ISSN 2158-5296

Analytical Approaches to World Musics

Analytical Approaches
to World Musics

ISSN 2158-5296

Simha 2025

AAWM Journal 13/No. 1 (2025)

African Meter: Measured Music Without “Measures”

Simha Arom

Time structuration, cyclic music, meter, rhythm, pulse, sub-Saharan Africa

In the measured music of sub-Saharan Africa, time is structured in three types of units:
• The beat, namely an isochronous yardstick, a series of equidistant points (often named “basic pulse”, “tactus level”, “regulative or reference beat”), which is equivalent to the medieval tactus.
• The elementary durations (“pulse”, “fastest pulse”, “elementary pulse-unit”, “density referent”,) which indicate the subdivision modes – binary, ternary or mixed – of the tactus.
• The cycle, highest level unit, within which a sequence of tactus acts as an armature and constitutes the framework of musical events characterized by repetition practices.
In this context, I propose to conceive meter as the whole of these units. From the outset, it is important to note that the tactus is not subject to any regular hierarchical grouping (by 2, by 3, etc.). In other words, the system ignores the notion of a “strong” beat that would necessarily be opposed to one or more “weak” beats; between the tactus and the cycle, there are no metric intermediates. As a corollary, the notion of “measure” is therefore absent here. This term only makes sense in the context of music predetermined by this mold. However, it is clear that, despite this obvious fact, most music theorists, as well as ethnomusicologists, transcribe African music using time signatures and Western-style bars, and consequently use a terminology that is largely inadequate for their analysis. Many misunderstandings arise from this situation. However, they could easily be cleared up if we were to accept the tactus as the cornerstone of this metric system, and not the fast pulses that are still favored by the ethnomusicological community. Contrary to the majority of ethnomusicologists who consider the “fast pulse” as guiding the performance of traditional sub-Saharan music, I am convinced that it is the tactus that assumes this function. This is the first point that I will attempt to demonstrate with arguments. After that, I will try to answer the question of which concepts to use – or to banish – to describe this heritage without betraying the conception that those who practice them have.

Simha Arom, PhD, is the Emeritus Research Director at CNRS (Centre National de la Recherche Scientifique), France

Click for DOI, citation, PDF version, and Appendix.

 

Introduction

In the music of sub-Saharan Africa, on which I have done fieldwork for more than thirty years, time is structured in three types of temporal units. Let me introduce these units, before putting forward the evidence for this. In my view, the types of temporal units are:

  1. The beat, namely an isochronous yardstick, a series of equidistant points (often named “basic pulse,” “main pulse,” “tactus level,” or “regulative” or “reference beat”), which is equivalent to the medieval tactus.
  2. The elementary durations (“pulse”, “fastest pulse”, “elementary pulse-units”, “density referent”) which derive from the subdivision modes—binary, ternary or mixed—of the tactus.
  3. A higher-level temporal unit, within which a sequence of tactus acts as a framework for the musical events, which are characterized by repetitions, or cycles.

[2] How can we understand the articulation between these three levels? And how should we transcribe these sub-Saharan musical pieces? I have two main hypotheses.  First, the tactus or “beat” is the most fundamental level. Second, these musical pieces should by no means be transcribed using the measures of Western staff notation, since this would introduce a higher-level temporality that would be foreign to them, and ignore the higher-level temporal unit that is specific to them.

 

About tactus and its subdivisions

 

[3] It is known that in sub-Saharan Africa, music is closely associated with dance. Any measured musical utterance—that is, one that lends itself to dance—is referred to, in the vernacular of performers, as a “song” (Arom and Martin 2015, 6). In this context, every musical piece is underpinned by a beat—i.e., the tactus—that corresponds to the basic step of the dance traditionally associated with it. However, very often, this beat, inherent to the musical piece in question, is not manifested: it remains present as an underlying sub-text. When this is the case, it is often difficult, sometimes even impossible, for an observer from outside the culture to determine with certainty where the beat occurs in the musical flow.

[4] In order to create transcriptions that respect the way the bearers of a given tradition conceive of this level of metric organization of their musical heritage, I always ask one of the performers to listen through headphones to the recordings that I made of each piece, and to systematically superimpose on it—without being given any other indication—hand claps. Thus, on a two-track stereo recorder, one track would be a copy of the initial recording, and the other, perfectly synchronized, the corresponding beat. This operation was always carried out in the presence of several elders considered locally as experts, as well as the musicians who had participated in the initial recording (Arom 1976).

[5] From 1964 to 1996, this process was applied to hundreds of recordings collected from some fifty ethnic communities, mostly from the Central African Republic, but also from Cameroon, Benin, Guinea, and the Democratic Republic of Congo. With a few rare exceptions, the musicians clapped the tactus without any hesitation. However, none of them ever materialized the subdivisions that are applied to the time durations created by the tactus, that is, the elementary units, i.e., the “fastest pulses” or “density referents.”

[6] In the African metric system, the tactus stratum can therefore be considered as the central element, especially since it is the only one that lends itself to both division and multiplication.

[7] Yet many ethnomusicologists and theorists, among others A.M. Jones (1959 vol. 1, 24), James Koetting (1970, 122), Kwabena Nketia (1974, 127), Robert Kauffman (1980, 396–7), David Locke (1982, 245), and Justin London (2004, 68, 167), consider the “fastest pulse” or “density referent” as the essential and the only regular pulse. In Koetting’s words, “[t]he fastest pulse is structurally fundamental, there being no standard substructure internal to it or between it and any pattern as a whole” (Koetting 1970, 122, my emphasis).

[8] Others, including Hewitt Pantaleoni, John Chernoff, Tellef Kvifte, and Kofi Agawu, disagree with this thesis—as I do. Let me quote them in turn. Pantaleoni asserts the following: “Throughout this paper ‘pulse’ is used in the sense of a slowly moving evenly spaced main beat…. Those who analyze African rhythms as various totals of an extremely brief unit of time, often refer to this small unit as a pulse, which is obviously not the same use of the term as my own.” (Pantaleoni 1972, 52, my emphasis). Chernoff likewise states that “[t]he pulse cannot profitably be defined as the fastest common unit of time that can unite the various rhythms. Rather, the main pulse […] represents the timing of dance steps” (Chernoff 1991, 1097, my emphasis).

[9] This is also the opinion put forward by Kvifte, who suggests that “[i]nstead of thinking about the pulse in the same way as a clock where all time units are built from one fast reference pulse … we might hold on to a tactus level … regarding the levels below as formed by dividing the beat and the levels above by adding the beats together” (Kvifte 2007, 75).

[10] Finally, Agawu also concurs: “For cultural insiders, identifying the gross pulse or the ‘pieds de danse’ (literally ‘dance feet’) occurs instinctively and spontaneously. Those not familiar with the choreographic supplement, however, sometimes have trouble locating the main beats and expressing them in movement…. No one hears a timeline without also hearing—in actuality or imaginatively—the movement of feet. And the movement of feet in turn registers directly or indirectly the metrical structure of the dance” (Agawu 2003,155, my emphasis). This is just common sense, because the movement that the feet of the dancer carry out while starting any dance is nothing else but the materialization of the tactus.

[11] Some cultures accept in the same piece two tactus that can be simultaneously divided, one in a binary way, and the other in a ternary way. However, the two superimposed beats have then a ratio of 2:3 or 3:2, and actors find them both adequate. This is notably the case with the Fulani of West Africa (Estreicher 1964), the Teke (Le Bomin 2004), the Fang and Tsogho of Gabon (Le Bomin 2017), and the Ewe of Ghana (Jones 1959). Under these conditions, it seems logical that the beat corresponding to the dance step should prevail as the primary tactus.

[12] Finally, for the sake of completeness, I should mention the phenomenon of aksak, i.e., asymmetrical rhythmic patterns, based on the alternation, at a fast tempo, of a certain number of binary cells – grouping two elementary units – with other cells, grouping three. Within measured music, aksak seems to provide the only case where the elementary units are structurally fundamental (Arom 2004). Apart from this exception, and in contrast to what many ethnomusicologists assert—when they consider these elementary units as guiding the performance of traditional sub-Saharan music—it is obvious that it is the tactus that assumes this function.

 

The higher-level temporal units

 

[13] It is important to emphasize a fact that I can assert on the basis of my fieldwork, which is that during any materialization of the tactus, the strikes are always of equal intensity. This means that the tactus is not subject to any regular hierarchical grouping, by 2, by 3, etc. In other words, African meter ignores the notion of a “strong” beat that would necessarily be opposed to one or more “weak” beats—which the notation with bars conventionally assumes. This remark highlights why one should not use time signatures and barlines in transcriptions of African music: indeed, using them would tacitly suggest that there exist stronger and weaker beats. Despite this obvious fact, most ethnomusicologists transcribe African music using time signatures and Western-style bars, and consequently use a terminology that is largely inadequate for the analysis of the musical pieces performed.

[14] In other words, and to make this point clearer,

  • The measure is a predetermined mold, the result of a writing convention that took shape and was used in specific contexts foreign to the practice of music in sub-Saharan Africa.
  • The term measure only makes sense in the context of music predetermined by this mold—that is, in Western music.
  • Time signatures and bar lines induce a hierarchy between strong and weak beats;
  • The notion of measure is foreign to all music in which such an opposition is absent, which is the case in the vast majority of sub-Saharan African music. Transcribing music using time signatures and bar lines is therefore misleading, because the mark of a regular strong beat induces the notion of a metric hierarchy.
  • As a corollary, the concept of “measure” is therefore absent from this music, in which another type of higher-level temporal unit plays a key part.

[15] Indeed, just as measures in Western music introduce a time unit larger than what the tactus indicates, sub-Saharan music also uses a time unit that is larger than the tactus, and which is made up of an integral number of tactus. This larger time unit is marked using repetitive musical events—to which I refer using the term “cycle.”

[16] Given the essentially cyclical character of this music, it would be easy not to use measures and time signatures in transcriptions, but rather to use markers which delineate these cycles. Among those who have attempted it, we can count David Locke, who avoids time signatures but keeps the bar line (Locke 2010), Gerhard Kubik, who indicates the number of elementary units per cycle (Kubik 2010, 143–166), and myself. In my own transcriptions, I designate the beginning of each cycle, as well as the numbered position of each of its tactus (Arom 1991).[1]

 

Conclusion: The use of inappropriate terminology

 

[17] The use of measures and time signatures in the transcriptions of music from sub-Saharan Africa has generated serious misunderstandings. However, these could easily be overcome if we were to accept the tactus as the cornerstone of its metric system, and not the fast pulses that are still favored by the ethnomusicological community. More precisely, in this context, I propose to conceive of meter as the combination of three components: first, the tactus, then the elementary durations that subdivide it, and, last but not least, the higher-level temporal units, i.e., the cycles. If one accepts my definition of these three metric strata, the terminology concerning the organization of time in the traditional music of sub-Saharan Africa—as well as in many other geo-cultural areas—could be considerably simplified and, at the same time, become much clearer.

[18] First of all, I suggest that we agree on the name of tactus or beat and that we give up the multitude of terms (strong beat, main beat, slow beat, downbeat, etc.) that have been used until now to designate the musical equivalent of the basic dance step.

[19] Second, we would have to admit that, in music that is transmitted orally, there is no meter per se. Meter is only manifested by the musical events that are grafted onto it; as the basis that, at the same time, makes periodicity possible and is manifested by periodic events, meter constitutes an abstract grid, acting as a background and framework to sound events considered as belonging to measured music. Therefore, a timeline does not in any way constitute a “metrical structure,” but rather an entity of another type, i.e., a rhythmic pattern.

[20] Third, In the absence of a strong beat, the term “measure” and, by corollary, the time signatures and their corresponding bar lines are to be banished.

[21] Fourth, every piece is underpinned by a single tactus (or rarely, by two, in which case their relationship makes them reducible to each other). Consequently, the idea of “heterometry”—i.e., on the horizontal axis, a metric scheme wherein the time signature changes frequently—is meaningless. Likewise, the concept of “polymetry”—by which I mean the simultaneous unfolding, on the vertical axis, of several parts not reducible to a single meter—is also meaningless.

[22] Fifth, in a context where the beats are not hierarchical, there can be neither strong nor weak beats. Therefore, the term “syncopation”—which designates an onset on a weak beat, and prolonged on the next strong one—should be replaced by off-beat.

[23] These proposals could contribute to the development of a unified terminology that would enable us to talk about the organization of time in sub-Saharan African music—and even in other geo-cultural areas—in a much more concise way, and as a corollary, to better understand each other.

 

Acknowledgements

 

I would like to express my gratitude to Karine Chemla, for her attentive reading of this text and for her very useful comments, Julien André, for agreeing to retranscribe the long excerpt of the piece “Sogba or Sogo Dance,” based on transcription of A.M. Jones in his Studies in African Music (Jones 1959, Vol 2, 100–111), and Gilles Monfort, for the page layout of these two scores, as they appear in the Appendix.

 

Appendix

 

This Appendix presents a long excerpt from the “Sogba or Sogo Dance,” transcribed by A. M. Jones and published in his Studies in African Music (1959, Vol 2: 100–111) and, below it, a new transcription of the same excerpt, based on the principles and recommendations made in the main text. This new transcription was made by ethnomusicologist Julien André, a virtuoso of the djembe, which he studied during long stays with Malinke masters. As a result, André is an expert in West African polyrhythm. He has just published Systématique de la polyrythmie malinké: Mali / Guinée (André 2023)—which appears to be the first book devoted exclusively to this musical process.  

A comparison of the two transcriptions highlights what can be achieved by adopting the principles I advocate. The new transcription, by revealing the sequencing of “Sogba,” in fact constitutes a first level of its formal analysis. Already at this stage, we can see that:

  • The part played by drum-master Atsimevu is the only one not bound by a strictly regular metrical framework. It can therefore indulge in genuine improvisation.
  • The double bell Gankogui, the rattle Axatse, and the drums Kidi and Kagan perform strict ostinati. Gankogui and Axatse, which constitute timelines, evolve in a cycle of four dotted quarter notes (i.e., four ternary beats, each grouping three elementary units), while Kagan’s part is limited to a single dotted quarter note, but systematically positioned off-beat.[2] All three do not admit the slightest variation, and are repeated unchanged until the end of the piece.
  • Kidi’s cycle, on the other hand, varies: at the beginning of the excerpt (timeline cycles 22 to 24), its formula counts four beats—Kidi-a—then (in cycles 25 to 36) it changes to an eight-beat—Kidi-b—before returning (in cycle 39) to four beats.
  • The vertical hemiola resulting from the interweaving of the two claps (one ternary, the other binary) also constitutes a strict—and double—ostinato.[3]
  • ‘Sogba’ presents six different periodicities, whose internal ratios are as follows:
  • 1:2 Kidi-b compared to Timelines Gankogui and Axatse
  • 1:4 Timelines Gankogui and Axatse compared to Kagan
  • 1:8 Song compared to Kidi-b
  • 1:16 Song compared to Timelines Gankogui, Axatse and Kidi-a
  • 1:64 Song compared to Kagan
  • 2:3 Clap 1 compared to Clap 2
  • As for the song, it alternates between cantor and choir. A.M. Jones’s transcription shows many changes in the time signatures, not only on the horizontal axis, but more surprisingly also on the vertical one. Thus, just in the first two bars of the extract—measures 21–22 of his “Pattern B”—the staff entitled “Song” starts in 3/8, becomes 3/4 and then returns to 3/8. Vertically, the two upper parts, Gankogui and Axatse, are indicated in 12/8, while the staff for “Clap 1” and “Clap 2” is marked by the superposition of 12/8 to 6/4.
  • Similarly, the two lower parts, Kigi and Kagan, are in 6/8 for the former and 3/8 for the latter in Jones’s transcription.

In view of the arguments put forward in this article, it’s clear that to transcribe a piece of music in the way Jones does is to deny one of the fundamental principles of the African metric system—the unshakeable regularity of the tactus—and thus to undermine the consistency of its internal architecture.

Click here to view the Appendix in the PDF version of this article

Brief Bibliography

 

Agawu, Kofi, 2003. Representing African Music. New York and London: Routledge.

André, Julien, 2023. Systématique de la polyrythmie malinké : Mali / Guinée. Paris: Éditions Peeters, Selaf. Collection “Ethnomusicologie” 10.

Arom, Simha, 1976. “The Use of Play-Back Techniques in the Study of Oral Polyphonies.” Ethnomusicology 20(3): 483–519.

Arom, Simha, 1984a. “The Constituting Features of Central African Rhythmic Systems. A Tentative Typology.” The World of Music 26(1): 51–67.

Arom, Simha, 1984b. “Structuration du temps dans les musiques d’Afrique Centrale : périodicité, mètre, rythme et polyrythmie.” Revue de musicologie 70(1): 5–36.

Arom, Simha, 1989. “Time Structure in the Music of Central Africa : Periodicity, Meter, Rhythm and Polyrhythm.” Leonardo, Journal of the International Society for the Arts, Sciences and Technology 22(1): 91–99.

Arom, Simha, 1991. African Polyphony and Polyrhythm: Musical Structure and Methodology. Cambridge: Cambridge University Press.

Arom, Simha, 1992 “ ‘À la recherche du temps perdu’: Métrique et rythme en musique.” In Les Rythmes : Lectures et théories, ed. Jean-Jacques Wunenburger. Paris : L’Harmattan, 195–205.

Arom, Simha, 1998. “‘L’arbre qui cachait la forêt’: principes métrique et rythmiques en Centrafrique.” Revue belge de musicologie 52: 179–195.

Arom, Simha, 2004. “L’aksak, principes et typologie.” Cahiers de musiques traditionnelles 17: 11–48.

Arom, Simha, 2005. “L’organizzazione del tempo musicale. Saggio di una tipologia.” In Enciclopedia della musica, Vol. V, ed. Jean-Jacques Nattiez. Torino : Einaudi, 1087–1103.

Arom, Simha and Denis-Constant Martin. 2015. L’Enquête en ethnomusicologie: préparation, terrain, analyse. Paris: Vrin.

Blom, Jan-Petter and Tellef Kvifte. 1986. “On the Problem of Inferential Ambivalence in Musical Meter.” Ethnomusicology 30(3): 491–517.

Chernoff, John. 1991. “The Rhythmic Medium in African Music.” New Literary History 22(4): 1093–1102. http://www.jstor.org/stable/469080

Estreicher, Zygmunt. 1964. “Le rythme des Peuls Bororo.” Les Colloques de Wegimont IV, 1958–1960. Paris: Les Belles Lettres, Bibliothèque de la Faculté de Philosophie et Lettres de l’Université de Liège, Fasc. 172I: 185–228.

Jones, A. M. 1959. Studies in African Music Vols 1 and 2. London: Oxford University Press.

Kauffman, Robert. 1980. “African Rhythm: A Reassessment.” Ethnomusicology 24(3): 393–415.

Koetting, James. 1970. “Analysis and Notation of West African Drum Ensemble Music.” Selected Reports in Ethnomusicology 1(3): 116–146.

Kubik, Gerhard, 2010. Theory of African Music, Vol. I. Chicago: University of Chicago Press.

Kvifte, Tellef. 2007. “Categories and Timing: On the Perception of Meter.” Ethnomusicology 51(1): 64–84.

Le Bomin, Sylvie. 2004. Musiques Bateke – Mpa Atege (Gabon). Paris : Sépia. 

Le Bomin, Sylvie. 2017. Systématique, catégorisation et évolution musicale en Afrique centrale. Habilitation to Direct Research, ED 227 MNHN et SU.

Locke, David. 1982. “Principles of Offbeat Timing and Cross Rhythm in Southern Ewe Dance Drumming.” Ethnomusicology 26(2): 217–246.

Locke, David. 2010. “Yewevu in the Metric Matrix.” Music Theory Online 16(4). http://www.mtosmt.org/issues/mto.10.16.4/mto.10.16.4.locke.php

London, Justin. 2001. “Rhythm.” In The New Grove Dictionary of Music and Musicians, 2nd edition, ed. J. Tyrrell and S. Sadie. Oxford: Oxford University Press, Vol. 21: 277–309.

London, Justin. 2004. Hearing in Time. Oxford: Oxford University Press.

Nketia, J. H. Kwabena. 1974. The Music of Africa. New York: Norton.

Pantaleoni, Hewitt. 1972. “Three Principles of Timing in Anlo Dancing Drumming., African Music 5(2): 50–63.

Temperley, David, 2000. “Meter and Grouping in African Music: A View from Music Theory.” Ethnomusicology 44(1): 65-96.

 

[1]. The Appendix illustrates the assertions made in this section, by comparing two excerpts of transcriptions of the same piece that follow the two types of opposite principles that I mentioned.

[2]. In the rewritten version, the first occurrence of all strict ostinato parts is enclosed in square brackets.

[3]. As a reminder, Gankogui and Axatse—which are both timelines—share the same metrical framework of four dotted quarter notes. What differentiates them is their rhythmic pattern; this confirms, if proof were needed, that a timeline is not a “metrical structure,” but definitely a rhythmic pattern.