1. Summary of capoeira recordings (1940–78) showing number of songs with one, two, three, or four berimbaus.

[31] The beginning date of our corpus is marked by the second Afro-Brazilian Congress, which took place in Salvador, Bahia, in 1937. Two documents from this event survived: a silent video of a capoeira demonstration and a musical transcription of a capoeira texture by classical Brazilian composer Camargo Guarnieri, published in 1946 by Brazilian folklorist Oneyda Alvarenga.^[8] This rare transcription is particularly relevant for our study as it includes both berimbau tuning and the song melody. Eunice Catunda’s (1952) transcription of pitched berimbau patterns and songs from a roda by Mestre Waldemar also supplements our analysis of tonicization practices.^[9] Other contextual sources include books (Carneiro 1936, 1937; Landes 1947; 1955–7; Rego 1968; Shaffer 1977), academic articles (Almeida 1942), and newspaper articles (Calvacanti and Verger 1947; Tavares and Verger 1948; Catunda 1952). These written sources provide information about berimbau construction, dimension, and playing technique that contextualizes our study of tuning practices.

[32] The 1940–69 period was marked by the appearance of eight sets of field recordings: foreign researchers who visited Bahia recorded six of them and the other two are of unknown origin. US African American linguist Lorenzo Dow Turner made the first of these during a seven-month visit to Bahia in 1940–41 (Turner 1940). Turner travelled to Brazil searching for African survivals in the language spoken in Afro-Brazilian practices, particularly Candomblé. He was one of several social scientists from the United States who visited Bahia during the first half of the twentieth century to study vestiges of African cultures.^[10] During his visit, Turner recorded some eighteen hours of music and speech, including an hour of capoeira music with capoeiristas Bimba (Manoel dos Reis Machado), Cabecinha (Fernando Cassiano), and Juvenal Cruz and their respective groups. These were likely recorded at Radio Sociedade, the only recording studio in Salvador at the time (Reis 2010: 93).^[11] Originally deposited at the Archives of Traditional Music at Indiana University, these recordings remained unknown to the community of capoeira practitioners and scholars until the late 1990s when they were discovered by the capoeira researcher and practitioner Vincent Brown.^[12] Turner recorded each of these mestres separately in November and December of 1940 and organized their performances in short tracks of three to six minutes. As can be inferred from the introductory speeches of each track and the lack of background noise typical of capoeira performances, all these recordings were demonstrations for the researcher as opposed to in situ performances. We are aware that the aspects of musical style that can be gleaned from these recordings will not fully reflect how music was performed in the context of the roda. Nonetheless, they provide the only material to study capoeira music from that decade.

[33] The second set of recordings is by the US American anthropologist Anthony Leeds, who visited Bahia in 1951 to conduct fieldwork for his doctoral dissertation (Leeds 1951; Leeds 1957). Although he devoted his research to a different topic (the culture of the cacao economy in Bahia), while in Salvador he recorded songs from various Afro-Brazilian genres, including some 16 minutes of capoeira music with the renowned Mestre Waldemar (Waldemar Rodrigues da Paixão) and other capoeiristas (probably Traíra and Antonio Messias dos Santos). These recordings are distributed in three tracks and also deposited at Indiana University’s Archives of Traditional Music. We were unable to discern whether the recordings were made during an in situ performance or if they were a controlled demonstration for Leeds. The latter seems more plausible because we cannot hear the typical yelling and excitement of a capoeira performance in the background, as happens in field recordings of the same mestre by French ethnologist Simone Dreyfus-Roche in 1955. In 1956, Dreyfus-Roche’s recordings were edited at the Musée de l’Homme in Paris and turned into the first capoeira LP under the name Brésil, Bahia, vol. 2 (Dreyfus-Roche 1956). This LP, which also remained mostly unknown to capoeiristas for the rest of the twentieth century, contains six tracks with a combined duration of some twenty-eight minutes.

[34] The fourth set of recordings were made by an unknown person in 1953 and feature Mestre Bimba and his group (Cunha 2020). Brazilian capoeirista Cristiano Mattos da Cunha (aka Cabeleira) claims to have purchased five 78-rpm discs containing these old recordings through the internet from an unidentified seller based in the Brazilian state of Paraná in 2019.^[13] Apart from the date of December 7, 1953 and Mestre Bimba’s signature written across the cover, there are no accompanying notes. However, examining the mestre’s biography we can infer that the recording was made in Bahia. Da Cunha digitized them, wrote liner notes, and released them as a 10-track CD titled Mestre Bimba e seus Tocadores in 2020. The tracks feature fragments of interviews with Bimba, a couple of instrumental sections, and many songs led by the mestre and accompanied by a berimbau and one or two pandeiros. The recordings were evidently detached from a roda context because each of the nearly three-minute tracks finishes with a cued ending.

[35] The fifth set of field recordings was made by US researcher Janice Marie Smith, who visited Bahia in 1960 and 1961 and recorded songs of various local music and dance traditions, including some fifteen minutes of capoeira music (Smith 1960). During her interview with Mestre Canjiquinha (Washington Bruno da Silva) on August 24, 1960, the mestre, accompanied by other famous capoeiristas of the period such as Bigodinho and Manoel, demonstrated various berimbau toques and sang a brief sequence of ladainha, chulas and corridos, evidently decoupled from physical action. Since 1961, these recordings have been deposited at the Archives of Traditional Music at Indiana University.

[36] The sixth set of field recordings is by Finish journalist and ethnographer Helinä Rautavaara, who interviewed separately Mestre Gato Preto (José Luiz Gabriel), Mestre Waldemar, and Mestre Pastinha (Vicente Ferreira Pastinha) at their respective capoeira schools in Salvador (Rautavaara 1963–4). Rautavaara recorded Gato Preto at the Mirante de Cabalar neighborhood (today Jardim Apipema) on December 31, 1963; Pastinha at the Pelourinho square on July 2, 1964 (but the actual lead singer was his student, Raimundo Natividade), and Waldemar at the Liberdade neighborhood, at some point in the first half of 1964. In addition to these interviews, Rautavaara recorded musical tracks with each of these groups, some of them during live rodas and others as musical demonstrations. These musical excerpts have a combined duration of approximately one hour and twenty-five minutes.

[37] The seventh set of field recordings was made by US ethnomusicologist Ralph Cole Waddey, who recorded capoeirista Arnol Concieção and his group in the Pernambués neighborhood of Salvador on November 11, 1966. The recordings include three spoken tracks plus some 35 minutes of capoeira music distributed over five tracks that seem to combine fragments of a street roda with purely musical demonstrations for the researcher (Waddey 1966).

[38] Lastly, in the late 1960s, an unknown person recorded a live roda of Grupo Bonfim in Rio de Janeiro, featuring Mestres Baiano and Zé Grande, two of the founders of this group (Grupo Bonfim 1994). According to Carioca capoeirista Rouxinol, Mestres André Lace and Cabide attended this roda and told him that it took place in 1969.^[14] Nilson Rossi recovered and remastered these recordings into an 11-track CD called Grupo Bonfim, RJ – Capoeira dos Anos 60s in 1994.^[15] Rouxinol asserts that the remastering process included layering in more singers for the choir. The louder and clearer added choral parts can be easily heard in all tracks but do not alter the original singing mode. Of the eleven tracks, we had access only to the first six.

[39] The 1960s was marked by the appearance of the first commercial capoeira albums. Because the number of albums in this decade is limited and most of them became seminal for subsequent groups of capoeira angola and regional styles, we analyzed all of these recordings. The list includes nine albums: Bimba (Machado 1962), Traíra (Nascimento 1963), Paraná (Santos 1963), Camafeu de Oxossi (Conceição 1967, 1968), Viva Bahia! No2 (Melo 1968), Caiçara (Moraes 1969), Pastinha (1969), and Guimarães do Berimbau (1969). Except for the Viva Bahia! Album, which was recorded live during a stage presentation at a ticketed venue in Salvador (Castro Alves theater), these albums were recorded in a controlled studio environment and feature a combination of instrumental tracks (berimbau toques) and sung tracks (including ladainhas, quadras, chulas, and corridos) with instrumental accompaniment. These first albums set a tradition of capoeira groups recording their own albums, which has continued until today.

[40] The four albums we selected from the 1970s document the musical practices of a more eclectic style of capoeira that emerged in this decade, known as capoeira contemporânea. These include Eu Bahia, featuring Bahian Mestre Onias Comenda (Comenda and Murundelê 1972); the album Quilombo – Jongo Basam & Capoeira Angola, featuring Mestre Carioca (Júlio César Figueiro) from Rio de Janeiro (Figueiro and Monteiro 1976); and two well-known albums by Mestres Suassuna (Reinaldo Ramos Suassuna) and Dirceu with the group Cordão de Ouro (Suassuna and Dirceu 1975, 1978), which were based in São Paulo at the time. These albums include tracks of traditional capoeira songs with musical arrangements of songs combining traditional capoeira instruments with melodic/harmonic instruments such as guitars, bass guitars, and flutes. These tonal instruments reveal some of the ways in which practitioners understood the relationship between berimbau tuning, song melody, and tonal harmony.

[41] Lastly, we included two movies and three documentary films featuring capoeira songs accompanied by berimbau. The two movies are Anselmo Duarte’s Pagador de Promesas (1962) and Glauber Rocha’s Barravento (1962), both featuring the singing voice of Mestre Canjiquinha. The three documentary films are Alexandre Robatto Filho’s Vadiação (1954), Pierre Kast’s La Capoeira de Mestre Bimba (1968), and Jair Moura’s Dança da Guerra (1968), where we hear the voices of Mestre Bimba and other famous Bahian mestres. Connected to his documentary, in the same year Moura released a soundtrack album lasting nearly an hour and featuring the singers seen in the documentary (Bimba, Tiburcinho, Totonho de Marê, and Noronha). For our analysis we focused on this soundtrack rather than on the short musical excerpts from the documentary.^[16] Although other TV shows, documentaries, and films from the period featured capoeira and/or the berimbau, their soundtracks are less useful for this study because they do not have any capoeira sounds (e.g., Araujo 1950),^[17] berimbaus can be heard but do not accompany songs (e.g., Santos 1960), the music is recycled from other known recordings (e.g., Peter 1960),^[18] or the capoeira musical fragments are too short or too stylized (e.g., Fontoura 1977). These latter materials are not discussed in this paper.

Analysis

[42] In our analysis, we systematically reviewed berimbau tuning and song mode in 448 song items contained in eight sets of field recordings, thirteen commercial albums, two movies, and three documentary films of the 1940–78 period (see Table 1). Appendix A shows the results of this analysis, including the toque and tuning of each discernible berimbau, name and mode of each song, and the changes to that mode when applicable. In this section, we select representative examples to illustrate our method of analysis and the patterns we have uncovered across this vast corpus.

Tonicizing the emphasized pitch in Bimba’s “Sim Sim Sim” (1962)

[43] The first and most obvious correspondence between berimbau tuning and song mode we found is when the singer uses the emphasized pitch of a single berimbau as the song’s fundamental. One of the most representative examples of this pattern is found on Mestre Bimba’s 1962 album (Machado 1962). With a bateria composed of one berimbau (played by Bimba) and two pandeiros (likely played by two of his students), this album canonized not only the instrumental ensemble of the capoeira regional style, but also its musical aesthetics, including ways of singing and song repertoire. Furthermore, being the first commercial capoeira recording led by a mestre, this album also pioneered a way of recording capoeira music featuring both sung and purely instrumental tracks. The instrumental tracks typically feature berimbau toques, and the sung tracks a combination of introductory songs (ladainhas or quadras with their subsequent chulas) and songs for the physical game (corridos). As seen in Appendix A, sides A and B of Bimba’s 1962 album are respectively devoted to toques and songs. For this analytical demonstration we will focus on the first corrido of side B’s second track, “Sim Sim Sim.”

[44] The berimbau used in this track is tuned to F2+40 / G2 and the toque, known among capoeira regional groups as toque de regional, emphasizes the raised note (G) by consistently articulating it on the downbeat (see Examples 5a and 5b; the pitches in Example 5a and all subsequent berimbau melograms are one octave above the actual pitch).

Example 5a. Berimbau tuning in “Sim Sim Sim,” Bimba (Machado 1962), track B2 (0:07–0:53).

Example 5b. Berimbau toque in “Sim Sim Sim,” Bimba (Machado 1962), track B2 (0:07–0:53).

[45] As we can see in Appendix A, Bimba sings all seven corridos of this track, approximately in G major, with minor oscillations between G and G+10 major. We infer that Bimba is therefore tonicizing the berimbau’s raised pitch (G2), which is the pitch emphasized in the toque. This is particularly apparent in “Sim Sim Sim,” which is in G major as shown in Examples 6a and 6b.

[46] Likewise, the previous track, B1, features a slightly sharper berimbau tuning (F♯2 / G2+50). Bimba accordingly sings the six quadras on this track slightly sharper, between G+25 and G+75 major. This further confirms that Bimba subtly adjusts his singing mode to tonicize the berimbau’s emphasized note.

Example 6a. Melogram of corrido “Sim Sim Sim,” Bimba (Machado 1962), track B2 (0:07–0:53).

Example 6b. Transcription of “Sim Sim Sim,” Bimba (Machado 1962), track B2 (0:07–0:53).

[47] We thus conclude that Bimba hears the emphasized berimbau note (the preso, in this case) as the fundamental for “Sim Sim Sim” and all the corridos and quadras on this album.

Tonicizing the non-emphasized pitch in Juvenal’s “Sim Sim Sim” (1940)

[48] Sometimes, mestres will tonicize to the non-emphasized pitch in the berimbau toque. We turn to another example of “Sim Sim Sim,” this time performed by Juvenal in Turner’s field recording from 1940. Interestingly, Juvenal introduces “Sim Sim Sim” as a song dedicated to Cabôclo entities in Afro-religious ceremonies. This early recording of this popular capoeira song confirms the circulation of musical items between capoeira and Afro-Brazilian religions (Diniz 2010). Throughout the track, a single berimbau plays with an E♭2+30 solto and F2 preso pitch. While the berimbau is accompanied by a pandeiro, the track features no choir, which is atypical of most capoeira recordings—Juvenal sings both the call and response parts of the song. Although Juvenal plays various berimbau patterns and variations in the instrumental sections, he uses a single toque to accompany the song in this and all his other tracks. As shown in Example 7b, that toque emphasizes the open berimbau note. While the pitch remains constant throughout the song, the tempo increases considerably.

Example 7a. Berimbau tuning in “Sim Sim Sim,” by Juvenal in Turner (1940), track 2 (0:20–0:30).

Example 7b. Berimbau toque in “Sim Sim Sim,” by Juvenal in Turner (1940), track 2 (0:20–0:30).

[49] As shown in Examples 8a and 8b, in this track Juvenal sings in F major, meaning that he tonicizes the open note of the berimbau, the pitch that is non-emphasized in the toque.

Example 8a. Melogram of “Sim Sim Sim,” by Juvenal in Turner (1940), track 2 (0:20–0:30).

Example 8b. Transcription of “Sim Sim Sim,” by Juvenal in Turner (1940), track 2 (0:20–0:30).

Tonicizing the emphasized pitch and its relative minor in Bimba’s (1962) “Valha me Nossa Senhora”

[50] Often connected by the word camará or a similar expression (camarada or camaradinha), the chulas function in the roda as a tag to the introductory solo song (ladainhas or quadras). It is therefore not surprising that transitions between the two are usually smooth and that the singers often maintain in the chulas the same mode used for the ladainha or quadra. There are, however, instances when singers seem to modulate to the relative major or relative minor when they switch to the chulas. Because some of the mestres who sing in this fashion tonicize a berimbau pitch during the ladainha/quadra, we argue that the same bow note functions as an indirect reference for the chulas, as its relative major or relative minor. Mestre Bimba, the most representative case, sings the six quadras of his track B1 (Machado 1962) approximately in G+50 major and consistently sings the chulas following each quadra approximately in E+50 minor. Therefore, from a purely analytical standpoint, we could say that he consistently modulates from a major mode to its relative minor in his quadra-chulas transitions.^[19] The first quadra of that track, “Quem Foi seu Mestre,” illustrates this process. As shown in Examples 9a and 9b, Bimba sings the quadra approximately in G+70 major but modulates to E+60 minor in the chulas.

Example 9a. Melograms of quadra “Quem Foi seu Mestre” and chulas, Bimba (Machado 1962), track B1 (0:27–0:39).

Example 9b. Transcription of quadra “Quem Foi seu Mestre” and chulas, Bimba (Machado 1962), track B1 (0:27–0:39). Sung pitches are about 70 cents sharper than notated in mm. 1–7 and approximately 60 cents sharper in mm. 8–11.

[51] The single berimbau heard in this track is tuned in F♯2-5 for solto and G2+50 for preso and plays a toque that emphasizes the raised pitch (see Examples 10a and 10b). In terms of the relationship between bow tuning and singing mode we see a complex case where, on the one hand, Bimba directly tonicizes the emphasized note while singing the quadra, as he does in the corrido “Sim Sim Sim” of track B2 that we discussed above. On the other, he modulates to E+60 minor in the chulas that follow, approximating the relative minor of the note emphasized by the toque.

Example 10a. Berimbau tuning of quadra “Quem Foi seu Mestre” and chulas, Bimba (Machado 1962), track B1 (0:27–0:39).

Example 10b. Toque of quadra “Quem Foi seu Mestre” and chulas, Bimba (Machado 1962), track A1 (0:27–0:39).

Tuning the berimbau a fifth above the song fundamental in Onias Comenda’s (1972) “Tava lá em Casa”

[52] There are other cases where singers seem to tonicize not the berimbau’s perceived fundamental but other scale degrees. Track B6 of the album Eu Bahia (Comenda and Murundelê 1972), sung by Onias Comenda, is a clear example. Unlike capoeira albums from the 1960s, in this recording Comenda combines traditional and non-conventional instruments and ways of singing. On the one hand, the ensemble includes a berimbau, atabaque, and agogô accompanying the well-known ladainha “Tava lá em Casa,” followed by chulas, all with melodies that any capoeirista would easily recognize. On the other, the ensemble is joined by a guitar that anchors the song’s tonality in F minor, and the various verses of the ladainha are first sung by the soloist and then repeated by the choir, something that is never done in traditional capoeira practice. In our analysis, we will focus on the first two lines of the ladainha: “Tava lá em casa, oiaiá, sem pensar nem imaginar.” In them, the guitar plays an ostinato alternating the chords E♭ major and F minor and aligned with a berimbau rhythm, which in turn alternates the pitches B♭2+20 and C3-20 (see Example 11a). As shown in Example 11b, these ostinati emphasize the guitar’s F minor chord and the raised berimbau note (C3-20).

Example 11a. Berimbau tuning in ladainha “Tava lá em Casa,” Eu Bahia (Comenda and Murundelê 1972), track B6 (0:16–0:22).

Example 11b. Berimbau toque and guitar accompaniment in “Tava lá em Casa,” Eu Bahia (Comenda and Murundelê 1972), track B6 (0:16–0:22).

[53] It is unsurprising that Comenda sang the ladainha in F minor because this is the tonality established by the guitar (see Examples 12a and 12b). Interestingly, the group decided to tune the emphasized berimbau pitch not to the song’s fundamental (F), but approximately to its fifth scale degree (C3-20). Given that this is a stylized musical arrangement, there is every reason to believe that this tuning was deliberate rather than coincidental.

Example 12a. Melogram of ladainha “Tava lá em Casa,” Eu Bahia (Comenda and Murundelê 1972), track B6 (0:16–0:22).

Example 12b. Transcription of ladainha “Tava lá em Casa,” Eu Bahia (Comenda and Murundelê 1972), track B6 (0:16–0:22).

Gradual mode migration in Caiçara’s (1969) “Tava em Casa” and Suassuna (1978)

[54] As expected, not all mestres maintain the same singing mode over time, even if berimbau tuning is constant. For instance, some singers may leap from one mode to another while changing songs, as happens in track 2 of the album Viva Bahia! (Melo 1968), where the singer suddenly moves from Ab-10 major in the corrido “Dá no Negô” to C-10 major in “O Besouro Preto,” and back to Ab major in “Quem Vem Lá.” As shown in Table 2, these two modal leaps occur within less than 40 seconds. Because the open note of the sole berimbau that accompanies the songs is A♭2-20 (which is also the note emphasized in the toque), we interpret the leap to C-10 major as a momentary deviation from the tonicization of the emphasized bow note.

Table 2. Contents, song mode, and berimbau tuning of Viva Bahia! (Melo 1968), track B2 (0:00–4:40).

[55] These sudden wide leaps are rare in capoeira.^[20] More commonly, a singer would shift the modal center gradually. The most common direction of mode migration in capoeira recordings is upwards, which is consistent with an aesthetic in capoeira performance that strives to increase intensity over time—as the singing mode moves upwards in pitch, so does the volume of the choir, who are prompted to respond in the upper register. In this section, we provide two examples of this modal migration.

[56] The first is in Caiçara’s track A1 (Moraes 1969), where he starts singing the ladainha “Tava em Casa” in B♭+20 major and finishes it in B major (see Example 13). As shown in Examples 14a and 14b, this song is accompanied by a single berimbau tuned at A2+10 for solto and B2+5 for preso playing the toque de angola, meaning that it emphasizes the preso note. We contend that the mestre gradually raised the pitch of the mode to better match the berimbau’s emphasized pitch (B2+5).

Example 13. Melogram of ladainha “Tava em Casa,” Caiçara (Moraes 1969), track 1A. First verse (0:35–0:40) above; last verse (1:29–1:35) below.

[57] Our second example is heard on Suassuna’s track A3 (Suassuna and Dirceu 1978), which represents a much more radical case of mode migration, spread over eight songs sung in about seven minutes. As shown in Table 3, the first gradual migration occurs in the ladainha, which starts in F+50 minor and ends F♯ minor. The beginning of the chulas features a modulation to the relative major (A-15 major), like those we heard in other recordings (e.g., Bimba [Machado 1962], track B1). From this point on, the singer begins a long and steady climb that culminates in B+40 major, five minutes and twenty-two seconds later, a gradual mode migration of a major second plus 55 cents! The relationship of this migration to berimbau tuning is less obvious than it is in Caiçara’s album. The two berimbaus in this track are tuned at F2-10 / F♯2+10 and E2+30 / F♯2-5 and the notes emphasized by the toque are the two soltos. We could then argue that throughout the ladainha the singer gradually moved towards the raised notes of these berimbaus (around F♯), meaning that he tonicized the non-emphasized note. However, after he has modulated to the relative major (A-15) in the chulas, the singer migrates away from the relative major of the berimbaus’ raised note. Clearly, singers may gradually migrate either towards or away from the tonicized bow pitch.

Table 3. Content, song mode, and berimbau tuning of Suassuna and Dirceu (1978), track A3.

[58] Microtonal rising in vocal music has been widely documented. Anthony Seeger (2004: 94), for instance, wrote that a rise in absolute pitch is common among various kinds of songs sung by the Suyá people in the Brazilian Amazon rainforest. In a chapter devoted to this phenomenon in his iconic book, we can see gradual microtonal rising ranging from a few hundred cents up to a fourth (Seeger 2004: 92, 94). While Seeger presents microtonal rising in Suyá songs as a “mystery” that cannot be explained through indigenous discourse, he seems to understand it as a widespread phenomenon in choral groups, including Euro-American choirs and Native American groups. He problematizes the assumption that gradual microtonal pitch variation necessarily results from the absence of “fixed-pitched instruments such as piano, organ, or woodwinds” by citing reports of Native American singers who deliberately raise the pitch of certain songs accompanied by melodic instruments (Seeger 2004: 93). He admits that “the absence of fix-pitched instruments made it quite possible that the phenomenon [microtonal change] was not an object of attention to them [the Suyá]” because flutes, for example, would have made pitch variations evident (ibid.). Although we do not have the kind of definitive testimonies that Frances Densmore (1956: 212) got from one of her Seminole informants, who asserted that he was taught by his grandfather to raise the pitch only of old war songs, our analysis of Caiçara’s ladainha “Tava em Casa” suggests that he and other capoeira singers may subtly raise tonality to converge towards a berimbau pitch. However, the analysis of the songs on track A3 of Suassuna’s 1978 album suggests the opposite: that singers may move away from a tonicized bow note. In this case, tonal considerations possibly give way to the abovementioned aesthetic of increasing intensity in the performance. This means that fixed-pitch instruments can coexist with microtonal rises in song pitch regardless of tonal awareness. Therefore, microtonal rising in capoeira does not imply that singers treat the berimbau as an unpitched percussion instrument but may instead be the singers’ response to tonal considerations and/or to capoeira’s aesthetics of intensity increase.

Full and partial adjacent tuning in Caiçara (1969), Waldemar (1951), and Traíra (1963)

[59] In our analysis, we observed various approaches to how mestres might tune their berimbaus in relation to one another. One common approach is when the preso note of one berimbau approximately matches the solto of the berimbau immediately above it in pitch. To our knowledge, practitioners do not have a name for this way of tuning. We have called it adjacent tuning. A good example is found in the case of the two berimbaus featured in track A2 of Caiçara’s 1969 album. As shown in Example 14a, the interval between the preso of the lower berimbau and the solto of the higher bow is quite small (30 cents). Although technically this gap is wider than the 25-cent margin of tolerance we use in this study, we consider it narrow enough for the mestre to have heard the two pitches as a near-unison. In Example 14b we see that the mestre voices the berimbaus in a reversed manner—the open note of the low berimbau corresponds to the raised note of the high berimbau and vice versa. Adjacent tuning makes sense within reversed voicing, because the berimbaus move from approximate unison to a wider interval that may range between a major second and a major third. In Caiçara’s case, the wide interval (between the low solto and high preso) is a major third minus 5 cents. Because Caiçara sings the five songs of this track approximately in A major (with subtle fluctuations between A-10 major and A+5 major), we could say that the berimbaus provide not only the fundamental, but also the second component of the A major triad (C♯3-10), which sets up the mode for the song. We believe that the mestre prioritized forming a major third with the two most distant berimbau pitches (A2-5 and C♯3-10) over creating a perfect unison with the two internal notes. This may explain the slightly wide gap of 30 cents between the low berimbau’s preso and the high berimbau’s solto.

Example 14a. Melograms of berimbau tunings in Caiçara (Moraes 1969), track A2.

Example 14b. Transcription of berimbaus’ toques in Caiçara (Moraes 1969), track A2.

[60] Adjacent tuning can also occur with three bows. In track 2 of Leeds’s 1951 field recording of Waldemar, for example, the gunga’s raised pitch (F♯2+30), nearly matches the medio’s open note (F♯2+50), and the medio’s raised pitch (G♯2+30) approximates the open note of the viola (G♯2+5) (see Example 15). Interestingly, the fundamental of Waldemar’s singing mode during the opening ladainha (which migrates from G minor to G+50 minor) sits in the middle of these three adjacently tuned berimbaus, but it does not match any of the bow pitches with the tolerance of 25 cents above or below that we have used in this study. We thus conceptualize it as “clashing” or “rubbing against” them, especially the F♯+50 and G♯+30 of the médio.

Example 15. Melograms of full adjacent berimbau tuning in Leeds’s 1951 recording of Waldemar, track 2.

[61] While Waldemar’s three berimbaus feature full adjacent tuning, more commonly we found examples of partial adjacent tuning, in which two of the three berimbaus are adjacent but the third is not. Track A1 of Traíra’s 1963 album (Nascimento 1963) illustrates this case: the two lower berimbaus are adjacently tuned, but the upper berimbau is not (see Example 16).

Example 16. Melograms of partial adjacent berimbau tuning in Traíra (Nascimento 1963), Track A1.

Chordal tuning of berimbaus in Guarnieri’s transcription (1937/46) and Traíra (1963)

[62] Camargo Guarnieri’s transcription of the capoeira demonstration at the Second Afro-Brazilian Congress in Salvador in January 1937 (but published in 1946 by Alvarenga), is the first known example of a transcription of berimbau pitches and rhythms in combination with song. It shows a chordal tuning where the open notes of the three berimbaus (G, C, and E) form a C major triad in second inversion and the raised notes move up diatonically by a second (A, D, F) to form a D minor triad, also in second inversion (see Example 17). According to Guarnieri, the chord formed by the open notes establishes the tonality for the singer, who sings the corrido “Esta Cobra te Morde” in C major. Guarnieri’s claim has some credibility because his transcription includes the three bows that appear in the pictures and footage of the 1937 exhibition, but also because the melody of the corrido is quite similar to the way in which it is sung in subsequent recordings and even nowadays. Unfortunately, we do not have access to the audio of this exhibition to verify the bow pitches and song mode in his transcription. In fact, after having examined comprehensively the audio and video recordings of capoeira music in the 1940–69 period, we find it improbable. Firstly, Guarnieri’s transcriptions of berimbau tuning suggest that all the bow pitches and the intervals formed by them can be represented with the 12-TET system. This is almost never the case in the recordings we have analyzed—pitches or intervals fitting 12-TET are the exception, not the rule. Secondly, Guarnieri shows the three berimbaus moving always in parallel. As shown in Appendix A, in most ensembles formed by three bows, at least two berimbaus move in opposite directions. For these reasons, we are hesitant to take Guarnieri’s transcription at face value as representative of a musical aesthetic followed by Bahian capoeiristas in the late 1930s. His score may be the result of adaptations made by his Western ear to the complex berimbau tunings and song modes he heard.

Example 17. Camargo Guarnieri’s transcription of the corrido “Esta Cobra te Morde” in Alvarenga (1946: 250). The pandeiro rhythm is misaligned by one quarter note.

[63] Interestingly, Brazilian composer Eunice Catunda, who was the only other author from the 1937–78 period who transcribed berimbau pitches and song mode, also suggests chordal tuning in her transcriptions. In an article published in 1952 and documenting her visit to Waldemar’s famous roda at the Liberdade neighborhood in Salvador, Catunda (1952: 17) transcribed the pitches of the two berimbaus she heard as E♭ / F and G / A♭ (see Example 18). Separately, but suggesting that the same bows accompany the melody, she transcribed the corrido “Eh Paraná” in E♭ major (see Example 18). Like Guarnieri, Catunda suggests that the open berimbau notes set up the singing mode chordally—E♭ and G form a dyad that sets up the E♭ major mode used by the singer. Although Catunda’s transcription avoids the unlikely parallel bow movement of Guarnieri’s transcription, we are still skeptical of it because it approximates all pitches and modes to 12-TET.

Example 18. Eunice Catunda’s transcription of berimbaus and the corrido “Eh Paraná” from Waldemar’s roda in Salvador (Catunda 1952: 17).

[64] There are, however, a few examples of chordal tuning in the audio recordings we examined. As noted in the previous section, Caiçara tunes two berimbaus adjacently, forming a major third between the open note of the low bow (A2-5) and the raised pitch of the higher berimbau (C♯2-10) (see Example 14a). Caiçara sings in A major, which is set up by this dyad. But chordal tuning can also occur with three berimbaus. Mestre Traíra’s track A1 (Nascimento 1963), for instance, features three berimbaus with partial adjacent tuning where the open note of the lowest berimbau (F2-10), the raised note of the medium-pitched berimbau (A♭+5), and the raised pitch of the high berimbau (C3+10) approximately form an F minor chord. (Examples 16 and 19 respectively show berimbau tuning and toques on this track.) As shown in Examples 20a and 20b, Traíra sings the ladainha and chulas of this track around F minor (with subtle oscillations between F+5 minor and F-25 minor).

Example 19. Berimbau toques in Traíra (Nascimento 1963), track A1.

Example 20a. Melogram of ladainha “Tava em Casa” and chulas by Traíra (Nascimento 1963), track A1 (2:00–2:12).

Example 20b. Transcription of ladainha “Tava em Casa” by Traíra (Nascimento 1963), track A1 (1:55–2:12).

[65] There are important similarities and differences between the berimbau tuning shown in Guarnieri’s transcription and that heard in Traíra’s track A1. Both feature chordal tuning and the tonic of the chord sets up the mode for singing. However, the manner of voicing the berimbaus to obtain those chords is different: in Guarnieri’s transcription, berimbaus play the same version of the toque in parallel to form tonic and supertonic chords. In Traíra’s case, the berimbaus do not form a chord in every metric position because some of them need to move in contrary motion to form it (see Example 19), and the minor third is approximate, unlike Guarnieri’s precise transcription in 12-TET. As can be seen in Appendix A, contrary motion is typical in ensembles with three berimbaus, not parallel motion as suggested by Guarnieri.^[21]

Singing without obvious relation to berimbau pitches in Pastinha’s (1969) “Eu Nasci pra Capoeira”

[66] In contrast to the multiple ways in which mestres creatively use berimbau pitches to guide their singing, we found many instances where singers seem to disregard bow notes, or perhaps relate to them in ways that are beyond our current understanding. As shown in Appendix A, examples abound and are chronologically scattered: From the field recordings of the 1940s and 1950s to albums and films from the 1960s and 1970s. Interestingly, Mestre Pastinha, who is broadly regarded as a “guardian” of the capoeira angola tradition and thus considered a model in performance aesthetics, appears to be one of the singers who more consistently sings in this manner. The opening of each of the five tracks of his 1969 album (one of the rare examples where his singing voice was recorded), a single berimbau accompanies Pastinha singing a short ladainha. After the introductory ladainha, the rest of the bateria (consisting of two or three additional berimbaus, agogô, one or two pandeiros, reco-reco, and possibly an atabaque) joins in while one of Pastinha’s students leads a sequence of ladainha, chulas, and corridos. Two of his students take on this leading role: Waldomiro Malvadeza (in tracks A1, A2, B1, and B2) and Raimundo Pequeno (in track A3). As shown in Appendix A, Pastinha sings four out of five ladainhas in a tonality unrelated to the pitches of the accompanying berimbau. The second track on the B side, where he sings the ladainha “Eu Nasci pra Capoeira,” is illustrative. As shown in Examples 21a and 22b, the opening berimbau accompanying the mestre’s song on this track is tuned at C3-5 for solto and C♯3 for preso. Surprisingly, the mestre sings in G♭+40 major, which is completely unrelated to both bow pitches (see Examples 22a and 22b).

Example 21a. Berimbau tuning of ladainha “Eu Nasci pra Capoeira,” Pastinha (1969), track B2 (0:25–0:38).

Example 21b. Berimbau toque in ladainha “Eu Nasci pra Capoeira,” Pastinha (1969), track B2 (0:25–0:38).

Example 22a. Melogram of ladainha “Eu Nasci pra Capoeira,” Pastinha (1969), track B2 (0:25–0:38).

Example 22b. Transcription of ladainha “Eu Nasci pra Capoeira,” Pastinha (1969), track B2 (0:25–0:38). All sung pitches are about 40 cents sharper than notated.

[67] Interestingly, the approach of another singer on this album, Waldomiro Malvadeza, is varied, sometimes tonicizing one of the bow pitches, and at other times not. After Pastinha’s “Eu Nasci pra Capoeira,” for instance, Malvadeza takes over to sing the ladainha “Eu Estava em Casa” followed by a series of chulas and corridos. As shown in Appendix A, he sings all these songs in around B major (with minor oscillations between B-20 major and B+10 major). This mode is also unrelated to both the initial berimbau (again, tuned at C3-5 / C♯3) and the second berimbau that enters once Pastinha finishes his ladainha (tuned at A♭2+30 / B♭2-15).

Listening to overtones

[68] As discussed, many of the overtones produced by the berimbau can sound as separate distinct pitches due to the bow’s inharmonicities. We therefore considered the possibility that some of these overtones (except for the loudest overtone, which we have treated as a working fundamental), may have guided singers. The ladainhas sung by Pastinha in his 1969 album are ideal to explore this potentiality because they feature a single berimbau without percussive accompaniment and are (with one exception), out of tune with regards to the perceived bow’s fundamentals. Indeed, we found that, in one case, a loud bow overtone detected by specialized software in the recording (different to the loudest overtone that we heard directly with our ears) matched the mestre’s central pitch in the ladainha. In track A1, he sings the ladainha “Maior é Deus” in G-20 major, while the berimbau plays the notes B2+55 and C♯3-5—that is, without any obvious tonal relationship between them (see Examples 23 and 24). However, when we looped the open berimbau note, Singscope and Audacity captured various partials, especially 385Hz (G4-31) (see Example 25). Pastinha, who seemingly disregards berimbau tuning in four of the five ladainhas he sings in this album, may have used this partial as the fundamental for “Maior é Deus.”

Example 23. Melogram of ladainha “Maior é Deus” by Pastinha (1969), track A1 (0:10–0:29).

Example 24. Melograms of berimbau tuning in “Maior é Deus” by Pastinha (1969), track A1 (0:10–0:29).

Example 25. Partials of open berimbau note in Pastinha (1969), track A1 (0:00-0:02). Bright yellow represents loudest overtones.

[69] Although we only checked overtone tonicization in tracks with a single berimbau, we also considered this form of tonicization when a single opening berimbau is later joined by more bows, but the singer (or subsequent singers) seems to maintain the tonal relationship to the initial berimbau. We observed this case in the same track of Pastinha’s album analyzed above (A1). After Pastinha sings his ladainha in G-20 major, tonicizing an overtone of the single berimbau that accompanies him, his student Waldomiro Malvadeza takes over by singing the ladainha “Bahia Nossa Bahia” in G-20 major, matching Pastinha’s tonality. Halfway through Malvadeza’s ladainha, three berimbaus join in with different tunings, but they do not seem to affect the singer, who continues singing the remaining of this track approximately in G major (see Table 4). Although, technically, Malvadeza sings most of his songs in this track accompanied by four berimbaus, one of which features the pitch G2-20, we believe that he established the tonality at the beginning of his ladainha following Pastinha’s overtone tonicization. For this reason, we posit that the six songs of this track are guided by the G-31 overtone.

Table 4. Track list, songs, song mode, berimbau tuning and overtones in track A1 (Pastinha 1969)

Discussion

[70] Through our hybrid transcription and analysis of the corpus of recordings from the 1937–78 period, we have discovered a great variety of approaches to how mestres sing in relation to berimbau notes. We have organized those approaches into a taxonomy of five categories, three of them with subcategories. The first approach is when mestres directly tonicize one of the berimbau pitches. To test our hypothesis that toques influence singers, we subdivided this first category into two subcategories: in one, singers tonicize the pitch emphasized by the toque; and in the other, singers tonicize the non-emphasized pitch. The second category is when singers use one of the bow pitches as an indirect reference—the berimbau is not tuned to the song’s fundamental but to a different scale degree. We identified two subcategories: a) when the berimbau pitch is the relative major or minor of the song mode; and b) when the bow note is the fifth above the song’s fundamental. The third category is when berimbau pitches form a triad or a dyad that clearly sets up the singing mode. The fourth is when singers tonicize not the perceived fundamental of a bow but one of its prominent overtones. Finally, the fifth category is when singers seem to disregard bow pitches altogether. This five-tiered taxonomy with its subcategories is presented schematically in Figure 1. Figure 2 exemplifies each of the categories and subcategories. Table 5 summarizes the results of our analysis of 448 songs, presenting the number of songs of each recording that fall into each of these five approaches. The ensuing discussion considers each of these categories and subcategories in turn.

Figure 1. Diagram of taxonomy of relationships between song mode and berimbau tuning.

Figure 2. Examples of taxonomy of relationships between song mode and berimbau tuning.

Table 5. Summary of approaches in capoeira recordings (1940–78).

Category 1: Direct Tonicization of a Berimbau Pitch

[71] Many singers around the world who accompany themselves with a musical bow tonicize one of the bow pitches. Díaz, Assunção, and Beyer (2021: 308, 311–14), for instance, have demonstrated that players of hungo and mbulumbumba musical bows from Angola typically tonicize the open note. Similarly, in the 1964 Symposium of Transcription held by the Society of Ethnomusicology, four transcribers (Garfias, Kolinski, List, and Rhodes) showed that a Hukwe singer accompanied by a musical bow in present-day northern Namibia used the bow notes as fundamentals for different parts of the song (England 1964). Accordingly, most of the capoeira singers we included in this study tonicize one of the berimbau pitches—the exceptions were Paraná (Santos 1963) and Comenda (Comenda and Murundelê 1972). As we can see in Table 5, direct tonicization of a berimbau note is the most common form of tonicization in the analyzed corpus of songs. Of the 294 songs that exhibited some form of tonicization, 180 directly tonicized a bow pitch—that is, about six out of ten.

[72] We hypothesized that singers who tonicize bow notes should exhibit a preference for the note emphasized by the berimbau toque (that is, the note articulating or syncopating the downbeat) because attention is likely to be drawn to this central metrical position. A global count confirms that this is the case: of the 180 songs where mestres tonicized a bow pitch, 127 tonicize the emphasized pitch. This means that if we use the toque as a predictor of what pitch singers tonicize, we will be correct about seven times out of ten.

[73] The phenomenon of gradual mode migration that we documented provides further support to the hypothesis that mestres deliberately tonicize certain berimbau notes. As explained earlier, some mestres may start singing “out of tune” and gradually correct the mode by slowly moving towards a mode whose fundamental coincides with a specific berimbau pitch. As shown in Table 6, the subtle adjustments made by Bimba (Robatto Filho 1954), Waldemar (Dreyfus-Roche 1956), Caiçara (Moraes 1969), and Suassuna (Suassuna and Dirceu 1978) are smaller than a semitone, yet they show that singers use this technique to converge towards a berimbau note. However, we also found instances where mestres started singing in tune with a berimbau note and gradually moved away from it, always upwards. For example, the singer in Suassuna’s 1978 album (track A3) continues gradually raising the pitch of the mode, after converging with a bow note (F♯2-5), thus diverging from this pitch (see Table 3). Although these cases of divergence are far outweighed by those of convergence, we believe that they exist for a good reason: when singers move the mode to higher registers, they do so in order to follow a capoeira aesthetic of increasing intensity over time, which corresponds with louder dynamics, faster tempi, and increased improvisation. More plainly, they prioritized increasing intensity over tonicizing a berimbau note.

Table 6. Examples of gradual mode migration towards a berimbau note.

Category 2: Indirect Tonicization of a Berimbau Note

[74] The second category of tonicization concerns singers who use a bow pitch as an indirect reference; that is, when the berimbau is not tuned to the song’s fundamental but to a different scale degree. We identified two main cases. The first, and more common, is associated with the modulations to the relative minor or major tonalities that frequently occur in the transition between the ladainha (or quadra) and the chulas. In this case, singers directly tonicize a bow pitch during the ladainha or quadra but when they modulate to the relative tonality in the chulas, their relationship to the referenced berimbau pitch becomes indirect. As shown in Table 5, many singers use this form of tonicization, including Comenda (Comenda and Murundelê 1972), Canjiquinha (Smith 1960), Bimba (Cunha 2020 [1953]; Machado 1962), and Camafeu de Oxossi (Conceição 1967).

[75] Modulation from one tonality to its relative major or minor can also happen within the ladainha or from the chulas to the subsequent corridos. The clearest examples of the former are by Caiçara, who sings the first verse of his ladainhas in tracks A3 and A4 of his 1969 album in the minor mode before modulating to the relative major, a tonality that is maintained for the rest of the song and carried over to the remaining songs of the track (see Appendix A). There are also various examples of modulation from the chulas to the corridos. One such example can be heard in tracks 3 and 4 of Cabecinha’s 1940 recording when the singer sings the chulas in the minor mode and shifts to the relative major at the outset of the subsequent corridos (Turner 1940). A more compelling case supporting the idea that singers use berimbau pitches as indirect reference is when singers open a track by indirectly tonicizing a berimbau note and then modulate to the relative tonality to directly tonicize that note. The 1953 field recording of Bimba features two examples on tracks 4 and 9 (Cunha 2020). On track 4, for instance, the mestre starts singing the quadra in the major mode (B♭+20 Maj) and modulates to the relative minor (G+20 min) in the chulas. Because the berimbau is tuned to G♭2-10 / G2+10, we conclude that Bimba first indirectly tonicized the bow’s raised note and then directly tonicized it. Clearly, singers do not need to initiate the quadra-chulas (or ladainha-chulas) sequence by making direct reference to a berimbau pitch.

[76] The second case of indirect tonicization is when the berimbau is tuned a fifth above the song’s tonality. We only found unambiguous evidence of this approach in three songs on Onias Comenda’s 1972 album, where the F minor tonality of all songs on track B6 is clearly established by a guitar but the berimbau is tuned at B♭2+10 / C3-10. Since the arrangements in this non-traditional capoeira album are deliberate, we infer that the raised berimbau note (which is the note emphasized by the toque) is purposefully tuned a fifth above the singing mode.^[22] The tuning and singing practices that capoeiristas use in arrangements including non-traditional instruments that provide harmonic accompaniment such as the guitar should not be taken as representative of traditional capoeira practice. We have included this case of indirect bow tonicization from Comenda in order to consider the possibility that some mestres may choose to sing in one tonality while tuning a berimbau note a fifth above the fundamental. In Appendix A we can see that various mestres may have used this relationship between bow tuning and song mode. For instance, in Turner’s 1940 recordings of Bimba, the mestre sings the eight songs of the third track approximately in C major, while the raised note of the loudest berimbau is G2.

[77] These forms of indirect tonicization are not surprising when we consider that singers typically start songs from the tonic, third, or fifth scale degrees. It is possible that some singers may intentionally match the first sung note to one of the bow pitches, thus resulting in either direct or indirect tonicization. After direct tonicization, indirect tonicization is the second most common approach we identified—of the 294 songs where singers used some form of tonicization, 49 used indirect tonicization; that is, 17 percent. While indirect tonicization has an important role in (or is a consequence of) the transition between songs, especially in the ladainha (or quadra)–chulas sequence, it may also allow singers to add melodic variety or move to a desired register.

Category 3: Chordal Tonicization

[78] The third category is when players tune berimbaus forming a dyad or triad that establishes a tonality for songs. The earliest example suggesting this approach is from Camargo Guarnieri’s transcription of the multi-bow ensemble led by Samuel Querido de Deus that accompanied a capoeira demonstration at the Second Afro-Brazilian Congress in Salvador, Bahia in 1937. Eunice Catunda provided another transcription suggesting chordal tonicization from her visit to Waldemar’s roda in Salvador in 1952. Earlier, we explained that we are skeptical of these two transcriptions because the tuning does not consider microtonal intervals and the parallel movement of the berimbaus in Guarnieri’s transcription is atypical of capoeira ensembles with three berimbaus. Guarnieri’s and Catunda’s transcriptions likely reflect the microtonal adjustments they made to force the capoeira sounds they heard into the 12-TET system that they had internalized during their training as pianists and composers in Western classical music. Yet, we found twenty songs in the corpus where capoeiristas clearly tuned berimbaus chordally. These can be divided into two cases. The first is when two berimbau pitches (the first and third scale degrees) form a dyad that establishes the tonality. The second is when three berimbau pitches form a complete triad (first, third, and fifth degrees) that serves as the song’s tonal center.

[79] Chordal tonicization can be seen as an elaborate form of direct tonicization, where singers not only tonicize the song’s fundamental, but also the third and, sometimes, the fifth. Because there are many more songs accompanied by ensembles with two berimbaus than with three (168 versus 69) and because coordinating two berimbau pitches is easier than coordinating three, it is not surprising that there are more cases of chordal tonicization with dyads than with full triads. Of the 19 songs that used chordal tonicization, 13 were tonicized to a dyad and 6 to a triad. Additionally, adjacent tuning of two berimbaus (whether they are part of a two-berimbau ensemble or are the two bottom bows of a larger multi-bow ensemble), which is relatively common among practitioners (e.g., Traíra [Nascimento 1963]; Natividade [Rautavaara 1963–4]; Zé Grande [Grupo Bonfim 1994 (1969)]; Caiçara [Moraes 1969]; Suassuna [Suassuna and Dirceu 1975]; Carioca [Figueiro and Monteiro 1976]), is partly conducive to dyad chordal tonicization because the interval between the open note of the bottom berimbau and the raised pitch of the bow immediately above it is likely to approximate a minor third or a major third.

[80] Although chordal tonicization is the least common approach we found (making up only 6.4% of the tonicized songs), it may explain the high esteem that capoeiristas have for certain historical recordings. As Diaz, Assunção, and Beyer (2021: 301–3) have explained, Mestre Cobra Mansa and members of the Associação de Capoeira Angola Dobrada from Belo Horizonte (Mestres Índio, Rogeiro, and Alcione), for instance, have modeled their musical approach on Traíra’s album. Furthermore, some capoeira angola mestres in today’s Bahia have become overtly interested in chordal tuning and singing in related tonalities, as is the case of Mestre Cabello from the group Barracão de Angola. We believe that the long-standing appreciation for Caiçara’s and Traíra’s recordings and the recent interest in chordal tuning by some capoeiristas can be partly attributed to their approach of singing in harmony with chordal tuning. The mestres who use this approach are few, probably because it requires a minimum degree of knowledge of Western harmony that is usually not part of traditional capoeira music training. While many capoeiristas may sing in parallel thirds without any explicit knowledge of Western harmony, it is less likely that they will tune the berimbaus chordally by intuition.

Category 4: Overtone Tonicization

[81] As explained in the section on methodology above, the fundamental is absent from the frequency spectrum of a berimbau tone, and as result, the loudest berimbau frequency that we can hear directly from a recording is always an upper overtone (Vilão and Melo 2014). Therefore, strictly speaking, the three categories of tonicization that we have proposed thus far are forms of overtone tonicization. Because the berimbau produces and amplifies many other overtones that often sound as “separate distinct pitches” due to the bow’s inharmonicities, we have conceptualized the loudest overtone of a berimbau tone as a working fundamental to differentiate it from its concurrent (and softer) inharmonic overtones. The fourth approach considers singers who tonicize to these other prominent overtones that can be detected through a spectrogram analysis, rather than the working fundamentals that we heard. Ethnomusicological studies suggest that many singers from sub-Saharan Africa use overtones as singing reference (England 1964; Johnston 1970; Rycroft 1975/76; and Díaz, Assunção, and Beyer 2021). While studying the relationship between singing and bow overtones is relatively straightforward in these African cases where a single bow normatively accompanies a solo singer, it is more difficult in multi-bow capoeira ensembles accompanied by percussion and featuring many singers, because the various inharmonicities of these instruments clash with each other, therefore limiting their capacity to serve as guidance for singers. For these reasons, we only considered the possibility of this form of tonicization for solo singers who (1) do not tonicize according to any of the three approaches already discussed; and (2) use a single berimbau. Additionally, we considered overtones in the 1940 recording of Bimba where only one of the two berimbaus is clearly heard (Turner 1940). This louder berimbau, surely played by Bimba himself, was captured by the single microphone that Turner used to record the mestre’s singing voice and was likely the one that influenced his song’s mode to a larger extent. As shown in Table 5 and Appendix B, 46 songs in the corpus utilize modes whose fundamentals are reasonably close to a bow overtone; that is, about 16% of the songs that used some form of tonicization.

[82] We recognize that many issues may arise when considering the possibility of overtone tonicization. First, it is possible that what we heard as tonicization of overtones may be chance, rather than deliberate pitch matching. As shown in Example 4, a single berimbau pitch may produce many inharmonic overtones that could randomly approximate the singing mode. This could have been the case of Mestre Pastinha, who used overtone tonicization in only one of his five ladainhas; in another he used direct tonicization and in the remaining three he did not use any form of tonicization. Because the mestre is not consistent in the tonicization of overtones, it is possible that this tonicization was coincidental. A similar claim could be made for other singers like Cabecinha, Juvenal, and Arnol, who used this approach only once. Bimba, on the other hand, seems to have demonstrated perception of berimbau overtones in three sets of recordings from three different decades. As shown in Table 5, the only form of tonicization that Bimba used in his 1940 (Turner) and 1968 (Moura) recordings was overtone tonicization. Likewise, in his 1953 recording, overtone tonicization was his most frequently used approach. Overall, Bimba was the singer who used overtones the most—he was responsible for 31 of the 46 songs that we postulate as being tonicized with this approach. Second, because of the berimbau’s extraordinary capacity to produce inharmonic overtones, singers theoretically have a wide array of overtones as candidates for tonicization. The challenge is determining which could have been the overtones that may have guided these singers. Our assumption was that these overtones were loud enough to have been picked up by the recording equipment of the time and subsequently detected by contemporary audio software. But it is possible that the overtones captured by Singscope were faint at the time of the recording and therefore did not affect singers in any meaningful way.

[83] It is for these reasons that we present overtone tonicization not as a definitive part of our taxonomy, but rather as a possibility—one that we cannot be certain of, or measure with the same level of precision as we did with regards to the other categories. While it may be possible to correct some of our findings through a study of acoustics, we believe that the general idea that singers may be intuitively guided by overtones should not be discarded, as the case of Bimba compellingly suggests.

Category 5: Non-Tonicization

[84] We failed to find any form of tonicization of berimbau notes or overtones in 154 of the 448 songs of the corpus; that is, in approximately one third of the songs. As shown in Table 5, most singers sang one or more songs without any obvious tonal relationship to berimbau pitches. The exceptions were Bimba (Robatto Filho 1954; Machado 1962, and Kast 1968), Totonho de Maré (Moura 1968), Canjiquinha (Smith 1960 and 1962a), Baiano (Grupo Bonfim 1994 [1969]), Zé Grande (Grupo Bonfim 1994 [1969]), and Comenda (Comenda and Murundelê 1972), who tonicized their songs in multiple ways but without a clear pattern that may explain why they are the only ones to avoid non-tonicization. Among non-tonicizers, Bimba (Turner 1940), Natividade (Rautavaara 1963–4), and Suassuna (Suassuna and Dirceu 1975) stood out, especially the latter two, whose ratio of non-tonicization over the total number of songs they recorded is above 70%. Again, there is no visible pattern that may account for their tendency to avoid tonicization.

[85] Non-tonicization may be explained by some singers’ tendency to sing in a comfortable register that allows them to project their voices loudly and clearly, regardless of bow tuning. As we can see in Appendix A, various mestres recorded most, or all their songs approximately in one tonality. In 1940, for instance, Bimba sang the first three minutes in B major, raised it to C major, and remained there for the remaining 20 minutes of the recording session, always returning to the same tonality after every break (Turner 1940). Other cases include Cabecinha, who sang his five tracks in A minor, and Gato Preto, who sang his six tracks in B major. Singers such as Cabecinha and Gato Preto tuned their berimbaus to match their comfortable singing range, but others such as Bimba (Turner 1940) may have not been able to do so because of the properties of the instrument, and this may account for his non-tonicization.^[23] Rather than shifting their singing mode to match berimbau tuning, they may have prioritized singing in a comfortable register over tonicization.

[86] Lastly, we would be remiss if we did not address Pastinha’s approach, given the status he has held among capoeira angola practitioners since the 1940s, when he emerged as a guardian figure and main codifier of this style. Despite his status, it is remarkable that Pastinha recorded only five songs with a combined duration of less than two minutes (five ladainhas on his 1969 album). Of those, one was directly tonicized, another was tonicized to an overtone, perhaps by chance, and the remaining three did not have any obvious form of tonicization. It is possible that Pastinha’s disregard for bow pitches while singing and the relative simplicity of his phrasing (as shown in Example 22b, his melodies lack the syncopation and melodic and rhythmic variety of his contemporaries) can be explained by his overreliance on his students and other capoeiristas who specialized in the music. It is known that during the 1950s and 1960s Pastinha appointed his student Waldomiro Malvadeza and mestres such as Waldemar and Gato Preto to be his mestres de bateria e de canto (Abreu 2003; Cruz 2022: 20). Their role was to prepare, tune, and play the instruments as well as lead songs during Pastinha’s rodas. In an interview with capoeira mestre and scholar Luiz Renato Vieira in 1989, Waldemar confirmed this, elaborating that, although Pastinha was perceived as the most authoritative capoeira angola figure in Bahia between the 1940s and 1960s (“capoeira’s president,” in his words), he was a modest singer and berimbau player and for that reason he let others oversee the music.^[24] Pastinha’s seemingly modest musical skills in capoeira are also surprising considering his musical training in Western music—during his youth, Pastinha enrolled for about seven years (1902–9) in the Escola de Aprendizes de Marinheiro da Bahia (Navy Academy of Bahia), where he received formal training in Western music theory and notation and learned how to play the cornet and the horn (Decânio 1996). Given that the mestre likely developed pitch-matching skills through ear training in this program, why are these skills not reflected in his capoeira recordings? We see three possibilities: First, because his focus was on capoeira’s physical movements and philosophy, he may have overlooked nuances of musical practice, such as the tonicization of berimbau pitches. Second, even if he paid attention to music, he may have treated the berimbau as a percussion instrument that produces rhythms, not pitches that could guide his singing. Lastly, by the time of the recording in 1969, his musical skills may have declined due to his advanced age (80) and lack of continued practice. Psychologists like Mari Riess Jones argue that skills such as beat entrainment and possibly pitch matching decrease as we age (2019: 102).

[87] In contrast, Mestre Bimba, the other iconic capoeirista from that period, who was consistently in charge of music during his rodas and recordings, is still well known for his musicality and, as we have discussed, the relatively rich archive we have of his recordings, spanning from 1940 to 1968, demonstrates a great variety of tonicization techniques.^[25] Like Pastinha, Bimba also had musical experience outside of capoeira; in this case, playing a chordophone called the viola in a local form of samba from Bahia, which is often performed immediately after capoeira rodas.^[26] But, differently to Pastinha, it seems that the experience of singing in tune with the chordal accompaniment of the viola sharpened Bimba’s pitch-matching skills and ultimately prompted him to better tonicize the berimbau in capoeira. A crucial difference between the two mestres’ musical experience outside of capoeira is that, while Bimba sustained the practice of playing viola throughout his adult life, Pastinha seems to have stopped playing the cornet and any other non-capoeira instrument once he left the Navy Academy in his youth. This factor, combined with the fact that Bimba was always in charge of playing the berimbau and singing in his group, while Pastinha delegated this role to others, may explain their different tonicization approaches. Plainly, consistent practice of singing in combination with berimbaus and even other melodic or harmonic instruments that do not belong to the capoeira tradition may affect tonicization.

Concluding Thoughts

[88] Our analysis of materials from the 1937–78 period has revealed a great variety of approaches to the question of how singers relate to berimbau tuning. We have organized that diversity into a taxonomy of five categories. Four of those categories postulate some kind of relationship between singing modes and berimbau pitches and one postulates that there is no relationship. Although we do not assume that old or new practitioners would deem the latter approach as “wrong” or aesthetically inferior, our analysis suggests that many mestres do have a preference for one or more of the four approaches that are based on an explicit relationship between singing modes and berimbau notes—as shown in Table 5, singers tonicize about two thirds of the time. One evidence pointing in this direction is that some mestres subtly raise the singing mode to match bow pitches, suggesting that they strive to sing in tune with the berimbaus. Another may be that the four tonicization approaches appear more frequently in studio/commercial albums, documentaries, and films than in field recordings—in studio recordings singers tonicized about 7 out of 10 songs, whereas in field recordings they did so 6 out of 10 times. In the context of a recording studio, mestres are able to hear themselves and re-record adjusting to their ideal musical aesthetics. This is not the case in field recordings, where mestres typically get a single opportunity to record. While some of the approaches proposed here may seem more musically sophisticated than others for the Western-trained ear (for instance, tuning berimbaus to form a tonic chord with selected pitches from each bow), as practitioners, we caution readers to refrain from such value judgements. The goal of this study has been to discover, by inference, the ideal aesthetic principles that may have guided recognized mestres from the past in their tuning and singing practices. We argue that the categories proposed in this article are, at least, part of those aesthetics and hope to verify them and possibly revise them through an ethnographic study with contemporary practitioners.

[89] This study was possible through the implementation of a method of hybrid transcription that takes advantage of both human transcription and specialized software capable of measuring and representing pitches at microtonal level. We offer this method as a tool for the study of musical traditions in which microtonal activity is an inherent part of singing and/or instrument tuning, such as Hindustani or Karnatic music, some styles of Indonesian gamelan, or modal musics from the Middle East. We are aware that specialized software is becoming better at direct automatic transcription (especially of monophonic music) and that eventually these programs may become more accurate tools to measure microtonal music. However, we believe that hybrid transcription offers two definitive advantages over automatic methods: First, by involving the transcriber’s ears and singing voices, the method better reflects the way in which the recorded musicians may have heard, experienced, and processed the music. Second, direct automatic transcription may never be able to disentangle melodies from thick textures formed by instruments of the same family in which similar timbre spectra obscure the identity of separate melodies. This difficulty is exacerbated when those recordings are made with poor quality equipment, as is the case of many early twentieth-century recordings of world music repertoires.

[90] Our systematic study of historical recordings through hybrid transcription has produced other kinds of data that, although revelatory on their own, call for ethnographic exploration. For instance, we found that overall, singers tend to sing in the major mode—about 80% of the time. A few mestres, however, such as Cabecinha (Turner 1940), Waldemar (Leeds 1951), and Guimarães (1969), preferred the minor mode, and others like Traíra (Nascimento 1963), Arnol (Waddey 1966), Camafeu (Conceição 1967), and Bimba (Moura 1968) sang about half of their songs in the major mode and the other half in the minor. When singers change mode, they tend to do it in the transition between ladainha/quadras and chulas. Finally, we noticed that, with a few exceptions (e.g., Waldemar [Leeds 1951], Traíra [Nascimento 1963], and Camafeu [Conceição 1967]), singers who more frequently use the minor mode do so in single-bow ensembles. We question whether these tendencies are indicative of styles and substyles of lineages or groups, or if they are the mark of individual singers. Since most contemporary practitioners identify themselves as belonging to lineages connected to certain singers analyzed in this article, an ethnographic study of their practices and of the ways in which they listen to these recordings could help to answer our questions about the possible relationship between tonicization practices and capoeira lineages.

[91] We cannot emphasize enough the importance of tolerance and flexibility in understanding the taxonomy we propose. As we have seen, when singers tonicize a bow pitch, they oscillate around that pitch, sometimes below, other times above it. It is possible that our findings implicate a microtonal equivalent of the concept of microtiming in groove-based musics. Charles Keil’s (1995) concept of “participatory discrepancies,” Chris Stover’s (2009) “beat span,” or Anne Danielsen’s (2010) “beat-bin,” may be useful analogies to think about subtle melodic tolerance. In light of what these scholars (especially Danielsen) have proposed in the field of microtiming, which is also applicable to musical timing in capoeira, we argue that a berimbau pitch (or a set of berimbau pitches) may support a range of modes for singers, rather than a single one. Although there is no consensus about the exact size of the span, our assumption of a 50-cents interval of pitch tolerance has resulted in the identification of tonicization in two thirds of the songs. It is possible that for some singers the span is larger (up to 100 cents, for example) thus increasing the number of tonicized songs.

[92] Finally, while we do not have evidence to establish whether the kinds of relationships postulated in our taxonomy were conscious or intuitive, we suspect that many of the analyzed singers, and probably their students and associates too, were able to tell when certain singing styles were congruent with berimbau notes, even if they did not have a precise vocabulary to name and teach those relationships. Again, a further ethnographic study may illuminate the extent to which the relationships we propose in our taxonomy resonate with the approaches of contemporary mestres.

Acknowledgements

We are grateful to UC Davis’s audio engineer Stephen Bingen for helping us understand the acoustic properties of the berimbau, particularly its broad overtone spectrum. We also thank mestres Cobra Mansa and Cabello, who helped us contextualize our analyses by sharing their approaches to singing and berimbau tuning with us.

References

Abreu, Frede. 2003. O Barracão do Mestre Waldemar. Salvador, BA: Zarabatana.

Adolfo, Antonio. 1997. Composição, Uma Discussão sobre o Processo Criativo Brasileiro. Rio de Janeiro: Lumiar.

Almeida, Bira (aka Mestre Acordeon). 1986. Capoeira: A Brazilian Art Form: History, Philosophy, and Practice. Berkeley, CA: North Atlantic Books.

Almeida, Renato. 1942. “O Brinquedo da Capoeira.” Revista do Arquivo Municipal de São Paulo 7 (84): 155–62.

Alvarenga, Oneyda. 1946. Melodias Registradas por Medios não-Mecânicos, vol 1. São Paulo: Archivo Folclórico da Discoteca Publica Municipal, Prefeitura do Municipio de São Paulo, Departamento de Cultura.

Angulo, Isa. 2008. “Following the Trail of the Snake: A Life History of Cobra Mansa ‘Cobrinha’ Mestre of Capoeira.” PhD diss., University of Maryland.

Babbitt, Milton. 1965. “The Structure and Function of Musical Theory: I.” College Music Symposium 5: 49–60.

Béhague, Gérard. 1973. “Latin American Folk Music.” In Folk and Traditional Music of the

Western Continents, edited by Bruno Nettl, 179–206. 2^nd ed. Englewood Cliffs, NJ: Prentice-Hall.

Beyer, Greg. 2004. “O Berimbau: A Project of Ethnomusicological Research, Musicological Analysis, and Creative Endeavor.” DMA diss., Manhattan School of Music.

______. 2016. A Musicalidade da ACAD. Arcomusical. DeKalb, IL: Self-published.

Biancardi, Emília. 2006. Raízes Musicais da Bahia. Salvador: Secretaría de Cultura e Turismo da Bahia.

Burns, Edward, and W. Dixon Ward. 1978. “Categorical Perception—Phenomenon or Epiphenomenon: Evidence from Experiments in the Perception of Melodic Musical Intervals.” The Journal of the Acoustical Society of America 63 (2): 456–68.

Calvacanti, Rodolfo Coelho and Pierre Verger. 1947. “ABC da Bahia.” O Cruzeiro, May 3, 1947: 57.

Catunda, Eunice. 1952. “Capoeira no Terreiro de Mestre Waldemar.” Fundamentos: Revista de Cultura Moderna 30: 16–18.

Carneiro, Edison. 1936. Religiões Brasileiras: Notas de Etnografia Religiosa. Rio de Janeiro: Civilização Brasileira.

______. 1991 [1937]. Negros Bantos: Notas de Etnografia e Folclore. 3^rd edition, Rio de Janeiro: Civilização Brasileira.

Cruz, Jose Luis Oliveira (aka Mestre Bola Sete). 2022 Capoeira Angola: Uma Filosofia de Vida (O Mestre e o Discipulo). 2^nd ed. Self-published.

Danielsen, Anna, ed. 2010. “Here, There and Everywhere: Three Accounts of Pulse in DAngelo’s ‘Left and Right’.” In Musical Rhythm in the Age of Digital Reproduction, edited by Anna Danielsen, 19–36. Farnham: Ashgate.

D’Anunciação, Luiz. 1971. “O Berimbau da Bahia.” Revista Brasileira de Folclore 11 (29): 24–33.

______. 1990. A Percussão dos Ritmos Brasileiros: Sua Técnica e sua Escrita. vol. 1: O Berimbau. Rio de Janeiro: Editora Europa.

Decânio, Ângelo. 1996. A Herança de Pastinha. A Metafísica da Capoeira. Comentários de Trechos Selecionados do Mestre. Salvador, BA: Self-published.

Densmore, Frances. 1956. Seminole Music. Washington, DC: United States Government Printing Office.

Díaz, Juan Diego, Matthias Röhrig Assunção, and Greg Beyer. 2021. “Arching over the Atlantic: Exploring Links Between Brazilian and Angolan Musical Bows.” Ethnomusicology 65 (2): 286–323.

Diniz, Flavia Cacheniski. 2010. “Capoeira Angola: Identidade e Trânsito Musical.” Master’s diss., Federal University of Bahia.

England, Nicholas, Robert Garfias, Miecyzslaw Kolinski, George List, Willard Rhodes, and Charles Seeger. 1964. “Symposium on Transcription and Analysis: A Hukwe Song with Musical Bow.” Ethnomusicology 8 (3): 223–77.

Estis, Julie M., Ashli Dean-Claytor, Robert E. Moore, and Thomas L. Rowell. 2011. “Pitch-Matching Accuracy in Trained Singers and Untrained Individuals: The Impact of Musical Interference and Noise.” Journal of Voice 25 (2): 173–80.

Galm, Eric. 2010. The Berimbau: Soul of Brazilian Music. Jackson, MS: University Press of Mississippi.

Johnston, Thomas. 1970. “Xizambi Friction-Bow Music of the Shangana-Tsonga.” African Music 4 (4): 81–95.

Jones, Mari Riess. 2019. Time Will Tell: A Theory of Dynamic Attending. New York: Oxford University Press.

Keil, Charles. 1995. “The Theory of Participatory Discrepancies: A Progress Report.” Ethnomusicology 39 (1): 1–19.

Koulaguina, Elena, Brandi Lee Drisdelle, Claude Alain, Stephan Grimault, Douglas Eck, François Vachon, and Pierre Jolicoeur. 2015. “The Perception of Concurrent Sound Objects through the Use of Harmonic Enhancement: A Study of Auditory Attention.” Attention, Perception & Psychophysics 77 (3): 922–9.

Kristensen, Bárbara and Gian C. R. Santos (aka Contramestre Latino). 2020. Capoeira Afinada. Santo Amaro, BA: Ecos de Santo Amaro.

Landes, Ruth. 1994 [1947]. The City of Women. Albuquerque, NM: University of New Mexico Press.

Larrain, Nicolás Rafael Severin. 2005. “Capoeira Angola: Música e Dança.” Master’s diss., Federal University of Bahia.

Leeds, Anthony. 1957. “Economic Cycles in Brazil: The Persistence of a Total Culture-Pattern: Cacao and Other Cases.” PhD diss., Columbia University.

Lewis, Lowell. 1992. Ring of Liberation: Deceptive Discourse in Brazilian Capoeira. Chicago, IL: University of Chicago Press.

Murphy, John. 2006. Music in Brazil: Experiencing Music, Expressing Culture. Oxford: Oxford University Press.

Nzewi, Meki. 1974. “Melo-Rhythmic Essence and Hot Rhythm in Nigerian Folk Music.” The Black Perspective in Music 2 (1) 23–8.

Ott, Carlos. 1955–7. Formação e Evolução Étnica da Cidade do Salvador. 2 vols. Salvador, BA: Manú Editora.

Querino, Manuel. 1922 [1916]. A Bahia de Outrora: Vultos e Factos Populares. 2^nd edition, Salvador, Bahia: Livraria Economica.

Rego, Waldelor. 1968. Capoeira Angola: Ensaio Sócio-Etnográfico. Rio de Janeiro: Itapoã.

Reis, Leonardo Abreu. 2010. “Cantos de Capoeira: Fonogramas e Etnografias no Diálogo da Tradição.” PhD diss., Pontifícia Universidade Católica do Rio de Janeiro.

Rycroft, Davis. 1975/76. “The Zulu Bow Songs of Princess Magogo.” African Music 5 (4): 41–97.

Seeger, Anthony. 2004. Why Suyá Sing: A Musical Anthropology of an Amazonian People. Champaign, IL: University of Illinois Press.

Silva, Vladimir. 2005. “Os Modos da Música Nordestina.” Piano Class: Revista de Música e Artes, published on February 4, 2005. Accessed May 6, 2024 https://www.hugoribeiro.com.br/biblioteca-digital/Silva-modos_musica_nordestina.pdf

Shaffer, Kay 1977. O Berimbau-de-Barriga e seus Toques. Rio de Janeiro: Ministerio de Educação e Cultura, Secretaría de Asuntos Culturais, Fundação Nacional de Arte, Instituto Nacional de Folclore.

Stover, Chris. 2009. “A Theory of Flexible Rhythmic Spaces.” PhD diss., University of Washington.

Tavares, Cláudio and Pierre Verger. 1948. “Capoeira Mata Um!” O Cruzeiro, January 10, 1948: 6–14.

Vilão, Rui, and Santino Melo. 2014. “Berimbau: A Simple Instrument for Teaching Basic Concepts in The Physics and Psychoacoustics of Music.” American Journal of Physics 82: 1149–56.

Audio Recordings

Comenda, Onias and Edinho Murundelê. 1972. Eu, Bahia – Atabaque e Berimbau. Phillips 6470.003, LP.

Cunha, Cristiano Mattos da (aka Cabeleira). 2020. Mestre Bimba e seus Tocadores. Remastered recordings of Mestre Bimba made by an anonymous person on December 7, 1953, in Salvador, BA. Self-published, CD.

Conceição, Ápio Patrocínio da (aka Camafeu de Oxossi). 1967. Berimbaus da Bahia. Musicolor LPK-20.108, LP.

______. 1968. Camafeu de Oxossi. Phillips P632.916L, LP.

Domingues, Henrique Foréis (aka Almirante). 1938. “As Cantigas dos Capoeiras da Bahia.” Presented on June 20 in the program Curiosidades Musicais, Radio Nacional del Brasil. Acervo Museo da Imagem e do Som. Accessed May 6, 2024. https://www.facebook.com/watch/?v=2260047400928285

Dreyfus-Roche, Simone. 1956. Brésil, Bahia, vol 2. Field recordings made in Salvador, Bahia October 31, 1955. Paris: CNRS/Musée de l’Homme MH16, LP.

Figueiro, Júlio César (aka Mestre Carioca) and Darcy Monteiro. 1976. Quilombo – Jongo Basam & Capoeira Angola. Tapecar LP-TC-083, LP.

Grupo Bonfim. 1994. Grupo Bonfim, vol 1.: Rio/RJ – Capoeira dos Anos 60s. Remastered recordings of a roda of this group in Rio de Janeiro in 1969. Self-published, CD.

Guimarães do Berimbau. 1969. Sambas de Roda e Capoeira. Rosicler R-7044, LP.

Leeds, Anthony. 1951 “Index of Brazil, Bahia. Antony Leeds 1951–1952.” Field recordings made in Salvador, Bahia, June and July 1951. Archives of Traditional Music, Indiana University, 54-073-F (ATL371, ATL373), wire magnetic recordings.

Machado, Manoel dos Reis (aka Mestre Bimba). 1962. Curso de Capoeira Regional. JS Discos JLP-101, LP.

Moraes, Antônio Conceição (aka Mestre Caiçara). 1969. Academia De Capoeira De Angola São Jorge Dos Irmãos Unidos Do Mestre Caiçara. AMC, AMCLP-5032, LP.

Moraes, Pedro Trindade (aka Mestre Moraes) and Pelourinho Group of Capoeira Angola (GCAP). 1996. Capoeira Angola from Salvador, Brazil. Smithsonian Folkways Recordings SFW40465, CD.

Moura, Jair. 1968. Soundtrack of Dança de Guerra, Accessed May 6, 2024. https://www.youtube.com/watch?v=uibmcKWOQMg

Nascimento, José Ramos do (aka Mestre Traíra). 1963. Mestre Traíra: Capoeira da Bahia. Rio de Janeiro: Editora Xauã EX.1-002, LP.

Pastinha, Vicente Ferreira (aka Mestre Pastinha). 1969. Mestre Pastinha e sua Academia: Capoeira Angola. Phillips R 765.097 L, LP.

Rautavaara, Helinä. 1963–4. Field recordings of Mestre Gato Preto and his group made on December 31, 1963 (at Mirante de Cabalar, Salvador); Mestre Pastinha and his group on July 2, 1964 (at Pelourinho square, Salvador); and Mestre Waldemar and his group in the first half of 1964 (at the Liberdade neighborhood, Salvador). Helinä Rautavaara Museum in Espoo, Finland and Finnish Literary Society in Helsinki (SKSÄ-HR), magnetic tape.

Santos, Oswaldo Lisboa dos (aka Mestre Paraná). 1963. Capoeira: Mestre Paraná. CBS Masterworks 56126, 45-rpm EP.

Smith, Janice Marie. 1960. Field recordings of Mestre Canjiquinha and his group at Itapõa, Salvador, Bahia on August 24, 1960. Archives of Traditional Music, Indiana University, 61-062-F (ATL 1738–1741), tape reels.

Suassuna, Reinaldo Ramos (aka Mestre Suassuna) and Dirceu. 1975. Capoeira Cordão de Ouro. Continental 104.405.102, LP.

Suassuna, Reinaldo Ramos (aka Mestre Suassuna) and Dirceu 1978. Capoeira Cordão de Ouro, vol 2. Continental 104.405.259, LP.

Turner, Lorenzo Dow. 1940. “The Dow Turner Collection in Brazil.” Field recordings of Mestres Cabecinha, Juvenal, and Bimba in Salvador, Bahia, November and December 1940. Archives of Traditional Music, Indiana University, Lorenzo Dow Turner Collection, 86-109-F, aluminum discs.

Melo, João. 1968. Viva Bahia! N^o 2. Phillips P632.923 L, LP.

Waddey, Ralph Cole. 1966. “Ralph Cole Waddey Recordings of Afro-Brazilian Music.” Field recordings of Mestre Arnol Conceição and his group at Pernambués neighborhood of Salvador, BA on November 11, 1966. Library of Congress, American Folklife Center, AFC 1971/025, 7-inch open-reel tape.

Audiovisual Sources

Anonymous. 1937. Capoeira Exhibition at the Second Afro-Brazilian Congress in Salvador, Bahia. Silent video. 1:26. Accessed May 6, 2024. https://www.youtube.com/watch?v=XjvOunvvQJ0

Araujo, Alceu Maynard. ca. 1950. “Veja o Brasil – Capoeira Angola.” TV program. 5:29. Accessed May 6, 2024. http://www.youtube.com/watch?v=QrLNIwZ1x50

Duarte, Anselmo. 1962. O Pagador de Promessas. Cinedistri. 91 min. Accessed May 6, 2024. https://www.youtube.com/watch?v=xGWveCsI2Wo

Farias, Lázaro. 2006. Mandinga em Manhattan: Como a Capoeira se Espalho pelo Mundo. 54:22. Accessed May 6, 2024. https://www.youtube.com/watch?v=K_y_84AwQI0

Fontoura, Antônio Carlos De. 1977. Cordão de Ouro. Embrafilme. 1:10:53. Accessed May 6, 2024. http://www.youtube.com/watch?v=VTnIV-ECPIE

Kast, Pierre. 1968. La Capoeira de Mestre Bimba. Filmed in 1966 in Salvador, Bahia and broadcast on Office Radio Télévision France on March 3, 1968. 6:11. Accessed May 6, 2024. https://www.youtube.com/watch?v=ZZjB1q6Vw_0

Moura, Jair. 1968. Dança de Guerra. 18:06. Accessed May 6, 2024. https://www.youtube.com/watch?v=q5w0l8nEylM

Peter, Solange. 1960. Voyage sans Pasaporte. Filmed on Jan 9, 1960 in Salvador, Bahia and broadcast on Radio Telévision Française. 12:47. Accessed May 6, 2024. https://fb.watch/iXr02EXm_s/

Robatto Filho, Alexandre. 1954. Vadiação. 8:14. Accessed May 6, 2024.http://www.youtube.com/watch?v=ObGj2e2bsAc

Rocha, Gauber. 1962. Barravento. Horus Filmes. 1:20:54. Accessed May 6, 2024. https://www.youtube.com/watch?v=sy60bm2Cn04

Santos, Luis Paulinho dos. 1960. Um Dia na Rampa. Filmed in Salvador, Bahia in 1960. 10:08. Accessed May 6, 2024. https://www.youtube.com/watch?v=Tdqb7a0ykCM

^[1] Capoeira has many styles and substyles. Today, most groups identify three: the traditionalist style of capoeira angola is considered by practitioners to be the closest to its African roots; capoeira regional, a variant invented by Mestre Bimba focused on fight efficiency; and capoeira contemporânea, a hybrid style combining the former two with other modern variations. The former two were developed in Bahia in the first half of the twentieth century and the latter in the larger cities of the Brazilian south, starting in the 1970s. In this article we use the term capoeira to refer to aspects that apply across styles, unless we use the specific style name.

^[2] Mestre Cobra Mansa, personal communication, Salvador (Bahia, Brazil), April 2006, August 2009, March 2012.

^[3] Kristensen and Santos (2020) analyzed berimbau tuning and song mode in the 1940 recordings by Turner, Dreyfus-Roche’s recordings of Waldemar in 1955, and four commercial albums from the 1960s (Traíra [Nascimento 1963], Camafeu de Oxossi [Conceição 1967], Caiçara [Moraes 1969], and Pastinha 1969). Their corpus for the 1940–69 period is thus incomplete. For the analysis of each recording, they considered the first song of each track only, which cannot be taken as representative of the entire set. More problematically, song mode was assumed to match the first sung note—as we know, capoeira songs can start in the tonic, but also in the third or fifth scale degrees. Lastly, berimbau notes were pigeonholed in the 12-TET system, which is inconsistent with berimbau tuning. Their work is, however, commendable not only because they are pioneers, but also for the inclusion of rich interviews with mestres discussing their tuning approaches and for the pedagogical section recommending tuning systems for the capoeira ensemble.

^[4] Some mestres refer to these three berimbaus differently. For Mestre Traíra (1963), for instance, the lowest-pitched bow is called berra-boi, the medium-pitched is gunga, and the highest-pitched is viola.

^[5] Authors such as Gérhard Béhague (1973), Antonio Adolfo (1997), and Vladimir Silva (2005) have explained that various folk genres from the Brazilian northeast, such as baião and frevo, use a particular set of scales called escalas nordestinas (northeastern scales). These scales include the equivalent of the Dorian, Lydian with a flattened seventh, and especially the Mixolydian. Since many of the singers discussed in this paper were likely immersed in these northeastern genres (Bahia is in the Brazilian northeast), they may have recorded some capoeira songs using escalas nordestinas.

^[6] Other authors agree with Shaffer’s downbeat placement; for instance, Camargo Guarnieri in Alvarenga (1946), Almeida (1986), Biancardi (2006), and Angulo (2008).

^[7] The corpus includes 4 Candomblé songs, 4 Caboclô songs, 1 maculelê song, and 20 samba de roda songs; all accompanied by berimbaus. Regardless of genre, songs not accompanied by berimbau were not included in the corpus.

[8] In 1938 Henrique Foréis Domingues (aka Almirante) hosted a program on Brazil’s National Radio featuring many capoeira songs from Bahia (Domingues 1938). Because none of these songs are accompanied by the berimbau, these materials are not considered in this study.

^[9] There are other transcriptions of capoeira music during this period published by Renato Almeida (1942), Carlos Ott (1955–7), and Kay Shaffer (1977). Because they only present unaccompanied song melodies or purely rhythmic berimbau patterns, we do not consider them in this study.

^[10] Other visitors include Ruth Landes, Donald Pierson, Melville Herskovits, and Frank Frazier. In fact, Turner visited Bahia with Frazier and was received by Pierson.

^[11] In some of his recordings, probably including those of capoeira, Turner wrote the note “Radio Studio” (Briand 2006).

^[12] According to historian Matthias Röhrig Assunção, after discovering Turner’s recordings in the late 1990s, Brown shared them with his mestre in London (Mestre Pastel) and with Assunção, who used to train with this group. Assunção shared these recordings with various mestres in Bahia, including Decânio and Nenel. Later, in the early 2000s, Decânio and others made the recordings public through the internet.

^[13] Online communication with the authors, September 28, 2022.

^[14] Online communication with Juan Diego Diaz, November 22, 2022.

^[15] A second edition of this album was released in 2012.

^[16] For our analysis we did not consider Tiburcinho’s songs because none of them was accompanied by the berimbau.

^[17] Alceu Maynard Araujo’s 1950 TV program Veja o Brasil shows images of a roda by Mestre Pastinha’s group but without capoeira musical accompaniment. Instead, the musical background is an orchestral arrangement of Ary Barroso’s “Na Baixa dos Sapateiros.”

^[18] Solange Peter’s TV documentary Voyage sans Pasaporte, which aired in France in 1960, recycled the ladainha “Torpedeiro Encoraçado” from Dreyfus-Roche’s 1955 field recordings of Waldemar, which we analyze separately.

^[19] We stress that our analysis of harmonic modulation is only a possibility and may not necessarily reflect how singers and berimbau players understand these ways of singing.

[20] Modal leaps are more common when a new singer takes over the soloist role, adjusting the mode to his/her vocal register.

^[21] While contrary motion is the most common manner of combining berimbau toques in ensembles with three berimbaus, parallel motion is more common in ensembles with two berimbaus, as in the cases of Bimba (1940), Camafeu (1967 and 1968), Noronha (Moura 1968), Caiçara (Moraes 1969), Malvadeza (1969), Suassuna (1975 and 1978), and Comenda (Comenda and Murundelê 1972).

[22] We also encountered one case where non-traditional melodic/harmonic instruments are added but without matching the singing mode or berimbau tuning. In the ladainha “Capoeira de São Salvador,” (Suassuna, 1975, track A1) a bass guitar is added, reinforcing the B major tonality through the alternation of the tonic and dominant chords. The song is, however, approximately in B♭ major (a whole semitone flat) and the three berimbaus form an A♭ minor chord.

^[23] Although a wide range of pitches may theoretically be produced on a given berimbau by changing the tension of the wire and the position of the gourd across the staff, in practice the choices are limited because capoeiristas prioritize maximum resonance, which can only be achieved with certain degrees of wire tension and gourd positions. Kay Shaffer (1977: 28) confirms this limitation, writing that “In general, there are one or two positions [of the gourd] that produce a better sound. Any other position produces a sound of inferior quality.” In our experience, the margin to vary pitch without compromising resonance is rarely above a major second.

^[24] Luis Renato Vieira interviewed Waldemar at the Liberdade neighborhood in Salvador in 1989. The key parts where Waldemar questions Pastinha’s musicality can be found in Abreu (2003: 53). Reportedly, Mestres João Pequeno and João Grande, two students of Mestre Pastinha and contemporaries of Waldemar, confirmed Waldemar’s claims about Pastinha’s musicality (online communication, Mestre Cobra Mansa, August 28, 2021).

^[25] Bimba’s berimbau skills were widely recognized, even beyond capoeira. For instance, he was one of the few Bahian musicians invited to collaborate with musicians from a US American diplomatic delegation that toured around various South American countries in 1961. Lázaro Farias’s documentary Mandinga em Manhattan (2006) briefly shows images of his encounter with jazz guitarist Charlie Byrd’s trio at a concert hall in Salvador (15:30–15:42). Unfortunately, the audio is too faint and brief for it to be included in our analysis.

[26] Personal communication, Mestre Cabello, November 27, 2023, Davis, CA.

Analytical Approaches to World Musics

Analytical Approaches
to World Musics

AAWM Journal 12/1 (2024)

**Singing in Tune: Berimbau Tuning and Song Mode in Capoeira (1937–1978)**

Juan Diego Díaz and Alex Rossi

Background

Berimbau

Song

Methodology

Hybrid transcription

Toques: Emphasized pitches and transcription

Tolerance

Scope and Repertoire

Analysis

Tonicizing the emphasized pitch in Bimba’s “Sim Sim Sim” (1962)

Tonicizing the non-emphasized pitch in Juvenal’s “Sim Sim Sim” (1940)

Tonicizing the emphasized pitch and its relative minor in Bimba’s (1962) “Valha me Nossa Senhora”

Tuning the berimbau a fifth above the song fundamental in Onias Comenda’s (1972) “Tava lá em Casa”

Gradual mode migration in Caiçara’s (1969) “Tava em Casa” and Suassuna (1978)

Full and partial adjacent tuning in Caiçara (1969), Waldemar (1951), and Traíra (1963)

Chordal tuning of berimbaus in Guarnieri’s transcription (1937/46) and Traíra (1963)

Singing without obvious relation to berimbau pitches in Pastinha’s (1969) “Eu Nasci pra Capoeira”

Listening to overtones

Discussion

Category 1: Direct Tonicization of a Berimbau Pitch

Category 2: Indirect Tonicization of a Berimbau Note

Category 3: Chordal Tonicization

Category 4: Overtone Tonicization

Category 5: Non-Tonicization

Concluding Thoughts

Acknowledgements

References

Audio Recordings

Audiovisual Sources

Analytical Approaches to World Musics

Analytical Approaches to World Musics

Díaz_Rossi

AAWM Journal 12/1 (2024)

Singing in Tune: Berimbau Tuning and Song Mode in Capoeira (1937–1978)

Juan Diego Díaz and Alex Rossi

Background

Berimbau

Song

Methodology

Hybrid transcription

Toques: Emphasized pitches and transcription

Tolerance

Scope and Repertoire

Analysis

Tonicizing the emphasized pitch in Bimba’s “Sim Sim Sim” (1962)

Tonicizing the non-emphasized pitch in Juvenal’s “Sim Sim Sim” (1940)

Tonicizing the emphasized pitch and its relative minor in Bimba’s (1962) “Valha me Nossa Senhora”

Tuning the berimbau a fifth above the song fundamental in Onias Comenda’s (1972) “Tava lá em Casa”

Gradual mode migration in Caiçara’s (1969) “Tava em Casa” and Suassuna (1978)

Full and partial adjacent tuning in Caiçara (1969), Waldemar (1951), and Traíra (1963)

Chordal tuning of berimbaus in Guarnieri’s transcription (1937/46) and Traíra (1963)

Singing without obvious relation to berimbau pitches in Pastinha’s (1969) “Eu Nasci pra Capoeira”

Listening to overtones

Discussion

Category 1: Direct Tonicization of a Berimbau Pitch

Category 2: Indirect Tonicization of a Berimbau Note

Category 3: Chordal Tonicization

Category 4: Overtone Tonicization

Category 5: Non-Tonicization

Concluding Thoughts

Acknowledgements

References

Audio Recordings

Audiovisual Sources

Analytical Approaches
to World Musics

**Singing in Tune: Berimbau Tuning and Song Mode in Capoeira (1937–1978)**