ThesisPDF Available

The Syntax of ‘Clave’ – Perception and Analysis of Meter in Cuban and African Music


Abstract and Figures

The continuity of cultural expression between Africa and the Americas has been represented by scholars in a number of disciplines. This study is intended to add to the musicological investigation of this subject by pointing to a fundamental unity in form of related music-syntactical qualities. Connecting the concept of clave shown operating as the deep-structural grammar of melo-rhythmic organization within Cuban music with West African musical practice, this thesis indicates a more encompassing cognitive significance for clave – encapsulated by a proposed syntax of ‘clave.’ It is asserted that this newly defined notion of ‘clave’ has important theoretical implications to the body of African musical practices in general, but in particular may serve as a cognitive model for the analysis and interpretation of other diasporic idioms, as well as for the large body of sub-Saharan music that has been identified in having had a major influence on musical practices in the Americas. With regard to the perception of temporal centrality and metrical dynamism for the African cultural sphere at large, it is suggested that there exists a unique structural rhythmic essence within its traditions that can be intuitively sensed by enculturated listeners and performers. The discussion originates in the author’s intuitions about such parallels, formed through extensive professional performance experience, and builds on previous scholarship concerning the structure and cognitive significance of African phrasing referents, also known as timelines. In particular, a fundamental correspondence among these timelines, due to their unified syntactical organization, is proposed. This syntactical dimension is defined in the form of a paradigmatic 5-note pattern resembling the Cuban clave, which functions as a “blueprint” for multiple morphologies. This background pattern, whether externalized or not, is experienced as a fundamental, internal metrical dynamism, imbuing the music with a particular kinetic energy and contrapuntal organization that extends into both temporal and tonal dimensions.
Content may be subject to copyright.
The Syntax of ‘Clave’ – Perception and Analysis
of Meter in Cuban and African Music
A thesis
submitted by
Bertram Lehmann
In partial fulfillment of the requirements
for the degree of
Master of Arts
August 2002
Adviser: David Locke
© 2002, Bertram Lehmann
The continuity of cultural expression between Africa and the Americas has been
represented by scholars in a number of disciplines. This study is intended to add to the
musicological investigation of this subject by pointing to a fundamental unity in form of
related music-syntactical qualities. Connecting the concept of clave shown operating as
the deep-structural grammar of melo-rhythmic organization within Cuban music with
West African musical practice, this thesis indicates a more encompassing cognitive
significance for clave – encapsulated by a proposed syntax of ‘clave.’ It is asserted that
this newly defined notion of ‘clave’ has important theoretical implications to the body of
African musical practices in general, but in particular may serve as a cognitive model for
the analysis and interpretation of other diasporic idioms, as well as for the large body of
sub-Saharan music that has been identified in having had a major influence on musical
practices in the Americas.
With regard to the perception of temporal centrality and metrical dynamism for
the African cultural sphere at large, it is suggested that there exists a unique structural
rhythmic essence within its traditions that can be intuitively sensed by enculturated
listeners and performers. The discussion originates in the author’s intuitions about such
parallels, formed through extensive professional performance experience, and builds on
previous scholarship concerning the structure and cognitive significance of African
phrasing referents, also known as timelines.
In particular, a fundamental correspondence among these timelines, due to their
unified syntactical organization, is proposed. This syntactical dimension is defined in the
form of a paradigmatic 5-note pattern resembling the Cuban clave, which functions as a
“blueprint” for multiple morphologies. This background pattern, whether externalized or
not, is experienced as a fundamental, internal metrical dynamism, imbuing the music with
a particular kinetic energy and contrapuntal organization that extends into both temporal
and tonal dimensions.
I am greatly thankful to my adviser David Locke, whose suggestions and diligent reviews
were critical to the development and formulation of the ideas brought forward in this
thesis. I also would like to thank Doug Freundlich and John McDonald for their time and
effort as members of the defense committee, as well as Tomie Hahn and Kofi Agawu for
their inspiring seminars during my studies at Tufts University. And I am of course
indebted to my many wonderful professional colleagues and friends that over the years
helped me, by learning from and with them, to be in the position to attempt an intellectual
endeavor such as this.
Bertram Lehmann
Arlington, Massachusetts, Summer 2002
Table of Contents
Abstract ..............................................................................................................................ii
Acknowledgements........................................................................................................... iv
Table of Figures............................................................................................................... vii
List of Audio Examples....................................................................................................ix
Chapter 1............................................................................................................................ 1
Introduction ................................................................................................................... 1
An original analytic stance and methodology ............................................................. 1
On the usage of commercially available aural data as a foundation of this discussion
..................................................................................................................................... 4
Issues of notation......................................................................................................... 5
General outline of this study ..................................................................................... 10
Fundamental Considerations of a Musical Syntax................................................... 13
Chapter 2.......................................................................................................................... 16
Clave in Cuban Music................................................................................................. 16
Perception of clave in context of the Ensemble Thematic Cycle.............................. 16
Syntactic features of clave in the context of Cuban music ....................................... 22
Paradigmatic musical examples of the “concept of clave” in Cuban music ............. 25
Clave - a “Syntactic Field”? ....................................................................................... 32
The West African Bell Pattern ................................................................................... 35
Previous cognitively oriented research ..................................................................... 37
Correspondences in morphology and syntax among West African timeline patterns
................................................................................................................................... 39
Drum language and the question of centrality .......................................................... 40
Chapter 3.......................................................................................................................... 48
A Comparative Overlay of Selected Ternary Patterns and the Notion of the
Syntactic Paradigm ..................................................................................................... 48
Bisectional periodicity and the “Syntactic Arc” ....................................................... 53
Flexibility in temporal categorization ....................................................................... 55
The “Syntactic Arc” .................................................................................................. 57
Cognitive aspects of timelines – the concept of invariants .......................................59
Definitions of a Paradigmatic Syntax of ‘Clave’ ......................................................61
Further aspects of syntactic coherence...................................................................... 64
Examples of paradigmatic pattern integration in Cuban music ................................ 65
Horizontal harmony................................................................................................... 74
Intonation type........................................................................................................... 76
Perception of syntactic “well-formedness” in Cuban Arará ..................................... 79
Aspects of schematic categorization – an example of timbral Stratification ............ 82
Paradigmatic Interpretations of Musical Examples ................................................ 85
Hierarchical levels of syntactic categorization.......................................................... 92
Chapter 4.......................................................................................................................... 96
The Syntactical Quality of Timelines with a Cyclical Length of 24 pulse units .... 96
Examples from Benin................................................................................................ 96
Syntactic resonance between Pattern A and the paradigm....................................... 99
Comparison of transpositions of Pattern A with the paradigm - Resultant Metric
Resonance................................................................................................................ 103
The interplay of Pulse Ground Alignment and Timeline Ground Alignment......... 107
Aspects of syntactic stability................................................................................... 110
Pattern A’s paradigmatic functionality ................................................................... 111
Discovering a set of two related metrical perspectives ........................................... 114
Conceptual analogies to the tonal dimension .......................................................... 118
Chapter 5........................................................................................................................ 121
Critical Reviews......................................................................................................... 121
Simha Arom ............................................................................................................ 121
William Anku.......................................................................................................... 130
Musical syntax and the unity of a musical community ......................................... 139
Conclusions and Further Implications of a ‘Syntax of Clave’.............................. 143
Appendix ........................................................................................................................ 146
Bibliography .................................................................................................................. 147
Discography ................................................................................................................... 158
Table of Figures
Figure: Page:
1 – Son and rumba clave 10
2 – Basic melo-rhythmic contour in rumba guaguancó 19
3 – Abakua basic ‘Ensemble Thematic Cycle’ 19
4 – Comparision of ternary and binary clave 21
5 – ‘3-side’ indicator 25
6 – Manteca horn riff 26
7 – Mozambique Campana 26
8 – Quinto “ride” 27
9 – Timba Bass Drum Pattern 27
10 – La Expressiva (Introduction) 28
11 – Gandinga, Sandunga, Mondongo 28
12 – West African bell pattern 35
13 – West African bell pattern variations 40
14 – Comparative Overlay 49
15 – Syntactic “wave” 58
16 – Syntactic “arc” 59
17 – Comparsa pattern integration 66
18 – Comparsa bell “skeleton” 68
19 – Brazilian paradigm 69
20 – Mozambique pattern integration 69
21 – Reverse comparsa pattern integration 70
22 – Cascara and clave integration 72
23 – Cascara accent pattern 74
24 – Arará (“Song for Ebioso”) 80
25 – Arará motivic unit 80
26 – Nagbiegu 88
27 – Mariama 88
28 – Ekun Rere 89
29 – Ekun Rere (stop-time section) 90
30 – Fantan Ni Mônè 93
31 – N’goni introduction to “Fantan Ni Mônè” 94
32 – Pattern A (vinli bell) 97
33 – Pattern B 98
34 – Pattern A commetric section A 102
35 – Pattern A commetric section B 102
36 – 1st transposition of pattern A 104
37 – 2nd transposition of pattern A 104
38 – 3rd transposition of pattern A 105
39 – 4th transposition of pattern A 105
40 – 5th transposition of pattern A 105
41 – 6th transposition of pattern A 106
42 – 7th transposition of pattern A 106
43 – Pattern A (PG and TLG alignment) 108
44 – Pattern C 112
45 – Pattern C (alternate metrical perception) 112
46 – Pattern C and pattern A (alternative alignment) 113
47 – Arará Afrekete (“Alua”) 114
48 – Paradigmatic alignment of pattern C and pattern A 115
49 – Paradigmatic integration of pattern C and pattern A
from two points of centrality 116
50 – Bòndó background paradigm 122
51 – Dì.kέto background paradigm 123
52 – Revised dì.kέto alignment 126
53 – Revised dì.kέto and mò.kóngò alignment 127
54 – Reproduction of Asante adowa bell pattern 133
List of Audio Examples
Track #1 – Abakuá
Track #2 – Quinto “ride”
Track #3 – “La Expresiva”
Track #4 – “Gandinga, Sandunga, Mondongo”
Track #5 – Comparsa Pattern integration
Track #6 – Comparsa bell “skeleton”
Track #7 – Brazilian Paradigm
Track #8 – Mozambique pattern integration
Track #9 – Reverse comparsa pattern integration
Track #10 – Cascara and clave integration
Track #11 – Cascara accent pattern
Track #12 – Arará (“Song for Ebioso”)
Track #13 – “Nagbiegu”
Track #14 – “Mariama”
Track #15 – “Ekun Rere”
Track #15 – “Ekun Rere” (stop-time section)
Track #16 – “Fantan Ni Mônè”
Track #16 – N’goni introduction to “Fantan Ni Mônè”
Track #17 – Pattern A (vinli bell) – fast
Track #18 – Pattern A (vinli bell) – slow
Track #19 – Pattern B
Track #20 – Pattern A and timeline
Track #21 – Pattern A and tactus
Track #22 – Pattern C
Track #23 – Arará Afrekete (“Alua”)
The Syntax of ‘Clave’ – Perception and Analysis of Meter in
Cuban and African Music
Chapter 1
An original analytic stance and methodology
I can look back to my mid-teenage years to find some of the origins of my long-
standing fascination with some of the issues we will encounter in this study. Of course at
that time I was only in the beginning stages of my musical development, yet I recall
already reading literature by authors such as Paul Watzlavick, Ireneus Eibl-Eiblsfeld and
Helga de-la-Motte-Haber concerning general topics of perception, cognition, behavioral
studies, and even music psychology in particular. Later, as a professionally performing
musician, this academic interest transformed into an awareness of a more practical nature.
During the past decade I also encountered new stylistic contexts that not only presented
instrumental challenges, but demanded my continuous cognitive re-adjustment as well.
For a drummer like me, a conceptual understanding of a musical style and its
associated rhythms is undoubtedly of paramount importance1. Such understanding (others
might call it “feeling” or “getting-in-the-groove”) is often a matter of hearing the music
in the “right” way, which sometimes requires me to spontaneously modify prior musical
A weak drummer’s performance in this respect can derail a musical performance more than
anything/anybody else.
intuitions that would otherwise sabotage my “navigation” of a musical texture. As a
result, my cultural growth often occurs in real-time during a rehearsal or a performance.
Ostensibly cerebral discussions about cognitive issues2 therefore do not only have
great relevance from the performer’s point of view, but cognition in general represents a
central component of a performer’s on-stage experience, while also contributing to her
“intuitions” during subsequent analytical reflection. My discussion of African music thus
begins with an attempt to reconcile my idiosyncratic intuitions gathered through intensive
practical involvement3 in African musical styles with the particular theoretical issues at
hand.4 As my thesis advisor David Locke once mused: “Everything is in the awareness of
the practitioners . . .!”5 In this sense, the heuristic integrity of this work is based
significantly on my experience as a performer and a listener.
“U.S. blues and jazz musicians often occupy a middle position. Neither African
nor Western musicians, they have figured out ways to reconcile the African and Western
concepts,” suggests Gerhard Kubik in Africa and the Blues (1999). I affirm that this
middle ground has yielded the particular analytical position from which my observations
Nzewi (1997: 24) also remarks on the centrality of this issue for a researcher’s “em-etic
approach” who is “capable of objective self-assessment”: “Cognition should be predicated on total
facultative and emotive submission to the affective object of perception. We, therefore, need to be seeking
a researcher/analyst of African music who is at the confluence of objective observation (by virtue of correct
and applicable scientific training) and subjective perception (by virtue of enculturation or genuine
penetration of the cultural intentions and creative manifestations of a cultural phenomenon).”
Washburne (1995) makes a similar point when stating that “I have determined to wear,
simultaneously, the hats of the researcher (fieldworker) and the actively performing musician. Thus,
playing salsa music has become my primary methodological tool . . . the process of data collection is often
conducted ‘on the bandstand’ . . . by the use of trial and error, imitation of insider’s behavior, and informal
‘hanging out’ sessions with other performers. [. . . ] My data acquisition process as both a performer and
researcher are then one and the same. Both require an acute awareness of one’s surroundings, an ease in
adapting to new situations, and an ability to organize, analyse, and interpret a large body of information.”
Aside from insights into African musical conceptions gained through their presence in jazz
drumming – my primary and so far most extensive account of experience – I have benefited from
performances in all kinds of music of the African diaspora, including Cuban-, Brazilian-, Trinidadian-, and
Afro-Argentinean traditions as well as selected popular and traditional African styles.
David Locke, lecture at Harvard University, March 3
, 2000.
originate.6 For it is at the meeting points of musical cultures that their individual traits
tend to be thrown into especially sharp relief. As failure may heighten a performer’s
awareness about musical structures7, the mistakes and confusion that arise in such
contexts are highly revealing of fundamental cultural cognitive biases involved.
John Sloboda (1985: 38) has proposed, that “[s]yntax becomes, in itself, an object
of aesthetic awareness.” My proposal for a syntactically oriented interpretation of African
rhythmic formulae is thus stimulated by an awareness of the aesthetic commonalities
arising primarily from such a syntax, and drawing on my practical experience with a
number of very specific musical styles from the African cultural sphere I will characterize
and verbalize this dimension of the music. In general, this study depicts the wider
patterns of musical organization and comprehension instead of being a meticulous
ethnographic analysis of a particular locale. Guided by my subjective judgment, I have
aimed to reconcile my original notions with Western and African concepts in the musical
Recounting his own experience of learning to play the Bulgarian gaida, Timothy Rice (1997)
argues eloquently for a new kind of “mediation . . . between the epistemological, methodological work of
explanation in the field and the ontological understandings of human and musical experience in the field
(106),” in which the researcher inhabits “a place untheorized by the insider-outsider distinction so crucial to
much ethnomusicological thinking (110).” Similarly, with respect to my methodological stance, I would
like to echo his notion of “Phenomenological Hermeneutics as a Foundation for Ethnomusicology (114)”,
whereby “[u]nderstanding, in this tradition, precedes explanation rather than being the product of it (115),”
and claim certain effects of ontological “self-transformation (116)” over the course of my performative
career : “Whereas Enlightenment philosophy leaves us with a certain confidence in a rational and fixed
subject moving through the world, analyzing and in some sense controlling it while keeping it at a distance,
hermeneutics suggests that the subject becomes a self in the encounter with the world of [musical] symbols
Examples of this process can, for instance, be seen in the technical difficulties musicians often
encounter when attempting to master music with a temporal matrix other than the fairly straight pulsation
often employed in Western art music (but not only there, cf. Polak 1998: 2), or even in much Pop and Rock.
For example when asked to play a Brazilian samba, many otherwise proficient, professional drummers are
often not able to transfer their overall motor control into this music’s typical lilting rhythmic structure.
Likewise, the highly differentiated timbral-melodic sequences common in Afro-Cuban drumming (i.e. in
the basic tumbao) often wreak havoc on an “uninitiated” player, who initially might have only an unclear
notion of their fundamental beat structure and metrical implication.
On the usage of commercially available aural data as a foundation of this discussion
Any a posteriori transcription of a musical performance will inevitably reflect the
transcriber’s idiosyncratic cognitive stance. One of the major drawbacks in the discussion
of any musical event whose performance does not originate from a prescriptive notational
template (i.e. a score ) provided by a composer, is its limited accessibility for verification
on part of the community of readers. This not to say that the musical structure perceived
in the aural dimension will necessarily match its symbolic representation on paper any
more closely in the former than in the latter case. But if we can imagine music’s
structural properties based on any notation, a composer-generated score is more
“authoritative”, i.e. closer to an idiosyncratic conceptual source, than an analyst’s
Hence, in the study of African music, one is not only confronted with the virtual
absence of such notational texts, but also the absence of the kind of mutually shared
listening experience and stylistic knowledge that is taken for granted in the musicological
discourse about Western art music. Furthermore, the sheer diversity of African traditions
on the continent and their extensions in the Americas makes familiarity with more than a
few unlikely, even among specialists. Consequently, while the landmark studies in
Africanist musical ethnography and theory rightly originated out of a large body of
personal fieldwork data collected by their respective authors (e.g. Nketia 1974, Jones
1959, Kubik 1994, Arom 1991, Euba 1990), in most cases their readership has had little
access to any of the concrete aural data at the source of a scholar’s observations.8 As we
shall see below, this has in some instances been limiting the potential for critical review
However, in keeping with new technological advances, inclusion of such data has become more
common these days, as in for example in Agawu (1995), Locke (1992), Kubik (1994), who all provide an
accompanying CD or tape of audio examples.
of the notions advanced by such works. Because first-hand verification about the degree
and type of an author’s mediation and cognitive abstraction of the sonic texture is
impossible, critical remarks tend to be confined to issues of the methodology employed,
academic scope, and a study’s general conclusions.9 Insofar as we are willing to assert
that the formulation of a theory about orally transmitted music is intrinsically tied to
musical cognition, this presents an unsatisfying situation.
Thus in this study, I will use sources that can be obtained commercially so that
my arguments can be related to a repertoire that is “open to the public”, so to speak.10
These are included on the accompanying audio CD, along with recordings of selected
examples performed by the author. While there are obviously drawbacks working with
musical examples that were not witnessed in person, I believe such methodology to be
appropriate and the examples’ illustrative salience to be sufficient for the purpose of this
study. Incidentally, the practice of working extensively with recorded sources of
paradigmatic stature has been (arguably successfully) used in musical analysis and jazz
education for decades.
Issues of notation
To retain visual clarity, I keep notation to the minimum needed to make a point. Since in
any case the effects of the syntactical phenomena under discussion can in only be
Kofi Agawu, in “Representing African Music” (1992), warns against an additional political
motivation in attributing original fieldwork an “iconic status”: “Instead of creating a public forum for the
open confrontation of ideas – such as publishing competing transcriptions of the ‘same’ musical objects, or
developing alternative interpretations of the same system of cultural expression – ethnomusicologists
writing on Africa have been more interested in establishing ownership of unstudied or little-studied musical
areas than in going over previously charted terrain (259).”
See a similar justification by Agawu (1986), who emphasizes a reader’s “opportunity to work
closely with the sound, indeed to verify the transcriptions and concepts discussed in this essay (65)” as
important to such methodology. For the same reason recent performance-oriented studies such as Locke
1990, 1992, and 1998 have been published with accompanying audio examples.
experienced in operation, I find schematic “extraction” of musical features to work better
than complete transcriptions. The reader must refer to audio examples whenever
indicated, as the notation only serves as a cognitive “highlighter,” so to speak, intended to
focus attention on the relevant parts of the musical texture. As such, these notations are
not intended to capture the kind of textural details found in ethnographically oriented,
notationally comprehensive representations.
In general, I follow established notational conventions, taking what Eric Charry
(2000: xxvii) has aptly called a “path of least resistance,” i.e., a “fundamental meter
comprising a fixed number of beats (and pulses making up those beats) that is simplest
. . . to feel and that usually does not change during the course of a single piece.” For
instance, I use a 12/8 time signature corresponding to the length of one timeline cycle
because the notion of a syntax of ‘clave’ is mostly explored from the perspective of its
relationship to a West African timeline morphology fundamentally rooted in an implicit
ternary pulsation subdividing each tactus beat. This 12/8 time signature must be taken
only in its explicitly numerical sense: it indicates a pulsation of 12 time-points based on
the durational value of an eighth note. As we shall see, I will suggest below that these
timelines are experienced in the context of a metrically “sculpted” underlying 4-beat
period. This primary feature is represented separately in form of schematic tactus beats,
shown below each the notation of each timeline cycle.
Overall, I have tried to avoid the visual “clutter” of a notation that always
indicates beat units by, for example, continuously beaming to the beat, and requiring rests
or ties that represent “empty” pulsations. While this practice effectively shows the exact
position of rhythmic events in relation to every pulse unit as well as simultaneously in
relation to the primary metrical frame of reference assumed, to some degree it also ties
one’s analytic perception to that frame. Instead, I have opted for what Simha Arom
(1991) calls a “rhythmic notation”, in which each duration is represented “by a single
sign, independently of whether it crosses from one pulsation to another”(228). Not only
does it look clearer, but once a pattern’s fundamental implication with the tactus level
has been understood, a syntactically accurate visualization can be achieved. For this
reason, rests are, with some exceptions, kept to a minimum. Ties have been avoided
altogether, save for certain ensemble transcriptions where they indicate an actually
sustained note, and in figures 36-42 where I have intended to keep the transposed
subunits’ internal durational structure more visible to the reader.
Although this study emphasizes a ternary metric context, the syntactic quality of
these rhythms clearly transcends that singular plane. When we cognitively experience
clave in form of a holistically perceived gestalt, any particular of the two basic metrical
orientations, implying an either binary or ternary grouped pulsation, looses significance
and should be de-emphasized. I prefer therefore to indicate this syntactic dimension
simply in this notationally “leaner” manner, where synchronization with the tactus level
is nevertheless available yet appreciation of the rhythmic structure is enhanced. I also
find this practice to better reveal multiple horizontal pattern relationships, as for instance
shown in the “Comparative Overlay of Selected Ternary Patterns” in Chapter 3.
I do however have one small point of modification with regards to what otherwise
may strike the reader as a beaming practice alluding to a 6/4 counter-meter: the
syntactically relevant mid-point of a timeline, and consequently that of the proposed
paradigmatic pattern itself, is marked by an eighth note rest so that readers will be aware
of the bisectional oscillation inherent in this structure. Likewise, in cases showing the
particular version of the clave used in rumba and other folkloric Cuban genres, the
pattern’s third note will be preceded by a rest that highlights the difference in position
from the other version’s third note (e.g., figures 1 and 3), and avoids the use of either tied
or double-dotted note values. Furthermore, the final, fifth attack of the ‘clave’ paradigm
proposed further on, which from the perspective of a notation in 6/4 would technically
need to be indicated as an off-beat attack, is instead visually synchronized with the 4-feel
of the tactus. By combining notational practices in this way, I have attempted to
graphically encode some aspects of this pattern’s syntactic directionality.11 Thus,
regarding specifically the notation of this paradigm and its related timeline patterns in this
study, the use of rests or ties is limited to illustrate analytical points relating to its
syntactic quality, and at times may break with the rigorous notational consistency perhaps
expected by the reader. Notations of other instruments’ parts, for instance in the ensemble
transcriptions, may follow more standard conventions.
Another liberty taken concerns the notation of examples that conform to a binary
metrical framework. Some may be seen employing a 2/2 (or “cut-time”) representation,
while others are shown in 4/4 time. In the first case one fundamental felt beat corresponds
to a half note, so in the case of a pattern spanning four such beats two bars are needed to
represent one full cycle. In the second case one beat equals a one quarter note, and only a
single bar is needed. Both practices are valid and common among musicians to notate the
kind of music under discussion, as far as it is notated at all.
That there can definitely be “flux” if not conceptual evolution with respect to notational
conventions not only within the musicological discipline as whole, but also within an individual scholar’s
practice should come as no surprise. A nice example of two different stances taken can be seen in Locke’s
manner of indicating the relationship between melody and bell pattern timeline in Ewe song (cf. Locke
1978: 151ff. and 1996: 97ff.).
The 2/2 fashion is often used in Latin-American musical styles such as salsa and
the numerous types of Latin-jazz. Due to the fact that it implies a subdividing pulsation of
eighth notes, it retains a high level of visual clarity and readability. This may be preferred
by the jazz-trained musicians often employed in these situations, in particular by horn
players that tend to have a background in big-band music where rhythmic notation mostly
consists of quarter and eighth note values. In any case, contemporary salsa arrangements,
small-group lead sheets, and many of the instructional publications on Latin music
nowadays utilize this format, especially when influenced by U.S.-based musical practices
and stylistic conventions.
However, notation indicating the quarter note as the fundamental unit of metrical
experience is also common, especially in the Latin American regional sphere itself. In
Cuba and Brazil, for example, this manner historically was used to publish sheet music of
native styles such as the danzón and son, and the choro and samba, respectively. And if
recent personal experience by the author is taken as an indicator, it continues to be
employed in both professional publishing and private use in these countries.12
The idea within these pages has simply been to familiarize the reader with both
conventions, sharpen his visual acuity to recognize and be able to identify paradigmatic
structures in both “disguises”, should he encounter them in other publications or in actual
musical practice.
For example, as far as Brazil is concerned, we find that sambas as well as samba-influenced
contemporary pop music continues to be notated in 2/4 time (e.g., in the ever-popular Songbook series
[Lumiar Editora, Rio de Janeiro] containing compilations of bossa novas, sambas, or music by individual
artists such as Djavan). Likewise, Cuban pianist Tony Peréz, a renown arranger and former member of
groups such as Irakere or vocalist’s Isaac Delgado’s ensemble, uses 4/4 time to write out his music
(personal communication). It has also been used in academia in this country, as the works of scholars such
as Fernando Ortiz (cf. 1950) attest to.
General outline of this study
Almost exclusively, the concept of clave13 so far has been discussed in the scholarly
literature within the particular context of traditional Cuban folkloric styles such as son,
rumba or the batá genre14, or their more contemporary offshoots within and outside the
Cuban island that are known to much of the world by the umbrella term salsa.15 In these
cases the focus is usually on a pattern called son clave or rumba clave (figure 1), which is
often performed by striking together two wooden bars that bear the same name – claves.
Figure 1 – Son and rumba clave
Yet, that the importance of the clave for the music-making process in both spontaneous
and compositionally mediated contexts in Cuban music extends beyond this rather
obvious foreground physical prominence might come as a surprise to many a non-
In this thesis, the italicized word clave refers to usage and meaning in the specific context of
Cuban music, while I use ‘clave’ whenever speaking in terms of its more abstracted function a syntactic
force within African music at large.
cf. Amira (1992: 23), Gerard (1989: 13-32, 60-71, 83).
cf. Mauleon (1993), Gerard (1989), Tozzi (1996 and 1999), Washburne (1998, 1995).
enculturated listener16, even though it is at the center of musical awareness, performance
conceptualization, and aesthetic judgment by enculturated culture exponents.17
The notion of clave as a structural phenomenon in the organization of primarily
the temporal dimension of such music already has also received some attention by
academic scholarship in recent years, and in rare cases scholars have extended their
analyses to related idioms such as jazz, or point to this notion’s historical and structural
connections to Caribbean music at large.18 And while occasional references indicate an
awareness of a nexus between features of African music and its New World extensions –
in particular concerning the notion of timelines or phrasing referents19 – Africanist
scholars have yet to significantly connect their work to the diaspora in the Americas.
In the following analysis I posit that clave operates as a sort of “deep structure”
of melo-rhythmic20 organization within the psychological, cognitive and aesthetic reality
I understand this listener to be an, of course rather idealized, individual, perhaps enculturated
instead in the folk or art music traditions deriving from a European/North American or Asian heritage, who
has had only limited exposure to music of specifically Cuban and/or otherwise immediate African heritage.
Gerard (1989) states that clave acts as a “centrifugal force throughout the performance of a
piece (13),” generating a “clave feeling [that] is perceptible in the phrasing of the melody and in other
percussion patterns” and as such “in the music whether or not the claves are actually being played”(14).
Washburne (1998) points out that “[i]n Latin music terminology the word clave refers . . . not
only to the instruments [i.e., two wooden sticks] but also to the . . . underlying rules which govern these
patterns. Competent salsa musicians develop a ‘clave sense’ . . . , whereby a subjective pulse serving as an
ordering principle is felt by the participants in a musical event”(162). Tozzi (1999) asserts that “clave
functions as a timeline and lets dissimilar rhythms appear coherent, since because of the existence of a
structuring timeline all polyrhythmic events are directed into a syntactical relationship. Afrocuban music
distinguishes between tradition-conforming and non-authentic rhythms on the basis of a tradition and
‘Sprachgemeinschaft’ [i.e., linguistic community] sprung from such syntax” (150). And Monson (2000) has
noted the haphazard use of clave in certain recordings by Art Blakey where “the ensemble is not really
‘thinking’ in clave”(339), but instead tends to “borrow rhythmic patterns that are ‘in clave’ without fully
internalizing clave as a structural principle of phrasing and rhythmic combination”(345).
E.g., see Kubik (1986: 126-129), Olivera-Pinto (1991: 190), Washburne (1995).
Nzewi (1974: 24) has proposed this terminology as referring to “a rhythmic organization that is
melodically conceived and melodically born . . . having a different orientation than the kind in which the
rhythm of a music has a more independent derivation and function. In West African folk music the rhythms
of the percussion instruments are firmly rooted in the melo-rhythmic essence, not in the abstract
depersonalised percussion function typical of Western percussive style.” He elaborates: “The notion of
rhythm as statistical computation which can have independent structure and performance existence, thereby
constituting a musical presentation as percussion, does not . . . belong to the African philosophy and
of Cuban music. Further, I argue that an expanded notion of ‘clave’ as a more widespread
syntactic force can serve as a cognitive model for the analysis and interpretation of other
diasporic idioms, as well as for the large body of sub-Saharan music itself that has been
identified in having had a major influence on musical practices in the Americas. I want to
suggest that there exists a certain structural rhythmic universality or essence21 within such
music, that, while not necessarily equally prominent in all cases, can be intuitively sensed
by members of an African “shared music contract,” as Harold Fiske (1996: 147) might
call this particular native cognitive stance.22 My observations are built a) on my own
subjective intuitions about such cognitive parallels, and b) on other scholarship that links
the morphological features of the Cuban clave (and it’s functionally similar siblings from
other diasporic cultures) with African timelines – a handy term coined by J.H.K. Nketia
(cf. 1974:131) to label the cyclic phrasing referents often played on a single idiophone
such as an iron bell.
Thus, the novel notion of this study lies in my assertion that a deep-structural
grammar of ‘clave’ has perhaps more cognitive salience, and thus analytic-theoretical
relevance, than usually recognized. I propose that such a grammar can apply to musical
situations beyond the particular context of Cuban musical expression, and to idioms
without an externalized timeline. In fact, I will try to show that ‘clave’ – rather than being
practice of music. [. . .] For the African, drums and bells are melorhythm instruments, not percussion”
(1997: 32ff.).
Nzewi (1997: 74) similarly posits the existence of “general, fundamental principles of creativity
and practice which enable us to identify a piece of music as African in the first instance, before
categorizing regional or cultural peculiarities.”
Also see Fiske’s (1996) notion of “ hard construction paradigm” (140), i.e., the application of
“culturally-encountered perceptual rules to incoming acoustic energy” (144). Swain (1994) identifies a
musical community as having “an intersubjective agreement about certain perceptual objects” or “beliefs”
(317), and Walker (1990) concludes that “[t]he role of culture and its belief systems in supplying us with
categories for our perceptual apparatus to utilize would seem, therefore, to be of paramount importance in
our auditory perceptual activity”(56).
just one equal component in a polyphonic musical texture – is experienced as a
fundamental, internal structural dynamism, which for the enculturated listener and
performer imbues the music with a particular kinetic energy and contrapuntal
organization, extending into both temporal and tonal dimensions.
Furthermore, in my understanding the notion of a ‘clave’ as an African cognitive
construct or “belief”23 points beyond its particular conceptualization in Cuban culture.
‘Clave’, I assert, has important theoretical implications to the body of African musical
practices in general,24 rivaling the structural salience usually attributed to harmonic
rhythm in European music. But at first it may be appropriate to discuss some fundamental
issues concerning the invocation of the theoretical construct ‘syntax’ for my purposes,
and to clarify its terms of usage.
Fundamental Considerations of a Musical Syntax
Joseph Swain (1995) provides us with a detailed explanation of the concept of syntax as it
may be applied to music in general, some of which I would like to abridge below:
syntax “reduce[s] information . . . the randomness of elements in the incoming
sound stream . . . constrain[ing] the order of items in general”(285), and
“extracting patterns . . . [from] the flow of information streaming through the
ears in real-time”(287).
as in linguistics, musical syntax works by “restricting admissible sounds” (i.e.,
pitches and rhythmic values) in a process of selecting “categorically discrete
In the sense discussed by Walker (1990) and Fiske (1996).
In trying to distinguish between clave in its particular Cuban conception and the general notion
of a culturally more encompassing syntax of ‘clave’, I have resorted to the italicizing and use of single
quotation marks, respectively.
sounds . . . organized into hierarchies . . . [and] bound together by syntactic
rules at various levels”(287).
syntax “now embraces almost any kind of abstracted relationship among
musical phenomena”(281).
syntax characterizes mostly “low-level or local relationships”(281), doesn’t
extend “much beyond the phrase”(283), and is “most powerful . . . not in the
big structural picture but in the details”(300).
“syntax creates its own ever-changing context built out of its own elements,
which then affects how the syntax itself functions”(287).
“[a] key virtue of syntax is . . . [its] unbreakable bond with a sense of
virtuosity, not a showy virtuosity in particular, but the kind that is demanded
as the standard equipment of a highly trained artist as an outstanding member
of a musical community,” it “demands a coordination of its various aspects
that must challenge the technique of a composer [or performer]”(299).
a “shared syntax . . . resides not just in the musical productions . . . but is in
the nature of an agreement between all members, producing or nonproducing,
of the musical community. [. . .] Music that cannot attract, develop, and
sustain this kind of bond can hardly claim to be syntactic”(303).
the “artistic raison d’être of syntax”(303) lies in the way in which it “mediates
tension and resolution in musical languages”(290), and whose differences
“can be heard not only in their means of organization . . . but in the tensions
and resolutions they create”(295).
tension and resolution thus “are essential effects of syntax, conceptually
separate from its function of information reduction but indissoluble from
syntax itself”(295), its “specific kinds . . . derived from cultural associations,
learned by listeners along with the syntax itself,” in which the musical phrase
constitutes the “locus of the most salient effects”(294).
syntax requires “finite discrete elements, rules of combination, absolute and
conventional markers, hierarchical organizations, tension and resolution, a
comprehending community”(304).
just as in linguistics there are “‘tiny’ rules, conditions of syntax that evade
generalization [and] we seem unable to specify completely the syntax of even
the best-known musical languages”(303), which will remain “elusive for the
foreseeable future, if not forever”(304).
The reader somewhat familiar with the structure of African music in general, or
even with the particular notion of clave in the Cuban musical context, might at this point
already have some intuition about the relevance of a syntactic-theoretical interpretation.
As we shall discuss in the musical analyses below, I see a number of the “keywords”
appearing above as having particular pertinence for the conceptualization of a larger,
culturally encompassing syntax of ‘clave’.25 But before we can assess any prevalent
syntactical dimensions, we first need to explain its morphological features as well as the
clave’s particular function in two folkloric styles that perhaps quite poignantly highlight
its properties – the Cuban rumba and abakua.26
A number of scholars have noticed the fundamental function of the clave for at least the Cuban
cultural context, in most cases noting its historical link and structural affinity to certain West African
musical traditions. A classic account is provided by Fernando Ortiz (1950), but in more recent research the
studies by Wolfgang Tozzi (1996, 1999) and Christopher Washburne (1995, 1997, 1998) have been
particularly insightful in focusing on ongoing popular music practice, and suggesting the theoretical
salience of a concept – or syntax – of clave.
As it would be beyond the scope of this study to establish a comprehensive assessment of clave
and its application within the context of an exclusively Cuban performance practice, I would like to refer
the reader to the following publications that define its functional and morphological features specifically
Chapter 2
Clave in Cuban Music
A frequent misconception about the significance of clave in a folkloric Cuban context
may be the impression that it only shapes the morphology of individual layers of a
polyrhythmic texture; that, by somehow resembling the clave’s durational intervals,
alignment or congruence is assured for the ensemble as a whole. However, as we shall
discuss later, individual patterns and the clave in many cases establish a particular
syntactic relationship that goes beyond a mere surface affinity, and that, on the more
fundamental level of a rhythm’s overall gestalt, such relationships within the melo-
rhythmic contour are even more crucial.
Perception of clave in context of the Ensemble Thematic Cycle
Meki Nzewi (1997: 44) has suggested that the circularity inherent in African
rhythm stems from a “traditional philosophy about form: a known quantity that recurs
with a different quality,” accruing “quantitative depth and qualitative energy”(42). His
concept of an Ensemble Thematic Cycle (henceforth referred to as ETC) nicely captures
this notion of a cyclical thematic essence that is “distilled from all the ensemble lines as
the identifying tune/statement/pattern of a piece”(49):
[T]he Ensemble Thematic Cycle is the span of an ensemble gestalt (gross durational content of
differentiated instrumental thematic gestalts) which recurs in essentially the same shape and time
but with continually changing sound quality. (44)
from such angle, and are geared towards performance practice: Rebecca Mauleón, Salsa Guidebook (1993),
and Charley Gerard, Salsa!: The Rhythm of Latin Music (1989).
And further:
[W]e define Ensemble Thematic Cycle as the significant musical form or module by which a piece
of African music is recognised. It is the aggregate sound of the layers of role-themes in an
ensemble. (44)
While it may be sufficient in general parlance to use more common terms such as
“rhythm,” “groove,” or “time-feel,” I find Nzewi’s terminology to be much more precise
and descriptive of the modes of aesthetic affect and cognitive processes that come into
play in the perception of such polyphonic textures.
A distinct “thematic span” arises from what Nzewi terms a “unilineal structure,
[in which] the different ensemble themes assigned to some of the instruments of an
ensemble are fractions of a fundamental melody or melorhythm or both”(45). The author
The mind perceives the cycling layers at simultaneous layers of affect: one primary affect of the
composite sound with defined length and depth, i.e. the Ensemble Theme; and the secondary
affects of the . . . individualistically composed layers of sound with varied length and characters,
i.e. role themes. (56)
In the context of each of the examples below, it is exactly this fundamental theme, or
timbral melody, that is of interest to us. For instance, the vocalization27 included in figure
2 below is an onomatopoetic approximation of the particular thematic essence known in
Cuba as rumba guaguancó. Its gestalt incorporates what Nzewi calls the “significant
notes/fragments of each layer,” (49) in this case mostly the open tones of the individual
parts of two conga drums.
UmBahKeemBahKeemBah…UmBahBahDahUm.” The asterisk in figure 2 marks the starting
point of the vocalization within the notated phrase. See also Gerard (1989: 66).
As can be seen and in the figures 2 and 3, in both cases the clave is part of this
texture and is perceived as a constant, unvarying phrasing referent. Also noticeable is the
particular kind of overall contrapuntal relationship that the compound thematic gestalt
maintains with the clave in each of these examples.28 This structural correlation is
paradigmatic for these types of music and subject to clear aesthetic judgment and
performative awareness29 by the musicians.
In the rumba30 example of figure 2, for example, we also note that the two
higher-pitched tones (the open tones of the segundo) mirror the clave attacks of the ‘3-
side’31 and help establish this characteristic melo-rhythmic theme, whose above-
mentioned vocalization even incorporates the two clave attacks which otherwise are not
part of this theme (see the underlined Bah’s) into one composite mnemonic phrase: Um
Bah Keem Bah Keem . . .”. This vocalization begins where indicated by the asterisk. It is
important to fully appreciate the interlocking of this unilineal melo-rhythmic contour with
the clave pattern as part of a distinct, integrated verbalized conceptualization, to get a
sense of the significance that the clave holds in the perception of this music.
Interestingly, as it has been noted (see Gerard 1989: 67) and can be heard on certain classic
Cubop recordings (e.g. the bigbands led by Cuban bandleader Machito around the 1940’s-50’s), this
particular contrapuntal relationship has itself been subject to a perhaps aesthetically motivated evolution. It
was also common in the past to have the segundo part line up with the ‘3-side’ of the clave (attacks 1,2, and
3) when a rumba rhythm was used. The contemporary folkloric practice however is as shown above.
Joseph Howard (1967: 239) accurately noticed this awareness in the case of Cuban
percussionists in the U.S., which “begun to play jazz only to find that in fast tempo they revert to clave
patterns; and their solos are nearly always within the clave framework. Like a first language, clave becomes
part of the subconscious: musicians brought up in the clave tradition hear music in ‘clave’ and express
themselves musically in reference to clave.”
Rumba is on of Cuba’s most prominent musical genres, a secular percussion-based music that is
often performed at informal gatherings in the street or elsewhere. In most cases, three conga drums are
utilized, the lowest sounding called salidor, the middle one segundo (or tres golpes), and the highest called
quinto. The basic melo-rythmic contour given in fig. 2 results from the particular interplay of open, muffled
and bass strokes played by the salidor and segundo, and the quinto is freely improvised with reference to
this ETC, and the dancing that may accompany it.
‘3-side’ and ‘2-side’ is common parlance among musicians when referring to the asymmetric
structure of this 5-note pattern, which implies a bi-sectional split-point that divides the clave in two “sides”.
The half with 3 notes is thus called ‘3-side’.
Figure 2 – Basic melo-rhythmic contour emerging from the interlocking tones of salidor and segundo
in rumba guaguancó, with parallel mnemonic vocalization and its correlation to the clave.
Incidentally, this vocalization also includes some further attacks by the segundo,
represented by the bracketed syllables “Bah Bah Dah Um, but I have chosen to omit
them in this schematic notation as to not visually obscure the basic theme. They usually
consist of muffled strokes (the “Bah’s” and “Dah”), as well as a low-pitched bass stroke
in the middle of the drumhead (the “Um” coinciding with the clave), and are less salient
to the perception of the fundamental contour.
Figure 3 - Abakuá basic Ensemble Thematic Cycle (“Abacua#3”)
Track #1
The abakuá ETC (figure 3) is particularly tricky to make sense of, so strong may
be the displaced metric “pull” exerted by the obiapá drum, and also by the ekón’s
alternate parsing of the 12/8 cycle, though this impression will depend greatly on the
subjective dynamic balance that a listener will perceive among the individual voices.32
This setting thus particularly highlights the clave’s function as the central cognitive
timing reference (timeline). In practice, listeners/performers focus on the clave to
withstand the draw of this insistent phenomenological counter-accentuation, and maintain
the implicit fundamental tactus33 of four dotted-eighth note beats per cycle.
Another, more general observation is that the morphological integrity of the clave
pattern seems to be the same, regardless of its fundamental surrounding timing matrix.
That matrix can either be characterized as more binary oriented in the case of the rumba,
or more as ternary for the abakuá. The clave, whether transcribed in duple or triplet
meter for analytical purposes, can mold itself onto the rhythm in either way, maintaining
an almost identical internal quality of its phrasing.34 In fact Jeff Pressing (1983: 42-3) has
noted that “by aligning the two patterns [shown in figure 4 below] synchronously and
expressing durations in units of 1/48 of a cycle . . . the basic clavé beat considered to
In this representation based on Los Muñequitos de Matanzas’ “Abakuá #5”, obiapá’s muted and
open tones and the kuchiyeremá open strokes are given a placement respective of their relative pitch levels
in the resultant theme. The ekón is an iron bell.
Throughout this study, I use the term tactus as a unit of duration in the way suggest by Arom
(1991: 206), who explains that “African practice . . . resembles the practice in use in the West during the
Middle Ages and the Renaissance, when the temporal organisation of was based solely on this neutral,
unmarked, intrinsic element . . . providing the performers with equidistant reference points.”
In a folkloric context, a distinction is sometimes made between the rumba clave with a more
binary timing inflection and a ternary sibling often called 6/8 clave. The former is commonly associated
with the rumba styles yambu, guaguanco, columbia as well as comparsa and songo, and the latter reserved
for rhythms like abakua, palo, bembe¸ or certain batá-toques. However, it seems that this terminological
distinction is more the result of a posteriori analytical reflection or pedagogical purpose than practical
necessity, as in practice none of these rhythms is limited exclusively to either one of these metric matrices.
Gerard (1989: 84) asserts that “the 6/8 form of the clave is nearly indistinguishable from the rumba clave,
although triplet values are used here instead of eighth notes.” There are nonetheless subtle timing
differences noticeable, mainly concerning clave attacks three and four (cf. figure 1).
Likewise, the often discussed distinction between a ‘3-side’ position and a ‘2-side’ position is not
made in these primarily folkloric genres. Only in the context of popular music, specifically in relation to a
song’s harmonic motion, does this become an issue (cf. Gerard 1989: 16 ) for a performer. In other words,
guaguanco, for instance, is not a rhythm in either ‘3-side’ or ‘2-side’ position, it just is! Only when its ETC
is merged with the formal harmonic structure of a particular song (e.g. as part of a salsa arrangement) does
it appear to take on such properties, in which case the cognitive forces inherent in the additional syntactic
level of harmonic motion rival the ETC’s own melo-rhythmic dynamism.
underlie virtually all the 4/4 Afro-Latin music of the Americas [i.e. the son clave] . . . is
one of the two closest approximations to 22323 [i.e., the durational intervals of the
paradigmatic pattern] that are possible in an L = 16 cycle (4/4 meter).”
Figure 4 - Comparison of ternary and binary clave (after Pressing [1983: 43])
Examining Pressing’s comparison, we note that based on such forty-eighth note
matrix the difference in inter-attack durations is at most 2/48th units (on the crucial 4th
attack), which, especially at faster tempos, makes for a rather small window of cognitive
distinction. He continues that, however small that distinction, it “is not merely a
theoretical supposition.”
African and Afro-Latin percussionists generally recognize that some patterns from 12/8/ and 4/4
meters may be used in either rhythmic framework (e.g. C.K. Ladzekpo, Pertout).
Furthermore, a number of most clearly African-derived Latin percussion patterns
exist in alternate 6/8 or 2/4 versions, or have such versions of the same pattern sounding
simultaneously [italics mine]. This is found, for example in the Cuban guaguanco and
bata drumming styles, in Brazilian candomblé, and in the rhumba columbia [sic] (Pertout). (ibid.:
And Wolfgang Tozzi (1996: 126) finds that in Cuban music “some rhythms sound as if
they were in 12 (6/8, 12/8) or 16 (8/8, 16/8) at the same time. These forms are therefore
equivalent.” It is this morphological fluidity of the clave, if not of timelines in general,
that appears to me as a particular structural strength – not only does the clave retain its
salience as a phrasing referent in a either case, it also allows for some “fuzzy” (but not
random) timing to take place within an ETC.
Syntactic features of clave in the context of Cuban music
It is of course a step further to claim syntactic qualities for what appears to be
primarily a surface structure in such music. This would imply assigning clave a particular
cognitive prominence among all the other events in a given ETC, creating a hierarchy of
structural strata in which a “concept of clave” (Gerard 1989: 60) would take a leading
position. But if the ubiquity of this structural phenomenon as one of the most enduring
Africanisms in the music of many traditions in the diaspora may be taken as an indicator,
we may be justified attributing more than superficial significance to its presence.
Keeping in mind Benjamin Boretz’ definition of musical syntax as “a model for
the determination of the interlocking structure of hierarchically connected relations”
(quoted in Swain 1995: 287), and Swain’s own characterization of “essential syntactic
elements” as being “discrete events bound by rules in hierarchical organizations (ibid.:
288), we begin to sense the potential of clave as perhaps serving as such a model. After
all, what could be better examples of an ‘interlocking structure’ of ‘discrete events’ than
our two samples above? What needs to be shown then, is how the proposed ‘deep
structure’ of a clave-syntax might manifest itself by means of such rules, what kind of
hierarchical organization said ‘events’ seem to reflect.
Several authors (Mauleon 1993, Gerard 1989, Washburne 1998) have already
shown the relationship of the underlying clave-pattern to each individual percussive
instrument in a typical Afro-Cuban ensemble, as well its significance for the generation
of idiomatically correct harmonic accompaniment patterns and improvised melodic
statements. However, the study “Kubanische Rhythmen im Jazz” by Tozzi (1999) marks
a cornerstone in the quest for a more theoretically motivated and interculturally oriented
understanding of the concept of clave.
His table of the “Basic Reference Points to Clave” (155) is a first attempt at
subjectively encoding the relative strength (“Spannungszustände,” i.e., tension-states) of
isolated, “single accents or strong phrase points” in relation to the particular metrical
framework generated by the clave (see Appendix p.146).35 By assigning those attacks a
phenomenological salience (and associated affective tension-level) referring to their
particular position within the clave-cycle,36 he establishes just the kind of hierarchy
among discrete events mandated by Swain’s definition of syntax. And by indicating an
implicit “fundamental or strongest relation to clave” (ibid.) in each case, Tozzi designates
a metrical identity, or weight, to each accent, suggesting a particular syntactic tension
with the clave for each position.37
In other words, a single note may, even in absence of any actually played clave,
nonetheless be perceived as implying not only such pattern in general, but also suggesting
a particular dynamic orientation within this framework. In this context, notes that would
coincide with any of the clave’s attacks, usually carry a high level of syntactic salience.
Locke (1992: 104) has made a related claim with respect to the 4-beat “ground,” which he
identifies as the paradigmatic metrical reference to the perception of the fundamental bell phrase in Ewe
music. This “stress pattern in 4-beat meter” is partly based on the points of convergence between the bell ,
imbuing each of the 4 beats with a particular quality. Tozzi’s notion may be seen as an extension of this
idea, in that it considers other possible time points beyond the basic metrical beat level.
His chart is limited to the binary metrical context of the son clave, however.
Tozzi arrives at the value for each particular position’s individual inferential gravity in relation
to the clave by means of an evaluation of its quantitative prominence and general salience, based on
individual instruments’ patterns as well as larger, formal-stylistic conventions. His transcriptions, as well
as the examples given by Mauleon (1993) throughout her work, are nonetheless indicative of such
prominence and provide a good basis for Tozzi’s theoretical abstraction.
More interesting however are those, that do not. It is those accents that by actually
“avoiding” an adjacent clave attack imply and mark its temporal position all the more
effectively, and contribute to the perception of syntactic flow. Thus, the establishment of
such flow does not depend on any quantitatively measurable matching or “echoing” of
the clave’s exact morphology on the musical surface, but can be also be created by
selecting attack points that carry a high level of implicative quality, or, as Tozzi calls it,
“level of strength.”
Such seemingly abstract considerations are in fact very much a part of the
musical discourse in the Cuban traditions, especially whenever this correlation with the
clave is violated by a musician during a performance, or, even worse, is written into a
tune or band arrangement. In these cases, the decision whether a particular note or short
phrase is “out-of-clave”, or cruzado (crossed), may at times hinge on the position of just
one attack in relation to the established clave direction. For instance, an accent played on
“3” of a given bar usually suggests that this bar correlates with the ‘2-side’ of a clave-
cycle. 38 However, if this accent was repeatedly (and with a particular prominence) played
on the ‘3-side’ of the cycle, it might be perceived as disrupting the clear directionality of
the clave established otherwise, because in general such directionality implies
phenomenological accentuation of the “2-and” (see Tozzi’s explication, p.155).
Certainly, musical conventions and clichés central to this genre also help establish
these expectations and often build on or reinforce general syntactic tendencies. The brief
phrase below may serve as an example, as it is often used in son or salsa music in form
of a pronounced ensemble figure at the end of a section or a tune. While the same
In figure 1, the clave is shown in what is referred to as ‘3-2’ orientation, with the
‘3-side’ spanning the first bar, and the ‘2-side’ in the second bar.
rhythmic phrase is entirely permissible to be played on both clave sides as part the typical
bass tumbao pattern in these styles, it would likely be considered incorrect if it for
instance appeared in an arrangement as an isolated horn-section figure on the ‘2-side.’
Figure 5 – ‘3-side’ indicator
Indeed, it is impossible to determine these syntactic implications beyond a basic
level of tendencies, despite the fact that in certain contexts there may be quite clear-cut
aesthetic judgments as to what constitutes a violation or a particularly elegant
incorporation of these normative rules at any moment within a piece of music.
Nevertheless, Tozzi’s relative scale of syntactic “weight” or tension, though it remains
rudimentary and somewhat abstracted, is based on and clearly reflects this syntactic
awareness and aesthetic judgment in Cuban music.39
Paradigmatic musical examples of the “concept of clave” in Cuban music
To enhance our understanding of this issue, I would like to present a few more examples
of such syntactically characteristic phrases in Cuban music, which traditionally imply a
sense of clave:
In a sense, Tozzi hence posits an equivalent for the Cuban musical context to the kind of
interpretative stance Lehrdahl and Jackendorf (1983) assume in their Generative Theory of Tonal Music,
who formulate a set of “well-formedness rules” for a Western classical tonal-durational context, upon
which listener-generated “preferences” come to bear.
This legendary Chano Pozo / Dizzy Gillespie composition is made from a
combination of interlocking melodic riffs, each of them essentially a diatonically
melodized drum or percussion pattern. ‘Horn riff 1’ is based on a campana (bell) pattern
used in Cuban son montuno or mozambique styles (figure 5), while ‘Horn riff 2’ is
reminiscent of melo-rhythmic phrases often improvised by the quinto drum in rumba.
Figure 6 – Manteca horn riff
Figure 7 – Mozambique Campana
Quinto “ride”
This pattern is one of the staple phrases of quinto improvisation in rumba
guaguancó and often employed as a sort of “holding pattern” between more active
figures. It is customarily articulated using a varying sequence of open tones and
closed slaps. 40
Ernesto Diaz, personal communication.
Also see Bauzó (1995: 18) for usage of the term “ride”
in the context of solo-improvisation by the quinto, the highest of the three congas generally played in
Figure 8 – Quinto “ride”
Track #2
“Timba Bassdrum”
In the contemporary Cuban salsa style called timba, the drumset commonly uses a
2-note bassdrum pattern (i.e., bar 2, below), or may play this complete 2-bar phrase
during the so-called despelote sections that are interjected percussion break-down’s of
heightened the rhythmic intensity. In that case, the single optional added attack on the “1-
and” of bar 1 may be taken by a listener as a cue towards establishing or reinforcing the
indicated syntactic correlation with the clave.
Figure 9 – Timba Bass Drum Pattern
“La Expresiva”
This is the opening horn line for the song “La Expresiva” by the Cuban group
N y G La Banda. While the clave is not played on the recording, its direction is
nonetheless almost immediately defined by the placement of the first few notes,
especially the “1-and” of the first bar, followed by the “1” of the second bar, which also
mark the tonal center (D major).
Figure 10 – “La Expresiva” (Introduction)
Track #3
“Gandinga, Sandunga, Mondongo”
Performed in the style of a classic Cuban descarga (jam session), this song’s
melody is simply a cyclical percussive riff utilizing a simple arpeggiated melodic motif.
Figure 11 “Gandinga, Sandunga, Mondongo”
Track #4
In addition to confirming the basic syntactic salience of individual attacks as
suggested in Tozzi’s chart, these examples also speak of a larger syntactic coherence at
the phrase level. It is important to note that even highly fragmented patterns such as
“Gandinga, . . .” or the “Quinto-Ride” reflect an underlying sense of clave and particular
alignment with it. As Tozzi (155) points out, a note on the “1” of a bar is often indicative
of a ‘2-side’ alignment, but it also would depend on the further context to determine the
“gravity” or tension-state this attack might be perceived to generate. For example, the “3-
and” in bar 1 of figure 11 above does not carry the same metrical implication, sounding
right after the ‘D’, as it does later in the fourth bar, where it is leading back (also tonally)
to the beginning of the phrase. Likewise, both the ‘D’ on “4-and” in bar three as well as
the ‘D’ on “2-and” in bar two are an equally strong indicators of the clave direction –
even though none of them coincide with the clave per se!
What matters is that these notes establish the correct syntactical relationship with
the clave, perhaps analogous to the way a leading tone in a melodic line may indicate a
tonic scale degree (and thus suggest a particular harmonic schemata), even though that
note may never sound. And only small a number of meaningfully placed notes may be
sufficient to conjure up the framework of clave as mental timing reference for a listener,
analogous to the way a sense of a particular tonality can be evoked by a melodic phrase
without ever providing all the pitch classes that may make up such tonality! Once the
temporal dynamism of clave is detected by a listener, even a single attack (see the
discussion of the “Timba bassdrum” example) may provide further cues towards the
particular alignment of the musical surface within this framework at any moment.
Therefore, syntactic perception of events on the musical surface in relation to
either the clave as a heard, or internally held mental reference represents a musical
dynamism, which requires a listener to establish a sense of some sort of structural
regularity among these two levels. At first this regularity may appear to be implicit or
hidden. Implicit – because a listener not familiar with this dynamism may not be able to
identify the “logic” with which, for example in case of figure 11, the melody notes are
placed to maintain contrapuntal tension with the clave, even if the clave is played. Hidden
– because in cases where there is no clave sounding, one of these levels is not itself
present on the musical surface and may not be suspected by such listener as being
structurally present at all. By only categorizing the audible events, the listener may thus
get no sense of any regularity, and not be able to establish any syntactic schema at all.
Yet, this process may also be influenced by additional tonal-harmonic
considerations that interact with such a primarily durational syntax. This is why, for
instance, in figures 6 and 11 the musical phrase as a whole is perceived as beginning with
the ‘2-side’ of the clave, despite the later argument for a paradigmatic flow from ‘3-side’
to ‘2-side’ within a clave or paradigmatic pattern cycle. In salsa music, where a fixed
relationship between harmonic rhythm and percussive rhythms is crucial to an
arrangement’s formal integrity, this specific alignment is commonly indicated at the top
of a chart as ‘2-3 clave’ or ‘3-2 clave.’ In above examples, the phrase periodicity
established by the cadential syntax of melodic and harmonic motion can be said to
“force” the listener towards the perceptual framework indicated, effectively overriding
the syntactic directionality of the clave. However, one unanswered question for further
cross-cultural research will remain what position these various constituent syntactic
forces may inhabit within a larger hierarchy of a unifying meta-syntax, and how fluid
such hierarchy may be internally depending on the listener.
Independent of these further issues, perception within the framework of clave
nonetheless enables a listener not only to, in terms of Swain’s general characterization of
musical syntax, “reduce . . . the randomness of elements in the incoming sound stream”
(1995: 285) by providing a cognitive association with another temporal reference
framework beyond a fundamental equidistant metrical grid, or other higher subdivisive
levels. It also creates a particular musical affect, in which a musical dynamism of
durational tension and resolution at the level of each clave cycle is the consequence of
such syntax, its “specific kinds . . . derived from cultural associations, learned by listeners
along with the syntax itself”(294).
In these pages I invoke the notion of a particular syntactic directionality that is
important to the perception of a listener steeped in African tradition. Sensitivity to such
‘forces’ within this metaphorical temporal space, in my opinion constitutes one of the
components that characterize the continuity of African musical expression across regions
within Africa, as well as in its diaspora. The significance of this syntax is part of the
(fluid) boundaries that mark these communities of listeners. A number of authors (e.g.
Brothers 1994; Wilson 1974; Cronbach 1981; Kubik 1986, 1999; Logan 1984, Roberts
1998, Monson 2000) similarly have tried to define the traits that would characterize the
“African/African American Idiom in Music” (cf. Brown 1989), especially in its process
of cultural history.41 My interpretation of ‘clave’ in a paradigmatic sense certainly is
intended as a further investigation of this grand zone of music and culture.
Finally, to conclude this chapter with a few, admittedly rather probing thoughts, I
would like to alert the reader to an interesting theoretical conception, which has been put
forward by Victor Grauer (1996) as part of his more general discussion of syntax and
semiotics in the arts. While it would be beyond the scope of this study to review the
author’s conception in all but its most basic terms, I nonetheless want to incorporate some
of his ideas into my argumentation.
Monson (2000: 17) has encapsulated this ongoing inquiry rather nicely: “A deeper
understanding of the way in which notions of cultural authenticity and legitimacy are necessarily reinvented
in each generation through process of intergenerational negotiation, contestation, and synthesis points to the
enduring importance of notions of authenticity and tradition in music of the African diaspora, without
reducing them to a racial essence, a standard of purity that is impossible to achieve, or the whipping boy of
poststructuralist cultural theories that are suspicious of any implicitly color-coded boundary.”
Clave - a “Syntactic Field”?
Central to Grauer’s theory is the notion of metaphorical “syntactic fields” that define an
equally metaphorical “space,” and signify a certain “meaning.” Specifically the author
argues that:
we cannot normally hear musical passages outside a kind of tonal “force-field” not unlike that
which produces, for language . . . ‘value.’ [. . . ] Thus, we hear music virtually, in terms of
essentially semiotic fields (gestalts produced by a systems of differences), not in terms of the
material, purely sensory experience of something we might want to call the sound themselves.
Such syntactic fields are perceived in a variety of aural and visual dimensions, and “any
object of perception can signify (take on meaning) only in relation to a controlling
syntactic field” (1.3.2). Since they constitute “a kind of invisible controlling spacelike
region or extent within which certain types of activities have the potential to take space
. . . we can consider a syntactic field as a controlling, determining fundamental entity or
function by means of which syntax and sign can be understood to operate” (1.3.4).
Within such gestaltist “perceptual ‘force fields’” that could also describe “the
‘space’ of musical time (e.g., a ‘metric space’),” events are “uniquely ‘colored’ by [their]
orientation,” which is expressed using the idea of “ ‘vector fields’ of physics, in which
the direction and strength of ‘forces’ at any given point can be symbolized by an arrow
and a number (2.4).” While in tonal space orientation is determined in relation to pitch
class, in temporal space “any such sound will take on a certain metric ‘coloration’
depending on its time-point class, which determines its place in the metric field (similarly
conceivable as a vector space) (2.6).” (Incidentally, this is what I have in mind
throughout this study when referring to a rhythmic pattern’s particular directionality).
But while Grauer may have primarily been thinking in terms of a Western concept
of meter as felt groupings within the flow isochronic pulses, such field characteristics
clearly can also be translated to the cyclic periodicity of the clave, whereby another,
albeit non-isochronous, reference grid is established in form of the syntactic arc proposed
at the outset of Chapter 3. Events are thus experienced in at least a twofold manner:
locally in relation to the nearest tactus beat, but also against the distinct circularity
ingrained within the paradigmatic quality of the timelines and the syntactically related
foreground surface.
Just as an analogue “spatial syntactic field is produced from a few carefully
placed lines,” as in, for example, “perspective space, where invisible ‘lines of force’
control the syntactic field of an entire picture (1.3.5),” the temporal organization within a
proposed syntactic field of ‘clave’ may be coordinated similarly. In trying to detail this
interplay between a basic paradigmatic pattern and other foreground patterns, my effort
has been to point to exactly these virtual lines of force within the context of such
rhythmic organization, and mark the perception of an ensuing ‘temporal resonance’ as the
result of such dynamism. The reader should indeed get a good sense of this affinity with
Grauer’s conceptions during the descriptions of the interplay between foreground and
background patterns, encountered in Chapter 3 under the heading ‘Examples of
Paradigmatic Pattern Integration in Cuban Music.” Employing Grauer’s terminology, the
basic paradigmatic pattern introduced further on in this thesis could hence be seen as
creating such a “metric field,” within which “a note occurring on the first beat of a
measure will sound different and carry a different musical meaning from the same note
occurring on the second beat, even if both are played in exactly the same way (1.3.7).”
Also associated with such notion of force-fields, whether tonal or temporal, is the
dialectic of “attraction” and “repulsion (2.15)” among events with respect to each other.
This corresponds to the kind of characterizations of syntactic pervasiveness I have
suggested in a number of the analyses below, e.g. the cáscara, comparsa or arará bell
patterns. Particularly, by describing a particular contrapuntal balance, interlocking flow,
and mutual implication of two patterns, and by pointing out the either implicitly
perceived or externalized paradigm. The syntactic meaning of such patterns is thus
always couched in terms of such dualism, and seems to echo what Grauer calls his second
semio-aesthetic principle: “before any perceptible can function as a sign, it must be
apprehended as a gestalt, i.e., a form or figure perceived against a ground (1.3.8).”
Important for our discussion will also be the idea that a particular musical syntax
is congruent with a particular community of listeners, constituting a shared music
contract as it has been called above. The perception of a ‘clave’ as a paradigmatic
syntactical feature of much African music and in the diaspora is thus like “every syntactic
field . . . a construct with an ideological determined basis . . . there is no such thing as a
passive or even neutral ground. The fields associated with all signifying processes are the
products of culture and reflect ideologically determined value systems enforced by
explicit or implicit rules (1.3.12-13).” Or, as Robert Walker (1990: 56) puts it, “ [a]ny
particular culture, as represented in both speech and music will supply the categories
within which to place auditory stimuli. [. . . ] It would seem that clear links are likely to
exist between a culture, its developed belief systems, our innate perceptual proclivities,
and our development of these proclivities.”
But before we move on to discuss in detail the ways in which a musical surface
may be indicative of such an expanded, paradigmatic notion of ‘clave’ or temporal
“hidden regularity”42, I would like to spend a moment analyzing the morphological
features and inherent structural dynamisms that mark an intrinsically related pattern, the
type of ternary timeline often referred to in academic parlance as “West African bell
pattern” or “standard pattern” (cf. Jones 1959).
The West African Bell Pattern
Figure 12 - West African bell pattern
The “West African bell pattern” is one of the most studied musical phenomena
among non-Western traditions in general, spawning numerous attempts by almost every
author in the field of African music to characterize its enigmatic structural properties and
functional multivalence. While not at all limited exclusively to being played on a bell,
this pattern tends to be perceived as a timeline pattern (also called phrasing referent, cf.
Nzewi 1997: 35) in a multitude of cultures of the sub-Saharan region, in which case it
usually sounds on a penetrating, high-pitched metal or wooden idiophone. In this
function, it seems to represent an – if not outright universal – yet nonetheless
quintessentially African approach to music-making for many scholars and listeners alike
We will see in Chapter 3 that, in their search for “a latent regularity that ‘translates the
theoretical rules into rules governing performance practice (33),” Cohen’s and Katz’ (1997) discussion of
the concept of “intonation type” or “intonation skeleton” within the maqām system of Arabic vocal music
provides an interesting and perhaps fitting analogy from a tonal perspective, as the authors suggest
extending their notion of “type” to other parameters (e.g. ratios of durations) as well.
It has also been the locus of much research tracing the historical and
morphological linkages (e.g. Kubik 1986: 124ff., Olivera Pinto 1991: 190, Tozzi 1999:
130) that are said to exist between African musical traditions on both sides of the
Atlantic,43 and especially its continuing utilization in Cuban music has been well
documented (Mauleon 1993: 49). In particular, the essential congruence of the Cuban
clave with its African parent has been established for both morphological and functional
dimensions (cf. Tozzi 1996: 39, Mauleon 1993: 50) and it is on this premise, on which I
would like to build the following analysis/argument.44
By and large, a lot of research on African timelines has focused on their linear,
horizontal structure, their circularity, and function as “phrasing referent” (Nzewi 1997)
for the timing of individual voices within an ensemble. Their temporal organization has
been captured quantitatively in arithmetic relationships (e.g. Pressing 1983a, Rahn 1987,
Kubik 1999: 54, Anku 1995, 2000), processes of linguistic representation ( Kubik
1986,1989,1999), and isomorphic associations with Western tonal-scalar systems (again
Pressing 1983a and Rahn 1987). Their qualitative and cognitive-affective dimensions
however, both horizontally and vertically, have received less attention (for an exception
see Magill and Pressing, 1997, and Pressing 1983a). It is these that I would like to
emphasize, since my own performance experience has much to do with sensing (and
Kubik (1994) particularly highlights the timelines’ enduring significance in this process: “Some
traits in a culture may also be immune to factors, external or internal that would , under different
circumstances, unleash culture change. For example, the so-called asymmetric time-line patterns (see
below) so prominent in the music of the West African coast and in West-Central Africa, as well as in
extensions of African cultures in Brazil and the Caribbean, are virtually immune to forces of change. They
may change their performance speed, their accentuation, the instruments or tools on which they are
performed (for example from iron bell to glass bottle), and they may gradually change their distribution
areas in Africa, thus disappearing from one tradition and reappearing in another, but they cannot change
their numerical structures. There is no force in the world which can change the internal mathematics of the
asymmetric time-line patterns, one of the most ingenious African inventions of the remote past (34).”
If, as we will discuss later, a certain syntactical coherence is asserted for a particular Cuban
rhythmic context that adheres to clave, we should be surprised to find the seeds to such organization in the
African timelines’ archetypal, paradigmatic morphology.
creatively manipulating) exactly these dimensions and their effects on the musical
Previous cognitively oriented research
David Locke (1992), for example, has given many comprehensive accounts of the
“rhythmic multideterminancy” inherent in the bell pattern shown in figure 12 for the
context of Ewe music, comparing it to “a work of visual or plastic art – a sculpture, a
mobile, or a gemstone” (103). But while he alerts us to the bell phrase’s maximal
perceptual plasticity in its alignment with the surrounding voices,45 he makes it clear that
it is nonetheless fundamentally rooted and primarily conceptualized against an
isochronous 4-beat “ground”(104). Thus, each of the bell’s attacks establishes a particular
type of tension with such matrix, the effect being a continuous in- and out-of-phase
movement between “figure” and “ground.” As this process, which will be of great interest
to us in this discussion, is confined to the periodicity of one bell cycle, we might say that
a dynamic “breathing” (or motion from tension to release ) is taking place within each
such period.
And Nissio Fiagbedzi (1977) observes that the fundamental sequence of steps
RLLR in Ewe dance follows the periodicity of a bell cycle as well: the second of the
consecutive right foot steps marks the beginning of each cycle, and “one may therefore
Locke (1982: 224ff.) explains that such “pattern may be heard differently depending on the
metric vantage point of the listener,” and, “moreover, listeners and dancers are free to change their metric
stance as influenced by acoustic accents, or as their aesthetic taste dictates and listening skills permit. But
no matter how prominently a countermeter is mentally or acoustically accented . . ., the primary
duple/quadruple stream of beats is never negated or replaced.” These potential metric multiplicities have
been one source of the common Western “mystification” of certain West African music as being “highly
complex” and metrically ambiguous. See also Agawu’s (1995) critical discussion of this issue.
deduce that the 12-unit patterns are each perceived as a complete cycle, the cycle
beginning on the first pulse and ending on the first pulse of the subsequent cycle (332).”
In an interesting study on motor coordination, Jonathan Magill and Jeff Pressing
(1997) have shown the cognitive models present in the performance of an asymmetric
bell pattern46 (in this case the Kete-variant related to the pattern shown in figure 12) by
one hand, while maintaining an equidistant underlying pulse structure in the other, to be
of an integrative and hierarchical nature. The authors suggest that superior timing does
not just depend on linear repetition of a timeline’s serial durational structure, but on
integration with a “central cognitive clock process that directly triggers the motor delay
processes of . . . one hand,” whereas the “motor delays in the other rhythmic stream . . .
played by the other hand, are cued from the immediately preceding ground-stream clock
element [i.e., a tactus beat] via an intervening ‘subpulse’ process (194).” Also, minor
fluctuations of such manually integrated timing process were detected in the consistency
of the combined pattern’s performance “depending on the figure/ground allocation”
between tactus and the bell pattern, 47 confirming the existence and physical effects of
“distinct cognitive models for the same polyrhythm” (ibid.) in musical performance.
While I cannot attempt here to discuss their specific findings in greater detail, I
would nonetheless like offer them as experimental confirmation of the assertion that,
whenever we are discussing the timelines’ musical qualities, we need to do so primarily
from the performer’s experience of a perceptually interdependent and cognitively
integrated organization in relation to a metrical “ground,” in particular the 4-beat time-
This study used a single participant, Ghanaian master drummer Noah K. Owusu from the Asante
ethnic group.
I.e., mean deviations from metronomic exactness of within-hand target intervals of a timeline
were different for each performance condition. (Magill and Pressing 1997: 205)
feel mentioned above.48 With this in mind, let us now proceed further by looking at the
syntactical and kinesthetic properties of some variations of the fundamental West
African bell pattern.
Correspondences in morphology and syntax among West African timeline patterns
First of all I would like to assert a syntactical correspondence between the, at least
three, versions of the West African bell pattern, also called “standard bell pattern” (cf.
Locke 1982: 224), that are shown in figure 13 below and which are often cited in the
literature as paradigmatic for the temporal structure of ternary timelines in general. While
one has been observed among the Ewe49 (e.g. Jones 1959, Pantaleoni 1972, Locke 1982),
the other is attributed to the Yoruba ( e.g. King 1961, Koetting 1970). Furthermore, it has
been noted that all are permutations or “modes” (Locke 1992: 101) of the same cyclic set
structure (Rahn 1987, 1996, Pressing 1983a), which can also generate a number of other
transpositions, including a third variant reported as being typical especially east of the
Niger river.50 That does not mean, however, that we should assume all such possible
transpositions to carry the same, if any, syntactic potential.
Which was also confirmed by Mr. Owusu as his “preferred performance orientation” in the
above study, and “yielded particularly metronomic results,” i.e., metric stability (ibid.: 206).
The name of an ethnic group used here to identify a particular pattern is of course not meant to
imply an exclusive practice by this group or a limited regional provenance, in fact each of these patterns
can be found in a wide variety of traditions in- and outside of sub-Saharan Africa.
Cf. Kubik (1986: 126). See also Monson (2000: 340), who refers to this manifestation as the
“‘long bell’ (or Congolese) variation of the classic 12/8 gankogui pattern.” Aditionally, Locke reports that
this morphology is also known in Ghana as being predominatly used be the Ga people (pers.
Figure 13 – West African bell pattern variations
As Kubik has remarked,51 regardless of the diverse theoretical abstractions by which
scholars have described the peculiar asymmetrical organization of such patterns, “the
mathematical content of the African time-line patterns exists independently of their
cognitive dimension.” In other words, such quantitative approaches will tell us relatively
little about the syntactical implications of these timelines since they do not note the
cognitive-affective component generated by particular constellations.
Drum language and the question of centrality
One common facet of West African music is the manner in which much of its
instrumental music, including percussive music, is based on speech patterns. Drum
patterns and phrases often closely model, approximate, or abstract speech rhythm, pitch
contour, and articulation of a culture’s language, using the particular drum’s timbral
potential. For example in the Ewe musical genre called agbekor, all supporting drum
rhythms have a corresponding lexical meaning (cf. Locke 1996: 95), and the whole
polyphonic ensemble texture is in fact an instrumental “enunciation” of several brief
Cf. Kubik (1999: 56). Also see Koetting (1986:61) . Temperley (2000) notes that “it is one thing
to show that a pattern has unusual properties, and another thing to show how these [abstract] properties
might explain the pattern’s success; so far, the set-theoretical approach has had little to offer towards the
latter question (82).”
verbal phrases. Aside from this level of direct semantic and syntactic correspondence
with language, such patterns are also often conceptualized as mnemonic phrases. These
mimic the instrument’s timbral characteristics in form of spoken non-sense syllables, and
efficiently convey a pattern’s melo-rhythmic and timbral structure, aiding in its
memorization and oral recall. Different traditions also seem to utilize varying mnemonic
or lexical phrases to represent the structure of the phrasing referents themselves, both in
isolation and in conjunction with the 4-beat tactus.
Locke (1996: 90) for instance reports the phrase “Matikpo matikpo kple ku dza
(“I will jump, I will jump on ‘ku dza’”) being used to teach the polyrhythmic integration
of the “standard pattern” bell phrase with the tactus as part of an Ewe children’s game,
and William Anku (1997: 218) points to a lexically meaningful resultant of two bell
rhythms plus an additional drum part for the Akan style adowa. However in both of these
cases, the beginning of the lexical phrase and the musical phrase are not necessarily
identical. Regarding the children’s game chant Locke shows that the first syllable “Ma”
of “Ma-ti-kpo” does not mark a coincidence of gankogui phrase and tactus, but outlines
the bell’s second attack point. On the contrary, the final “dza” coincides with beat “1” of
the underling 4-feel and is additionally marked by a handclap.
And according to Gerhard Kubik (1999: 55), the “Yoruba version” shown in
figure 11 is conceptualized by the Yoruba using the mnemonics
kŋ kŋ k l kŋ k
but the author also points to at least two alternative regional practices (1986: 125,
1999: 56). Both of these integrate the above mnemonic phrase with the 4-beat tactus
differently, placing one of the unaccented or softer (“‘weakest’ timbre-value”) syllables
on the “starting point of the underlying cycle and metrical scheme.”
In his discussion of Ewe agbadza, Robert Kwami (unpublished) presents an
mnemonic variant of the same pattern as ka ka-gan ka ka ka-gan (i.e.,
q eq q q eq )
that in
performance also aligns itself with the first beat of the underlying 4-beat on the final
muted syllable gan. Locke (1982: 225) also has noted that this way the pattern is taught to
students, and how the bell is actually started at the beginning of a piece.
Likewise, James Koetting (1986: 61) reports sang si sang sang si sang si
q eq q eq q
; si is a press-stroke on the bell) as used by the Ashanti in Kete music.
Here, the final syllable sang coincides with the first of the four tactus beats.
At this point we notice that the initial syllable in all these timeline’s verbal
representations does not coincide with the musically perceived “starting point” of these
patterns’ cycle, in fact many of them tend to place the final syllable in such place!
Consequently, insofar as there appears to be no categorical alignment between the
beginning of a timeline’s mnemonic or lexical phrase representation with a tactus beat, it
is unlikely that we have much reason to assign to it a particular value in terms of forming
a point of musical centrality, and I argue that in analysis it thus should not serve as the
assumed cognitive boundary of each cyclic revolution in the sense that I will try to
explain below. If anything, the recurring marking of the first bell attack and the first
tactus beat by these last syllables, points of rhythmic release of the utterances or semantic
unit, should be noticed and taken as an indicator for this position’s cognitive significance.
For the moment, then, I would like to assert that there may exist a difference
between a pattern’s “native” mnemonic conceptualization and its phenomenologically
relevant syntactic quality, a quality that is a result of the perceptual integration with the
underlying isochronic tactus framework. Thus, regardless of how mnemonic or lexically
based verbalizations align themselves with the structure of a pattern, they should not be
mistaken as implying a particular metrical centrality of such pattern in the music as well!
While they surely represent a cognitive dimension for a performer, and hence may posses
a subjective reality, such dimension is as such not perceivable on the musical surface and
exists outside the kind of syntactic organization at the heart of this discussion.
It therefore can be problematic to base a structural analysis of an entire
polyphonic texture on a visual representation resulting from this lexical aspect of pattern
morphology, as for instance Anku (1997) as proposed. His analysis concedes a particular
point of structural centrality (labeled regulative time point [1995: 177]) as bearing on the
music’s perception, yet refuses to assign it also cognitive centrality in a metrical sense. In
the resulting notation all individual drum patterns are thus shown in correlation to a
timeline that is essentially the same as those in figure 11, but its suggested point of
centrality does not coincide at all with the indicated centrality of the ensemble rhythm at
large (1997: 220ff.). As a consequence the potential for a vertically integrated, metrically
oriented appreciation of the music as heard is greatly diminished, as I will point out later
in a later chapter (see pp .124ff.).
But the question arises, if neither the beginning of a lexical, mnemonic phrase nor
any other mathematically reasonable properties of a timeline indicate a central point of its
periodic cycle52, what does? Indeed, why look for a such point in the first place since
there seems to be agreement among scholars that the question of where is “one” in
Note that, for example, Anku’s (2000) “arbitrary standard” of a ‘prime form’ as a “set rotation
[of a specific set rhythm, i.e., timeline] with the least time interval arrangement at the beginning in
‘unordered’ form” (in this case a 1221222 sequence), in order to asses a pattern’s “embedded rotational
possibility,” can make no claims in this direction.
general does not really seem to present itself from a African point of view (cf. Charry
2000: xxviii), an impression that seems to be corroborated especially well by the unique
modal qualities of asymmetrical phasing referents, as timelines have been called, which
appear to defy the implication of a primary, or most salient, point of centrality.
Victor Zuckerkandl (1956: 167) has made the crucial point that in the perception
of music, “directly or indirectly . . . tones always communicate to the listener the basic
beat that regulates their motion.” Such beat sets up a “sympathetic oscillation” within the
listener that is “never . . . a mere dividing of the time flux into equal parts; inevitably,
without any difference in accentuation, the parts will join into little groups and form
measures.” This measured perception of time quality, not just of its quantitative fractions,
implies a motion where “further . . . forward in time” always also means “back [towards
a] starting point,”(168), as the author puts it. Based on this notion, I suggest that the
perception of circularity as a fundamental characteristic of African rhythmic structure
does thus not automatically preclude, nor undermine, the notion of centrality, because in
the perception of circularity we nonetheless experience a process of “‘away from-back
to’,” in which “the [count of] ‘one’ that closes one cycle simultaneously begins
And while Zuckerkandl’s conception emphasizes the point that the semantic
distinction of this position as marking either the “end” or “start” of a cycle is probably a
moot one, is does clearly assert its cognitive function. Indeed, skepticism towards any
potential, “silently” implicated, assumptions regarding the cognition of such boundaries,
The notion of a point of cognitive centrality within a circular structure such as the diatonic set
(whether it implies a scale of durations or frequencies) is of course nothing new, as according to the
“cognitive-structuralist” theory discussed for instance by Cross (1997: 366), in the tonal dimension it forms
the basis for the perception of a distinct tonality emerging from distinct pitch classes.
which can result from particular modes of visual representation (i.e., notating the
progression of events within a segment of time in form of a “measure” read from left to
right) does need to be taken seriously. But such undisputed pitfalls in the attempt to
demonstrate a cognitive phenomenon should not be taken as calling the phenomena itself
into question.
Furthermore, as Zuckerkandl explains, the very idea of “circularity” in music
itself does not quite match a listener’s dynamic experience of temporality or his sense of
time passing, “since in time there can be no real going back, and hence, strictly speaking,
no real cyclical motion either.” He instead proposes the notion of a “wave” as better
metaphor for this process, which I would liken to my personal concept of the “syntactic
arc” that will be introduced later.
Returning to the issues of our particular topic, I believe that the question of
centrality, or what Charry has called “rhythmic frames of reference”(2000: xxviii), is
obviously of theoretical significance and analytical concern. The fact that African
informants seem not to refer to it explicitly nor tend to engage terms like “one,”
“downbeat,” or “beginning of a phrase,” does not automatically invalidate its phenomenal
reality, neither for Africans nor for non-Africans, as Charry would like to deduct from the
“apparent lack of a definitive starting point in a cycle”(ibid.). Even though we certainly
can say that the perceived phrase length of an individually observed rhythmic pattern
played by, let’s say, a supporting drum, may, in the context of the actual or implicit
metrical subdivision imparted by the tactus, appear to be somewhat “suspended” and out-
of-phase with other such phrases, it is a step further to deny a unified point centrality to
the musical texture as a whole. I would argue further that insofar as we want to speak of
any musical texture as having a salient gestalt, and in translating its spatial parameters
into the perception of “boundaries” within the unilineal temporal domain, such centrality
is in fact indispensable!
However in analytical practice authors tend to acknowledge, at least in their
transcriptions, but most likely already in the prior subjective cognitive experience leading
up to such representation, one perspective that, in the words of Charry (ibid), “is the
simplest . . . to feel,” adding that “[t]his choice also should not be confused with how the
African performer may feel it, although our perceptions here too probably converge most
of the time”(italics mine).
For the specific context of the West African timeline pattern under discussion,
again, it seems that one particular metrical perspective appears to be primary. And
regardless of the surrounding polyrhythmic texture or specific countermetric challenges,
as far as the syntactically defining bell cycle is concerned, “[t]he moment of articulated
and felt musical resolution [or perhaps cyclical revolution] is beat one, stroke 1” of the
gakokoe (Locke 1992: 104).
Hence with respect to the examples in figure 13 above, I suggest that even though
each may receive a different lexical or mnemonic representation, these three timeline
variants not only just share a distinct diatonic scalar design, i.e., represent three
morphologies in a collection of twelve possible modes of the same circular structure of
durational intervals (cf. Rahn 1996: 76ff.), but are also syntactically equivalent to each
other on a higher level – they are in vertical alignment54 with a yet to be defined
Tozzi (1999: 156) similarly notes the particularly “non-linear, vertical conception” of Cuban
music versus for instance jazz practice, which “additionally demands a particular, complementing,
dovetailing behavior of each instrumental line,” and “each line establishes the syntactically correct tension
with the clave.”
paradigmatic “background”, which implies a particular centricity that in turn suggests a
preferred overall metrical orientation. Note that “alignment” in this case does not
exclusively mean literal congruence or synchronicity among attacks, but rather sharing
the same essential dynamic syntactic directionality.
On the way towards the formulation of a more detailed description of the
syntactical qualities inherent in timelines of West Africa and the diaspora, I now want to
turn in my analysis to a larger body of related examples and discuss their structural
Chapter 3
A Comparative Overlay of Selected Ternary Patterns and the Notion of the
Syntactic Paradigm
The following ‘comparative overlay’ (figure 14) puts the patterns discussed so far in
context of some more examples of phrasing referents collected by the author (Lehmann
2000) as part of an informal survey of about 100 timelines compiled from audio
recordings and other sources. The data is limited to ternary morphology, but could also be
presented from a binary viewpoint.55 This visually aligned superposition of a
representative number of rhythms should facilitate visual verification of their structural
affinities. Despite the fact that these examples are not necessarily directly socio-culturally
related, I propose that their fundamental correspondence is due to a unifying syntactical
organization operating in the background. In particular, I suggest this syntactical
dimension to be defined by the structure shown as pattern 1.1. It consists of five attack
points (AP) per cycle and functions as the paradigmatic “blueprint” for all other
Laz Ekwueme (1975: 31-32) has presented a similar overlay, but bases his comparison only on
the linear, additive coherence of the patterns. Nonetheless, he discerns two “standard rhythm patterns in
West African music” that are defined by a common denominator of binary and ternary subunits (3+3+2 and
2+2+3+2+3) from which a number of “relatively non-essential” variations can be generated “by breaking
up the dotted rhythms [ i.e., dotted quarter notes],” a process Pressing (1983a) has further labeled “element
fission.” In a sense, then, Ekwueme, too, acknowledges the existence of a deep-seated paradigmatic pattern
structure, and even notes a “binary balance, whether or not the initial points of both halves are acoustically
stressed,” as arising from an always cognitively present “duple pulse dividing the whole pattern” in its
middle. In his opinion the faster four-beat tactus hence indicated below pattern 1.14 constitutes “a sort of
temporal prolongation” (32) of such background binary subdivision.
Figure 14 - Comparative Overlay
First we notice the general alignment among most of the patterns’ attack points, and
specifically the explicit congruencies with the syntactic paradigm in pattern 1.1.
Furthermore, I would call attention to a number of features that appear to justify
concluding such a shared morphology:
Patterns feature a bisection of their periodicity. The first half often takes the form
of an explicit or implied 3:2 ratio of the timeline to the tactus beats, versus the
second half where we find a 1:1 ratio (i.e., unison) concluding the phrase. I
suggest a tendency towards a shift from contrametric to commetric sections.56
Almost all the patterns start the first half “with a sound” (beat one of the
underlying 4-beat tactus), while avoiding to mark the beginning of the second
bisection (note the eighth note rest right before paradigmatic attack point 4 in
many places). Even when there seems to be a sound occurring on that beat, the
dynamic accentuation with which these patterns are often performed may
highlight the implicit structure of pattern 1.1 and treat such event as a “ghosted” 57
pickup into attack point 4. For instance, in pattern 1.3 the quarter notes
immediately following each of the two eighth notes will thus sound slightly
“Contrametric” and “commetric” are terms first introduced by Kolinski (1973: 497), and also
entertained by Arom (1991: 208). However, “metric” in this sense only refers to an underlying tactus,
which strictly speaking is not really a meter, but simply an isochronic pulse.
“Ghosted” in this context means that the attack is part of the kinesthetic flow of the pattern, but
may not receive equal dynamic stress as the surrounding strokes, and be subject to a slight timbral
Patterns that depart from the sequential presentation of attack points 1, 2 and 3 (an
implicit 3:2 ratio) by incorporating beat 2 of the underlying pulse (see 1.7-1.10)
still keep a clear reference to attack points 4 and 5. Perhaps initially less
syntactically indicative as 1.1-1.5, their bisectional quality is nonetheless retained.
For example, 1.8 may be seen as a resultant, incorporating 1.2 with all the attacks
of the 4-beat tactus, and in fact often is played in conjunction with it (for example
the guagua pattern in rumba columbia).
Variations of the fundamental attack point scheme tend to be limited to a few
notes per example, the sense of bisection is never quite compromised, and most
changes seem to occur around attack points 3 and 5. Elaboration of the scheme in
one part of the bisection is stabilized by relative simplicity in the other (e.g. 1.6 or
Attack point 3 seems to enjoy greater flexibility in its positioning. Three scenarios
appear common: a single attack farther away (1.1, 1.3 etc.) or closer (1.2, 1.7) to
the midpoint of the bisection, or a double attack including both positions (1.4, 1.5.
etc.) which may or may not be shaped by dynamic accent. This may be taken as
an indication for functional equivalence of all these positions.
Attack point 4 almost never is embellished or omitted, leaving the silent interval
between it and attack point 5 intact.
I would like to stress that in this interpretation the focus is neither on a pattern’s
internal arithmetical categorization (whether it is made up from units of 2’s and 3’s, 7’s
and 5’s), nor its particular additive sequencing (is it 7 + 5, 5 + 7, 2 + 2 + 3 + 2, etc.?), but
on its relationship to a paradigmatic background skeleton (i.e., 1.1), whose metric
salience is itself partly founded in a process of cognitive integration with the fundamental
tactus. Each of the paradigm’s five attack points are assumed to be experienced as
occupying a unique metrical position, with their individual “value” valence based on their
position in relation to isochronic regularity of the underlying 4-beat tactus. As such, they
have in common a distinct cadential flow from tension to release that spans the entire
phrase, from its first attack to the stroke that coincides with the fourth beat of the
isorhythmic tactus, and peaks at around halfway through the cycle.58
“Metrical,” the alert reader may ask – didn’t the discussion so far simply seem to
focus on rhythmic patterns and their structural appearances? Indeed so, yet, we may be
about to realize that the common (Western) conception of meter as a strictly periodic
grouping of “events into equivalence classes” and as “a regular alternation of strong and
weak beats” (Parncutt 1994: 412) may be cutting short its interpretative potential for this
In fact, Robert Kauffman (1980: 411ff.) has suggested that timelines in West
African music are “rhythmically regulative” in such metrical sense, essentially bestowing
a syncopated function on whatever isochronic pulse may be present. And while the
“polychronic aspect of African music has led to much ambiguity in terminology, . . . it
In Ewe music, for example, Locke (1996: 93) finds this fundamental dynamism to be borne out
by the characteristic interplay of the parts of gankogui, axatse, and kaganu that “create a distinctive quality
of musical temporal experience” in which “the moments when bell and beat fall together – beats 4 and 1,
bell tones 6 and 1 – are specially marked in musical awareness.”
Cohen and Katz (1997: 41) elaborate on this issue: “The simple musical metres in the West,
which indicate a clear separation between rhythm and metre, serve as background for hierarchical
organization on high levels. By contrast, the prevalent predetermined non-Western rhythmic patterns do not
differentiate between rhythm and metre and do not provide a simple background for further elaboration;
instead, to various degrees, they form part of the foreground, which as such partakes in improvisation.”
would seem more appropriate to consider it monometric [as] . . . the other lines are
[perceived] in polyrhythmic relationship” to the timeline. “Perhaps rhythm and meter are
used too interchangeably in Western parlance to allow us to distinguish precisely between
polymeter and polyrhythm in African music.” Well put!
This notion is also reinforced by Locke (1992: 104), who has pointed specifically
to one metrically implicating stress pattern, or time-feel, which the gakokoe phrase
imparts to the tactus, saying that “[s]ome readers may react skeptically . . . because it
seems surprisingly similar to a Western 12/8 meter.”
Bisectional periodicity and the “Syntactic Arc”
The bi-sectionality of timelines has been noted by many scholars (e.g. Nketia
1974, Ortiz 1950: 276) and also forms part of the theories of Rahn (1987), Arom (1991),
and Pressing (1983). It depends on the integrated perception of the timeline and the 4-
beat tactus, creating two dynamically interacting subunits of diverging and converging
metrical charge. The first subunit pursues an out-of-phase trajectory or 3:2 ratio with the
tactus by means of a paradigmatic 3-note cell q q q .60 The third of these attacks
establishes the highest tension with the underlying pulse by virtue of not being
Pressing (1983), in reference to studies by Wendell Garner, has pointed out that “pattern
descriptions tend . . . to avoid breaking up runs of . . . perceived aggregate subunits”(50) and “possible
perceived starting points could be accurately predicted by the interaction of two principles:
1. Pattern descriptions tend to end with the longest gap between elements (gap principle);
2. Pattern descriptions tend to begin with the longest run of elements (run principle)”(49).
Note that this supports my suggestion of 12 q q q Ò q q ' as a perceptually preferred orientation and
suggests it may be deep-structurally and syntactically paradigmatic for all three above patterns, their
phenomenological and metrical salience essential being equivalent.
immediately resolved,61 and is also phenomenologically accented because of its increased
duration (3 pulse subdivisions vs. 2 in case of the first two attacks), as well as non-
continuation of the q run. The second subunit, in its most fundamental form q q ' ,
consists of a delayed continuation of the kinetic energy built up by the first three attacks,
and a resolution of the tension with the underlying tactus that has developed by mid-point
of the pattern. This dynamic periodicity, regardless of the prominent metrical implication
of a timeline (i.e., ternary or binary structure, or a mix of these) or the number of its
constituent events (e.g. 5 or 7 attacks in a cycle of 12 elementary pulses), is at the heart of
the syntactic quality of such patterns. If the directional energy of the initial half of a
timeline is compromised, this arc of motional energy will not be allowed to develop, and
the music loses its defining syntactic quality.
Regarding this bi-sectional quality, Kobla Ladzekpo and Hewitt Pantaloni (1970:
11) may have been the first scholars to not only realize “that the intricate polyrhythms of
Anlo ensemble drumming are the result of varied relationships between each drum and
the rhythmic cantus firmus of gankogui,” but also to suggest “the presence of an
ensemble fluctuation in two different qualities of the total motion.” Specifically, they
were referring to the “rhythmic duets” of bell and rattle, and noted “an alteration between
rhythmic unity and disparity, and . . . an effect of rushing together and pulling apart.”
While the authors did not employ the notion of a “syntax” per se, their usage of such
metaphorically descriptive terms like “thick” and “sparse” points towards an intrinsic
tensional dynamism, whose “rhythmic effects are probably fundamental to the
organization of the music.”
Perhaps analogous to the notion of a melodic suspension in Western tonal music.
Depending on the particular tradition, this mid-point may be rhythmically framed
in a couple of closely related “intensities,” the third attack of the first subunit positioned
closer or further away from the third tactus pulse (
q q q
Ò q q ' ), thereby creating
a phenomenologically “harder” ( i.e.,
q q ee Ò q q e ;
q q e Ò q q ' ;
q ' e Œ e Ò Œ
q q Π) or softer suspension (
q q q Ò q q ' ;
q ' e Œ q Ò Œ q q Œ ).
Flexibility in temporal categorization
We have seen that from the perspective of notational practice timelines seem to – more or
less – fit a familiar ternary or binary metrical grid (for example 12 or 16 units long), of
which, according to Richard Parncutt (1994: 425), each “basic time unit is assumed to be
perceived as a temporal category (C)” within the pulsation of such units. But it is clear
that in reality performers and listeners perceive timelines as “highly integrated wholes in
their own right” (Rahn 1996: 71), e.g., holistically perceived musical gestalten, which are
cognitively experienced as invariant schemes involving a distinct vertical (i.e., metrical)
integration, rather than as a string of discrete serial events.62 Thus, analytical
methodologies departing from the assumption that each such temporal category does in
fact correspond to a cognitive unit, may, in a certain way, miss the proverbial “forest for
the trees.” However, judging by recent thinking on certain aspects of tonal
conceptualization in African music, a new interpretative approach could be emerging:
Dowling and Harewood (1986: 187) confirm that the “dual structure of underlying beat and
superimposed rhythm is fundamental to the cognitive organization of music in general from very early
what on the musical surface might appear to measure perfectly in discrete cognitive and
analytical units, nonetheless may have little or nothing to do with such categorization.
Such is, for instance, the case of the ‘blue notes’, suggested by Kubik (1999: 123-
25) to be examples of African-derived practices of “microtonal pitch modifications”
reflecting “flexible pitch areas” rather than the “separate conceptual pitch units” derived
from comparison with an “extrasystemic parameter: the European diatonic scale,” that
Western analyses make them out to be. Furthermore, Kubik concludes that “[t]he whole
concept of blue notes therefore introduces an artificial split in the blues singers’ pitch
resources.” I am tempted to extend part of this rationale to the temporal domain; can
perhaps attack point 3 of the paradigmatic pattern be understood as a conceptually unified
“flexible time area”?
The fact that timelines usually present either a “soft” or “hard” a version of this
transitional point, points to its implicit syntactical significance and indicates a qualitative
difference among both versions.63 However, despite such division into two categorically
different types of the paradigm, I would argue that functionally they are equivalent. This
means that even though in conventional musicological terms they may be thought of as
representing two discrete temporal categories (C) out of twelve possible units (i.e.,
pulses 5 and 6), these two categories are actually cognitively experienced as one syntactic
Depending on the context, the placement of attack point (AP) 3 is thus subject to a
larger temporal scatter, while others remain more stable: AP’s 2 and 4 may shift as well,
We can experience this qualitative difference very clearly in the case of the Cuban son clave and
the rumba clave (see figure 1), whereby the latter would represent the “hard” version and the former the
“soft.” Each imparts a distinct feel on each respective style. Within a song, these two types of phrasing
referents are traditionally never alternated, as this would alter the unique syntactic tension of either clave
with the particular ETC.
but AP’s 1 and 5 never move in relation to the tactus – they always coincide with it (see
also cf. Locke 1992: 104)! Because the paradigmatic pattern allows for a dynamic
hierarchy of temporally flexible “intonations”64 of some of its internal attack points in
relation to the fundamental tactus, as mentioned earlier, it can mold itself onto multiple
metrical “states” without compromising its syntactic coherence. In particular, AP’s 1 and
5 define and stabilize the boundaries of this continuum by providing important anchor
points through their static relationship with the tactus. However AP’s 2, 3 and 4, which
never coincide with a tactus beat, represent more “elastic” areas of timing. Their relative
positions in relation to the 4-beat tactus are fixed but absolute values are variable.
The ensuing metrical dynamism thus spans tactus beats 1 through 4, leaving the
space between beats 4 and 1 of each cycle as a “caesura” of sorts. A fundamental ratio of
3:1 is established at this level of cyclical structure, with a long “active” and a short “rest”
period (see figure 15).
The “Syntactic Arc”
We may be inclined to debate whether to interpret the “at-rest-section” as neutral or as a
syntactic anacrusis to the “active” region, but I would in any case propose to assign a
different qualitative value. A fitting analogy could be that of a “breathing cycle,” with
dynamic motion of breathing in and out followed by a short pause. Another image I
would like to suggest may be that of a waveform, with one wavelength corresponding to
a timeline cycle, yet the crest and trough following the asymmetric distribution:
This notion stems from the work of Cohen and Katz (1997), also discussed later in this chapter,
who discovered a hidden “regularity in intonation” based on an “intonation skeleton” in the Arabic
maqam system, despite the otherwise typically “large scatter in the sizes of each of the diatonic intervals,
even among the average sizes” (34).
active rest
Figure 15 – Syntactic “wave”
The syntactic idiosyncrasy of the paradigmatic timeline and its related patterns should
thus be understood as a system of interacting dynamisms, powered by directional
energies relating to both the serial grouping of the notes in a pattern, and their vertical
integration with the beats of a tactus. These dynamic cognitive schemata combine to
form a cyclical system of high momentary (short range) complexity and directionality,
typical of a large number of non-Western musical traditions (cf. Cohen’s and Katz’
[1997] investigation into the intonational regularities of Arabic music).65
I have attempted to combine these various aspects to the perception cyclical
dynamism dynamic under the notion of the “syntactic arc” in figure 16 below:
Interestingly, the sequential mnemonic conceptualization of native or acculturated performers
mentioned earlier appears to be somewhat independent from above syntactic system, yet not necessarily at
odds with it. Rather, the fact that vocalizations seem to initiate from what would be called ‘attack point 2’
may point to that note’s structural significance, being the first out-of-phase event of a phrasing referent
with the 4-beat fundamental pulse (and as such the first temporal marker in establishing the suggested
syntactic periodicity). Hence, the mnemonics could be said to reference the horizontal progression from
initial syntactic tension towards a release point (‘attack point 1’) in form of an individuating sequence of
temporal/phonetic markers. It would be misleading though, to ascribe further metrical salience to the initial
syllable of a particular sequence by implying a phenomenal accent in the musical texture itself. On the
contrary, it seems that the final velar-fricative, softer syllables (e.g. lo or gan) indicate a particular
alignment with the fundamental tactus. (cf. Kubik 1986).
3:1 ratio: active rest
Figure 16 – Syntactic “arc
Cognitive aspects of timelines – the concept of invariants
But how can the timelines’ function be described? On a practical level, the notion
of “phrasing referent” (cf. Nzewi 1997: 35) certainly applies, and it has been accurately
observed (cf. Locke 1978, 1982, 1996; Pantaleoni 1972a, 1972b; Ladzekpo and
Pantaleoni 1970) how in a percussive texture such as the Ewe’s atsiagbekor the
supporting parts, master drum, dance, and vocal voices time their entrances and maintain
alignment in relation to specific beats of the bell.66
But it is especially the concept of invariants by Jay Dowling and Dane Harwood
(1986: 160ff.) that strikes me as an appropriate cognitive model. According to the
authors, invariants:
are part of a listener’s apprehension of the global organization of a piece.
are “structural constancies underlying surface change in local features.”
Likewise, the clave in Afro-Cuban traditions serves an identical purpose and ensemble
interaction and specific formal arrangements (breaks, endings, etc.) are cognizised and verbalized with it in
“syntactic arc”
bisectional vertical tension state
(in relation to tactus)
attack points numbers (AP’s)
bisectional serial tension state
native sequential cognitive
parsing and mnemonic
representation of full bell phrase
essential syntactic oscillation
“need not be literally represented in the surface pattern,” i.e., can be inferred by a
As we have discussed, all above characterizations hold true with respect to clave and the
notion of timelines at large. But there are some more issues suggested by Dowling and
Harwood that concern us:
a “conflict of different sets of invariants impinging on a given point sets up
structural tension.”
“[t]he meaning of a note lies in the sets of invariants it invokes . . . the meaning of
a note arises from those invariants whose tendencies it violates as well as those
whose tendencies it expresses.”
A “conflict of invariants” is certainly the case with respect to the polyrhythm among
timeline and tactus, i.e., the horizontal parsing of time by a phrasing referent, and the
implicated underlying isochronic pulse(s), and the syntactic arc in figure 16 above may
be understood as expressing such structural tension. However, usage of the term
‘conflict’ should be met with caution. Perhaps a less antagonistic notion such as invariant
system may be utilized, in which a number of invariants contribute to an overall syntactic
A key point to Dowling’s and Harwood’s notion also is that invariants can be
inferred after only brief exposure to piece of music, so pervasive may be their syntactic
coherence (in the tonal dimension we may think of modality or a particular scalar
structure). The authors further explain that such schemata on a general level embody
“knowledge structures developed in our experience of the [musical] world,” but may also
represent more specific knowledge of specific local relationships of tonal or temporal
patterns. Moreover, as in the case of our timeline paradigm, the features of a musical
schemata may be “cognitively important even though they may not be salient in the
stimulus” (125).
Definitions of a Paradigmatic Syntax of ‘Clave’
Whether through enculturation or a subsequent acculturation, the perception and
cognition of timeline schemata clearly involves a process of cultural learning. We can
invoke the concept of a rule structure to interpret their temporal units and assign
meaning.67 And if we may recall Swain’s (1995: 303) statement that a “shared syntax . . .
resides not just in the musical productions . . . but is in the nature of an agreement
between all members, producing or non-producing, of the musical community,” it
becomes apparent that the proposition of a “syntax of clave” may be seen as a culturally
unifying set of cognitive schemata, which define the boundaries of such a community. 68
As such, it is observable to members of such community (i.e., the “community of clave
perceivers”) not only through an externalization of its key morphological units on the
While Cross (1997) has discussed the notion of schemata in its tonal dimension, we can certainly
substitute a temporal context for his quotation of Krummhansl and his further comments: “‘[I]t is
presumed that there is a more abstract, invariant hierarchy of stability that is typical of a musical style more
generally, and that this more abstract hierarchy is an important characteristic contributing to the perceived
stability of each tone with in a complex musical sequence.’ Listeners brought up within a particular musical
culture will progressively form this ‘tonal hierarchy’ as they are exposed more and more to the music of
their culture, as well as in the course of any formal musical training thy undergo. The formation of the
‘tonal hierarchy’ is thus held to depend both on processes of acculturation – most likely non-conscious –
and on explicit and conscious learning” (361). Also see Snyder (2000: 100ff.) on the issue of schemata in
music, and his definition of musical culture as consisting of “music-related schemas – evolving concepts
and practices – that were shared by a group of people . . . [which] would form a context for the perception
and understanding of music.”
In this context, I follow Swain’s (1994) definition of a musical community as consisting of “a
group of listeners within a culture for whom a perceptual object of music is real and has practical value.
There are as many theoretical communities as there are useful perceptions. [. . . ] As far as the notion that
musical community is tied to competence, it is nothing new. Its novel aspect is that small communities,
overlooked as statistically insignificant in experimental research, could yet exercise great influence on a
musical culture and therefore endow their characteristic perceptions with significance that goes undetected
in empirical studies” (311).
musical surface itself (such as the rumba clave pattern or another type of phrasing
referent), but its effects may also be “indirectly”69 experienced through inference, based
on the particular syntactic coherence on part of the texture in general.
The timelines shown earlier in figure 14 reveal a distinctive intrinsic syntactic
unity, because, despite the fact that they may differ morphologically to some extent, they
mediate metrical tension and resolution at the local level in essentially the same way. As
“intersubjective” perceptual objects, 70 they share a common significance and are
assigned a common “practical value” expressing a larger “aesthetic coherence” or
“belief” (Swain 1994: 317) as well as “aesthetic awareness” (Sloboda 1985: 38). Their
precise deep-structural permanence frees up the surface rhythmic structure, perhaps
analogous to the interaction of harmonic and melodic syntaxes in tonal music described
by Swain (1995: 293).
The musical surface of such “timeline music” hence retains a fundamental family
resemblance across cultural domains. We posit the following essential axioms about such
Cognition of a timeline’s syntactic tension is dependent on integrated
perception with the tactus.
The syntactic tension of the timeline/tactus compound is perceived as starting
with the diverging subunit and concluding with the converging half. It is
See Swain (ibid., 309) on the “untenable distinction between direct and indirect observation” of
music theoretically predicated phenomena in the music itself.
Swain (ibid., 317) asserts that “[m]ost complex perceptual objects are not universal, but are
dependent on a community for which they are real and practical; those outside the community can be said
to be incompetent in perceiving such objects.” And Dowling and Harewood (1986) suggest that “[w]hen
we begin to think about the listener’s comprehension of global properties of a piece, we also need to take
into account the listener’s attitudes to listening and the listening strategies those attitudes invoke – the
listener’s set. Two dimensions along which those attitudes might be differentiated are (1) depth of cognitive
understanding, and (2) degree of subjective (vs. objective) involvement”(166).[italics mine]
confined to one cycle-length and repeats itself in alignment with each
timeline’s periodicity.
Syntactic tension builds from a point of temporal juncture along an arc of
countermetrical energy, what Locke has called a counter meter 6/4” (1982:
224) or “onbeat 6-feel” (1996: 105), between the paradigm’s first three
diverging attacks and the tactus. It releases at a second point of juncture, i.e.,
the 4th tactus beat. The arc thus spans from beats 1 to 4. The release cannot
precede the tensioning, that’s why the paradigmatic structure does not start
with a “rest” or silence.71 Attack point ‘1’ thus commands the highest
“psychological centrality” (cf. Sloboda 1985: 42) among all points.
The syntactic dynamism will be highest if the tension of the first half is not
already partially released by coinciding with the 3rd tactus beat (e.g. as in
q q q Ò q ' q ' ), but carried over (e.g.
q q ee Ò q q ' ). The “avoided” third
tactus beat is thus particularly highlighted by implication, and the bipartite
perception strengthened.
Hence, my suggestion is that once a listener is enculturated to a certain degree in above
perception norms, he or she will also be guided by a sensitivity for such syntactical
properties. This makes categorical perception of a texture’s metrical orientation not only
However, as this statement refers only the level of syntactic dynamism between the
paradigmatic pattern and the tactus, this is not to say that in the overall syntactic balance of a musical
texture this dynamism is necessarily the prevalent one. Therefore, in presence of other salient syntactical
forces such as harmonic cadence or melodic phrase structure, it may be perceived as reversed (in the
context of Cuban music commonly referred to as “reverse clave” or “2-3 position. This implies a perceptual
perspective that places the point of metrical centrality on the beginning of the ‘2-side’ of the clave, which,
however, is not articulated by that pattern itself.
possible, but a resource that can be actively exercised even in ostensibly ambiguous
contexts of competing phenomenological saliencies.
Further aspects of syntactic coherence
If we return for a moment to the ‘comparative overlay’ table of figure 14, we notice how
closely some patterns (e.g. 1.2, 1.3, 1.4, 1.5, or 1.14.) resemble the paradigm’s (1.1)
structural composition, coinciding more or less exactly with its attack points. Perhaps not
surprisingly, these patterns are among the most commonly used phrasing referents in a
variety of sub-Saharan traditions, and elsewhere in the African diaspora. They retain a
high level of horizontal asymmetry and contrametric tension with the fundamental 4-beat
tactus and have a high paradigmatic syntactic value. Others, notably 1.7-1.9, while still
close to 1.1, “neutralize” the inherent tension between paradigm and tactus somewhat by
incorporating select beats of the tactus, or additional “filler notes” inserted between the
primary attack points into their morphology. While they may function as a phrasing
referent as well, they do so on a syntactically less implicative level, by partially reducing
the directionality of individual segments of the paradigm. Thus, these patterns are often
found at a – syntactically speaking – lower hierarchical level, and sound in conjunction
with a version of the paradigmatic pattern (we might call such versions archetypes). For
example pattern 1.8, which is called guagua or palitos in Cuban music, is played together
with the rumba clave (1.2) in the rumba columbia, or on a hoe blade called guataca in the
ritual music of palo. Nonetheless, it correlates highly to the paradigmatic “ideal” and
contributes to an overall syntactic coherence, though strictly speaking, it should already
be considered foreground material.
Interestingly, as in pattern 1.6., an attack point of the background paradigm (1.1)
can also be sharply defined even by its very absence. Agawu (1986) has already
commented on the use of the ‘silent downbeat’ in African music, which “creates tension
by propelling the music towards and beyond the nominal downbeat (72).” As such,
pattern 1.6 could be interpreted as a foreground elaboration of 1.1, even though it differs
only slightly. The phenomenological salience of attack point ‘1’ (the “downbeat”) is
established through the suspension of temporal category C =12 into C = 2,72 its strength
drawn from what Agawu calls “a powerful rhetorical use of silence (72).” We need to
understand that herein lies an important connection among the various morphologies of
timelines: they need not to literally match the paradigm’s structure, but need to imply it,
the same way in example 1.6 attack point ‘1’ is referenced by its very omission!73
Incidentally, Pantaleoni (1970b: 75) reports that, when in Anlo drumming “this felt point
is avoided, . . . the drummers say the stroke has been put ‘inside the time’ (‘vugbedemi
or ‘vumamla’).” In other words, timelines and other foreground patterns that may not
appear to be connected on such levels nonetheless adhere to the dynamism predisposed
by such syntactic forces!
Examples of paradigmatic pattern integration in Cuban music
For the reader to be able to experience and appreciate this dynamism in the more
concrete terms of an actual musical context, I would now like to examine a few more
With reference to Parncutt (1994: 425), here C denotes one temporal unit within a metrical
matrix consisting of 12 beats (i.e., 12 eighth notes in 12/8 meter).
One audio sample documenting pattern 1.6 actually starts with pattern 1.2 and then continues to
alternate between 1.5 and 1.6 (cf. “Culte Congo: Palo Monte.” On Cuba: Les Danses des Dieux, Ocora C
559051, CD recording, track #4).
pattern structures with regards to their particular syntactic “well-formedness”74 or
specific resonance with the paradigmatic pattern.
The comparsa bell pattern, typically played on one campana (i.e., cowbell) using
its “mouth” for the low pitch and the top for a high pitch, may also be performed on two
separate bells. It stems from the traditional music played during Cuban carnival,
predominantly in eastern Cuba. In the context of popular music it is often played by the
timbale player, who may also concurrently play the clave indicated below.
Figure 17 – Comparsa pattern integration
Track #5
The two patterns are always aligned in this way, and if either pattern were sounded
individually by different players, this alignment would be established among both.
Particularly, I would like to draw attention to the distinct interlocking sequence of
attacks in the ‘3-side’ of the pattern, which seem to create a figure-ground relationship
The term well-formedness has been adapted by analogy from Lehrdahl’s and Jackendoff’s
(1983) Generative Theory of Tonal Music, where it is one of a set of two rule structures that are said to
guide a listener’s cognition of a given piece in (Western) tonal music. However, its usage here is by no
means to imply the particular criteria (i.e., Western metrical structural conventions) associated with it in
that theory.
between bell and clave. Beginning with the event marked by the asterisk, the two voices
seem to oscillate around each other, each filling in the space left by the other. The
syntactic energy generated by this propulsive “dance” is partly balanced out on the ‘2-
side’ due to the simultaneous attacks of both lines. I say partly, because the second of the
attacks on the high pitch level (circled) is effectively generating a stress point of higher
phenomenological salience than the one coinciding with the tactus, due to its longer
duration. It challenges the commetric “meaning” of the preceding attack and possesses a
higher rhythmic “gravity”, so to speak. Recalling Tozzi’s notion above, this translates
into a higher local tension-state or syncopation with respect to both clave and tactus,
while maintaining the proper syntactic flow between the clave and bell.
This metaphorical notion of rhythmic gravity is created by a composite of various
cognitively ascribed saliencies, such as duration, pitch level, and timbral quality. Like a
heavy physical object in space, it can “warp” the metric interpretation of a texture by
“pulling” the listener’s metrical abstraction process into its sphere. If this occurs in the
context of an already clearly established metrical scheme, a high level of local tension
will be generated. The syntactic dynamism implicit in the integration of these two
patterns in figure 17 is thus not necessarily rooted in any specific synchrony among its
constituent parts, but in their contrapuntal, mirror-like implication of each other – a
mutual invocation if you may – and hence the figure-ground analogy earlier. In general,
African music brilliantly exploits this process for aesthetic purpose.
Often in these kind of patterns, a note with syntactically defining properties
actually receives further accentuation in performance, either by way of additional
dynamic accent or through relative dynamic accent. The latter is achieved when the
preceding or succeeding notes are “ghosted”, or, “negatively” accented so to speak. In a
‘short-long’ durational sequence for instance, the short note may be ghosted or even
completely omitted without altering the essential syntactic coherence of the phrase.75
Below we see a version of this pattern, which has been reduced to its fundamental
syntactic structure. At this point we may note that there is only one point of simultaneous
attack between the background paradigm (i.e., rumba clave) and the bell, the ‘and’ of
beat “1”. Despite that, a listener or performer experienced in the Cuban tradition would
have a little problem identifying the correct relationship between these parts.
Figure 18 – Comparsa bell “skeleton”
Track #6
The existence of a simple alternating pitch scheme of low-high differentiation
adds yet another level of syntactic directionality, conferring a phenomenological accent
pattern that clearly marks the fundamental unit of the tactus. Whether high pitched tones
are interpreted by a listener as a metrical marker, e.g. an implied “backbeat” accentuation
on tactus beats 2 and 4, depends on an individually variable cognitive predisposition.
Incidentally, one of the most ubiquitous patterns in Brazilian music features a
structure identical to the two examples presented here, the full version often sounds on a
We will encounter this effect again in the cáscara example below.
two-tone bell called ago-go, while the durational “skeleton” has paradigmatic status in a
variety of contexts and is played by tonal-harmonic instruments (e.g. guitar) and
percussion (e.g. tamborim) alike, or may operate only as a background structure.
Typically though, it is performed and conceptualized with a slightly altered ‘high-low’
melodic contour and dynamic accentuation on the higher pitched notes:
Figure 19 – Brazilian paradigm
Track #7
The next example is also from a comparsa style, the more recent mozambique, in
which two bass drums called bombo play highly syncopated sequences of muffled and
open strokes (here, x denotes the muted strokes which are achieved by striking the drum
with one stick while the free hand mutes the head), that again give the effect of distinct
high and low pitch strata. This pattern may be played by the secondary bombo:
Figure 20 – Mozambique pattern integration
Track #8
Again, the tension created by the counterpoint of clave and bombo is build up in the ‘3-
side’ section, and releases on the final of the three open strokes (the “1” in this notation).
As in the case of pattern 1.6 in figure 14 we note that here, too, the initial beat of the ‘3-
side’ is implicated by its omission, while the “empty” downbeat of the ‘2-side’ is being
In general, and applying to both a ternary and binary metrical context, the
syntactic stability of interlocking patterns is reinforced by such schemata, like for
instance the tendency of a foreground figure to avoid marking a note position coinciding
with a metrical stress-point of the paradigmatic background, and vice versa. In music that
adheres to this arrangement, a delicate contrapuntal balance is established, resulting in a
particular degree of and identifiable quality of rhythmic tension. Were we to reverse the
alignment of for instance the comparsa bell pattern discussed above, a less “driving”, i.e.,
syntactically satisfactory compound pattern would be the result:
Figure 21 – Reverse comparsa pattern integration
Track #9
For once, we note the higher level of simultaneous attacks between bell and clave, a
metrical redundancy of sorts, that neutralizes the contrapuntal “handoff” characteristic of
the first example. Moreover, the syncopated relationship of the circled note in the bell
with the preceding clave attack (the second attack of the ‘2-side’) also has been lost in
this case. And in above version the alignment of the ‘2-side’ establishes virtually none of
the motive energy with the bell. On the contrary, the two parts are locked in rigid
synchronicity that make the phrasing of the composite pattern stiff and “square.”
The quality of a rhythmic organization influenced by a syntax of ‘clave’ can thus
clearly not only be the result of synchronicity or simultaneous alignment among parts, as
a superficial analysis of this issue sometimes might have it. Rather, the syntactic “best-
fit” involves context-dependent rules of polyphonic sequential combination, that, as they
are integrated into a cognitive whole or gestalt, generate a scheme of vertical tension as
well. An interesting approach to further research may be along the lines of John
Sloboda’s (1985: 38) discussion of Sundberg and Lindblom’s notion of “prominence
contour”76, which assigns each melodic “element”77 a hierarchical prominence value, and
a generative function in the temporal and tonal dimension based on such value. Perhaps a
similar contour could be created for the short-range directionality of the clave/timeline
The cáscara presents another good example to portray the deep-structural
relationships within the clave , even if this pattern doesn’t quite boast the same
syntactical permeation as the bell and bombo pattern above.
Sloboda explains: “In speech, such contour is used to assign stress, timing, and intonation to a
sentence. In this music grammar, the contour is used to generate appropriate chords and durations (38).”
Sundberg and Lindblom used this notion in the context of eight-bar Swedish nursery tunes in
4/4, with each element representing one quarter note.
Figure 22 – Cascara and clave integration
Track #10
Crucial for the correct articulation and feel of this pattern (typically played on the side of
the timbales in Cuban music) are the dynamic accents indicated for attacks 1,3,4,6,7 and
9. These accents may also constitute a pattern by themselves, though usually the
remaining unaccented attacks are being played, albeit at a lower dynamic. Yet it actually
is this newly emerging pattern that we need to investigate for its syntactic properties, as it
frames each fundamental tactus beat in a very particular fashion and either reinforces or
negates the additional asymmetric directionalities of the implicit clave. 78 The dashed
lines point to a possible interpretation of such syntactically fluid integration. Note that the
cáscara accents are partly out-of-phase with the clave attacks, but still reference its
fundamental structure.
The resulting counterpoint is perhaps not as intricate as in the comparsa
examples, but nonetheless existent.79 Revealing as well is the way the cáscara
Whether the clave is actually sounding or not shouldn’t make a difference from a cognitive
standpoint. A performer/listener enculturated in these perception norms will supply this invariant schema
either way, and the sensation of syntactic tension will be generated. Likewise, consider that the tactus is an
equally abstract, if perhaps more easily maintainable, mental template that may or may not be externalized
audibly or otherwise.
The reader may perhaps wonder about the varying time signatures involved in the notated
examples. I feel it is good practice to observe the same pattern morphologies in various metrical contexts to
counter the particular preconceptions or “cognitive habits” sometimes associated with a particular
rhythmically frames the isochronous anchor points of the tactus, especially its second and
fourth beats (circled). As the brackets indicate, the resulting integrative flow is
qualitatively slightly different in each case, with attacks 3 and 4 providing for a “harder”
framework than the more commetric 7-8-9 sequence. Both have in common that the
cáscara emphasizes the eighth-note immediately following, creating a local point of
tension with the tactus. Their phenomenological quality however arises from a different
syntactic constellation in each case. While attack number 4 seems to grow out of the
syncopated directionality set up by the second attack of the clave, number 9 clearly
bounces off the metrical “weight” of tactus beat four and the concurrent clave attack. The
immediate temporal context surrounding each attack, especially the distance from its
preceding note and that note’s own metrical salience,80 is thus highly important.81 As
such, AP 3 sets up 4 differently than 7 does 9, creating motivic units of different metrical
implication, both from the perspective of the clave and the tactus (attack 8 is concurrent
with the tactus and clave and does not generate additional syntactic energy).
In general, we may examine how the combined sequence of cáscara accents and
the clave form a distinct relationship at each of the four tactus beats: only on the first do
all three invariants coincide. The second one is completely empty, while the third and
fourth each receive attacks by either the cáscara or the clave:
representation. Flexibility in this respect may make spotting familiar musical structures across the great
variety of notational practices more successful.
Attack point 3 coincides with the clave, but neither with the fundamental tactus nor the next
subdividing level below. AP 7 coincides with the clave as well, but also with the next subdivision below the
tactus level.
This kind of low-level relationship among events is equally common in the tonal dimension,
where harmonic and melodic “meaning” are often tied to immediate context.
Figure 23 – Cascara accent pattern
Track #11
By comparison, the comparsa bell skeleton shown earlier has no coincidences between
clave and the bell foreground at any tactus beat at all. We can conclude that in each case
these concurrent timeline patterns not only exhibit a distinct horizontal alignment
resulting in a specific linear directionality, but each also relates to the fundamental tactus
in a unique way. For a performer this interplay of syntactic energies in both horizontal
and vertical dimensions translates into a distinct kinesthetic sensation, that is felt perhaps
most immediately if all its components are physically maintained by a single player. But
a listener may also pick up on this sense of flow.
Horizontal harmony
In assessing the syntactic quality of these pattern structures, we are presented with a
intricate web of musical energies in both horizontal and vertical dimensions, and at least
three strata of invariant schemata, two of which are often actively supplied by a
listener/performer: background paradigm (clave), foreground (e.g. bell or cáscara), and
tactus. While we can experience them as independent linear structural components in a
particular configuration, there may also be an impression of a “harmony” in a temporal
sense, based on a) the proportion of compound “weight” created vertically by coordinated
attacks among all three strata around a given point and b) the flow of horizontal
directionalities in relation to the harmonic “progression” established by a).
In a way, the cognition of a particular structural identity of simultaneous events found in
the pitch domain (e.g. in their consonance) and the relationship of adjacent notes to
them, perhaps can be said to find an analog in what Nzewi (1997: 54) has called
“harmony, perceived in horizontal affect,” a holistic sense of calibrated tension and
proportionality between the linearly progressing, but also vertically integrated patterns.
Hence, the repetition of events within an essentially unchanging period sets up a
structural tension among its constituent parts, in which “[h]armonic effect/affect
transpires from one Ensemble Starting Point of the ETC to the next (53),” in the above
cases consisting of one clave cycle.
Thus unlike in the tonal understanding of harmony, 82 this notion entails the idea
of harmonic congruence as a particular temporal association of notes, the result of an
integration of all the syntactic dynamics present. The harmonic identity of a group of
notes, in the tonal dimension cognitively bounded by their – more or less – simultaneous
vertical attack point, is translated into a – more or less – stable horizontal temporal
resonance, similar perhaps to the “block harmonic sense” suggested by Nzewi (54).83
For the reader critical of such attempts of conceptual translation, I would like to offer the
following remark by Swain (1995): “The power of any analogical argument stems from an unfortunately
vague sense of aptness, some essential similarity between two things that invites other comparisons in the
effort to make the better known reveal secrets of the lesser known (302).”
Nzewi further explains that:
“[t]his block harmonic sense continues with changing internal aesthetic enhancements until a significant
thematic change occurs in any of the component themes constituting an ETC, commonly the recognized
principal theme/layer. This would constitute a movement, a new progression, from one thematic/harmonic
block to another . . . Within this horizontal logic of block harmonic thinking, the underlying sensibilities
about vertical, note-by-note traditional concord and discord, particularly where definite pitches are
I argue that all of the examples presented so far retain a harmonic quality in this
sense, maybe slightly varying in external character, but unified in its relationship to the
syntactic centrality provided by the clave and expressing the aesthetic preferences that
may be subsumed under such musical concept. If we keep in mind that a particular
temporal gestalt is generated as result of the listener/performer’s intuitive appreciation of
such harmonic quality, we may get closer towards recognizing its salience and function
as a cognitive cue in the process of metrical organization and in the hierarchical
construction of a musical texture, a position already well established for the notion of
harmony within the context of tonal music.84
Intonation type
Another interesting theoretical conception connected to our discussion also emerges by
analogy to the pitch organization in the Arabic māqam system. Cohen and Katz (1997)
There must be some regularity that allows performers and listeners to classify the songs by
māqam, albeit unwittingly. Meticulous examination with the aid of a melograph revealed . . . a
latent regularity that ‘translates’ the theoretical rules into rules governing performance practice . . .
this regularity reflects the cultural preference for short-range directionality and great complexity,
which in turn increases the possibilities for improvisation. (33)
While the authors focus on issues of micro-tonal pitch structure, especially the hidden
“complex regularity of intonation,” their concept of an “intonation type” (36) may be
analogous to my notion of a syntactic paradigmatic structure. They affirm that “the term
involved, is not compromised . . . [yet][h]armonic thoughts and processes are . . . conceptually linear and
intuitive in African music (1997: 54)
Cf. Swain’s (ibid.) discussion of the characteristic circular relationship of metric and harmonic
tension, p. 298.
‘type’ can refer to other parameters as well,” whose particular factors may be examined
with respect to their contribution to directionality and complexity, each based on a
different set of compound variables.85 I believe that the “latent regularities” I have tried to
interpret as part of a syntax of ‘clave’ may similarly represent a “type”, albeit controlled
primarily by temporal factors. Certainly, a number of the variables (cf. p. 51) suggested
by Cohen and Katz have been part of our discussion already.
I have already pointed to the relatively local level (i.e., the clave cycle) at which a
rhythmic syntax is most effective – its short-range directionality, so to speak. The high
degree of complexity at the same level should come as no surprise to anyone familiar
with Cuban or West African “timeline music”. As such, Cohen’s and Katz
characterizations of their particular repertoire’s directionality, complexity, and its
contributing musical factors, share many communalities with the African repertoire and
perhaps indicate analogies in cognitive processing:
an “ideal of suspensive momentary directionality” with “maximum momentary
complexity”, i.e., “focus on the moment.”
intervallic relations among events (pitch or durational) are maintained as relative
ratios rather than absolute values (a “skeleton” or scheme).
exact interval sizes are context-dependent; all realizations derived from a single
“type” are considered equivalent.
the phrase “is the main musical unit that can be considered in terms of
Cohen and Katz (1997: 31) list the following compound variables: “(1) quantity of elements on
the various levels; (2) their degree of definition; (3) their degree and kind of distinctness; (4) their
dependence on or independence of context; (5) the kinds of hierarchies; (6) the levels at which schemes
appear; (7) concurrence/non-concurrence among musical phenomena; and (8) the relationship between
learned and ‘natural’ parameters. All of theses are subject to cognitive constraints.”
numerous complex “schemes [exist] in various parameters on the immediate
“types” provide an “option of maximum improvisation within the limits of the
The authors suggest an overall classification of their repertoire into the theoretical
framework of the ‘māqam system’86, “which the performers use but are not able to
articulate,” and a second modal framework (depending on the first) consisting of “groups
of songs that are like genres . . . well known to the bearers of the culture who are able to
discuss these genres (32).” In this respect, I see parallels to the concept of clave in Cuban
music as an analogue theoretical framework, whereby explicitly theorized awareness
among performers/listeners may be limited,87 but a highly developed sense towards its
operational properties and aesthetic effects is commonplace. Genres like rumba, son,
abakuá, or many others in the Cuban tradition would then be defined by the modal
framework, involving rules governing, for instance, their particular melo-rhythmicity,88
hierarchies of functional stratification among instruments, or extramusical/contextual
Rules governing the māqam system relate mainly to aspects of pitch organization (cf. Cohen and
Katz 1997: 32): “The māqam ‘determines’ which notes (degrees of a scale) or intervals are relatively stable
and which ‘require’ a large scatter [. . .]. Similarity between the skeletons of the songs in a single māqam is
obtained by preserving the ratio between the sizes of the consecutive seconds in the scale (the ratio is
expressed as ‘greater than’, ‘less than’, or ‘equal to’, rather than as a precise number) (34).”
Snyder (2000) notes that schematic cognitive frameworks that are relevant for the formation of
syntactic information may be part of “semiactivated memory contexts” and “an example of implicit
memory use” (104). They thus not necessarily “exist in the form of explicit knowledge (102)” but “operate
unconsciously to contextualize current experience (95).”
With respect to Nigerian folk music, Nzewi (1974: 27) has coined the term “cultural
instrumental melo-harmonic ethos” that is produced by lineal and vertical melo-rhythmic interdependencies
within an ensemble.
To conclude our investigation into the factors that make for a particular “well-
formedness” within a syntax of clave, let us look at a few more musical examples which
may exhibit a little more ambiguity in this respect.
Perception of syntactic “well-formedness” in Cuban Arará
The figure below shows a bell pattern used in the arará tradition of Cuba. It has been put
in the context of two clave orientations, so the differences in alignment are clearly
visible. On the source recording, the tactus level is marked by shekeres, so establishing
the fundamental 4-beat ground is not a problem. Both melodic structure and formal
segmentation (in lead voice and choral response sections) indicate a cyclical starting
point as implied by the notation below. Less obvious is the existence of a background
paradigm, let alone its directional orientation. An acoustic externalization of clave in the
sense suggested below is never actually played, even though a different bell pattern is
briefly sounded in the beginning, which hints at what could be called the “preferred clave
direction” below. But what in this timeline may be seen as indicating the “preferred”
orientation over its “reverse” counterpart? While the boundaries between these
distinctions are not etched in stone, a certain syntactic pervasiveness seems to prevail in
one case but not the other.89 It can not be due to quantitative properties such as number of
coincidences of attacks among the bell pattern and the clave – in both alignments there
are three.90
In experiencing this “reverse orientation”, the bell pattern is felt as beginning on attack point “1”
of the reverse clave, i.e., from its midpoint, and coinciding with the third tactus/pulse beat. Its serial
rhythmic contour (sequence of inter-onset-intervals between events) is hence subjectively changed, as the
new alignment with the clave suggests a new paradigmatic point of temporal centrality and schematic
The fact that the suggested rumba clave does not sound and may only be present as a cognitive
schema, can not be said to automatically take away from its cognitive salience for a listener /performer,
Figure 24 – Arará (“Song for Ebioso”)
Track #12
One factor relevant to the syntactic quality of the preferred orientation, is its repetition of
a short motivic unit (see boxes above and figure 25 below), which by itself creates a
strong linear energy culminating in a unison attack (circled):
Figure 25 – Arará motivic unit
The crescendo mark in this case is meant to express a dynamic build-up in motivic
tension as well as physical movement intensity between hands, 91 which is released on the
final attack.
Two of these motivic units combine to a longer phrase which resolves on tactus
beat “4”, tracing the essential “syntactic arc” proposed earlier. In contrast, the “reverse
clave” direction does not create such nested dynamism – but not only that: it also creates
however. As previously remarked, such schemata need not be found on the musical surface to be relevant
from an experiential perspective.
To be experienced when actually performing both patterns together.
an alignment with the bell whereby its structurally meaningful AP’s “2” (representing the
anticipation of a fundamental pulse beat) and “5” (coinciding with a pulse beat and
closing the syntactic arc) receive no externalization. In addition, the coincidence of the
first of the two double-eighth-note attacks of the bell with the clave (attack points “4” and
“1”) does not establish the kind of forward directionality that is characteristic of the
syntactic congruence among both patterns when they link up on the second attack (AP’s
“2” and “4” in the preferred orientation). That may be, because these second eighth-notes
have a stronger phenomenological accent based on their higher durational “weight” (since
the distance to the next attack is actually a quarter note), which in turn creates a rhythmic
grouping boundary at the level of primitive perceptual grouping (cf. Snyder 2000,
p.33ff.). This by itself not only creates a salient tension with the tactus because the
release of motive energy is suspended (i.e., syncopated), but in the “preferred case” also
synchronizes with the implicit metrical tension established by the clave in its relation to
the tactus.
Furthermore, in performance (see also the cáscara example) such longer second
note in a pair of attacks often receives a light dynamic accent,92 which is vital for a
pattern’s “forward drive.” In the reverse direction on the other hand, concurrent attacks
From the point of physical motion, these combinations of two strokes in a short-long ratio may
also be seen as forming a distinct kimene (cf. Thiago de Olivera-Pinto [1991: 125] ), i.e., the shortest
individual motional unit making up a pattern, which also embodies a certain directionality or “goal-
orientation.” In this particular case, due to the percussive stroke-mechanics involved, a dynamic/motor
accent can be more easily imparted – and accurately timed – on the second and final event forming the
kineme (making the first one more of a “pick-up” motion) than on the first.
The notion that sound patterns and movement patterns may be experienced as cognitively highly
integrated and conceptually unified, has been advanced by a number of scholars (e.g. Olivera-Pinto 1991,
Baily 1985, 1992, Kubik 1979) and should be considered an element in the formation of a particular
musical-syntactic tradition.
Baily (1992: 149) proposes that “[t]here is a precise isomorphism between music structure and
movement structure: every nuance in the micro-structure of the sound pattern reflects a subtle adjustment of
the motor pattern,” while Kubik (1979: 229) asserts that the “change in the motional pictures brings about a
change, even if only slight, in the exact ‘spacing’ of the notes to be struck,” destroying “the original
accentuation,” and leads to “a sense of a lack of drive.
between clave and bell in general precede syncopated attacks, in a sense putting the
resolution of motion before any motive energy has actually been projected.93 The result is
a slightly “limping” contrapuntal sequence, which, as a whole, lacks the linear
directionality, fluid psychomotor integration and cohesive syntactic identity formed by
the preferred alignment.
Aspects of schematic categorization – an example of timbral stratification
We have seen that a cognitive approach to clave – indeed a concept of ‘clave’ – as a
syntactically relevant notion of schematic categorization needs to incorporate various
categorical hierarchies and syntactic relationships between the acoustic events involved,
both from the perspective of a passive listener, as well as a performer’s active agency in
constructing them. The intensities and types of syntactic momentum established by these
various schematic categorizations which are constructed from the musical surface
(whether actually present acoustically or through inference), shape our experience of it,
and are highly diverse.
They result from the interaction of short-range grouping of rhythmic units of
largely horizontal directionality with a more consciously constructed vertical integration
that may be categorized as a higher–level process of syntactic tension and release. Many
musical parameters can be responsible for this linearity, and while we seem to be
concerned mostly with issues from the temporal domain, by no means are these the only
defining ones. Pitch stratification for example, even a relative simple high-low
On the contrary, the preferred clave direction in this case creates a strong tension among both
patterns at the beginning of the cycle, because the first clave attack point is avoided by the bell pattern (a
similar case to the comparsa bell discussed earlier), but the point of bisection is made audible.
relationship, can provide for additional cognitive segmentation that may exert its own
directionality, and have an influence on the overall syntactic stability.
Regarding this particular issue, Pressing (1983b) has made some interesting
observations regarding the play of kidi (one of the mid-size support drums in a number of
Ewe ensembles) in agbadza, a dance and musical style of the Ewe. When analyzed in
relation to the position of the standard bell pattern’s individual attacks (cf. figure 12), the
boundaries created by the points of timbral transition from units of press (muted) strokes
to units of bounced (open) strokes seem to follow a distinct “tonal logic”(14). Pressing
states that “[p]oints of tonal transition”(ibid.) appear more frequently at certain positions
within the bell cycle than at others, as some areas of the cycle receive a higher amount of
bounce strokes, while others are characterized by a higher incidence of muted strokes.94
For instance, attack points 1 and 3 of our paradigmatic pattern are emphasized by such
transitions, which, in almost all cases (the author samples a total of 15 different kidi
variations), are “left relatively free of bounce strokes”(ibid.). Indeed, such consistent
relations are what a musical syntax is all about, and it appears that even patterns of
simple, purely registral changes of pitch and timbre can create their own directionalities,
which in turn may guide further higher-level schema formation.
According to Pressing statistics, then, the bell cycle induces the musical
organization of the kidi with a distinct formalism and grammatical demands: “[T]he
average distribution of bounce and mute strokes with respect to the gankogui cycle . . .
should be quite non-random, since all kidi patterns are . . . intended to be conceptualized
with respect to the bell pattern (13).” And despite the fact that there still is a large
Pressing determines a bounce-to-mute ratio for each of the twelve positions within a cycle (see p.
13), which renders at sort of very generalized registral melodic contour, based on the average of each
stroke-type’s occurrence.
structural variety among specific kidi phrases in agbadza, overall the ‘high-low’
“registral melodies”(5) created by kidi appear to be shaped by such higher-level (albeit
still at the level of a single cycle) syntax of “tonal relations” in relation to the bell,
specifically by the manner “[t]one is used in a fashion which emphasises certain
consistent relations with the timeline, primarily that of reinforcement of distinctive
positions of tonal change”(15).
But that would imply that also the other way around, from both the perspective of
where these timbral changes happen and what type of transition exactly takes place (cf.
p.14-5), one should be able to infer the properties of the bell in case of its absence. That
in fact makes for a significant realization in the context of this study: the play of kidi
shaped by the timeline’s structure – exhibits a subtle dynamism that in turn mirrors the
bell’s linearity in a metaphorical way!95
The significant point for this discussion is not so much what kind of specific
structural references are in operation in this particular musical case, but rather that they
can be made at all. Therefore, the most interesting aspect is that a position of change
based on a simple timbral dichotomy, a parameter of an entirely different acoustic
quality, in itself can help in forming abstracted schemata of higher-level categorization –
which then even may suggest another, equally only cognitively implied scheme (in this
case the bell) in a entirely different domain (in this case durational intervals). While this
has been acknowledged for the corpus of Western tonal music (e.g. changes in harmony
motion may indicate metrical segmentation), Pressing’s study is unique in drawing that
analogy in context of a purely percussive texture, and correlating the qualitative effects of
Perhaps the term morphological metaphor could be an appropriate notion for such a relationship
– one pattern creating a metaphorical reference for another
a simple two-timbre pattern categorization (i.e. mute/bounce strokes) to the formation of
a separate cognitive entity.
At least as far as agbadza is concerned then, we can assert that the syntactic
pervasiveness of a timeline’s linearity may also be experienced indirectly, suggesting its
(and by functional analogy also our paradigmatic pattern’s) real significance as a deep-
structural organizational force and fundamental cognitive schema.
Paradigmatic Interpretations of Musical Examples
The kind of rhythmic resonance under discussion in this thesis may also be conjured up in
absence of such timeline, through a pattern’s syntactically related structure, additional
cues in its articulation, or through cues provided by other instruments’ patterns in an
ensemble. That is what I would now like show for the following examples of music, in
which a listener’s sensitivity towards this resonance may be instrumental in establishing a
unified metrical coherence with regards to the both its cyclical centrality (“where is the
downbeat”) and – perhaps even more important – its fundamental pulse or tactus (“where
and what is a beat”).
Both are discrete issues and solving one question doesn’t necessary provide an
answer to the other. And I posit both need to be solved as part of the process of
categorization of a musical texture.96 While in some cases many possible frames of
reference appear to emerge at once, as we have seen in the above patterns from Benin, I
suspect that eventually one primary perspective can be extracted given an adequately
sensitized perception. This can be said to be true across different diachronic musical
Independently, such categorization already starts at a neuropsychological level, and Vercoe
(1997) has pointed out its effects as “prime generators of human perceived rhythm and meter (317).”
parameters, whether temporal or pitched. It is thus highly likely that at any given moment
only one cognitive perspective is adapted by the listener.97
Maintaining a stasis among multiple perspectives does not make much sense from
a point of cognitive efficiency, i.e., memory constraints, since, as Snyder has pointed out,
[. . .] organization of memory and the limits of our ability to remember have a profound effect on
how we perceive patterns of events and boundaries in time. Memory influences how we decide
when groups of events end and other groups of events begin, and how these events are related. It
also allows us to comprehend time sequences of events in their totality, and to have expectations a
out what will happen next. Thus, in music that has communication as its goal, the structure of the
music must take into consideration the structure of memory – even if we want to work against that
structure. (2000: 3)
That does not preclude the possibility that an experienced listener can actively chose
different categorization schemes, or even alternate among them in rapid succession (Cook
1994: 89). In addition, it is possible that the points of centrality established by different
syntactic dynamisms from a variety of dimensions – tonal, rhythmic, timbral, dynamic
stress – may be integrated according to the constraints of a higher “meta-syntax”, so to
This meta-syntax should be as culturally predicated as the constituent sub-
syntaxes themselves, and I propose that this is what happens in the formation of a overall
metrical abstraction. For a listener or performer from a certain cultural background, one
Whether this cognitive perspective initially includes the sensitivity for the particular syntactic
dynamism presented here or not, is of secondary concern. Perhaps the reader’s perception, i.e., “understan-
ding” of the various examples of music discussed so far may already have changed by the end of this
chapter, just by having been alerted to the presence of such syntax. But for this theoretical discussion to be
valid it is not imperative that it need to. As Cook (1994: 89) has asserted, “the aim of music theory . . . is to
go beyond perception.” It’s “a particular way of thinking about” a piece, that may “change the way people
experience the music.” It is not a psychological model of perception itself. And Swain (1995) concedes that
“the character of each hierarchical theory of music arises from the syntax imagined [. . .] Discovering or,
indeed, creating a musical syntax guarantees nothing in the way of music’s value or perceptibility – that
requires an external criterion – but it does guarantee the possibility of comprehension, at least.” (italics
syntactic schema may override other competing schemas, which in musical practice
means that for instance one performer may prefer to establish a metrical centrality on the
basis of tonal relationships among events or their serial grouping structure, while another
may chose dynamic contour, timbral contour, and yet another, as suggested in this work,
the dynamic organization afforded by the syntactic forces of ‘clave’.
First, I would like to turn to a couple of audio examples in which the presence of
our paradigmatic syntactic organization is fairly obvious, and often cued by pattern
structures that resemble one of the paradigmatic timelines’ organization completely or in
part. However, the difference is that a timeline per se is not externalized in the music, so
the categorization of events must take place in relation to an abstracted schema. The point
I am trying to make here is not that such music is hence necessarily conceived of in
explicit relation to such timeline (even though it may very well be), but to simply indicate
a syntactical coherence that may connect it with a larger corpus of African musical
traditions. Recognition of such coherence may in turn be helpful in a theoretical
interpretation of its structure and help decide, if only after reflective analysis, among
competing meanings. In other words, possibilities could be narrowed down to
probabilities. At this point I highly recommend listening to the sound examples included
on the accompanying CD, as the transcriptions are not intended to give a complete
representation of the music, but tend only to highlight parts that are especially salient for
the perception of syntax in the sense discussed throughout this chapter.
As for the first example (figure 26) , a section of a piece98 from the Dagomba
tradition called Nagbiegu, the gungon’s pattern in conjunction with the isolated attacks by
The lead lunga phrase transcribed here happens at ~ 1:30min. into the recording.
the support lunga99 not only creates a formal segmentation into periods of C=24 units,
but also invokes the metrical directionality inherent in the syntactic paradigm by
articulating two of its AP’s (see box).
Figure 26- “Nagbiegu”
Track #13
Next a fragment of the call-and-response cycle between a solo voice and a chorus
of one-stringed horsetail fiddles called nyanyer from the Fulani people of Senegal.
Figure 27 – “Mariama”
Track #14
Again we notice a correspondence with an implicit paradigm because of the low drum’s
pattern, but also a due to a more subtle association with the phenomenological accents
See Locke (1990: 26ff.) for a detailed description of the gungon, as well as the lunga
(ibid.: 30ff.), and their associated playing techniques.
created by the pitch contour and durational organization of the nyanyer response phrase.
The alignment of its internal sub-phrase boundaries with attack point ‘3’ of each cycle
seems to be of particular significance in establishing this syntactic correlation.
Figure 28 – “Ekun Rere”
Track #15
Figure 28 shows an excerpt of a juju song from Nigeria by I.K. Dairo and his
group. The “stop-time” section that occurs at beginning the song and which otherwise
alternates with the section notated in figure 29 below, is characterized by an initial sense
of metrical ambiguity. In particular, the introductory phrases played by the guitar in
figure 29 are only in retrospect heard as corresponding to a timeline. Instead, as can be
seen in the brief excerpt, in absence of such metrical categorization the melodic contour
first suggests hearing the Db-Ab ascent as an agogic accent, or perhaps the final pitch Ab
at the end of such phrase as a downbeat. In any case, neither scenario quite relates well to
the metric directionality established by the ensemble attacks (a Gb pitch). Consequently,
a strong indication of whether to integrate those into a cognitive schema that places the
centrality on the first or second attack remains somewhat elusive.
Figure 29 – “Ekun Rere” (stop-time section)
Track #15
At this point, neither tactus pulse nor metrical centrality seem settled. Yet the chorus
sections maintain this “ambiguity”, and the contrametric accents of the shekere’s pattern
– in conjunction with its almost ternary phrasing – seem to lead to a displaced perception
of the tactus (the accent is be perceived as indicating the pulse).
But is this really a structural ambiguity, where two or more theoretical
interpretations would be equally viable? If so, their meanings would need to be
“comparably or equally plausible”(Agawu 1994: 89), and we would have to be “prepared
to specify the plausibility of each meaning in relation to every other meaning (ibid.).”
Granted, a listener not accustomed to such textural organization and not expecting the
potential syntactic correlations frequently observed in the music of Africa and its
diaspora may miss cues that could offer further information towards a solution. And even
with such knowledge and intuitive sensitivity, a single interpretation may not form
instantly but may only come after – intensive – reflection. For instance, we may continue
to hear the circled Gb as possessing downbeat “gravity,” so to speak. But as Agawu
asserts, “[once] a specified context and a specific metalanguage intervene . . . the
interpretation of a musical event as ambiguous in the strict sense becomes untenable
In this case one subtle hint at a “context” is in fact provided right from the
beginning of the tune in form of a talking drum pattern, which albeit in the background,
can be heard articulating the attacks of the very timeline indicated in the transcription (to
draw attention to this literal correspondence, I have notated the complete “standard
pattern,” not just the paradigmatic version indicated elsewhere). That same
correspondence is also maintained in its accompaniment of the complete ETC (see boxed
set of notes). Identification of such implicit timeline may in turn immediately serve to
infer a particular metrical scheme and its constellation with respect to the melodic
structure of the guitar. And of course, once such theoretical viewpoint is elected the
alignment with such paradigm is noted throughout the ensemble’s voices. Hence, hearing
– i.e., understanding its “meaning” – the tune right from the start with such
categorization may allow us the resist the perceptual “pull” exerted by the metrical
directionality (itself an amalgamation of tonal and durational linearities) of the repeating
chorus phrase or the guitar’s later improvisations.
At this point, the reader might again ask why the paradigmatic pattern is suddenly
indicated as “starting” from its half way point. Didn’t much of this discussion suggest a
distinct centrality which would anchor the inherent dynamic “flow” from release to
tension and back? Well, that still remains true, but as we realized earlier during the
discussion of the Cuban examples, it does not follow that other syntactic directionalities
need to be in phase alignment with it. Their effective juxtaposition may in fact be key to
the overall “depth” of a texture. Undoubtedly, this particular transcription also pays
reference to this author’s cognitive “baggage”, which is still informed by a sensitivity to
Western tonal syntax, leading to the indicated diachronic perception of the chorus
phrase’s linearity and its boundaries. A person not steeped in this particular tradition may
in turn not be cognitively influenced by it and formulate another representation.
Yet, while there may remain a real (but nonetheless idiosyncratic!) sense of
ambiguity in terms of the aggregate centrality within such cyclical periodicity, it would
require equal weighing of such directionalities out of context. In actual music, however,
there is always a context and the parameters that can define such context(s) are of course
pretty much limitless. Indeed, our discussion of the syntactic quality inherent in timelines
provides insight into only one of many. The next question is what hierarchic significance
a listener assigns (subconsciously or voluntarily) to a particular context.
Hierarchical levels of syntactic categorization
That musical cognition relies on hierarchical categorization in both physical and abstract
dimensions (though a potentially tenuous distinction in itself) is by now a commonplace.
Less clear and understood is the process how such categorizations are weighted against
each other, or further integrated into meta-categories (as for instance in the case of
meter). Even less clear is the way various musical cultures may go about such weighing,
where such cognitive boundaries may be. One of the tenets of this discussion then is that
for listeners enculturated in the perception of such rhythmic syntax, categorization
according to its implicit rules indeed receives a higher cognitive salience. For such
listeners, this salience may rival the significance that, for instance, categorization
schemas established by tonal syntaxes in Western music hold for a Western listener. I
offer such perception as a “key”100 to musical meaning and a clue towards having a
culturally relevant cognitive experience. In other words, ‘clave’ may provide for an
utterly disambiguating experience, whether from a pre-theoretically “naïve” perspective
or in analytical retrospect.101
Figure 30– “Fantan Ni Mônè”
Track #16
As a good example of why such significance may be justified, I offer a brief
analysis of a song from the wassoulou tradition of Mali (figure 30), as performed by
Oumou Sangare.102 As in the previous example, the introduction of the piece (played on
Perhaps not surprisingly, “clave” in Spanish means “key”!
Agawu accurately remarks that “[t]o consign something to the category ‘pre-theoretical’ is, of
course, to court disaster since, in some views, we can never be ‘outside’ theory”(1994: 95).
The notation shows the music about 3 min. into the tune when the electric guitar switches to the
line indicated.
the bass lute n’goni) sets up a metrical directionality that, if arduously maintained, turns
out to be at odds with the texture further on.103 Figure 31 below shows the metrical
alignment in which the n’goni’s pattern should be heard in order for the directionalities to
match later. While in I.K. Dairo’s song that may have only concerned the point of
cyclical centrality, here the positioning of the fundamental tactus itself is at stake.
Overall in figure, we notice the general correspondence of many patterns with
attack point “4” of the paradigmatic pattern that almost takes on the quality of a
phenomenal downbeat, but attack point “1” is clearly implied as well, even if only by a
subtle dynamic accent as in the electric guitar’s melodic loop. The reason why in this
Figure 31 – N’goni introduction to “Fantan Ni Mônè”
Track #16
piece the second half of the paradigm appears to be the point of overall metrical centrality
once again, may lie in the perception of boundaries in the harmonic motion from
Aminor/E to D, within the overall D pentatonic scale. These boundaries’ points of
Such directionality is created by interpreting the pitch G as marking the tactus beats, a
perception that may be triggered by the additional cadential motion from E to D, a move from relative
dissonance to consonance (6
to 5
scale degree in the key of “G”). The synchronous attack of the second
E of longer duration with the tonal center of G may help to promote its salience as a point of metrical
centrality, despite the fact that this places the all the A’s, which are also salient because of their registral
prominence and dynamic accentuation, in an off-beat position. The timbral stratification and differences in
articulation between these pitches (A and G sound much drier and shorter) are instrumental, too, perhaps
conferring greater weight to categorizations built around the open sounding E and D pitches.
centrality, mostly due to the electric bass’ melodic structure, are out of phase by half a
cycle with the paradigmatic pattern.
To make sense of this piece of music, we benefit highly from a theorizing vantage
point that takes into account the possibility, if not probability, of temporal organization
along an implicit timeline. For once, it seems to be an important factor if just from the
point of vertical congruence among the various voices. But furthermore it also provides
important cues as to the most fundamental notion of musical temporal structure: its most
salient tactus pulsation. Awareness of such syntactic dynamism encoded in the musical
organization is thus a knowledge that we can bring to the interpretative process of
selecting among multiple meanings.
In the process, we may discover that ‘clave’ will shape and realign previously
“chaotic” or meaningless perceptions of other patterns in the music, or even reveal new
syntactic relationships altogether. An example in this case may be the pattern of the metal
scraper, whose “fuzzy” timing between binary and ternary phrasing really “locks in” with
the tactus and generates a tremendously expressive “swing,” if seen from the perspective
Chapter 4
The Syntactical Quality of Timelines with a Cyclical Length of 24 Pulse Units
Examples from Benin
So far, our examples of relationships between a deep-structural syntax represented by an
incarnation of the fundamental paradigmatic pattern and foreground timelines have been
limited to the length of one cycle. There are, however, some cases of phrasing referents
that seem to operate over the span two cycles of our paradigmatic pattern, i.e., at a
periodicity of C=24. Arom (1991) for instance has noted examples in pygmy musical
culture, but there is also a group of such patterns to be found in the music of the Fon and
Gun people, geographically centered on what currently is the Republic of Benin. Perhaps
not surprisingly, these patterns are also utilized by the culturally and geographically
adjacent Ewe, and the descendents of Fon/Gun culture in the Americas, most notably the
arará of Cuba. We will see that a syntax of ‘clave’ may also provide a deep-structural
background for cyclical patterns spanning beyond a fundamental periodic length of C=12
or C=16. In the following two chapters I therefore want to clarify and test the ideas put
forward so far in the context of these, on first sight seemingly unrelated pattern
morphologies, and then apply some points of this discussion to a critique of two previous
investigations into African rhythm. This section has been particularly informed by my
previously mentioned stance as both theorist and performer, and the following analysis
was inspired by a strongly felt sense of syntactic unity that I experienced during
performance and experimentation with this material.
Pattern A in figure 32, known in Benin as the timeline for a rhythm called vinli,
spans two cycles of the basic paradigmatic structure.104 A variation of it, pattern B
(figure 33), which can be heard on a source recording from the region (cf. Track #19;
from Benin: Rhythms and Chants for the Vodun, 1991) is slightly different, but otherwise
retains a fundamental congruence.
Figure 32 – Pattern A (vinli bell)
Tracks #17 (slow) and #18 (fast)
Guitarist Lionel Loueke [personal communication, 2002], a native of Benin who, in an
interesting twist of transcultural mediation, has developed an original way of incorporating these patterns
into the context of jazz performance, confirmed the usage of pattern A as typical for the type of traditional
music of Benin called vinli or zinli (also pronounced as zinlin). This rhythmic pattern structure is known
as the vinli bell and usually played very slowly on metal claves (similar to Ewe atokes) by either two
people, or by one person on a single double bell. It is conceived of in relation to the implicit standard 12/8
bell cycle, which is kept in mind by performers even though the vinli bell itself spans two periods of this
12-pulse cycle. Vinli is from the region around the small city of Abomey (Fon people) in the middle of
Benin, and, unlike other types of music from the coast or the western parts that have influences from
neighboring Togo, Ghana, or Nigeria, is considered being more typical to Benin.
Mr. Loueke recalls playing vinli as a child, and in fact it continues to be played very much by
children in the context of a percussion ensemble called kaleta during the New Year’s season between
November and January. The kaleta ensemble is headed by an experienced older drummer, who plays the
lead drum rhythms, and the children sing and play the non-varying supporting drum parts. There are also
professional ensembles that play the vinli rhythm in a different, ceremonial context during, for example,
overnight wakes honoring a deceased person. Mr. Loueke suggests a connection between the vinli rhythms
and ritual music of the vodun, however it seems that the tempo at which pattern A is played in vinli is
slower than that of similar patterns in the music for the vodun.
While the “standard pattern” timeline indicated here is not necessarily sounded, I argue
that it underlies these patterns’ morphology and provides for a particular “best fit” from a
syntactical perspective.
Note that over the two cycles of the standard 12/8 bell pattern A has three
occurrences of a motif consisting three attacks (see brackets).105 On the recording
mentioned above, asan rattles perform an accompanying pulse of either 6 beats (see
figure 33), or at the twice-as-fast 12:24 ratio. This practice is common for this repertoire
and such pulses may also be played by additional bells, as we soon shall encounter. The
“inexorable flow of dotted quarter-note beats” (Locke 1982: 224), which I have referred
to as tactus throughout this chapter, thus appears to be replaced by a very salient
Figure 33 – Pattern B (“Procession”)
Track #19
Arom (1991: 239) calls such a pattern structure “uniform,” “based on the repetition of a single
configuration or cell, whose position with respect to the pulsation [i.e., tactus] varies at each recurrence
(brackets mine).
“competing” pulse that chunks the pattern in a way as to make it seem of binary metric
orientation.106 On the recording we can hear the supporting drums play a melo-rhythmic
motif that seems to suggest a dotted-quarter note tactus of an 8:24 ratio, but which
remains ambiguous as to whether these accents actually represent an on- or off-beat
accentuation. Hence, a definite metric schema initially remains highly elusive in this case.
The realization of an overall metric orientation thus depends on the discovery of a very
different kind of particular and subtle “cohesion” among all the patterns and their
combined linearities.
Syntactic resonance between Pattern A and the paradigm
Looking at the relationship between pattern A and the timeline, we note a very particular
counterpoint that “individualizes” each of the three motifs’ fit with a section of the bell.
While the first unit is marked by total synchronicity with some of the timeline’s structural
attacks (i.e., attack points 1, 2 and 4), the next one displays an entirely interlocking
alignment and the final unit has elements of both. Furthermore, over a period of two 12/8
timeline cycles, the total of nine attacks in pattern A sample nine out of a possible 12
distinct metric positions of the paradigmatic cycle without any redundancy. That means
that the directionality generated by these three seemingly symmetric pattern-units as a
whole not only maintains a high syntactic consistency with respect to each of the two
constituent underlying paradigmatic cycles, but additionally manages to establish a
Locke (1982: 231) has pointed to similar example found in the master drum’s play in Ewe
Atsiagbekor, and remarks that such patterns nevertheless “are felt primarily in relation to the fundamental
ternary quadruple meter, not the countermeters. Hence, in each of its three appearances the short motive
lies in a different relation to the main beats.” In our discussion, however, we need to remember that it is the
audible timeline itself that alludes to such countermeter, making conceptualization with reference to such
implicit 4-beat meter all the more difficult.
unique syntactic flow over the entire C=24 period. We get the impression that the area of
tension-release (between bell phrase attack points 2-13) has been widened to include an
extended “middle” section, hence retaining the essential asymmetrical oscillation of
“active” and “rest” sections and fundamental syntactic dynamism (cf. Track #20 –
Pattern A and timeline).
Incidentally, if we examine pattern A against a slower moving 4-beat tactus (four
dotted half-note beats per C=24 period), the same 3:1 “active-rest” ratio suggested earlier
for the C=12 cycle length seems to recur. Each such tactus beat is framed by pattern A’s
attacks analogous to the way the dotted quarter tactus of the C=12 span is framed by the
paradigmatic pattern in figure16: commetric coincidence on beats one and four,
contrametric syncopation at beats two and three (cf. Track #21 – Pattern A and
tactus). While I don’t think the tension implicit in this ratio may be quite as cognitively
salient at this hierarchically lower tactus level, it may be possible that the syntactic
tension inherent in a syntax of clave may have a sort of general “fractal” quality. In other
words, at various levels of hierarchic “magnification” the same essential metric
dynamisms are present. The attacks of pattern A in figure 32 marked underneath with a
diagonal line indicate the positions of the implicit paradigmatic pattern’s attacks
comprising such further potential deep-structural relationship.107
So far we have noticed about pattern A’s morphology that, on its surface, it seems
to the ear to project a strong binary meter (i.e., 3x’s 4/4 = 2x’s 12/8) in no simple ratio to
neither a C=12 nor a C=24 periodicity, let alone implying a 4-beat tactus at odds with
The fact that there are more events than the original 5 attacks may not be of relevance for the
invocation of such cognitive schema, just like at the C=12 level the paradigm can be invoked by a variety
of surface structures. What matters is whether grouping boundaries are perceived that match the paradigm’s
inter-attack intervals in relation to the fundamental 4-pulse tactus. And as we have seen earlier, these
interval values are not absolute but rather variable according to context.
such linearity. Yet, even here a compelling resonance with the overall syntactic
directionality established by the paradigmatic pattern may be said to exist. That is not to
say that such resonance may be easily apparent in the actual musical texture, in fact it
may only exist in a listener’s mental abstraction.108 However, it nevertheless can form an
important aspect as far as a performer’s cognitive clock is concerned.
One indication for this is my impression that accurate timing of this pattern in
performance seems to be easier if undertaken from within the perspective of two such
paradigmatic cycles.109 Timing against either of the two implicit isochronic tactus
schemas alone (the faster dotted-eighth note tactus or its half-speed variant) results in less
stability – perhaps because in this case the cognitive integration of pattern and tactus
results in a syntactically less pervasive conceptualization, and thus less a steady motor
pattern. Instead, it would have to rely more on additive, horizontal timing procedures
based on the perception of a common density referent pulse (i.e., the eighth-note pulse)
for accuracy.
We can hear in Track # 20 the way the outwardly “autonomous” linearity of
the repeating q q . q . unit is nonetheless tied into a specific dynamic relationship of both
synchrony and figure-ground oscillation with the timeline, similar to the ones we already
It may however be established indirectly, as for instance in audio example #5 (figure 26,
Pattern B), where operational “proof” of its background structure might be inferred from details in the
cyclical pattern structure of the supporting drum parts suggesting a 4-beat tactus, or master-drum motivic
See Pressing (1997) for more research on cognitive pattern integration and timing processes
with respect to asymmetric West African timelines. He investigates two cognitive models of bi-manual
polyrhythmic performance: 1) isochronic strokes in one hand (“figure”) are timed against an asymmetrical
internal clock, a “ timeline-ground (TLG)”; 2) the timeline is timed from an isochronic “pulse-ground
(PG)” (p.196). The current context is slightly different in that both patterns (pattern A and the timeline) are
asymmetrical, and we interpret pattern A as a “figure” to be integrated with either a PG or a paradigmatic
TLG cognitive model.
noticed in our previous examples like the comparsa bell or the cáscara. As far as this
contrapuntal dynamism is concerned, it appears that it is marked first by a motive flow
away from initial synchronization with the durational contour of the timeline towards off-
beat juxtaposition, and then back to synchrony.110 Figures 34 and 35 below show the
commetric111 sections that in this case frame this flow, like “bookends” of sorts:
Figure 34 – Pattern A commetric section A Figure 35 – Pattern A commetric section B
Note that attack points 1 and 13 of the timeline are the only locations where both
patterns also synchronize with another metrical scheme, the tactus at the 8:24 ratio, i.e.,
the four fundamental beats implied by each of the two 12/8 bell patterns. This
convergence of several patterns at one point, or within a small region of relative
synchrony among otherwise diverging directional momentums, may be crucial to the
The alternation between areas of synchronization and syncopation of attacks among hands, as
part of an asymmetrical timeline’s cognitive integration (and physical externalization) with an isochronic
pulse, has been found to generate a “hand-independent but consistent pattern of tempo variation in the half
of the cycle characterized by synchrony of hands. The other syncopated half showed tempo stability
(Pressing 1997: 218).” Such alternation may thus be an important indicator of overall syntactic coherence,
which in turn facilitates expressive timing control as well as global tempo stability. This particular
dynamism may be one of the aspects that I have elsewhere described as a “best-fit” and fluid motor
coordination between two schemas, regardless whether they are actually simultaneously externalized or
partly maintained as a mental template.
In this context, I propose commetric to mean alignment of pattern A not with the tactus but
with the timeline’s attacks, to which I attribute a metrical salience of its own. From this perspective, pattern
A’s synchronous attacks with the timeline (i.e., at beats 1,2,4, 10 and 13) carry the lowest metrical tension
and are most conforming to the timeline’s linearity. There thus at least two levels of metrical reference
active, tactus and timeline, each with its own commetric or contrametric quality. This can be understood as
an expanded notion of Kolinski’s (1973) definition .
perception of a unified metric centrality (“downbeat”), against which syntactic motion
(i.e., tension and release) is experienced.112
Earlier, I already suggested the significance of this type of syntactic motion
during the discussion of the implicit metrical tension which results from a perceptual
integration of the basic paradigmatic pattern with the isochronic tactus (see p.59), and
subsumed it under the notion of the “syntactic arc.” Regarding figure 32, it seems then
that a deep-structural syntax of ‘clave’ furthermore may be defined by multiple,
simultaneous “arcs” at various levels of schema-relationships, whereby patterns
generated by physical events in the auditory experience are abstracted “silently” by
cognitive frameworks supplied by a listener/performer.
Comparison of transpositions of Pattern A with the paradigm - Resultant Metric
Let us now look at the transpositions (n) to which pattern A (n=0) can be subjected
before it reaches identical alignment with the timeline. Here we notice that actually only
seven, not twenty-three, forward transpositions are required. Since the length of each of
the three subunits is only C=8, and all three units are identical (see brackets in fig. 32), an
eighth transposition would therefore present the same alignment as n=0. 113But that
means that perception of pattern A as forming a C=24 periodic cycle must depend on
some other cue of metric centrality. I suggest that this centrality is achieved by further
In Western musical traditions such centrality is usually delineated by converging
directionalities (i.e., cadences) of several syntactic conventions at the same time, for instance through
harmonic cadence, melodic contour, or dynamic accentuation
To maintain the visual integrity of the shifting ‘quarter note-dotted quarter note-dotted quarter
note’ subunit in fig, 35-41, I have made an exception to my practice of avoiding ties altogether. That way,
dotted quarter notes that straddle a barline are retained as a integer durational value.
integration of pattern A’s phrase structure with the implicit metrical dynamism of a C=12
timeline paradigm, in this case shown as the “standard pattern.”
By comparing the resultant metric resonance of at least the three parameters that
have concerned us so far (namely pattern “figure,” paradigm “background,” and tactus
“ground”) for these transpositions, we may see why the original configuration of figure
32 creates a particularly “well-formed” association among these linearities. The
perception of such resonance originates in a listener’s experience of the particular
interplay of the directional forces exerted by these parameters in the music.
Figure 36 – 1
transposition of pattern A
Figure 37 – 2nd transposition of pattern A
Figure 38 –3
transposition of pattern A
Figure 39 – 4
transposition of pattern A
Figure 40 – 5
transposition of pattern A
Figure 41 – 6
transposition of pattern A
Figure 42 – 7
transposition of pattern A
While space constraints do not permit me to discuss each transposition in
individual detail, a few general observations shall be made nevertheless. First of all, each
transposition creates a specific constellation among these three parameters, as far as the
sequence of coincidences among their attack points are concerned.114 In other words,
even though the relationship between two of the parameters, C=12 timeline and tactus,
remains consistent, for all eight possible alignments (n=0-7) a unique pattern integration
results. Furthermore, we note that the amount of simultaneous attacks among all three
patterns varies from one to two such instances per transposition (see circled notes).
N=4 is of course identical to the original constellation notated with the second half first, but
since the alignment between the parameters is otherwise unchanged we will not consider this transposition
for this discussion.
But what, we may ask, distinguishes pattern A’s integration with the timeline
paradigm and tactus shown in figure 32 from above transpositions? What in the
constellation of all these parameters leads to a perception of centrality, and this point of
centrality in particular. I propose that this quality may be the effect of a particular
consonance among these metrical schemata that is perceived in the music, and which
results in the kind of syntactically “well-formed” integration already discussed earlier
with respect to Cuban clave.
To elucidate this interplay of what I suggest are separate lower-level metrical
directionalities resulting from pattern A’s interaction with the timeline and with the
tactus, I have attempted a rudimentary schematic encoding of these individual metric
schemata in form of a simple cognitive dichotomy of on-beat and off-beat relationships
between the involved parameters. By indulging for a moment in this admittedly
idiosyncratic nomenclature, I hope, in the spirit of Swain’s remark below, to nevertheless
give some insight into the subtle syntactic dynamism that may be the result of this
differentiated sequence of alignments:
Discovering or, indeed creating a musical syntax guarantees nothing in the way of music’s value
or even perceptibility – that requires an extra criterion – but it does guarantee the possibility of
comprehension, at least. (Swain 1995: 287)
The interplay of Pulse Ground Alignment and Timeline Ground Alignment
An interesting regularity exhibited by pattern A lies in its relationship to the isochronic
tactus beats, specifically in the order of synchronous or coinciding (C) and syncopated
attacks, the latter of which can further be distinguished into anticipations (A)(an attack
preceding a tactus beat by one unit of the C=24 matrix) and delays (D) (an attack
immediately following on the next unit).
Figure 43 – Pattern A (PG and TLG alignment)
Pattern A’s attacks will group into a syncopated sequence consisting of CCC AAA and
DDD units. Each of the eight possible positions are distinguished by a different starting
point from within this sequence, in the original position n=0 (i.e., figure 32) that would
be CAAADDDCC. We can call this a pulse ground alignment code (PG), basically a
visual representation of a metrical tension state.
A similar representation can be construed for the relationship between pattern A
and the timeline, here utilizing only the two categorical distinctions C for coinciding and
S for syncopated notes. This timeline ground alignment code (TLG) does not
demonstrate the internal regularity of the pulse ground model; other than that there are
always a total of five coinciding and four syncopated attacks, their sequential order
changes in each case. Here we note a CCCSSSCSC sequence.
While it is difficult, and perhaps only possible through extensive neuro-cognitive
research,115 to quantify the cognitive “value” exhibited by each of these two metrical
abstractions, I suggest that pattern A’s “well-formedness” is due to the particular
interplay of these alignment schemata. If for example we look at the correspondence
between these metrical directionalities in terms of where an attack of pattern A is
commetric in relation to both PG and TLG, we notice that n=0, n =3, n=4, and
transposition n=7 have two of such points – all others only one (which I have circled in
each transposition). Yet, only n=0 and n =3 place this coincidence on the first and last
tactus beats of the C=24 cycle, thereby creating the syntactic arc shown in figure 32.
In the case of transpositions n=4 and n=7 we presuppose that further syntactic
categorization taking place simultaneously would establish the grouping perception
suggested in the notation, overriding the particular syntactic dynamism under discussion.
However n=3, I believe, does not have the directional stability in either PG or TLG
alignment as does n=0, since it lacks the initial run of three C attacks with the timeline,
and, on its crucial second attack, fails to establish the syntactically paradigmatic A
relationship with the pulse ground.
In the field of cognitive-linguistics there have been recent approaches towards such
quantification by constructing a “cost” for different types of syntactic integration through the measuring of
event-related brain potentials, and a correspondence to musical cognitive processes has been suggested (cf.
Patel, Gibson, Ratner, Besson & Holcomb, 1998).
Aspects of syntactic stability
If the syntactic dynamism I have sought to describe here is indeed the result of the
integrated perception of pattern A within two metrical directionalities established in
relation to the timeline and tactus, such dynamism moreover may benefit from a
juxtaposition of relatively stable areas of consecutive coincident or syncopated attacks,
rather than from a constantly alternating combination of single events.
The fact that perceptual grouping of repeating cyclical patterns generally seems to
favor the run principle, i.e., the first event in the longest running group of events of
identical value being perceived as the initial unit, has already been remarked earlier in
respect to the paradigmatic pattern itself (also cf. Pressing 1983a :49). Perhaps in this
case long runs of commetric attacks in either pulse- or timeline ground alignment are
preferred, and their centrality for an emergent overall syntactic property strengthened
when the commetricities of pulse ground and timeline ground alignments combine. The
unique side to this in our discussion is that here the notion of ‘commetric’ may entail both
metric abstraction in relation to an isochronic pulse (i.e. the tactus), and additionally in
relation to an asymmetric pattern schema (TLG). Incidentally, n=3 seems a weaker
pattern than n=0 from this perspective.
Furthermore, we note that n=6 also has a long run of commetric attacks, albeit in
its pulse ground alignment. Even though the PG and TLG code combination of its first
two attacks resemble n=3, by maintaining a clear directionality in the PG alignment
(actually in a sort of “prime form” sequence: CCCAAADDD) this transposition of
pattern A (n=0) manages to create a different, but syntactically similarly pervasive
alignment. Here the “run” directionality is established with the pulse ground, while the
timeline ground provides a consistent syncopated relationship (note that this is the exact
opposite to n=0’s first four attacks).
Could perhaps this transposition’s directionality thus serve as a background paradigm in
its own right?
Pattern A’s paradigmatic functionality
It does in fact appear so with respect to pattern C in figure 44 , which is part of the same
Fon and Gun repertoire as is pattern A. Incidentally, this pattern is also used in the
kadodo section of the Ewe dance adzogbo,116 a genre said to have been adopted from
Fon/Gun traditions by the former.117 And Faith Conant (1988: 112) has also documented
its usage in the region of Lomé, Togo, where it accompanies a section of adjogbo called
Schmalenberger (1998: 39) remarks on the perceptual multiplicity of this bell pattern, yet
confirms that the dancers mark the tactus’ beats indicated above, which are also played by the axatse.
Pantaleoni (1972) also confirms the dotted quarter tactus pulse that makes “the dancers go out of phase
with [the bell] . . . and return to being in phase with the bell at the end of the third playing of the bell
pattern. Since another bell pattern twelve units long was also used for the same dancing . . . it seems clear
that the true length of the bell pattern for kadodo is twenty-four units.”
In the case of adzogbo, the attacks of pattern C marked here by the quarter notes with a
triangular note head are usually executed as a muted stroke, i.e., the bell is muffled by pressing the stick
into it, or pressing its cone against the body of the performer.
Figure 44 - Pattern C (“Polyrhythm”)
Track #22
Since the 2xC=12 timeline is usually not sounded, pattern C (played on an iron
bell called gan) by itself initially also seems to, like pattern A, suggest a binary metrical
orientation, which on the recording is reinforced by a steady half note pulse on another
bell called gogbe.
Figure 45 – Pattern C in alternate metrical perception
If figure 44 above is a correct interpretation of such structural relationship, it would mean
that the single attacks of dotted-quarter duration in pattern A’s n=6 transposition would
“get doubled-up,” generating118 a new surface pattern but maintaining the same essential
grouping structure of what Pressing (1983a) has termed “perceived aggregate subunits.”
Pressing (1983a: 41) has called this transformation process “element fission”
However, the structural relationship between patterns A and C could also be
imagined in another manner shown in figure 46. In this scenario, the attack points of
pattern A’s original position (n=0) would be partly the result of a figure-ground reversal
with pattern C’s attacks consisting of its complementary attacks, and including every
other of the gogbe’s attacks (circled). This interpretation exhibits a particularly fluid
kinesthetic flow, if performed bimanually.119
Figure 46 – Pattern C and pattern A alternative alignment
With just the aural data at hand it is of course futile to attempt definite conclusions as to
the nature of any specific derivative, transformational or other generative processes that
may link these patterns, let alone their particular cultural-historical evolution.
Nonetheless, at the level of a C=24 periodicity we have found a number of syntactical
relationships that seem to be based on a deep-structural relationship with the
paradigmatic pattern’s fundamental properties.
Kinesthetic flow alone is of course not proof of a corresponding structural organization or
cognitive integration by a performer/listener, but should also not be discarded as coincidental. As already
mentioned earlier in this text, research suggests a particular balance between synchronous and syncopated
areas of attacks may be instrumental in stabilizing global timing, while allowing for expressive nuance
within such timing (cf. Magill and Pressing 1997).
Discovering a set of two related metrical perspectives
To conclude our investigation into the syntactical affinities and structural correspondence
sparked by pattern A’s morphology, I would like to finally turn to yet another, culturally
related pattern (figure 47), that functions as somewhat of an analytical “keystone” to the
theorization of pattern A’s and pattern C’s syntactical connection. It is played on the ogán
as part of the Afro-Cuban arará120 rhythm called afrekete, and structurally very similar to
pattern C. But we again also notice the persistent congruence with pattern A’s grouping
structure. The connection to the implicit C=12 timeline is revealed by the vocal parts, for
instance through the timing of the recurring “A-so-ma-do”. Its stressed syllables “ma” and
do” suggest alignment with the paradigmatic bell as indicated.
Figure 47 – Arará Afrekete (“Alua”)
Track #23
A closer look reveals that the ogán pattern is indeed identical to pattern C, save
for the final two attacks.121 If we extend its phrase all the way through by adding these
Arará was the name given in Cuba to descendents of the Ewe, Fon and Gun from the region of
Dahomey, present-day Benin (cf. Rodriguez 1994: 97 and Diaz, Garibaldi and Spiro 1999: 32).
Incidentally, in the audio example from Benin these two attacks serve as a starting cue for the
entire ensemble, most likely played on the side of the masterdrum. These two final commetric attacks (in
relation to the pulse ground, i.e., tactus) of pattern A thus serve the dual purpose of locking in the
fundamental pulse for the whole ensemble (that otherwise remains highly elusive in terms of actual
two missing attacks (circled), a third possible alignment (figure 48) between patterns A
and C emerges:
Figure 48 – Pattern C and pattern A paradigmatic alignment
In this interpretation we notice that pattern C seems to be structurally elaborated on
pattern A’s aggregate subunits, by the process of element fission (the dotted-quarter units
are broken up). But this elaboration may perhaps also be seen as a durational anacrusis to
each following attack, reminiscent of the cáscara’s accent pattern). Listening closely to
the afrekete audio example, we can in fact hear a slight dynamic emphasis and timbral
change marking each long-duration (the quarter notes or dotted-quarter notes in figure
47) that may indicate such conceptualization.122
More significantly though, patterns A and C appear to be “two sides of the same
coin,” each side being marked by one of the symmetrical halves of the implicit C=24
timeline. Position X marks the starting point of pattern A, position Y the starting point of
pattern C. Also, in relation to the C=24 period, both patterns continue to support to the
notion of “~ ½ filling ratios” or “theory of bisection” that has been suggested as an
ubiquitous feature of African timelines’ morphosyntax of C=12 or C=16 length.
externalization in the music, save for a supporting drum pattern of muffled and open strokes that sounds
reminiscent of typical Ewe kidi vocabulary), but already foreshadow its linear organization.
The afrekete pattern and its dynamic accentuation in fig. 40 corresponds to pattern #2 given by
Diaz, Garibaldi and Spiro (1999: 32) in their comprehensive table of arara bell patterns, which, to no
surprise, also includes under #3 our pattern C of this discussion.
Interestingly though, at the point where pattern A begins (the point of metrical
centrality established by its own horizontal linearity and its commetric relationship to
both the C=12 timeline ground, the 4-beat tactus, and additionally the gogbe’s and
rattles’ patterns, which all coincide at that point), it is overlapped by pattern C’s
perceptual grouping unit (see boxed notes in figure 49), and vice-verse!
Figure 49 – Pattern C and A paradigmatic integration from two points of centrality
The C=24 cyclic period in this example can hence be established from two points of
potential metrical centrality, X and Y, depending on which one serves as the cognitive
reference point.
Patterns A and C share the same quality of syntactic integration with the standard
bell pattern (i.e., the paradigmatic pattern) in terms of the C=12 periodicity that
constitutes each half of C=24 (X-Y; Y-X), and at any point within their entire C=24
phrase length, patterns A and C are “in clave,” so to speak, with the timeline, maintaining
a commetric relationship to it. They are also in vertical durational consonance123 with one
another. In fact, pattern C’s structure is a slight elaboration of pattern A’s scheme of
2+3+3+2+3+3+2+3+3. However, in their respective performance contexts, perception of
their individual linear “gestalt” (experienced as a cycle of three of the subunits marked in
See Rahn (1996: 76) on the notion of consonance/dissonance in the rhythmic domain.
figure 49) puts them at different positions within the larger C=24 continuum, and they are
experienced as the discrete rhythms that we first encountered in figures 32 and 44.
For example, perception of pattern A places position X at the beginning of the
overall syntactic flow within the C=24 period, and pattern C – even though that has its
own “autonomous” point of centrality at Y – in the second half of such arc.124 I suggest
that in this case, X’s centrality is solidified because pattern A avoids coinciding with
either PG or TLG at its half-cycle point Y, which marks the bisectional division of the
C=24 periodicity.
In a sense, the cognitive experience of pattern C as starting from halfway through
the C=24 cycle would thus be analogous to the situation which I already mentioned
earlier regarding the C=12 or C=16 cycle, where such experience is, particularly in the
vernacular of the Afro-Cuban tradition, described as hearing the pattern in “2-3” position.
While I do not want to suggest that in this present case this analogy can be as easily
recognized right away, if at all, I believe that it nevertheless presents an analytically
meaningful explanation of pattern C’s morphology.
From this perspective, I therefore suggest that a) pattern A functions as an
additional, higher-level paradigmatic structure or ‘clave’, in relation to which pattern C is
a particular foreground elaboration, and b) that the gestalt of pattern C nonetheless retains
a perceptual independence as to its own metrical centricity, based on its alignment with
the C=12 timeline at point Y. Pattern C’s gestalt can thus be described as derived from,
and syntactically linked to pattern A on one level, while equally on another level
syntactically linked to the C=12 timeline.