n œ œ œ œ œœ ..

∑ c œœœœ

œ.
∑ c n œœ œœ œœ œœ œ .
œœ œœ # œœ œœ œœ œœ ˙
c ˙

#œ œ œ œ œ.
∑ c
f
œ œ œ.
∑ c œ œ
f
∑ c #œ œ œ œ œ.
a2
A Guide to Schillinger’s
Œ
n˙
c
˙
Theory of Rhythm
Œ b ˙˙ c ˙˙
f
Œ ˙˙ c ˙˙ Second Edition
f
Πc
f˙ œ œ
n˙ ˙
œ œ c

œ œ ˙˙ c ˙˙

Frans Absil
œ #œ œ n˙. c ˙
˙. ˙
2015
œ #œ œ ˙. c ˙
b ˙˙ .. ˙˙
˙. ˙
c

b ˙˙ .. c ˙˙

∑ c

∑ c

c
b˙. œ œ
b 13/ # 11 b 13/ # 11
œ #œ œ ˙.
E7 E7

c |
.
œ œ œ œ c n n œœ œœ n œœ œœ œœ .
œ œ œ œ œ œ œ œ bœ nœ œ
∑ c

Œ b˙ c Œ bœ

y y y y y y Y
c æ

Œ ˙ c œ œ
æ æ æ
f c ˙
∑
Suspended cymbal
fæ
n ˙˙
œ n ˙˙ c nœ œ œ œ œ.
Colophon
This book is a guide to Book I The Theory of Rhythm from The Schillinger
System of Musical Composition. It follows the chapters from the original book.
Rhythm creation techniques will be discussed in detail and musical examples
will be presented.

This document was created using the public domain LATEX computer typeset-
ting program. Diagrams were created using the LATEX picture environment.
Musical examples were created using the MakeMusic Finale 2014 music nota-
tion software. Score examples were imported into the document as PDF file
format figures, using the graphicx package. The navigation links (printed
in blue) in the PDF file (use a PDF reader) were created using the hyperref
package from the LATEX distribution.

Full version (2nd Ed, Rev 2.2):

This is the full version of the e-book.

Note to the document owner:

This copy of the full edition has been issued on October 7, 2020 to
Seth Dempsey ([email protected]) for personal use only. The owner
agrees to respect all copyright restrictions.

A Guide to Schillinger’s Theory of Rhythm, Second Edition, Frans Absil

2006-2016
c F.G.J. Absil, the Netherlands.1
URL Website: http://www.fransabsil.nl

All rights reserved. No part of this publication may be reproduced or transmitted in any form or by
any means, electronic or mechanical, including photocopying, recording, or any information storage
and retrieval system, without prior permission in writing from the Author.

ii Copy issued to Seth Dempsey ([email protected])

Contents

List of Figures vi

Preface ix

1 Introduction 1
1.1 Document structure overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Notation system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.1 Graphing music . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.2 Integer number representation of ticking metronomes . . . . . . . . . . 6
1.3.3 Forms of periodicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Interference of periodicities 9
2.1 Binary synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1 Uniform binary synchronization . . . . . . . . . . . . . . . . . . . . . . 9
2.1.2 Non-uniform binary synchronization . . . . . . . . . . . . . . . . . . . 11
2.1.3 Overview of non-uniform binary synchronization and fractioning re-
sultants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2 Grouping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.1 Grouping by the common product . . . . . . . . . . . . . . . . . . . . . 18
2.2.2 Superimposition of a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.3 Superimposition of b . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.4 Alien measure grouping . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.3 Characteristics of the resultant . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.3.1 The natural nucleus of a score . . . . . . . . . . . . . . . . . . . . . . . . 22

3 The techniques of grouping 25

4 The techniques of fractioning 29

4.1 The process of fractioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.1.1 Fractioning group: 3 and 2 . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.1.2 Fractioning group: 4 and 3 . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.1.3 Fractioning group: 5 and 2 . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.1.4 Fractioning group: 5 and 3 . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.1.5 Fractioning group: 5 and 4 . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.1.6 Fractioning group: 6 and 5 . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.1.7 Fractioning group: 7 and 3 . . . . . . . . . . . . . . . . . . . . . . . . . 33

iii
CONTENTS

4.1.8 Fractioning group: 7 and 4 . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.1.9 Fractioning group: 7 and 5 . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.1.10 Fractioning group: 7 and 6 . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2 Grouping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5 Composition of groups by pairs 39

5.1 Balancing adjacent groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.2 Expanding adjacent groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.3 Contracting adjacent groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

6 Utilization of three or more generators 45

6.1 The technique of synchronization of three generators . . . . . . . . . . . . . . 45
6.1.1 Interference group: 5, 3 and 2 . . . . . . . . . . . . . . . . . . . . . . . . 45
6.1.2 Interference group: 7, 4 and 3 . . . . . . . . . . . . . . . . . . . . . . . . 46
6.2 Grouping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

7 Resultants applied to instrumental forms 51

7.1 Instrumental rhythm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
7.2 Applying the principles of interference to harmony . . . . . . . . . . . . . . . 53
7.3 Combination of instrumental form techniques . . . . . . . . . . . . . . . . . . 56
7.3.1 Extension 1: Doubling on multiple staves . . . . . . . . . . . . . . . . . 56
7.3.2 Extension 2: Interference between time and instrumental groups . . . 58
7.3.3 Extension 3: The combination of techniques . . . . . . . . . . . . . . . 63

8 Coordination of time structures 67

8.1 Distribution of a duration group through instrumental and attack groups . . 67
8.2 Synchronization of an attack group with a duration group . . . . . . . . . . . 69
8.3 Distribution through the final duration group . . . . . . . . . . . . . . . . . . . 70
8.4 Synchronization of an instrumental with an attack group . . . . . . . . . . . . 70

9 Homogeneous simultaneity and continuity 79

9.1 Variation through circular and general permutations . . . . . . . . . . . . . . . 79
9.1.1 The set of permutations . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
9.1.2 Permutations in continuity and simultaneity . . . . . . . . . . . . . . . 83
9.2 Permutations applied to musical parameters . . . . . . . . . . . . . . . . . . . 84
9.3 Application to attack-duration groups . . . . . . . . . . . . . . . . . . . . . . . 86

10 Generalization of variation techniques 95

10.1 Higher order permutations of a two-element group . . . . . . . . . . . . . . . 95
10.2 Higher order permutations of a three-element group . . . . . . . . . . . . . . . 98
10.2.1 Combinations by two elements . . . . . . . . . . . . . . . . . . . . . . . 98
10.2.2 Combinations by three elements . . . . . . . . . . . . . . . . . . . . . . 102

11 Composition of homogeneous rhythmic continuity 107

11.1 Splitting through the simplest divisor . . . . . . . . . . . . . . . . . . . . . . . 107
11.2 Splitting through individual measures . . . . . . . . . . . . . . . . . . . . . . . 109
11.3 Splitting through the individual attacks . . . . . . . . . . . . . . . . . . . . . . 112

iv Copy issued to Seth Dempsey ([email protected])

 CONTENTS

12 Distributive powers 113

12.1 Continuity of harmonic contrasts . . . . . . . . . . . . . . . . . . . . . . . . . . 113
12.2 Rhythmic counterthemes by distributive powers . . . . . . . . . . . . . . . . . 114
12.2.1 The square of an attack-duration group . . . . . . . . . . . . . . . . . . 115
12.2.2 The cube of an attack-duration group . . . . . . . . . . . . . . . . . . . 127
12.2.3 The generalization of all powers . . . . . . . . . . . . . . . . . . . . . . 134

13 Evolution of rhythm styles 135

13.1 Evolution on the fractional level . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
13.2 Evolution on the factorial level . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

14 Rhythms of variable velocities 145

14.1 Acceleration series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
14.2 Application of acceleration series . . . . . . . . . . . . . . . . . . . . . . . . . . 149
14.2.1 Acceleration in uniform groups . . . . . . . . . . . . . . . . . . . . . . . 149
14.2.2 Acceleration in non-uniform groups . . . . . . . . . . . . . . . . . . . . 150
14.2.3 Rubato . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
14.2.4 Fermata (hold) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

15 Conclusion 153

Bibliography 155

Index 156

Copy issued to Seth Dempsey ([email protected]) v

CONTENTS

vi Copy issued to Seth Dempsey ([email protected])

List of Figures

1.1 Overview of the book structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Overview of rhythm terminology in diagram . . . . . . . . . . . . . . . . . . . 5
1.3 Plotting a rhythm using the Schillinger System graphing method . . . . . . . 5
1.4 Plotting the time instants of a ticking metronome along a time axis . . . . . . 6
1.5 Two metronomes ticking at different time intervals . . . . . . . . . . . . . . . . 7

2.1 Uniform binary synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2 Non-uniform binary synchronization . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Grouping and superimposition . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4 The natural nucleus of a musical score . . . . . . . . . . . . . . . . . . . . . . . 23

3.1 Grouping of non-uniform interference patterns . . . . . . . . . . . . . . . . . . 26

3.2 Grouping of non-uniform interference patterns (cont’d) . . . . . . . . . . . . . 27

4.1 Fractioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2 Fractioning (cont’d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.3 Grouping of fractioning patterns . . . . . . . . . . . . . . . . . . . . . . . . . . 38

5.1 Balancing, expanding and contracting a pair of adjacent groups. . . . . . . . . 44

6.1 Synchronization of three generators . . . . . . . . . . . . . . . . . . . . . . . . 47

6.2 Grouping of the three-generator interference group 5 ÷ 3 ÷ 2 . . . . . . . . . . 49
6.3 Grouping of the three-generator interference group 7 ÷ 4 ÷ 3 . . . . . . . . . . 50

7.1 The dependence of the instrumental rhythm on the number of elements in the
pitch and attack series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
7.2 Instrumental rhythm example . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7.3 Interference applied to harmony example . . . . . . . . . . . . . . . . . . . . . 57
7.4 Interference to harmony: doubling on multiple staves, Case 1 . . . . . . . . . 59
7.5 Interference to harmony: doubling on multiple staves, Case 2 . . . . . . . . . 60
7.6 Interference between time and instrumental groups . . . . . . . . . . . . . . . 62
7.7 Instrumental form created through combination of techniques . . . . . . . . . 64

8.1 The process of coordination of time structures . . . . . . . . . . . . . . . . . . 68

8.2 Synchronize an attack group with a duration series . . . . . . . . . . . . . . . 71
8.3 Pitch attack group distribution over multiple instruments in a score . . . . . . 73
8.4 Synchronization of a three-layer instrumental group with an attack and group 74

vii
LIST OF FIGURES

8.5 Synchronization of a three-layer instrumental group with an attack group

(cont’d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
8.6 Synchronization of a four-layer instrument with an attack group . . . . . . . . 78

9.1 Variation through permutations . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

9.2 Variation through permutation of attack-duration groups, two staves . . . . . 88
9.3 Variation through permutation of attack-duration groups, three staves . . . . 89
9.4 Variation through permutation of attack-duration groups, four staves . . . . . 92

10.1 Higher order permutations of a two-element group . . . . . . . . . . . . . . . 96

10.2 Higher order permutations of a two-element group: example in rhythm notation 97
10.3 Higher order permutations of a three-element group: combinations by two
elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
10.4 Higher order permutations of a three-element group: combinations by three
elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

11.1 Composition of a homogeneous rhythmic continuity by splitting . . . . . . . 108

11.2 Homogeneous rhythmic continuity by splitting. Example 1 . . . . . . . . . . . 110
11.3 Homogeneous rhythmic continuity by splitting. Example 2 . . . . . . . . . . . 111

12.1 Overview of rhythmical power series . . . . . . . . . . . . . . . . . . . . . . . 114

12.2 Distributive powers. The square of a binomial . . . . . . . . . . . . . . . . . . 117
12.3 Distributive powers. Example: the square of a binomial. . . . . . . . . . . . . 120
12.4 Combination of the square of a binomial with the original and the fractioning
pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
12.5 Distributive powers. The square of a trinomial . . . . . . . . . . . . . . . . . . 124
12.6 Distributive powers. Example: the square of a trinomial. . . . . . . . . . . . . 126
12.7 Combination of the square of a trinomial with the original and the fractioning
pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
12.8 Distributive powers. The cube of some binomials and trinomials . . . . . . . 128
12.9 Distributive powers. Example: the cube of a binomial and trinomial. . . . . . 131
12.10Combination of the cube of a binomial with the original, square and the frac-
tioning pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

13.1 Iterative process of permutation and synchronization . . . . . . . . . . . . . . 136

13.2 Fractional evolution of rhythm families. Determinants n = 2, 3, 4, 5 . . . . . . 137
13.3 Fractional evolution of rhythm families. Determinants n = 6, 7 . . . . . . . . . 139
13.4 Fractional division of rhythm families. Determinants n = 2, 3, 4, 5, 6, 7 . . . . 140
13.5 Fractional evolution of rhythm families. Determinants n = 8, 9 . . . . . . . . . 141
13.6 Fractional division of rhythm families. Determinants n = 8, 9 . . . . . . . . . . 142
13.7 Factorial evolution of rhythm families. Power series: square . . . . . . . . . . 143

14.1 Acceleration series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

14.2 Growth series in musical notation . . . . . . . . . . . . . . . . . . . . . . . . . . 148

viii Copy issued to Seth Dempsey ([email protected])

Preface

This is the second edition of the book A Guide to Schillinger’s Theory of Rhythm. This doc-
ument describes a system for creating musical rhythm. Combining a mathematical basis
with evolutionary growth processes provides various techniques for creating rhythmical
patterns. The content will be useful for composers and arrangers that need a toolset for
creating rhythms when they want to create alternatives to regular, short patterns and repet-
itive loops. These techniques are most useful for writing film, dance and computer game
music.

A book about rhythm

The Schillinger System of Musical Composition [3] was published originally in 1946 and as a
reprint in a 1640 pages two-volume book set in 1978. The books were compiled by the ed-
itors from source material by Joseph Schillinger (1895–1943) and from student notes. The
system consists of 12 books on various aspects of musical composition. These books are not
part of the standard literature; they are in general not being considered as serious studies in
musicology. The mathematical notation and number theory approach did not help spread-
ing the information. However, for a certain period they were considered as a toolbox for
composers and arrangers in radio, film, dance and theatre music. The numerous techniques
in these books will develop craftmanship.
The system begins with a book on the Theory of Rhythm. Rhythm is coordinated mu-
sical time; the theory of rhythm provides techniques for generating basic patterns, how to
develop and vary these patterns using the analogy of natural growth processes, and how
to apply the resultant rhythms to multiple parts in a musical score. The book also contains
musings about rhythmical styles and evolution, and there are predictons about the emer-
gence of future rhythm families. Book I from the Schillinger System consists of 14 chapters;
this Guide to Schillinger’s Theory of Rhythm closely follows the structure from the book and
covers the chapter subjects using mathematical, graphical and musical notation.1
Why did I write this guide to the theory of rhythm? First, hardopies of the Schillinger
books are hard to find; the volumes are out of print. Secondly, they are hard to read, due
to the confusing mathematical notation, the exhaustive presentation of combinations and
permutations of integer number sets.2 Finally, some techniques definitely needed more ex-
amples to clarify the application. Although the guide started as a summary of the Schillinger
1
Regarding the theory of rhythm there is a companion publication called Encyclopedia of Rhythms [4]. It is a
reference volume with single and double staff notation of all possible rhythmical patterns, created within the
Schillinger System.
2
In the pre-computer era (the 1930s and 1940s) the many pages with numbers may have served as a handbook
with lookup tables.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
Preface

book on rhythm, the second edition of this guide is significantly longer than the original text
with its 95 pages.
So this guide is both a replacement for and an addition to the original document. The
mathematical and graphical notation were somewhat adapted, many new musical notation
examples were created. The techniques and examples are explained in great detail, to the
best of my understanding. Application tips are provided for various musical styles. Website
statistics are evidence of the demand for such a book. The PDF version of the first edition
received a significant number of monthly hits. The first edition was incomplete, covering
only half of the chapters from the original Schillinger book on rhythm. Therefore the time
seemed appropriate to complete this work and publish a second edition. Studying the tech-
niques from the book will help the composer and arranger in creating, developing and ap-
plying rhythms. If you feel threatened by the danger of two-measure, repetitive rhythmical
loops, then this guide may trigger your creativity and provide more than sufficient alterna-
tives for creating interesting rhythms, that still contain homogeneity and consistency. When
inspiration is failing, try this recipe book to overcome your writer’s block.

Document update history

Here is an overview of the document history:

Vs. 1.1, September 2006: Chapter 1 to 4 completed (text plus figures), Chapter 6 text only.8

Vs. 1.2, March 2011: Chapter 5 added, additions to Chapter 6, minor edits to other chapters.
Total: 37 pages.

Second Edition, May 2015, Vs. 2.1: Chapters 7–14 added. Layout and styling updates. Ad-
ditional examples and figures in Chapters 2–6. Errors corrected. Total: 159 pages in the
full version.
Revision 2.2, September 2015: Errors corrected and additions, based on reader suggestions
(mainly in Chapter 1, see the overview of the terminology, and in Chapter 7, text and
figures). Mathematical notation improved for consistency and easier reading. Full
version total: 167 pages.

x Copy issued to Seth Dempsey ([email protected])

Chapter 1

Introduction

The Theory of Rhythm from the Schillinger System of Musical Composition [3] deals with tem-
poral aspects of music. Rhythm is coordinated use of music time through note attacks and
durations. In the field of musical rhythm there are not many textbooks. However, here’s
a number of suggestions for books that discuss rhythm and temporal aspects of music: see
[1, 2].5
This introductory chapter will provide a structural overview of the content of this guide
and familiarize the reader with the integer number approach to rhythm notation. The orig-
inal Schillinger Theory of Rhythm contains 14 chapters that describe rhythmical aspects and
techniques on different levels. Studying the many techniques from these chapters carries the
risk of losing the total picture. Therefore we start with an overview that is not part of the
original book.

1.1 Document structure overview

The overview of the document structure is shown in a diagram in Figure 1.1. In Schillinger’s
Theory of Rhythm three levels may be discerned:

1. There is the lowest level, the source level where rhythmical patterns are generated.
There are several techniques for creating interference patterns from two or three clocks
or metronomes. These metronome clocks will be ticking at constant time intervals,
but variable speeds may also be used. The resultant rhythmical pattern from these
generators, i.e., an attack-duration group is the output at this level.

2. Using the generator source pattern, there are various techniques that yield variation
and development. Subdivision of the original rhythm attack-duration series, grouping
patterns into measures at a specified time signature or meter, multiplying the resultant
by a set of coefficients, all these approaches will create homogeneous variation and a
rhythmical continuity at an intermediate level.

3. At the highest level there is the application of a rhythmical resultant to a musical instru-
ment or a number of parts in a score. Patterns may be combined in parallel, distributed
into a simultaneity over multiple staves. Then there is also the evolution of rhythms:
from the basic patterns families of rhythms will evolve with coherent and unique char-
acteristics. This brings along the aspect of rhythmical style.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 1. INTRODUCTION

In the diagram the chapters are positioned on these levels. Most chapters remain on a
single level, but some of them will cover aspects on multiple levels. So, a clear separation
is not possible, but the positioning will indicate the focus of the chapter. The diagram will
help the reader to keep the big picture, while studying the detailed techniques in the book.

1.2 Terminology
In this book about rhythm we will use much of the original nomenclature from the Schillinger
System of Musical Composition books. An overview of the most frequently used terminology,
symbols and equivalent meaning in other music literature may help in reading the document
and understanding the fundamental concepts. The list below may serve as a glossary of
terms.
Generator. A generator is a rhythm source, creating note attacks at a fixed, constant time
interval. Symbol: A, B, . . ., with time intervals a∆t, b∆t, . . ., respectively. The perfect
example of a rhythm generator is a metronome, ticking at a specific Beats-Per-Minute
(BPM) setting.
Attack. The attack is the beginning of a note, i.e., a new event in a rhythmical sequence.
Symbol: a.1 Note that often the note duration is implicit when using the term attack.
So the rhythm 3 + 1 + 2 + 1 + 1 + 4 consists of 6 attacks, or Na = 6. In most chapters
we will see attack groups in the time domain; however, they may also be applied as a
pitch distribution series.
Resultant. The resultant is the output of a rhythm generation process, such as interference.
Symbol r. The resultant consists of an attack-duration series, i.e., a number of note
events with durations. It is a pattern in the time domain. So we will write r = 3 + 1 +
2 + 2 + 1 + 3 or r = 3∆t + ∆t + 2∆t + 2∆t + ∆t + 3∆t, an attack-duration series with
6 elements and a total duration of Tr = 12∆t, i.e., 12 time units.
Time unit. The time unit is the smallest division in the time domain. Symbol: ∆t. Note
durations and rhythms are expressed as integer multiples of this time unit. So a note
duration may be 3∆t, a single measure may contain 12∆t time units, and a rhythm
written in short as 3 + 1 + 2 + 1 + 1 + 4 consists of six notes with durations 3∆t + 1∆t +
2∆t + 1∆t + 1∆t + 4∆t.
Periodicity. Many rhythmical phenomena have a periodic character; they will repeat after a
certain, constant period of time. E.g., the rhythm 2 + 1 + 1 + 2 + 1 + 1 + 2 + 1 + 1 is
repeating after 4 time units. We may write this also as (2+1+1)+(2+1+1)+(2+1+1)
which helps in recognizing the periodic nature. In this example the period is 4∆t; it is
4 time units long.
Interference. Rhythm is often the result of a combination of two or more periodic processes,
running simultaneously, but each with a different period. Think of two metronomes,
each with a different BPM setting. The combination of time events from the separate
processes is called interference.
1
Notation ambiguity cannot be totally avoided in this book. Here we have an example of such a double
meaning. The symbol a represents both the note attack and the time interval setting of the rhythm generator A.
The context should resolve the ambiguity.

2 Copy issued to Seth Dempsey ([email protected])

 1.2. TERMINOLOGY

Source/Generator Variation/Development Application/Style

Integer number clocks Time unit t, meter Parts, Score

2. Interference
- 3. Grouping technique
-
2 generators a, b
Resultant r
n
Meter, time signature n
Recurrence, N T
4. Fractioning
- 5. Group by pairs
2 generators
Grouping T Balancing rB , expanding rE ,
contracting rC
6. Three generators
-
N generators - 7. Instrumental forms
Interference - Interference attack/harmony
Grouping Combination
-
- 8. Coordinated structures
H
H
HH Distribution instr./attack
H Synchronization
HH
- 9. Homogeneous variation H
Permutations, sets {a, b}
Rests, accents, split-unit
?
10. Generalization of variation
Higher order permutations
and combinations, sets {a, b, c}

- 11. Homogeneous continuity

Splitting by common divisor,
measure, attack

- 12. Distributive powers

n
Meter n , fractional level
Square r2 , cube r3 , factorial
- 13. Evolution of families

14. Variable velocities Determinant n

Unbalanced binomials
Acceleration series
Fractional-factorial level
Variable tempo

Figure 1.1: Overview of the book structure

Copy issued to Seth Dempsey ([email protected]) 3

CHAPTER 1. INTRODUCTION

Synchronization. The synchronization process is closely related to the interference phenome-

non. Two or more periodic processes (clocks, metronomes) start at a certain time in-
stant with a simultaneous first time event (an attack). When the duration of two peri-
ods have a simple ratio, then after a while both phenomena will return to the beginning
of the pattern. For example, one rhythmic pattern has a duration of 3 time units, the
other has the period 2∆t, then after 6 time units both patterns return to the start of a
new period. The first pattern will have played twice, the other pattern three times.

Grouping. Rhythms are grouped into fixed duration units, i.e., measures of a certain dura-
tion. Grouping is relevant for the notation of music, the counting of time units by
musicians or in DAW software.

Measure, bar. A measure or bar of music (synonyms) is a rhythm time division unit. Mea-
sures are the result of regular grouping into units with duration TM .

Meter, time signature. The meter and time signature are synonyms,
n
indicating the subdivi-
sion of a measure into smaller units. The notation is m , with n and m two integer
numbers. The lower integer m indicates the time unit ∆t within the measure, n is the
total numberh ofi time units in a single measure. So we have TM = nm = n∆t. Exam-
ples are the 44 time signature, with the quarter note time unit and four beats to the
h i h i
3 6
measure, the 4 Waltz, and the 8 meter with six units of 8th notes in a bar.

Recurrence. The result of synchronization is the recurrence of the combined periodic phe-
nomena after a certain time interval. Symbol: Tr , TR .2 See the example under synchro-
nization, which repeats after TR = 6∆t. Often we will be calculating the recurrence of a
certain rhythm in combination with grouping into a meter. E.g., the resultant r = 2 + 1
with duration Tr = 3∆t, when grouped into a meter of two time units TM = 2∆t (one
bar contains two time units) will have recurrence after TR = 6∆t = 3TM , i.e., after
three full meaures.

Instrumental form. An instrumental form is a pattern in the pitch domain. An attack pattern
is mapped onto an ordered set of pitches, either a melody (a sequence of pitches) or a
harmonic structure (simultaneous pitches). In this book about rhythm an instrumental
form is created by applying an attack-duration group (time domain) to a given pitch-
attack distribution pattern (pitch domain). This may happen on a single staff or on
multiple staves in a musical score.

Some of the terminology may seem puzzling now, but should become clear when study-
ing the subsequent chapters in this book. A graphical representation of the technical terms
is shown in diagram in Fig. 1.2.

1.3 Notation system

In his Theory of Rhythm Schillinger introduces alternative systems of rhythmic pattern no-
tation: numbers, graphs and musical notes. This document will also use all three notation
2
The symbol Tr is also the length of a resultant, an attack-duration series. Creating a rhythmic resultant from
a binary synchronization process is also a recurrent process.

4 Copy issued to Seth Dempsey ([email protected])

 1.3. NOTATION SYSTEM

A + B
 generators with periodicity a∆t and b∆t

interference pattern  synchronization process 

- attack-duration group r with Na attacks

 - periodicity Tr  periodicity (TR )
?note: a attack duration ‘RHYTHM’
resultant ?
r: d dd d dd ?

n
m
m
=-∆t
-
6 6 time t
time unit recurrence:
 TM = n∆t - synchronization of resultant
measure, bar  grouping attack-duration pattern r with period Tr
meter, time signature and grouping by TM into measures

Figure 1.2: Overview of rhythm terminology in diagram.

Figure 1.3: Plotting the rhythm 3 + 1 + 2 + 2 + 1 + 1 + 1 + 1 + 4 on graph paper according to

the Schillinger System books. A vertical line segment indicates an attack. The note duration
is represented by the horizontal line segments.

systems. In the original book the usefulness of graphs is stressed, because of the visualiza-
tion aspect.

1.3.1 Graphing music

In the book the analogy between acoustic waveforms (periodic patterns, displaying the
sound amplitude vs. time) and durations (stressed accents) is used to introduce the square
wave graphical notation of musical attacks. So the Schillinger books use the graphic rep-
resentation of a rhythm, as shown in Fig. 1.3. On a regular grid the duration each note is
represented by a horizontal line segment.
Here we will use the analogy of clocks or metronomes ticking (short pulses of sound) at
regular intervals, as shown in Fig. 1.4. The time instants of the ticking will be represented
as symbols (circles) in a diagram or as numbers. Most of the techniques (the mathematical
processes, the arithmetic) will be done in integer number calculations; don’t be afraid, this
is all very simple and the analogy of the ticking metronomes will help to understand the
concepts and results.

Copy issued to Seth Dempsey ([email protected]) 5

CHAPTER 1. INTRODUCTION

“tick”
b b b b b b b b b b
-
t

Figure 1.4: Plotting the time instants of a ticking metronome along a time axis

1.3.2 Integer number representation of ticking metronomes

This guide is about rhythm, i.e., coordinated musical time. Notes in a rhythm will have an
attack (the start of the note) and a duration (the note length). Ticking clocks or metronomes
will generate attacks at specific time instants, determined by the clock speed or periodicity.
Rhythmical patterns are written along a time axis. Calculations with attacks and dura-
tions become easier with integer numbers: 0, 1, 2, 3, 4, . . .. This implies that the metronome
periodicity is an integer multiple of a certain time unit ∆t. The representation of the rhythm
can be in numbers (mathematical notation), in graphics (dots, lines on graph paper) or in
musical notation. To become familiar with the mathematical notation of rhythm, let us first
have a look at forms of periodicity.

1.3.3 Forms of periodicity

Uniform periodicity can be achieved with a single metronome generating pulses at a constant
rate. This is called monomial periodicity, since it is determined by a single coefficient. The
pulses are generated at discrete time intervals ∆t, i.e., the smallest rhythmical time unit.
Attacks will occur at multiples of this smallest time unit. We can write the time instant of the
i-th attack ti (from a total series of N attacks or pulses) in mathematical form as follows

ti = (i − 1)∆t, i = 1, 2, . . . , N. (1.1)

Note that the first attack occurs at t = 0 (and not t = 1 ∆t). That may seem a bit odd, but
later we will see that this makes the arithmetic a lot easier to understand. We may represent
the whole series of ticks as a vector ~t (note the small arrow over the symbol t) and therefore
we write
~t = [0 1 2 . . . (N − 1)] · ∆t. (1.2)
Let us look at an example of uniform periodicity with ticking metronomes.

Example 1.1
Monomial periodicity: attack series.
The idea of attack series will be illustrated by considering three ticking metrono-
mes, each ticking at a fixed time interval ∆t.

• Consider metronome A ticking five times at intervals of one time unit (e.g., 1
second intervals). Then we have N = 5 and ∆t = 1 and the series of attacks
is written as3
~tA = [0 1 2 3 4].
3
In general, the ticking interval for a generator (metronome) A is written as a∆t with reference to a common
smallest time unit ∆t. In this and part of other chapters we write for brevity ∆tA = a∆t.

6 Copy issued to Seth Dempsey ([email protected])

 1.3. NOTATION SYSTEM

b b b b b b b b b b bB
bbbbbA
-
0 10 20 30 t

Figure 1.5: Two metronomes A and B ticking at different time intervals ∆tA = 1 and ∆tB = 3
time units. Tickmarks on the time axis represent the time instants.

The tick pattern is sketched in Fig. 1.5. This is the four-on-the-floor kick drum
pattern in electronic dance music (EDM).
• Another metronome B, generating 11 (N = 11) pulses at three time unit
intervals (∆t = 3) will yield an attack series

~tB = [0 1 2 . . . 10] · 3 = [0 3 6 . . . 30].

The tick pattern is shown in the figure above the A metronome. This could
be the pizzicato contrabass pattern on the downbeat of every measure in a
waltz.
• As a last example we will consider metronome C generating N pulses at n
time unit intervals. The number N has no specific value, and this in general
means that we have a very long or infinitely long sequence of ticks. In that
case the attack series is

~tC = [0 1 2 . . . N − 1] · n = [0 n 2n . . . (N − 1)n].

Note that we have only indicated the time instant of the beginning of a musical event,
i.e., an attack (e.g., the staccato tones from a xylophone). The notation in the Schillinger
book uses +-signs between the elements in an attack series, because the numbers in the
series there also indicate the duration of the attacks. The duration of an attack a is the time
interval between two ticks, i.e., a = ti+1 − ti = ∆t.4 As an analogy, consider an electronic
keyboard, where you would press a specific key at the abovementioned time instant and
keep the key depressed, until the next time event occurs. The result obviously is a series
of repeated pitches with a specific duration, i.e., an attack-duration group. If the duration
series is meant we will use the series with the +-signs between the terms.

Example 1.2
Monomial periodicity: duration series.
The aspect of the duration of an attack series is illustrated with the metronomes
used already in the previous example.
4
Setting the note length, i.e., the duration, equal to the ticking interval implies a legato playing. There are no
rests or pauses between the metronome ticks, such as in staccato playing.

Copy issued to Seth Dempsey ([email protected]) 7

CHAPTER 1. INTRODUCTION

• Returning to the metronomes A to C from Ex. 1.1 we obtain the following

attack-duration series:
N1
X 5
X
rA = ∆t1 = 1 = 1 + 1 + 1 + 1 + 1,
1 1
N
X2 11
X
rB = ∆t2 = 3 = 3 + 3 + 3 + 3 + 3 + 3 + 3 + 3 + 3 + 3 + 3,
1 1
N
X3 N
X
rC = ∆t3 = n = n + n + n + · · · + n.
1 1

8 Copy issued to Seth Dempsey ([email protected])

Chapter 2

Interference of periodicities

This chapter introduces a basic technique for generating attack series; in the Schillinger text
these are also called an attack group. The technique is based in the interference pattern that
results when two clocks or metronomes tick at a different constant time interval. These
attack series may then be grouped using a specified number of time units per measure.

2.1 Binary synchronization

Suppose there are two clocks or metronomes A and B ticking at different constant time 2
intervals ∆tA = a∆t and ∆tB = b∆t, where a and b are integer numbers and ∆t is the
musical time unit (e.g., a quarter note 41 or an 8th note 81 ). We assume that metronome B
is ticking faster than metronome A, and therefore ∆tB < ∆tA . Metronome B is called the
minor generator, metronome A is called the major generator.
Starting these two clocks at the same time instant will yield an attack pattern, called
the resultant r. The attack pattern will repeat after Tr = ab∆t time units. The process of
combining the two metronomes, the two monomial periodicities, is called interference and
since there are two clocks we call this binary synchronization. When metronome A and B
produce an attack series, the interference process is notated as a ÷ b and the resultant attack-
duration series is written as ra÷b .
We determine the resultant time series by finding the combination (in mathematical
terms, the union) of the two attack series ~tA and ~tB , i.e.,
~tr = ~tA ∪ ~tB . (2.1)
Given the attack series ~tr we determine the duration series r by listing the difference between
two consecutive terms in the attack series
ri = tr,i+1 − tr,i . (2.2)

2.1.1 Uniform binary synchronization

In the case of uniform binary synchronization the faster metronome ticks at intervals of ∆tB = 1
time unit. Remember, the time unit may be any musical note duration, e.g., ∆tB = 41 note or
∆tB = 18 note.
In this special case the attack-duration patterns will repeat after ∆tA time units. The book
discusses the three most common cases of this interference a ÷ b.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 2. INTERFERENCE OF PERIODICITIES

rbrb B rbbrbb B rbbbrbbb B

r r A r r A r r A
- - -
0 5 t 0 5 t 0 5 10 t
(a) r2÷1 (b) r3÷1 (c) r4÷1

Figure 2.1: Uniform binary synchronization. Two metronomes A and B are ticking at dif-
ferent time intervals. a): ∆tA = 2 and ∆tB = 1, b): ∆tA = 3 and c): ∆tA = 4 time units.
Accented attacks are indicated by closed circles, and the pattern is repeated once.

Case 1: 2 and 1
Now ∆tA = 2 and ∆tB = 1 time units and using Eq. 1.2 we obtain the attack series
~tA = [0], ~tB = [0 1],

repeating itself after 2 time units. We determine the resultant from the combination of these
two attack series and get the following attack and duration series
~tr = [0 1], r = 1̂ + 1,

where 1̂ indicates an accented attack occurring every ∆tA time units. The synchronization
process is illustrated in Fig. 2.1.a.

Case 2: 3 and 1
For ∆tA = 3 and ∆tB = 1 and using Eq. 1.2 we get the attack series
~tA = [0], ~tB = [0 1 2],

repeating itself after 3 time units. We determine the resultant from the combination of these
two attack series and get the following attack and duration series
~tr = [0 1 2], r = 1̂ + 1 + 1,

shown in Fig. 2.1.b.

Case 3: 4 and 1
Finally, ∆tA = 4 and ∆tB = 1 and using Eq. 1.2 yields the attack series
~tA = [0], ~tB = [0 1 2 3],

with recurrence after 4 time units. We determine the resultant from the combination of these
two attack series and get the following attack and duration series (see Fig. 2.1.c).
~tr = [0 1 2 3], r = 1̂ + 1 + 1 + 1.

The process of uniform binary synchronization leads to trivial rhythmic results: a series
of attacks with equal duration where at regular intervals the attack receives an accent. More
interesting resultants are obtained in the next section.

10 Copy issued to Seth Dempsey ([email protected])

 2.1. BINARY SYNCHRONIZATION

Table 2.1: List of generator combinations for non-uniform binary synchronization. The first
number is the major generator value a, the second number is the minor generator value b

Major Interference group (combination of generators)

3: 3÷2
4: (2 ÷ 1) 4÷3
5: 5÷2 5÷3 5÷4
6: (3 ÷ 1) (2 ÷ 1) (3 ÷ 2) 6÷5
7: 7÷2 7÷3 7÷4 7÷5 7÷6
8: (4 ÷ 1) 8÷3 (2 ÷ 1) 8÷5 (4 ÷ 3) 8÷7
9: 9÷2 (3 ÷ 1) 9÷4 9÷5 (3 ÷ 2) 9÷7 9÷8

Application Tip:
The uniform binary synchronization technique can be used in combination in
a multi-staff, layered score. Use the slower metronome rhythm, i.e., the major
generator A, in the first layer for the attacks with long duration. Then increase
the rhythmical drive by adding either the 1:2 (doubling), 1:3 (tripling) or the 1:4
(quadruple) minor generator in the second layer. In a three-layer approach use
the slower tick tempo with 1:2 in layer two and 1:4 in layer three. This may be
useful in interactive game music, using fade-in and fade-outs to mix the layers. The
combination of 1:2 and 1:3 creates somewhat more complex cross-rhythms.

2.1.2 Non-uniform binary synchronization

In the case of non-uniform binary synchronization we have two metronomes A and B with
∆tA > ∆tB (both an integer number) and ∆tB 6= 1. The resultant attack series is repeated 2
after Tr = ∆tr = ∆tA ∆tB = ab∆t time units and the attack-duration group r will contain
non-equal durations.
On [3] p. 14 in the Schillinger book we find the list of practical combinations of genera-
tors, here reproduced (somewhat re-arranged and complemented) as Table 2.1. We note that
only the lower left triangular region of the table is filled; this is due to the fact that we must
have ∆tA > ∆tB . Then there is a number of terms between brackets, e.g., (2 ÷ 1). These are
either uniform synchronization cases (where the minor generator value b = 1), or combina-
tions that can be reduced to simpler ratios using the common denominator. For example for
∆tA = 4, (a = 4) and ∆tB = 2, (b = 2) we may write
a ÷ b = 4 ÷ 2 = (2 · 2) ÷ (1 · 2) = 2 ÷ 1,
and for ∆tA = 6, (a = 6) and ∆tB = 4, (b = 4) we may write
a ÷ b = 6 ÷ 4 = (3 · 2) ÷ (2 · 2) = 3 ÷ 2.
Now we will consider the combinations listed in the table and determine the resultant r
for each pair (a, b). These will be shown in graphical and in musical notation. The musical
notation will be shown in Chapter 3, when the aspect of grouping has been discussed in
more detail.

Copy issued to Seth Dempsey ([email protected]) 11

CHAPTER 2. INTERFERENCE OF PERIODICITIES

r bbb r r bb b bb r
r b b B r b b b B
r b A r b b A
- -
0 5 t 0 5 10 t
(a) r3÷2 (b) r4÷3

r b bbb b r r b bb bb b r
r b b b b B r b b b b B
r b A r b b A
- -
0 5 10 t 0 5 10 15 t
(c) r5÷2 (d) r5÷3

r bb b b b bb r r bb b b b b b bb r
r b b b b B r b b b b b B
r b b b A r b b b b A
- -
0 5 10 15 20 t 0 5 10 15 20 25 30 t
(e) r5÷4 (f) r6÷5

Figure 2.2: Non-uniform binary synchronization. The resultant ra÷b of two generators A
and B that tick at different time intervals. a): ∆tA = 3, ∆tB = 2, b): ∆tA = 4, ∆tB = 3, c):
∆tA = 5, ∆tB = 2, d): ∆tA = 5, ∆tB = 3, e): ∆tA = 5, ∆tB = 4, f): ∆tA = 6, ∆tB = 5 time
units. Accented attacks are indicated by closed circles.

Interference generators: 3 and 2

For ∆tA = 3 and ∆tB = 2 and using Eq. 1.2 we get the attack series

~tA = [0 3], ~tB = [0 2 4],

repeating itself after Tr = ab∆t = 3 · 2 = 6 time units. We determine the resultant from
the combination of these two attack series, Eq. 2.1, which yields the following attack and
duration series
~tr = [0 2 3 4], r = 2 + 1 + 1 + 2,

i.e., four attacks (Na = 4) and two note duration values. The non-uniform generator interfer-
ence process is illustrated in Fig. 2.2.a. The diagram shows the attacks as a repeating series
of dots and circles along a time axis t. The top row shows the resultant r, obtained from the
union of the two generator attack series.

Interference generators: 4 and 3

For ∆tA = 4 and ∆tB = 3 and using Eq. 1.2 we get the attack series

~tA = [0 4 8], ~tB = [0 3 6 9],

with recurrence after Tr = 4 · 3 = 12 time units. We determine the resultant from the com-
bination of these two attack series, Eq. 2.1, which yields the following attack and duration

12 Copy issued to Seth Dempsey ([email protected])

 2.1. BINARY SYNCHRONIZATION

series
~tr = [0 3 4 6 8 9], r = 3 + 1 + 2 + 2 + 1 + 3,
i.e., six attacks (Na = 6) and three note duration values. The synchronization resultant r is
shown in Fig. 2.2.b. When we discuss grouping in Chapter 3, we wll see that the total pattern
duration Tr = 12∆t is convenient, since the number can be divided by 2, 3, 4 and 6.

Interference generators: 5 and 2

With ∆tA = 5 and ∆tB = 2 as input into Eq. 1.2 we get the attack series

~tA = [0 5], ~tB = [0 2 4 6 8],

with total pattern length Tr = 5 · 2 = 10 time units. We determine the resultant from the
union of these two attack series, Eq. 2.1, which yields the following attack and duration
series
~tr = [0 2 4 5 6 8], r = 2 + 2 + 1 + 1 + 2 + 2,

i.e., Na = 6 and two note duration values. The resulting pattern is shown in Fig. 2.2.c.

Interference generators: 5 and 3

With ∆tA = 5 and ∆tB = 3 Eq. 1.2 the two attack series becomes

~tA = [0 5 10], ~tB = [0 3 6 9 12],

with recurrence after Tr = 5 · 3 = 15 time units. The union of these two attack series, Eq. 2.1
yields the attack and duration series

~tr = [0 3 5 6 9 10 12], r = 3 + 2 + 1 + 3 + 1 + 2 + 3,

i.e., Na = 7 and three note duration values. The resultant rhythm is shown in Fig. 2.2.d.

Interference generators: 5 and 4

The final case for this major generator is ∆tA = 5 and ∆tB = 4. Using Eq. 1.2 leads to the
attack series
~tA = [0 5 10 15], ~tB = [0 4 8 12 16],

with total pattern duration Tr = 5 · 4 = 20 time units. With Eq. 2.1 we find the following
attack and duration series

~tr = [0 4 5 8 10 12 15 16], r = 4 + 1 + 3 + 2 + 2 + 3 + 1 + 4,

i.e., Na = 8 and four note duration values. The resulting pattern is shown in Fig. 2.2.e. The
major generator a = 5 was combined with three values of the minor generator b = {2, 3, 4}.
Comparing these cases we note that the latter number is determining the range of duration
values and the number of attacks in the series; the resultant series contains more elements
when b increases and consists of duration values between 1∆t and b∆t.

Copy issued to Seth Dempsey ([email protected]) 13

CHAPTER 2. INTERFERENCE OF PERIODICITIES

Interference generators: 6 and 5

For ∆tA = 6 and ∆tB = 5 and using Eq. 1.2 we get the attack series

~tA = [0 6 12 18 24], ~tB = [0 5 10 15 20 25],

repeating itself after Tr = 6 · 5 = 30 time units. We determine the resultant from the union of
these two attack series, using Eq. 2.1, which yields the following attack and duration series

~tr = [0 5 6 10 12 15 18 20 24 25], r = 5 + 1 + 4 + 2 + 3 + 3 + 2 + 4 + 1 + 5,

i.e., Na = 10 and five note duration values. The resultant rhythm is shown in Fig. 2.2.f.

Interference generators: 7 and 2

When ∆tA = 7 and ∆tB = 2, Eq. 1.2 yields the attack series

~tA = [0 7], ~tB = [0 2 4 6 8 10 12],

with recurrence after Tr = 7 · 2 = 14 time units. The resultant from Eq. 2.1 leads to the
following attack and duration series

~tr = [0 2 4 6 7 8 10 12], r = 2 + 2 + 2 + 1 + 1 + 2 + 2 + 2,

i.e., Na = 8 and two note duration values.

Interference generators: 7 and 3

With ∆tA = 7 and ∆tB = 3 as input into Eq. 1.2 the attack series is

~tA = [0 7 14], ~tB = [0 3 6 9 12 15 18],

repeating after Tr = 7 · 3 = 21 time units. From the combination of these two attack series,
using Eq. 2.1, the resultant attack and duration series becomes

~tr = [0 3 6 7 9 12 14 15 18], r = 3 + 3 + 1 + 2 + 3 + 2 + 1 + 3 + 3,

i.e., Na = 9 and three note duration values.

Interference generators: 7 and 4

For ∆tA = 7 and ∆tB = 4 Eq. 1.2 produces the attack series

~tA = [0 7 14 21], ~tB = [0 4 8 12 16 20 24],

repeating itself after Tr = 7·4 = 28 time units. Determine the resultant from the combination
of these two attack series, Eq. 2.1 and find the attack and duration series

~tr = [0 4 7 8 12 14 16 20 21 24], r = 4 + 3 + 1 + 4 + 2 + 2 + 4 + 1 + 3 + 4,

i.e., Na = 10 and four note duration values.

14 Copy issued to Seth Dempsey ([email protected])

 2.1. BINARY SYNCHRONIZATION

Interference generators: 7 and 5

With ∆tA = 7 and ∆tB = 5 and using Eq. 1.2 the two attack series are
~tA = [0 7 14 21 28], ~tB = [0 5 10 15 20 25 30],
with recurrence after Tr = 7 · 5 = 35 time units. The union of both according to Eq. 2.1 gives
the resultant attack and duration series
~tr = [0 5 7 10 14 15 20 21 25 28 30], r = 5 + 2 + 3 + 4 + 1 + 5 + 1 + 4 + 3 + 2 + 5,
i.e., Na = 11 and five note duration values.

Interference generators: 7 and 6

The final case for this major generator is ∆tA = 7 and ∆tB = 6 time units. Using Eq. 1.2 the
pair of attack series is
~tA = [0 7 14 21 28 35], ~tB = [0 6 12 18 24 30 36],
with total pattern duration Tr = 7 · 6 = 42 time units. We determine the resultant from
the combination of these two attack series, Eq. 2.1, which yields the following attack and
duration series
~tr = [0 6 7 12 14 18 21 24 28 30 35 36], r = 6 + 1 + 5 + 2 + 4 + 3 + 3 + 4 + 2 + 5 + 1 + 6,
i.e., Na = 12 and six note duration values. Major generator a = 7 enables five combinations
b = {2, 3, 4, 5, 6}. Again note the dependence of the resultant pattern length and the range of
note durations on the value of the minor generator b.

Interference generators: 8 and 3

For ∆tA = 8 and ∆tB = 3 and using Eq. 1.2 we get the attack series
~tA = [0 8 16], ~tB = [0 3 6 9 12 15 18 21],
repeating itself after Tr = 8 · 3 = 24 time units. We determine the resultant from the com-
bination of these two attack series, Eq. 2.1, which yields the following attack and duration
series
~tr = [0 3 6 8 9 12 15 16 18 21], r = 3 + 3 + 2 + 1 + 3 + 3 + 1 + 2 + 3 + 3,
i.e., Na = 10 and three note duration values. The pattern duration Tr = 24∆ provides many
grouping options, since the number 24 is divisible by 2, 3, 4., 6, 8 and 12.

Interference generators: 8 and 5

When ∆tA = 8 and ∆tB = 5, Eq. 1.2 produces the two attack series
~tA = [0 8 16 24 32], ~tB = [0 5 10 15 20 25 30 35],
repeating itself after Tr = 8 · 5 = 40 time units. We determine the resultant from the com-
bination of these two attack series, Eq. 2.1, which yields the following attack and duration
series
~tr = [0 5 8 10 15 16 20 24 25 30 32 35], r = 5 + 3 + 2 + 5 + 1 + 4 + 4 + 1 + 5 + 2 + 3 + 5,
i.e., Na = 12 and five note duration values.

Copy issued to Seth Dempsey ([email protected]) 15

CHAPTER 2. INTERFERENCE OF PERIODICITIES

Interference generators: 8 and 7

The third and final case for major generator a = 8 is ∆tA = 8 and ∆tB = 7 and using Eq. 1.2
we get the attack series

~tA = [0 8 16 24 32 40 48], ~tB = [0 7 14 21 28 35 42 49],

with recurrence after Tr = 8 · 7 = 56 time units. Ths is a pattern of considerable length.

Determine the resultant from the combination of these two attack series, Eq. 2.1, and find
the following attack and duration series

~tr = [0 7 814 16 21 24 28 32 35 40 42 48 49], r = 7+1+6+2+5+3+4+4+3+5+2+6+1+7,

i.e., Na = 14 and seven note duration values.

Interference generators: 9 and 2

The last practical major generator value is a = 9, with five possible combinations for non-
uniform binary synchronization. For ∆tA = 9 and ∆tB = 2 and using Eq. 1.2 we get the
attack series
~tA = [0 9], ~tB = [0 2 4 6 8 10 12 14 16],

repeating itself after Tr = 9 · 2 = 18 time units. The resultant from the union of these two
attack series according to Eq. 2.1 consists of the attack and duration series

~tr = [0 2 4 6 8 9 10 12 14 16],
r = 2 + 2 + 2 + 2 + 1 + 1 + 2 + 2 + 2 + 2,

i.e., Na = 10 and two note duration values.

Interference generators: 9 and 4

With ∆tA = 9 and ∆tB = 4 Eq. 1.2 produces the attack series

~tA = [0 9 18 27], ~tB = [0 4 8 12 16 20 24 28 32],

with recurrence after Tr = 9 · 4 = 36 time units. The grouping options are many, since 36
allows division by 2, 3, 4, 6, 9 and 12. Determine the combination of these two attack series,
Eq. 2.1 and find the resultant attack and duration series

~tr = [0 4 8 9 12 16 18 20 24 27 28 32],
r = 4 + 4 + 1 + 3 + 4 + 2 + 2 + 4 + 3 + 1 + 4 + 4,

i.e., Na = 12 and four note duration values.

Interference generators: 9 and 5

Enter ∆tA = 9 and ∆tB = 5 into Eq. 1.2 and find the pair of attack series

~tA = [0 9 18 27 36], ~tB = [0 5 10 15 20 25 30 35 40],

16 Copy issued to Seth Dempsey ([email protected])

 2.1. BINARY SYNCHRONIZATION

with total pattern duration Tr = 9 · 5 = 45 time units. With Eq. 2.1 the resultant attack and
duration series is

~tr = [0 5 9 10 15 18 20 25 27 30 35 36 40],
r = 5 + 4 + 1 + 5 + 3 + 2 + 5 + 2 + 3 + 5 + 1 + 4 + 5,

i.e., Na = 13 and five note duration values.

Interference generators: 9 and 7

For ∆tA = 9 and ∆tB = 7 and using Eq. 1.2 the generator attack series are

~tA = [0 9 18 27 36 45 54], ~tB = [0 7 14 21 28 35 42 49 56],

repeating itself after Tr = 9 · 7 = 63 time units. The combination of these two attack series
according to Eq. 2.1 yields the attack and duration series

~tr = [0 7 9 14 18 21 27 28 35 36 42 45 49 54 56]
r = 7 + 2 + 5 + 4 + 3 + 6 + 1 + 7 + 1 + 6 + 3 + 4 + 5 + 2 + 7,

i.e., Na = 15 and seven note duration values.

Interference generators: 9 and 8

Finally, when ∆tA = 9 and ∆tB = 8 using Eq. 1.2 produces the attack series

~tA = [0 9 18 27 36 45 54 63], ~tB = [0 8 16 24 32 40 48 56 64],

repeating itself after Tr = 9 · 8 = 72 time units. Although this is a long pattern, it allows
many regular groupings, since 72 can be divided by 3, 4, 6, 8, 9 and 12.
The union of these generator attack series, Eq. 2.1, leads to the resultant attack and dura-
tion series

~tr = [0 8 9 16 18 24 27 32 36 40 45 48 54 56 63 64]
r = 8 + 1 + 7 + 2 + 6 + 3 + 5 + 4 + 4 + 5 + 3 + 6 + 2 + 7 + 1 + 8,

i.e., Na = 16 and eight note duration values. The five combinations for major generator
a = 9 once again show the effect of the value of the minor generator b on total pattern
length, number of attacks and range of durations in the resultant series.
Considering all cases for larger values of the major generator, say a > 5, there is also a
sense of balance within the resultant; this is mainly determined by the value of the ratio of the
longest to shortest duration. For example, take the (a, b) = (9, 8) generator combination, this
ratio is 8 : 1. This implies that as the minor generator value b increases, the patterns become
more unbalanced. Another determining factor for internal balance is the succession of two
duration values with a large difference, such as 5 + 1, 8 + 1, 7 + 2.

Copy issued to Seth Dempsey ([email protected]) 17

CHAPTER 2. INTERFERENCE OF PERIODICITIES

2.1.3 Overview of non-uniform binary synchronization and fractioning resul-

tants
Table 2.2 presents an overview of the most important characteristics of the non-uniform
binary synchronization attack-duration groups. For each combination of generators {a, b}
the resultant pattern r has total length Tr , and consists of Na attacks. Also the distribution
of the note durations is shown: e.g., N2∆t is the number of notes in the pattern with length 2
time units.
Use this table when looking for an attack-duration series with specific characteristics.
The second half of the table presents these characteristics for the resultants r obtained through
the fractioning technique, discussed in Chapter 4.

2.2 Grouping
Grouping is the selection of the meter or time signature; the resultant will be divided into a
number of measures with each measure containing a fixed number of time units. In the case
of binary synchronization there are three options for grouping, discussed in the following
subsections.
Note that the next chapter, Ch. 3, is completely devoted to the aspect of grouping. In
addition, note that apart from the three options mentioned here, Schillinger also considers
what he calls alien measure grouping (see [3], Ch. 7, p. 33).

2.2.1 Grouping by the common product

The common product is the number of time units after which the resultant attack pattern re-
peats itself. So, grouping by the common product implies a single measure duration

TM = ∆tA · ∆tB = ab ∆t. (2.3)

Note that we will get unfamiliar time signatures, when we reach the higher numbers for
the generator combinations. Therefore, on [3], p. 10, there is an upper limit to this grouping
technique: do not group the resultant by the common product whenhTMi >h 15∆t, i or, equiv-
18 21
alently ab > 15. This would lead to unlikely time signatures such as 8 , 4 ; these can be
h i h i
6 3
replaced by the more common 8 and 4 , respectively.

2.2.2 Superimposition of a
Superimposition of a means that the major generator A will determine the meter. A single
measure (bar) will have the duration

TM = ∆tA = a ∆t. (2.4)

The selection of the grouping time unit TM is independent of the generator synchronization
process; it is a new degree of freedom. There is relation to the rhythm generators through
the use of the factors a and b. But, as this section shows, we have several alternatives with
measures of longer or shorter duration. Refer to the diagram in Fig. 1.2 to see the relation
between the attack-duration group (the rhythmic pattern) and the grouping process.

18 Copy issued to Seth Dempsey ([email protected])

 2.2. GROUPING

Table 2.2: Overview of the characteristics of the resultants r for non-uniform binary synchro-
nization and fractioning attack-duration groups. For combinations of two generators {a, b}
the total duration of the pattern Tr , the number of atacks in the series Na and the distribution
of the note duration values Nn∆t are listed.

Resultant Note duration distribution

ra÷b Tr Na N∆t N2∆t N3∆t N4∆t N5∆t N6∆t N7∆t N8∆t

Non-uniform binary synchronization

3÷2 6 4 2 2
4 ÷ 3 12 6 2 2 2
5 ÷ 2 10 6 2 4
5 ÷ 3 15 7 2 2 3
5 ÷ 4 20 8 2 2 2 2
6 ÷ 5 30 10 2 2 2 2 2
7 ÷ 2 14 8 2 6
7 ÷ 3 21 9 2 2 5
7 ÷ 4 28 10 2 2 2 4
7 ÷ 5 35 11 2 2 2 2 3
7 ÷ 6 42 12 2 2 2 2 2 2
8 ÷ 3 24 10 2 2 6
8 ÷ 5 40 12 2 2 2 2 4
8 ÷ 7 56 14 2 2 2 2 2 2 2
9 ÷ 2 18 10 2 8
9 ÷ 4 36 12 2 2 2 6
9 ÷ 5 45 13 2 2 2 2 5
9 ÷ 7 63 15 2 2 2 2 2 2 3
9 ÷ 8 72 16 2 2 2 2 2 2 2 2

ra÷b Tr Na N∆t N2∆t N3∆t N4∆t N5∆t N6∆t N7∆t N8∆t

Fractioning
3÷2 9 7 5 2
4 ÷ 3 16 10 6 2 2
5 ÷ 2 25 21 4 21
5 ÷ 3 25 17 19 4 2
5 ÷ 4 25 13 7 2 2 2
6 ÷ 5 36 16 8 2 2 2 2
7 ÷ 3 49 37 29 4 4
7 ÷ 4 49 31 21 4 4 2
7 ÷ 5 49 25 9 12 2 2
7 ÷ 6 49 19 9 2 2 2 2 2

Copy issued to Seth Dempsey ([email protected]) 19

CHAPTER 2. INTERFERENCE OF PERIODICITIES

2.2.3 Superimposition of b
Superimposition of b means that the minor generator B will determine the time signature

TM = ∆tB = b ∆t. (2.5)

Now, obviously there are less time units in a measure, compared to the grouping by the
major generator.

2.2.4 Alien measure grouping

In the case of non-uniform binary synchronization the potential of alien measure grouping is
determined by the set of divisors of the total pattern length Tr . For example, the combination
(a, b) = (6, 5) has total duration Tr = 30∆t time units and can be factorized into 30 = 1×30 =
2 × 15 = 3 × 10 = h i5 × 6.h The
i last factorization pair enables regular grouping and time
signature, such as 54 and 68 meter. The first factorization (30) is impractical. The second
and
h i thirdh factorization
i pairs yield alien groupings
h i h andi time signatures, such as the regular
2 3 10 15
4 and 8 meter, and and the irregular 8 and 8 meters.

Example 2.1
Grouping and superimposition.
The grouping and superimposition process is illustrated in Fig. 2.3. Four non-
uniform binary synchronization cases ra÷b are considered, each with the three
possible grouping mechanisms: grouping by the common product ab∆t, super-
imposition of the major generator a∆t, and superimposition of the minor gen-
erator b∆t. The synchronization process is discusssed in Section 2.1.2 and the
resultants were already shown in Fig. 2.2.

• Case 1: r3÷2 . This non-uniform binary synchronization pattern leads

to three possible time signatures, determined
h iby the grouping approach:
grouping by TM = ab∆t = 6∆t and meter 68 , superimposition of TM =
h i
3
a∆t = 3∆t with meter 4 (waltz), and superimposition of TM = b∆t = 2∆t
h i
2
and meter 4 . The meter for these three alternatives is determined by the
value of the time unit; here we have chosen t = 81 , 14 and 14 , respectively (8th
note and quarter note).
• Case 2: r4÷3 . This yields three groupings and timeh signature
i proposals:
12
grouping by TM = ab∆t = 12∆t and time signature 8 , superimposition
h i
4
of TM = a∆t = 4∆t with time signature 4 , and superimposition of TM =
h i
3
b∆t = 3∆t, e.g., 4 meter.
1
• Case 3: r5÷2 . With time unit t = 8the three groupings
h i
and potential time
5
signatures are: grouping by TM = ab∆t = 10∆t and 4 , superimposition
h i
5
of TM = a∆t = 5∆t and 4 , and superimposition of TM = b∆t = 2∆t and
h i h i
2 5
2 . Note the irregular meter 4 when grouping by the major generator a.

20 Copy issued to Seth Dempsey ([email protected])

 2.3. CHARACTERISTICS OF THE RESULTANT

ab∆t ab∆t
a∆t a∆t
b∆t b∆t
r bbb r r bb b bb r
- -
0 5 t 0 5 10 t
(a) r3÷2 (b) r4÷3

ab∆t ab∆t
a∆t a∆t
b∆t b∆t
r b bbb b r r b bb bb b r
- -
0 5 10 t 0 5 10 15 t
(c) r5÷2 (d) r5÷3

Figure 2.3: Grouping and superimposition for non-uniform binary synchronization attack
series ra÷b . Four cases are shown. a): r3÷2 , b): r4÷3 , c): r5÷2 , d): r5÷3 . For each case three
groupings are shown: grouping by the common product ab∆t, superimposition of the major
generator a∆t, and superimposition of the minor generator b∆t. This yields three meters or
time signatures per case.

• Case 4: r5÷3 . And finally, the threeh groupings

i and time signatures are:
15
grouping by TM = ab∆t = 15∆t and 8 , superimposition of TM = a∆t =
h i h i
5 3
5∆t and 4 , and superimposition of TM = b∆t = 3∆t and 8 .

The example shows that synchronizaton and grouping are independent processes. For
each pair of generators we obtain many options for selecting a time signature (duple, triple,
irregular meter). In musical staff notation this will be further illustrated in Chapter 3.

2.3 Characteristics of the resultant

Although in his book this is not a separate section, Schillinger concludes the section with a
few remarks on the characteristics of the resultant, summarized here as:
• The recurrence of the full pattern after T = ∆tA · ∆tB = ab ∆t time units. Then the
attack-duration group will repeat itself.
• The rhythmic symmetry of the duration series about the centre (beginning and end are
mirrored about the middle of the series). In other words, the time-reversed pattern,
i.e., the retrograde rhythmical series, is equal to the original interference pattern.
• The use of both complete and incomplete resultants in music. There is no requirement
for using full resultant series only; we may select a fraction (a subset of consecutive el-
ements) from the attack-duration group. The synchronization process sort of generates
a total complex that may be used as rhythmic source material.

Copy issued to Seth Dempsey ([email protected]) 21

CHAPTER 2. INTERFERENCE OF PERIODICITIES

• The analogy with cross-rhythm when two generators create an interference pattern.

2.3.1 The natural nucleus of a score

On p. 9 of [3], there is a comment about the combined use of all the groupings, which
Schillinger calls the natural nucleus of a musical score. There he writes that:

1. the duration pattern with the smallest time unit ∆t (the common denominator) may be
used for arpeggio or ostinato figures;

2. the resultants, grouped by either the major generator a or the minor generator b may
be used for chords (harmonic structures);

3. the resultant ra÷b represents the rhythmic structure of the melody;

4. and the longest duration note with length ab ∆t may be used for sustained notes, such
as a pedal point.

This will be demonstrated in an example.

Example 2.2
The natural nucleus of a musical score.
The process of creating a musical score from a natural nucleus is illustrated
in Fig. 2.4, using the non-uniform binary synchronization resultants r4÷3 and
r5÷3 . In this example a single chord is used to generate all components (melody,
chords, pedal and arpeggio) in a score fragment. Note the small liberty in the
duration of the attacks by the interspresed rests, and the slight rhythmic modifi-
cation of the arpeggio on the right (the techniques from these books are not meant
to be used rigidly, but to provide tools and stimulate creativity).
h i The time signa-
ture on the left for the grouping by the major generator a is 44 : a single measure
has duration TM = 4∆t, h i consisting of four time units, i.e., four quarter notes. On
the right the meter is 54 and TM = 5∆t, with grouping by the minor generator b.

22 Copy issued to Seth Dempsey ([email protected])

 2.3. CHARACTERISTICS OF THE RESULTANT

˙. ˙
4:3 5:3
œ œ œ œ. ˙ œ
M: r
b
&b c
˙ ‰ n n # 43 ˙ . ˙ œ ˙. œ ˙ J ‰

j j
b ˙
& b c ˙˙
œœ ..
œ. ‰ ˙˙˙ œœ ..
œ. ‰ ˙˙˙ œœ ..
œ. ‰ n n # 43 ˙˙˙ ... œœœ ... ‰ œœœ ˙˙˙ ... œœœœ ‰ ˙˙˙˙ ˙˙˙ œœœœ ‰
˙. œ. œ ˙. ˙
HS: a

? b c ˙˙˙ œœ œœ
œ ‰ œ
œœ .
œ ..
˙
‰ ˙˙
œœ œœ œœ . ˙
œ ‰ œ œ .. ‰ n n # 43 ˙˙
œœ ˙˙ ..
œ ‰ ˙.
˙˙
˙
œœ ˙˙ ..
œ ‰ ˙.
˙˙
˙
œœ
œ ‰
HS: b b J J J J J
œ œ œ œ œ œ œ œ œ œ
? b c œ œ œ œ œ n n # 43 Œ œ
œ Œ œ Œ œ œ Œ œ œ j‰
Arp: ∆t b œ
C m7 G6

? b c ‰ n n # 43 ˙ . j
Ped: ab∆t b ˙. ˙. ˙. ˙ œ ‰
w w ˙ œ.

Figure 2.4: The natural nucleus of a musical score combining all the groupings. These rep-
resent the melody (M, upper staff, grouped by the resultant r) the harmonic structures (HS,
grouped by both a and b), the arpeggios (Arp, based on the smallest time unit) and the pedal
point (Ped, bottom staff, grouped by the common product ab). On the left we see 4 ÷ 3
grouped by a = 4, TM = 4∆t, on the right we see 5 ÷ 3 grouped by b = 3, TM = 3∆t (the time
unit is the quarter note t = 14 ). Note the use of rests between attacks, reducing the attack
durations.

Copy issued to Seth Dempsey ([email protected]) 23

CHAPTER 2. INTERFERENCE OF PERIODICITIES

24 Copy issued to Seth Dempsey ([email protected])

Chapter 3

The techniques of grouping

This chapter focusses on the grouping of the resultant ra÷b . An overview in musical notation
is presented for the practical time signatures listed in Table 2.1 for the non-uniform binary
generator combinations in the synchronization process.1 We briefly repeat the grouping op-
tions from Section 2.2 of the interference resultant ra÷b (a is the major, b the minor generator
for binary synchronization). Here we will consider the regular grouping options only:

1. Grouping by the common product ab. A single measure contains ab time units, i.e.,
TM = ab∆t, and this wll determine the meter. Use this time signature only for reason-
able values of ab (the practical limit is ab < 15).

2. Grouping by the major generator a (previously called superimposition by a). The frag-
ment contains b measures with length TM = a∆t, and we will obtain the rhythmic
effect of syncopation.

3. Grouping by the minor generator b. Now we get a measures with length TM = b∆t.

The grouping of the 19 combinations from Table 2.1 is shown in musical notation in
Fig. 3.1 and 3.2. We use the time signatures from [3], p. 14 in Schillinger’s book.
Verify that grouping by the common product is only shown for values ab < 15; time
signatures for larger values will hamper reading of the musical notation. Note the rhythmic
symmetry about the middle of the series of durations. The following pairs of generators
generate two note lengths with ratio 2:1 (half note and quarter note, or quarter note and 8th
note): 3:2, 5:2 and 7:2. We see three note duration values for the pairs: 4:3, 5:3 and 7:3. The
range of note durations therefore is determined by the minor generator b, as was already
discussed in Section 2.1.2.
The differences in durations (short-long) are maximum at either beginning or end of the
series, with more even durations in the middle (this was already noted in Section 2.1.2). For
grouping by either the major or the minor generator, check the number of measures and note
that there are no slurred notes across bar lines.

1
Time signature and meter are considered synonyms in this text, as should was stated in the introduction in
Section 1.2.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 3. THE TECHNIQUES OF GROUPING

3:2 (ab) (a) (b)

ã 68 œ œ œ œ
J J 43 œ œ œ œ ˙ œ œ ˙ 42 ˙ œ œ ˙

4:3 (ab) (a) (b)

ã 8 œ . Jœ œ œ œJ œ . 43 œ . œ œ œ œ œ . c ˙ .
12 œ ˙ ˙ œ ˙. 3
4 ˙. œ ˙ ˙ œ ˙.

5:2 (ab) (a) (b)

10 5 2
ã 8 œ œ Jœ Jœ œ œ 4˙ ˙ œ œ ˙ ˙ 4˙ ˙ œ œ ˙ ˙

5:3 (ab) (a) (b)

15 5 3
ã 8 œ . œ Jœ œ . œJ œ œ . 4 ˙ . ˙ œ ˙. œ ˙ ˙. 4 ˙. ˙ œ ˙. œ ˙ ˙.

5:4 (a) (b)

5
ã4 ˙ œ ˙. ˙ ˙ ˙. œ ˙ c˙ œ ˙. ˙ ˙ ˙. œ ˙

6:5 (a)
ã 8 œ . œ Jœ œ . œJ œ œ . œ . œ Jœ œ . œJ œ œ . 43 ˙
6 œœ ˙ œ œ œœ œ œ ˙ œœ ˙

(b)
ã 45 ˙ . ˙ œ˙ ˙ ˙. ˙. ˙ ˙ œ ˙ ˙.

7:2 (ab) (a) (b)

14 7 œ œ œ œ œœ œ œ 2
ã 8 œ œ œ Jœ Jœ œ œ œ 8 J J 4˙ ˙ ˙ œ œ ˙ ˙ ˙

7:3 (a) (b)

ã 78 œ . œ . œJ œ œ . œ Jœ œ . œ .
3
4 ˙. ˙. œ ˙ ˙. ˙ œ ˙. ˙.

7:4 (a) (b)

ã 78 ˙ œ . œJ œ œ œ œ œ œ Jœ œ . ˙ c˙ ˙. œ ˙ ˙ ˙ ˙ œ ˙. ˙

7:5 (a) (b)

ã 78 œ . œ œ œ . ˙ œœ
J J
œ .œ ˙ œ . œ œ œ . 45 ˙ . ˙ ˙ ˙ . ˙ œ ˙. ˙ œ ˙ ˙. ˙ ˙. ˙

7:6 (a)
ã 78 ˙ œ œ œ. œ œ ˙
J
œ. œ. ˙ œ œ œ. œ œ ˙
J
7:6 (b)
6 3
ã 8 ˙. œ œ œ. œ œ œ. œ. œ. œ. œ œ œ. œ œ ˙.
J J J J 4 ˙. œœ ˙ œ ˙ œ œœ œ ˙ œ ˙ œ œ ˙ .

Figure 3.1: Grouping of non-uniform interference patterns

26 Copy issued to Seth Dempsey ([email protected])

 8:3 (a) (b)
ã 88 œ . œ . œ Jœ œ . œ . œJ œ œ . œ . 43 ˙ . ˙. ˙ œ ˙. ˙. œ ˙ ˙. ˙.

8:5 (a)
8
ã 8 ˙ œJ œ . œ œ œ . Jœ ˙ ˙ œ œ. œ œ œ. œ ˙
J J
(b)
5
ã 4 ˙. ˙ ˙. ˙ ˙. ˙ œ ˙ ˙ œ ˙. ˙ ˙ ˙. ˙. ˙

8:7 (a)
8
ã8 ˙ œ. œ ˙.
J
œ ˙ œ œ.
J
˙ ˙ œ. œ ˙
J
œ ˙. œ œ. ˙
J
(b)
7
ã8 ˙ œ. œ œ. œ.
J
œ œ œ. œ. ˙ ˙ œ. œ. œ œ ˙ œ œ ˙
J
œ.

9:2 (ab) (a) (b)

18 9 2
ã 8 œ œ œ œ Jœ Jœ œ œ œ œ 8 œ œ œ œ Jœ œJ œ œ œ œ 4 ˙ ˙ ˙ œ œ ˙ ˙ ˙

9:4 (a) (b)

9
ã 8 ˙ ˙ œJ œ . œJ œ . œ œ œ œ œ . œJ ˙ ˙ c ˙ ˙ œ ˙. ˙ ˙ ˙ ˙ ˙. œ ˙ ˙

9:5 (a)
9
ã 8 œ. œ ˙ œ œ œ. œ. œ œ. œ œ œ. œ œ. œ ˙ œ œ.
J J
(b)
5
ã 4 ˙. ˙ ˙ œ ˙. ˙ ˙. ˙ ˙ ˙. ˙. ˙ œ ˙ ˙. ˙

9:7 (a)
ã 98 ˙ œ. œ œ. œ ˙ œ. ˙. œ œ œ. œ œ ˙.
J J
œ. ˙ œ. œ œ œ. ˙

(b)
ã 78 ˙ œ. œ œ œ. ˙ œ. ˙ œ œ ˙
J
œ. œœ ˙
J
œ. ˙ œ. œ œ ˙ œ.

9:8 (a)
ã 98 ˙ . œ œ ˙.
J
œ œ ˙.
J
œ. œ. œ œ œ. œ. œ œ œ. œ. ˙.
J J
œ œ ˙.
J
œ œ ˙.
J
(b)
8
ã8 ˙ œ œ ˙. œ ˙. œ. œ ˙
J
˙ ˙ ˙ œ œ. ˙.
J
œ ˙ œ. œ ˙
J

Figure 3.2: Grouping of non-uniform interference patterns (cont’d)

Copy issued to Seth Dempsey ([email protected]) 27

CHAPTER 3. THE TECHNIQUES OF GROUPING

28 Copy issued to Seth Dempsey ([email protected])

Chapter 4

The techniques of fractioning

This chapter introduces another basic technique for generating attack series with two gen-
erators a and b, called fractioning and notated as a ÷ b. The interference and the grouping
process are discussed.

4.1 The process of fractioning

The process of fractioning is also based on combining two generators, the major generator a
and the minor generator b (two integer numbers, with a > b). Again, we will use the analogy
of ticking clocks or metronomes. For the fractioning technique we will use one metronome
of type A, ticking at intervals of ∆tA time units, and using Nb metronomes of type B, each N
generating a ticks at a time interval of ∆tB time units, with

Nb = a − b + 1. (4.1)

We will synchronize these metronomes by starting the first B metronome at the same
time instant as metronome A. At each subsequent tick of metronome A we start another
metronome B, until all Nb metronomes are ticking. The resultant attack series t~r is de-
termined by the combination, the union, of all metronomes (compare this with Eq. 2.1 for
binary synchronization)

~tr = ~tA ∪ ~tB1 ∪ ~tB2 ∪ . . . ∪ ~tBN , (4.2)

b

and the duration series will repeat itself after Tr = (∆tA )2 time units.
The combinations for fractioning are listed in Table 2.1; here we will consider a number
of examples and determine the resultant ra÷b . Not all combinations from the table will be
considered here. See the example in Section 4.2 for the musical notation. A summary of the
characteristics of the fractioning patterns is found in the lower half of Table 2.2.

4.1.1 Fractioning group: 3 and 2

For a = 3 and b = 2, the interference pattern 3 ÷ 2 will repeat after Tr = a2 = 9 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 3 − 2 + 1 = 2.
Each B clock will generate a = 3 ticks. The process is shown in Fig. 4.1.a. The attack series

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 4. THE TECHNIQUES OF FRACTIONING

for the A and B generators are

~tA = [0 3 6],
~tB1 = [0 2 4],
~tB2 = [3 5 7].

Applying Eq. 4.2 the resultant attack and duration series are

~tr = [0 2 3 4 5 6 7],
r = 2 + 1 + 1 + 1 + 1 + 1 + 2,

i.e., seven attacks (Na = 7) and two note duration values (in length ratio 2:1). Note the
symmetry of the attack pattern about the centre.

4.1.2 Fractioning group: 4 and 3

For a = 4 and b = 3, the interference pattern 4 ÷ 3 will repeat after Tr = a2 = 16 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 4 − 3 + 1 = 2.
Each B clock will generate a = 4 ticks. The attack series for the A and B generators are

~tA = [0 4 8 12],
~tB1 = [0 3 6 9],
~tB2 = [4 7 10 13],

as shown in Fig. 4.1.b.

Applying Eq. 4.2 the resultant attack and duration series are

~tr = [0 3 4 6 7 8 9 10 12 13],
r = 3 + 1 + 2 + 1 + 1 + 1 + 1 + 2 + 1 + 3,

i.e., 10 attacks (Na = 10) and three note duration values. There are more shorter notes in the
centre of the attack pattern.

4.1.3 Fractioning group: 5 and 2

For a = 5 and b = 2, the interference pattern 5 ÷ 2 will repeat after Tr = a2 = 25 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 5 − 2 + 1 = 4.
Each B clock will generate a = 5 ticks. The attack series for the A and B generators are

~tA = [0 5 10 15 20],
~tB1 = [0 2 4 6 8],
~tB2 = [5 7 9 11 13],
~tB3 = [10 12 14 16 18],
~tB4 = [15 17 19 21 23],

as shown in Fig. 4.1.c.

30 Copy issued to Seth Dempsey ([email protected])

 4.1. THE PROCESS OF FRACTIONING

b bbbbbb r b bb bbbbb bb r
b b b B2 b b b b B2
b b b B1 b b b b B1
b b b A b b b b A
- -
0 5 10 t 0 5 10 15 t
(a) r3÷2 (b) r4÷3

b b bbbbbbbbbbbbbbbbbb b r
b b b b b B4 b b bb bbbbbbbbbb bb b r
b b b b b B3 b b b b b B3
b b b b b B2 b b b b b B2
b b b b b B1 b b b b b B1
b b b b b A b b b b b A
- -
0 5 10 15 20 25 t 0 5 10 15 20 25 t
(c) r5÷2 (d) r5÷3

b bb bbb bb bbb bb r
b b b b b B2
b b b b b B1
b b b b b A
-
0 5 10 15 20 25 t
(e) r5÷4

b bb bbb bb b bb bbb bb r
b b b b b b B2
b b b b b b B1
b b b b b b A
-
0 5 10 15 20 25 30 35 t
(f) r6÷5

Figure 4.1: Fractioning. The resultant ra÷b (top row) of two generators A and B that tick at
different time intervals. a): ∆tA = 3, ∆tB = 2, b): ∆tA = 4, ∆tB = 3, c): ∆tA = 5, ∆tB = 2,
d): ∆tA = 5, ∆tB = 3, e): ∆tA = 5, ∆tB = 4, f): ∆tA = 6, ∆tB = 5 time units.

Copy issued to Seth Dempsey ([email protected]) 31

CHAPTER 4. THE TECHNIQUES OF FRACTIONING

Applying Eq. 4.2 the resultant attack and duration series are
~tr = [0 2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 23],
r = 2 + 2 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 2 + 2,

i.e., Na = 21 and two different note durations. Again the note length shows the tendency to
increase from the pattern centre towards both ends.

4.1.4 Fractioning group: 5 and 3

For a = 5 and b = 3, the interference pattern 5 ÷ 3 will repeat after Tr = a2 = 25 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 5 − 3 + 1 = 3.
Each B clock will generate a = 5 ticks. The attack series for the A and B generators are
~tA = [0 5 10 15 20],
~tB1 = [0 3 6 9 12],
~tB2 = [5 8 11 14 17],
~tB3 = [10 13 16 19 22],

as shown in Fig. 4.1.d.

Applying Eq. 4.2 the resultant attack and duration series are
~tr = [0 3 5 6 8 9 10 11 12 13 14 15 16 17 19 20 22],
r = 3 + 2 + 1 + 2 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 2 + 1 + 2 + 3,

i.e., Na = 17 and three different note lengths.

4.1.5 Fractioning group: 5 and 4

For a = 5 and b = 4, the interference pattern 5 ÷ 4 will repeat after Tr = a2 = 25 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 5 − 4 + 1 = 2.
Each B clock will generate a = 5 ticks. The attack series for the A and B generators are
~tA = [0 5 10 15 20],
~tB1 = [0 4 8 12 16],
~tB2 = [5 9 13 17 21],

as shown in Fig. 4.1.e.

Applying Eq. 4.2 the resultant attack and duration series are
~tr = [0 4 5 8 9 10 12 13 15 16 17 20 21],
r = 4 + 1 + 3 + 1 + 1 + 2 + 1 + 2 + 1 + 1 + 3 + 1 + 4,

i.e., Na = 13 and four different note durations.

In Section 2.1.2 we already discussed the effect of the value of the minor generator b on
the resultant r; this effect is also present in the fractioning process. Compare the patterns for
r5÷2 to r5÷4 and note that the range of note durations is determined by the value of b.

32 Copy issued to Seth Dempsey ([email protected])

 4.1. THE PROCESS OF FRACTIONING

4.1.6 Fractioning group: 6 and 5

For a = 6 and b = 5, the interference pattern 6 ÷ 5 will repeat after a2 = 36 time units and
we determine the number of B minor generators using Eq. 4.1; this yields Nb = 6 − 5 + 1 = 2.
Each B clock will generate a = 6 ticks. The attack series for the A and B generators are

~tA = [0 6 12 18 24 30],
~tB1 = [0 5 10 15 20 25],
~tB2 = [6 11 16 21 26 31],

as shown in Fig. 4.1.f.

Applying Eq. 4.2 the resultant attack and duration series are

~tr = [0 5 6 10 11 12 15 16 18 20 21 24 25 26 30 31],
r = 5 + 1 + 4 + 1 + 1 + 3 + 1 + 2 + 2 + 1 + 3 + 1 + 1 + 4 + 1 + 5,

i.e., Na = 16 and five note durations.

4.1.7 Fractioning group: 7 and 3

For a = 7 and b = 3, the interference pattern 7 ÷ 3 will repeat after Tr = a2 = 49 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 7 − 3 + 1 = 5.
Each B clock will generate a = 7 ticks. The attack series for the A and B generators are

~tA = [0 7 14 21 28 35 42],
~tB1 = [0 3 6 9 12 15 18],
~tB2 = [7 10 13 16 19 22 25],
~tB3 = [14 17 20 23 26 29 32],
~tB4 = [21 24 27 30 33 36 39],
~tB5 = [28 31 34 37 40 43 46].

This is shown in Fig. 4.2.a.

Applying Eq. 4.2 the resultant attack and duration series are

~tr = [0 3 6 7 9 10 12 13 14 15 16 17 18 19 20 21 22 23 24 25
26 27 28 29 30 31 32 33 34 35 36 37 39 40 42 43 46],
r = 3+3+1+2+1+2+1+1+1+1+1+1+1+1+1+1+1+1+1+
1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 2 + 1 + 2 + 1 + 3 + 3,

i.e., a long series of 37 attacks (Na = 37) with only three different note durations. Once, again,
it is the value of the minor generator b that determines the set length of note durations. Note
the risk of rhythmical monotony at the pattern centre, due to the 25 consecutive notes with
one time unit duration.

Copy issued to Seth Dempsey ([email protected]) 33

CHAPTER 4. THE TECHNIQUES OF FRACTIONING

b b bb bb bbbbbbbbbbbbbbbbbbbbbbbbbb bb bb b r
b b b b b b b B5
b b b b b b b B4
b b b b b b b B3
b b b b b b b B2
b b b b b b b B1
b b b b b b b A
-
0 5 10 15 20 25 30 35 40 45 50 t
(a) r7÷3

b b bb bb bbb bbbbbbbbbbbbbb bbb bb bb b r

b b b b b b b B4
b b b b b b b B3
b b b b b b b B2
b b b b b b b B1
b b b b b b b A
-
0 5 10 15 20 25 30 35 40 45 50 t
(b) r7÷4

b b b b b bb b bbbb bb bbbb b bb b b b b r
b b b b b b b B3
b b b b b b b B2
b b b b b b b B1
b b b b b b b A
-
0 5 10 15 20 25 30 35 40 45 50 t
(c) r7÷5

b bb bbb bb b bb b bb bbb bb r
b b b b b b b B2
b b b b b b b B1
b b b b b b b A
-
0 5 10 15 20 25 30 35 40 45 50 t
(d) r7÷6

Figure 4.2: Fractioning (cont’d). The resultant ra÷b of two generators A and B that tick at
different time intervals. a): ∆tA = 7, ∆tB = 3, b): ∆tA = 7, ∆tB = 4, c): ∆tA = 7, ∆tB = 5,
d): ∆tA = 7, ∆tB = 6 time units.

34 Copy issued to Seth Dempsey ([email protected])

 4.1. THE PROCESS OF FRACTIONING

Application Tip:
Let’s revisit the aspect of rhythmic monotony that the fractioning resultant r7÷3
contained in the middle (the 25 repeated single time unit duration notes). As-
sume that we apply this attack series to a single percussion instrument. Then
indeed the result would sound boring and create a hammering rhythm. We have
achieved a longer duration coordinated rhythm through a simple two-generator
scheme, but it does not seem to make much sense musically.
However, have another look at Fig. 4.2, use a more liberal approach and re-
consider the original idea of the five beating metronomes B1 , . . . , B5 (the minor
generator). Now there’s a nice opportunity opening up: use five woodblocks,
boobams or a set of toms to play the separate ~tBi patterns over a ~tA bass ostinato
(another percussion instrument or a pitched instrument). Or how about a five-
pitch note pattern for a melody or harmony instrument? Assign a single pitch
1
to each B metronome. Next, consider a 16th note time unit t = 16 and use these
B metronome rhythms to create arpeggio (monophonic) synthesizer patterns; re-
member the 1980s?
One difficulty remains though: the grouping of this attack series into a regular
meter requires multiple repeats before we achieveh irecurrence. The pattern lends
itself more to an irregular meter grouping, e.g., 78 , as found in Balkan or Arabic
folk music. Well, there is also the option of changing the meter along the pattern.1
Or, maybe better, consider the fractioning with the major generator a = 8; see the
resultants of r8÷n below.

4.1.8 Fractioning group: 7 and 4

For a = 7 and b = 4, the interference pattern 7 ÷ 4 will repeat after Tr = a2 = 49 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 7 − 4 + 1 = 4.
Each B clock will generate a = 7 ticks. The attack series for the A and B generators are

~tA = [0 7 14 21 28 35 42],
~tB1 = [0 4 8 12 16 20 24],
~tB2 = [7 11 15 19 23 27 31],
~tB3 = [14 18 22 26 30 34 38],
~tB4 = [21 25 29 33 37 41 45],

as shown in Fig. 4.2.b.

Applying Eq. 4.2 the resultant attack and duration series are

~tr = [0 4 7 8 11 12 14 15 16 18 19 20 21 22 23 24
25 26 27 28 29 30 31 33 34 35 37 38 41 42 45],
r = 4+3+1+3+1+2+1+1+2+1+1+1+1+1+1+
1 + 1 + 1 + 1 + 1 + 1 + 1 + 2 + 1 + 1 + 2 + 1 + 3 + 1 + 3 + 4,
1
The approach suggested here is in fact the subject of later chapters in the theory of rhythm, or discussed in
other books in the Schillinger System.

Copy issued to Seth Dempsey ([email protected]) 35

CHAPTER 4. THE TECHNIQUES OF FRACTIONING

i.e., Na = 31 and four different note lengths. Here also the pattern centre contains 13 notes
with unit duration; as a standalone pattern this yields rhythmic monotony.

4.1.9 Fractioning group: 7 and 5

For a = 7 and b = 5, the interference pattern 7 ÷ 5 will repeat after Tr = a2 = 49 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 7 − 5 + 1 = 3.
Each B clock will generate a = 7 ticks. The attack series for the A and B generators are

~tA = [0 7 14 21 28 35 42],
~tB1 = [0 5 10 15 20 25 30],
~tB2 = [7 12 17 22 27 32 37],
~tB3 = [14 19 24 29 34 39 44],

as shown in Fig. 4.2.c.

Applying Eq. 4.2 the resultant attack and duration series are

~tr = [0 5 7 10 12 14 15 17 19 20 21 22 24
25 27 28 29 30 32 34 35 37 39 42 44],
r = 5+2+3+2+2+1+2+2+1+1+1+2+
1 + 2 + 1 + 1 + 1 + 2 + 2 + 1 + 2 + 2 + 3 + 2 + 5,

i.e., Na = 25 and four different note lengths. Note that the duration of 4∆t time units is
missing from the series.

4.1.10 Fractioning group: 7 and 6

For a = 7 and b = 6, the interference pattern 7 ÷ 5 will repeat after Tr = a2 = 49 time units.
We determine the number of B minor generators using Eq. 4.1; this yields Nb = 7 − 6 + 1 = 2.
Each B clock will generate a = 7 ticks. The attack series for the A and B generators are

~tA = [0 7 14 21 28 35 42],
~tB1 = [0 6 12 18 26 32 38],
~tB2 = [7 13 19 25 31 37 43],

as shown in Fig. 4.2.d.

Applying Eq. 4.2 the resultant attack and duration series are

~tr = [0 6 7 12 13 14 18 19 21 24
25 28 30 31 35 36 37 42 43],
r = 6+1+5+1+1+4+1+2+3+
1 + 3 + 2 + 1 + 4 + 1 + 1 + 5 + 1 + 6,

i.e., Na = 19 and six different note lengths.

36 Copy issued to Seth Dempsey ([email protected])

 4.2. GROUPING

4.2 Grouping
Grouping is the selection of the meter or time signature; the resultant will be divided into a
number of measures with each measure containing TM time units.
In the case of fractioning there are again three options for grouping:

1. The resultant attack pattern repeats itself after Tr = (∆tA )2 time units. So, grouping
by a2 implies
TM = (∆tA )2 = a2 ∆t. (4.3)
The fractioning pattern consists of a single measure. Like the case for binary synchro-
nization (see Section 2.2), there is an upper limit to this grouping technique: do not
group the resultant by a2 when TM > 15∆t.

2. Superimposition of a means that the major generator A will determine the time signa-
ture
TM = ∆tA = a ∆t. (4.4)
The pattern length now is a measures.

3. Superimposition of b means that the minor generator B will determine the time signa-
ture
TM = ∆tB = b ∆t. (4.5)
This yields syncopated rhythms for the fractioning attack series. When repeating the
grouped resultant attack series until stopping at a complete measure, we have achieved
recurrence, i.e., the attack series is completed at the end of a full measure, just before
the bar line. The pattern now contains a higher total number of measures, compared
to the previous case (superposition of a).

Example 4.1
Grouping of fractioning patterns.
Several groupings of the resultant fractioning patterns discussed in Section 4.1
are shown in musical notation in Fig. 4.3. The smallest time unit is either the 8th
or the 4th note duration (t = 18 or 14 ).

• Case 1: r3÷2 . Note that the grouping by either a2 (TM = a2 ∆t) or a (TM =
a∆t) again leads to symmetrical attack patterns. Grouping the resultant by
b (TM = b∆t) leads to syncopated rhythm patterns; the end of each resultant
pattern is indicated by the breathing sign (0 ).
• Case 2: r5÷2 . Only the superimposition by a (TM = a∆t) is shown. The
pattern is a = 5 measures long.
• For Case 3: r5÷3 , Case 4: r5÷4 and Case 5: r6÷5 there is grouping by either
the major generator a (TM = a∆t) or the minor generator b (TM = b∆t).

Copy issued to Seth Dempsey ([email protected]) 37

CHAPTER 4. THE TECHNIQUES OF FRACTIONING

,
ã 98 œ œJ œ œ œ œJ œ 43 ˙ œ œ œ œ œ ˙ 42 ˙
frac(3:2), (a^2) (a) (b)
œ œ œ œ œœ œœ œ œ œ œ œ œ ˙

ã 85 œ œ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œJ œ œ œ. œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ.
frac(5:2), (a) frac(5:3), (a)

J JJ JJJJJ JJ J

3 ,
ã 8 œ.
(b)
œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ
J J J J J J

,
ã œ œ œ œ œ œ œ œJ œ œ œ œ œ œ œ œ œ œœœ œœœ œœœ œœ œœ œ.
J J J J J

,
ã 85 ˙ œ œ. œ œ œ œ œ œ œ œ. 42 ˙ œ œ. œ œ œ œ œ œ œ œ. œ œ. œ œ.
frac(5:4), (a) (b)
œ˙
J J J J J J J J J

, ,
ã œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ . œJ

ã œ . œJ œ . œJ œJ œ œJ œ œ œ œ . œJ ˙ 68 œ . œ œ ˙ œ œ œ. œ œ œ œ œ. œ œ ˙ œ œ œ.
frac(6:5), (a)

J J J J

,
ã 85 œ . œ œJ ˙ œ œ œ. œ œ œ œ œ. œ œ ˙ œ œ œ. œ œ œ œ œ œ œ œ œ œ œ.
(b)
œ˙ œ˙
J J J J J J J

, ,
ã œ œ œ . œ œ œ . œ œ . œ œJ œ œ œ œ œ œ œJ œ œ œJ œ œ œ œ œ œ œJ œ œ . œ œ . œ œ œ . œ œ

,
ã œ . œ œ œJ œ œ œ œ œ œ œ œ . œ œ ˙ œ ˙
J
œ ˙
J
œ œ œ. œ œ œ œ œ. œ œ ˙
J J J J
œ ˙ œ
J J

Figure 4.3: Grouping of fractioning patterns

38 Copy issued to Seth Dempsey ([email protected])

Chapter 5

Composition of groups by pairs

In this chapter we will see the process of composing rhythmic resultants in pairs. Based on
the interference of two generators A and B with different tick intervals a∆t and b∆t pairs
of adjacent groups will be formed. The two members in the group are generated with the
techniques described in Section 2.1 (binary synchronization) and Section 4.1 (fractioning).
Binary synchronization yields the resultant ra÷b , fractioning generates the resultant ra÷b .
The pairing of resultants is a method to generate a longer rhythmic continuity. The uni-
fying element is the set of two generators A and B, that determine the note durations in the
adjacent group. There are three approaches to the combination of these resultants, labeled
as balance, expansion, and contraction. For all three approaches we will demonstrate measure
grouping by a time units only.

5.1 Balancing adjacent groups

A balanced pairing is given by

rB (a, b) = ra÷b + ra÷b + a(a − b). (5.1)

In fact we are juxtaposing three sub-patterns with length a2 , ab and a(a − b) time units,
respectively. The total duration of this balanced adjacent group therefore is

Tr = a2 + ab + a(a − b) = a2 + ab + a2 − ab = 2a2 ,

balancing the first half of the pattern with an equally long second half. The result is that
the duration symmetry about the pattern centre will be disturbed; this may be considered
an improvement to the rhythm. There is also less obvious risk of short note rhythmical
monotony in the pattern centre.
The balancing resultant sounds unnatural when a ≥ 2b, as is the case in generator com-
binations such as (a, b) = (5, 2) or (9, 4). In that case a balanced pairing is achieved by using

rB (a > mb) = ra÷b + m ra÷b + (a2 − mab), (5.2)

where m is an integer number determined by the next lower integer rounding of the ratio
a2 /(ab). This can be explained as follows: the total duration of the balanced adjacent group
must be Tr = 2a2 time units. The first half is the fractioning resultant ra÷b ; therefore we will

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 5. COMPOSITION OF GROUPS BY PAIRS

determine the number of times that the entire non-uniform binary synchronization resultant
ra÷b fits into the second half; this is the integer number m. The remaining duration (a2 −mab)
in the second half is the last element in this rhythmical pattern.

Example 5.1
Balancing resultant: rB (a, b).
For the following examples we will present the duration series, using Eq. 5.1
or Eq. 5.2, and the results from Section 2.1.2 (interference through non-uniform
binary synchronization) and Section 4.1 (fractioning). The resultants are shown
in musical notation in Fig. 5.1.

• Generator time units a = 3, b = 2. The ratio a/b = 3/2 = 1.5; therefore the
adjacent group pair is determined with Eq. 5.1.

rB (3, 2) = r3÷2 + r3÷2 + 3(3 − 2)

= (2 + 1 + 1 + 2) + (2 + 1 + 1 + 1 + 1 + 1 + 2) + 3,

i.e., a total of 12 attacks (Na = 12), repeating after Tr = 2a2 = 2 × 9 =

18 time units. Looking at the pattern we may note that the second half
suggests a variation of the first group, with some developmental character
and concluding with a sustained note. The note durations in all elements
but the last are the same, i.e., {∆t, 2∆t}, supporting the homogeneity of the
adjacent group.
• Generator time units a = 4, b = 3. The ratio a/b = 4/3 = 1.33; therefore this
adjacent group pair is also determined with Eq. 5.1.

rB (4, 3) = r4÷3 + r4÷3 + 4(4 − 3)

= (3 + 1 + 2 + 2 + 1 + 3) +
+(3 + 1 + 2 + 1 + 1 + 1 + 1 + 2 + 1 + 3) + 4,

i.e., a total of 17 attacks (Na = 17), repeating after Tr = 2a2 = 2 × 16 = 32

time units.
• Generator time units a = 5, b = 3. Since a/b = 1.67 once again we use Eq. 5.1
for calculating the balanced adjacent group attack series.

rB (5, 3) = r5÷3 + r5÷3 + +5(5 − 3)

= (3 + 2 + 1 + 2 + 1 + 1 + 1 + 1 + 1 +
+1 + 1 + 1 + 1 + 2 + 1 + 2 + 3) +
+(3 + 2 + 1 + 3 + 1 + 2 + 3) + 10,

i.e., Na = 25, repeating after Tr = 2a2 = 2 × 25 = 50 time units. Since the

ratio a/b = 5/3 = 1.67 we begin to notice the unbalanced long last element
in the pattern with duration 10∆t.
• Generator time units a = 5, b = 2. Since the ratio a/b = 5/2 = 2.5 > 2,
we will generate the adjacent group with Eq. 5.2, the alternative approach
to balancing the adjacent groups. Determine the value of m from the ratio

40 Copy issued to Seth Dempsey ([email protected])

 5.2. EXPANDING ADJACENT GROUPS

a2 /(ab) = 25/(5 × 2) = 25/10 = 2.5; rounding this to the next lower integer
number yields m = 2, a double statement of the two-generator interference
pattern. Therefore the balancing pattern is
rB (5, 2) = r5÷2 + 2r5÷2 + (25 − 2 × 10)
= (2 + 2 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 +
+1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 2 + 2) +
+(2 + 2 + 1 + 1 + 2 + 2) + (2 + 2 + 1 + 1 + 2 + 2) + 5,
i.e.,Na = 35 repeating after Tr = 2a2 = 2 × 25 = 50 time units. The length
of the sustained note at the end is acceptable, compared to the maximum
duration in the synchronization resultant: this ratio is 5/2 = 2.5.
• Finally, we demonstrate the generator time units a = 7, b = 3. The ratio
a/b = 7/3 = 2.335 > 2, leading to the application of Eq. 5.2. Determine the
value of m from the ratio a2 /(ab) = 49/(7 × 3) = 49/21 = 2.33; rounding
this to the next lower integer number yields m = 2. Therefore the balanced
adjacent pair rhythmical pattern is
rB (7, 3) = r7÷3 + 2r7÷3 + (49 − 2 × 21)
= (3 + 3 + 1 + 2 + 1 + 3 + 1 + 1 + 2 + 2 + 1 +
+1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 +
+1 + 2 + 2 + 1 + 1 + 3 + 1 + 2 + 1 + 3 + 3) +
+(3 + 3 + 1 + 2 + 3 + 2 + 1 + 3 + 3) +
+(3 + 3 + 1 + 2 + 3 + 2 + 1 + 3 + 3) + 7,
i.e., Na = 52, repeating after Tr = 2a2 = 2 × 49 = 98 time units. The
sustained note at the end has a duration of 7 time units.

5.2 Expanding adjacent groups

The expanding pairing of adjacent groups is given by
rE (a, b) = ra÷b + ra÷b , (5.3)
i.e., the addition (juxtaposition) of two sub-patterns of non-equal length. The total length is
Tr = ab + a2 . Since a is the major generator time interval with a > b, and therefore ab < a2
the second part of the adjacent group is longer than the first. This is the reason for labeling
such a pair as expanding. The length ratio of the two parts obviously is a2 /(ab) = a/b, i.e.,
the ratio of the major to minor generator time interval.

Example 5.2
Expanding resultant: rE (a, b).
For the following examples we will present the duration series, using Eq. 5.3, and
the results from Section 2.1.2 and Section 4.1. The resultants are shown in musical
notation in Fig. 5.1.

Copy issued to Seth Dempsey ([email protected]) 41

CHAPTER 5. COMPOSITION OF GROUPS BY PAIRS

• Generator time units a = 3, b = 2.

rE (3, 2) = r3÷2 + r3÷2

= (2 + 1 + 1 + 2) + (2 + 1 + 1 + 1 + 1 + 1 + 2),

i.e., Na = 11, repeating after Tr = ab + a2 = 3 × 2 + 9 = 6 + 9 = 15 time

units. The length ratio of the two groups in the pair is a/b = 3/2 = 1.5.
• Generator time units a = 5, b = 3.

rE (5, 3) = r5÷3 + r5÷3

= (3 + 2 + 1 + 3 + 1 + 2 + 3) +
+(3 + 2 + 1 + 2 + 1 + 1 + 1 + 1 + 1 +
+1 + 1 + 1 + 1 + 2 + 1 + 2 + 3),

i.e., Na = 24, repeating after Tr = ab + a2 = 5 × 3 + 25 = 15 + 25 = 40 time

units. The note durations are the same in both parts {∆t, 2∆t, 3∆t}. The
longer second part sounds like a rhythmical elaboration of the first part.

5.3 Contracting adjacent groups

A contracting pairing of adjacent groups is given by

rC (a, b) = ra÷b + ra÷b . (5.4)

The total length is Tr = a2 + ab, with the fractioning resultant in the first part longer than
the binary synchronization resultant. This aspect leads to the contracting character of the
adjacent group. It is a kind of mirrored counterpart of the expanding group.

Example 5.3
Contracting resultant: rC (a, b).
For the following examples we will present the duration series, using Eq. 5.4, and
the results from Section 2.1.2 and Section 4.1. The resultants are shown in musical
notation in Fig. 5.1.

• Generator time units a = 4, b = 3.

rC (4, 3) = r4÷3 + r4÷3

= (3 + 1 + 2 + 1 + 1 + 1 + 1 + 2 + 1 + 3) +
(3 + 1 + 2 + 2 + 1 + 3),

i.e., Na = 17, repeating after Tr = a2 + ab = 16 + 4 × 3 = 16 + 12 = 28

time units. The shorter second part suggests a rhythmical summary of the
fractioning resultant in the first part. The series of single time unit attacks in
the centre of the fractioning group has been eliminated in the second group.

42 Copy issued to Seth Dempsey ([email protected])

 5.3. CONTRACTING ADJACENT GROUPS

• Generator time units a = 5, b = 3.

rC (5, 3) = r5÷3 + r5÷3

= (3 + 2 + 1 + 2 + 1 + 1 + 1 + 1 + 1 +
+1 + 1 + 1 + 1 + 2 + 1 + 2 + 3) +
= (3 + 2 + 1 + 3 + 1 + 2 + 3),

i.e., a resultant containing 24 attacks (Na = 24) and repeating after Tr = a2 +

ab = 25 + 5 × 3 = 25 + 15 = 40 time units. The length ratio of the fractioning
to the interference part is a/b = 5/3 = 1.67. Again, the second group is
commenting on the fractioning pattern, skipping the single time unit attacks
in the centre and sort of replacing it by a single attack wth duration 3∆t.

Note from the examples, that:

• The second group of the resultant pair always is a varied version of the first group.
This creates both homogeneity and variation.

• In balancing (rB ) and contracting (rC ) the pair, there is more rhythmic activity, i.e.,
shorter note durations in the first half of the resultant. The balanced pairing concludes
with a sustained, long duration note.

• When the time interval difference between the two generators becomes large, the reg-
ular balanced adjacent group pairing (rB ) has unnatural characteristics, such as long
series of equal short notes in the first half and a very long closing note. This was noted
by Schillinger and corrected by the special case of balancing adjacent pairs.

• The expanding pair (rE ) has more rhythmic activity (shorter notes) in the second half
of the resultant.

Copy issued to Seth Dempsey ([email protected]) 43

CHAPTER 5. COMPOSITION OF GROUPS BY PAIRS

3 "˙ . c ˙. œ ˙.
r_B(3,2) r_B(4,3)
ã4 ˙ œ œ œ œ œ ˙ "˙ œ œ ˙ œ ˙ œ œ œ œ ˙ "

ã ˙. œ ˙. 85 œ . œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ. "
r_B(5,3)
œ ˙ ˙ "˙
J J J

ã œ. œ œ œ. œ œ œ. " œ. œ œ. œ 45 ˙
r_B(5,2)
˙ œ œ œ œ œ œ œ œ œ œ œ
J J

㜠œ œ œ œ œ ˙ ˙" ˙ ˙ œ œ ˙ ˙ ˙ ˙ œ œ ˙ ˙" ˙ œ

ã 78 œ . œ . œJ œ œJ œ . œJ œJ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œJ œJ œ . œ œJ œJ œ . œ . "
r_B(7,3)

ã œ. œ. œ œ œ.
J
œ œ œ.
J
œ. " œ. œ. œ œ œ.
J
œ œ œ.
J
œ. " ˙ œ.

ã 43 ˙
r_E(3,2)
œ œ ˙" ˙ œ œ œ œ œ ˙

5
ã 8 œ. œ œ. œ œ œ ." œ . œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ.
r_E(5,3)
œ œ
J J J J J

ã c ˙. œ ˙. " ˙. œ ˙.
r_C(4,3)
œ ˙ œ œ œ œ ˙ œ ˙ ˙

ã 85 œ . œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ. " œ. œ œ. œ œ œ.
r_C(5,3)
œ œ
J J J J J

Figure 5.1: Balancing (rB ), expanding (rE ) and contracting (rC ) a pair of adjacent groups.
These groups are based on A and B generator combinations, ticking at (a, b)∆t time inter-
vals.

44 Copy issued to Seth Dempsey ([email protected])

Chapter 6

Utilization of three or more generators

This chapter introduces a basic technique for generating attack series, by considering the
interference pattern of more than two clocks or metronomes that each tick at a different
interval. These attack series may then be grouped using different numbers of time units per
measure.
The three generators will form a family of rhythms when they are based on the same series
of growth, also called summation series or Fibonacci series, shown in Table 6.1.1
We will limit the combinations to practical sizes, and consider only the combinations
discussed in the next section.

6.1 The technique of synchronization of three generators

Now there are three clocks or metronomes A, B and C ticking at different regular time 3
intervals ∆tA , ∆tB and ∆tC . We assume that metronome C is ticking faster than metronome
B, and metronome B is ticking faster than metronome A. So we have ∆tC < ∆tB < ∆tA .
We will derive two resultants: r and the alternative r0 . The interference pattern will
repeat after Tr = ∆tC · ∆tB · ∆tA time units. The resultant r is determined analogous to the
case of binary synchronization (see Section 2.1.2), by finding the combination, the union, of
three attack series
~tr = ~tA ∪ ~tB ∪ ~tC . (6.1)
The alternative resultant r0 is determined by synchronising the complementary factors: this
is a set of three alternative generators, but now the metronomes tick at intervals of ∆tB · ∆tC
(the complementary factor of generator A), ∆tA ·∆tC (the complementary factor of generator
B), and ∆tA · ∆tB (the complementary factor of generator C).
We will see in the examples that resultant r yields series of more attacks and shorter
duration, whereas the alternative r0 , creates rhythmic patterns with less attacks and longer
durations. Both the r and r0 pattern have the same length.

6.1.1 Interference group: 5, 3 and 2

For ∆tA = 5, ∆tB = 3 and ∆tC = 2 time units and using Eq. 1.2 we get the attack series
~tA = [0 5 10 . . . 25], ~tB = [0 3 6 . . . 27], ~tC = [0 2 4 . . . 28],
1
The summation series will return in Chapter 13, where the evolution of rhythm families is discussed.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 6. UTILIZATION OF THREE OR MORE GENERATORS

Table 6.1: The summation series serving musical purposes. Each row in the table is a Fi-
bonacci summation series with the third and higher column number element being the sum
of the two previous elements.

1 2 3 5 8 13 ...
1 3 4 7 11 18 ...
1 4 5 9 14 23 ...

repeating itself after Tr = 2 · 3 · 5 = 30 time units. We determine the resultant r from the
combination of these attack series, the union according to Eq. 6.1, which yields the following
attack and duration series
~tr = [0 2 3 4 5 6 8 9 10 12 14 15 16 18 20 21 22 24 25 26 27 28],
r = 2+1+1+1+1+2+1+1+2+2+1+
1 + 2 + 2 + 1 + 1 + 2 + 1 + 1 + 1 + 1 + 2.
The resultant number of attacks is 22 (Na = 22), with two different note duration values.
The attack series is shown in Fig. 6.1.a. The resultant will be written as r5÷3÷2 (descending
order) or r2÷3÷5 (ascending order).
The resultant r0 is obtained with the complementary clocks, with ∆tA0 = 6, ∆tB 0 = 10
and ∆tC 0 = 15 and using Eq. 1.2 we get the attack series
~tA0 = [0 6 12 . . . 24], ~tB 0 = [0 10 20], ~tC 0 = [0 15],

repeating itself after Tr = 30 time units and shown in Fig. 6.1.b. The combination of these
attack series yields the following attack and duration series
~tr0 = [0 6 10 12 15 18 20 24],
0
r = 6 + 4 + 2 + 3 + 3 + 2 + 4 + 6.
This is a series of eight attacks (Na = 8), with four different note duration values.

6.1.2 Interference group: 7, 4 and 3

For ∆tA = 7, ∆tB = 4 and ∆tC = 3 time units and using Eq. 1.2 we get the attack series
~tA = [0 7 14 . . . 77], ~tB = [0 4 8 . . . 80], ~tC = [0 3 6 . . . 81],

repeating itself after Tr = 7 · 4 · 3 = 84 time units. We determine the resultant r from the
combination of these attack series, Eq. 6.1, which yields the following attack and duration
series
~tr = [0 3 4 6 7 8 9 12 14 15 16 18 20 21 24 27 28 30 32 33 35 36 39 40 42
44 45 48 49 51 52 54 56 57 60 63 64 66 68 69 70 72 75 76 77 78 80 81],
r = 3+1+2+1+1+1+3+2+1+1+2+2+1+3+3+1+2+2+1+2+1+
3+1+2+2+1+3+1+2+1+2+2+1+3+3+1+2+2+1+1+2+3+
1 + 1 + 1 + 2 + 1 + 3.

46 Copy issued to Seth Dempsey ([email protected])

 6.1. THE TECHNIQUE OF SYNCHRONIZATION OF THREE GENERATORS

b bbbbb bbb b bbb b bbb bbbbb r

b b b b b b b b b b b b b b b C
b b b b b b b b b b B
b b b b b b A
-
0 5 10 15 20 25 30 t
(a) r5÷3÷2

b b b b b b b b r0
b b C0
b b b B0
b b b b b A0
-
0 5 10 15 20 25 30 t
0
(b) r6÷10÷15

b bb bbbb b bbb b bb b bb b bb bb bb b bb bb bb b bb b bb b bbb b bbbb bb r

b b b b b b b b b b b b b b b b b b b b b b b b b b b b C
b b b b b b b b b b b b b b b b b b b b b B
b b b b b b b b b b b b A
-
0 10 20 30 40 50 60 70 80 t
(c) r7÷4÷3

b b b b b b b b b b b b r0
b b b C0
b b b b B0
b b b b b b b A0
-
0 10 20 30 40 50 60 70 80 t
0
(d) r12÷21÷28

Figure 6.1: Synchronization of three generators. The resultant ra÷b÷c and the alternative
complement resultant r0 of three generators A, B and C that tick at different time intervals.
a): ∆tA = 5, ∆tB = 3, ∆tC = 2, b): ∆tA0 = 6, ∆tB 0 = 10, ∆tC 0 = 15, c) ∆tA = 7, ∆tB =
4, ∆tC = 3, d) ∆tA0 = 12, ∆tB 0 = 21, ∆tC 0 = 28 time units.

Copy issued to Seth Dempsey ([email protected]) 47

CHAPTER 6. UTILIZATION OF THREE OR MORE GENERATORS

The resultant number of attacks Na = 48, with three different note duration values. The
attack series is shown in Fig. 6.1.c.
The alternative resultant r0 is obtained with the complementary clocks, with ∆tA0 =
12, ∆tB 0 = 21 and ∆tC 0 = 28 and using Eq. 1.2 we get the attack series

~tA0 = [0 12 24 . . . 72], ~tB 0 = [0 21 42 63], ~tC 0 = [0 28 56],

repeating itself after Tr = 84 time units, and shown in Fig. 6.1.d. The combination of these
attack series yields the following attack and duration series

~tr0 = [0 12 21 24 28 36 42 48 56 60 63 72],
0
r = 12 + 9 + 3 + 4 + 8 + 6 + 6 + 8 + 4 + 3 + 9 + 12.

This is a series of Na = 12 attacks, with six different note duration values.

6.2 Grouping
The previous section discussed the generation of a rhythm using three generators: A, B, and
C. The synchronization process yields two alternative attack-duration groups with resultant
r or r0 .
These rhythms can be grouped into measures in six alternative ways. The length of a
single measure is now based on either the generator time intervals {a, b, c} or the comple-
mentary factors {bc, ac, ab}. This means that a single measure may contain either TM =
∆tA , ∆tB , ∆tC , ∆tB × ∆tC , ∆tA × ∆tC , or TM = ∆tA × ∆tB time units. As before we will
limit the grouping to a practical maximum limit of TM ≤ 15∆t.
For the interference group 5÷3÷2 this yields grouping by either TM = 2, 3, 5, 6, 10 or 15
time units. The result is shown in musical notation in Fig. 6.2. For the interference group
7 ÷ 4 ÷ 3 this yields grouping by either TM = 3, 4, 7 or 12 time units, and the result is shown
in Fig. 6.3.

48 Copy issued to Seth Dempsey ([email protected])

 6.2. GROUPING

5
r(5:3:2), (a)
ã8 œ œ œ œ œ œ
J
œ œ œ œ œ
J
œ œ
J
œ œ œ œ œ œ œ œ œ
J

3
r(5:3:2), (b)
ã8 œ œ œ œ œ œ
J
œ œ œ
J J
œ œ œ œ
J J
œ œ œ œ
J J
œ œ œ œ œ
J

ã 28 œ
r(5:3:2), (c)
œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ

ã 68 œ
r(5:3:2), (bc)
œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ
J J J J J J J J

ã 10
r(5:3:2), (ac)
8 œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ

ã 15
r(5:3:2), (ab)
8 œ œ œ œ œ œ
J
œ œ œ
J J
œ œ
J
œ œ
J
œ œ œ œ
J J
œ œ œ œ œ
J

ã 85 œ . œ œJ ˙ œ œ . œ . œ ˙ œ œ œ . 38 œ . œ . œ . œ œ œ . œ . œ œ œ . œ . œ .
r'(5:3:2), (a) r'(5:3:2), (b)

J J J

ã 28 œ
r'(5:3:2), (c)
œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ
J J

ã 68 ˙ . œ. œ. œ. œ. ˙.
r'(5:3:2), (bc)
œ œ œ œ
J J

ã 10
15
8 ˙. œ œ. œ. œ ˙. 8 ˙. œ. œ œ œ. œ. œ œ œ. ˙.
r'(5:3:2), (ac) r'(5:3:2), (ab)
˙ ˙
J J

Figure 6.2: Grouping of the three-generator interference group 5 ÷ 3 ÷ 2. Time signatures

are determined by the (complementary) generator time intervals T = {a, b, c, bc, ac, ab}∆t (6
options), with a practical limit TM < 15∆t.

Copy issued to Seth Dempsey ([email protected]) 49

CHAPTER 6. UTILIZATION OF THREE OR MORE GENERATORS

ã 78 œ . œœ œ œœ œ . œ œœœœ œ œ . œ . œ œ œ œ œ œ œ . œœ œ œ œ . œ œ œœœœ œ œ . œ . œœœœœ œ œ . œœ œœ œ œ .

r(7:4:3), (a)

J J J J J J J J J J J J J J

4
ã 8 œ . œJ œ œ œ œJ œ . œ œ. œ. œ œ œ œ œ œ œ. œ
r(7:4:3), (b)
œ œœ œ œ
J J J J J

㜠œ œ œ. œ œ œ œ œ œ œ œ. œ. œ œ œ œ œ œ œ. œ œ œ œ œ œ.
J J J J J J J

3
ã 8 œ . œJ œ œ œ œ œ . œ œJ œJ œ œ œJ œ . œ . œJ œ œ œJ œ œJ œ . œJ œ œ œJ
r(7:4:3), (c)

ã œ. œœ
J
œœ
J
œ œ œ œ. œ. œœ
J
œ œ œœ
J J
œ. œ œ œ œ œ œ.
J

12
ã 8 œ . œJ œ œ œ œ œ . œ œ œ œ œ œ œ. œ. œ œ œ œ œ œ œ. œ œ œ œ œ.
r(7:4:3), (bc)

JJ J J J J J J

ã œJ œ œ œ œ œ œ œ. œ. œ œ œ œ œ œ œ. œ œ œ œ œ œ.
J J J J J

7 œ. œ. œ œ œ. ˙ œ. ˙ œ. œ œ. œ. œ. œ. œ œ. ˙ œ.
r'(7:4:3), (a)
ã8 ˙ ˙
J J
˙

ã ˙ œ. œ œ œ. œ. ˙ 42 ˙ œ œ. ˙
r'(7:4:3), (b)
˙ ˙ ˙ ˙ ˙ ˙ ˙ œœ
J

œ. œ ˙ 38 œ . œ . œ . œ . œ .
r'(7:4:3), (c)
ã˙ ˙ ˙ ˙
J
˙ ˙ ˙ ˙

ã œ . œ . œ . œ . œJ œ œ . œ . œ . œ . œ . œ . œ . œ . œ œJ œ . œ . œ . œ . œ .

ã œ . œ . œ . œ . 12
8 ˙. ˙. œ. œ. œ. œ œ ˙. ˙. ˙. ˙. œ œ œ. œ. ˙. œ. ˙.
r'(7:4:3), (bc)

J J

Figure 6.3: Grouping of the three-generator interference group 7 ÷ 4 ÷ 3

50 Copy issued to Seth Dempsey ([email protected])

Chapter 7

Resultants applied to instrumental

forms

This chapter discusses the application of a rhythmic resultant to a given pitch sequence (ap-
plication to melody) or to a given set of simultaneous pitches (application to harmony). The
harmonic application is limited to the coordinated attack patterns of pitch subsets from a
single static chord structure.
The techniques are based on a given time rhythm, i.e., a resultant set of attack-durations,
obtained through one of the techniques from the previous or following chapters, and a given
instrumental rhythm. An instrumental rhythm is a pre-determined ordered set of pitches
(melody) or a given sequence of simultaneous pitches (elements, subsets from a given har-
mony).1
The process in this chapter involves interference of the time rhythm with the instrumen-
tal sequence pattern. The result of this process is a pitch-attack series with Npa elements and
a specific distribution over multiple staves in a musical score. This result is called an instru-
mental form, a subject to which Schillinger devotes an entire book (i.e., Book 8) in Volume 2
of [3].

7.1 Instrumental rhythm

The essential step in this approach is that we are determinining the synchronization pattern
of an attack series (with Na elements and note durations) on a single staff with an ordered
set of pitches Np . The resulting pattern is an instrumental part from a musical score.
The synchronization process depends on the ratio of number of attacks to number of
pitches Na /Np , or its inverse Np /Na . When there are integer ratios, such as Na /Np = 1, 2, 3
we find short recurrent pitch-attack series Npa , where the pitch pattern fits an integer num-
ber of times into one attack-duration series r. The consequences for achieving recurrence
through synchronization are illustrated in Fig. 7.1. The diagram shows a pitch pattern, con-
sisting of two pitches, i.e., Np = 2. There are three examples of attack-duration patterns
that lead to integer attack to pitch ratios, Na = 2, 4, 6, and two rhythm resultants r with
non-integer ratios, Na = 3, 5. The integer ratio cases yield recurrence after a few repeats of
1
Note that this chapter introduces processes in the pitch domain. In previous chapters all operations took
place in the time domain only, such as the interference of generators or the synchronization of an attack-duration
pattern with a grouping scheme.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

statements of the pitch pattern

? ? ? ? ?
 recurrence
u
Np = 2 u
Integer ratios:
Na = 2 e e Na /Np = 1 ⇒ Npa = 2

Na = 4 e e e e Na /Np = 2 ⇒ Npa = 4

Na = 6 e e e e e e Na /Np = 3 ⇒ Npa = 6

Non-integer ratios:
Na = 3 e e e Na /Np = 1.5 ⇒ Npa = 6

Na = 5 e e e e e Na /Np = 2.5 ⇒ Npa = 10

6 6 statements of the resultant r

Figure 7.1: The dependence of the instrumental rhythm on the number of elements in the
pitch Np and attack series Na . Closed circles represent the ordered pitch series, open circles
represent rhythmic attacks (no durations indicated)

the pitch distribution pattern; for Na = Np there is the perfect match after a single state-
ment, and Npa = 2. The non-integer ratios require three or more repeats of the pitch pattern
Npa ≥ 6. The diagram illustrates the process for the case of more attacks than pitches, i.e.,
Na ≥ Np . In principle we may have to deal with the reverse situation, where there are more
pitches than attacks, but the approach remains the same. Note durations are irrelevant for
this synchronization process; note lengths only become relevant when we add the grouping
into measures process or want to write out the rhythm.
Let’s demonstrate the pitch-attack synchronization with an example in musical notation.

Example 7.1
Create an instrumental rhythm from interference of an attack-duration series
with an ordered pitch set.
For the following examples we will apply various duration series to a given or-
dered pitch set. The duration series were created using the techniques of (non)-
uniform binary synchronization from Section 2.1.2, interference of three gener-
ators (Section 6.1) and the composition of groups by pairs, shown in Chapter 5.
Various groupings (see Chapter 3) will yield the accompanying time signature
and the total number of measures. The results are shown in musical notation in
Fig. 7.2.

• The given two-pitch set Np = 2 is shown in m. 1. First there is interfer-

ence of a constant duration, i.e., uniform binary synchronization
h i time series
4
∆t = 1 unit in m. 2. For a grouping at time signature 4 this yields the char-
acteristic march bass pattern in the lower octaves; each measure contains

52 Copy issued to Seth Dempsey ([email protected])

 7.2. APPLYING THE PRINCIPLES OF INTERFERENCE TO HARMONY

two statements of the pitch distribution pattern since we have Na /Np = 2.

Then there are two cases of two-generator patterns, the interference resul-
tant r3÷2 with Na = 4 and the fractioning r5÷4 with Na = 13. The latter
would require two statements of r to achieve recurrence at Npa = 26; here
we took the liberty of opening and ending on the same tonic pitch c with
Npa = Na = 13. Finally we apply the three-generator interference resultant
r5÷3÷2 and Npa = Na = 8 = 4Np , four statements of the pitch pattern.
• The three-pitch set Np = 3 is shown in m. 17. The constant duration pat-
tern yields an arpeggio with Npa = 6. This example also shows various
groupings (see the meter in m. 20-21 and 28–32), the interference pattern
r5÷2 with Na = 6 and thus Npa = 6, and a version with contracting adjacent
groups rC (4, 3) (see Section 5.3). This attack-duration group has Na = 16
attacks; we disregard the strict approach and instead create a more musical
alternative, stressing the root of the G major triad.
• The four-pitch example Np = 4 starts in m. 33 and contains an example
of fractioning r6÷5 with Na = 16 and thus four statements of the pitch-
distribution pattern Npa = 16 = 4NP . The balancing adjacent groups
rB (4, 3) pattern with Na = 17 (see Section 5.1) encompasses 4 statements
of the pitch pattern plus a repeat of the opening pitch. The three-generator
alternative resultant r0 (5 ÷ 3 ÷ 2) with Npa = Na = 8 = 2NP is demonstrated
for two complementary factor groupings at TM = 10∆t and TM = 15∆t.

What these examples demonstrate is that with careful design, combining appropriate
pitch-distribution and attack-duration patterns we are able to obtain short patterns, with a
riff character. Sometimes we use the freedom to deviate from the strict calculus in order to
obtain musically sensible results.

7.2 Applying the principles of interference to harmony

The result of the interference process applied to a harmony is an instrumental form dis-
tributed over multiple staves. The typical example is the piano part two-staff score with
the bass in the left hand and the other parts as simultaneous attacks in the right hand. The
harmonic structure H is split into two subsets: if the harmonic structure contains Ns pitches
(parts), i.e., H(Ns ), then the lower staff will use Nl < Ns pitches and the upper staff will
receive Nu = Ns − Nl pitches.
First we will determine the pitch-distribution pattern over the combination of two har-
mony subsets in both hands; this yields the total number of piano pitch attacks Np . Let’s
assume that the bass part plays a sequence of single pitches in rhythmic arpeggio style (an
ordered series with Nl elements), while the upper staff has a fixed number of simultaneous
Nu pitches (played as accented chords). This leads to a kind of ‘oom-pah’ piano part. In
this special case of the single pitch left hand bass line the pitch attack distribution series has
Np = 2Nl elements, since each left hand bass pitch is answered by an accented chord in the
right hand.2
2
Note that we are mapping a simple uniform resultant to the pitch domain as a special case. The next chapter

Copy issued to Seth Dempsey ([email protected]) 53

CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

2p 4:1 = 4t r(3:2) r=frac(5:4)

2 4 œ 3 5 4
&4 œ œ œ 4 œ œœœ 8 ˙ j œ. jœ œ œ jœ. j˙
1
4 œ œ œ œJ œ Jœ Jœ œ 4
r'(5:3:2)

4 ˙ ˙ ˙. ˙ w ˙ 3
9
&4 w Ó 4
˙ ˙ œ ˙ w
3p 6:1=6t r(5:2),(a) r_C(4,3)

3 6 5 4 .
17
&4 œ œ œ 8 œœœœœœ 4 ˙ ˙ œ
œ ˙
˙ 4 ˙ œ
˙ œ œ œ œ ˙
r'(5:3:2), (b)
" ˙. 68 j 44
24
& œ ˙. ˙ œ. œ œ œ j
œ ˙ œ ˙. ˙. œ. œ. œ œ. ˙.
4p r=frac(6:5), (a)
j j
& 44 œ œ œ œ 68 j j 44
33
œ œ œ œ œ œ œ.
œ. œ œ ˙ œ œ. œ œ œ. œ˙
r_B(4,3)
"
& 44 ˙ . " 10
40
˙ œ œ œ ˙ ˙. œ œ ˙. 8
œ œ œ ˙. ˙ ˙ w
r'(5:3:2), (ac) r'(5:3:2), (ab)

& 10
15 j j œ œj œ . ˙ .
48
8 ˙. ˙ œ œ. œ. œ ˙ ˙. 8 ˙. œ. œ œ œ. œ œ

Figure 7.2: Instrumental rhythm example for a 2-, 3- and 4-part ordered pitch set (see m. 1, 17
and 33). Several attack patterns are considered for interference with these pitches: uniform
binary synchronization with constant duration t, synchronization of two or three generators,
fractioning and composition of balancing groups by pairs. Note the various meter groupings
(see the time signature).

54 Copy issued to Seth Dempsey ([email protected])

 7.2. APPLYING THE PRINCIPLES OF INTERFERENCE TO HARMONY

In the following example the rhythmic pattern uses three inputs:

1. A harmonic structure H(Ns ), with Nl pitches in the bass part. The latter number deter-
mines the attack distribution pattern: e.g., root only in the bass (Nl = 1) corresponds
to Np = 2 (left hand - right hand, or LH–RH in piano terminology). Root and fifth in
the lower staff (Nl = 2) implies a Np = 4 pattern (LH–RH–LH–RH) with four attacks
in the pitch domain.

2. A rhythmic attack-duration group with resultant r and Na attacks. This can be any
technique from the previous chapters, such as binary synchronization ra÷b , fraction-
ing ra÷b (Section 2.1.2), group pairing rB , rE , rC (Chapter 5), three generator synchro-
nization ra÷b÷c (Section 6.1). In all odd-numbered statements of the pitch distribution
pattern odd-numbered attack-duration elements from the resultant, ra÷b or ra÷b , are
assigned to the lower staff, even-numbered attack-duration elements are assigned to
the upper staff. For even-numbered statements the assignment is swapped (see the
examples below).

3. Various time signature groupings TM may be used. This is the last of the three degrees
of freedom: the pitch-distribution pattern with number Np , the attack-duration series
r with Na attacks and grouping into measures with duration TM .

Example 7.2
Interference applied to harmony.
Before selecting a rhythmic resultant, we must check the numbers. In case the
total number of attacks in the rhythmc resultant r is not an integer multiple of
Np , the pattern becomes longer. We have to repeat r a number of times before
returning on the first beat of a new measure and achieving recurrence (see the
diagram in Fig. 7.1 which explains the effect of the ratio Na /Np ). Finding the
right combination of resultant r and time signature (grouping) for a given Np
harmonic structure attack pattern is a design consideration. In the example we
will consider three cases for different chord structures. The results are shown in
musical notation in Fig. 7.3.

• A four-part harmony H(4p) with root in the bass Nl = 1, corresponding

to Np = 2, two attacks in the pitch domain for the Am chord (see m. 1–
5). Three resultants are applied, two uniform binary and one non-uniform
binary synchronization patterns. The 2:1 resultant has constant note dura-
tion and yields a polka rhythm with Npa = Na = Np = 2. In the uni-
form binary synchronizaton resultant r4÷1 the foxtrot rhythm emerges with
Npa = Na = 4 = 2Np . The 7:2 resultant with Na = 8 requires an irregu-
lar meter grouping TM = 7∆t for the reccurence to coincide with the end
of a single rhythmic pattern, after three statements of the pitch-distribution
pattern and thus Npa = Na = 8 = 4Np .

will generalize this principle; there we may apply any resultant to subsets of pitches from a chord, distributed
over a number of staves.

Copy issued to Seth Dempsey ([email protected]) 55

CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

• A five-part harmony H(5p) with root and fifth in the bass Nl = 2, corre-
sponding to Np = 4 for the Dmadd9 chord (see m. 6–14 with four attacks
in the pitch domain). The 4:1 uniform binary synchronization pattern in-
strumental form is shown in m. 6 with Npa = 4 = Na = Np (perfect
match, one chord pitch-distribution pattern fits into one measure of this
uniform rhythm). The r8:5 non-uniform binary synchronization in m. 10–
14 has Na = 12, three statements of the pitch-distribution pattern and
thus Npa = 3Np = 12. The latter pattern contains various afterbeat and
syncopated accents, as does the r6÷5 fractioning instrumental form (total
Npa = Na = 16 = 4Np attacks).
• A seven-part harmony H(7p) with root, fifth and dominant 7th in the bass
[9/13
Nl = 3, corresponding to Np = 6 for the G7 chord (see m. 15–41 with 6
attacks in the pitch domain). The r7÷6 non-uniform binary synchronization
resultant in m. 16–36 has Na = 12 and h i requires two statements (note the
break in m. 26) at the time signature 24 with TM = 4∆t, before the recur-
rence landing on the first downbeat; the total number of attacks is Npa =
24 = 2Na = 4Np . The contracting adjacent group rC (5, 3) with Na = 24 was
carefully designed to achieve a short pattern with Npa = 24 = Na = 4Np
and once again demonstrates the appearance of afterbeat and syncopated
attacks.

This example nicely demonstrates the versatily of the Schillinger approach to rhythm
generation. A single harmony already generates interesting, longer rhythmical patterns.
Experimenting with these techniques is a useful training to break away from the standard
regular time division one- or two-measure loops with the inherent danger or monotony in
riff-based music.

Application Tip:
For clarity reasons Example 7.2 shows non-overlapping notes in the two layers,
i.e., the bass and treble clef staves. This is done in order to easily read the rhyth-
mic pattern, resulting from the attack-duration series. However, in order to create
real music there is no reason not to use overlapping note durations between the
layers. So feel free to use longer sustained notes in each layer. Since the rhythm
is applied to a single chord, there will never be a harmonic conflict. The use
of overlapping durations will be even more evident in the multi-staff examples
below (see the extensions in the following section).

7.3 Combination of instrumental form techniques

There are several extensions to the two basic approaches described in the previous sections.

7.3.1 Extension 1: Doubling on multiple staves

This approach yields a score with more than two staves. Typically, the upper harmony staff
will be doubled at the higher octave and the rhythmic resultant will now be distributed over

56 Copy issued to Seth Dempsey ([email protected])

 7.3. COMBINATION OF INSTRUMENTAL FORM TECHNIQUES

j
H(4p),2a 2:1 4:1 7:2 H(5p),4a,4:1

& 42 Œ œœœ Œ œœ 44 Œ œœ Œ œœ 78 Œ œœ Œ œœ ‰ œœ Œ œœ 44 Œ œ Œ œ 68
1

œ œ œ œ œ œ œ œœ œœ

? 42 œ Œ œ Œ 44 œ Œ œ Œ 78 œ Œ œ ‰ œJ Œ œ Œ 44 œ Œ œ Œ
Am
68

r=frac(6:5) 8:5
r r r r r
& 68 Œ œœœ ! œœœ .. ! œœœ ‰ œœœ ‰ . 44 Ó
7
! œœ Œ œœœ ! œœœ ! œœœ œœœ ‰ œœ .. Œ œœ œœ .. ‰
œ . œ. œ œ.
D m(add9)
? 68 œ œ ! œ ! œ ‰ . œR ‰ œ ! œ œJ ! œ Œ œ! Œ
R 44 ˙ œ‰Œ œ Œ Œ. œ
R R R J J
H(7p),6a,6:1 7:6
j j 68 ‰ b œj ‰ Œ b œj ‰ Ó
œœœ Œ . Œ œœœ Œ . œœœ ˙˙˙ œœœ 42 Ó
12
& ˙˙˙ Ó œœœ ‰ b œœ Œ
œœœ œ œ œœœ œœ
G 13(b 9)
?Ó ˙ ‰ œ. œ Œ œ. ‰ Ó 68 ‰ œ ‰ œJ ‰ 42 ˙ Œ œ
œ œ ‰ œJ ˙

j " j
œœœ Œ .
20
& Œ b œœœ b ˙˙˙ Œ b œœ
œœ œœœ .. ‰ b ˙˙˙ œœœ Œ Œ Œ b œœœ ‰ Œ
œ œ ˙ œ .. ˙ œ œ
?œ Œ ‰ œ. Ó œ Œ Œ. œ " Œ ‰ œ.
J œ ˙

29
j 44
& Œ b œœœ Œ Œ b œœœ .. ‰ Œ b œœœ œœœ Œ b ˙˙˙ œœœ ‰ b œœœ ˙˙˙
œ œ .. œ œ ˙ œ œ ˙
?œ Œ ˙ Œ ‰ œj œ Œ Œ œ Œ Œ ‰ œJ Œ " 44

r_C(5,3)
j j ‰ b œœ .. Œ œjœ ‰ Œ b œjœ ‰ ‰ œœ ..
& 44 Œ . b œœœ œœœ ‰
37
œœœ ‰ b œœœ ‰ œœœ ‰ œœœ ‰ œœœ ‰ b œœœ ‰ œœœ Œ œœ .. œœ œœ œœ ..
œ œ œ œ œ œ œ œ œ
? 44 œ ‰ ‰ œ ‰ œ ‰ jŒ œ Œ œ œ Œ. œ œ
œ . ‰ ‰ œJ Œ œ œ J J œ ‰ œJ œ ‰ J J ‰ Œ

Figure 7.3: Interference applied to harmony. There is interference between three chord struc-
tures with H(4p), Np = 2, H(5p), Np = 4 and H(7p), Np = 6 and the rhythmic resultants r.
The example uses binary synchronization, fractioning, group pairing, and three-generator
interference. Various meter groupings (time signatures) were used.

Copy issued to Seth Dempsey ([email protected]) 57

CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

three (or more) staves. There will be single bass part, since in general there the doubling
(into the lower octave) does not make sense and will confuse the overall feeling of rhythm.
The attacks are then distributed over the total number of staves.

Example 7.3
Interference to harmony: doubling on multiple staves.
Note the three-staff system, with the middle staff harmony copied into the upper
staff at the higher octave.
• A six-part harmony H(6p) with root, fifth and third in the bass Nl = 3 and
therefore Np = 9 for the F chord, i.e., 9 distributed attack in the pitch do-
main. Three resultants are applied, the 9:1 uniform binary, the r7:3 with
Npa = Np = Na = 9 and the r9÷4 non-uniform
h i binary synchronization pat-
2
tern with Na = 12 , see Fig. 7.4. The 4 meter for grouping r9÷4 (m. 5–31)
leads to interference with three statements of the attack series, before there
is recurrence on the downbeat. Thus Npa = 36 = 3Na = 4Np . Note the
constant half note durations in the middle layer; this coincidence is caused
by the deliberately selected generator combination (or coincidentally, that
depends on the composer’s intention).
• A seven-part harmony H(7p) with doubled root and fifth in the bass Nl =
3, corresponding to Np = 5 for the Am6/9 chord, see Fig. 7.5. The attack
distribution is chosen freely here: the attack series starts with the middle
harmony layer.3 The bass notes are on the offbeats, with the root doubled at
the lower octave. There is r5÷1 uniform binary with a single statement, r9÷7
non-uniform binary synchronization with Na = 15, and r4÷3 fractioning
h i
with Na = 10. The 38 time signature for the 9:7 synchronization leads
to three attack pattern statements before achieving recurrence, thus
h i Npa =
15 = Na = 3Np . The r4÷3 fractioning pattern at time signature 44 yields
Npa = 10 = Na = 2Np .

7.3.2 Extension 2: Interference between time and instrumental groups

n  
In this approach the time signature m will interfere with the rhythmic resultant r in order
to achieve recurrence through synchronization.4 This yields multiple measure patterns with
repeated attack series before there is recurrence at the end of a full measure. Another form
of interference is the combination of a single resultant with a time-shifted variant. This is the
first case of the combination of resultants in parallel layers. This combination introduces a new
degree of freedom: a different starting point of the two resultants in the lower and upper
layer (the parts P1 and P2 on different staves).
3
This freedom of creating pitch distributions is in fact the subject of the following chapter. Here we demon-
strate a simple form of that generalization to get a feel for what is to come.
4
Most of the examples in the previous section were constructed with fine-tuned numbers in the sense that the
time signature matched a low integer multiple of the atttack-duration pattern. Here we will the generalization
of this interference process.

58 Copy issued to Seth Dempsey ([email protected])

 7.3. COMBINATION OF INSTRUMENTAL FORM TECHNIQUES

œœ œœ œœ œ œ œ. œœ
78 Œ . Œ . œœ Œ Œ . œœ ‰ Œ . œœ ..
H(3p),9a,9:1 7:3 9:4

& 49 Ó œ Ó œ Ó œ 42 ! œ‰Œ
1
! !
J J

9 7 Œ. œ. ‰ Œ œ. Œ ‰ œœ .. Œ . 2
& 4 Œ œœ Œ Œ œœ Œ Œ œœ Œ 8 œœ .. œœ .. 4 ! ˙˙ ! ˙˙
œ œ œ œ. ˙ ˙
? 49 œ Ó 78 œ Œ. Œ.
42 ˙
F
Ó œ Ó Œ. ‰ œ Œ. Œ J ! ‰ œ. !
œ œ.

œœ œœ .. ˙˙ œœ œœ œœ .. ˙˙
& œ Œ œ. ˙ œ‰Œ œ Œ œ. ˙
9
! ‰ ! ! ! ! ! ‰ !
J

& ! ˙˙ ! ˙˙ ! ! ˙˙ ! ˙˙ ! ˙˙ ! ˙˙ !
˙ ˙ ˙ ˙ ˙ ˙
? Œ œ .
! Œ ‰ j ! ! ˙ ! ‰œ ! Œ ! Œ ‰ œj ! !
œ œ

œœ œœ œœ .. ˙˙
œ ‰ Œ œ œ. ˙
23
& ! ! ! Œ ! ‰ !
J

& ! ˙˙ ! ˙˙ ! ˙˙ ! ˙˙ !
˙ ˙ ˙ ˙

?˙ ! ‰ ! Œ œ ! Œ
œ
‰ J ! !
œ.

Figure 7.4: Interference applied to harmony. Doubling on multiple staves, Case 1: inter-
ference between the chord structure with H(6p), Np = 9 and either binary synchronization
or fractioning rhythmic resultants r. The upper staff with harmony parts is doubled at the
higher octave.

Copy issued to Seth Dempsey ([email protected]) 59

CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

# œœœœ œœ # œœœœ .... œ # œœœœ ....

H(7p),5a,5:1 9:7

Œ œœ œœœ
& 45 Ó 38
1
! ! ! ‰ !

5 3 œ. j
& 4 # œœœœ Œ Ó Œ 8 # œœœ .. œœ .. œœ ‰ ‰ ! ! ! !
. œœ .. œœ
6

? 45 Œ œ 38 ‰ œœ
A min 9

œ Œ œ Œ ! ! ! Œ œ œ. !
J

œ œ. œ œ ..
‰ œœ
# œ œœœ ... œœœ ‰ # œœœ .. œœœœ .... œ
‰ # œœœ œœœœ ....
9
& ! ! ! !

& # œœœœ ... œœœ ..

œ ..
! ! ! ! ! # œœœœ ... ! ! !
. .
? ! ! œ ‰ ‰ ! Œ œ ! ! ! œ. œ ‰ ‰ !
œ J œ. œ
J J

# œœœœ œœœœ .... œœœœ .... # ˙˙˙˙ œ # ˙˙˙˙ ˙ ..

frac(4:3)

44 Œ œœœ Œ # ˙˙˙ ..
20
& Œ J ! Ó

! ! ! 44 ˙ . Œ !
& # ˙˙˙ .. # œœœœ Œ Ó !
.
?œ ‰ ! ! 44 Ó . œ Ó œ Œ Œ œœ Ó œ Œ Ó
œ

Figure 7.5: Interference applied to harmony. doubling on multiple staves, Case 2: inter-
ference between the chord structure with H(7p), Np = 5 and either binary synchronization
or fractioning rhythmic resultants r. The upper staff with harmony parts is doubled at the
higher octave.

60 Copy issued to Seth Dempsey ([email protected])

 7.3. COMBINATION OF INSTRUMENTAL FORM TECHNIQUES

Example 7.4
Interference to harmony: interference with time signature.
This form of interference occurs when convoluting a pitch distribution series
(with Np pitches, depending on the given layered harmony and bass) with an
attack-duration series (the resultant
n
r with Na elements) and a time signature
(grouping at a specific meter m ). This may require multiple statements of the
series before achieving recurrence on a downbeat.

• This type of interference has already occurred before, when both the num-
ber of note durations is larger than the number of attacks for the harmony
structure, i.e., Na > Np , and this does not match the number of time units
in the meter. This is a fairly general case and we have seen it in Example 7.2
and Example 7.3.
• The reverse situation with Na < Np is shown in Fig. 7.6, m. 2–3. A five-
part harmony H(5p) with root and fifth in the bass Nl = 2, corresponding
to Np = 4 for the Cm117 chord; there are 4 attacks in the pitch
h i domain. The
three-attack series Na = 3 (2t + t + t) is grouped into the 44 meter.5 This
interference process requires three repetitions of the attack pattern, corre-
sponding to four repetitions of the note duration series before we achieve
recurrence on the downbeat. Thus we obtain Npa = 12 = 4Na = 3Np .
• The second example, m. 4–8, for r5÷2 with Na = 6 requires three repetitions
of the harmonic attack pattern and two statements of the duration series
before achieving recurrence after five measures, i.e., Npa = 12 = 2Na = 3Np .

Example 7.5
Interference to harmony: time-shifted variant of a single resultant.
The example in Fig. 7.6 m. 9–17 demonstrates the time-shifting of a resultant in
combination with the original. The source is a 4 attack pitch distribution pattern
(Np ).

• The rumba rhythm in m. 9–11 is the result of interference of the 3+3+2

attack-duration series in the bass (Na = 3) with the time-shifted permu-
tation 2+3+3 in the upper harmony layer.6 The pattern repeats itself after
two measures.
0
• In m. 12–17 there is interference between the three generator pattern r2÷3÷5
with Na = 8 and a time-shifted copy (see Section 6.1 for determining the
alternative resultant r0 ). The numbers are Npa = 8 = Na = 2Np . Most at-
tacks are non-coincident, except for two attacks in the middle of the pattern
(m. 14–15). The combination creates a sort of rhythmical counterpoint.

5
Note that the three element 2+1+1 attack-duration group is not the result of a two-generator synchronization
process. It may be a subset of a longer synchronization process, such as r3÷2 .
6
Permutations are discussed in depth in Chapter 9 on variation.

Copy issued to Seth Dempsey ([email protected]) 61

CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

j j j j
H(4p),4a t:2+1+1 5:2
b
& b b 44 Œ œœœ Œ œœœ
1
Œ œœœ ‰ œœœ ‰ œœœ Œ œœœ ‰ œœœ ‰ œœœ Ó ˙˙˙ Œ œœœ Ó
C m7(add11)

? b 4 œ Œ œ Œ ‰ œJ Œ j ‰ œj Œ
bb 4 œ œ ‰ œ œ ‰
J ˙ Ó œ Œ ˙

j j
t:2+3+3
b Œ œœœ œœœ .. œœœ œœœ œœœ œœœ .. œœœ œœœ Œ Ó 68
6
& b b ˙˙˙ Ó ˙˙˙ Œ œœœ Ó ˙˙˙
. .
t:3+3+2
˙ ˙ œ œ œ œ. j
? bb Ó
b Ó œ Œ Ó œ. J œ œ œ ! 68

j j
r'(2,3,5)
b 6
& b b 8 Œ. œœ .. œœ .. œœ .. œœ œœ œœ .. œœ .. œœ œœ .. œœ .. œœ .. Œ .
12
œœ
œ. œ. œ. œ œ œ. œ. œ œ œ. œ. œ.
r'(2,3,5)

? b 6 ˙. œ. œ œ. œ j ˙. !
bb 8 J œ œ. œ œ.

Figure 7.6: Interference between time and instrumental groups. There is a mismatch between
the number of elements in the harmonic pitch distribution series Np , the number of elements
in the duration series Na and the interference process with the time signature. This leads to
multiple statements of the patterns, before recurrence is achieved (see m. 2–8). Shown is a
five-part chord structure H(5p) with Np =4. In m. 9–17 there is interference between a given
duration series and a time-shifted permutation or copy of r in the upper layer.

62 Copy issued to Seth Dempsey ([email protected])

 7.3. COMBINATION OF INSTRUMENTAL FORM TECHNIQUES

7.3.3 Extension 3: The combination of techniques

A full score with multiple staves is created from a single harmonic structure by combining
various techniques from this chapter. E.g., a single staff instrumental rhythm may be com-
bined with one or more instrumental forms of interference applied to harmony. The resulting
score looks begins to show similarities with the example of the natural nucleus, discussed in
Section 2.3.

Example 7.6
Instrumental form: Combination of techniques.
In Fig. 7.7 there is an instrumental rhythm (top), a doubled harmony layer (mid-
dle) and a given source bass-harmony staff pair (bottom). The chord structure is
H(6p), the Amadd4 chord, with root and fifth in the bass part (Nl = 2). We will
create three instrumental forms from this source, using a combination of tech-
niques and patterns.

• In m. 2–4, there are constant durations in the lower and middle layer: the
bass-harmony lower layer has a 4:1 uniform binary synchronization pattern
with one-measure duration, thus Np = Na = 4. The middle layer uses
∆t = 3t constant duration, while the upper instrumental rhythm uses the
r9÷2 non-uniform binary synchronization with Na = 10 applied to three
pitches in the ordered subset {d, e, g} from the harmony layer Np = 3. The
horizontal bracket above the staff shows the length of a single statement
(where we have not reached full recurrence). Note the different time scales
in each layer, with the shortest notes in the top layer. The middle layer acts
as a sustained harmonic background. Since it is doubling in the same octave
as the bottom layer, we must inspect the attack series for simultaneous at-
tacks, cross-rhythms, and use a different instrumentation, in order to make
them audibly discernable as separate elements.
• The second instrumental form in m. 5–8 uses a balanced adjacent pair
grouping rB (4, 3) with Na = 17 in the lower layer. The middle layer has
non-uniform binary synchronization 3:2, while the top layer uses fraction-
ing r4÷3 with Na = 10 to create an instrumental rhythm. The top layer with
the shortest time unit (16th notes) repeats three times and is a sort of osti-
nato riff pattern. The middle layer has two statements at half note triplet
time unit. Again, no full recurrence is achieved in this example; see the
return of the root a at the end in the bottom staff.
• The third instrumental form in m. 9–14h i demonstrates maximum activity in
the lower layer (8th note time unit in 68 time signature). The rhythm is r3÷2
with Na = 7; there is no full recurrence, note the root a at the start and end
of the pattern. The middle layer shows a time-shifted alternative resultant
0
for the three-generator pattern r2÷3÷5 with Na = 8. The top layer uses a
balanced adjacent pair group rB (3, 2) with Na = 12.

Copy issued to Seth Dempsey ([email protected]) 63

CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

˙ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ
9:2

4 ˙ ˙
1
&4
3 3 3 3 3 3 3
3t

4 ! ˙˙ .. œœ ˙˙ ˙˙ œœ ˙˙ ..
&4 ˙˙ .. œœ ˙˙ ˙˙ œœ ˙˙ ..
H(6p) 4a,t

4 ww œœ Œ œœ Œ œœ Œ œœ Œ œœ Œ œœ Œ
&4 ww œœ œœ œœ œœ œœ œœ
A m7(add4)
? 44 w Œ Œ œ Œ Œ œ Œ Œ œ
w œ œ œ

. œ œ œ œ œ œ œ œ œ. œ. œ œ œ œ œ œ œ œ œ. œ . œ œ œ œ œ œ œ œ œ . œ . œ œ œ œ œ œ œ œ œ . 68
frac(4:3)

&œ
5

3:2

68
& wwww ˙˙ ˙˙ ww ww ˙˙ ˙˙ ww
3 3 3 3

˙˙ ˙˙ ww ww ˙˙ ˙˙ ww
j j j j
r_B(4,3)

œœ Œ œœ ‰ œœœ ... œœ Œ œœ ‰ œœ ‰ œœ ‰ œœœ ... 68

& Œ. œœ œœ œ.
Œ.
œœ œœ œœ œœ œ.
Ó

? ‰ œ j œ. ‰ œJ ‰ œj Œ œ Œ. 68
œ. Œ œ Œ. ‰ œ J ˙

œ œ œ œ œJ œJ œ ˙.
r_B(3,2)
. œ œ œ œ. œ. œ œ.
& 68 œ
œ
9
J J J J
j j
r'(5:3:2), (b)

& 68 Œ . œœ ..
œœ ..
œœ ..
œœ ..
œœ ..
œœ ..
œœ œœ œœ ..
œœ œœ œœ ..
œœ ..
œœ ..
œœ
œœ
œœ œœ ..
œœ œœ ..
œœ ..
œœ ..
œœ ..
œœ ..
Œ.

j j j j j j
frac(3:2)

& 68 Œ œœœœ ‰ œœ ‰ œœ Œ
œœ œœ
œœ ‰ œœ ‰
œœ œœ
œœ ‰
œœ
œœ
œœ
Œ œœœ ‰ œœœ ‰
œ œ
œœ Œ
œœ
œœ ‰ œœ ‰
œœ œœ
œœ ‰
œœ
œœ
œœ

? 68 Œ œj ‰ œJ ‰ œj Œ Œ œj ‰ œJ ‰ œj Œ
œ ‰ œ ‰ œ ‰ œ œ ‰ œ ‰ œ ‰ œ

Figure 7.7: Instrumental form created through combination of techniques. The six-part chord
structure H(6p) is used in three layers: the bass-harmony lower layer (bottom), the doubled
harmony layer (middle) and the instrumental rhythm (top). Three instrumental forms are
shown.

64 Copy issued to Seth Dempsey ([email protected])

 7.3. COMBINATION OF INSTRUMENTAL FORM TECHNIQUES

With this third extension, the combination of techniques, we may (have to) apply creativ-
ity and artistic taste to generate or combine patterns that occasionally and locally deviate
from the strict procedures described in the first half of this chapter. The degrees of freedom
in pitch, time and grouping domain provide many options for creating complex rhythmic
assemblies from a simple source.

Copy issued to Seth Dempsey ([email protected]) 65

CHAPTER 7. RESULTANTS APPLIED TO INSTRUMENTAL FORMS

66 Copy issued to Seth Dempsey ([email protected])

Chapter 8

Coordination of time structures

In this chapter the aspect of coordinated time, i.e., rhythm is discussed further. Schillinger
here splits coordinated time into two application cases: simultaneity for multiple parts in a
score and continuity in a single part. The latter case was discussed in Chapter 7, Section 7.1,
where the interference of a time series over either a single staff with a number of pitches
was discussed (yielding an instrumental rhythm) or over a set of staves with a given chord
structure. In that chapter also a number of extensions were presented, including doubling
of harmony parts. However, the attacks then were moving one-by-one between the staves.
In this chapter there is a generalization of that principle to the case of multiple staves.
There will be two independent resultants at work. One has to do with the distribution of
attacks over the staves in a score, the other with the coordination of rhythm over the entire
score. And, as usual, there is the grouping degree of freedom, where multiple time signa-
tures may be selected. All this leads to a multitude of cases that will generate shorter or
longer examples with coordinated time.
The section titles below are copied from the Schillinger volumes [3]; they may be puz-
zling at first sight, but they will become clear as we interpret them in terms of what we have
learned in earlier chapters.

8.1 Distribution of a duration group through instrumental and at-

tack groups
This section shows the principle of multiple independent generators that are applied to the
general case of a set of staves in a score. In the case of a single staff, the number of pitches on
that staff is relevant.1 The time structure coordination process is illustrated in Fig. 8.1. The
upper diagram shows the single staff case with an ordered set of pitches, the lower diagram
shows the general case of multiple instruments in a score.
The coordination of time structures involves three subprocesses, three recurrence steps:

1. Attack distribution recurrence. This implies the synchronization of the pitch attack dis-
tribution group rp with Np attacks with the number of instruments NI in the score. In
general Np > NI and the instruments will play more than one attack in a single pattern
statement. The lower diagram shows four instruments NI = 4 and three subgroups
1
Remember that attack groups and attack series are synonyms. The section titles in this chapter from the
Schillinger book use the attack groups alternative.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 8. COORDINATION OF TIME STRUCTURES

Attack duration Time signature

series grouping
h i
N
∆t
Ordered ra , Na , Tr | TM = N ∆t |
pitch ? ?
set Attack Full
Np Npa NM
c c c c c - duration - measure -
recurrence Tpa recurrence Ttot

(a)

Attack distribution Attack duration Time signature

group series grouping
b
bbb
h i
N
bb ∆t
Instruments rp , Np ra , Na , Tr | TM = N ∆t |
? ? ?
NI Attack NIp Attack Npa Full NM
- distribution - duration - measure -
recurrence rIp recurrence Tpa recurrence Ttot

(b)

Figure 8.1: The process of coordination of time structures through interference and recur-
rence. a): Synchronization of an attack-duration group for a single staff with ordered pitch
set, b): Synchronization of an instrument group with and attack group for multiple instru-
ments in a score. The final subprocess for both cases is the grouping into measures for a
specific meter.

68 Copy issued to Seth Dempsey ([email protected])

 8.2. SYNCHRONIZATION OF AN ATTACK GROUP WITH A DURATION GROUP

of attacks rp = 2a + 3a + a, Np = 6. Recurrence requires more than one statement of

the pitch attack distribution group. In general NIp = N1 Np with N1 > 1. In the sketch
we have NI = 4 and thus NIp = 4Np = 24 attacks in a specific pitch order rIp . This
corresponds to N1 = 4, four statements of the pitch attack distribution pattern.

2. Attack-duration recurrence. In this step the rhythm will be determined from the attack-
duration series ra with Na elements and a total duration of Tr time units. We synchro-
nize the NIp element pitch attack group with the Na note duration series. This also
in general will imply multiple statements of both, i.e., N2 ra and N3 rIp , before there
is recurrence. This determines the total set of durations which now has Npa = N2 Na
elements and the total duration Tpa = N2 Tr time units.
h i
N
3. In the time signature grouping process we choose the time unit ∆t and a meter ∆t .
The duration of a single measure is TM = N ∆t. We synchronize the Npa duration
pattern with the measure length until we find recurrence after NM full measures; this
may require multiple statements N4 Npa . The total duration thence is Ttot = NM TM =
N4 Tpa .

In the case of the single staff with ordered pitches in Fig. 8.1.a the first subprocess is miss-
ing. Instead we synchronize the attack-duration series rt with the Np elements pitch set. The
entire process is strongly dependent on a careful selection of the h attack
i distribution group
N
ra , the attack duration series rt and the time signature grouping t . There is some calculus
involved. When the integer values N1 , N2 , N3 or N4 become large numbers, we will have
very long patterns with lots of offbeat and afterbeat attacks. The potential of achieving in-
teresting non-standard rhythmic patterns then has to be balanced against the overall length
until the pattern repeats.2

8.2 Synchronization of an attack group with a duration group

This is the single staff case, shown in Fig. 8.1.a. The process is identical to the technique of
creating an instrumental rhythm, as discussed in Section 7.1. The number of pitches on the
staff is Np and it is an ordered set. There is one resultant r, that determines both the number
of attacks Na and the attack-duration group.

Example 8.1
Synchronize an attack group with a duration series.
Figure 8.2 shows the synchronization process between durations and attacks for a
given ordered set of six pitches Np = 6 (m. 1). It is a simple diatonic theme. Note
that the pitch c occurs twice in the set. All duration sets are generated through
non-uniform binary synchronization, using the two-generator technique from
Section 2.1.2. The example demonstrates the effect of the number of elements
in the duration series Na on the instrumental rhythm (see also Section 7.1.).
2
Creating the examples in the following sections meant careful design, some trial-and-error, and lots of num-
ber checking.

Copy issued to Seth Dempsey ([email protected]) 69

CHAPTER 8. COORDINATION OF TIME STRUCTURES

• Case 1: Na = Np (see m. 2–8). The resultant r4÷3 = 3t + 1t + 2t + 2t +

1t + 3t produces six attacks, Na = 6.3 Therefore there is recurrence
h i after
h i
one statement of the pitch set. Two groupings are shown, with 4 and 64
4

meter at quarter note time unit t = 14 .

• Case 2: Na < Np (see m. 9–17). In the duration series r2÷1 = 2t + 1t + 1t + 2t
there are
h ifour attacks, Na = 4. Recurrence requires four pitch set statements
with a 44 meter (indicated by the breathing signs) and two statements with
h i
a 32 grouping. The time unit is the quarter note t = 1
4 and the half note,
respectively.
• Case 2: Na > Np (see m. 18–38) The 9:4 non-uniform binary synchronization
pattern contains 12 attacks: r9÷4 = 4t + 4t + 1th+i3t + 4th+i2t + 2t + 4t + 3t +
1t + 4t + 4t, Na = 12. The two groupings at 68 and 44 both lead to two
1
statements of the pitch set before recurrence takes place, at 8th note t = 8
and quarter note t = 14 time unit, respectively.

8.3 Distribution of a synchronized duration group through the fi-

nal duration group
This section is revisiting the aspect of grouping and using a specific meter. It refers to the
last recurrence subprocess in Fig. 8.1. The final duration group is the number of time units
4
hin ithe time signature, i.e., a single measure contains T = N t time units. h For
i example the
4 1 6
4 meter is equivalent to T = 4t with a quarter note (t = 4 ) time unit, 8 meter implies
T = 6t = 2 × 3t with 8th note (t = 18 ) time units.
Grouping options were already discussed in Chapter 3. There is nothing new in this
section, except for Schillinger introducing the notion of the final duration group. Example 8.1
shows the effect of two final duration groups (two meters) for each case.

8.4 Synchronization of an instrumental group with an attack group

This is the case of a score with multiple instruments on separate staves, which will require
the three steps sketched in Fig. 8.1.b. There is more freedom now, as we need a resultant
rp (with a total of Np attacks) to determine the distribution of attacks over the number of
instruments NI in the score and a second resultant ra (an Na element attack-duration group)
to coordinate the rhythm over the entire score. And finally, the time signature grouping
selection will determine the number of measures until recurrence.
The double use of the term attack group in this context is confusing. On a single staff it
has been used to control the time domain, in the case of a score it is used also to distribute
3
Here we write the resultant as rt = 3t + 1t + . . ., including the variable t = ∆t in the time, in order to prevent
confusion with the attack pattern resultant rp = 3a + 1a that determines the distribution of attacks in the pitch
domain.
4
Here we write t = ∆t, indicating the time unit for rhythmic division.

70 Copy issued to Seth Dempsey ([email protected])

 8.4. SYNCHRONIZATION OF AN INSTRUMENTAL WITH AN ATTACK GROUP

6p r(4:3) r(4:3)

& 62 ˙ ˙ ˙ 44 ˙ . œ œ ˙. 43 ˙ . œ ˙
1
˙ ˙ ˙ ˙ ˙

r(3:2)
, ,
44 ˙ œ œ
œ œ ˙
œ ˙ œ œ ˙ œ œ ˙
7
&˙ ˙. ˙ ˙ ˙

, r(3:2)
,
˙ 32 w ˙
&˙ œ œ œ œ ˙ w
15
˙ ˙ w ˙

r(9:4)
,
&˙
w ˙ 68 œ . œ œ œ œ œ. œ. œ œ œ œJ œ .
21
w ˙ w J J J
r(9:4)
,
j j
& œ . œ œ œ œ œ . 44 w
w w w
˙ ˙ ˙. œ w
28
œ ˙. w

Figure 8.2: Synchronize an attack group with a duration series. The six-pitch Np = 6 ordered
attack set is synchronized with a duration series r. Several grouping cases are shown for
Na = Np (m. 2–8), Na < Np (m. 9–23) and Na > Np (m. 24–38). The two latter cases require
multiple statements in order to achieve recurrence. Horizontal brackets indicate the first
statement of the attack-duration group.

Copy issued to Seth Dempsey ([email protected]) 71

CHAPTER 8. COORDINATION OF TIME STRUCTURES

attacks over instruments on separate staves. Hopefully, the examples below will clarify this
issue.
Schillinger demonstrated this process with single pitches on each staff in order to con-
centrate on the aspect of rhythm. In the example here there are also multiple pitches on the
staff; that might better illustrate how the coordinated time techniques may be applied to
generate musical scores.

Example 8.2
Synchronization of a three-layer instrumental group with an attack and dura-
tion group.
[9/13
The example is based on an extended dominant chord G7 distributed over
three instruments in a score (NI = 3). From bottom to top the layers contain one
(I1 : root), two (I2 : third and dominant 7th) and three (I3 : the extensions plus
doubled third) pitches. The synchronization is demonstrated for two cases.

• Case 1: the pitch distribution attack series is rp = 3a + 2a (Np = 5), i.e., three
attacks in one layer followed by two attacks in the next layer. Recurrence
requires three iterations of the attack group: (I1 + I2 ) + (I3 + I1 ) + (I2 + I3 )
with a total of NIp = 15 attacks, as shown in Fig. 8.3.a. We demonstrate
three possible attack-duration synchronizations, see Fig. 8.4:5
h i
5
1. Uniform distribution t, grouping by TM = 5t, i.e., 4 . The time unit
1
is a quarter note t = 4.
All attacks will be one time unit long, and the
example is three measures long (m. 2–4).
2. Non-uniform duration series ra = 3t + t + 2t + t + t; note that this is not
the resultant of a two-generator mechanism, but a traditional rhythm.
The number of elements in the time series is equal to the number of
1
elements in the attack series Nha i= Np . The time unit is t = 16 , grouped
by T = 2 × 4t = 8t, yielding a 24 meter. Again, there is recurrence after
three measures (m. 5–7).
3. The duration series is the fractioning resultant ra(4÷3) = 3t + t + 2t +
t + t + t + t + 2t + t + 3t, with Na = 10 and total duration Tr = 16t.
Synchronization of attack and duration series requires three statements
of the resultant N2 = 3, and Npa = 2NIp = 30, total duration h Ti tot =
1 4
3Tr = 48t. Grouping with TM = 8t at time unit t = 8 yields a 4 time
signature and a total of six measures (m. 8–13).
• Case 2: the pitch distribution attack group now is generated through non-
uniform binary synchronization, i.e., rp(3÷2) . This yields p~ = 2a + a + a + 2a,
Np = 6, see Fig. 8.3.b. Once again, we need three attack group iterations to
achieve recurrence: (I1 +I2 +I3 +I1 )+(I2 +I3 +I1 +I2 )+(I3 +I1 +I2 +I3 ) with a
total of NIp = 18 attacks. Three attack-duration synchronization results are
shown in Fig. 8.5.
h i
6
1. Uniform distribution t, time unit t = 81 , grouping by TM = 6t, i.e., 8 ,
see m. 1–3.
5
Horizontal brackets in the score indicate the first statement of the attack distribution or the duration group.

72 Copy issued to Seth Dempsey ([email protected])

 8.4. SYNCHRONIZATION OF AN INSTRUMENTAL WITH AN ATTACK GROUP

bbb b b I3 b b bb b b I3
bb bbb I2 b bb bb b I2
bbb bb I1 bb bb b b I1
- -
0 5 10 15 a 0 5 10 15 a
(a) rp = 3a + 2a, NI = 3 (b) rp(3÷2) , NI = 3

b b I4 bb b b b b I4
b I3 bb bbb b I3
b I2 b bbb bb I2
bb I1 bbb b bb I1
- -
0 5 a 0 5 10 15 20 a
(c) rp(3÷2) , NI = 4 (d) rp(4÷3) , NI = 4

Figure 8.3: Pitch attack group distribution over multiple instruments in a score. a): rp =
3a + 2a, NI = 3, b): rp(3÷2) , NI = 3, c): rp(3÷2) , NI = 4, d): rp(4÷3) , NI = 4.

2. Non-uniform binary synchronization with ra(3÷2) = 2t + t + t + 2t,

Na = 4 with total duration Tr = 6t. Attack-duration recurrence requires
Npa = 2NIp = 36, i.e., nine statemens of ra with a total duration of
Tpa = 54t. Grouping by TM = 9t at the time unit t = 81 then leads to a
six-measure pattern, m. 4–9, with NM = 6, Ttot = Tpa .
3. Finally, we use an expanding adjacent groups resultant rE (5, 3) with
Na = 24 and total duration Tr = 40t (see Section 5.2). Recurrence is
achieved after Npa = 4NIp = 72, i.e., three statemens of rE (5, 3) and a
total duration of T = 120t. We group at TM = 8t, time unit t = 81 ,
h i pa
time signature 44 into a total of NM = 15 measures, see m. 10-24. This
example demonstrates that we may extend the duration of notes, see
e.g., the sustained chords in the upper layer in m. 10–11 and 14. The
rhythm is determined by the attacks, the beginning of the notes. Since
the entire example is based on a single given chord, there will not be a
harmonic clash when we extend the note durations.

Note that the order of the attack subgroups has always been from the instrument on the
bottom staff in the score to the upper staff instrument, or equivalently, from layer 1 to layer
NI . In the example this also implies from lower to higher pitches. This, however, is another
degree of freedom. In a different staff order or with a modified attack distribution group
(that skips or swaps certain instruments) the coordination process would have resulted in a
different recurrence pattern. The time structure coordination technique and the subprocesses
would have remained identical, though.
The next example demonstrates how the same principle may be applied to multiple lay-
ers for a single instrument; either a four-note arpeggio pattern for a monophonic instrument
or a chord voicing for a polyphonic instrument as is the case here.

Copy issued to Seth Dempsey ([email protected]) 73

CHAPTER 8. COORDINATION OF TIME STRUCTURES

j
3a+2a, t,T=5t 3a+2a, 3t+t+2t+t+t, T=2x4t

& 44 # www 45 Ó. # œœœ œœœ 42 # œœœ ... œœœ œœœ ‰ Œ ‰ # œœœ œœœ 44
1
∑ # œœœ œœœ œœœ Ó ∑

? 4 ww 5 œœ œœ œœ œœ œœ 2 œœ œœ œœ .. œœ œœ 4
4 4 Ó. ∑ Ó 4 Œ ‰ ∑ J‰ 4
G 13(b 9)
? 44
45 œ œ œ Ó 42 œ . œ œj ‰
Ó. ∑ Œ ‰ ∑ 44
w œ œ œœ

j j
3a+2a,r=frac(4:3),T=8t

& 44 Ó. # œœœ ... œœœ œœœ Œ # œœœ œœœ ...

8
∑ # œœœ œœœ œœœ Ó # œœœ œœœ ∑ Ó

? 44 Ó . œœ œœ œœ .. œœ œœ œœ œœ .. œœ œœ œœ
∑ J Œ Ó J ∑ Ó

? 44 j j Ó.
œ. œ œ Œ
Ó ∑
œœœ Ó
∑
œ œ. œœ

68
Figure 8.4: Synchronization of a three-layer instrumental group with an attack group. The
∑
score&distributes an ∑extended G
∑ [9/13
7
∑ three instruments.
over ∑ ∑ Case 1 is∑ based on∑ the pitch
attack distribution group 3a + 2a, Np = 5. Shown are a uniform (m. 2–4), a non-uniform
? ∑ ∑ ∑ ∑
(m. 5–7) duration series and a fractioning attack-duration pattern r4÷3 (m. 8–13).
∑ ∑ ∑ ∑ 68

? ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ 68

74 Copy issued to Seth Dempsey ([email protected])

 8.4. SYNCHRONIZATION OF AN INSTRUMENTAL WITH AN ATTACK GROUP

j j j j
a(3:2)=2a+a+a+2a,t,T=6t a(3:2),r(3:2),T=9t

& 68 Œ . # œœœ ‰ ‰ Œ # œœœ Œ . # œœœ œœœ ‰ ‰ œœœ œœœ 98 Œ . ‰ # œœ Œ . Œ . # œœ ‰ Œ . # œœœ œœœ Œ .

1
œœ œœ
œ œ œ œ

? 68 Œ œœ Œ . œœ œœ œœ
‰ ‰ œ œ Œ.
œœ 9 œœ œœ œœ œœ œœ œ
‰ œ Œ.
J J ‰ ‰ 8 Œ. J Œ Œ. J Œ. J Œ.

? 68 ‰‰ Œ. j‰ ‰ Œ jŒ . 98 jŒ . j Œ . Œ jŒ . Œ. j‰ ‰ Œ .
œœ œœ œ œ œ œ œ œ œ œ

j j j
a(3:2),r_E(5,3),T=8t

& Œ . # œœœ ‰ Œ . Œ . # œœœ ‰ ‰ Œ . # œœœ œœœ Œ . œœœ œœœ 44 Ó

7
Œ # œœœ ˙˙˙ Ó Œ # œœœ Ó

? ‰ œœ Ó . œœ œœ œœ œœ œ
Œ œJ Œ . 44 Ó œ
‰ œJ Œ
œœ .. œœ œœ œœ œœ
J Œ. J Œ. Ó J Œ ‰ J

? j Œ. j Œ. ‰ Œ. Œ. ‰ Œ. 44 j j‰ Œ ‰ j Ó
œœ œœ œ œ œ. œ œ œœ ˙ ˙

j j j j
# œœœ ... œœœ œœœ Œ ‰ # œœœ œœœ Œ ‰ # œœœ Ó
13
& # œœœ œœœ Œ œœœ œœœ Œ Œ # œœœ ˙˙˙ Ó ‰ # œœœ Œ Ó

? Œ ‰ œœ Ó ˙˙
Ó
œœ .. œœ œœ œœ ˙
‰ œJ œ ˙
˙˙ œ œœ
Œ œJ ‰ Œ œ œ
J J J‰ Œ Ó

? Œ j‰ Œ . j j Ó.
œ œ œ Œ œ œ œ. œ ˙ Ó Œ œ ˙ œœŒ œœŒ

j j j j j j
# œœœ ... œœœ œœœ ‰ Œ Œ # œœœ ‰ Œ œœœ œœœ Œ ‰ # œœœ œœœ œœœ ...
19
& # ˙˙˙ Ó ‰ # œœœ œœœ ˙˙˙ Ó # ˙˙˙

? Œ . œœ œœ œœ .. Ó Œ
œœ ˙˙
Ó
œ ˙
Œ œ ˙
œœ œœ
Œ
œœ œœ
Œ
œ
Œ œ
˙˙
J J

? Œ Ó ‰ œj Œ Ó j j j
œ ˙ œ. œ œ Œ ‰ œ œ Œ ‰ œ Ó ˙ Ó

Figure 8.5: Synchronization of a three-layer instrumental group with an attack group

[9/13
(cont’d). The three-layer G7 with pitch attack distribution Case 2 (rp = 3 ÷ 2) is syn-
chronized with a uniform (m. 1–3), a non-uniform binary synchronization r3÷2 (m. 4–9) and
an expanded group pairing rE (5, 3) (m. 10–24) attack-duration series.

Copy issued to Seth Dempsey ([email protected]) 75

CHAPTER 8. COORDINATION OF TIME STRUCTURES

Example 8.3
Synchronization of a four-layer instrument with an attack and duration group.
The example is based on a minor chord Dm47 distributed over four layers in one
instrument (NI = 4). This is a typical fingering example for the two hands
in a keyboard part. The voicing in the two middle layers is in fourths. The
instrument-attack-duration synchronization is shown in Fig. 8.6:

• Case 1: the attack group is generated through non-uniform binary synchro-

nization rp(3÷2) . The resultant is rp = 2a + a + a + 2a (Np = 6), i.e., six
attacks distributed over four layers, as shown in Fig. 8.3.c. We demonstrate
four possible attack-duration synchronizations:
h i
6
1. Uniform distribution t, grouping by TM = 6t, i.e., 4 . The time unit is
a quarter note t = 14 , see m. 1.
2. With a deliberately chosen duration series ra = 2t + t + t + t + 2t + t
1
the number Na = Np = 6. With grouping TM = 16t, at t = 16 time unit
there is recurrence after one measure (m. 2), NM = 1, Ttot = 16t.
3. With the non-uniform binary synchronization duration series ra(5÷4) ,
the resultant is ra = 4t + t + 3t + 2t + 2t + 3t + t + 4t (Na = 8, Tr = 20t).
The attack pattern will repeat after three statements when Npa = 24 =
3Na = 4Np , and the total duration is Tpa = 3 × 20t = 60t. With grouping h i
TM = 12t at t = 16 1
this yields a total of NM = 5 measures at meter 34
(m. 3–7).
4. The duration series is the fractioning resultant ra(4÷3) = 3t + t + 2t + t +
t + t + t + 2t + t + 3t, with Na = 10 and total duration Tr = 16t. Syn-
chronization of attack and duration series requires three statements of
the resultant, Npa = 30 = 3Na = 5Np , with a total duration is Tpa = 48t.
1
Grouping with T = 16t at time unit t = 16
h i M
yields a total of NM = 3
measures at 44 time signature (m. 8–10). Note the note duration over-
lap between the two staves for easier reading; as stated before the over-
all rhythm is determined by the note attacks.
• Case 2: the attack group now is generated through non-uniform binary syn-
chronization resultant rp(4÷3) . This yields rp = 3a + a + 2a + 2a + a + 3a,
Np = 12. We need two attack group iterations to achieve recurrence:
(I1 + I2 + I3 + I4 + I1 + I2 ) + (I3 + I4 + I1 + I2 + I3 + I4 ) with a total
of NIp = 24 attacks, see Fig. 8.3.d. Here also we demonstrate four synchro-
nization results:
1
1. We start with uniform note duration
h i
distribution t. At time unit t = 8
12
and grouping by TM = 12t, i.e., 8 meter this leads to recurrence after
NM = 2 measures, see m. 11–12.
2. Next there is the non-uniform binary synchronization with ra(3÷2) =
2t + t + t + 2t, Na = 4 with total duration Tr = 6t. Attack-duration
recurrence requires Npa = 24 = 6Na = NIp , i.e., six statemens of the
duration series with a total duration of Tpa = 36t. Grouping by TM = 9t

76 Copy issued to Seth Dempsey ([email protected])

 8.4. SYNCHRONIZATION OF AN INSTRUMENTAL WITH AN ATTACK GROUP

1
at
h i
the time unit t = 8 then leads to a four-measure pattern (NM = 4) in
9
8 time signature, see m. 13–16.
3. The fractioning pattern ra(6÷5) implies Na = 16 attacks with a total du-
ration of Tr = 36t (see the numbers in Section 4.1). Now there will
be recurrence after Npa = 48 = 3Na = 2NIp , with a total duration of
1
Tpa = 108t. Grouping with TM h= i12t at time unit t = 16 leads to a
pattern of NM = 9 measures in 34 time signature (m. 17–25). Some
notes were extended for easier reading and creating overlap between
both hands.
0
4. Finally, we use the three generator alternative resultant ra(5÷3÷2) , ra =
6t+4t+2t+3t+3t+2t+4t+6t, with Na = 8 and total duration Tr = 30t
(see the numbers in Section 6.1). When Npa = 24 = 3Na = NIp there
is recurrence after a total of Tpah =i 90t. When we group at TM = 10t,
time unit t = 18 , time signature 54 the synchronization pattern is nine
measures long (NM = 9), see m. 26-34.

This example shows that using small time units, such as the 16th note
1
(t = 16 ), leads to various syncopated and afterbeat patterns, that may be
suitable as keyboard accompaniment patterns in a 16th (Funk)-Rock setting.

Copy issued to Seth Dempsey ([email protected]) 77

CHAPTER 8. COORDINATION OF TIME STRUCTURES

a(3:2)=2a+a+a+2a a(3:2),2t+t+t+t+2t+t,T=16t a(3:2),r(5:4),T=12t

& 46 Ó Œ œœ œ œ 44 Œ œœ œ œ Œ œœ œ œ 43 Œ Œ œœ œ œ
1
Œ Œ œœ œ . œ Œ œœ œ œ
D m7(add4)
?6 œ 4 œœ Œ œ 3 œ. ‰ ≈ œr œ œœ Œ ‰ œj œ . œœ Œ Œ ≈ œj.
4 œ œ œ Œ Ó 4 œœ œ œœŒ 4 œ œ œ. Œ

a(3:2),r=frac(4:3),T=16t
7
j 44 Œ ‰ œ œ œj ‰ ≈ œj. œ . œ Œ ‰ . œr œ œ ‰ œj œ œ . 12
& œ ‰ œ .. œ œ œœ œ œ Œ 8
œ œ œ. œ œ
? œ 4 œ œ œ œ ‰ ≈ œr œ œœ ‰ Œ 12
œ œ Œ Œ 4 œ . œ œJ ‰ ≈ œ œ œJ ‰ Œ œ œœŒ œ œ. œ 8

a(4:3),t,T=12t a(4:3),r(2:1),T=9t
j j j
& 12
8 Œ . ‰ œœ œœ œ œ ‰ Œ . j 9
œœ œœ œœ œ ‰ ‰ Œ œœ œ œ œ 8 Œ . Œ . œœ œœ œ œ Ó .
j j
œœ œœ œœ œ Œ .
11

? 12 œ jœ œ œ ‰ œ œ œœ œœ ‰ Œ . 98 j jœ Œ . œ œœ œœ œœ Œ . Œ . œ œj
8 œ œ œ œJ Œ Œ œ œ œ œ Œ . œ œ œ œ Œ. JJ

a(4:3),r=frac(6:5),T=12t
jj ‰ ≈ œr œ . œ œ ≈ œr œ œ œ œ ‰ ≈ œr œ . œ œ
& Œ. 43 Œ Œ
16
œœ œ œ œ Œ œ Œ Œ
œ œ. œ œ œ œ

? œœ œœ Ó . œ
43 œ œ œ œ œJ Œ ‰ œj œœ œœ .. œœ Œ Œ œ œ ‰ œj œ œ .
J œ œ œ œ œJ Œ

a(4:3),r(5:3:2),T=10t

r ‰ ≈ œr œ . œ œ 5
22
& ≈ œœ œœ œ œ œ œ Œ
œ œ. œ
≈ œ œ œ œ œ ˙
œ 4 ∑ Ó ‰ œ. œ
œ. œ
? œœ Œ Œ œœ œ ‰ œj œ œœ .. œœ
45 ˙ .
œ œ œ œ œJ Œ Œ Œ ˙ œ ˙ Ó

28
&˙ ˙. ∑ Ó ‰ œ. œ ˙˙ ˙. ∑ Ó ‰ œ œ. ˙ ˙.
œ. œ œ
? ∑ ˙˙ œœ ˙˙ Ó ∑ œœ ˙˙ Ó ∑
˙. ˙. ˙

Figure 8.6: Synchronization of a four-layer instrument with an attack group. The score
distributes an extended Dm47 over two staves. Case 1 uses the attack distribution group
rp(3÷2) , Np = 6, Case 2 on the pattern rp(4÷3) .

78 Copy issued to Seth Dempsey ([email protected])

Chapter 9

Homogeneous simultaneity and

continuity (variations)

This chapter introduces the concept of variation for short patterns. The variation technique
is used to prevent monotony in the case of repetition of shorter patterns with just a few
attacks. We have seen in the previous chapters how sets of generators with higher values for
the periodicity (slowly ticking metronomes) may generate long attack-duration patterns (see
Chapters 2 to 6). Interference between instrument or score attack groups, duration series
and time signature grouping also may lead to long patterns before recurrence takes place
(see Chapter 7 and 8.) Limited musical perception memory will disguise the repeats in these
cases.
For shorter patterns, typically between two and five elements in the attack-duration
group, we may apply the technique of permutation in order to create variation. The origi-
nal pattern plus the set of variations may then be used either in series to create continuity or
in parallel on multiples staves to create simultaneity. The examples will also demonstrate the
combination of both continuity and simultaneity. The permutation approach will be applied
to various attack group parameters such as duration (note length), note vs. rests, dynam-
ics (accented notes), split-unit groups and groups in general (higher order permutations).
The use of a simple starting attack pattern as the basis for the variations guarantuees the
homogeneous character of the result.

9.1 Variation through circular and general permutations

Consider an ordered set with N elements; e.g., abc is a set with N = 3 elements, where a
is the first, b the second and c the third element. There are several options for changing the
element order: these are called permutations.
We may discern two types of permutations for a set containing N different (non-equal)
elements:
1. the set has N ! = 1 × 2 × . . . × N general permutations. This set of general permutations
contains all possible orderings of the elements in the original set.
2. the set has N circular permutations. Imagine positioning the elements on a disk, starting
at the top; then rotate the disk by one element in clockwise direction and the last ele-
ment appears on top. This is the first circular permutation. Repeat the rotation process

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

until the first element returns on top. Alternatively, we may change the direction and
move in counterclockwise direction to obtain the circular permutations in a different
order.
The set size of possible permutations will change when the original set contains a number
of equal elements, but let’s first consider the case for dissimilar elements.

9.1.1 The set of permutations

Permutations of a set of two elements
The two-element set is {a, b}, N = 2. The mathematics tell us that there are N ! = 2! = 1×2 =
2 general and N = 2 circular permutations. Well, here they are:
{ab, ba},
where the general and the circular permutations are identical, since we can only swap the
two elements.1

Example 9.1
Permutation of a two-element set.
The two-element group permutation is shown in musical notation in Fig. 9.1
m. 1–2, where it is applied to the attack-duration group r = 3t + t, a = 3t, b =h t.i2
The time unit is the quarter note t = 14 , the grouping time signature is 44 .

Permutations of a set of three elements

The number of permutations increases for N = 3, i.e., the set {a, b, c}. There are N ! = 3! =
1 × 2 × 3 = 6 general permutations
{abc, acb, bac, bca, cab, cba}
and N = 3 circular permutations
{abc, bca, cab}.
Note that the circular permutations are included in the set of general permutations.

Example 9.2
Circular and general permutations of a three-element set.
See the three-element group permutations in rhythmic musical notation in
Fig. 9.1 m. 3–11, where it is applied to the attack-duration group r = 3t +
t + 2t, i.e., a = 3t, b = t, c = 2t. First the three circular permutations
{abc, bca, cab} are shown (m. 3–5), then there are the six general permuta-
h i
tions. The time unit is t = 18 note, the grouping time signature is 68 .

1
Note that the possible permutations are represented as a comma-separated superset of ordered sets. Later
we will see how this superset is used in music as either a continuity or simultaneity.
2
Note that we have used the symbol a once again. Now it represents the duration of the first element in an
attack-duration group.

80 Copy issued to Seth Dempsey ([email protected])

 9.1. VARIATION THROUGH CIRCULAR AND GENERAL PERMUTATIONS

ab=3t+t ba abc=3t+t+2t bca cab abc acb bac

& 44 | . Û |. 68 Û . Û Û Û Û Û. Û Û Û Û Û. Û Û Û. Û Û Û Û Û Û
1
Û
J J J J J J J J
bca cab cba abcd=5t+t+4t+2t bcda

& Û Û Û. Û Û Û Û. 12
8 Û. Û Û Û. Û Û Û Û Û Û Û.
9
Û Û Û Û Û
J J J J J J J J J
cdab dabc abcd abdc acbd

& Û. Û Û Û. Û Û Û Û Û. Û Û Û Û. Û. Û Û Û. Û Û Û. Û Û Û Û Û. Û. Û Û Û. Û Û
14

J J J J J J J J J
acdb adbc adcb bacd

& Û. Û Û Û. Û Û Û. Û Û Û Û Û Û. Û. Û Û Û Û Û Û Û Û Û. Û. Û Û
19

J J J J J J J J
dcab dcba

Û Û. Û. Û Û. Û.
23
&’’’’ ’’’’ Û Û Û Û Û Û
J J J J

Figure 9.1: Variation through permutations. Circular and general permutations are shown
in rhythmic notation for a two-element (N = 2, m. 1–2), three-element (N = 3, m. 3–11)
and four-element attack-duration group (N = 4, m. 12–26). For the four-element set not all
general permutations are shown, as indicated by the slashes.

Copy issued to Seth Dempsey ([email protected]) 81

CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

Permutations of a set of four elements

Using the same approach for a set of N = 4 elements {a, b, c, d} yields N ! = 4! = 1×2×3×4 =
24 general permutations. Here they are:

{abcd, abdc, acbd, acdb, adbc, adcb, bacd, badc,

bcad, bcda, bdac, bdca, cabd, cadb, cbad, cbda,
cdab, cdba, dabc, dacb, dbac, dbca, dcab, dcba}.

There are N = 4 circular permutations

{abcd, bcda, cdab, dabc}.

It will be obvious that for larger sets the number of general permutations increases expo-
nentially. For the next larger set, with N ≥ 5, there are 120 general permutations. Using
the complete permutation set in a rhythm is most unlikely. I tis more likely to use a specific
subset in a musical application.

Example 9.3
Circular and general permutations of a four-element set.
The four-element group permutations in musical notation are shown in Fig. 9.1
m. 12–27, where it is applied to the attack-duration group r = 5t + t + 4t + 2t, a =
5t, b = t, c = 4t, d = 2t. The four circular permutations {abcd, bcda, cdab, dabc}
are shown in m. 12–15, then the incomplete set of 24 general permutations h isi
shown. The time unit is t = 8 note, the grouping time signature is 12
1
8 .

Permutations of a set containing equal elements

Until now all set elements were different. Let’s consider how the number of permutations
decreases when there are equal elements in the set. For N = 3 this implies two equal ele-
ments Ne = 2, i.e., {a, a, b}. Now the reduced set of permutations becomes

{aab, aba, baa},

with the number of permutations given by

N! 3! 6
= = = 3. (9.1)
Ne !(N − Ne )! 2! 1! 2

Next case is three equal elements, Ne = 3 in a set with N = 4 elements: {a, a, a, b}. The
number of permutations is

N! 4! 24
= = = 4,
Ne !(N − Ne )! 3! 1! 6

i.e.,
{aaab, aaba, abaa, baaa}.

82 Copy issued to Seth Dempsey ([email protected])

 9.1. VARIATION THROUGH CIRCULAR AND GENERAL PERMUTATIONS

When there are two equal elements {a, a, b, c} in this N = 4 set, i.e., Na = 2, Nb = 1, Nc = 1,
we find
N! 4! 24
= = = 12, (9.2)
Na !Nb !Nc ! 2! 1! 1! 2
permutations, i.e.,

{aabc, aacb, abac, abca, acba, baca, bcaa, cbaa, acab, caab, baac, caba}.

For the case two equal subsets {a, a, b, b}, i.e., N = 4, Na = 2, Nb = 2 there are
N! 4! 24
= = = 6,
Na !Nb ! 2! 2! 4
permutations, i.e.,
{aabb, abab, abba, baba, bbaa, baab}.

9.1.2 Permutations in continuity and simultaneity

Given the set of general or circular permutations we may use them in music either as a con-
tinuity in the time domain or as a simultaneity in the pitch domain, distributed over multiple
staves. There is also the combination of both approaches.
Let’s illustrate this first in mathematical notation. For the two-element set (N = 2) we
find
Part Continuity Simultaneity Combination
P1 : ab ba or ba ab a or b ab ba or ba ab
P2 : b a ba ab ab ba
with the individual parts P1 and P2 are displayed as rows.
Repeating this approach for the circular permutations of the three-element set (N = 3)
we find a number of options for the single part continuity

Part Continuity
P1 : abc bca cab or abc cab bca or bca abc cab or
bca cab abc or cab abc bca or cab bca abc,
i.e., the sequence of three circular permutations can be ordered as a continuity on a higher
level in six possible ways, see also Section 9.2.
The simultaneous use of the circular permutations yields the following three-part options

Part Simultaneity
P1 : abc or abc or bca or bca or cab or cab
P2 : bca cab abc cab abc bca
P3 : cab bca cab abc bca abc.
Here the six possible vertical distributions over the three parts in the score are shown. Fi-
nally, the combination approach for the circular permutation of the three-element set implies
(not all options are shown)

Part Combination of continuity and simultaneity

P1 : abc bca cab or abc cab bca or abc bca cab
P2 : bca cab abc bca abc cab cab abc bca
P3 : cab abc bca cab bca abc bca cab abc.

Copy issued to Seth Dempsey ([email protected]) 83

CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

The essential aspect here is that all parts are playing their individually different continuity.
This gives the impression of a number of independent parts in a rhythmical counterpoint.
However, the resulting multi-part continuity pattern is based on a single small pattern; this
is why Schillinger calls it homogeneous variation. We have achieved another degree of freedom
to create a rhythmic pattern.

9.2 Permutations applied to musical parameters

Now that we have seen the process of obtaining the permutation sets and the creation of
a time continuity or a vertical distribution over multiple parts, let’s decide to what musi-
cal parameter we will apply the pattern. Mostly these patterns are either a two-element
attack-duration series, called a binomial, or a three-element series, called a trinomial.3 In a
binomial the periodicity of the pattern is determined by two coefficients, in a trinomial by
three coefficients (we encountered the monomial periodicity at the beginning of this book,
see Section 1.3.3).
Schillinger introduces a number of new aspects in this chapter. Here is his application
proposal:

Durations: Permutation of an attack-duration series ra is the predictable application after

reading the previous chapters. Suppose a two-element duration series, e.g., the bino-
mial ra = 2t + t, the continuity based on permutations yields rc = (2t + t) + (t + 2t)
or rc = (t + 2t) + (2t + t). However, there are other parameters that may undergo
permutation and variation.

Rests: When the duration series consists of a combination of note attacks and rests, we may
also apply permutations. Suppose we have the trinomial ra = t + 3t + 2r̄, where r̄
indicates a rest of a certain number of time units, we may create a continuity based on
general permutations. For example rc = (t + 3t + 2r̄) + (3t + t + 2r̄) + (2r̄ + 3t + t) + (t +
2r̄ + 3t), a continuity of four permutations. A possible simultaneity based on circular
permutations would be
 
(t + 3t + 2r̄) + (2r̄ + t + 3t) + (3t + 2r̄ + t)
rs =  (2r̄ + t + 3t) + (3t + 2r̄ + t) + (t + 3t + 2r̄)  .
 
(3t + 2r̄ + t) + (t + 3t + 2r̄) + (2r̄ + t + 3t)

Accents: Suppose there are accented notes in the duration series. These are equivalent to
rests in the pattern. Suppose we have the trinomial ra = t̂ + t + 2t, where t̂ indicates
the accented note, we may create a continuity based on circular permutations. For
example rc = (2t + t̂ + t) + (t̂ + t + 2t) + (t + 2t + t̂). A possible four-part combination
of continuity and simultaneity based on general permutations would be
 
(t̂ + t + 2t) + (t̂ + 2t + t) + (t + t̂ + 2t)
 (t + t̂ + 2t) + (t̂ + t + 2t) + (2t + t + t̂) 
rcs =  .
 
 (t̂ + 2t + t) + (2t + t + t̂) + (t + 2t + t̂) 
(2t + t + t̂) + (t + 2t + t̂) + (t̂ + t + 2t)
3
The labels binomial and trinomial might have been introduced earlier in this text, but adhering to the original
Schillinger book, they are being used from here on regularly.

84 Copy issued to Seth Dempsey ([email protected])

 9.2. PERMUTATIONS APPLIED TO MUSICAL PARAMETERS

Note that the number of parts Np = 4 here is greater than the number of permutations
in the continuity. Some parts may have simultaneous rests, but each part is playing an
independent rhythm. By now it should be obvious that we have a wealth of possibili-
ties to create rhythmic multi-part scores.

Split-unit groups: This is another new technique for creating variation. A given resultant
attack-duration series may be split into one or more subgroups. These subgroups with
their elements will be treated as constant, but we probably will have more options for
permutations than before. Let’s illustrate that with the simplest example, where the
original resultant is the binomial r = 2t + 2t. This pattern has no permutations. Now,
split one element into smaller time units: r = (t + t) + 2t. The two subgroups can now
be permutated, yielding the set {(t + t) + 2t, 2t + (t + t)} and this can be used for a
continuity or simultaneity
!
(t + t) + 2t + 2t + (t + t)
rc = [(t + t) + 2t] + [2t + (t + t)] or rcs =
2t + (t + t) + (t + t) + 2t

Another example is the trinomial r = 3t + 2t + 3t, with the first element split into
subgroups. There we have the options r = (2t + t) + 2t + 3t or r = (t + t + t) + 2t + 3t.
With these split-unit groups we may create a continuity such as rc = [(2t + t) + 2t +
3t] + [2t + 3t + (2t + t)] + [3t + (2t + t) + 2t] or a simultaneity
 
(t + t + t) + 2t + 3t
rs =  2t + (t + t + t) + 3t  .
 
3t + 2t + (t + t + t)

Groups in general: Schillinger now introduces permutation at a higher aggregation level.

Implicitly we have already encountered that option in Section 9.1.2, where the three-
part combination of continuity and simultaneity was shown. The simplest possible
option here is the two-attack binomial resultant r = 2t + t (total duration T = 3t). The
two permutations in continuity are ac = (2t + t) + (t + 2t) and bc = (t + 2t) + (2t + t).
Permutation of this higher level two-element set {a, b} yields rc,1 = ac bc = [(2t + t) +
(t + 2t)] + [(t + 2t) + (2t + t)] and rc,2 = bc ac = [(t + 2t) + (2t + t)] + [(2t + t) + (t + 2t)] with
total duration T 0 = 12t.4 And thus we may create a continuity rc = (rc,1 rc,2 ) + (rc,2 rc,1 )
with a total duration of T ” = 24t or a simultaneity
!
rc,1 rc,2 = ac bc + bc ac
rs =
rc,2 rc,1 = bc ac + ac bc
000
with total duration T = 12t. We have achieved a self-scaling property that may be
repeated forever at higher aggregation levels. This variation technique will be further
discussed in Chapter 10.

After all this mathematics it is appropriate to demonstrate the application of permuta-

tions in musical notation.
4
The notation of permutations used pairs ab. In this application section we will use that notation as an
abbreviation for r = ab = at + bt. The same holds for the higher level continuity rc = rc,1 rc,2 .

Copy issued to Seth Dempsey ([email protected]) 85

CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

9.3 Application to attack-duration groups

We will apply the permutation technique to two- (binomial), three- (trinomial) and four-
element attack-duration groups. The permutations may affect any of the five musical pa-
rameters listed in Section 9.2. The figures will group the examples in musical notation on
two, three or four staves in a score.

Example 9.4
Permutation of an attack group with two elements.
This example is based on two-element (N = 2) attack-duration groups with re-
sultant r = ab.

• The first case uses the binomial r = 2t + t, a = 2t, b = t and is shown

in rhythm notation in Fig. 9.2m. 1–8 as a combination of continuity (rc =
a
ab) and simultaneity rs = . Permutation is applied to various musical
b
parameters:
1. The note duration permutations are ab = (2t h i + t) + (t + 2t) and ba =
(t + 2t) + (2t + t), with time unit t = 4 and 34 meter. This is shown as a
1

continuity
 rc = ab on the upper staff in m. 1–2. Thesimultaneity in m. 1
a ab
is rs = , the combination in m. 1–2 is rcs = .
b ba
2. When the second attack is replaced by a rest, the pattern becomes r =
2t + r̄.5 Measure 3 shows h ithe continuity and simultaneity rcs at time
unit t = 8 and grouping 68 .
1

3. The accent on the second attack b = t̂ is demonstrated in m. 4.6 Accented

and regular attacks will sound simultaneously.
4. In order to demonstrate the split-unit group approach, we consider the
resultant r = 2t + 2t, a = 2t, b = 2t. Since both elements are equal
there is no permutation, hence no potential for creating variation. By
splitting the first element into two smaller units a = (t + t) we obtain
non-equal
h i elements. The rcs permutation is shown in m. 5–6 at group-
2
ing 4 . Remember that we keep the two smaller units together as a
constant subgroup; therefore the permutation is t + 2t + t is not allowed
here (that would be a three-element attack group).
5. Finally, in m. 7–8 we see the higher order permutation, that comes with
the
h i
groups in general approach. The time unit is t = 14 , the grouping at
6
4 meter. The upper staff shows the continuity rc,u = rc,1 rc,2 = abba =
[(2t + t) + (t + 2t)] + [(t + 2t) + (2t + t)], with rc,1 = ab = (2t + t) + (t + 2t)
and rc,2 = ba = (t+2t)+(2t+t). The lower staff contains rc,l = rc,2 rc,1 =
abba

baab, and the score has the combination rcs = .
baab
5
The rest with duration t is is indicated in the score as the text label [t] at the top, and is equivalent to r̄ in the
text. This alternative notation is simpler and faster in the music notation software.
6
The accent label above the staff is written as t’, and is equivalent to t̂ in the text.

86 Copy issued to Seth Dempsey ([email protected])

 9.3. APPLICATION TO ATTACK-DURATION GROUPS

• In Fig. 9.3, m. 1–2 the binomial attack-duration group r = 3t + t (N = 2, a =

3t, b = t) is distributed over three staves, with pitches based on the extended
tonic chord G69 , voiced in perfect fourths. There are two pitches per instru-
ment, with the higher pitch assigned to the a-element, and the other to b.
The distribution is based on the scheme
 
ab + ba
rcs =  ba + ab + ab + ba  .
 
+ ba + ab

Here we have deliberately used more freedom in creating a distributed

rhythm.

Example 9.5
Permutation of an attack group with three elements.
Now we look at an attack-duration group with three elements abc, N = 3. Several
variation through permutation approaches will be demonstrated:

• The first case is based on a major chord C distributed over two layers in
the score. The permutation patterns are shown in Fig. 9.2, m. 9–29. This is
another typical fingering example for the two hands in a keyboard part. The
attack-duration group is the trinomial r = 2t+t+3t, a = 2t, b = t, c = 3t with
time unit t = 18 , shown in the lower staff. It has an attack-group al = 1a +
1a + 1a (one attack for root, third and fifth in upward arpeggio style). The
upper part has three two-pitch attack groups, somewhat like a descending
chord arpeggio. The continuity of the three circular permutations of the
note durations rc = (abc) + (bca) + (cab) are shown in m. 9-11. The two
latter measures contain considerable cross-rhythm, although all measures
have simultaneous upper and lower staff attacks on the first beat.
• The same figure shows the introduction of a rest in the three-element group
r = 2t + r̄ + 3t. The bass part is now a two-attack group (root and third),
as
h iis the upper part. In m. 12–14 the afterbeat attacks create a nice groove in
6
8 time signature.

• In m. 15–23 there is an accent on the first attack element, r = 2t̂ + t + 3t with

time unit t = 18 . The lower staff uses a two-attack pitch distribution group
Np,l = 2 (root and fifth), the upper part a three-attack upward arpeggio of
theh extentions
i (3, 6 and 9) in perfect fourth voicing, Np,u . The grouping
4
at 4 meter yields syncopations. The three circular permutations yield an
attack-duration group continuity rc = (2t̂ + t + 3t) + (3t + 2t̂ + t) + (t + 3t +
2t̂), with Na = 9 attacks and total duration Tr = 18t. Synchronizing this
duration group with the lower attack group (see Section 8.2) leads to two
pattern statements with Npa = 18 = 2Na = 9Np,l = 3Np,u for recurrence
(the attack numbers were carefully chosen in the example). Synchronization

Copy issued to Seth Dempsey ([email protected]) 87

CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

r=2t+t 2t+[t] 2t+t' (t+t)+2t (2t+t)+(t+2t)

& 43 | Û Û | 68 Û ‰ ‰ Û 42 Û Û Û Û Û Û 46 | Û Û | Û | | Û 68
1
Û ÛÛÛ
JJ
>>
?3 Û 6 2 6 6
4 | | Û 8 ‰Û Û ‰ ÛÛ Û Û 4 Û ÛÛ ÛÛÛ 4 Û | | Û | Û Û | 8
J J
> >
j j œ œ œj œ
6

œ. œœ œ œœ œœ œ . œ. œœ ‰ œœ ‰ œœ .. ‰ œœ œœ œœ j
C C 9

& 68 œ . 44 œ œœœœ
9

œ œ. œœ œœ œ. œ.
J J >
r=2t+t+3t 2t+[t]+3t 2t'+t+3t

? 6 œ œ œ. j œ œ œJ œ ‰ œ . j œ ‰ 44 œ œ œ œ œ
8 J œ œ œJ œ. ‰œ œœ œ.
>

> > j . j >

& œ œ œ œ œ œ œ œj œ œJ œj œ . œ œ œ œ œ . œ œ œ œ œ œ œ œ œ j œ j
16

> > œ œ œ œ >œ J œ œ .

> > >
? œJ œ œJ œj œ . œ >œ œj œ . œ . >œJ œ œ œ œ œ >œ œ œ œ œ œ œJ œ œJ j > j
œ œ . œ œ œ œ . œ . œJ
>
> > > > >

>
(t+t)+[2t]+2t
j
& œ œ œ. 43 œœœ œœœ Œ
23

œ œœœ Œ œœœ œœœ œœœ œœœ Œ œœœ œœœ Œ œœœ œœœ œœœ œœœ œœœ œœœ Œ
> œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ
? œ œ œ œ œ >œ
G 13

43 œ œœ Œ œ Œ œ œ œœ œ Œ œœŒ œ Œ œ œ œ

r=4t+2t+3t+3t

12
& œœœ œœœ œœœ Œ 8 Û . Û Û Û. Û. Û Û Û Û Û Û Û. Û. Û. Û. Û Û Û. Û. Û Û Û.
29

œ œ œ J J J J J J
? Π12
8 Û Û Û Û Û Û Û. Û. Û. Û. Û Û Û. Û. Û Û Û. Û. Û Û Û. Û.
œœœ J J J J J J

j j j j j
[4t]+2t'+3t+(2t+t)

& Œ. œœœ ...

œœœ œœœ œœœ œœœ œœœ œœœ œœœ œœœ ‰ Œ . œœœ ... œœœ œœœ Œ . ‰ œœœ œœœ œœœ Œ . ‰ œœœ œœœ ...
34
‰ œœœ
> > > >
>œ œ .
? >œ œJ œ œJ œœ œ œJ Œ . ‰ >œ œ œJ Œ . Œ. ‰ œ œ
> . œ œ
D 13sus

‰ Œ. œ. ‰ J

Figure 9.2: Variation through permutation of attack-duration groups. Two-, three- and four-
element duration groups (N = 2, 3, 4) are assigned to two staves.

88 Copy issued to Seth Dempsey ([email protected])

 9.3. APPLICATION TO ATTACK-DURATION GROUPS

4 œ . œj œj œ . 6 ¿ . ¿ ¿j ¿ . ¿j ¿ ¿ ¿j ¿ ¿j ¿ ¿j ¿ . ¿j ¿ ¿j ¿ ¿j ¿ ¿ .
1 ab=3t+t abc=3t+2t+t acb bac bca cab cba

&4 ∑ 8
ba ab ab ba

& 44 j œ . œ . j jj 68 ∑ ∑ ∑ ∑ ∑ ∑
œ œ œ. œ œ œ.
œ œ. œ. œ 6
ba ab
?4 ∑ ∑ ∑ ∑ ∑ ∑ ∑
4 J J 8
6
G 9

j j j j
abc bca cab cab abc bca abc=3t+t+([2t]+2t)
4 œœ œœ .. œœ œœ .. ˙˙ œœ .. œœ œœ .. œœ ˙˙ œœ .. œœ Œ œœ
9
& œ. œ œ œ œ œ. œœ œœ 4 J J J J J
bca cab abc bca cab abc bca
j j j j 4 j j j j j j j
& œ œ œ. œ œ œ œ œ. œ œ 4 œœ .. œœ ˙˙ œœ œœ .. œœ œœ .. ˙˙ œœ .. œœ œœ Œ œœ œœ œœ ..

? œj œ œj œ œ . j j
cab abc bca abc=4t+3t+t bca cab cab

œ œ œ œ œ. 44 j j j j Œ œ j
˙ œ. œ œ. œ ˙ œ œ. œ œ. œ. œ
F/C Am

(2t+t+t)+(2t'+[t]+t+2t)+([t]+t)

œœ 43 œj œ œj œ ¿ ¿ ¿ >¿ ≈ ¿ ¿ ≈ ¿ ¿ ¿ ¿ >¿ ≈ ¿ ¿ ≈ ¿
16 bca cab cab abc

& œœJ Œ œœJ œœJ œœ .. Œ œœ œœ ..

J
j
cab abc bca bc ab ab
j j 3 œœ ≈ œœ œœ ≈ œœ œœ œœ œœ œœ ≈ œœ œœ œœ œœ œœ œœ
& Œ œœ œœ .. œœ œœ .. œœ Œ œœ 4 œ. œ œ œ œœ œœ œœ œ œœ œ œœœ
> > >
> >œ œ >œ
abc bca abc=4t+6t+2t ab bc bc
? j j j j 43 œ œ . j
œ. œ Œ œ œ Œ œ œ œ. œ œ œ œ œ ≈ œ œ ≈ œ œ ≈ ≈ œ œ
A m7(add4)

abac=t+(1/3t+1/3t+1/3t)+t+[t]

¿ ¿ ¿ >¿ ≈ ¿ ¿ ≈ ¿ 4 œ œ œ œ œ œ œœ œœ œœ œ œ œœ œœ œœ œœ œœ œœ œ
21 abc caba acab baca

& 4 œœœ Œ Œ œ Œ Œ œ
3 3 3 3
bc caba acab baca abac

& ≈ œœœ œœœ œœœ ≈ œœœ œœœ ≈ œœœ 44 Œ œ Œ œ Œ œ Œ

œ œ œ œœ œœ œ œ œ œ œ œ œœ œœ œ œ œ œœ
3 3 3 3

> œ œ œ
? ≈ œ
ab acab baca abac caba
44 Œ
œ œ œ >œ ≈ œ œ œ œ œœœ œœœœ Œ œ œ œœœœ Œ Œ œ œœœœ
3 3 3 3
A m11

Figure 9.3: Variation through permutation of attack-duration groups. Two-, three- and four-
element duration groups (N = 2, 3, 4) are distributed over three staves in the score.

Copy issued to Seth Dempsey ([email protected]) 89

CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

with the time signature TM = 8t results in a nine-measure pattern (NM = 9),

since Ttot = 72t = 9TM = 4Tr . This illustrates the combination of two
techniques, each already generating longer patterns: synchronization and
permutation.
• In Fig. 9.2, m. 24–29, there is a three-element group r = a + b + c with a split-
unit group for the first element a = (t + t), a rest for the second element
b = 2r̄ and a regular attack for the third element c = 2t. Not splitting the
first element would imply that a = c = 2t, two equal elements, and a total
set of only three permutations. The duration series is applied to an extended
dominant 7th chord G13 7 with two pitches in the lower staff (root and fifth in
the left hand) and a four-pitch chord in the upper staff. The lower staff has
an attack group rp = a + 2a, i.e., three attacks; this was deliberately chosen,
in agreement with the three attacks Na = 3 in the duration series.
The resultant attack-duration trinomial r = (t + t) + 2r̄ + 2t goes through all
general permutations according to the scheme
!
abc + bac + cba + bca + cab + acb
rcs = .
cab + cba + acb + abc + bca + bac
h i
The time unit is t = 81 the meter is 34 , a pattern with a total length of
NM = 6 measures.
• All six general permutations of the three-element group with attack-
duration group r = 3t+2t+t are shown in Fig. 9.3, m. 3–8. The ×-noteheads
indicate that this could be a cymbal pattern. The time unit is t = 81 , group-
h i
ing is at time signature 68 . Some permutations lead to a syncopated or
afterbeat rhythm. The continuity rc = abc + acb + bac + bca + cab + bca.
• The three circular permutations of the same attack-duration group r = 3t +
2t + t are shown as a combination of continuity and simultaneity in the
same figure in m. 9–11. The distribution is based on the inverted F/C triad,
distributed over a score with three staves using the scheme
 
abc + bca + cab
rcs =  bca + cab + abc  .
 
cab + abc + bca

• In Fig. 9.3, m. 12–14 the same circular permutation approach is applied to a

five-part minor chord Am, with two pitches in the upper and middle staff.
The attack-duration group
h i is r = 4t + 3t + t (see bottom staff), time unit
t = 8 and grouped at 44 .
1

• This pattern is modified in m. 15–17, swapping elements and introducing a

split-unit group. The attack-duration group is r = 3t + t + (2r̄ + 2t), with
a = 3t, b = t and the split-unit element c = (2r̄ + 2t). This creates the
impression of a rhythmical counterpoint in three parts.
• Another three-element simultaneity is shown
h i in m. 18 for r = 4t + 6t +
2t, which groups into a single measure at 34 meter with the three circular
1
permutations distributed vertically. The time unit is the 16th note t = 16 .

90 Copy issued to Seth Dempsey ([email protected])

 9.3. APPLICATION TO ATTACK-DURATION GROUPS

• The next case, m. 19–21, introduces a new combination with more freedom.
The basic pattern is the attack-duration group from the previous example
r = 4t + 6t + 2t, but now containing three split-unit groups with accents
and rests r = (2t + t + t) + (2t̂ +
h r̄i + t + 2t) + (r̄ + t). The time unit is
t = 16 , the grouping remains at 34 meter. The upper staff does not use
1

permutations, but plays an ostinato percussion rhythm with total duration

Tu = 12t. The two lower staves are based on the extended minor chord Am47
and use two-element subgroups in permutations, according to the scheme
!
bc + ab + ab + bc
rcs = ,
ab + bc + bc + ab

which also uses the higher level permutation of the groups in general ap-
proach: rc = (bc + ab) + (ab + bc). The middle staff plays three-part chords
in perfect fourths, the bass line has three pitches with root assigned to the
a-element, b is the chord fifth and the 7th goes to the c-element. The dura-
tion of the (bc + ab) pattern is Tl = 18t; synchronization of the upper staff,
lower staves and the time signature therefore requires one repeat to achieve
full measure recurrence. This freedom and combination of techniques is not
discussed in the Schillinger text; it is a new interpretation and creative use
of the permutation and homogeneous variation approach.
• Distribution of a three-element group over four staves is demonstrated in
Fig. 9.4, m. 1–6. The attack-duration group is theh trinomial
i r = 5t + 3t + 2t,
1 10
with time unit t = 8 , grouped at time signature 8 . The example is meant
to show the possibility of using variable density: sometimes there are two
instruments playing, sometimes three while the setting is moving through
the score. All six general permutations are used, but first there are three
circular permutations in the lead part, later the other three permutations are
used. This yields the following distribution scheme
 
abc + cab + acb + bac
 bca + cab + cba + bac 
rs,c = .
 
 abc + bca + acb + cba 
+ abc + acb
Note that all four instruments are playing their own continuity rc with two
or four patterns, but without a repeat.

Example 9.6
Permutation of an attack group with four elements.
In this case the resultant attack-duration group consists of four elements abcd,
N = 4. There are 24 general permutations; using all of these will create very long
patterns, so it is more likely to use either a subset from the general permutations,
or use the four circular permutations. We will apply these alternatives to different
musical parameters.

Copy issued to Seth Dempsey ([email protected]) 91

CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

œ. œ œ. œ œœ œœ .. œœ œœ .. œœ .. œœ œœ œœ .. œœ .. œœ œœ .. œœ 4
abc=5t+3t+2t cab acb bac

& 10
8 œ. œ œ. œ
1
∑ ∑ 4
bca cab cba bac

& 10
8 œœ .. œœ œœ .. œœ œœ œœ .. œœ œœ .. ∑ œœ œœ .. œœ .. œœ œœ .. œœ œœ .. œœ ∑ 44

œœ .. œœ œœ .. œœ œœ .. œœ œœ .. œœ œœ .. œœ œœ œœ .. œœ œœ .. œœ .. œœ 4
abc bca acb cba
? 10
8 ∑ ∑ 4
abc acb
? 10
8 ∑ ∑ œœ .. œœ œœ .. œœ ∑ ∑ œœ .. œœ œœ œœ .. 44

2t+([t]+t)+2t+([t]+t) abab abba aabb

& 44 ¿ ‰¿ ¿ ‰¿ ¿ ‰¿‰¿ ¿ ¿ ¿ ‰¿‰¿

7
∑ ∑ ∑ J J J J
j j j j j j
abab abba aabb baab

& 44 ∑ ∑ ¿ ‰¿ ¿ ‰¿ ¿ ‰¿ ‰¿ ¿ ¿ ¿ ‰¿ ‰¿ ‰¿ ¿ ¿ ‰¿
abab abba aabb baab baba
?4 ∑ ¿ ‰¿ ¿ ‰¿ ¿ ‰¿ ‰¿¿ ¿ ¿ ‰¿ ‰¿ ‰¿ ¿ ¿ ‰¿ ‰¿ ¿ ‰¿¿
4 J J J J J J J J
abab abba aabb baab baba bbaa
? 4 ‰ j ‰ j ‰ j‰ j j j j j
4 ¿ ¿¿ ¿ ¿ ¿ ¿¿ ¿ ¿ ‰¿‰¿ ‰¿ ¿ ¿ ‰¿ ‰¿ ¿ ‰¿ ¿ ‰¿ ‰¿¿ ¿
abcc
> > [t]+t+t+t
œ œ ‰ œ ‰ œ œ œ œ ‰ œ œ œ œ ‰ œ œ œ œj ‰ 12
abbb

œ œ ‰ œ ‰ œ ‰ œJ œ
13 2t+(t+t)+([t]+t')+([t]+t) cabc babb bbab bbba

&œ 8
bcca abcc babb bbab bbba abbb

& œ œ ‰ œ ‰ œj œ œ œ œ ‰ œ ‰ œ œ ‰ œ œ œ œ ‰ œ œ œ œ ‰ ‰ œ œ œ 12
8
> >
? ‰ >œ ‰ œ œ œ œ œ œ ‰ >œ ‰ œ œ œ œ ‰ œ œ œ œ œ œ 12
‰ ‰ œ œ œ ‰ œ
ccab bcca bbab bbba abbb babb

J 8
œ œ ‰ >œ ‰ >œ ‰ œ œ œ œ œ œ ‰ ‰ œ œ œ œ ‰ œ œ œ œ ‰ œ 12
cabc ccab bbba abbb babb bbab
?‰ œ
J œ œ 8

> > >

aabbb=3t+3t+2t+2t+2t' abbba bbaab

& 12
8 ¿. ¿. ¿ ¿ ¿ ¿. ¿ ¿ ¿ ¿ ¿.
17
¿ ¿ ¿¿ ¿¿ ¿¿ ∑
J J J J J J
abbba bbaab aabbb

& 12
8 ¿. j j
¿ ¿ ¿ ¿ ¿.
j j
¿ ¿ ¿ ¿
j j
¿ ¿ ¿ ¿ ∑ ¿. ¿. ¿ ¿ ¿
> > >
> ¿. ¿. ¿ ¿ >¿ ¿ . > .
? 12 ¿ ¿J ¿J ¿ ¿J ¿ ¿¿ ¿ ¿ ¿¿ ¿
bbaab aabbb abbba

8 J ∑ J J
> > >
aabbb abbba bbaab
? 12
8 ∑ ¿. ¿. ¿ ¿ ¿ ¿. ¿ ¿ ¿ ¿ ¿. ¿ ¿ ¿¿ ¿¿ ¿¿
J J J J J J

Figure 9.4: Variation through permutation of attack-duration groups. Three- and four-
element duration groups (N = 3, 4) are distributed over four staves in the score.

92 Copy issued to Seth Dempsey ([email protected])

 9.3. APPLICATION TO ATTACK-DURATION GROUPS

• In Fig. 9.2, m. 30–33 there is the four-element

h i group r = 4t + 2t + 3t + 3t,
1 12
time unit t = 8 , time grouping at 8 . So a = 4t, b = 2t, c = 3t, d = 3t. For
the varied pattern we use the four circular permutations, according to the
two-staff scheme
!
abcd + bcda + cdab + dabc
rcs = ,
bcda + cdab + dabc + abcd

which yields a four-measure pattern, shown in rhythmical notation.

• The previous example is modified in m. 34–37 in the sense that we introduce
a rest for the first element a = 4r̄, the second element attack is accented
b = 2t̂ and the fourth element becomes a split-unit group d = (2t + t) with
9/sus4
non-equal units. The extended dominant chord D7 is distributed over a
lower layer with three pitches and attack group rp,l = a(p1 ) + a(p2 ) + a(p3 ) +
a(p2 ). Again, this example was constructed with Np,l = Na = 4. The upper
layer plays three-part chords. Homogeneous variation was based on the
four circular permutations using the scheme from the previous example.
The result might be used as a rhythmic groove in Afro-Cuban style Latin
music.
• In Fig. 9.3, m. 22–25, there is the four-element group r = abac = t+(h13 t+
i 3
1
t+
1
3 t)+t+ r̄, with a triplet elements b, time unit t = 41 and grouping at 44 . This
attack-duration group is distributed over three staves, the chord structure is
9/11
Am7 . The example uses a continuity of four circular permutations rc =
abac + caba + acab + baca. This application demonstrates two effects: first
there is the division of a time unit into triplets, and then there is the variable
density setting within the staff and between staves. This implies that at any
time instant there are between two and four pitches sounding. Each staff
uses different pitches for the a- and b-element.
• Distribution over a four-staff score in demonstrated in Fig. 9.4. In m. 7–12
there is the attack-duration group r = abab = 2t + (r̄ + t) + 2t + (r̄ + t), with
a split-unit b-element. This is a typical kick drum pattern in Rock music.
Since both a and b occur twice, or, Na = Nb = 2, there are

N! 4! 24
= = =6
Na ! Nb ! 2! 2! 4

general permutations. These are shown as a continuity in the lower staff

rc = abab + abba + aabb + baab + baba + bbaa; note the position of the synco-
pated and afterbeat 8th notes. The upper parts imitate this pattern, yielding
a variable density setting that combines more and more permutations in
parallel as a simultaneity.
• In the same figure, in m. 13–14, we find the attack-duration group with split-
unit elements r =h abcc
i
= 2t + (t + t) + (r̄ + t̂) + (r̄ + t), time unit t = 18 ,
grouped at meter 44 . Ignore the accent on the first c-element, which is not
considered a separate element for this example. We use the four circular

Copy issued to Seth Dempsey ([email protected]) 93

CHAPTER 9. HOMOGENEOUS SIMULTANEITY AND CONTINUITY

permutations as a combination of four-part simultaneity and two patterns

in continuity, using the scheme
 
abcc + cabc
 bcaa + abcc 
rcs =  .
 
 ccab + bcca 
cabc + ccab
Note how in this syncopated pattern the accents are moving through the
parts; this has the flavour of a Latin music percussion section. In fact, the
rhythm in the first staff rc = abcc + cabc is a variation of the cascara pattern
for timbales.
• The final four-element group in this example is in m. 15–16, with the attack-
duration group r = abbb = r̄+t+t+t. Since Nb = 3, there are N !/(Na ! Nb !) =
4!/(1! 3!) = 24/6 = 4 general permutations, which are also the circular per-
mutations. These are shown as a continuity rc = abbb + babb + bbab + bbba,
with two pitches in each staff. Together these form a minor 6-pitch scale
d − e − f − g − a − d, giving the setting a minimal music cell flavour.

Example 9.7
Permutation of an attack group with five elements.
In this case the resultant attack-duration group consists of five elements abcde
(N = 5). This has not been discussed in the text. There are 120 general per-
mutations and 5 circular permutations. We will consider a special case of the
five-element attack group only.
• In Fig. 9.4, m. 17–21 we use the five-element
h i group r = 3t + 3t + 2t + 2t + 2t,
1 12
time unit t = 8 , time grouping at 8 . This is the characteristic fandango
dance rhythm, also used in Renaissance music and Afro-Cuban Latin music.
So a = 3t, b = 2t (ignore the accent on the last note in the score; it serves easy
reading of the example). This pattern has
N! 5! 120
= = = 10
Na !Nb ! 2! 3! 12
general permutations. We will use three permutations {aabbb, abbba, bbaab}
and distribute these over the four staves, according to the scheme
 
aabbb + abbba + bbaab
 abbba + bbaab + aabbb 
rs,c =  ,
 
 bbaab + aabbb + abbba 
aabbb + abbba + bbaab
a four-measure pattern shown in rhythmical notation. This demonstrates a
case where the constant three-instrument density moves through the staves,
creating an empty measure between phrases. Each part is playing a conti-
nuity rc without repeats, with an imitative counterpoint feeling.

94 Copy issued to Seth Dempsey ([email protected])

Chapter 10

Generalization of variation techniques

This chapter elaborates the concept of higher order permutations as a generalization of varia-
tion techniques. The technique was in fact introduced at the first higher level in Section 9.2,
where it was called Groups in general. Here we will investigate the sets of permutations that
arise when we extend the technique to higher levels.
We will learn that the number of possiblities grows exponentially, when we perform
higher order permutations. Therefore, in this chapter we will discuss only simple cases for
the two- and three-element basic groups.

10.1 Higher order permutations of a two-element group

The starting point is the core two-element group at the lowest level L0 : {a0 , b0 } = {a, b},
where {a, b} might be an attack-duration group, e.g., the binomial {a, b} = {3t, t} with re-
sultant r = 3t + t. The core elements may contain notes, rests, accented notes, or split-unit
groups (see the list of attributes in Section 9.2).
Combining the elements at the first application level yields a continuity rc with two per-
mutations rc (L1 ) = {a1 , b1 } = {a + b, b + a}. What happens when we extend this approach
to higher levels is illustrated in Fig. 10.1.
The number of elements in the n-th layer Ne,n = 2n , creating elongated but unique
rhythms at the higher level. The original attack-duration group a + b returns as a subgroup
at each level, but in different combinations with its retrograde, i.e., the time-reversed pattern
b + a. Higher order permutations of a two-element group will now be illustrated with an
example in rhythmical notation.

Example 10.1
Higher order permutations of a two-element group.
This example is based on two-element attack-duration groups at level L0 : {a, b}.
Figure 10.2 shows four staves, that represent the higher order levels L1 , . . . , L4 .

• The first case uses the two-element attack-duration pattern {a, b}, a = 3t, b =
t with time unit t = 81 and is shown in Fig. 10.2 m. 1–6. The first higher order
permutation, level L1 consists of the two possible permutations a+b = 3t+t
and b + a = t + 3t (the figure uses the abbreviate notation ab = a + b; the
measures with rests separate the patterns). The order of the elements is

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 10. GENERALIZATION OF VARIATION TECHNIQUES

Level Ln : an = an−1 + bn−1 = . . . bn = bn−1 + an−1 = . . .

.. ..
. .
Level L4 : a4 = a3 + b3 = (abbabaab) + . . . b4 = b3 + a3 = (baababba) + . . .

Level L3 : a3 = a2 + b2 = (abba) + (baab) b3 = b2 + a2 = (baab) + (abba)

Level L2 : a2 = a1 + b1 = (ab) + (ba) b2 = b1 + a1 = (ba) + (ab)

Level L1 : a1 = a0 + b0 = (ab) b = b0 + a0 = (ba)

i
P
PP1 1

6 PP 6
Level L0 : a0 = a, b0 = b

Figure 10.1: Higher order permutations of a two-element group. Continuity rc (Li ) at level
Li , i = 1, . . . , 4 for the two-element group {a, b} at the lowest level L0 . Abbreviated notation:
(ab) = a + b; the order is relevant for the rhythm.

relevant. At level L2 the permutations of the L1 patterns h i are L2 : {(a +

4
b) + (b + a), (b + a) + (a + b)}. Grouped at meter 4 this leads to one-
measure patterns. Permutation of these two patterns produces the set L3 :
{[(a + b) + (b + a)] + [(b + a) + (a + b)], [(b + a) + (a + b)] + [(a + b) + (b + a)]}, a
pair of two-measure patterns. Finally, at level L4 only one pattern is shown
rc = [(a+b)+(b+a)+(b+a)+(a+b)]+[(b+a)+(a+b)+(a+b)+(b+a)]. The
self-scaling property is visible with the original a- and b-element returning
in various positions and groupings along the continuity rc .
• The second case is shown in Fig. 10.2 m. 7–15 and is based on a two-element
attack-duration group L0 : {a, b} with the split-unit group a = (2t + t + t)
and the accented note b = 2t̂ at time unit t = 18 .1 Again, the process is
demonstrated as higher order permutations at levels L1 , . . . , L4 . At the first
level this yields the set L1 : {a +
h bi= (2t + t + t) + 2t̂, b + a = 2t̂ + (2t + t + t)},
each a one-measure pattern at 34 grouping.
• The third example is shown in m. 16–21 for L0 : {a, b} with two split-unit
groups a = (4t + t + t) and b = (2t + 2r̄ + 2t̂), the latter containing
h i a rest
1 12
and an accented note. The time unit is t = 16 , grouped at 8 . At L4 the
first permutation L4 : rc = [(a + b + b + a) + (b + a + a + b)] + [(b + a +
a + b) + (a + b + b + a)] is a four-measure
h i
pattern. At time unit t = 81 (the
augmentation variation) and 34 time signature this pattern could be used
as a waltz rhythmical background.

1
Once again, the notation in the score is slightly different from the text. An accented note is notated as t0 in
the score and as t̂ in the text. The rest is notated as r in the score and corresponds to r̄ in the text.

96 Copy issued to Seth Dempsey ([email protected])

 10.1. HIGHER ORDER PERMUTATIONS OF A TWO-ELEMENT GROUP

& 44 œ . œJ œJ œ . œ œ. œ. œ œ œ. œ. œ œ. œ œ œ. 43
1 L4: abbabaabbaababba

J J J J JJ ∑ ∑

& 44 œ . œJ œJ œ . œ œ. œ. œ œ œ. œ. œ œ. œ œ œ. 43
L3: abbabaab baababba

J J ∑ ∑ J J JJ
? 44 œ . œ œ œ . œ œ. œ. œ 43
L2: abba baab

JJ ∑ ∑ ∑ J J ∑

? 44 œ . œ Ó œ œ. Ó 43
L1: ab = 3t+t ba

J ∑ ∑ ∑ J ∑

& 43 œ œ œ >œ >œ œ œ œ >œ œ œ œ œ œ œ >œ >œ œ œ œ œ œ œ >œ œ œ œ >œ >œ œ œ œ 12

7 L4: abbabaabbaababba

∑ 8

& 43 œ œ œ >œ >œ œ œ œ >œ œ œ œ œ œ œ >œ >œ œ œ œ œ œ œ >œ œ œ œ >œ >œ œ œ œ 12

L3: abbabaab baababba

∑ 8

? 3 œ œ œ >œ >œ œ œ œ >œ œ œ œ œ œ œ >œ 12

L2: abba baab

4 ∑ ∑ ∑ ∑ ∑ 8

? 3 œ œ œ >œ >œ œ œ œ 12
L1: ab=(2t+t+t)+2t' ba

4 ∑ ∑ ∑ ∑ ∑ ∑ ∑ 8

& 12
8 œ œ œ œ ‰ >œ œ ‰ >œ œ œ œ œ ‰ >œ œ œ œ œ œ œ œ ‰ >œ œ ‰ >œ œ œ œ œ œ œ œ ‰ >œ
16 L4: abbabaabbaababba

& 12
8 œ œ œ œ ‰ >œ œ ‰ >œ œ œ œ œ ‰ >œ œ œ œ œ œ œ œ ‰ >œ
L3: abbabaab

∑

? 12
8 œ œ œ œ ‰ >œ œ ‰ >œ œ œ œ
L2: abba

∑ ∑

? 12
8 œ œ œ œ ‰ >œ Ó .
L1: ab =(4t+t+t)+(2t+2r+2t')

∑ ∑

œ œ œ ‰ >œ œ ‰ >œ œ
19
&œ œœ ∑ ∑

œ ‰ >œ œ œ œ œ ‰ >œ œ œ œ œ ‰ >œ œ ‰ >œ œ œ œ

baababba

& ∑ œœœ

? œ ‰ >œ œ œ œ œ ‰ >œ
baab

∑ œœœ ∑

? œ ‰ >œ œ œ œ Ó.
ba

∑ ∑

Figure 10.2: Higher order permutations of a two-element group {a, b}: example in rhythm
notation.

Copy issued to Seth Dempsey ([email protected]) 97

CHAPTER 10. GENERALIZATION OF VARIATION TECHNIQUES

10.2 Higher order permutations of a three-element group

The starting point now is a three-element group L0 : {a0 , b0 , c0 } = {a, b, c} and resultant
r = at + bt + ct, where each element may consist of a note, rest, accented attack, or split-unit
group. We will see that the higher order permutations contain a wealth of rhythmic varia-
tions. In fact the number will become so large, that Schillinger in his book only considers
the second higher order level L2 . He also writes the possible cases now as combinations of
elements, where the order is irrelevant. For each combination the number of permutations
Np is indicated.
We will consider the two possible cases of combinations by two or by three elements at
the first level L1 .

10.2.1 Combinations by two elements at the first higher order level

The two-element combinations at the first higher order level are threefold L1 : {a1 , b1 , c1 } =
{ab, ac, bc}. Each of these combinations has two permutations, e.g., a1 = (a + b) or the
retrograde ar1 = (b + a). At the next higher order level, L2 we may use either two-element or
three-element combinations, discussed as two separate cases.

Level 2: Combination by two elements

At the second higher order level L2 we use combinations of the two elements from the first
higher order level. This yields the following set

a1 + ar1 ar1 + b1 b1 + br1 br1 + c1 c1 + cr1

a1 + b1 ar1 + br1 b1 + c1 br1 + cr1
a1 + br1 ar1 + c1 b1 + cr1
a1 + c1 ar1 + cr1
a1 + cr1

How should we interpret this set? The first combination at this level is L2 : a1 + ar1 =
(a + b) + (b + a), i.e., a rhythmic pattern with four elements. This combination has two
permutations {a1 + ar1 , ar1 + a1 } or, equivalently {(a + b) + (b + a), (b + a) + (a + b)}. This
property holds for all combinations in the set. There are 15 two-element combinations at L2 ,
each with two permutations, so the total set contains 15 × 2 = 30 cases.
Three of the combinations use two elements only: {a1 + ar1 , b1 + br1 , c1 + cr1 }, which are
based on the combinations {a, b}, {a, c} and {b, c}, respectively. These cases are analogous
to the two-element group higher order permutations, discussed in Section 10.1.
The other combinations use all three elements {a, b, c} and have a double occurrence of
one group element. For example, L2 : a1 + b1 = (a + b) + (a + c) (a is used twice) or
L2 : br1 + cr1 = (c + a) + (c + b) (now c is used twice). Higher order permutations of a
three-element group with combinations by two elements at level L1 will be demonstrated in
an example. We will not show all the combinations and permutations in musical notation;
however, there will be some possible patterns at the level L3 .

Example 10.2
Higher order permutations of a three-element group: combinations by two

98 Copy issued to Seth Dempsey ([email protected])

 10.2. HIGHER ORDER PERMUTATIONS OF A THREE-ELEMENT GROUP

elements.
This example is based on three-element attack-duration groups with resultant
r = abc at level L0 . These are shown in musical notation on a single staff; the
pitches a, b, and c are chosen arbitrarily. These are used to make reading of the
rhythmical patterns easier.

• The first case uses the three-element set L0 := {a, b, c} = {2t, t, 3t} at time
unit t = 81 and is shown in Fig. 10.3 m. 1–37. The three two-element
combinations at the first higher order are L1 : {a1 , b1 , c1 } = {ab, ac, bc} =
{(2t + t), (2t + 3t), (t + 3t)} (m. 2–4). Note that these combinations have
different lengths of 3t, 5t and 4t, respectively. At the next level L2 the ex-
haustive set of 15 combinations is shown in m. 5–21. Each lower level com-
bination is used in the original (e.g., a1 = a + b) and retrograde form (as
ar1 = b + a, labeled with an accent a01 in the staff). For example, the first com-
bination is L2 : a1 + ar1 = (a + b) + (b + a) = (2t + t) + (t + 2t). Each of these
combinations has two permutations, not shown in the figure. Various me-
ters are used to enable easy reading. Some combinations use an incomplete
original set of elements (shown as two pitches only).
Then, at the next higher level L3 some possibilities are demonstrated in
m. 22–37. At this level again we may choose various combinations. This
example shows two combinations by two elements, L3 : (a1 + ar1 ) + (ar1 + a1 )
with total lenght T = 12t and L3 : (a1 + b1 ) + (a1 + c1 ), T = 15t and one
combination by three elements L3 :h (bi1 + a1 ) + (br1 + c1 )h +i (cr1 + ar1 ). Note
that the grouping with the meters 44 in m. 24–26 and 68 in m. 30–35 re-
quires a pattern repeat before there is recurrence at the full measure. These
are clear evidence of achieving a non-repetitive and interesting longer, i.e.,
multi-measure rhythmical pattern from a basic set of elements. This is ho-
mogeneity in variation through the application of higher order combina-
tions and permutations.
• The second case is based on the trinomial L0 : {a, b, c} = {2t, (t + t̂), (2t +
r̄ + t)} with two split-unit groups, shown in Fig. 10.3 m. 38–62, at time unit
t = 81 . The three two-element combinations at level L1 are shown in m. 39–
41. Nine of the possible two-element combinations at level L2 are shown
in m. 42–50. For example, L2 : a1 + b1 = (a + b) + (a + c) = [2t + (t +
t̂)] + [2t + (2t + r̄ + t)}] with total duration T = 10t. This combination
has two permutations {a1 + b1 , b1 + a1 }. From these combinations, three
possible examples at level L3 are shown, two based on the combination by
two elements, L3 : (a1 + b1 ) + (c1 + a1 ) and L3 : (b1 + cr1 ) + (cr1 + c1 ), and one
using a combination of three elements L3 : (a1 + b1 ) + (b1 + c1 ) + (c1 + a1 ).
These
h i combinations were deliberately chosen with a specific meter (here
4
4 ), in order to have recurrence after one statement.

• Finally, we demonstrate the case L0 : {a, b, c} = {r̄ + t, ( 32 t + 23 t + 23 t), (t +

1 1 1
2 t + 2 t)} with three split-unit groups in m. 63–79. The time unit is t = 8 .
All three combinations at level L1 are shown in m. 64–66, and a subset of
combinations by two elements at level L2 in m. 67–71. These are L2 : {a1 +
b1 , a1 + c1 , b1 + c1 , c1 + a1 , cr1 + a1 }. Two patterns at the next higher order

Copy issued to Seth Dempsey ([email protected]) 99

CHAPTER 10. GENERALIZATION OF VARIATION TECHNIQUES

level are L3 : {(a1 + b1 ) + (b1 + c1 ) + (cr1 + a1 ), (a1 + c1 ) + (cr1 + a1 )}. The latter
combination by two elements at L3 : rc = (a1 + c1 ) + (cr1 + a1 ) requires three
statementsh i(two repeats) in order to achieve recurrence at the full measure
in meter 64 .

After carefully studying these examples the potential of this variation technique and the
plethora of options to create longer rhythmical continuities rc from a given set of simple
source elements should be obvious. There is no reason to be afraid of composing boring
repetitive rhythms; suddenly we find ourselves in a mer à boire, from which to choose tasteful
and suitable combinations.2

Level 2: Combination by three elements

Alternatively, we use combinations of three elements at the level L2 . Now the set becomes

a1 + ar1 + b1 ar1 + b1 + br1 b1 + br1 + c1 br1 + c1 + cr1

a1 + ar1 + br1 ar1 + b1 + c1 b1 + br1 + cr1
a1 + ar1 + c1 ar1 + b1 + cr1 b1 + c1 + cr1
a1 + ar1 + cr1 ar1 + br1 + c1
a1 + b1 + br1 ar1 + br1 + cr1
a1 + b1 + c1 ar1 + c1 + cr1
a1 + b1 + cr1
a1 + br1 + c1
a1 + br1 + cr1
a1 + c1 + cr1

The first element is L2 : a1 + ar1 + b1 = (a + b) + (b + a) + (a + c), a rhythmic pattern with

six elements. It consists of three non-equal subgroups and therefore has six permutations.
The set contains 10 + 6 + 3 + 1 = 20 combinations. The total number of cases therefore is
20 × 6 = 120.
All combinations contain the full original three-element group {a, b, c}. The first com-
bination L2 : a1 + ar1 + b1 contains a three times, b twice and c once. This count holds for
(r) (r) (r)
all combinations, except for eight combinations of L2 : a1 + b1 + c1 where each element
occurs twice. Already we can see how the number of possibilities for higher order permuta-
tions increases enormously.

Application Tip:
In the discussion in the text we have covered the higher order permutations as
separate levels. However, when you look at the examples, you will discover that
a juxtaposition of the levels creates application potential in development sections
or solo cadenzas in longer pieces. In such a case you would most likely start with
the original level L0 core material as a rhythmic cell and gradually work your
way up to the higher levels, e.g., L0 → L1 → L2 → . . . → Ln . Or you could
2
Warning: do not use this technique when you are specializing in EDM, techno or house music. Could cost
you your job and will likely lead to sleeping under the bridge.

100 Copy issued to Seth Dempsey ([email protected])

 10.2. HIGHER ORDER PERMUTATIONS OF A THREE-ELEMENT GROUP

& 68 œ œJ œ . 38 œ œ 85 œ œ . 48 œ œ . 68 œ œ ,œ œ 44 œ œ œ œ . œ œ œ . œ 78 œ œ œ œ .
1 L0: abc L1: a1=ab b1=ac c1=bc L2: a1+a1' a1+b1 a1+b1' a1+c1

J J JJ J J JJ

j j
& œ œJ œ . œJ 44 œJ œ œ œ œ . œ œ œJ œ œ 78 œ œ œ œ . œ œ œ . œ 85 œ œ . œ . œ 98
9 a1+c1' a1'+b1 a1'+b1' a1'+c1 a1'+c1' b1+b1'

J J J J J

& 98 œ œ . œJ œ . œ œ . œ . œJ œ . œ œJ œ . œ . œ œ . œJ 42 œJ œ . œ . œJ 68 œ œJ œJ œ œJ œ œ œJ 44
16 b1+c1 b1+c1' b1'+c1 b1'+c1' c1+c1' L3: (a1+a1')+(a1'+a1)

j , j
& 44 œ œ œ œ œ œ œ œ œJ œ œ œ œ œ œ œ œ œJ 85 œ œJ œ œ. œ œ œ œ. 68 œ œ œ œ
24 (a1+b1)+(a1+c1)

J J

j , j 4 œ œ œ œ œ œ . œj œ œ œ
& œ œ œ œ œ œ . œ œJ œ œJ œ œ œ œ œ œ . œ œ. œ œ œ
31 (b1+a1)+(b1'+c1)+(c1'+a1')

4 J J J

> œ œ 2 L1: a1=ab

38 L0: abc
> 3 b1=acœ œ c1=bc
> œ œ 4 L2: a1+a1'
> > 5
a1+b1
> œ œ
& œ œ œ ‰ J 4 œ œ œ 4 œ ‰ J œ œ ‰ J 4 œ œ œ œ œ œ 4 œ œ œœ ‰J

œ œ > c1'+a1
œ œ > c1'+a1'
œ œ > 6 b1+c1
œ > œ œ b1+c1' >
‰ J œ œ ‰ J œ œ ‰ œJ œ ‰ œJ œ œ
œ
44 c1+a1
&œ ‰Jœ œœ ‰Jœ œ œœ ‰Jœ œœœ 4 œ

œ œ > œ œ b1'+c1'
œ œ œ œ > 4 L3: (a1+b1)+(c1+a1)
> >, >
œ 4 œ œœœ œ ‰œœ œ ‰œ œ œœœ œœ
49 b1'+c1
& ‰ œ œ œ ‰ ‰ œ ‰ œ
J J J J J J

> > > >

& œ œ ‰ œJ œ œ ‰ œJ œ œ œ œ œ œ œ œ ‰ œJ œ œ ‰ œJ œ œ œ ‰ œJ œ œ ‰ œJ œ œ œ œ œ ‰ œJ œ
54 (a1+b1)+(b1+c1)+(c1+a1) (b1+c1')+(b1'+c1)

61
œ >œ œ > œ œ 3 L0: abc
jœ œ œ œ œœ 2 ‰ jœ œ œ ‰ jœ œœ œ œ œ œ œœ 4
L1: a1=ab b1=ac c1=bc

& ‰ œ œ ‰ J œ œ œ ‰ J 4 ‰ œ 4 œ œ 4
3 3 3

4 j j j j j j 3
& 4 ‰ œ œœœ‰ œ œ œœ ‰ œ œœœœœœœ œœ ‰ œ œ œœœœœœ œœ œœœœ œœ‰ œ œœœ œ œœœœœ‰ œ œœœ 4
67 L2: a1+b1 a1+c1 b1+c1 c1+a1 c1'+a1

3 3 3 3 3 3 3 3

3 ‰ œj œ œ œ ‰ œj œ œ œ ‰ œj œ œ œ œ œ œ œ œ œ,œ œ œ œ œ œ ‰ œj œ œ œ 6 ‰ œj œ œ œ œ œ œ œ œ œœ œ œ œ œ œ
72 L3: (a1+b1)+(b1+c1)+(c1'+a1) (a1+c1)+(c1'+a1)

&4 4
3 3 3 3 3 3 3

j j j , j j
& ‰ œ œ œ œ ‰ œ œ œ œ œ œ œ œ œœ œ œœœ œ œ ‰ œ œ œ œ ‰ œ œ œ œ œ œ œ œ œœœ œœœ œ œ ‰ œ œ œ œ
77

3 3 3 3 3 3 3 3 3

Figure 10.3: Higher order permutations of a three-element group {a, b, c}: combinations by
two elements.

Copy issued to Seth Dempsey ([email protected]) 101

CHAPTER 10. GENERALIZATION OF VARIATION TECHNIQUES

temporarily return to a lower level, then pick a different combination and build
a second rhythmical development, e.g., L0 → L1 → L2 → L1 → L2 → L3 →
. . . → Ln . Use a walking path through the levels to create tension and relaxation,
or build up to a climax.
The Schillinger text stops at the combinations by three elements at level L2 , but since we
have a six-element set L1 : {a1 , ar1 , b1 , br1 , c1 , cr1 }, there might also be combinations by four or
more elements. Such an extended set of possibilites is discussed however in the next section,
where three elements are combined at L1 .

10.2.2 Combinations by three elements at the first higher order level

The three-element combination at the first higher order level is L1 : {abc}. This combination
has six permutations: {a1 , b1 , c1 , d1 , e1 , f1 } = {a + b + c, a + c + b, b + a + c, b + c + a, c + a +
b, c + b + a}. At the second higher order level L2 we may use combinations of two, three,
four, five and six elements from the first higher order level.

Level 2: Combination by two elements

The set of combinations of two elements at the second order level L2 contains 15 elements
a1 + b1 b1 + c1 c1 + d1 d1 + e1 e1 + f1
a1 + c1 b1 + d1 c1 + e1 d1 + f1
a1 + d1 b1 + e1 c1 + f1
a1 + e1 b1 + f1
a1 + f1
The first combination L2 : a1 + b1 = (a + b + c) + (a + c + b) is a rhythmic pattern with six
elements. All elements occur twice in the pattern; at this level there obviously do not exist
combinations with an incomplete set of source elements. Each combination consists of two
subgroups, implying two permutations. Therefore, the total number of cases is 15 × 2 = 30.

Level 2: Combination by three elements

The combination of three element from level L1 yields the following set at the next higher
level L2
a1 + b1 + c1 b1 + c1 + d1 c1 + d1 + e1 d1 + e1 + f1
a1 + b1 + d1 b1 + c1 + e1 c1 + d1 + f1
a1 + b1 + e1 b1 + c1 + f1 c1 + e1 + f1
a1 + b1 + f1 b1 + d1 + e1
a1 + c1 + d1 b1 + d1 + f1
a1 + c1 + e1 b1 + e1 + f1
a1 + c1 + f1
a1 + d1 + e1
a1 + d1 + f1
a1 + e1 + f1
The first combination L2 : a1 +b1 +c1 = (a+b+c)+(a+c+b)+(b+a+c) is a rhythmic pattern
with nine elements. The three subgroups in each combination yield six permutations. The
total number now is 20 × 6 = 120 cases.

102 Copy issued to Seth Dempsey ([email protected])

 10.2. HIGHER ORDER PERMUTATIONS OF A THREE-ELEMENT GROUP

Level 2: Combination by four elements

Using a combination of four elements at level L2 produces the set of combinations

a1 + b1 + c1 + d1 b1 + c1 + d1 + e1 c1 + d1 + e1 + f1
a1 + b1 + c1 + e1 b1 + c1 + d1 + f1
a1 + b1 + c1 + f1 b1 + c1 + e1 + f1
a1 + b1 + d1 + e1 b1 + d1 + e1 + f1
a1 + b1 + d1 + f1
a1 + b1 + e1 + f1
a1 + c1 + d1 + e1
a1 + c1 + d1 + f1
a1 + c1 + e1 + f1
a1 + d1 + e1 + f1

The first combination L2 : a1 + b1 + c1 + d1 = (a + b + c) + (a + c + b) + (b + a + c) + (b + c + a) is

a rhythmic pattern with 12 elements. The four subgroups in each combination yield 4! = 24
permutations. The total number therefore is 15 × 24 = 360 cases.

Level 2: Combination by five elements

With a five elements combined the full set at level L2 is

a1 + b1 + c1 + d1 + e1 b1 + c1 + d1 + e1 + f1
a1 + b1 + c1 + d1 + f1
a1 + b1 + c1 + e1 + f1
a1 + b1 + d1 + e1 + f1
a1 + c1 + d1 + e1 + f1

The first combination L2 : a1 + b1 + c1 + d1 + e1 = (a + b + c) + (a + c + b) + (b + a + c) +

(b + c + a) + (c + a + b) is a rhythmic pattern with 15 elements. Each combination has five
subgroups, i.e., 5! = 120 permutations. The total number of cases therefore is 6 × 120 = 720.

Level 2: Combination by six elements

Finally, we may combine all lower level elements into a single combination al level L2 :
a1 +b1 +c1 +d1 +e1 +f1 = (a+b+c)+(a+c+b)+(b+a+c)+(b+c+a)+(c+a+b)+(c+b+a),
i.e., a rhythmic pattern with 18 elements and 720 permutations.

Example 10.3
Higher order permutations of a three-element group: combinations by three
elements.
This example demonstrates the combinations by three elements, derived from
a three-element attack-duration group with resultant r = abc at level L0 . The
single staff notation with three pitches is shown in Fig. 10.4.

• The first case uses the three-element set L0 := {a, b, c} = {4t, 3t, t} at time
unit t = 18 , see m. 1. All six permutations of the first level L1 combination

Copy issued to Seth Dempsey ([email protected]) 103

CHAPTER 10. GENERALIZATION OF VARIATION TECHNIQUES

by three elements {a1 , b1 , c1 , d1 , e1 , f1h} =i {abc, acb, bac, bca, cab cba} are
4
shown in m. 2–7. With grouping at 4 some or these permutations yield
syncopated or afterbeat patterns. A combination by two elements at the
next higher level is L2 : a1 + c1 = (a + b + c) + (b + a + c), which has
two permutations. These are combined into a continuity at level L3 : rc =
(a1 + c1 ) + (c1 + c1 ) in m. 8–11.
At level L2 there are many possible combinations by three elements. The
three-element subset with c1 as the first element is L2 : {c1 + d1 + e1 , c1 +
d1 + f1 , c1 + e1 + f1 } and shown in m. 12–20. A continuity at the next higher
level L3 , based on the three-element combination L2 : b1 +c1 +d1 in the three
circular permutations, is L3 : (b1 + c1 + d1 ) + (c1 + d1 + b1 ) + (d1 + b1 + c1 ),
see m. 21–29.
• The second case is based on the trinomial L0 := {a, b, c} = {(2t+ 21 t+ 12 t), (r̄+
t̂), t} with two split-unit groups.3 The time unit t = 18 and the L0 resultant is
shown in Fig. 10.4 m. 30. The six three-element combinations at level h i L1 are
shown in m. 31–36. Like the first case, here the grouping at meter 68 , leads
to syncopated and afterbeat rhythmic patterns.
From the 24 possible combinations by four elements at level L2 the complete
subset with b1 = (a + c + b) as first element is {b1 + c1 + d1 + e1 , b1 +
c1 + d1 + f1 , b1 + c1 + e1 + f1 , b1 + d1 + e1 + f1 } and is shown in m. 37–
52. For the five circular permutations for the combination by five elements
L2 : a1 + b1 + c1 + e1 + f1 , see m. 53–77. Finally, the retrograde version of the
only possible combination by six elements L2 : f1 + e1 + d1 + c1 + b1 + a1 is
shown in m. 78–83.

In summary, combining two- and three-element attack-duration groups with notes, rests,
accented notes and split-unit groups opens up a huge potential for creating rhythmic conti-
nuities rc , when higher order permutations are used. These have the self-scaling property,
i.e., the original attack-duration group will return at the higher level in varied context.4
This means homogeneity and variation in one approach, while preventing repetition and
monotony.

Application Tip:
This technique will be most useful in film, video and game music, when an at-
mosphere, a mood must be underscored. Especially the latter application do-
main, writing background music for a specific level in a game requires looking
for means to combine consistency with variation. Short-time repetitions must
be prevented, because these become an obvious distraction from the immersion
feeling and will bore the player to digital death. Well, here the higher order per-
mutations of a basic rhythm may prove to be a useful technique in the composer’s
toolbox.

3
The label in the score does not indicate the splitting of the 8th note into two 16th notes.
4
A nice analogy is the composition of the two helical strands in the DNA molecule. The genetic information
is determined by the sequences of the four elementary nucleotides labeled G,A,T and C.

104 Copy issued to Seth Dempsey ([email protected])

 10.2. HIGHER ORDER PERMUTATIONS OF A THREE-ELEMENT GROUP

& 44 ˙ œ . œJ ˙ œ . œJ ˙ œ œ. œ . œj œ . œJ œ . œJ ˙ œ œ . œj œ . œ œ. ˙
1 L0: 4t+3t+t L1: abc acb bac bca cab cba

J J J

œ . œJ œ . œj œ . œJ œ . œj œ . œJ ˙ œ . œJ œ . œj œ . œJ œ . œJ ˙ œ œ . œj œ .
8 L3: a1c1+c1a1 L2: c1d1e1

&˙ J

j œ œ. œ ˙ œ œ. ˙ j œ œ œ. j œ œ. ˙
& œ. œ. œ œ.
15 c1d1f1 c1e1f1
œ œ. J J J œ œ. J J J

œ œ. œ . œj œ . œJ œ . œJ ˙ œ . œj œ . œJ œ . œJ ˙ œ œ. œ . œJ ˙
21 L3: b1c1d1+c1d1b1+d1b1c1
&˙ J ˙ J

> > acb > bac >

œ œ. œ . œj œ . œ 68 œ œ œ ‰ œ œ œ œ œ ‰ œ œ œ œ œ œ ‰ œ ‰ œ œ œ œ œ œ ‰ œ œ œ œ œ
28 (2t+t)+(r+t')+t L1: abc bca

&˙ J J >

> > > >

& œJ œ œ œ ‰ œ œ ‰ œ œ œ œ œ œ œ œ ‰ œ ‰ œ œ œ œ œ œ ‰ œ œ œ œ œ œJ œ œ œ ‰ œ œ œ œ œ ‰ œ
35 cab cba L2: b1c1d1e1 b1c1d1f1

> > >

œ >œ œ > > >

œ œ œ œ œ œ œ ‰ œ ‰ œ œ œ œ œ œ œJ œ œ œ ‰ œ œ ‰ œ œ œ œ
42 b1c1e1f1

& ‰ œ œ œ œœ ‰ œ œ œœ ‰ œ
> > >

œ > >œ œ œ œ œ > œ > > >

‰ œ ‰ œ œ œœ œ œœ‰ œ œ œ œœœ ‰ œ ‰ œ œ œ œœœ
49 b1d1e1f1 a1b1c1e1f1

& œ œœ ‰ œ ‰ œ œ œœ J
>

œ
56
œ > œ ‰ >œ œ œ œ œ œ œ ‰ >œ œ œ œ œ œ ‰ >œ ‰ œ œ œ œ œ œ œ œ œ œ ‰ œ
f1a1b1c1e1

&J œ œ œ ‰ œ ‰ œ œ œ œ J
> > >

> > >

& œJ œ œ œ ‰ œ œ ‰ œ œ œ œ œ œ œ ‰ œ œ œ œ œ œ ‰ œ ‰ œ œ œ œ œ œ ‰ œ œ œ œ œ œ œJ œ œ œ ‰ œ
63 e1f1a1b1c1 c1e1f1a1b1

> > > >

œ > >œ œ > b1c1e1f1a1 > >

œ œ œ œ œ ‰ œ ‰ œ œ œ œ œ œ œJ œ œ œ ‰ œ œ ‰ œ œ œ œ
70
& ‰ œ œ œ œ œ œ œ ‰ œ œ œ œ ‰
> >

> œ f1c1e1b1d1a1
> > > >
& œ œ œ œ œ ‰ œ œ œ œ ‰ œ œ œ œ œ œ œJ œ œ œ ‰ œ œ œ œ œ ‰ œ ‰ œ œ œ œ œ œ œ œ ‰ œ œ
77
‰
> >

Figure 10.4: Higher order permutations of a three-element group {a, b, c}: combinations by
three elements.

Copy issued to Seth Dempsey ([email protected]) 105

CHAPTER 10. GENERALIZATION OF VARIATION TECHNIQUES

106 Copy issued to Seth Dempsey ([email protected])

Chapter 11

Composition of homogeneous
rhythmic continuity

In this chapter Schillinger returns to the combination of simultaneity and continuity, that
was introduced in Section 9.1.2. In Chapter 9 that concept was applied at the lowest level
attack-duration group, by considering the set of possible permutations, and applying it to
different musical attributes (notes, rests, accents, etc).
In Chapter 10 the starting point was a small attack-duration group with two or three el-
ements. A longer continuity rc was created from this core cell by considering higher order
permutations, i.e., at higher aggregation levels. In this chapter the process follows the re-
verse path: the starting point is a longer series of attack-durations and we will apply splitting
in order to achieve smaller units. These smaller groups then serve as the basis for a growth
process, both in the time domain (continuity) and in a parallel distribution over multiple
parts in a score (simultaneity).
Splitting as another mechanism for generating growth in simultaneity and continuity
may be done:
• through the simplest divisor. This yields multi-measure units and the minumum quan-
tity of material to be evolved.
• at the measure level. This is an intermediate case.
• at the individual attack level. This leads to the maximum quantity of source material
for evolving.
The splitting process alternatives are shown in diagram in Fig 11.1. These three approaches
are covered in more detail in the next sections.

11.1 Splitting through the simplest divisor

Remember from Chapter 2, 4 and 6 that interference of two or three generators and the
process of fractioning all lead to symmetrical patterns in the sense that the time-reversed
(retrograde) rhythmical pattern is equal to the original series.
This means that for an even number of attacks we can split the pattern at the centre, and
obtain two smaller attack-duration units that can be combined in simultaneity (i.e., parallel
distribution over two staves) and continuity (i.e., juxtaposed permutations).

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 11. COMPOSITION OF HOMOGENEOUS RHYTHMIC CONTINUITY

a bcd
b bbb r3÷2 = 2t + t + t + 2t
-
0 5 t
Splitting in half (division by two, 3t units)
- rc : a1 b1 b1 a1
P1 : b b b b b b b b a1
P2 : b b b b b1
Splitting by measures (unit length 2t)
- rc : a1 b1 c1 b1 c1 a1 c1 a1 b1
P1 : b b b b b b b b b b b b b1
P2 : b b b b a1
P3 : b b b b c1
Splitting by attacks
- rc : a bcd bcd a cda b d a bc
P1 : b b b b bbb b bb b b b b bb c
P2 : b b b b bb b b b b bb b bbb b
P3 : b b b b b b bb b bbb bbb b a
P4 : b b b b b bbb bbb b bb b b d
Vertical: simultaneity (parts P ). Horizontal: continuity rc

Figure 11.1: Composition of a homogeneous rhythmic continuity by splitting an attack-

duration group. The non-uniform binary synchronization pattern with interference attack-
duration group r3÷2 = 2t + t + t + 2t (4 attack elements {a.b.c.d}), is split into two halves,
per measure and by individual attacks. This creates different continuities and multiple staff
simultaneities.

We will illustrate the three splitting approaches graphically in Fig. 11.1 with the simples
case possible. It is the non-uniform binary synchronization pattern r3÷2 = 2t + 1t + 1t + 2t,
presented in Section 2.1.2. It has four attacks and the total duration is Tt = 6t; therefore it is
an unrealistic case. The division into two halves creates two units {a1 , b1 } = {(2t+t), (t+2t)};
these can be permutated to create a continuity rc = (a1 +b1 )+(b1 +a1 ) with total duration Tt =
12t, and combined in a two-part (2P ) simultaneity. The splitting by measure leads to three
units {a1 , b1 , c1 } = {2t, (t + t), 2t} with a length each of 2t. Now the circular permutation
process creates a continuity rc = (a1 + b1 + c1 ) + (b1 + c1 + a1 ) + (c1 + a1 + b1 ) with total
duration Tt = 18t, and combined in a three-part (3P ) simultaneity. Finally, splitting by
individual attacks using the four original units {a, b, c, d} = {2t, t, t, 2t} creates a continuity
rc = (a + b + c + d) + . . . + (d + a + b + c) with total duration Tt = 24t, and a four-layer (4P )
simultaneity.
A more realistic case is shown in the next example.

Example 11.1
Composition of homogeneous rhythmic continuity: splitting through the sim-
plest divisor.
The approach of splitting the original attack-duration group through the simplest

108 Copy issued to Seth Dempsey ([email protected])

 11.2. SPLITTING THROUGH INDIVIDUAL MEASURES

divisor is shown in Fig. 11.2 in rhythmic notation. Three cases will be discussed;
the first and last are splitting the rhythmic pattern in half, the other divides it into
three equally long parts.
• The source attack-duration group is the fractioning pattern rc = 6 ÷ 5 =
5t + t + 4t + r̄ + t + 3t + t + 2t + 2t + t + 3t + t + t̂ + 4t + t + 5t. This pattern
has 16 attacks and a total duration of Tt = 36t at time unit t = 81 . The
pattern has been slightly modified by introducing a rest and an accented
note. The simplest divisor is 2, splitting the group into two equal parts, the
units a1 = (5t+t+4t+ r̄+t+3t+t+2t) and b1 = (2t+t+3t+t+ t̂+4t+t+5t).
These may be combined into a continuity rc = (a1 + b1 ) + (b1 + a1 ), using h i
the two permutations. This is shown in m. 1–12 at grouping by meter 68 .
Combination of both units as a simultaneity creates a score with two staves.
• The second case is based on the non-uniform binary synchronization pattern
r9÷4 = 4t + 4r̄ + t + 3t + 4t + 2t + 2t̂ + 4t + 3t + t + 4t + 4t. This group
has 12 attacks and a total duration of Tt = 36t; one attack has been replaced
with a rest and there is an accented
h i note. The example in m. 13–39 uses time
1 4
unit t = 4 and grouping at 4 . Here we may divide into either two or three
equal parts; the example demonstrates the latter splitting option. The result
consists of three smaller units {a1 , b1 , c1 } = {(4t + 4r̄ + t + 3t), (4t + 2t + 2t̂ +
4t), (3t + t + 4t + 4t)}, each with duration 12t, that can be used in continuity
and simultaneity. The three-part score shows the vertical distribution of the
continuity rc = (a1 + b1 + c1 ) + (b1 + c1 + a1 ) + (c1 + a1 + b1 ), a 3 × 9 = 27
measures long pattern achieved through circular permutation.
• In Fig. 11.3, m. 1–4 the source pattern is the three-generator interference
0
pattern r5÷3÷2 = 6t + 4t + 2t + 3t + 3t + 2t + 4t + 6t̂, consisting of 8 attack handi
a total duration of Tt = 30t. Time unit is t = 18 , the grouping is at meter 68 .
Division by two leads yields to smaller units that are no longer an integer
number of full measures: both a1 = 6t + 4t + 2t + 3t and b1 = 3t + 2t + 4t + 6t̂
have a duration of T = 15t, corresponding to 2 12 measures. The example
shows the simultaneity in two parts
!
a1 + b1 = (6t + 4t + 2t + 3t) + (3t + 2t + 4t + 6t̂)
rs = .
b1 + a1 = (3t + 2t + 4t + 6t̂) + (6t + 4t + 2t + 3t)

When both an odd number of attacks and the total duration of the attack-duration group
are an integer multiple of three, than the simplest divisor obviously is three. Looking at
the examples in the previous chapters, there are not many combinations of generators that
will satisfy this condition. One such case was illustrated in the example above. All cases
consisted of smaller scale units that were longer than one measure.

11.2 Splitting through individual measures

Splitting may be applied at a smaller timescale by considering individual measures. Here
the grouping of the time units into measures becomes relevant as it may lead to different

Copy issued to Seth Dempsey ([email protected]) 109

CHAPTER 11. COMPOSITION OF HOMOGENEOUS RHYTHMIC CONTINUITY

& 68 œ . ‰ œ œ. œ œ. œ >œ ˙ œ.
1 frac(6,5), a1 b1
œ œ ˙ œ œ œ œ œ
J J J J J

& 68 œ œ œ. œ >œ ˙ œ. œ. ‰ œ œ.
b1 a1
œ œ œ œ ˙ œ œ
J J J J J

œ œ. œ >œ ˙ œ. œ. ‰ œ œ. 44
7 b1 a1

&œ J
œ œ
J
œ œ ˙
J J
œ œ
J

& œ. ‰ œ œ. œ œ. œ >œ ˙ œ. 44
a1 b1
œ œ ˙ œ œ œ œ œ
J J J J J

4 œ ˙. >˙ ˙.
13 r(9:4), a1 b1 c1

&4 w ∑ w ˙ w œ w w

& 44 w >˙ ˙. œ ˙.
b1 c1 a1
˙ w œ w w w ∑

4 >˙
& 4 ˙. œ ˙.
c1 a1 b1
œ w w w ∑ w ˙ w

>˙ ˙. œ ˙.
22 b1 c1 a1

&w ˙ w œ w w w ∑

& ˙. œ ˙. >˙
c1 a1 b1
œ w w w ∑ w ˙ w

œ ˙. >˙ ˙.
a1 b1 c1

&w ∑ w ˙ w œ w w

& ˙. œ ˙. >˙
31 c1 a1 b1
œ w w w ∑ w ˙ w

œ ˙. >˙ ˙.
a1 b1 c1

&w ∑ w ˙ w œ w w

>˙ ˙. œ ˙.
b1 c1 a1

&w ˙ w œ w w w ∑

Figure 11.2: Homogeneous rhythmic continuity by splitting. Example 1, splitting through

the simplest divisor. Division by two is applied to the fractioning pattern r6÷5 , division by
three to the non-uniform binary synchronization group r9÷4 .

110 Copy issued to Seth Dempsey ([email protected])

 11.2. SPLITTING THROUGH INDIVIDUAL MEASURES

& 68 ˙ . œ . œ œ œ ." œ . œ œ œ . >˙ . ˙. œ. œ œ œ . >œ . œ œ œ. ˙.

1
r'(5:3:2), a1 b1 a1 b1 c1 d1 e1

J J J J
& 68 œ . œ œJ œ . >œ . œ . "œ . œ . œ . œJ œ œ . œ . œJ œ œ . >œ . d1
œ œ œ. ˙. ˙.
b1 a1 b1 c1 e1 a1

J
6 œ . >œ . œ œ œ. ˙. ˙. œ. œ œ
& 8 ∑ ∑ ∑ ∑ ∑
c1 d1 e1 a1 b1

J J
& 68 ∑ ∑ ∑ ∑ ∑ œ œ œ. ˙. ˙. œ . œ œ œ . >œ .
d1 e1 a1 b1 c1

J J
6
& 8 ∑ ∑ ∑ ∑ ∑ ˙.
e1
˙.
a1 b1 c1
>
œ. œ œ œ. œ. œ œ œ.
d1

J J

& œ . œJ œ œ . >œ . d1
œ œ œ. ˙. ˙. ’ ’ ˙. ˙. œ. œ œ œ. >œ . d1
œ œ œ.
11
b1 c1 e1 a1 e1 a1 b1c1

J J J
& œ . >œ . œ œJ œ. ˙. ˙. œ. œ œ ’ ’ ˙. œ . œ œ œ . >œ . œ œ œ. ˙.
c1 d1 e1 a1 b1 a1 b1 c1 d1 e1

J J J
& œ œJ œ . ˙ . ˙. œ . œ œ œ . >œ . ’ ’ œ . œ œ œ . >œ . œ œ œ . ˙ . ˙.
d1 e1 a1 b1 c1 b1 c1 d1 e1 a1

J J J
& ˙. ˙. œ . œ œ œ . >œ . œ œ œ . ’ ’ œ . >œ . œ œ œ . ˙ . ˙. œ. œ œ
e1 a1 b1 c1 d1 c1 d1 e1 a1 b1

J J J J
& ˙. œ . œ œ œ . >œ . œ œ œ . ˙ . ’ ’ œ œ œ. ˙. ˙. œ. œ œ œ . >œ .
a1 b1 c1 d1 e1 d1 e1 a1 b1 c1

J J J J

& ˙. œ . œŒ œ . >œ . œ œ œ . >˙ . ’ ’ >˙ . ˙. œ . œŒ œ . >œ . œ œ œ .

22

J J J J
& œ. >
œ Œ œ. œ. œœ >
œ. ˙. ˙. ’ ’ ˙. œ. >
œŒ œ . œ . œ œ >
œ. ˙.
J J J J
& Œ œJ œ >œ œ œ œœ œ >œ
œ. œœ œ. œ œ œ. ’ ’ œ . œJ Œ œ . >œ . œ œ œ . >˙ . ˙.
J J J J J J J
& œ. >œ . œ œ œ . >˙ .
˙. œ. œ Œ ’ ’ Œ œJ œ >œJ œ œJ œœ œ >œ œ . œ œ œ. œ œ œ.
J J J J J J
& >œ . œ œ œ . >œ . œ .
J
œ. œ. œ. œ Œ œ.
J ’ ’ œ . >œ . œ œJ œ . >˙ . ˙. œ. œ Œ
J
& œ œJ œ . >˙ . ˙.œ . œŒ œ . >œ . ’ ’ >œ . œ œJ œ . >œ . œ . œ . œ . œ. œ Œ œ.
J J
>
& œ . œJ œ œ . œœ œ . œœ œ‰ ‰ œ œ >œ œ œ ’ ’ œ œJ œ . >˙ . ˙. œ. œŒ œ . >œ .
J J J J J
& >˙ . ˙. œ . œŒ œ . >œ . œ œ œ .
J J ’ ’ œ . œJ >œ œ . œœ œ . œœ œ‰
J J
‰ œ œ >œ œ œ
J J

Figure 11.3: Homogeneous rhythmic continuity by splitting. Example 2, three-generator

0
group r5÷3÷2 simplest divisor and splitting by measures or attacks. Not all elements in the
continuity are shown, as indicated by the slashes.

Copy issued to Seth Dempsey ([email protected]) 111

CHAPTER 11. COMPOSITION OF HOMOGENEOUS RHYTHMIC CONTINUITY

smaller rhythmical units that can be combined.

Example 11.2
Composition of homogeneous rhythmic continuity: splitting by measures.
The splitting by measures technique is shown in Fig. 11.3.
• We return to the three-generator interference pattern r5÷3÷20 = 6t + 4t + 2t +
3t + 3t̂ + 2t + 4t + 6t in
h iFig. 11.3, m. 6–10, now considering each split measure
6
at grouping meter 8 as a separate unit. So we have {a1 , b1 , c1 , d1 , e1 } =
{6t, 4t + 2t, 3t + 3t̂, 2t + 4t, 6t}. These may combined into a continuity
using circular permutations rc = (a1 + b1 + c1 + d1 + e1 ) + (b1 + c1 + d1 +
e1 + a1 ) + . . . + (e1 + a1 + b1 + c1 + d1 ) (not all permutations are shown in
m. 5–21, see the slashes). The simultaneity is a five-part score
 
a1 + b1 + c1 + d1 + e1

 b1 + c1 + d1 + e1 + a1 

rs =  c1 + d1 + e1 + a1 + b1 .
 
 
 d1 + e1 + a1 + b1 + c1 
e1 + a1 + b1 + c1 + d1
Note the synchronized attack at the downbeat of each measure; the rhyth-
mic patterns within the measure are different for each part.

11.3 Splitting through the individual attacks

At the smallest scale each attack may be considered a unit; we are splitting the original
rhythmic group by individual attacks. This provides the maximum amount of material for
the growth process of creating a homogeneous continuity (through permutation) and simul-
taneity (by distribution over multiple parts). Let’s return to our previous example and see
what the result will be when we split by attacks.

Example 11.3
Composition of homogeneous rhythmic continuity: splitting by attacks.
0
The result of splitting the three-generator interference pattern r5÷3÷2 by individ-
ual attacks is shown in Fig. 11.3, m. 22–32.
• We have modified the number of accented notes and introduced a rest:
0
r5÷3÷2 = 6t + 4t + 2r̄ + 3t + 3t̂ + 2t + 4t + 6t̂.
• Splitting by individual attacks yields 8 units, that may be combined into a
continuity rc or an 8-part simultaneity rs . Only the first and last permutation
are shown in rhythmical notation.
• Each part is playing a different rhythm; the combination in the score con-
tains a number of parts with many syncopated notes.

112 Copy issued to Seth Dempsey ([email protected])

Chapter 12

Distributive powers

This chapter covers the application of distributive powers in rhythm. It is divided into two
sections: the first has an analytical character. It discusses the possible power series that
may be used either for the subdivision of rhythmical units within the measure or for the
length of musical phrases. The second section is a technique for writing continuities and
counterthemes applying distributed powers to a given attack-duration group.

12.1 Continuity of harmonic contrasts

The first application of power series in rhythmical continuity is to the grouping meter nn ,
 

where n is called the rhythm determinant. An overview of such series is shown in Fig. 12.1
for determinants n = 2, . . . , 9. Taking powers of the denominator value 1/n determines the
subdivision into shorter measure segments of the basic time unit n = t; Schillinger calls
this fractional continuity. In the figure this subdivision is shown to the left of the centre as
1/nk , k = 1, 2, 3, . . . (k is also an integer number). So, for example, taking the most familiar
n h2i
grouping n = 2 , the subdivision of the measure is into two smaller units 1/n1 = 21 , the
next smaller is 1/n2 = 41 , and so forth until the 4th power 1/n4 = 16 1
.
k
Applying the power series n to the numerator term yields the number of measures used
in a pattern. Schillinger labels this as factorial continuity. These are the values shown in the
n h2i
figure to the right of the centre. For the meter n = 2 the result is a set of patterns with
length n1 = 2, n2 = 4, n3 = 8 and n4 = 16 measures. The power series stop at both ends
when the resulting values are either too large or small for application in music.
Schillinger then discusses which power series have been used in music. The most fre-
quently used power series is the binary system: multiplication and division by 2; doubling
the number of measures in a musical phrase and halving the note duration. The former leads
to regular phrases of either 2, 4, 8 or 16 measures in classical music. This power series also
covers the n5 = 32 measure chorus in popular songs. Division by powers of two yields the
familiar whole note – half note – quarter note – 18 note – 32 1
note durations.
h i
Already the pow-
h i
3 9
ers of three are less frequently used: there are the triple division 4 and 8 meter and the
triplet subdivision ( 13 )1 .
A rhythmical phrase with a length of three or nine measures is used
most infrequently. Division by and multiplication by 4 are similar to the 22 power term. In
some traditional folk music there is the occasional division by 5 and 7. However asymmetric

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 12. DISTRIBUTIVE POWERS

1 1 1 1 n
... n4 n3 n2 n n n n2 n3 n4 ...
1 1 1 1 2
... 16 8 4 2 2 2 4 8 16 . . .
1 1 1 3
... 27 9 3 3 3 9 27 . . .
1 1 1 4
... 64 16 4 4 4 16 64 . . .
1 1 5
... 25 5 5 5 25 . . .
1 1 6
... 36 6 6 6 36 . . .
1 1 7
... 49 7 7 7 49 . . .
1 1 8
... 64 8 8 8 64 . . .
1 1 9
... 81 9 9 9 81 . . .

← Fractional continuity (within the measure) | Factorial continuity (groups of measures) →

Figure 12.1: Overview of rhythmical power series for creating fractional (division by powers
of the determinant 1/nk ) and factorial continuity (multiplication by powers nk ). Each row
represents a power series family for a specific determinant n.

phrases are still fairly rare in music.1

h i
6
There are also some hybrid power series. For example in 8 there is first division by 2
( 68
= + 3
8
3
8 ),
then by three = + + ( 38 1
8
1
8
1
8 ),
then by two = + 16 1
( 81 1
). The 12-measure Blues
16
chorus, with its three subphrases of four measures each, is a multiplication of first by 4, then
by 3 (12 = 3 × 4). Schillinger blames the limited use of the power series in fractional and
factorial continuitues on music notation schemes
h i and indicates some potential for the future.
9
In the text he claims that the determinant 9 is in the making. He points to swing and shuffle
h i h i
8 9
music as the hybrid of 8 and 9 , a binary division followed by a triple division.2

12.2 Composition of rhythmic counterthemes by means of distri-

butive powers
In this section there is a technique of applying the distributive power approach to an attack-
duration group. Although the word counterthemes in the section title suggests that the ap-
proach is meant for creating a simultaneity, it may equally well be used for creating a rhyth-
mic continuity.
Distributive powers will be used first for squaring an attack-duration group. Then the
third power, the cube is explained. Finally, this approach is generalized to higher order
powers.
1
In classical music there are examples of phrase lengths of odd numbers, such as three, five or seven measures.
Look for irregular phrases in the compositions by Johannes Brahms. In popular music there is the seven-measure
A-phrase in Yesterday, the Paul McCartney song.
2
Swing music was a current popular style at the writing of the Schillinger volumes. Now, almost a century
later use your own analytical skills to determine which power series are dominating the popular music domain.

114 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

12.2.1 The square of an attack-duration group

An attack-duration group with resultant r, a binomial, trinomial or multinomial, with total
duration T = nt (n is the determinant) may be squared, written as r2 . The multinomial
rhythmical pattern has a periodicity determined by multiple coefficients, typically more than
three.3 The length of the square r2 is T 2 = (nt)2 time units. However, squaring in the
distributive power approach requires that we keep all the separate multiplication terms. This
may seem a little confusing, but will become clear when we present the numerical examples.

The square of a binomial

The source is the binomial attack-duration group r = a + b. Equivalently this is written as

r = at + bt with the sum a + b = n the determinant of the group. Squaring this group in the
distributive power approach means

r2 = (a + b)2 = (a + b)(a + b)
= a(a + b) + b(a + b) = (a2 + ab) + (ba + b2 ) = a2 + ab + ba + b2 . (12.1)

It is essential to keep the separate terms; the square of the binomial leads to a group with
four attacks.4 Musically, it means that we have obtained two timescaled copies of the original
pattern. Let’s look at the simplest possible binomial example.

Example 12.1
The square of a binomial: simplest case
The binomial attack-duration group is r = 2t + t with length T = 3t. We
write the distributive square using Eq. 12.1: r2 = (2t + t)2 = (2 + 1)(2t + t) =
2(2t + t) + 1(2t + t) = (4t + 2t) + (2t + t) = 4t + 2t + 2t + t with four attacks.
Note that we have used the distributive coefficients {2, 1} for the multiplication
without the t time unit. The total length is now T 2 = 9t = 6t + 3t, two sub-
groups where the first is twice the length of the second group. This scaling effect
is the consequence of the ratio (a/b) = (2/1) = 2 of the elements in the binomial.

This effect of the distributive power approach s another self-scaling property. In Chapter 9
we discovered this property when using the groups in general approach in Section 9.2. There
it had to do with the return of smaller scale rhythmic cells at a higher aggregation level.
Here it is the duration pattern, that returns on a different timescale, depending on the power
series.
The binomial has two permutations; therefore the square in our example may be either
r2 = (2t + t)2 = (4t + 2t) + (2t + t) or r2 = (t + 2t)2 = (t + 2t) + (2t + 4t). Such a continuity
is great for creating either a rhythmical acceleration or for the reverse, going from a busy
rhythm to longer durations and calming down, e.g., closing with a fermata (see Chapter 14
for variable velocity patterns and acceleration series).
An overview of practical binomials for squaring is shown in the second column in Ta-
ble 12.1. The first column shows the determinant n, the sum of the two elements in the
binomial attack-duration group.

Copy issued to Seth Dempsey ([email protected]) 115

CHAPTER 12. DISTRIBUTIVE POWERS

Table 12.1: Practical distributive powers, i.e., square and cube, applied to binomials a + b
and trinomials a + b + c with determinant n. Each binomial pair {a, b} has two permutations.
Trinomials have either three or six permutations. Combinations in brackets, such as (3 + 2 +
1), are not mentioned in Schillinger’s book.

Determinant Binomial Trinomial

(Sum) n Square Cube Square Cube
n
(a + b)2 (a + b)3 (a + b + c)2 (a + b + c)3
 
n

3 2+1 2+1
4 3+1 3+1 2+1+1 2+1+1
5 3+2 3+2 2+2+1 2+2+1
4+1 3+1+1 3+1+1
6 5+1 (3+2+1) 3+2+1
4+1+1
7 4+3 3+2+2 3+2+2
5+2 3+3+1
6+1 (4+2+1)
5+1+1
8 5+3 3+3+2 3+3+2
7+1 (4+3+1)
(5+2+1)
6+1+1
9 5+4 (4+3+2)
7+2 4+4+1
8+1 5+2+2
(5+3+1)
(6+2+1)
7+1+1

116 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

b bbb
(2 + 1)2
b b b b
(3 + 1)2
-
0 10 20 30 40 50 60 70 80 t

b b b b
(4 + 1)2
b b b b
(3 + 2)2
b b b b
(5 + 1)2
-
0 10 20 30 40 50 60 70 80 t

b b b b
(6 + 1)2
b b b b
(5 + 2)2
b b b b
(4 + 3)2
-
0 10 20 30 40 50 60 70 80 t

b b b b
(7 + 1)2
b b b b
(5 + 3)2
-
0 10 20 30 40 50 60 70 80 t

b b b b
(8 + 1)2
b b b b
(7 + 2)2
b b b b
(5 + 4)2
-
0 10 20 30 40 50 60 70 80 t

Figure 12.2: Distributive powers. The square of a binomial r2 = (at + bt)2 with determinant
n = a + b; each pair has two permutations. Note the scaling effect, casued by the element
value ratio a/b.

Copy issued to Seth Dempsey ([email protected]) 117

CHAPTER 12. DISTRIBUTIVE POWERS

All binomial squares are shown in diagram in Fig. 12.2. The figure shows one permuta-
tion only as a continuity, with the greatest number as the a-element, i.e., a > b. Blue lines in
the diagrams indicate the division of the four attacks into two subgroups. The scaling effect
is obvious; compare the r2 = (5 + 3)2 with r2 = (7 + 1)2 , both combinations with determi-
nant n = 8. The latter suggests an enormous rhythmical acceleration, whereas the former is
a smoother, more balanced change in durations.

Example 12.2
The square of a binomial: various cases.
We will perform the calculations for a number of two-element attack-duration
groups with resultant r = (at + bt). All cases are shown in diagram in Fig. 12.2,
the musical notation is presented in Fig 12.3:

• Case 1, determinant n = 4: r = 3t + t (a = 3, b = 1, T = 4t). The distributive

second power is calculated with Eq. 12.1 and yields the square

r2 = (3 + 1)(3t + t) = 3(3t + t) + 1(3t + t) = (9t + 3t) + (3t + t).

This four-attack pattern with total duration Tt = T 2 = 16t consists of

two
h i subgroups with durations 12t, 4t, respectively. This is suitable for a
4
4 grouping. The other permutation is r = t + 3t, which has the square
r2 = (t + 3t) + (3t + 9t). Both permutations are shown in the upper staves
in m. 1–4 of Fig. 12.3.
• Case 2, determinant n = 5: r = 4t + t (a = 4, b = 1, T = 5t). The square of r
is

r2 = (4 + 1)(4t + t) = 4(4t + t) + 1(4t + t) = (16t + 4t) + (4t + t).

Compared to the first case we note the increased difference in length be-
tween the two subgroups; the ratio of longest to shortest attack
h i now is
16t/1t = 16. The total duration Tt = 25t, i.e., 5 measures at 54 time sig-
h i h i
3 4
nature, as shown in m. 10–14. Recurrence at meter 4 or 4 requires three
or four statements of the pattern, respectively.
• Case 3, determinant n = 7: r = 4t + 3t (a = 4, b = 3, T = 7t). The square of
this binomial is

r2 = (4 + 3)(4t + 3t) = 4(4t + 3t) + 3(4t + 3t) = (16t + 12t) + (12t + 9t).

The two subgroups are now more in balance 28t + 21t, Tt = 72 t = 49t and
the longest to shortest ratio is h16t/9t
i = 1.77. The determinant n = 7 re-
7
quires irregular meter such as 8 in folk music. This is illustrated in two
permutations in Fig. 12.3 m, 15–21.
3
Until now, we have almost exclusively encountered binomials and trinomials.
4
Do not add the ab + ba = 2ab terms, as is taught in calculus courses, because then we lose one attack.

118 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

• Case 4, determinant n = 9: r = 7t + 2t (a = 7, b = 2, T = 9t). The square of

this binomial is

r2 = (7 + 2)(7t + 2t) = 7(7t + 2t) + 2(7t + 2t) = (49t + 14t) + (14t + 4t).
h i
3
The total duration is Tt = 81t, a 27-measure pattern in 4 , as shown in
m. 22–48.

The original group may also be combined with the square in a simultaneity. In order to
achieve synchronization the original group durations must be multiplied with the determi-
nant n. This leads to a timescaled, augmented version of the original pattern nr in the top
part and the square r2 in the lower part
!
P1 : nr = n(a + b) = na + nb
rs = . (12.2)
P2 : r2 = (a2 + ab) + (ba + b2 )

Both now have the duration T = n2 t time units. The combination in simultaneity leads to
the style known as isorhythm, used in the Ars Nova period in Medieval music. A rhythmic
pattern sounds in parallel with its augmented version ra (in the augmentation the durations
are multiplied) or with the diminution (durations halved or in some other fraction). As a
combination various permutations may be used in each part; for the binomial there are four
possibilities; this potential was discussed in Section 9.2.
However, there is another candidate for synchronization with the square. Remember the
fractioning pattern from Chapter 4. The length of the fractioning pattern ra÷b is T = a2 ,
the square of the major generator value. This enables us to the combination of the square
r1 = (a + b)2 and determinant n = a + b with the fractioning pattern r2 = rn÷m .
Combining the three approaches in simultaneity leads to a three-part score
 
P1 : nr = n(a + b) = na + nb
2 = (a2 + ab) + (ba + b2 )  .
rs =  P2 : r (12.3)
 
P3 : rn÷m

We may use any subset from this set in any vertical order in the score; there are three subsets
of two parts {(P1 , P2 ), (P1 , P3 ), (P2 , P3 )} in two possible vertical orderings, or six vertical
distributions of the three parts.

Example 12.3
Combination of the square of a binomial with the augmented original and the
fractioning pattern.
The starting point is the binomial, the two-element attack-duration group with
r = (at + bt), determinant n = a + b. In this example the square r2 will be
combined with the augmented, timescaled version of the original r and the frac-
tioning pattern rn÷m . The result of the synchronization process will be shown for
a number of cases:

Copy issued to Seth Dempsey ([email protected]) 119

CHAPTER 12. DISTRIBUTIVE POWERS

& 44 w œ ˙. ˙ . œ 45 w œ w 85
1
w œ w œ w œ œw
(3t+t)^2 (3t+2t)^2

& 44 œ ˙ . ˙ . œ w w 45 w œ w œ w œ œw w œ 85
(t+3t)^2 (2t+3t)^2

4 12t+4t 5 15t+10t œ 85
&4 w w w w 4 w œ w œ w œ w œ w

& 44 w w w w 45 w œ w œ w œ w œ w œ 85
4t+12t 10t+15t

& 44 ˙ . œ ˙ œ œ œ œ ˙ œ ˙. 45 ˙ ˙ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ ˙ ˙ 85
frac(4,3) frac(5,2)

& 44 ∑ ∑ ∑ ∑ 45 ˙ . ˙ œ ˙ œ œ œ œ œ œ œ œ œ ˙ œ ˙ ˙. 85
frac(5,3)

& 44 ∑ ∑ ∑ ∑ 45 w œ ˙ . œœ ˙ œ˙ œ œ ˙. œw 85
frac(5,4)

& 85 œ . œ. œ œ. œ œ œ œ 78 ˙ œ . œ. œ œ œ. œ.
10
œ œœ œ ˙ ˙
(4t+t)^2 (4t+3t)^2

J J
& 85 œJ œ œ œ. œ œ œ. œ œ. œ œ . œ 78 ˙ œ . œ œ. œ ˙ œ. ˙ œ.
(t+4t)^2 (3t+4t)^2

œ. œ. œ œ œ. 43 ˙ . ˙. ˙. ˙. ˙. ˙. ˙.
19
&˙ ˙
(7t+2t)^2

& œ. œ œ ˙ œ. ˙ œ. 43 ˙ . œ ˙ ˙. ˙. ˙. ˙. ˙.
(2t+7t)^2

& ˙. ˙. ˙. ˙. ˙. ˙. ˙. ˙. ˙.
29
œ ˙

& ˙. ˙. ˙. ˙ œ ˙. ˙. ˙. ˙. ˙. ˙.

& ˙. ˙. ˙. ˙. ˙. ˙. ˙. ˙. ˙.
39
˙ œ

& ˙. ˙. ˙. ˙. ˙. ˙. ˙. ˙. ˙. ˙.

Figure 12.3: Distributive powers. Example: the square of a binomial r2 = (at + bt)2 , as
permutations and combinations in simultaneity. Note the combinations with the augmented
original ra and the fractioning pattern rn÷m (top).

120 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

• Case 1, determinant n = 4: r = 3t + t (a = 3, b = 1, T = 4t). There are two

permutations {r1 , r2 } = {(a + b), (b + a)}. The distributive second power
yields the square

r12 = (3 + 1)(3t + t) = 3(3t + t) + 1(3t + t) = (9t + 3t) + (3t + t)

r22 = (1 + 3)(t + 3t) = 1(t + 3t) + 3(t + 3t) = (t + 3t) + (3t + 9t).

This four-attack pattern has total duration Tt = T 2 = 16t. The first pattern
is an accelerating rhythm, the second has increasing durations between the
attacks. The timescaled, augmented version of the original is

r1a = n(at + bt) = 4(3t + t) = 12t + 4t

r2a = n(bt + at) = 4(1t + 3t) = 4t + 12t.

Recall the fractioning pattern rn÷m from Section 4.1

rn÷m = r4÷3 = 3t + t + 2t + t + t + t + t + 2t + t + 3t,

a symmetrical pattern where the retrograde is the original and the short-
est durations occur in the middle. These five patterns are displayed in
Fig. 12.4.a, and in the five-part score in Fig. 12.3, m. 1–4. Combinations
of two to five of these options in any vertical order can be based on the de-
liberate design of an accelerating or slowing-down rhythmical pattern. The
synchronization of all five leads to attack on almost any time unit.
• Case 2, determinant n = 5: r = 3t + 2t (a = 3, b = 2, T = 5t). Squaring the
two permutations yields

r12 = (3 + 2)(3t + 2t) = 3(3t + 2t) + 2(3t + 2t) = (9t + 6t) + (6t + 4t)
r22 = (2 + 3)(2t + 3t) = 2(2t + 3t) + 3(2t + 3t) = (4t + 6t) + (6t + 9t).

The two timescaled versions of the original r are

r1a = n(at + bt) = 5(3t + 2t) = (15t + 10t)

r2a = n(bt + at) = 5(2t + 3t) = (10t + 15t).

Now we have three options for the fractioning pattern

r5÷2 = 2t + 2t + t + t + t + t + t + t + t + t + t +
t + t + t + t + t + t + t + t + 2t + 2t
r5÷3 = 3t + 2t + t + 2t + t + t + t + t + t +
t + t + t + t + 2t + t + 2t + 3t
r5÷4 = 4t + t + 3t + t + t + 2t + t + 2t + t + t + 3t + t + 4t.

These three contain more attacks than either the square or the augmented
original group. The seven
h i patterns are shown in Fig. 12.4.b and in the score
in Fig. 12.3 m. 5–9 in 54 . Again, one may choose any combination of a
subset of these seven alternative patterns in a multi-part score.

Copy issued to Seth Dempsey ([email protected]) 121

CHAPTER 12. DISTRIBUTIVE POWERS

b b b b r2
1
b b b b r22
b b r1a
b b r2a
b bb bbbbb bb r4÷3
-
0 5 10 15 t
(a) r = 3t + t

b b b b r12
b b b b r22
b b r1a
b b r2a
b b bbbbbbbbbbbbbbbbbb b r5÷2
b b bb bbbbbbbbbb bb b r5÷3
b bb bbb bb bbb bb r5÷4
-
0 5 10 15 20 25 t
(b) r = 3t + 2t

Figure 12.4: Combination of the square of a binomial r2 (determinant n = a + b) with the

augmented original ra , their permutations and the fractioning pattern rn÷m . Shown are two
cases for the binomials r = (3t + t) (a) and r = (3t + 2t) (b).

122 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

The square of a trinomial

Taking the distributive second power of a trinomial r = (a + b + c) yields the square

r2 = (a + b + c)2 = (a + b + c)(a + b + c)
= a(a + b + c) + b(a + b + c) + c(a + b + c)
= (a2 + ab + ac) + (ba + b2 + bc) + (ca + cb + c2 )
= a2 + ab + ac + ba + b2 + bc + ca + cb + c2 . (12.4)

For a three-element group the square contains nine attacks. Practical trinomials for the dis-
tributive second power are shown in the fourth column of Table 12.1. In the Schillinger book
there are only trinomials with a doubled element, such as (2 + 2 + 1)2 ; the possible combi-
nations with three different elements, such as (4 + 2 + 1)2 are discarded. The explanation
for this is given in Chapter 13, where he discusses families of rhythms that are the result of
an evolutionary process. Figure 12.5 shows the graphical representation of the square of a
subset of these trinomials.
The original trinomial is now reproduced at either two or three different timescales, de-
pending on the ratios a/b and a/c. Only one permutation is displayed, with the coefficients
in decreasing order, i.e., with a > b > c Each combination with two equal elements has three
permutations, e.g., (5 + 2 + 2)2 , (2 + 5 + 2)2 , (2 + 2 + 5)2 , trinomials with three different
elements have 6 permutations.

Example 12.4
The square of a trinomial.
We will perform the calculations for a number of three-element attack-duration
groups with r = (at + bt + ct) and determinant n = a + b + c. We will select a
number of cases, shown in diagram in Fig. 12.5 and calculate the results using
Eq. 12.4:

• Case 1, determinant n = 6: r = 4t + t + t (a = 4, b = c = 1, T = 6t), with two

equal elements. The distributive second power yields the square

r2 = (4 + 1 + 1)(4t + t + t) = 4(4t + t + t) + 1(4t + t + t) + 1(4t + t + t)

= (16t + 4t + 4t) + (4t + t + t) + (4t + t + t).

This nine-attack pattern with total duration Tt = T 2 = 36t contains three

subgroups with durations 24t, 6t, 6t, respectively. There are three permuta-
tions of the subgroups: these are {(16t+4t+4t)+(4t+t+t)+(4t+t+t), (t+
4t + t) + (4t + 16t + 4t) + (t + 4t + t), (t + t + 4t) + (t + t + 4t) + (4t + 4t + 16t)},
shown as a simultaneity in Fig. 12.6, m. 1–9 and applied h ito an extended
Dm7 chord. The pattern will group into 12 measures at 38 time signature
11
h i
4
or nine measures in 4 meter.
• Case 2, determinant n = 6: r = 3t + 2t + t (a = 3, b = 2, c = 1, T = 6t), with
three different elements. The square of r is

r2 = (3 + 2 + 1)(3t + 2t + t)

Copy issued to Seth Dempsey ([email protected]) 123

CHAPTER 12. DISTRIBUTIVE POWERS

b b b b bbb bb
(2 + 1 + 1)2
b b b bbb bbb
(3 + 1 + 1)2
b b bb b bbbb
(2 + 2 + 1)2
-
0 10 20 30 40 50 60 70 80 t

b b b b bbb bb
(4 + 1 + 1)2
b b b b b bb bb
(3 + 2 + 1)2
-
0 10 20 30 40 50 60 70 80 t

b b b b b b b b b
(3 + 2 + 2)2
b b b b b b b b b
(3 + 3 + 1)2
b b b b b bb bb
(4 + 2 + 1)2
-
0 10 20 30 40 50 60 70 80 t

b b b b bbb bb
(6 + 1 + 1)2
b b b b b b b b b
(3 + 3 + 2)2
-
0 10 20 30 40 50 60 70 80 t

b b b b b bb bb
(6 + 2 + 1)2
b b b b b b b b b
(5 + 2 + 2)2
b b b b b b b b b
(4 + 3 + 2)2
-
0 10 20 30 40 50 60 70 80 t

Figure 12.5: Distributive powers. The square of a trinomial r2 = (at + bt + ct)2 with determi-
nant n = a + b + c; trinomials with two equal elements have three permutations, trinomials
with three different elements have six permutations.

124 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

= 3(3t + 2t + t) + 2(3t + 2t + t) + 1(3t + 2t + t)

= (9t + 6t + 3t) + (6t + 4t + 2t) + (3t + 2t + t).

The three subgroups may be used in six permutations; the three circular
permutations are shown in Fig. 12.6, m. 10–21, applied to a G69 chord. Note
that all permutations have a simultaneous attack in hm.i 10 and 18. The total
duration Tt = 36t, corresponding to 12 measures in 34 .
• Case 3, determinant n = 8: r = 3t + 3t + 2t (a = b = 3, c = 2, T = 8t), a
familiar rhythm pattern in Latin music , e.g., in the rumba, and in popular
music at time unit t = 18 . The square of this trinomial is

r2 = (3 + 3 + 2)(3t + 3t + 2t)
= 3(3t + 3t + 2t) + 3(3t + 3t + 2t) + 2(3t + 3t + 2t)
= (9t + 9t + 6t) + (9t + 9t + 6t) + (6t + 6t + 4t).

The three subgroups have

h lengths
i 24t + 24t + 16t, Tt = 82 t = 64t, leading to
an 8 measure pattern in 88 grouping. The three permutations for r = 3t +
3t̂ + 2t with an accented note are shown in rhythmical notation in m. 22–29;
note the syncopations and afterbeats with occasional simultaneous attacks.

Also the square of the trinomial can be used as a continuity rc = r2 or as a simultaneity

with the augmented version of the original group nr = n(a + b + c) = na + nb + nc or with
the fractioning pattern rn÷m .

Example 12.5
Combination of the square of a trinomial with the original and the fractioning
pattern.
In this example we will look at the combinations of the trinomial r = (at + bt +
ct) = (4t + 1t + 1t), determinant n = a + b + c = 6. The square r2 has duration
Tt = 36t and will be combined with the augmented, timescaled version of the
original r and the fractioning pattern rn÷m .

• Determine the square r2 and its permutations (see Example 12.4, Case 1)

r12 = (16t + 4t + 4t) + (4t + t + t) + (4t + t + t)

r22 = (t + 4t + t) + (4t + 16t + 4t) + (t + 4t + t)
r32 = (t + t + 4t) + (t + t + 4t) + (4t + 4t + 16t).

• The timescaled versions of the original r are

r1a = n(at + bt + ct) = 6(4t + t + t) = 24t + 6t + 6t

r2a = 6t + 24t + 6t
r3a = 6t + 6t + 24t.

Copy issued to Seth Dempsey ([email protected]) 125

CHAPTER 12. DISTRIBUTIVE POWERS

œ œ ˙ ˙˙ œœ œœ ww ww ww ww ww ww ww
(t+t+4t)^2

& 44 œœ œœ ˙˙ 43
1
˙ œ œ w w w w w w w
(t+4t+t)^2

& 44 œ ˙ .. œœ œœ ˙˙ ˙˙ ˙˙ ww ww ww ˙˙ ˙˙ ˙˙ œœ œœ ˙˙ ..
3
œœ 4
œ ˙
(4t+t+t)^2
? 44 w w w w w w w œ œ ˙ ˙ œ œ 43
(
D m7 add11 )

˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ œœ ˙˙ .. ˙˙ .. ˙˙ ..
(t+3t+2t)^2

& 43 œœ ˙˙ œœ ˙˙ 44
10

(3t+2t+t)^2
3 4
& 4 ˙. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. œœ ˙˙ ˙˙ .. ˙˙ œœ 4
˙.
(2t+t+3t)^2
? 43 ˙ . œ ˙ ˙ .. ˙ .. ˙ œ ˙ .. ˙ .. ˙ .. ˙ .. ˙ .. ˙ .. ˙ .. 44
˙. œ ˙ ˙ ˙ ˙ œ ˙ ˙ ˙ ˙ ˙ ˙ ˙
6
G 9

(3t+3t'+2t)^2

& 44 | Û Û. | Û |. Û Û. | Û |. |. 68
22
| Û | |
J J
(3t'+2t+3t)^2 > > >
4 Û Û. Û. Û | 6
&4 | |. Û Û |. | Û Û. Û. Û | 8
> J J > J J
(2t+3t+3t')^2 >
? 44 | | Û |. |. Û | Û. Û | |. Û | Û. Û | 68
J J
> > >

˙. ˙. ˙ œ œ ˙ ˙ œ œ ˙ œ œ
(4t+t+t)^2/(t+4t+t)^2

& 68 œ ˙
30
œ ˙ œ ˙. ˙. œ ˙ œ ˙ œ
J J J J
(24t+6t+6t)/(6t+24t+6t)

& 68 ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ .. ˙˙ ..
frac(6,5)
? 68 œ . œ œ ˙ œ œ œ. œ œ œ œ œ. œ œ ˙ œ œ œ.
J J J J
Em

Figure 12.6: Distributive powers. Example: the square of a trinomial r2 = (at + bt + ct)2 with
determinant n = a + b + c.

126 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

b b b b bbb bb r12
bb bb b b bb b r22
bbb bbb b b b r32
b b b r1a
b b b r2a
b b b r3a
b bb bbb bb b bb bbb bb r6÷5
-
0 5 10 15 20 25 30 35 t

Figure 12.7: Combination of the square of a trinomial r2 with the augmented original ra ,
their permutations and the fractioning pattern tn÷m . Shown is the trinomial r = (4t + t + t)
with determinant n = 6.

• Now we have one option for the fractioning pattern (see Section 4.1)

r6÷5 = 5 + 1 + 4 + 1 + 1 + 3 + 1 + 2 + 2 + 1 + 3 + 1 + 1 + 4 + 1 + 5.

The seven patterns are shown in Fig. 12.7. Use any combination of a subset
of these seven alternative patterns in a multi-part score. A five-part combi-
nation, using two permutations of the square (r12 , r22 ), two augmented vari-
ants (r1a , r2a ) and the fractioning pattern
h i is shown in Fig. 12.6 m. 30–35, ap-
6
plied to an Em chord in meter 8 . The fractioning pattern is the busiest
rhythm.

12.2.2 The cube of an attack-duration group

The distributed third power of an attack-duration group is the cube of the resultant r3 . The
procedure is the same as for the square; after completing the distributive second power we
repeat the multiplication process once more. This is an iterative procedure.

The cube of a binomial

The cube of the binomial r = (a + b) with total duration T = n is

r3 = (a + b)3 = (a + b)(a + b)2 = (a + b)(a2 + ab + ba + b2 )

= (a3 + a2 b + aba + ab2 ) + (ba2 + bab + b2 a + b3 )
= (a3 + a2 b + a2 b + ab2 ) + (ba2 + ab2 + b2 a + b3 ), (12.5)

a series of 8 attacks with total duration n3 . The calculation for the easiest case r = 2t + t
yields r3 = (2t + t)3 = (2 + 1)(4t + 2t + 2t + t) = (8t + 4t + 4t + 2t) + (4t + 2t + 2t + t).
The three practical binomials for the cube are shown in the third column of Table 12.1, some
examples are shown graphically in the upper half of Fig. 12.8.

Copy issued to Seth Dempsey ([email protected]) 127

CHAPTER 12. DISTRIBUTIVE POWERS

(2 + 1)3
b b b bb bbb

(3 + 1)3
b b b b b b b b

(3 + 2)3
b b b b b b b b
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

b b b b b b b b b b b b b bbb bbb b b b bbb bb (2 + 1 + 1)3

(3 + 1 + 1)3
b b b b b b b b b b b b b bbb bbb b b b bbb bb

(2 + 2 + 1)3
b b b b b b b b bb b b b b b b b bb b bb b bbbb
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

Figure 12.8: Distributive powers. The cube of some binomials r3 = (at + bt)3 and trinomials
r3 = (at + bt + ct)3 . The cube of a binomial consists of eight attacks in two subgroups, the
cube of the trinomial contains three subgroups with nine attacks each.

128 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

Also the cube leads to a continuity with two subgroups of four attacks and two permuta-
tions. Obviously for determinants n ≥ 3 the total duration of the cube explodes; the example
(3t + 2t)2 already leads to a total duration of (53 )t = 125t time units. Therefore the table has
no entries beyond n = 5.

Example 12.6
The cube of a binomial.
We will perform the calculations for a number of two-element attack-duration
groups with r = (at + bt). All practical cases for cubing are shown in diagram in
Fig. 12.8:

• Case 1, determinant n = 3: r = 2t + t (a = 2, b = 1, T = 3t). The distributive

third power according to Eq. 12.5 yields the cube

r2 = (2 + 1)(2t + t) = 2(2t + t) + 1(2t + t) = (4t + 2t) + (2t + t)

r3 = (2 + 1)(2t + t)2 = 2(4t + 2t + 2t + t) + 1(4t + 2t + 2t + t)
= (8t + 4t + 4t + 2t) + (4t + 2t + 2t + t).

This accelerating eight-attack pattern with total duration Tt = T 3 = 33 t =

27t consists of two subgroups
h i with durations 18t, 9t, respectively. This im-
3
plies 9 measures in 8 grouping. The other permutation is r = t + 2t with
the cube r3 = (t + 2t + 2t + 4t) + (2t + 4t + 4t + 8t). Both are shown in
Fig. 12.9, m. 1–9, applied to an Am7 chord. There is also the distribution of
the duration group over the two-part attack patterns (see the technique in
Section 8.1).
• Case 2, determinant n = 5: r = 3t + 2t (a = 3, b = 2, T = 5t). The cube of
this binomial is

r3 = (3 + 2)(3t + 2t)2 = 3(9t + 6t + 6t + 4t) + 2(9t + 6t + 6t + 4t)

= (27t + 18t + 18t + 12t) + (18t + 12t + 12t + 8t).

For the square see Example 12.3 Case 2. The two subgroups are now more in
balance 75t + 50t, Tt = 53 t = 125t and the longest to shortest note duration
ratio
h iis 27t/8t = 3.37. The determinant n = 5 requires irregular meter such
as 58 in folk music, yielding a 25-measure rhythmical pattern. The cube is
demonstrated in two simultaneous permutations in a pair of staves for the
pattern r = 3t+2r̄ in m. 10–34, where we see sustained 2- and 3-part chords,
interspersed with long rests.
• Case 3, determinant n = 4: r = 3t + t (a = 3, b = 1, T = 4t). The cube of r is

r3 = (3 + 1)(3t + t)2 = 3(9t + 3t + 3t + t) + 1(9t + 3t + 3t + t)

= (27t + 9t + 9t + 3t) + (9t + 3t + 3t + t).

For the square see Example 12.2 Case 1. The ratio of longest to shortest
attack now is 27t/1t = 27. The total duration Tt = 64t, with subgroupsh i
of lenght 48t and 16t, respectively. This corresponds to 8 measures in 44

Copy issued to Seth Dempsey ([email protected]) 129

CHAPTER 12. DISTRIBUTIVE POWERS

grouping and time unit t = 18 . This is shown in the upper staff of Fig. 12.9,
m. 35–42. Here the cube is applied to an attack-duration group r = 3t + r̄,
containing a rest (see Section 9.2 for the technique of using accented notes
and rests) and an Em11 7 chord. Two permutations are shown on the upper
staff.

Creating a simultaneity now opens up a number of possibilties for combining the original
group, the square, the cube and the fractioning pattern. Each of these must be properly
scaled, i.e., multiplied with the a power of the determinant n to achieve synchronization
through augmentation. The possibilities may be written as a four-part score

n2 r = n2 (a + b)
 
P1 :
 P2 : nr2 = n(a + b)2 
rs  . (12.6)
 
P3 : r 3 = (a + b)3
 
P4 : nrn÷m

Remember that each part has a number of permutations, increasing the potential for combi-
nation even further.

Example 12.7
Combination of the cube of a binomial with the original, square and the frac-
tioning pattern.
The example uses the binomial r = (at + bt) = (3t + t), determinant n = a + b = 4.
In this example the cube r3 will be combined with the augmented, timescaled
versions of the original r, the augmented square r2 and the fractioning pattern
rn÷m = r4÷3 .

• The cube is r3 = (3t + t)3 = (27t + 9t + 9t + 3t) + (9t + 3t + 3t + t) with two

permutations {r13 , r23 }, see Example 12.6 Case 3.
• The timescaled version of the square nr2 has two permutations

r12,a = 4[(9t + 3t) + (3t + t)] = (36t + 12t) + (12t + 4t)

r22,a = 4[(t + 3t) + (3t + 9t)] = (4t + 12t) + (12t + 36t).

• The permutations of the augmented version of the original pattern n2 r are

r1a = n2 (at + bt) = 16(3t + t) = 48t + 16t

r2a = n2 (bt + at) = 16(1t + 3t) = 16t + 48t.

• The augmented fractioning pattern is (see Section 4.1)

nrn÷m = nr4÷3
= 4(3t + t + 2t + t + t + t + t + 2t + t + 3t)
= 12t + 4t + 8t + 4t + 4t + 4t + 4t + 8t + 4t + 12t.

130 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

œœ .. œœ .. œœ œœ œœ . œœ .. œœ œœ œœ .. œœ œœ œœ œœ 5
(2t+t)^3, a+a (3t+2r)^3, a+a

& 38 . . . . .
1

J J J J 8 œ œ œ œ œ œ œ œ

3 j j
(t+2t)^3, 2a+2a
j j
(2r+3t)^3,a+a

& 8 œ œ œœ œœ œœ .. œ œ œ. œœ .. œœ œœ œœ .. œœ .. 85 ∑ Œ . œœ œœ .. œœ œœ .. œœ
œ œ œ œ œ. E m7(add4)
A m7

œ. œ œ. œ œ. œ œ. Œ
& œ. œ œ Œ.
14
∑ ∑ ∑ ∑ ∑

& ∑ ∑ Œ œœ .. œœ .. œœ œœ .. œœ œœ .. œœ ∑ ∑ Œ œ .. œœ .. œœ œœ .. œœ
œ

œ. œ œ. œ œ ‰Œ 44
& œ. œ œ. œ œ. œ œ. Œ
25
∑ ∑ ∑

& œœ .. œœ ∑ ∑ ∑ Œ . œœ œœ .. œœ œœ .. œœ œœ .. œœ œœ .. œœ œœ .. œœ 44

j j
4 w w w œ ˙ œ˙ . ‰ Ó Ó ˙ œ ˙w œ Œ . ww œw ‰ Œ œ . ‰
35 (3t+r)^3/(r+3t)^3

& 4 ‰ œ . Œ . œJ w Œ. J Ó Ó Œ
4(3t+r)^2/4(r+3t)^2
Ó Ó
& 44 wwÓ ˙ www ww wwÓ ˙ ˙˙w w www ˙˙w

? 44 w w w w w w
16(3t+r)

∑ ∑
4frac(4,3),a+r+a+r
? 44 w ˙ Ó w Ó ˙ Ó ˙ ∑ ˙ Ó ∑
E m7

> ˙ > > > > >

Œ œ ˙ Œ œ œ ‰ œj œ ‰ œJ ˙ Œ œ œ ‰ œj œ ‰ œJ
43 (2t+r+t')^3,a+a+a
&w Ó ˙ Œ œ ˙
> > >
j j
(t'+2t+r)^3

& œj œ ‰ œ œ Œ œ ‰ œ ˙ Œ ˙ ˙ Ó œ Œ j
œ œ ‰ œ œ œ Œ >œ œ ‰
> >œ > œ > ˙ >˙ > >
>
> >œ >œ
? ‰ œJ œ ‰ >œ œ Œ œ >œ > ˙
(r+t'+2t)^3
˙ ‰Jœ ‰ œJ Œ œ ˙ Œ ˙ Œ œ Ó ˙ w
J > > >
Em
4frac(4,3)
?¿ ŒÓ Ó ¿ Œ ¿ Œ Ó ¿ Œ ¿ Œ ¿ Œ ¿ Œ ¿ Œ Ó ¿ Œ ¿Œ ∑

Figure 12.9: Distributive powers. Example: the cube of a binomial r3 = (at + bt)3 and
trinomial r3 = (at + bt + ct)3 . Permutations and combinations in simultaneity.

Copy issued to Seth Dempsey ([email protected]) 131

CHAPTER 12. DISTRIBUTIVE POWERS

b b b b b b b b r3
1
bb b b b b b b r23
b b b b r12,a
b b b b r22,a
b b r1a
b b r2a
b b b b b b b b b b nr4÷3
-
0 5 10 15 20 25 30 35 40 45 50 55 60 65 t

Figure 12.10: Combination of the cube of a binomial r3 with the augmented original ra , the
augmented square r2,a and the fractioning pattern rn÷m . Shown is the binomial r = (3t + t),
determinant n = 4.

The seven patterns are shown in Fig. 12.10. Use any combination of a subset
of these seven rhythmical patterns in a multi-part score. The example with
a combination of six is shown in Fig. h i 12.9, m. 35–42, applied to the resultant
r = 3t + r̄, with an Em7 chord in 44 . The upper staff combines two permu-
tations of the cube (r13 , r23 ), the second staff from the top has the timescaled
squares (r12,a , r22,a ), the third staff has one augmented original (r1a ) and the
bottom staff contains the fractioning pattern.

The cube of a trinomial

The cube of the trinomial r = (a + b + c) is

r3 = (a + b + c)3 = (a + b + c)(a + b + c)2

= a(a2 + ab + ac + ba + b2 + bc + ca + cb + c2 ) +
b(a2 + ab + ac + ba + b2 + bc + ca + cb + c2 ) +
c(a2 + ab + ac + ba + b2 + bc + ca + cb + c2 )
= (a3 + a2 b + a2 c + aba + ab2 + abc + aca + acb + ac2 ) +
(ba2 + bab + bac + b2 a + b3 + b2 c + bca + bcb + bc2 ) +
(ca2 + cab + cac + cba + cb2 + cbc + c2 a + c2 b + c3 )
= (a3 + a2 b + a2 c + a2 b + ab2 + abc + a2 c + acb + ac2 ) +
(ba2 + b2 a + bac + b2 a + b3 + b2 c + bca + b2 c + bc2 ) +
(ca2 + cab + c2 a + cba + cb2 + c2 b + c2 a + c2 b + c3 ). (12.7)

There are 27 attacks, subdivided into three subgroups, whose durations depend on the ra-
tios a/b and a/c. Practical trinomials for the distributive third power are shown in the last
column of Fig. 12.1. This is a small set of 6 trinomials, of which 5 have two equal elements.

132 Copy issued to Seth Dempsey ([email protected])

 12.2. RHYTHMIC COUNTERTHEMES BY DISTRIBUTIVE POWERS

Some examples are shown graphically in the lower half of Fig. 12.8. These may have three
or six permutations. Again, note the total pattern duration of the cube as a continuity.

Example 12.8
The cube of a trinomial.
We will perform the calculations for the cube of the trinomials r = (at + bt + ct)
shown in diagram in the lower half of Fig. 12.8. The three cases are:

• Case 1, determinant n = 4: r = 2t + t + t (a = 2, b = c = 1). The distributive

power approac used iteratively yields the third power, i.e., the cube

r2 = (2 + 1 + 1)(2t + t + t)
= (4t + 2t + 2t) + (2t + t + t) + (2t + t + t)
3
r = r(r2 ) = (2 + 1 + 1)(2t + t + t)2 =
= (2 + 1 + 1)[(4t + 2t + 2t) + (2t + t + t) + (2t + t + t)]
= [(8t + 4t + 4t) + (4t + 2t + 2t) + (4t + 2t + 2t)]
+[(4t + 2t + 2t) + (2t + t + t) + (2t + t + t)]
+[(4t + 2t + 2t) + (2t + t + t) + (2t + t + t)].

The 27-attack pattern with total duration Tt = T 3 = 43 t = 64t consists

of three subgroups
h i with durations 32t, 16t, 16t, respectively. This implies
8 measures in 44 grouping at time unit t = 18 , as illustrated in Fig. 12.9,
m. 43–53. Here the original resultant contains a rest and an accented note
r = 2t + r̄ + t̂ and there is an attack group distribution over an Em chord.
The arpeggio patterns in the three upper staves demonstrate the three per-
mutations in simultaneity, combined with a percussion rhythm in the lower
staff, based on the augmented fractioning pattern rn÷ma = nrn÷m = 4r4÷3 .
Note how busier rhythms are moving through the upper parts against a
more regular and steady lower part.
• Case 2, determinant n = 5: r = 3t + t + t (a = 3, b = c = 1). The cube of r is
calculated in two steps

r2 = (3 + 1 + 1)(3t + t + t)
= (9t + 3t + 3t) + (3t + t + t) + (3t + t + t)
3
r = r(r2 ) = (3 + 1 + 1)(3t + t + t)2 =
= (3 + 1 + 1)[(9t + 3t + 3t) + (3t + t + t) + (3t + t + t)]
= [(27t + 9t + 9t) + (9t + 3t + 3t) + (9t + 3t + 3t)]
+[(9t + 3t + 3t) + (3t + t + t) + (3t + t + t)]
+[(9t + 3t + 3t) + (3t + t + t) + (3t + t + t)].

The total duration of this 27-attack group Tt = 53 t = 125t, divided into three
subgroups 75t + 25t + 25t = 125t. This leads to a visible unbalance when
comparing the length of the first with the other subgroups. This case is not
shown in musical notation.

Copy issued to Seth Dempsey ([email protected]) 133

CHAPTER 12. DISTRIBUTIVE POWERS

• Case 3, determinant n = 5: r = 2t + 2t + t (a = b = 2, c = 1). The cube of

this binomial is

r2 = (2 + 2 + 1)(2t + 2t + t)
= (4t + 4t + 2t) + (4t + 4t + 2t) + (2t + 2t + t)
3
r = r(r2 ) = (2 + 2 + 1)(2t + 2t + t)2 =
= (2 + 2 + 1)[(4t + 4t + 2t) + (4t + 4t + 2t) + (2t + 2t + t)]
= [(8t + 8t + 4t) + (8t + 8t + 4t) + (4t + 4t + 2t)]
+[(8t + 8t + 4t) + (8t + 8t + 4t) + (4t + 4t + 2t)]
+[(4t + 4t + 2t) + (4t + 4t + 2t) + (2t + 2t + t)].

The three subgroups add up to 50t + 50t + 25t = 125t, with an improved
balance between the subgroups.

In a synchronized simultaneity we may once again combine the cube P1 : r3 = (a+b+c)3

with the timescaled version of the original group P2 : n2 r = n2 (a + b + c) and the augmented
square P3 : nr2 = n(a + b)2 or the fractioning pattern P4 : nrn÷m , in any permutation.

12.2.3 The generalization of all powers

The process of taking the distributive power of a binomial, trinomial or multinomial may
be extended to higher powers, beyond the cube (third power). In general this implies the
N -th power rN = (a + b + . . .)N . In practice the calculations are done in an iterative process,
written as
rN = r(rN −1 ). (12.8)
As we have seen in the caluculations the cube of a binomial involves the distributive multi-
plication process applied to the square, i.e., r3 = (at + bt)3 = r(r2 ) = (a + b)(at + bt)2 . This
approach is repeated at higher order powers: r4 = r(r3 ), etc.
These may yield continuities and simultaneities after proper timescaling. The essential
point is that all the higher order powers are derived from a single source attack-duration
group r; this acts as a binding element stylewise and will create a homogeneous continuity
or simultaneity.

134 Copy issued to Seth Dempsey ([email protected])

Chapter 13

Evolution of rhythm styles (families)

This chapter is about families of rhythms, as they can be derived from a specific determi-
nant. It connects the previous chapters. Chapter 9 and 10 covered the technique of variation
through permutation on the smaller timescale and at higher aggregation levels. Chapter 11
showed the technique of splitting as another approach to small-scale variation. Then Chap-
ter 12 explained a mechanism for creating longer multi-measure rhythmical patterns using
distributive powers.
In this chapter there is the concept of the evolution of rhythms from a common starting
point, given by the determinant n or meter grouping nn . The starting point of the evolution
 

is the subdivision of a group with determinant n = r into a binomial with a major generator
a and minor generator b, such that r = a + b.
Starting from this binomial we can proceed with the subdivision approach, i.e., evolution
on the fractional level. Using distributive powers there is the potential for evolution on the
larger scale, i.e., on the factorial level. The evolution is determined by a sequence of the two
processes permutation and synchronization, as illustrated in Fig. 13.1.
On the fractional level we write the two possible permutations of the binomial {a + b, b +
a}, as illustrated by the two attacks on the right at the ouput of the permutation process
block. Synchronization yields three attacks. Then there is a test to check for uniform distri-
bution with all attacks of equal duration. We stop the process when the ultimate subdivision
into equal time units has been reached. Otherwise we iterate the process for the new trino-
mial or multinomial.
On the factorial level there is a similar approach. We raise the binomial to the distributive
second power, i.e., the square r2 = (a + b)2 . This leads to four attacks, to which we then also
apply iterative permutation and synchronization until the uniform distribution is reached.
The integer numbers for the new elements we obtain along this process establish a family
of rhythms, that carry specific characteristics. We will now consider this evolution process
on the fractional and factorial level for practical determinants between n = 2, . . . , 9. We will
use a subscript to indicate the generations in the evolution process: r0 is the parent pattern,
which yields the first generation of children with resultant r1 after one iteration cycle of
permutation and synchronization.

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 13. EVOLUTION OF RHYTHM STYLES

Determinant n
r Fractional - Factorial r2
Binomial r = a + b b b |
- Trinomial r = a + b + c
Multinomial r = a + b + . . .
?
Permutation
b b |
b b |
?
Synchronization
b bb |
?
HH
 Uniform HH
 H
 H bbbbb|
No Hdistribution?
HH 
H  
HH
?Yes
Stop

Figure 13.1: Iterative process of permutation and synchronization. The initial binomial pair
{a, b} with determinant n will produce generations of a rhythm family on either the frac-
tional or the factorial level.

13.1 Evolution on the fractional level

Subdividing the determinant n = 2 leads to the only possible binomial r0 = a + b = 1 + 1,
as shown in Fig. 13.2. This is a uniform distribution a = b = 1; there are no permutations of
this rhythmical pattern. There is no evolutionary process and we have the only member of
this family; this is a trivial case.
The permutation-synchronization process starts to make sense at determinant n = 3.
There is one possible subdivision into the binomial r0 = a + b = 2 + 1, where a = 2 is the
major generator and b = 1 the minor generator. When we synchronize the two permutations
we reach uniform distribution r1 = 1 + 1 + 1 and stop the evolution.
For determinant n = 4 there is one non-uniform distribution r0 = a + b = 3 + 1, with
major generator a = 3 (see Table 2.1 in Section 2.1.2 for the set of unique combinations for
a given major generator). As Fig. 13.2 shows (top right), the synchronization of the two
permutations yields the next generation trinomial r1 = 1 + 2 + 1, which is a non-uniform
pattern with two equal elements (a1 = 2, b1 = 1, r1 = b+a+b). Therefore we enter the second
iteration of the evolution process; now permutation and synchronization reach uniformity
r2 = 1 + 1 + 1 + 1. This is another family of rhythms.
Determinant n = 5 is the first to have two possible subdivisions into binomials r0 =
a + b = {3 + 2, 4 + 1}. For both cases the lower half of the figure illustrates the evolution
process: r0 = 3 + 2 → r1 = 2 + 1 + 2 → r2 = 1 + 1 + 1 + 1 + 1 and r0 = 4 + 1 → r1 =
1 + 3 + 1 → r2 = 1 + 1 + 1 + 1 + 1. We may discern the general characteristic of the first
generation resultant: it will always be a trinomial with two equal elements, specific for that
family. This is the reason why Table 12.1 in Schillinger’s book only listed these trinomials

136 Copy issued to Seth Dempsey ([email protected])

 13.1. EVOLUTION ON THE FRACTIONAL LEVEL

h i h i h i
2 3 4
2 3 4 Determinant

bb|1+1 b b|2+1 b b|3+1 r0 Binomial a + b

bb |1+2 bb |1+3 Permutation
↓ Synchronize
bbb|1+1+1 bb b|1+2+1 r1 Trinomial
b bb|2+1+1 Permutations
..
bbb |1+1+2 .
↓ Synchronize
bbbb|1+1+1+1 Uniform distribution
n=2 n=3 n=4

h i h i
5 5
5 5 Determinant

b b |3+2 b b|4+1 r0 Binomial a + b

b b |2+3 bb |1+4 Permutation
↓ Synchronize
b bb |2+1+2 bb b|1+3+1 r1 Trinomial
bb b |1+2+2 b bb|3+1+1 Permutations
..
b b b|2+2+1 bbb |1+1+3 .
↓ Synchronize
bbbbb|1+1+1+1+1 bbbbb|1+1+1+1+1 Uniform distribution
n=5 n=5

Figure 13.2: Fractional evolution of rhythm families through iterations of permutation and
synchronization. Determinants n = 2, 3, 4, 5. The parent generation is the binomial r0 = a+b,
the i-th child generation is labeled ri .

Copy issued to Seth Dempsey ([email protected]) 137

CHAPTER 13. EVOLUTION OF RHYTHM STYLES

and discarded the trinomials with three different elements. In his theory of rhythm, the latter
are standalone patterns and not part of an evolving family.
Determinant n = 6 has one non-uniform binomial r0 = 5 + 1, see Fig. 13.3. There are two
generations of non-uniform descendants r1 = 1 + 4 + 1 and the quintinomial, five-element
r2 = 1 + 1 + 2 + 1 + 1. Uniform duration distribution is reached at child generation r3 .
Determinant n = 7 also leads to three generations, but now there are three parent binomials
r0 = {4 + 3, 5 + 2, 6 + 1}. The rhythmical asymmetry increases (i.e., the ratio of longest to
shortest duration) towards the last binomial, which might be characterized as most unbal-
anced.
The rhythm families of determinants n = 2, . . . , 7 are shown in musical notation in
Fig. 13.4. The parent generation r0 , the starting binomial is shown on the top staff. The
second staff from the top shows the retrograde permutation, the third staff has the result of
the first synchronization step. In case of a non-uniform child generation r1 , the three permu-
tations are shown on the grouped staves 3 to 5 (remember that two elements are equal and
therefore the set of permutations is 3!/2! = 3). The five permutations of the second genera-
tion quintinomial r2 are shown on grouped staves 6–10. The uniform distribution after the
last synchronization process is shown on the bottom staff.
Also for the determinants n = 8 and n = 9 the iterative process stops at the third gener-
ation, as shown in Fig. 13.5. The former has two parent binomials r0 = {5 + 3, 7 + 1}, the
latter has three starting non-uniform binomials 9 = {5 + 4, 7 + 2, 8 + 1}. Note how all 2nd
generation quintinomials have either 4 equal elements, r2 = b + b + a + b + b or the pattern
r2 = b + a + b + a + b (two equal larger values a and three equal smaller values b). These
rhythmical patterns are shown in musical notation in Fig. 13.6.

13.2 Evolution on the factorial level

The evolutionary process may also be applied on the factorial level by using distributive
powers. These powers may be applied to the parent generation binomial, i.e., the square
r02 = (a + b)2 or to the trinomial from the first generation of children r12 = (b + a + b)2 .
After calculating the distributive power the resulting longer pattern will then undergo
the iterative process of permutation and synchronization, shown in diagram in Fig. 13.1. For
the simplest case, the square of a binomial, this is shown in numbers in Fig. 13.7 for deter-
minants n = 3, 4 and one binomial for n = 5, r02 = (3 + 2)2 . The patterns have duration n2
and the first generation r1 has four attacks (see Section 12.2.1 for the calculations involved).
The resultant has the general characteristic r1 = a + b + b + c, with twelve permutations.
For n = 2 and n = 3 the synchronization process yields uniform distribution at this child
generation r1 , and the evolution stops. For n = 5 the next generation is the 11 attack group
r2 = 4 + 2 + 3 + 1 + 2 + 1 + 2 + 1 + 3 + 2 + 4, with numerous permutations.
The same process may be illustrated for the evolution of the cube of a binomial, the
square of a trinomial but the number of permutations will be beyond handling and limit
practical musical applications. Although the resulting child generation patterns are unique
to the family it is questionable whether the listener will recognize or experience this charac-
teristic.
Schillinger in his book illustrates the evolution process with examples of rhythmical pat-
terns in classical and popular music and once again tries to prove the potential and h rising
i
importance of the determinants n = 6, 9, through the subdivision and grouping into 66 and

138 Copy issued to Seth Dempsey ([email protected])

 13.2. EVOLUTION ON THE FACTORIAL LEVEL

h i
6
6 Determinant

b b|5+1 r0 Binomial a + b
bb |1+5 Permutation
↓ Synchronize
bb b|1+4+1 r1 Trinomial
b bb|4+1+1 Permutations
..
bbb |1+1+4 .
↓ Synchronize
b b b b b | 1 + 1 + 2 + 1 + 1 r2 Quintinomial
b b b b b | 1 + 2 + 1 + 1 + 1 Permutations
..
b bbbb|2+1+1+1+1 .
..
bbbb b|1+1+1+2+1 .
..
bbbbb |1+1+1+1+2 .
↓ Synchronize
b b b b b b | 1 + ... + 1 Uniform distribution
n=6

h i h i h i
7 7 7
7 7 7

b b |4+3 b b |5+2 b b |6+1

b b |3+4 b b |2+5 b b |1+6

b bb |3+1+3 b b b |2+3+2 b b b |1+5+1

bb b |1+3+3 b b b |3+2+2 b b b |5+1+1
b b b|3+3+1 b b b |2+2+3 b b b |1+1+5

bb bb b|1+2+1+2+1 b bbbb |2+1+1+1+2 b bbbb |1+1+3+1+1

b bb bb|2+1+2+1+1 bbbb b |1+1+1+2+2 bbbb b |1+3+1+1+1
bb bbb |1+2+1+1+2 bbb b b|1+1+2+2+1 bbb b b|3+1+1+1+1
b bbb b|2+1+1+2+1 bb b bb|1+2+2+1+1 bb b bb|1+1+1+1+3
bbb bb |1+1+2+1+2 b b bbb|2+2+1+1+1 b b bbb|1+1+1+3+1

b b b b b b b | 1 + ... + 1 b b b b b b b | 1 + ... + 1 b b b b b b b | 1 + ... + 1

n=7 n=7 n=7

Figure 13.3: Fractional evolution of rhythm families. Determinants n = 6, 7

Copy issued to Seth Dempsey ([email protected]) 139

CHAPTER 13. EVOLUTION OF RHYTHM STYLES

& 42 œ œ 43 ˙ œ 44 4=3+1
˙. œ 45 ˙ . ˙ w œ 46
2 = 1+1 3=2+1 5=3+2 5=4+1

& 42 ∑ 43 œ ˙ 44 œ ˙ . 45 ˙ ˙. œ w 46

& 42 ∑ 43 œ œ œ 44 4=1+2+1
œ ˙ œ 45 ˙ œ ˙ œ ˙. œ 46
5=2+1+2 5=1+3+1

& 42 ∑ 43 ∑ 44 ˙ œ œ 45 œ ˙ ˙ ˙. œ œ 46

& 42 ∑ 43 ∑ 44 œ œ ˙ 45 ˙ ˙ œ œ œ ˙. 46

& 42 ∑ 43 ∑ 44 œ œ œ œ 45 œ œ œ œ œ œ œ œ œ œ 46

& 46 ˙ . ˙ œ 74 7=4+3
w ˙. ˙. ˙ ˙ w ˙ œ 48
6=5+1 7=5+2 7=6+1

& 46 œ ˙ ˙. 74 ˙ . w ˙ ˙. ˙ œ ˙ w 48
6 6=1+4+1 7 7=3+1+3 8
&4 œ w œ 4 ˙. œ ˙. ˙ ˙. ˙ œ ˙. ˙ œ 4
7=2+3+2 7=1+5+1

& 46 w œ œ 74 œ ˙ . ˙. ˙. ˙ ˙ ˙. ˙ œ œ 48

& 46 œ œ w 74 ˙ . ˙. œ ˙ ˙ ˙. œ œ ˙ ˙. 48

& 46 œ œ ˙ œ œ 74 7=1+2+1+2+1
œ ˙ œ ˙ œ ˙ œ œ œ ˙ œ œ ˙. œ œ 48
6=1+1+2+1+1 7=2+1+1+1+2 7=1+1+3+1+1

& 46 œ ˙ œ œ œ 74 ˙ œ ˙ œ œ œ œ œ ˙ ˙ œ ˙. œ œ œ 48

& 46 ˙ œ œ œ œ 74 œ ˙ œ œ ˙ œ œ ˙ ˙ œ ˙. œ œ œ œ 48

& 46 œ œ œ ˙ œ 74 ˙ œ œ ˙ œ œ ˙ ˙ œ œ œ œ œ œ ˙. 48
6 7 œ œ ˙ œ œ œ ˙. 8
&4 œ œ œ œ ˙ 4 œ ˙ ˙ ˙ œ œ œ œ 4

& 46 œ œ œ œ œ œ 74 œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ 48

Figure 13.4: Fractional division of rhythm families. Determinants n = 2, 3, 4, 5, 6, 7. The

parent generation binomial combinations r0 = a + b are shown in the top staff. Lower staves
illustrate the results of the iterative permutation-synchronization process.

140 Copy issued to Seth Dempsey ([email protected])

 13.2. EVOLUTION ON THE FACTORIAL LEVEL

h i h i
8 8
8 8 Determinant

b b |5+3 b b|7+1 r0 Binomial a + b

b b |3+5 bb |1+7 Permutation
↓ Synchronize
b b b |3+2+3 bb b|1+6+1 r1 Trinomial
b b b |2+3+3 b bb|6+1+1 Permutations
..
b b b |3+3+2 bbb |1+1+6 .
↓ Synchronize
b bb bb |2+1+2+1+2 bbb b b | 1 + 1 + 4 + 1 + 1 r2 Quintinomial
bb bb b |1+2+1+2+2 bb b b b | 1 + 4 + 1 + 1 + 1 Permutations
..
b bb b b|2+1+2+2+1 b bbbb|4+1+1+1+1 .
..
bb b bb |1+2+2+1+2 bbbbb |1+1+1+1+4 .
..
b b bb b|2+2+1+2+1 bbbb b|1+1+1+4+1 .
↓ Synchronize
b b b b b b b b | 1 + ... + 1 b b b b b b b b | 1 + ... + 1 Uniform distribution
n=8 n=8

h i h i h i
9 9 9
9 9 9

b b |5+4 b b |7+2 b b|8+1

b b |4+5 b b |2+7 bb |1+8

b bb |4+1+4 b b b |2+5+2 bb b|1+7+1

bb b |1+4+4 b b b |5+2+2 b bb|7+1+1
b b b|4+4+1 b b b |2+2+5 bbb |1+1+7

bb bb b|1+3+1+3+1 b b bb b |2+2+1+2+2 bbb bb|1+1+5+1+1

b bb bb|3+1+3+1+1 b bb b b |2+1+2+2+2 bb bbb|1+5+1+1+1
bb bbb |1+3+1+1+3 bb b b b |1+2+2+2+2 bb bbb|5+1+1+1+1
b bbb b|3+1+1+3+1 b b b b b|2+2+2+2+1 bbb bb |1+1+1+1+5
bbb bb |1+1+3+1+3 b b b bb |2+2+2+1+2 bbbb b|1+1+1+5+1

b b b b b b b b b | 1 + ... + 1 b b b b b b b b b | 1 + ... + 1 b b b b b b b b b | 1 + ... + 1

n=9 n=9 n=9

Figure 13.5: Fractional evolution of rhythm families. Determinants n = 8, 9

Copy issued to Seth Dempsey ([email protected]) 141

CHAPTER 13. EVOLUTION OF RHYTHM STYLES

& 48 w œ ˙. w ˙. œ 49 ˙ . ˙ w w ˙. ˙ w w œ
8=5+3 8=7+1 9=5+4 9=7+2 9=8+1

8 9
& 4 ˙. œ w œ ˙. w 4 w ˙ ˙. ˙ ˙. w œw w

& 48 ˙ . ˙ ˙. œ ˙. ˙. œ 49 w œw ˙ ˙. ˙ ˙ œ ˙. w œ
8=3+2+3 8=1+6+1 9=4+1+4 9=2+5+2 9=1+7+1

8 ˙ œ œ 49 œ w
& 4 ˙ ˙. ˙. w w ˙. ˙ ˙ ˙ w ˙. œœ

& 48 ˙ . ˙. ˙ œœ˙ w 49 w w œ ˙ ˙ ˙ ˙. œ œ ˙. w

8 8=2+1+2+1+2 œ œ 49 œ ˙ . œ ˙ . œ œ ˙ ˙.
&4 ˙ œ˙ œ˙ œœw œ ˙ ˙ œ˙ ˙ œœ
8=1+1+4+1+1 9=1+3+1+3+1 9=2+2+1+2+2 9=1+1+5+1+1

& 48 œ ˙ œ ˙ ˙ œw œ œ œ 49 ˙ . œ ˙. œœ ˙ œ˙ ˙ ˙ œ ˙. ˙ œœœ

8 œ œ œ œ 49 œ ˙ . œ œ ˙.
&4 ˙ œ˙ ˙ œ w œ˙ ˙ ˙ ˙ w œœœœœ

& 48 œ ˙ ˙ œ˙ œœœœw 49 ˙ . œ œ ˙. œ ˙ ˙ ˙ ˙ œ œ œ œ œ ˙ ˙.

& 48 ˙ ˙ œ˙ œ œœœw œ 49 œ œ ˙ . œ ˙. ˙ ˙ ˙ œ˙ œ œ œ ˙. ˙ œ

& 48 œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ 49 œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ œ

Figure 13.6: Fractional division of rhythm families. Determinants n = 8, 9

142 Copy issued to Seth Dempsey ([email protected])

 13.2. EVOLUTION ON THE FACTORIAL LEVEL

(2 + 1)2 (3 + 1)2

b b b b|4+2+2+1 b b b b|9+3+3+1
b b bb |4+2+1+2 b b bb |9+3+1+3
b bb b |4+1+2+2 b bb b |9+1+3+3
b b b b|2+4+2+1 b b b b|3+9+3+1
.. ..
. 12 permutations . 12 permutations
bb b b |1+2+2+4 bb b b |1+3+3+9
↓ synchronize ↓ synchronize
b b b b b b b b b | 1 + ... + 1 b b b b b b b b b b b b b b b b | 1 + ... + 1
n=3 n=4

(3 + 2)2

b b b b |9+6+6+4
b b b b |9+6+4+6
b b b b |9+4+6+6
b b b b |6+9+6+4
..
. 12 permutations
b b b b |4+6+6+9
↓ synchronize
b b b bb bb bb b b |4+2+3+1+2+1+2+1+3+2+4
..
. 69300 permutations
b b b b bb bb bb b |4+4+2+3+1+2+1+2+1+3+2
↓ synchronize
b b b b b b b b b b b b b b b b b b b b b b b b b | 1 + ... + 1
n=5

Figure 13.7: Factorial evolution of rhythm families. Power series: square, applied to a bino-
mial a + b. Three cases are shown for determinants n = 2, 3.

Copy issued to Seth Dempsey ([email protected]) 143

CHAPTER 13. EVOLUTION OF RHYTHM STYLES

h i
9
9 meters. This aspect of musical style analysis was already discussed in Section 12.1. He
provides an analysish of
i swing music as evidence of the emergence of triple-division, albeit
4
in a hybrid form in 4 , where one measure contains 4 × ( 82 + 18 ) (swing) or 4 × ( 18 + 18 + 81 )
(shuffle) subgroups at the fractional level.

144 Copy issued to Seth Dempsey ([email protected])

Chapter 14

Rhythms of variable velocities

This chapter is about rhythmical patterns with an accelerando or ritardando character. The
use of acceleration series is the mechanism for creating the feeling of speeding up or slowing
down. Schillinger calls these variable velocity rhythms, which is a confusing term in our
MIDI era where note velocity refers to the speed at which a key is depressed on a MIDI
input controller.
Essentially, acceleration series are affecting the duration of a sequence of attacks over
a continuous, steady beat. Gradually increasing the note duration creates more and more
widely spread attacks, with a feeling of slowing down. Reversing such a series creates an
accelerando. Schillinger illustrates a number of variable time musical phenomena related to
these acceleration series or the variable ratios between longer and shorter notes.

14.1 Acceleration series

Acceleration series are the mathematical equivalent of a natural growth process. Rhythms
created with techniques from the previous chapters, such as interference (see Chapter 2) or
fractioning of two generators (see Chapter 4) yield rhythms with a limited set of durations
in ordered sequences. Depending on the determinant n they yield attack patterns with du-
rations of 1t, 2t, 3t, . . . time units in a characteristic ordered series, the resultant r (remember
the retrograde symmetry property with longer notes at both ends). Acceleration series, on
the other hand, keep growing in duration; each new term in the series has a longer duration
than its predecessor.
There are various types of mathematical series of integers, that may used to model or
represent a (natural) growth process. What characterizes these series is the speed of growth,
it is the amount of duration growth, i.e., the difference in duration between two consecutive
terms in the series.
A number of acceleration series are listed in Table 14.1. These are all infinite series of
integer numbers ni , where i is the index of the i-th element in the series. The mathematical
formula will sometimes apply from the third element in the series onward, i.e., i > 2. Not
all series start at initial value n0 = 1; see the third and fourth example of the geometrical pro-
gressions. The natural harmonic series contains increasing integer numbers at constant speed,
i.e., ni − ni−1 = 1. The power series have the highest growth rate, a characteristic that we
may recognize from Chapter 12, where distributive powers were used to create variation
with longer rhythmical patterns. The summation series group is also known as Fibonacci series

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 14. RHYTHMS OF VARIABLE VELOCITIES

Table 14.1: Acceleration series. The first column has the name of the series, the second
column shows the mathematical formula for creating the infinite series, and the last column
presents the result in integer numbers. ni is the integer value of the i-th element in the series.

Type ni , i = (1), 2, . . . , ∞ Series element values

Natural harmonic series ni = ni−1 + 1 1, 2, 3, 4, 5, 6, 7, . . .

Arithmical progressions ni = ni−1 + m, m = 2, 3, . . . 1, 3, 5, 7, 9, . . . (m = 2)
1, 4, 7, 10, 13, . . . (m = 3)
ni = ni−1 + (i − 1) 1, 2, 4, 7, 11, 16, 22, . . .
Geometrical progressions ni = m ni−1 , m = 2, 3, . . . 1, 2, 4, 8, 16, . . . (m = 2)
1, 3, 9, 27, 81, . . . (m = 3)
3, 6, 12, 24, 48, . . . (m = 2)
2, 6, 18, 54, 162, . . . (m = 3)
Power series ni = mi , m = 2, 3, . . . 2, 4, 8, 16, 42 . . . (m = 2)
3, 9, 27, 81, . . . (m = 3)
5, 25, 125, 625, . . . (m = 5)
Summation series ni = ni−1 + ni−2 1, 2, 3, 5, 8, 13, 21, . . .
1, 3, 4, 7, 11, 18, 29, . . .
1, 4, 5, 9, 14, 23, 37, . . .
Prime number series 1, 2, 3, 5, 7, 11, 13, 17, 19, 23, . . .

and is a growth series that may be observed in nature; the distance between branches and
leaves on plants, the growth of snail shells, the uncurling of a fern spiral, etc. We already
encountered these natural growth series in Chapter 6, where we looked at three-generator
combinations. The Fibonacci series has another property: for very large values of i, ap-
proaching infinity i → ∞, the ratio of two consecutive numbers in the series becomes the
Golden Section, also known as the golden ratio

ni
 
= 1.618033988 . . . .
ni−1 i→∞

This number is used in architecture and in musical form, where it determines the ratio be-
tween geometrical dimensions and section lengths. This may be an intuitive or deliberate
choice by the composer.
The acceleration series are shown graphically in Fig. 14.1. This representation provides
a better view on the speed of growth of each series; the power series have the strongest
acceleration effect. Vertical lines in these diagrams indicate options for regular grouping by
either 3 or 4 time units into multiple full measures. These series are also shown in musical
notation, see Fig. 14.2. In m. 1–10 the time unit is t = 18 and the first five or six terms in the
series are shown.

146 Copy issued to Seth Dempsey ([email protected])

 14.1. ACCELERATION SERIES

Natural harmonic series r = 1 + 2 + 3 + 4 + ...

bb b b b b b b b b b b b b b b
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

Arithmical progression r = 1 + 3 + 5 + 7 + ...

bb b b b b b b b b b b
r = 1 + 4 + 7 + 10 + . . .
bb b b b b b b b b
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

Geometrical progression r = 1 + 2 + 4 + 8 + ...

bb b b b b b b
r = 1 + 3 + 9 + 27 + . . .
bb b b b b
r = 2 + 6 + 18 + 54 + . . .
bb b b b
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

Power series r = 2 + 4 + 8 + 16 + . . .
bb b b b b b
r = 3 + 9 + 27 + 81 + . . .
b b b b b
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

Summation series r = 1 + 2 + 3 + 5 + ...

bb b b b b b b b
r = 1 + 3 + 4 + 7 + ...
bb b b b b b b b
r = 1 + 4 + 5 + 9 + ...
bb b b b b b b
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

Prime number series r = 1 + 2 + 3 + 5 + 7 + ...

bb b b b b b b b b b
-
0 10 20 30 40 50 60 70 80 90 100 110 120 t

Figure 14.1: Acceleration series. All start at t = 0 and have continuously increasing note
durations. The rate of growth is different for each series. Blue vertical lines indicate regular
grouping options

Copy issued to Seth Dempsey ([email protected]) 147

CHAPTER 14. RHYTHMS OF VARIABLE VELOCITIES

& 44 œJ œ œJ œ œ œ œ œ. œ ˙ œ œ. ˙
1
˙ ˙ Ó
Natural harmonic series: 1+2+3+4+5+...

J J ∑ ∑ ∑ ∑ ∑
4
& 4 œJ œ . ˙ œ œ. ˙ w œœ. ˙ ˙ Ó
Arithmical progression: 1+3+5+7+...

J J ∑ ∑ ∑ ∑ ∑
4
& 4 œJ œ . œJ œ . ˙ ˙ ˙. œ w œ. œ ˙ w ˙ Ó
1+4+7+10+13+...

J ∑ ∑ ∑
4
& 4 œJ œ œJ œ . œJ ˙ œ. œ w ˙ œ. ‰
Geometrical progression: 1+2+4+8+16+...

J ∑ ∑ ∑ ∑ ∑ ∑
4
& 4 œJ œ . ˙ ˙ œ œ. w w w
1+3+9+27+...

J ∑ ∑ ∑ ∑ ∑

& 44 œ ˙ . w w œ ˙. w w w w w w
2+6+18+54+...

4 ˙. ˙. ˙. Œ
&4 œ ˙ œ œ w œ w w w
Power series: 2+4+8+16+32+...
∑ ∑

& 44 ˙ . œ ˙ œ. œ w w w œ Ó.
3+9+27+81+...

J ∑ ∑ ∑ ∑

& 44 œJ œ œJ œ œ œ. œ ˙ œ.œ ˙ w w w ˙ œ‰Œ

Summation series: 1+2+3+5+8+...

J J J ∑ ∑ ∑
4
& 4 œJ œ . ˙ ˙ œ. œ w œ ˙. w ˙ Ó
1+3+4+7+11+...

J ∑ ∑ ∑ ∑
4
& 4 œJ œ . œJ œ . œ ˙ . œ.œ ˙ w œ œ. ˙ w w
1+4+5+9+14+...

J J ∑ ∑ ∑
4 œ. œ ˙ œ ˙. œœ. w
& 4 œJ œ œJ œ œ ˙ œŒÓ
Prime numbers: 1+2+3+5+7+11+...

J J ∑ ∑ ∑ ∑

œ. œ.
11 1+2+3+5+8+13=32
& œJ œ œ œ
J
œ œ ˙
J
œ ˙
J
w

&w ˙ œ œ. ˙ œ œ. œ œ œ œ œ
13+8+5+3+2+1=32

J J J J

& ˙. œ ˙. ˙. 12
15 (1+3+5)(3t+t)
œ w w w w w œ w 8

& 12
8 w. œ. œ œ ˙. w. w. ˙. œ œ œ. w. œ. œ œ ˙.
24 (8+4+2)(2t+t+3t)

J J J

Figure 14.2: The first five or six elements in the growth series are shown in musical notation.

148 Copy issued to Seth Dempsey ([email protected])

 14.2. APPLICATION OF ACCELERATION SERIES

14.2 Application of acceleration series

The use of an acceleration series may be at either a continuous steady beat or in a variable
tempo phrase.

14.2.1 Acceleration in uniform groups

First we consider the case of a monomial, a single-element generator with time unit t. This
leads to a uniform, steady beat attack pattern as we have seen in Section 2.1.1. Based on this
time unit we select an acceleration series (or the reverse) to create a variable velocity effect
at some point in the music.
Note that all the acceleration series consist of integer numbers. The reason for this is
twofold: at the time of writing of the Schillinger’s books the personal computer did not exist
and calculating the numbers had to be doable on paper.1 The other reason is the use of the
series over a continuous steady beat; the time unit remains constant and in musical notation
there is the grouping at the measure level, using a fixed time signature nn .
Choosing the appropriate growth series and the number of elements in the series allows
grouping at full measures. For example, adding the first six terms in the summation h series
i
yields the sum 1 + 2 + 3 + 5 + 8 + 13 = 32, which leads to 8 measures at meter 44 and
time unit t = 41 , a regular phrase length. The natural harmonich series
i
also has this property
4
1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 = 36, which yields 9 measures at 4 time signature and t = 41 ,
h i h i
or 12 measures at 34 or three measures at 12 1
8 and t = 8 . The first four prime numbers
add up to 1 + 2 + 3 + 5 + 7 = 18. The power series also have this grouping property. All
these grouping examples lead to a written-out ritardando. These grouping options into full
measures are indicated in Fig. 14.1 as blue vertical lines.

Application Tip:
In Schillinger’s book the potential of this technique in film music, stage and dance
productions is mentioned. In such productions certain events in a cue or scene
may have to be timed to occur on specific beats at a constant musical tempo,
while at the same time this sequence has to suggest speeding up or slowing
down. Selecting the appropriate acceleration series may offer a solution in these
situations.2

Acceleration and ritardando may also be combined, distributing these contrasting effects
over the parts in a score. The combination can be used for a climax effect.

Example 14.1
Combination of written-out acceleration and deceleration.
The summation series r = t + 2t + 3t + 5t + 8t + 13t with total duration
1
There obviously also exist growth series based on rational, non-integer numbers. An example of these is
the exponential growth pattern in the Risset rhythm. This is a clever rhythm, named after French composer Jean-
Claude Risset, that suggests a continuous acceleration or slowing down over a steady tempo. The effect is used
in Electronic Dance Music (EDM).
2
The alternative is a cue at variable tempo, using a click track. That is great for recording the music, but is
more difficult in a live performance situation. In such cases the acceleration series over a steady tempo makes
the job easier for the conductor and musicians; synchronization of the attacks is simpler to achieve.

Copy issued to Seth Dempsey ([email protected]) 149

CHAPTER 14. RHYTHMS OF VARIABLE VELOCITIES

T = 32t is used to demonstrate the combination of a written-out deceleration

with an acceleration, see Fig. 14.2, m. 11–14. The tempo is constant, the vari-
able speed impression is a consequence of the changing note duration. This
four-measure fragment might conclude a musical phrase and prepare for a
transition. The slowing down in the upper staff might start after a number
of steady shorter notes, preferably stressing the time unit. A series of four
8th notes just before this phrase makes the deceleration effect more noticeable.

14.2.2 Acceleration in non-uniform groups

In non-uniform groups we may apply the acceleration series to the individual elements of
a binomial, trinomial or multinomial. Multiply the separate elements from the multinomial
at + bt + ct + . . . by each term from the series {n1 , n2 , n3 , . . .}, analogous to the distributive
power process discussed in Section 12.2 to find the resultant rhythmical pattern

r = (n1 + n2 + n3 + . . .)(at + bt + ct + . . .)
= n1 (at + bt + ct + . . .) + n2 (at + bt + ct + . . .) + n3 (at + bt + ct + . . .) + . . . .

The result is a multi-timescale repeat of the original pattern with an overall ritardanco effect.

Example 14.2
Acceleration in non-uniform groups.
The starting point is a multinomial at + bt + . . . rhythmical pattern. We select
an acceleration series and apply it to this non-uniform group. We will look at a
number of cases.

• Case 1: binomial at + bt = 3t + t and arithmical progression n = 1, 3, 5. This

yields the resultant

r = (n1 + n2 + n3 )(at + bt) = (1 + 3 + 5)(3t + t)

= 1(3t + t) + 3(3t + t) + 5(3t + t) = (3t + t) + (9t + 3t) + (15t + 5t),

a six-attack ritardando rhythmical

h i pattern with total duration Tt = 9 × 4t =
4
36t. This is suitable for a 4 grouping into nine measures, as shown in
m. 15–23 from Fig. 14.2.
• Case 2: trinomial at + bt + ct = 2t + t + 3t and power series n = 8, 4, 2 for an
accelerando effect. Now the resultant is

r = (n1 + n2 + n3 )(at + bt + ct) = (8 + 4 + 2)(2t + t + 3t)

= (16t + 8t + 24t) + (8t + 4t + 12t) + (4t + 2t + 6t).

h ihas total duration Tt = 14 × 6t = 84t,

This accelerating nine-attack pattern
grouping into seven measures at 12 8 meter, see Fig. 14.2 m. 24–30.

150 Copy issued to Seth Dempsey ([email protected])

 14.2. APPLICATION OF ACCELERATION SERIES

• Case 3: square of a binomial (at + bt)2 = (2t + 3t)2 , with summation series
n = 1, 2, 3, 5. This will require the correct order of multiplications to obtain
the resultant
r0 = (at + bt) = (2t + 3t)
r02 = (at + bt)2 = (2t + 3t)2 = (2 + 3)(2t + 3t)
= (4t + 6t) + (6t + 9t)
r = (n1 + n2 + n3 )r02 = (1 + 2 + 3 + 5)[(4t + 6t) + (6t + 9t)]
= (4t + 6t + 6t + 9t) + (8t + 12t + 12t + 18t) +
(12t + 18t + 18t + 27t) + (20t + 30t + 30t + 45t).
The resultant is a 16-attack group with total duration Tt h=i11 × 52 t = 11 ×
25t = 275t. This might be grouped into 55 measures at 58 time signature,
but the numbers already indicate that this is a very long pattern, where the
perception of a coordinated deceleration might be confused by the high note
duration values involved. This extreme example is not very practical in a
musical application.

14.2.3 Rubato
The rubato is a case where the tempo is changing freely, with local accelerations and decel-
erations, while on the longer timescale there still is a steady tempo. Schillinger identifies
the rubato as a mechanism with a general tendency to deviate in the performance from a
balanced binomial towards a more unbalanced ratio.
This means that in a case of two equal attack-durations one of the pair is elongated by an
amount ∆t, where this difference is compensated in the second attack. For example, when
there is a sequence of two equal 8th notes at time unit t = 18 the rubato implies
r = (t + t) → (1 + ∆)t + (1 − ∆)t or (1 − ∆)t + (1 + ∆)t.
The value of ∆t will differ from pair to pair. A limit situation could be where ∆t = 12 t, which
means that the two equal 8th notes have become a dotted 8th-16th note pair.
A similar phenomenon is observed in the interpretation of swing music which is notated
as (pairs of) equal, steady 8th note patterns. The interpretation leads to an elongation of
the first 8th note of each pair (the onbeat note) with accompanying shorter second note (the
afterbeat note). The elongation ∆t in the swing style interpretation is tempo-dependent,
with in inversely proportional relationship
1
∆t ∝ ,
BP M
where BP M is the tempo in beats-per-minute. At slower tempos, i.e., moderate swing, ∆t →
1 3 1
16 corresponding to the dotted 8th-16th pair r = 4 t + 4 t. This is also known as the bounce
swing pattern. At medium swing tempos this leads to the familiar triplet interpretation
where ∆t = 61 t, r = 23 t + 13 t. At fast, up-tempo swing ∆ → 0, with almost equal 8th notes.
The difference with the rubato is that the swing elongation is constant and applied to every
pair, whereas the rubato is flexible and variable on the local timescale.

Copy issued to Seth Dempsey ([email protected]) 151

CHAPTER 14. RHYTHMS OF VARIABLE VELOCITIES

14.2.4 Fermata (hold)

Schillinger makes a distinction between the full stop fermata and the temporary delay created
by a fermata in the middle of a phrase.

The full stop fermata

The full stop fermata is found at the beginning or end of a phrase, at a moment of climax,
or when there is a change with the entrance of a new theme. The written-out length of the
fermata should be aiming at a more balanced group of measures. This means that it is a
multiple of the note durations in the previous measures, such that a low ratio in fermata to
previous measures is achieved, e.g., 1 : 2. Or the actual phrase length as executed becomes
regular, typically an even number of measures such as 4 or 8.

The temporary delay fermata

This fermata type is written in the middle of a rhythmical phrase. The performance of the
fermata leads to a lengthening of the fermata attack and then will yield lower ratios of longer
to shorter note durations. For example, in a sequence of dotted 8th-16th note pairs with a
3 1
fermata on one of the first notes the ratio is likely to shift in the sense ( 16 + 16 ) → ( 28 + 18 );
hence the ratio has decreased from 3 : 1 → 2 : 1. Similarly, a fermata in the rhythm ( 82 + 18 ) →
( 83 + 28 ) leads to the ratio decrease 2 : 1 → 3 : 2 = 1.5.

152 Copy issued to Seth Dempsey ([email protected])

Chapter 15

Conclusion

We have come to the conclusion of this detailed guide to Schillinger’s Theory of Rhythm. The
previous 14 chapters have adhered to the naming and sectioning of the original book. They
have provided a collection of techniques for creating and modifying rhythmical patterns on
different timescales.
We have seen a number of source mechanisms, using interference or fractioning of either
two or three uniform generators, ‘ticking’ at constant intervals. These mechanisms create
resultants, i.e., attack-duration groups with multiple elements, ranging from non-uniform
binomials to long series of multinomials with variable note duration.
Some chapters have focused on variation techniques. Permutations and splitting were
introduced to create homogeneous variability locally on the small timescale; it implies re-
ordering the elements of an attack-duration group. Grouping by pairs and distributive pow-
ers enabled the creation of longer rhythmical patterns from a given source pattern. Rhythm
creation at longer timescales involved juxtaposing several variants in a continuity.
The application to music was also covered: grouping attack-duration patterns into regu-
lar divisions based on the time signature, i.e., the meter. Thus the phenomenon of recurrence,
the repeat of a rhythmical pattern at the full measure on the downbeat, introduced itself. The
distribution of a rhythm using a specific, ordered attack sequence for a single instrument was
discussed. We also saw the distribution of attacks over multiple parts. The combination of
several variants in parallel generates a simultaneity, a multi-part score. As an obvious re-
minder of the potential of the techniques there were demonstrations of the application of
resultants to musical attributes suchs as accents, rests and split-unit groups.
Some chapters considered the evolution of rhythm families on the subdivision, fractional
level (within the measure) and on the longer timescale, factorial level (multiple measures).
Schillinger provides style analysis and made predictions about the future prominence of
certain rhythm families and meters.
This guide used simple mathematics with integer numbers, graphical representations
and musical notation to illustrate the techniques and the various rhythmical aspects. Quite
a number of examples were created, that do not appear in the original Schillinger book.
Application tips, comments and suggestions were interspersed in the text.
Studying these techniques will create a toolbox for the composer and arranger. Mastering
the content of his guide to the theory of rhythm will undoubtedly add to the craftmanship
and skills of the creative musician. There is no need to limit yourself to short repetitive
rhythmical loops, because you did not know how to deviate from the trodden path and

2006-2015
c F.G.J. Absil, http://www.fransabsil.nl
CHAPTER 15. CONCLUSION

find interesting, yet coherent and homogeneous alternatives. Play and experiment with the
techniques in moments of lacking inspiration. Most likely, some great idea will pop-up
while doing rhythmical sketches. The result might be someting that is or is not covered in
this book; finding a personal style and be creative is the end goal anyway. May this book
help you on your way.

154 Copy issued to Seth Dempsey ([email protected])

Bibliography

[1] Grosvenor Cooper and Leonard B. Meyer. The Rhythmic Structure of Music. The Univer-
sity of Chicago Press, Chicago, MI, 1960. ISBN 0-226-11522-4. ix + 212 pp.

[2] Fred Lerdahl and Ray Jackendoff. A Generative Theory of Tonal Music. The MIT Press
Series on Cognitive Theory and Mental Representation. The MIT Press, Cambridge, MA,
1983. ISBN 0-262-62049-9. xiv + 368 pp.

[3] Joseph Schillinger. The Schillinger System of Musical Composition, volume I and II of Da
Capo Press Music Reprint Series. Da Capo Press, New York, fourth edition, 1946. ISBN
0-306-77521-2 and 0-306-77522-0. xxiii + 1640 pp.

[4] Joseph Schillinger. Encyclopedia of Rhythms: Instrumental Forms of Harmony. Da Capo

Press Music Reprint Series. Da Capo Press, New York, 1976. ISBN 0-306-70782-9. xxv +
250 pp.

155
Index

accelerando, 145 density, 91, 94

acceleration series, 145 determinant, 113, 135
accented note, 10, 84, 95, 109 development, 100
afterbeat, 56, 69, 77, 90, 125 diminution, 119
Arabic music, 35 distributive powers, 113
arpeggio, 22, 35, 53, 73, 87, 133 duration, 6, 7
Ars Nova, 119
attack, 2, 5, 6 EDM, 7, 149
attack group, see attack series evolution, 135
attack series, 9, 29, 51 evolution
augmentation, 119, 130 factorial, 135
fractional, 135
balance, 17
Balkan music, 35 fandango, 94
bar, see measure fermata, 152
binary synchronization Fibonacci series, 45, 145
non-uniform, 11, 18, 108, 109 film music, 104, 149
uniform, 9 final duration group, 70
binomial, 84, 135 folk music, 35
Blues chorus, 114 foxtrot, 55
bounce, 151 fractioning, 18, 29, 109, 119
Funk-Rock music, 77
cascara, 94
chord, 22 game music, 104
chorus, 113 game music
climax, 149 interactive, 11
combinations, 98 generalization, 95
common denominator, 22 generator, 2, 11
common product, 18 generator
complementary factor, 45 major, 9, 29, 135
continuity, 67, 79, 83, 107, 114, 118, 129 minor, 9, 29, 135
continuity geometrical progressions., 145
factorial, 113 Golden Section, 146
fractional, 113 group pair
counterpoint, 84, 90, 94 balancing, 39
countertheme, 114 contracting, 42
cross-rhythm, 11, 22, 63, 87 expanding, 41
cube, 127 grouping, 4, 9, 18, 25, 29, 37, 45, 48, 70
grouping
dance, 149 alien measure, 18

156
 INDEX

groups in general, 85 rests, 84, 95, 109

resultant, 2, 9, 11, 29, 45
harmonic structure, 22 retrograde, 21, 95, 98, 104
homogeneity, 43, 79, 84, 99 rhythm family, 45, 135
riff, 53, 56
instrumental form, 4, 51
Risset rhythm, 149
instrumental rhythm, 51, 69
ritardando, 145
integer numbers, 6
Rock music, 93
interference, 2, 9, 29, 45
rubato, 151
isorhythm, 119
rumba, 61, 125
Latin music, 93, 94, 125
self-scaling property, 85, 104, 115
measure, 4, 18, 37 shuffle, 114
Medieval music, 119 simultaneity, 67, 79, 83, 107, 114, 119, 130
melody, 22 solo cadenza, 100
meter, 4, 18, 37 split-unit group, 85, 95
meter splitting, 107
irregular, 21, 55 square, 115
metronome, 5, 9, 45 summation series, 45, 145
minimal music, 94 superimposition, 18, 20, 25, 37
multinomial, 115, 135 swing music, 114, 144, 151
symmetry, 21
natural harmonic series, 145 synchronization, 4, 29, 51, 119, 135
natural nucleus, 22, 63 synchronization
binary, 9, 25
offbeat, 69 syncopation, 25, 37, 56, 77, 87, 125
oom-pah, 53
ordered set, 51, 67, 69, 79 time signature, 4, 18, 25, 37, 70
ostinato, 22, 35, 63, 91 time unit, 2
trinomial, 84, 135
pedal point, 22
periodicity, 2, 6 union, 9, 29
periodicity
variable density, 93
monomial, 6
variation, 43, 79, 99
uniform, 6
vector, 6
permutation, 61, 79, 135
permutations waltz, 4, 7, 20, 96
circular, 79
general, 79
higher order, 95
polka, 55
popular music, 125
power series, 145

quintinomial, 138

recurrence, 4, 21, 51, 55, 67

Renaissance music, 94

Copy issued to Seth Dempsey ([email protected]) 157