100% found this document useful (1 vote)

236 views

Classification of Musical Objects For Analysis and Composition

Classical music

Uploaded by

Akram Mohamed

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (1 vote)

236 views

Classification of Musical Objects For Analysis and Composition

Classical music

Uploaded by

Akram Mohamed

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 156

Computational Music Science

Linshujie Zheng
Guerino Mazzola

Classification
of Musical Objects
for Analysis and
Composition
Computational Music Science

Series Editors
Guerino Mazzola, School of Music, University of Minnesota, Minneapolis,
MN, USA
Moreno Andreatta, Music Representation Team, IRCAM - CNRS, Paris, France

Advisory Editors
Emmanuel Amiot, Laboratoire de Mathématiques et Physique, Université
de Perpignan Via Domitia, Perpignan, France
Christina Anagnostopoulou, Department of Music Studies, National and
Kapodistrian University of Athens, Athens, Greece
Yves André, CNRS, Sorbonne Université - Université de Paris, Paris, France
Gerard Assayag, Music Representation Team, IRCAM - CNRS, Paris, France
Elaine Chew, King's College London, London, UK
Johanna Devaney, Brooklyn College, Brooklyn, NY, USA
Andrée C. Ehresmann, Faculte des Sciences Mathematiques, Université de Picardie
Jules Verne, Amiens, France
Thomas M. Fiore, Department of Mathematics and Statistics, University of
Michigan–Dearborn, Dearborn, MI, USA
Harald Fripertinger, Institut für Mathematik und Wissenschaftliches Rechnen,
Karl-Franzens-Universität, NAWI-Graz, Graz, Austria
Emilio Lluis-Puebla, Faculdad de Ciencias, Universidad Nacional Autónoma
de México, México City, Mexico
Mariana Montiel, Department of Mathematics and Statistics, Georgia State
University, Atlanta, GA, USA
Thomas Noll, Escuela Superior de Música de Cataluña (ESMUC), Barcelona, Spain
John Rahn, School of Music, Music Bldg, University of Washington, Seattle,
WA, USA
Anja Volk, Department of Information and Computing Sciences, Urtrecht
University, Utrecht, The Netherlands
About this series - The CMS series covers all topics dealing with essential usage of
mathematics for the formal conceptualization, modeling, theory, computation, and
technology in music. The series publishes peer-reviewed only works.

Comprehensiveness - The series comprises symbolic, physical, and psychological

reality, including areas such as mathematical music theory, musical acoustics,
performance theory, sound engineering, music information retrieval, AI in music,
programming, soft- and hardware for musical analysis, composition, performance,
and gesture. The CMS series also includes mathematically oriented or computational
aspects of music semiotics, philosophy, and psychology.

Quality - All volumes in the CMS series are published according to rigorous peer
review, based on the editors' preview and selection and adequate refereeing by
independent experts.

Collaboration - The editors of this series act in strong collaboration with the Society
for Mathematics and Computation in Music and other professional societies and
institutions.

Should an author wish to submit a manuscript, please note that this can be done by
directly contacting the series Editorial Board, which is in charge of the peer-review
process.

THE SERIES IS INDEXED IN SCOPUS

Linshujie Zheng • Guerino Mazzola

Classification of Musical
Objects for Analysis
and Composition
Linshujie Zheng Guerino Mazzola
School of Music School of Music
University of Minnesota University of Minnesota
Minneapolis, MN, USA Minneapolis, MN, USA

ISSN 1868-0305 ISSN 1868-0313 (electronic)

Computational Music Science
ISBN 978-3-031-30182-7 ISBN 978-3-031-30183-4 (eBook)
https://doi.org/10.1007/978-3-031-30183-4

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland
AG 2023
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Dedicated to Sina Asiedu Mazzola

Photo and  by Ida Mazzola

Preface

This book deals with the classification of musical objects. This means that we
thematize the project of exhibiting all “types” of such objects up to equivalence
following structural similarity. Of course, it is critical to make precise which
musical objects we are addressing. Nobody can at the time being pretend to
include all possible objects. For example, our attention does not include psy-
chological aspects.
Also are we focusing on objects that can be described in frameworks of
mathematical conceptualization. More precisely, we shall deal with two con-
ceptual approaches: compositions and gestures. Both concepts were introduced
in Mazzola’s work, also in view of the software implementation of musical ob-
jects. Local and global compositions are essentially implemented in Mazzola’s
prestor and RUBATOr software.
The gestural classification topic is new here, and it deals with two topics:
To begin with, local and global gestures are introduced in a quite parallel
procedure to the structures of compositions. Local and global compositions
have been classified in a precise sense, namely by exhibiting algebraic spaces,
so-called schemes, a.k. as algebraic varieties in a more traditional language.
Classes of local and global compositions can be represented as points in such
schemes. But the remarkable difference between compositions and gestures is
that the classification of gestures is far from settled. This is due to the purely
mathematical framework of gestures. Their topological aspects cannot—at the
time being—be classified yet. We describe these problems and will establish as
much as possible—from our perspective—to specify what is missing.
The classification topic will in both cases, compositions and gestures, be
introduced in view of the musical signification of this endeavor. Classification is
in fact essential for the creative work of composers, most of the classical musical
techniques are intimately related to classification, be it in harmony, counter-
point, and motivic work. We shall present these correlations in a discussion of
musical creativity.
The prerequisites we ask the reader to comply with are introductory
courses in algebra [2], category theory [22], and topology [21].

vii
viii Preface

We are pleased to acknowledge the support from Springer’s science editor

Francesca Bonadei for writing such a demanding treatise.

Minneapolis, December 2022 Guerino Mazzola and Linshujie Zheng

Contents

Part I Initial Orientation

1 The Basic Problem of Classification . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.1 Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 The Overall Challenge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Part II General Formal Concepts

2 Ontology, Oniontology, and Creativity . . . . . . . . . . . . . . . . . . . . . . 7

2.1 Ontology and Oniontology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Ontology: Where, Why, and How . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Oniontology: Facts, Processes, and Gestures . . . . . . . . . . . . . . . . . 8
2.4 A Short Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5 Oniontology for Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.6 Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3 Formal Representation of Musical Structures . . . . . . . . . . . . . . . 13

3.1 Scientific vs. Creative Value of Classification . . . . . . . . . . . . . . . . . 13
3.2 Software-oriented Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4 Denotators over General Categories . . . . . . . . . . . . . . . . . . . . . . . . 15

4.1 Formal Definition of Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.2 Formal Definition of Denotators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.3 The Category of Denotators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.4 First Examples of Denotators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.4.1 First Examples of Denotators . . . . . . . . . . . . . . . . . . . . . . . . 19
4.4.2 Hyperdenotators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.4.3 Gestures over Topological Categories . . . . . . . . . . . . . . . . . . 25

ix
x Contents

5 Composition Denotators and Classification . . . . . . . . . . . . . . . . . 27

5.1 Some Classical Composition Denotators . . . . . . . . . . . . . . . . . . . . . 27
5.2 The Role of Classification in Tonal Modulation Theory . . . . . . . . 28
5.3 Modulations in Beethoven’s op. 106 . . . . . . . . . . . . . . . . . . . . . . . . . 30

6 Gestural Denotators: A First Overview . . . . . . . . . . . . . . . . . . . . . 33

6.1 The Symmetric Concept Architecture between Compositions
and Gestures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6.2 The Role of Yoneda’s Lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.3 Software Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

7 The Escher Theorem for Compositions and Gestures . . . . . . . 37

7.1 The Escher Theorem and Escher Categories . . . . . . . . . . . . . . . . . . 37
7.1.1 First Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Part III Local Classification

8 Local Composition and Gesture Classification . . . . . . . . . . . . . . 41

8.1 Classification of Local Compositions . . . . . . . . . . . . . . . . . . . . . . . . 41
8.1.1 Compositions as Sequences of Module Elements . . . . . . . . 42
8.1.2 Eliminating Diagonal Elements . . . . . . . . . . . . . . . . . . . . . . . 42
8.1.3 Transforming Sequences to Linear Maps . . . . . . . . . . . . . . . 43
8.1.4 The Grassmann Scheme for Local Compositions . . . . . . . . 44
8.2 Classification Problems for Local Gestures . . . . . . . . . . . . . . . . . . . 45

9 Classification of Chords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

10 Motif Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
10.1 Three Element Motives in Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 49
10.1.1 An Example from Classical Music . . . . . . . . . . . . . . . . . . . . 50
10.2 Three Element Motives and a Jazz Example . . . . . . . . . . . . . . . . . 51
10.3 A General Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
10.3.1 A RUBETTEr for Generic Melodies . . . . . . . . . . . . . . . . . . 53
10.4 Calculation of the Classes of n-element Motives
in Z212 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

11 Third Chain Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

12 Harmony through Third Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

13 Counterpoint Worlds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

14 Strong Interval Dichotomies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Contents xi

15 Microtonal Contrapuntal Theories . . . . . . . . . . . . . . . . . . . . . . . . . . 65

15.1 The Category of Strong Dichotomies . . . . . . . . . . . . . . . . . . . . . . . . 65
15.2 Towers of Strong Dichotomies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

16 Dodecaphonic Rows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Part IV Global Classification

17 Global Composition and Gesture Classification . . . . . . . . . . . . . 71

17.1 Why Global Compositions and Gestures . . . . . . . . . . . . . . . . . . . . . 71
17.1.1 Global Compositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
17.1.2 The Nerve of a Global Composition or Gesture . . . . . . . . . 81
17.2 Classification through Modules of Affine Functions . . . . . . . . . . . . 81
17.2.1 Local Compositions and Functions . . . . . . . . . . . . . . . . . . . . 81
17.2.2 Morphisms and Function Spaces . . . . . . . . . . . . . . . . . . . . . . 82
17.2.3 Global Standard Structures and Resolutions for
Compositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

18 The Classification Theorem for Global Compositions . . . . . . . 85

18.1 Preliminary Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
18.1.1 Characterization of Interpretations . . . . . . . . . . . . . . . . . . . 85
18.1.2 An Example of a Non-interpretable Global Composition 85
18.2 Non-Interpretable Compositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
18.2.1 Global Dodecaphonic Classes . . . . . . . . . . . . . . . . . . . . . . . . 91
18.3 Classification of Global Compositions . . . . . . . . . . . . . . . . . . . . . . . 91
18.3.1 The Resolution of a Global Composition . . . . . . . . . . . . . . 91

19 The Classification Problem of Global Gestures . . . . . . . . . . . . . . 95

19.1 Local Gesture and Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
19.1.1 Restriction of Gesture Types . . . . . . . . . . . . . . . . . . . . . . . . . 96
19.1.2 Global Standard Structures and Resolutions for Gestures 97
19.2 A Conjectured Classification Theorem . . . . . . . . . . . . . . . . . . . . . . . 98
19.2.1 A Hypergesture for Human Bodies . . . . . . . . . . . . . . . . . . . . 98

20 Singular Homology of Hypergestures . . . . . . . . . . . . . . . . . . . . . . . 103

20.1 Homology via Hypergestures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
20.2 Homology and Counterpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

21 Local Gestures, Structures of Knots, and Local Gestures

as Local Compositions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
21.1 Local Gestures and Knot Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
21.2 Local Gestures as Local Compositions . . . . . . . . . . . . . . . . . . . . . . . 108
21.2.1 Characterstic Differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
21.2.2 Conclusion and Future Topics . . . . . . . . . . . . . . . . . . . . . . . . 111
xii Contents

Part V Classification and Creativity

22 Gestural Interpretation and Future Developments . . . . . . . . . . 115

22.1 Gestural Modulation and Creativity . . . . . . . . . . . . . . . . . . . . . . . . 115
22.2 Classification Problems for Performance Gestures . . . . . . . . . . . . . 117
22.3 Mirror Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
22.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
22.3.2 The Logic of Spatial Mirrors . . . . . . . . . . . . . . . . . . . . . . . . . 119
22.3.3 Temporal Specification of Chords . . . . . . . . . . . . . . . . . . . . . 119
22.3.4 Construction of First Species Counterpoint . . . . . . . . . . . . 120
22.3.5 Lead Sheet in Jazz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
22.3.6 Dance Gestures Associated with Musical Structures . . . . . 121
22.3.7 Conclusion and Future Topics . . . . . . . . . . . . . . . . . . . . . . . . 121
22.4 Perspectives of Future Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

Part VI References, Index

23 Classification Lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

23.1 List of Local Denotators in Z12 , Chord Classes . . . . . . . . . . . . . . . 127
23.2 Third Chain Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
23.3 List of Local Denotators of Cardinality Two and Three in Z212 . . 139
23.3.1 Two Tone Motifs in Z212 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
23.3.2 Two Tone Motifs in Z5 × Z12 . . . . . . . . . . . . . . . . . . . . . . . . 139
23.3.3 Three Tone Motifs in Z212 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
23.4 List of Local Denotators of Cardinality Four in Z212 . . . . . . . . . . . 142
23.5 List of Modulation Chords (Pivots) . . . . . . . . . . . . . . . . . . . . . . . . . 150

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Part I

Initial Orientation
1
The Basic Problem of Classification

Summary. This initial chapter deals with the very definition of classification.
Such a definition must specify the structures to be dealt with and also the
relationships among these structures.
–Σ–
Classification of a determined domain of object means that one should
specify the structures, together with relationships among such structures. It
is evident that not any type of musical object can at present be classified, we
will have to determine the object types that have the chance to be controlled
by precise conceptualization. In our approach in this book, we shall exclude
a significant class of objects that escape a precision as needed by our formal
methodologies. In the next chapter, we shall expose the ontological limits of
our approach.

1.1 Structures
The type of structures we can deal with will be defined by a mathematical
description, which was successful in mathematical music theory and its software
reification. This description is given by what we call denotators, which are
objects or points in spaces that are called forms.
This approach is a restatement of the classical conceptualization in object-
oriented programming, as given by classes and their instantiations, called ob-
jects. Denotators and forms are canonically restated in terms of objects in
specific mathematical categories.
The remarkable advantage of our approach will not only include struc-
tures, such as local compositions, e.g., chords or rhythms, structures that were
dealt with our book [27]. It will also include local and global gestures, a new
type of structures that were not present there. The significant point of this
extension is that the conceptual approach of denotators and forms allows for
a natural transfer to global gestures, by a covering local gestures without any

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 3

L. Zheng, G. Mazzola, Classification of Musical Objects for Analysis and Composition,
Computational Music Science, https://doi.org/10.1007/978-3-031-30183-4_1
4 1 The Basic Problem of Classification

major structural restatement. Gestures are a canonical extension of classical

denotators and forms.

1.2 Relationships
The conceptual setup of denotators enables a natural conceptualization of re-
lationships among denotators. Classification deals with the understanding of
how denotators can be compared to each other. This feature is a consequence
of the fact that denotators can be viewed as objects in determined categories.
Whenever one works in a category, its objects admit morphisms between any
two of its objects.
Classification of objects within a given category then means to determine
isomorphism classes in such a category. This means that one determines classes
of objects, which are isomorphic to each other. Isomorphism classes define a
partition of the object class of the category into classes of isomorphic objects.
In mathematics, this task can be an easy task, but it can also present unsolvable
problems. Many important categories have not been understood with regard to
their classification objective.
For example, in the category of vector spaces over a determined field, the
dimension of a vector space is the defining invariant of a vector space, i.e., two
vector spaces are isomorphic iff their dimensions coincides. The classification
of finite simples groups (groups without non-trivial normal subgroups) is also
achieved, but this work only results from a huge collaboration of leading group
theorists [50]. But, for example, the category of knots (closed continuous curves
in topological spaces) has not been classified yet [16].
In our context, there are many subcategories of denotators, whose classifi-
cation has not yet been achieved. Nevertheless, some important subcategories,
namely zero-addressed local composition were classified. But local gesture de-
notators are not classified at the present.

1.3 The Overall Challenge

Whatever might be the status of classification of denotators, the semantics
of classification is not only a mathematical objective, but plays a crucial role
for the creative challenge of realizing musical works. We shall learn that most
systematic approaches to musical composition, analysis, and performance are
intimately connected to classification objectives. For example, tonal modulation
presupposes the classification of tonalities, and counterpoint is built upon the
classification of sets of consonant and dissonant intervals, but also more modern
approaches, such as dodecaphonic and serial composition methods, are derived
from classification of melodic units, called rows in this case.
We shall claim and prove that classification is a driving force backing
musical creativity.
Part II

General Formal Concepts

2
Ontology, Oniontology, and Creativity

Summary. This short chapter introduces first the global architecture of ontol-
ogy of music, which this book is going to use extensively as an initial classifying
perspective.
–Σ–

2.1 Ontology and Oniontology

This section is about ontology of music, including three dimensions: realities,
semiotics, and communication. It also includes the extension of ontology to
the fourth dimension of embodiment. We call this extension “oniontology” for
reasons that will become evident soon.

2.2 Ontology: Where, Why, and How

Ontology is the science of being. We are therefore discussing the ways of being
that are shared by music. As shown in Figure 2.1, we view musical being as
spanned by three ‘dimensions’, i.e., fundamental ways of being. The first one
is the dimension of realities. Music has a threefold articulated reality: physics,
psychology, and mentality. Mentality means that music has a symbolic reality,
which it shares with mathematics. This answers the question of “where” music
exists.
The second dimension, semiotics, specifies that musical being is also one
of meaningful expression. Music is also an expressive entity. This answers the
question of “why” music is so important: it creates meaningful expressions, the
signs that point to contents.
The third dimension, communication, stresses the fact that music exists
also as a shared being between a sender (usually the composer or musician), the
message (typically the composition), and the receiver (the audience). Musical
communication answers the question of “how” music exists.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 7

Fig. 2.1: The three-dimensional cube of musical ontology. Guerino Mazzola

Fig. 2.2: The hypercube of musical oniontology. Guerino Mazzola

2.3 Oniontology: Facts, Processes, and Gestures

Beyond the three dimensions of ontology, we have to be aware that music is

not only a being that is built from facts and finished results. Music is strongly
also processual, creative, and living in the very making of sounds. Musical per-
formance is a typical essence of music that lives, especially in the realm of
improvisation, while being created. The fourth dimension, embodiment, deals
with this aspect; it answers the question “how to come into being?” It is ar-
2.6 Creativity 9

ticulated in three values: facts, processes, and gestures. This fourth dimension
of embodiment gives the cube of the three ontological dimensions a threefold
aspect: ontology of facts, of processes, and of gestures. This four-dimensional
display can be visualized as a threefold imbrication of the ontological cube, and
this, as shown in Figure 2.2, turns out to be a threefold layering, similar to an
onion. This is the reason why we coined this structure “oniontology”—it sounds
funny, but it is an adequate terminology.

2.4 A Short Characterization

The four dimensions of musical oniontology can be put together to present a
short characerization of music:
Music embodies
meaningful communication
and
mediates physically
between its
emotional and symbolic layers.

2.5 Oniontology for Classification

Our classification business will not deal with all of these oniontological dimen-
sions. In particular, the psychological dimension will not be addressed. This
means in particular that, for example, psychological arguments for the selec-
tion of consonances are not dealt with. We shall only address mathematical
constructions of consonance-dissonance concepts. Also, modulations will not be
addresses as psychological processes, but only in their shape of mathematical
models, despite the fact that such models may be motivated by psychological
rationales. And denotators may also represent physical entities, such as Fourier
or FM constructions, but we will not address these genuinely physical phe-
nomena, we shall only deal with their mathematical models using denotators.
Finally, gestures as typical entities of embodiment will only be represented in
their mathematical form.
In other words, the oniontological dimensions will only be addressed
through their formal representation via denotators. Nevertheless, the original
oniontology of denotators will be referred to when we discuss their impact on
the creative implications within musical constructions in composition, analysis,
and performance.

2.6 Creativity
We want to describe here the essentials of our approach to creativity [36] in
practical terms to be used in the book.
10 2 Ontology, Oniontology, and Creativity

Our approach to creativity is not a psychological trick nor does it incite a

state of ecstasy, possibly boosted by drugs and similar devices. It is a process of
discovery and invention that begins with an open question and continues with
a run through a sequence of well-defined operational steps. It might not be
successful in finding a solution to the question, but we can start over and repeat
the processual run until we find a solution. Creativity cannot be guaranteed, but
there are good reasons for approaching it based upon a clear strategy. Again,
we discuss many examples of successful creativity in the theory part and hope
that the reader will accept our approach here as a basis for our practical work.
The process of creativity is composed of the following steps:
1. Exhibiting the open question
2. Identifying the semiotic context
3. Finding the question’s critical sign or concept in the semiotic context
4. Identifying the concept’s walls
5. Opening the walls and displaying its new perspectives
6. Evaluating the extended walls
Let us briefly describe them, together with a
simple example: the invention of 3M’s Post-It (see
Figure 2.3).
Step 1, exhibiting an open question. Creativ-
ity should start with the will to find something
new that will answer an open question. It has a
target.
Example: In 1968, Dr. Spencer Silver, a chemist
Fig. 2.3: The Post-It at Minnesota’s 3M company, developed an adhe-
notes. sive that did not really glue. He did not know what
to do with it. What would be the usage of such a
substance?
Step 2 consists of identifying the context of this question. This is a natural
requirement: We want to know what is the overall position of the question, since
for one individual the question might be open but trivial, while for another it
might be a big challenge. Also, we stress that the context is a semiotic one. What
does this mean? It means that the context should be one of things, thoughts,
and signs that mean something. The open question then would be one that
creates new contents when answered. It would extend the given meaningful
context. So creation in our understanding is about creating new contents, not
just forms.
Example: The context is the chemical industry. Including a variety of more
than 50,000 3M products.
Step 3 is the moment of focusing on a specific location in the context, a
critical concept or thing where we guess that the open question could be made
more precise.
Example: The critical concept here was “adhesive.” Dr. Silver had invented
an adhesive that was critical—it did not glue as required, but only “half of it.”
2.6 Creativity 11

Step 4 is very important. It asks us to identify the concept’s walls. Walls

are a metaphor for properties, characteristics, and specificities of the concept
that circumscribe the concept in a more or less explicit form. It is a delicate
task here, since some properties might be so subtle that one is barely capable
of recognizing them.
Example: The walls of the concept “adhesive” was that one would expect
that an adhesive has to glue, meaning 100 percent not just somewhat. It was a
wall because it was so evident that an adhesive had to glue, by its very definition!
Step 5 asks us to consider these walls and to try to ‘soften’ and to ‘open’
them. This means that we ask to what extent these walls are necessary for the
critical concept, and whether we could possibly find ways to open them and to
recognize new perspectives on the other side of the given walls.
Example: The softening of the “glue” wall meant that one had to put into
question this strong requirement of 100 percent gluing. It was Arthur Fry, who
attended one of Silver’s seminars, who successfully opened that wall. He sang in
a church choir and was frustrated with the paper bookmarks he used to mark the
songs in his hymnal because they would not stay put. Fry realized that Silver’s
adhesive met exactly his needs.
Step 6 is the terminal step. It consists of the evaluation of wall extensions
we may have found in step 5, and then the judgment of this evaluation’s result.
Is it a successful extension of the original critical concept or didn’t we find an
answer to the open question? If we are successful, everything is OK; otherwise,
we have to go back to step 4 and find new walls and new extensions, or even
to step 3 and look for new critical concepts.
Example: In 1980 Post-It Notes were introduced nationwide in the U.S.
and soon became a big success.

Remark 1 It is, of course, easy to give an example of a creative strategy

that very probably would not be successful. Suppose that an open question in
number theory is given. The critical concept could be “number.” And the wall
one could discover could be “the number’s color” when you write it down on
paper, on a blackboard, or on a computer screen. You could have the ‘ingenious’
idea that the number’s color could be extended from the usual black, blue, or
red to any fancy color. But it is quite clear that such an extension would not
solve the original number-theoretical question.

Remark 2 We should not conclude this informal introduction to our creativ-

ity process scheme without pointing out the deep impact of representing ideas
and methods in a sensual way (touch, smell, taste, hearing, sight). We have
already established the sight- and touch-oriented metaphor of a wall described
in steps 4, 5, and 6. When unfolding creative processes, we should always em-
bed our activity in an object-oriented environment, where we feel at home
and comfortable. Humans cannot think in abstract categories without using
sensual metaphors. Even the most abstract mathematical thoughts are always
embodied in objects that can be manipulated by human gestures. This is one of
12 2 Ontology, Oniontology, and Creativity

the secrets of successful mathematicians. Einstein, for example, admitted that

his thoughts were always performed with an intense chewing activity. He was
literally chewing ideas!
Last but not least, we should be aware that probably the most creative
force is love for life. After all, this is the strongest motor of human propagation.
For a successful creation, the atmosphere should be a warm and loving one 1 ;
creation out of hate or indifference is never for life. A creation is a mental baby;
it needs a ‘mother’ and her carrying body.

1
Sidney Lanier says, “Music is love in search of a word.”
3
Formal Representation of Musical Structures

Summary. We discuss what classification (the determination of isomorphism

classes) means for understanding and creating music. We analyze the scientific
versus artistic perspective of classification, in particular its often subconscious
presence while musical creativity is operated. A basis of any serious investiga-
tion of musical objects/structures is the project of their reliable and precise
conceptual description. And also being useful for the software management of
music. Such a representation should include not only the objects, but also their
possible relationships.
–Σ–

3.1 Scientific vs. Creative Value of Classification

The scientific classification using formal representation of musical structures
seems to contradict the creative process when constructing musical works. The
reason is that creativity is frequently viewed as being an activity full of vague-
ness and a fuzzy multitude of possibilities and options. No doubt that such cat-
egories intervene. But the transition from phantasies to musical objects must
be driven by tools that realize those options qua musical objects. To recur to
our definition of creativity, the point is about an efficient and concrete recog-
nition and transgression of the walls of the previous framework. The softening
and opening of walls must comply with a concrete trajectory to learn about
the potential success of such an extension.
This approach must rely upon an optimal representation of musical con-
cepts or objects to be instantiated in the present creative movement. There
is a moment within the creative procedure, where vagueness has to yield to
concrete precision to generate an intended musical extension. For this reason,
a formal representation of critical concepts, together with their classification,
becomes mandatory to reach the extension of the box’s walls.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 13

3.2 Software-oriented Perspectives

It becomes increasingly relevant to apply a technological methodology to classi-
fied structures to reach the intended creative objective. More concretely speak-
ing, creative activity needs to face the software-driven implementation of mu-
sical concepts. In view of what was said in the preceding section, it seems
important to understand the necessity of a formal representation of musical
concepts or objects as a basic style of realistic creativity.
Let us look at a representative example, the famous Structures pour deux
pianos, 1952 and 1961, by Pierre Boulez [5, 6]. The creative work of this com-
poser was mainly focused on formal manipulations, but without technological
support, Boulez had to perform lengthy calculations by hand without a hope
to redo those calculations in reasonable time. We have reconstructed this com-
position, now for twelve instruments, using our RUBATOr software [27], see
also Figure 3.1. This re-creative work could realize Boulez’s idea in seconds (af-
ter having programmed the software tools). This gave us a much more general
perspective upon this composer’s creative methodology of serial composition.

Fig. 3.1: The Boulez network in the Rubato software.

4
Denotators over General Categories

Summary. We recapitulate the successful concept architecture of forms and

denotators, which was also essential in software developments, in particular
for the software RUBATOr . However, with regard to the newer concepts of
musical gestures, we now extend forms and denotators from module categories
to include directed graph topoi as well as topological categories.
–Σ–

4.1 Formal Definition of Forms

Summary. The definition of a form is given, including name, type, coordinator,
and the space functor identifier. In our new approach, the basic category Mod
of modules will be enriched by a general category Cat, especially with regard to
the category Digraph of digraphs that is crucial for the formalism of gestures.
–Σ–
Recall that Cat denotes the topos of contravariant functors F u : Cat →
@

Ens from the category Cat to the category Ens of sets. Recall also that for
functors F u in Cat@ , an object X which is an argument of F u is called an
address of F u, and a morphism f : X → Y in Cat is called an address change.
Recall the notation X@F u for the value set of F u at address X. Recall finally
the subobject classifier Ω in the topos Cat@ [23].

Definition 1 A form F is a quadruple F = (N F, T F, CF, IF ) where

(i) NF is a string of Unicode characters; it is called the name of F and denoted
by N (F ).
(ii) TF is one of the symbols1
1
They turn out to symbolize five types of operators, but in the definition, we just
need the symbols, i.e., the character strings.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 15

1. Simple,
2. Syn,2
3. Power,
4. Limit,
5. Colimit;
it is called the type of F and denoted by T (F ).
(iii) CF is one of the following objects according to the previous symbols:
A. For Simple, CF is an object M of Cat,
B. for Syn, and Power, CF is a form,
C. for Limit and Colimit, CF is a diagram D of forms;
it is called the coordinator of F and denoted by C(F ). The diagram D is a
diagram of functors F un(Fi ), as defined in (iv), for a family (Fi )i of forms.
(iv) IF is a monomorphism of functors IF : F u X in Cat@ , with this data:
1. For Simple, X = @M ,
2. for Syn, X = F un(CF ),
3. for Power, X = Ω F un(CF ) ,
4. for Limit, X = lim(D),
5. for Colimit, X = colim(D);
it is called the identifier of F and denoted by I(F ), whereas its domain F u
is called the space (functor) of F and denoted by F un(F ). The codomain
of the identifier is called the frame space of the form.
To denote a form F , we inherit the notation of the setup established in naive
context [27, Ch. 6] and add the identifier below the arrow:

N ame −→ T ype(Coordinator). (4.1)

Identif ier

A more formulaic representation of a form can be written by

N ame : Identif ier.T ype(Coordinator).

4.2 Formal Definition of Denotators

Summary. Based on the definition of a form (Subsection 4.1), we define de-
notators.
–Σ–

Definition 2 Let M be an address in Cat. A M -addressed denotator is a triple

D = (N D, F D, CD) where
2
In the generic definition of a form, the synonym form type is superfluous. We
however maintain this type for semiotic reasons: synonymy is a proper type of
reference which is meant to be different from any other reference mode.
4.2 Formal Definition of Denotators 17

(i) ND is a string of Unicode characters; it is called the name of D and denoted

by N (D).
(ii) FD is a form; it is called the form of D and denoted by F (D).
(iii) CD is an element of M @F un(F (D)); it is called the coordinates3 of D and
denoted by CT (D).
According to the naive setup, a denotator is denoted by

N ame : Address F orm(Coordinates) (4.2)

So, in the full notation, a denotator D with form F is symbolized as follows:

N (D) : Address N (F ) −→ T ype(Coordinator)(Coordinates), (4.3)

Identif ier

but this clumsy writing is only used if absolutely unavoidable. Again, as already
described for the naive setup, we shall write

myN ame : myAddress@AF orm(myCoordinates) (4.4)

for a denotator of form AF orm and coordinates myCoordinates, or even

myN ame : @AF orm(myCoordinates) (4.5)

if myAddress is also clear.

Naming. The naming policy is identical with the naive situation. Recall that
we can have the empty denotator name.
Address. The address object M is an important generalization, however, we
shall not overstress it since in the everyday language we speak of denotators
independently of their address, as if there were just one ambient space
instead of an entire space functor.
Form. The form of a denotator englobes the whole recursion information as
well as the form’s functor. It is the latter which contains the coordinates.
Coordinates. The coordinates are one “point” 4 or element of the form’s func-
tor at the given address M . Let us look at the shape of coordinates as a
function of the particular form:
1. For Simple, the coordinator is an object N of he basic category Cat,
and the coordinates CT (D) are identified (via the identifier!) with an
element of M @N , i.e. a morphism CT (D) : M → N of Cat objects.
2. For Syn, the coordinates CT (D) identify with an element of

M @F un(C(F (D))).

So the coordinates are described by recursion to the coordinator C(F (D)).

3
This is a plural in singular mode.
4
In algebraic geometry, an element of M @F for a functor F ∈ Mod@ is called an
“M -valued point of F”.
18 4 Denotators over General Categories

3. For Power, the coordinates CT (D) identify with an element of

∼
M @Ω F un(C(F )) → Sub(@M × F un(C(F ))),

i.e. a subfunctor of @M × F un(C(F )).

4. For Limit, suppose that the diagram D of forms has vertex forms Fi
and vertex functors F ui := F un(Fi ) at vertexes i, as well as natural
transformations m(f ) : F utail(f ) → F uhead(f ) for arrows f in D. Recall
([27, Appendix G.2.2.1]) that at a given address M , we have a natural
isomorphism
∼
M @lim(D) →
Y
{x ∈ M @F ui | m(f )(xtail(f ) ) = xhead(f ) , all arrows f of D}
i

so that the coordinates of Limit denotators are special tuples in the

product of all M @F ui . In other words, the limit denotators are canoni-
cally related to the product denotators from the naive setup. Therefore
the general setup does resemble the naive one as being focused on a
general, but fixed address.
5. For Colimit, let us inherit the notation of the preceding situation.
Recall ([27, Appendix G.2.2.1]) that for a given
` address M , we have
the equivalence relation ∼ on the coproduct i M @F ui , generated by
the relation x ∼ y, x ∈ M @F ui and y ∈ M @F uj , iff there is an arrow
f in D with i = tail(f ), j = head(f ), and m(f )(x) = y. We then have
a natural isomorphism
∼
a
M @colim(D) → M @F ui / ∼
i

and recognize that the naive coproduct is just the basic space of the
Colimit denotator space before dividing through the equivalence rela-
tion ∼, and again, selecting a general, but fixed address M . So—up to
general addressing—the coordinates of a Colimit denotator are those
of the coproduct modulo an equivalence relation defined by the dia-
gram’s arrows.

4.3 The Category of Denotators

Denotators can be understood as objects of the category Deno(Cat) of deno-
tators over Cat as follows.
Let

D = DN ame : DAddress@DF orm(DCoordinates),

4.4 First Examples of Denotators 19

E = EN ame : EAddress@EF orm(ECoordinates)

be two denotators over Cat. This includes two elements (the coordinates)

DCoordinates : DAddress → F un(DF orm),

ECoordinates : EAddress → F un(EF orm).

A morphism f : D → E is a pair f = (Add(f ), F un(f )), where Add(f ) :
DAddress → EAddress is a morphism in Cat and F un(f ) : DF orm →
EF orm is a natural transformation (a morphism of functors) such that the
following square diagram commutes:
DCoordinates
@DAddress −−−−−−−−−→ DF orm
 

@Add(f )y

yF un(f )
ECoordinates
@EAddress −−−−−−−−−→ EF orm
To understand this setup, recall that the Yoneda Lemma identifies N @F
for the evaluation of a presheaf F at N with the set N at(@N, F ) = @N @F
of natural transformations from the representable presheaf @N to F. The hor-
izontal arrows in the above square diagram are the Yoneda-associated natural
transformations for the two evaluations DCoordinates, ECoordinates.

4.4 First Examples of Denotators in Theory and Practice

With the definition of morphisms among denotators over Cat we may start
investigating the task of classification in Deno(Cat), i.e., the determination of
classes of isomorphic denotators. Before we investigate this major task, it is
useful to present a variety of denotators, which are significant in theory and
practice.

4.4.1 First Examples of Denotators

We want to present first examples of denotators, starting with the category

Cat = Mod of modules, and then addressing gestural denotators over the
category Digraph of digraphs.

4.4.1.1 Denotators over the Category Mod

The first example is for form

Loudness : Id.Simple(ZST RG).

where Id denotes the identity identifier, with module ZST RG, the Z monoid
algebra over the monoid ST RG of strings (words) over an alphabet ST RG that
20 4 Denotators over General Categories

could be the ASCII alphabet or the extended alphabet Unicode. A denotator

could be
mezzof orte : 0@Loudness(mf )
denoting the mezzoforte symbol mf. The address 0 (zero module) means that
we are given an affine map x : 0 → ZST RG, which defines its image point
x(0) = mf .
For the module Z2 we get a form

HiHat : Id.Simple(Z2 ),

and we may consider the denotator

HihatOpen : 0@HiHat(1)

to denote the open state 1 of a HiHat, whereas the value 0 would denote a
closed HiHat.
For module Z, we get the pitch form

P itch : Id.Simple(Z),

and we may consider the denotator

myP itch : 0@P itch(p)

to denote a pitch of value p.

For module R, we get the onset form

Onset : Id.Simple(R),

and we may consider the denotator

myOnset : 0@Onset(o)

to denote an onset of value o.

With these forms, we may define a simple version of a note by the form

N ote : Id.Limit(Onset, P itch)

where a denotator

myN ote : 0@N ote(thisOnset, thisP itch)

would denote a note with two zero-addressed coordinates thisOset, thisP itch
in the simple forms Onset, P itch, which are a diagram without arrows and two
Forms.
We may use a pause form of limit type over the forms Onset and Duration

Duration : Id.Simple(R).
4.4 First Examples of Denotators 21

We then define
P ause : Id.Limit(Onset, Duration),
and then a form GeneralN ote would be

GeneralN ote : Id.Colimit(N ote, P ause)

such that a general note would be either a note or a pause.

For a score (of a very simple type), we would define

SimpleScore : Id.P ower(GeneralN ote)

whose zero-addressed denotators would be

myScore : 0@SimpleScore{n1 , n2 , . . . nk }

a collection of k notes n1 , n2 , . . . nk . Here Id is the canonical embedding of the

functor 2X in the topos-theoretically given functor Ω X , see [27, Ch. 6.2.3].
Forms may also be defined in a circular way, as illustrated by the following
example, which formalizes a hierarchical score variant, also used implicitly in
Schenker analysis, called M acroScore:

M acroScore : Id.P ower(N ode),

with
N ode : Id.Limit(N ote, M acroScore)
A macroscore denotator is a collection of nodes, which are each defined by an
“anchor note” and a “satellite” collection. Trills are typical denotators of such a
form. The circularity boils down to a finite information when satellites become
empty sets.
To represent a Fourier decomposition of a periodic function of time t,
X
f (t) = A0 + Ai sin(2πif + P hi )
i>0

we define the form

F ourier : Id.Limit(Basis, F requencey, Spectrum)

with the basic value A0

Basis : Id.Simple(R),

the frequency f in
F requency : Id.Simple(R),
and
Spectrum : Id.Limit(Amplitude, P hase, Spectrum),
22 4 Denotators over General Categories

with
Amplitude : Id.Simple(R),
and
P hase : Id.Simple(R).
Here the circularity is strictly infinite as required for the Fourier formula.
One may also represent the Frequency Modulation formula
X
f (t) = Ai sin(2πfi + P hi + M odulatori )
i

by the representation of this sum via

F M Object : Id.P ower(Carrier)

with
Carrier : Id.Limit(Support, F M Object),
where

Support : Id.Limit(Amplitude, F requency, P hase, M odulator).

The form M odulator is synonymous to FMObject:

M odualator : Id.Syn(F M Object).

Because of the powerset in this form setup, finite solutions are feasible, as
usually realized for FM soft- and hardware, typically realized by Yamaha’s
DX7 synthesizer.
Denotators over non-zero addresses are useful to denote important con-
structions in modern compositional methodologies, such as the serial method
that is built upon a sequence of rows in a number of musical dimensions, such
as pitchclass, duration, loudness, and attack. Pierre Boulez’s (in)famous com-
position Structures pour deux pianos (1952 and 1961) is a good example [5].
He considers these four forms

P itchClass : Id.Simple(Z12 )

DurationClass : Id.Simple(Z12 )
LoudnessClass : Id.Simple(Z12 )
AttackClass : Id.Simple(Z12 )
A serial composition refers to the form

SerialComp : Id.Limit(P itchClass, DurationClass, LoudnessClass, AttackClass),

and the denotators in a serial composition are quadruples of denotators in

SerialComp:
4.4 First Examples of Denotators 23

mySerialComp : Z11 @SerialComp(p, d, l, a)

which means that each of the four coordinates is Z11 -addressed. In other words,
each coordinate is given by the images of the twelve affine basis vectors e0 =
0, e1 = (1, 0 . . . 0), . . . e11 = (0, . . . 0, 1).
This approach allows Boulez to consider address changes c : Z11 → Z11 in
order to generate derived rows by address changes, i.e., without any transforma-
tion on the four parameter spaces. Boulez with this approach avoids the prob-
lem of performing transformations on duration, loudness, and attack, which
would make no musical sense! See [20] for a thorough discussion of Boulez’s
serial composition.

4.4.1.2 Denotators over the Category Digraph

We present examples of integral curves of performance, gestures over topologi-

cal categories, sets of gestures, gestures in topological vector spaces, glissandi,
crescendi, tempo curves (fermata etc.)
Simple forms over the category Digraph refer to the coordinators, which
are digraphs. For classical gestures, these digraphs are derived from topological
→
−
spaces X as follows. We consider the digraph X , whose vertices x are the
elements x of X, and whose arrows a are the continuous functions a : I → X,
I = [0, 1] ⊂ R, with the head h(a) = a(1) and tail t(a) = a(0). A simple form
over X is the form
→
−
Xgestures : Id.Simple( X )
with Id being the identity identifier. A ∆-addressed gesture in this form is

aGesture : ∆@Xgestures(g)
→
−
with coordinate g : ∆ → X a digraph morphism.
A concrete example is an integral curve i : I → REHLD of a performance
−−−−−→
field, represented as a gesture g :↑→ REHLD : a 7→ i in the four-dimensional
real vector space REHLD of onset E, pitch H, loudness L, and duration D.
A second example is the triply articulated hand movement of a pianist:
go down to the keyboard, stay there for the duration of a sound, go up away
from the keys. Here we take the representation of the hand by five finger tips
and one carpus, each in the spacetime R4 , adding up to a movement in R24 .
−−→
This defines a gesture g : ∆ → R24 of address ∆ = • → • → •, one curve for
every arrow.
For a non-simple form, take
−−−−−→
CurveSets : Id.P ower(REHLD )

whose denotators

aCurveSet :↑ @CurveSets{c1 , c2 , . . .}
24 4 Denotators over General Categories

denote sets {c1 , c2 , . . .} of integral curves ci : I → REHLD as represented by

−−−−−→
↑-addressed gestures ci :↑→ REHLD . Here, the identifier Id is not the identity,
but the embedding of the subfunctor 2X within the full powerset functor Ω X ,
see [27, Ch. 6.2.3].

4.4.2 Hyperdenotators

This section presents the construction of hyperdenotators for both categories,

Mod and Digraph.

4.4.2.1 Hyperdenotators over the Category Mod

For a module M , the set of M -addressed denotators in a simply form Spl :

Id.Simple(N ) over module N identifies with the set M @N = Hom(M, N ) of
homomorphisms f : M → N (when neglecting their names). But M @N is
again a module, and we may consider denotators at address R in the simple
form SplM : Id.Simple(M @N ). These denotators correspond to elements of
the set R@M @N , which is isomorphic5 to the set (RM )@N . Such denotators
are called hyperdentators since they are denotators whose coordinates are again
denotators, denotators of denotators.
For example, if R = Zd , M = Z11 , N = Z12 , we may interprete deno-
tators in T welveRows : Id.Simple(Z11 @Z12 ) with address Zd as d-tuples of
twelvetone rows, i.e. as hyperdenotators in Z12 .

4.4.2.2 Hyperdenotators over the Category Digraph

→
−
To define hypergestures, take the simple form Gst : Id.Simple( X ) over a topo-
logical space X. For a digraph ∆, the ∆-addressed gestures
→
−
N g : ∆@Gst(g : ∆ → X )
→
−
are elements of the morphism set ∆@ X . But since ∆ is the colimit of its arrows
and vertexes, we have
→
− ∼ →
−
∆@ X → lim ↑ @ X .
∆
→
− ∼
But ↑ @ X → I@Top X, the set of I-addressed curves in X. Giving this set the
→
−
compact-open topology, we may define a topology on ∆@ X by the limit of the
→
− →
−
compact-open topologies on the ↑ @ X for the ∆’s arrows. This turns ∆@ X into
→
−
a topological space, which we denote by ∆ @ X, and we may consider gestures
→
− →−
in Γ @ ∆ @ X, these are gesture of gestures, i.e., hypergestures.
5
denotes the tensor product in the category of modules with affine transforma-
tions.
4.4 First Examples of Denotators 25

4.4.3 Gestures over Topological Categories

In this section, we deal not only with topological spaces for gestures, but more-
over with topological categories. A topological category is a category which
is internal in the category of topological spaces, i.e., the set of morphisms is
a topological space, the domain and codomain maps are continuous, and the
composition of morphisms is continuous. Continuous maps in gesture theory
will be replaced by continuous functors.
We first need to turn the unit interval into a topological category ∇. As
a topological space it is the set ∇ = {(x, y)|x, y ∈ I, x ≤ y}, with the topology
inherited from R × R, and domain d(x, y) = (x, x), codomain c(x, y) = (y, y).
For a topological space X we define the graph category, again denoted by X,
as the category, whose objects are the elements of X, whereas the morphisms
are the elements of X × X, the topology being the product topology. Clearly,
continuous curves f : I → X correspond to continuous functors f : ∇ → X.
The category TopCat of topological categories with continuous functors
has morphism sets for pairs of topological categories K, L, denoted by K@L,
which are subsets of the sets of general functors f : K → L. The functor
Top → TopCat identifying topological spaces and their continuous functions
with corresponding graph categories is fully faithful.
For a topological category X and a digraph Γ , we define the space digraph
→
−
X of X with the arrow set being the topological functors ∇ → X with the
vertices as the restrictions of arrows to the initial and final arguments in ∇.
→
−
The gestures of skeleton Γ are then the set Γ @ X . To turn this latter set into
→
− ∼
a topological category, observe that ↑ @ X → ∇@X. This latter space becomes
a topological category as follows: The objects are the continuous functors f :
∇ → X, the morphisms are continuous natural transformations between such
functors, which are defined by continuous maps n : I → X. The topology on
∇@X is essentially the compact-open topology, see [27, Ch. 62.1] for details.
Now, since Γ = colim ↑, the colimit of its arrows, we may define the limit
→
− ∼ →
−
topology colim ↑ @ X → lim ∇@X. This defines the topological category Γ @ X .
With this construction, we may step over to hypergestures in topological
categories.
5
Composition Denotators and Classification

Summary. In this chapter we discuss some denotators over Mod, traditionally

called (local) composition denotators. We also give examples of their classifica-
tion in music theory of tonal modulation.
–Σ–

5.1 Some Classical Composition Denotators

We consider denotators in forms F N ame : Id.P ower(M ) with the canonical
Id = 2@M ⊂ Ω @M whose denotators DN ame : A@F N ame{x1 , x2 , . . . xk } are
finite sets of A-addressed elements xi : A → M . Such denotators are called
A-addressed local compositions or simply compositions in M . Such denotators
were the first objects in mathematical music theory since its initial development
(see [25]) in the 80th of the 20th century.
Morphisms between local compositions

K : A@KF orm{x1 , . . . xk } in KF orm : Id.P ower(M )

and
L : A@LF orm{y1 , . . . yl } in LF orm : Id.P ower(N )
were defined by commutative squares for a fixed address A and two carrier
modules M, N :
K −−−−→ A@M
 

fy

yA@h
L −−−−→ A@N
where f is a set map that is a restriction of the evaluation A@h of an affine
module morphism h : M → N at A, and where the horizontal arrows are set
inclusions.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 27

This morphism concept is slightly more general than what we had defined
for denotator morphisms. The common information is the part A@h, but the
difference here is that f needn’t be surjective, f is only landing in L, not onto
L. However, when we later deal with isomorphism, this difference disappears.

5.2 The Role of Classification in Tonal Modulation

Theory

Let us now discuss the classification role in tonal modulation. This theory
usually deals with so-called tonalities and the mechanism of constructing mod-
ulations between such tonalities.
To begin with, modulations deal with tonalities that qua denotators are
isomorphic to each other. One starts with scales, such as the major scales, and
then develops mechanisms that involve isomorphisms between such scales. A
scale is defined to be a zero-addressed denotator S@0Chroma{s1 , s2 , . . . s7 }
with form of Chroma : Id.P ower(Z12 ). The C-major scale would be

Cmaj : 0@Chroma{0, 2, 4, 5, 7, 9, 11}

with the standard interpretation of 0 (the Cmajor tonic) for C, 2 for D etc., 11
for B. The isomorphisms would be the affine inversion isomorphisms T t .(−1)
and the affine transpositions T t . The group of these isomorphisms on Chroma
is denoted by T ±, ± standing for ±1 where T t ± (x) = t ± x. By definition,
a major scale is the image T t ± (Cmaj). For example, T 8 (Cmaj) = Emaj,
the major scale over E. The scale Cmaj has a unique automorphism T 4 .(−1).
Every major scale has a unique inversion automorphism. More generally, the
image T t (−1)(Xmaj) is the major scale whose third is the image of the tonic
of Xmaj.
Modulation deals with scales in the orbit of Cmajor under the group T ±.
This is a simple first background of music theory that starts from classifica-
tion. A tonality is a scale together with a covering of the scale by seven stan-
dard triads, so-called degrees, i.e., X ⊂ Xmaj that for Cmaj consists of these
three-element degree sets: ICmaj = {0, 4, 7}, IICmaj = {2, 5, 9}, IIICmaj =
{4, 7, 11}, IVCmaj = {5, 9, 0}, VCmaj = {7, 11, 2}, V ICmaj = {9, 0, 4}, V IICmaj =
{11, 2, 5}. Observe that this setup does not identify a tonic with special func-
tions. If we had to specify a tonic, i.e., an element of a major scale, this would
define a mode, but this is not addressed in the present setup.
What is significant for the definition of a tonality is that it exceeds the
simple scale and adds seven parts in the sense of the covering of a geographic
atlas by charts. This is the origin of a concept of global compositions (and
gestures) that will be dealt with Chapter 18.
The configuration of the seven degree chords that is defined by their inter-
sections is geometrically represented by the following construction: Represent
each degree of Xmaj by a point in three-space. Connect two degrees if they
5.2 The Role of Classification in Tonal Modulation Theory 29

have a non-empty intersection, and connect three degrees by a triangular sur-

face having these degree points as its vertices. Then we get what is called the
harmonic band Xmaj 3 . This band turns out to be a Möbius strip, see Figure
5.1.

Fig. 5.1: The harmonic band of triads.

Observe that every degree in a tonality is mapped to a degree of another

tonality under a connecting isomorphism T t ±. More precisely, for T t , every
degree XY maj is mapped to the degree of same name in the image tonality.
For T t (−1), we have these maps: I 7→ II, II 7→ V, III 7→ IV, IV 7→ III, V 7→
II, V I 7→ III, V II 7→ V II. This means that under T t ± we not only have
isomorphisms of scales, but also of tonalities.
Arnold Schönberg in his harmony book [44] described the configuration of
the lines connecting intersecting triadic degrees, but he did not make evident
the present geometric shape issued by the triangles of three intersection degrees.
It can be shown that the non-orientability of the harmonic band is a geometric
reason for the failure of Hugo Riemann’s attempted harmony theory, see [27,
CH. 13.4.2.1].
After these preliminary thoughts, we need to explicate the concept of a
tonal modulation m : Xmaj 3 → Y maj 3 using the classification, i.e., the fact
that these two tonalities are isomorphic to each other under the two isomor-
phism in the group T ±. The modulation should “materialize” such a connecting
isomorphism since symmetries are not collections of notes, the latter being a
conditio sine qua non for a concrete musical object. So the question arises
about how to generate sets of notes that represent connecting isomorphisms.
This problem is similar to the theory of physical forces, where one aims at mate-
rializing such forces by quanta, e.g., the electromagnetic force being represented
by photons.
We want to follow the modulation model as proposed by Schönberg in his
harmony book. He divides the modulatory process into three stages: (A) neu-
30 5 Composition Denotators and Classification

tralization of the tonality of departure, by playing neutralization degrees, (B)

playing the pivotal degrees of the targeted new tonality, (C) playing cadential
degrees of the new tonality.
To this end we have to define what is a cadence in our context.

Definition 3 A cadence in a tonality Xmaj 3 is a set C of degrees such that

no other scale contains all of these degrees. A minimal cadence is a cadence
with a minimal number of degrees.
There are five minimal cadences in Xmaj 3 , namely C1 = {II, III}, C2 =
{II, V }, C3 = {III, IV }, C4 = {IV, V }, C5 = {V II}. In stage (A), we may
play any combination of degrees that does not contain a cadence, for example
I, V I. The open problem here is the calculation of the set of pivotal degrees in
stage (B).
The main structure here is dealing with the pivotal degrees of a mod-
ulation. Formally speaking a modulation m : Xmaj I 3 → Y maj 3 is a pair
m = (s, C) where s ∈ T ± such that s(Xmaj 3 ) = Y maj 3 , s being called the
modulator and C is a minimal cadence of Y maj 3 . For such a modulation m, a
modulation quantum M is a set of pitch classes with the following properties:
1. We have s(M ) = M , i.e., the modulator s is an “inner symmetry” of M ;
2. The cadence C of Y maj 3 is a set of degrees, all of which are contained in
M , i.e., the quantum M is large enough to contain C;
3. the intersection of T ±∩Sym(Y maj ∩M ) is trivial and Y maj ∩M is covered
by degrees of Y maj 3 ;
4. the quantum M is a minimal set with properties (1) and (2).
Observe that the last property does not guarantee property (3). Therefore
the existence of M is not automatic. With such a quantum, if it exists, the
Y maj 3 degrees in M are by definition the pivotal degrees of modulation m. We
call m quantized if it admits a modulation quantum M .

Theorem 1 For every modulation m : Xmaj 3 → Y maj 3 , there exists a mod-

ulation quantum M , i.e., every modulation is quantized.
The list of all possible quantized modulations is shown in Section 23.5.

5.3 Modulations in Beethoven’s op. 106

Beethoven’s “Hammerklavier” Sonata op. 106 is a challenge for modulation
theory. We find some modulations that are executed very fast, with a minimum
of notes, although they connect tonalities that are far from each other in terms
of fourth distance. For example, in the Allegro movement, there is a modulation
B[ → G[ between distant tonalities in measures 238-239. But there are other
modulations that have a huge anatomy that is difficult to understand without
a deeper understanding of the modulatory process.
5.3 Modulations in Beethoven’s op. 106 31

Fig. 5.2: A complex modulation Gmaj → E[maj.

Let us look at an example of such a complex modulation. It is an excellent

example not only of Beethoven’s virtuosity, but also an illustration of the pres-
ence of the modulator within the composed score, i.e., of the presence of the
classificatory mechanism of isomorphisms between categorical objects, tonali-
ties in our case. Consider the modulation Gmaj → E[maj in measures 124-129
of the Allegro movement, see Figure 5.2. All modulations in the Allegro move-
ment have been analyzed and completely understood using our modulation
theory, see [27, Section 28.2].
Let us look at the modulation Gmaj → Emaj[ in measures 124-129 of the
Allegro movement. This modulation is bipartite (first part: measures 124-127,
second part: measures 128-129). Before we encounter the fundamental degrees
V II − V − V II in E[ in part two, according to our modulation table in Section
23.5, we hear note g as an octave interval: pedal and stationary voice in the first
part. The pitches of the first part, when transposed into the octave spanned by
the two g notes, show a regular melodic structure, see Figure 5.2, bottom.
This structure has two parts: the first in measures 124-125, and the sec-
ond in measures 126-127. They are related to each other by the inversion at d[,
which is the same as the inversion at g in pitch classes. This first part of the
32 5 Composition Denotators and Classification

modulation makes evident the inversion Id[ before we see the fundamental de-
grees in the second part of the modulation. But why this preliminary inversion?
It is the modulator for the modulation in our model, Id[ (G(3) ) = E[(3) . This
strategy is a beautiful compositional realization of what our model specifies.
The model does predict fundamental degrees, and it does so on the basis of
modulation forces that are provided by modulator symmetries. Beethoven not
only writes down the fundamental degrees, but also makes evident the modu-
lator symmetry in the first part of the modulation.
Our interpretation in this analysis does not assume that Beethoven has
performed his construction using the ideas of our model. But he might have
done so instinctively; one cannot know such hidden layers of creativity. This
situation is parallel to what happens in physics. We discover physical laws, but
we cannot know whether a divine creator (if this is the underlying cosmolog-
ical hypothesis) has constructed the universe according to these laws, which
are our way to understand nature. Nevertheless, the laws hold, and so does
our modulation model for the critical system of modulations in Beethoven’s
composition.
Concluding this chapter, we should add that our model also holds for
other compositions by Beethoven, for example for modulations in the Cavatina
movement of String Quartet op. 130.
6
Gestural Denotators: A First Overview

Summary. We discuss the formalism of denotators in digraph categories to

represent gestures generalizing the already successful denotator approach. This
is also a first step to a software-oriented implementation of gestural perspec-
tives. This generalization generates a powerful conceptual symmetry between
compositions and gestures. Both contexts start with simple forms, be it over
the category of modules or the category of digraphs. Presheaves over these cat-
egories then introduce higher compound concepts by one and the same topos-
theoretical technique starting from Yoneda’s Lemma.
–Σ–

6.1 The Symmetric Concept Architecture between

Compositions and Gestures
It is remarkable that the two conceptual threads, starting from the category
Mod of modules for local compositions or from the category Digraph of di-
graphs, show a symmetric shape, namely the fact that compound forms in both
cases are built from limit, colimit, and powerset constructions. Only simple
forms are different here, compound forms are deduced from the same topos-
theoretical tools.
This symmetry does however not imply that the classification of local
compositions and gestures shares the same properties. We will see in Chapter
19 that classification of gestures is highly problematic, not because of philo-
sophical aspects, but because the mathematical categories of gestures are far
from being understood, in opposition to the categories of local compositions,
where important classification theorems are available.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 33

6.2 The Role of Yoneda’s Lemma

In both categories the Yoneda Lemma plays a crucial role. In both cases the
Lemma’s statement that for a presheaf F : Cat → Set and an argument X in
Cat, the evaluation X@F of a F at X is identified with the set N at(@X, F ) =
@X@F of natural transformations from the representable presheaf @X to F .
This enables the definition of the categories of denotators for local compositions
and for gestures. This fact unites the two types of denotators, but it does not
guarantee an equally demanding classification procedure.
The difference in classification is focused upon simple forms, which are
much more problematic for gestures than for local compositions.
Let us look at two concrete examples, where gestural classification be-
→
−
comes problematic. The first example deals with the space ↑ @ REHLD of one-
−−−−−→
arrow gestures g :↑→ REHLD in the classical topological space of symbolic onset
E, pitch H, loudness L, and duration D. Such gestures intervene for perfor-
mance theory, which describes the performance functions ℘ : REHLD → Rehld
with codomain Rehld and physical parameters e, h, l, d corresponding to the
symbolic E, H, L, D. The performance maps ℘ are diffeomorphisms1 , and the
images ℘ ◦ g of g describe the performance of determined points (notes) in
∼
the symbolic space REHLD . We therefore obtain an isomorphism g → ℘ ◦ g of
gestures, both with domain digraph ↑, that are generated by the performance
map ℘. Such isomorphisms of gestures are difficult to classify since the classi-
fication of curves modulo homeomorphisms of underlying topological spaces is
not accomplished yet. In performance theory, the diffeomorphisms ℘ are not
arbitrary homeomorphisms, they are vice versa given to define isomorphisms,
which is not the proper classification objective.
A second, simpler, but even more dramatic, example is the classification
→
−
of simple loop gestures g : → X . Their classification is equivalent to the
classification of knots, which is far from settled. Loops in topological spaces are
critical structures, which we shall discuss in Chapter 21.1.
We should observe that the classification of zero-addressed local composi-
tion powerset denotators in a module is much simpler than the same problem of
zero-addressed gestures within a topological space. The former have the linear
algebra structures, which we shall discuss later, but for gestures, the classifica-
tion of collections of points in a topological space is far from understood.

6.3 Software Perspectives

Software for local compositions has been developed in the RUBATOr environ-
ment, programmed in Java [26]. The formal software representation of simple
gestures would be the starting point of an implementation of gestures since
1
A diffeomorphism is an isomorphism in the category, whose morphisms are differ-
entiable maps.
6.3 Software Perspectives 35

compound forms are implemented literally by the same methods and objects
as for local compositions. The critical question here is to represent a simple
→
−
gesture g : ∆ → X for digraph ∆ and topological space X. The software repre-
sentation of a digraph is not difficult, one has to define a vertex set plus a set of
arrows. But the topological space X is more problematic. More precisely, how
would one define continuous curves c : I → X? The map c must be given by a
formula that defines the map’s evaluation at arguments in I. This amounts to
implementing a set of possible formulas, polynomial, sinusoidal, exponential,
say. But the evaluation of such formulas must also be feasible, yielding function
values in X. This presupposes that X is given terms that may describe values
of available functions. In other words, the implementation of simple gestures
presupposes a collection of space constructions, together with a corresponding
collection of continuous functions. The software construction should provide
the user with flexible methods of constructing continuous functions c : I → X.
For local compositions in RUBATOr construction methods rely on standard
linear algebra constructions of modules via direct sums, projections, and the
like. For gestures, we need to construct very different objects.
7
The Escher Theorem for Compositions and
Gestures

Summary. The Escher Theorem establishes an isomorphism (classification)

among hyperdenotator (denotators of denotators) sets for a sequence of ad-
dresses, i.e., modules or digraphs. It establishes an isomorphism between dif-
ferent permutations of address sequences when applied to simple forms.
–Σ–

7.1 The Escher Theorem and Escher Categories

A hyergesture is a gesture of gestures. This is reasonable since the set of ges-

→
−
tures Γ @ X is a topological space/category [35], and therefore, one may con-
→
−
sider gestures with values in Γ @ X. The case of modules, leading to hyperde-
∼
notators, follows from the adjunction (with affine morphisms) M @(R@N ) →
(M R)@N , where M R = M ⊗ R ⊕ M ⊕ R is the affine tensor product.
The two hyperstructural constructions share a commutativity property, which
we coined Escher Theorem in [33], a proof for hypergestures can be found in
[35].

Theorem 2 If R, M, N are three modules then there is a canonical isomor-

phism of hyperdenotator sets
∼
R@M @N → M @R@N.

If Γ, ∆ are two digraphs, and X a topological category, then there is a canonical

isomorphism of topological categories of hypergestures
→
− →
− ∼ →
− →−
Γ @∆@X → ∆@Γ @X

This may be generalized to arbitrary permutations of addresses for hyperde-

notators and hypergestures: Let π be a permutation of the set {1, 2, . . . k} of

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 37

natural numbers, and take a sequence (Mi )i=1...k of modules as well as a se-
quence (∆i )i=1...k of digraphs. Further, let M be a module and X a topological
category. Then we have canonical isomorphisms of hyperdenotator modules or
topological categories, respectively:
∼
M1 @M2 @ . . . Mk @M → Mp1 @Mp2 @ . . . Mpk @M
and
→
− →
− →
− ∼ →
− →
− →
−
∆1 @ ∆2 @ . . . ∆k @ X → ∆p1 @ ∆p2 @ . . . ∆pk @ X.
The Escher Theorem unites local composition denotator theory with ges-
ture theory. We might also call more intuitively the categories Mod and
Digraph Escher categories. It is an open problem to set up an axiomatic for-
malism unifying these categories in a wider context, i.e., making precise what
are the general consequences of an Escher Theorem for such categories. This
problem is comparable to the situation for general homology theories.

7.1.1 First Examples

For modules and their hyperdenotators, refer to the dodecaphonic example in
∼
Section 4.4.2.1. We have Zd @Z11 @Z12 → Z11 @Zd @Z12 , which means that d+1-
tuples of dodecaphonic rows may be interpreted as 12-tuples of d + 1-phonic
rows. This generates a reinterpretation of rows in serial music: For example
with d = 3, we get four rows in Boulez’s Structures pour deux pianos, which
now may be interpreted as a sequence of 12 rows, each for the four parameter
rows in pitch, duration, attack, and loudness.
For digraphs ∆, Γ and topological categories X, the Escher isomorphism
− →
→ − ∼ − →
→ −
∆@Γ @X → Γ @∆@X
→
−
means that a ∆ hypergesture in Γ @ X can be viewed as a Γ hypergesture
−
→
in ∆ @ X. For example, if ∆ = Γ =↑ and X = REHLD , the exchange of
→
− →
−
the two copies of ↑ in ↑ @ ↑ @ REHLD means to interpret a deformation
of an performance (qua integral curve as described in Section 4.4.1.2) as a
performance of a deformation, which is a remarkable reinterpretation of given
hypergstures.
In practice the Escher Theorem enables an important application for col-
laborative performance. Suppose that two musicians are interacting through
→
− →
−
morphisms f between their hypergestures: f : ∆ @ . . . X → Γ @ . . . Y . The
codomain of f is a hypergestural structure with initial digraph Γ . If the Escher
Theorem is applied to the codomain, one might get another initial digraph,
and the morphism f would change into a morphism whose digraph map differs
from f ’s digraph map. Musically, this means to change the perspective of the
codomain of morphisms. So the musician who represents the domain of f views
his/her action under a different interpretation of the involved codomain and its
musician’s role. In jazz improvisation this is a well known procedure, focusing
on different aspects of the addressed musician’s behavior.
Part III

Local Classification
8
Local Composition and Gesture Classification

Summary. Local compositions and gestures are the basic structures in classical
musical concept spaces. They are essentially subsets in modules, such as chords,
motifs, and rhythms, or else digraph maps from a digraph to a topological
space/category. They are local since they are not built as configurations of
smaller subsets.
We describe in detail the theorem of classification of local compositions,
including its proof.
–Σ–

8.1 Classification of Local Compositions

We present and discuss the theory of classification of local compositions, which

we first discussed 1985 in our first book [25]. In this approach, we dealt with
what now are called orbifolds, and what is essential, though not mathematically
explicated, in a more recent approach by Dmitri Tymoczko [46], who did not
know my German book.
The classification in this section deals with simple musical objects, namely
local compositions, which are denotators d : 0@M odulF orm(M ) for simple
forms M oduleF orm : Id.P ower(M ) for modules M and Id the identity. There
are approaches to the classification of such local compositions for non-zero ad-
dresses and for general modules, we refer to [27, Ch. 11.2] for those approaches.
But here we want to present a context, which for special modules shows the
general methods without being mathematically too demanding.
The modules in our approach will be modules over a semi-simple com-
∼
mutative ring R, which means that R → k1 × . . . kn is a direct product of
a finite sequence of commutative fields ki . We shall not prove those facts for
such modules, which we use here, but simply inform the reader about what can
be assumed for such modules. So we consider modules M , where we have (1)
the usual structure of an additive commutative group, yielding sums m + n

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 41

of module elements m, n, and (2) the scalar multiplication r.m of a mod-

ule element m with a ring element r, with the usual distributive property
(r + s).(m + n) = r.m + s.m + r.n + s.n.
We classify local, zero-addressed compositions (K, M ), i.e. sets K ⊂ M of
elements of the M . Morphisms between such objects are set maps f : K → L
which are the restriction of an affine module morphism1 h : M → N , i.e.,
f (k) = h(k) for all k ∈ K. Isomorphisms are isomorphisms in this category,
i.e., morphisms f which have right and left inverses g with f ◦ g = IdL and
g ◦ f = IdK .
For a local composition (K, M ), we have its associated module RK ⊂ M ,
which is generated by all differences k − k0 , k ∈ K with respect to a fixed
element k0 ∈ K. The module RK is independent of the chosen fixed element
k0 since for another fixed element k1 , we have k − k1 = (k − k0 ) − (k1 − k0 ). This
association (K, M ) 7→ RK is in fact a functor from the category of our local
compositions to the category Mod. For the present semi-simple rings R, it can
be shown that every local composition (K, M ) is isomorphic to a generating
local composition (L, N ), which means that RL = N . We shall therefore restrict
our discussion to generating local compositions. For example, the major scales
are (Xmaj, Z12 ) are generating local compositions (attention: the ring Z12 is
not semi-simple, it is the product Z3 × Z4 , where the second factor is not a
field).
The classification of these local compositions splits into a sequence of
intermediate bijections of classifying sets as follows:

8.1.1 Compositions as Sequences of Module Elements

We start with a local composition (K, M ) having t + 1 elements. Consider the

symmetric group St+1 of permutations of natural numbers 1, 2, . . . t + 1. Then
K may be interpreted as a sequence K. = (k1 , . . . kt+1 ) of its elements, i.e.,
an element of M t+1 modulo the action of St+1 , the σ ∈ St+1 taking k. =
(k1 , . . . kt+1 ) to σ.k. = (kσ(1) , . . . kσ(t+1) ). The elements of the sequence K.
must be pairwise different. The subset of M t+1 of pairwise different element
sequences is denoted by M t+1 . The local compositions with t + 1 elements are
the orbits in M t+1 /St+1 of the St+1 action.

8.1.2 Eliminating Diagonal Elements

Here we reduce the sequence K. to the set of its differences. We first consider
the short exact sequence of R-modules

0 → ∆M → M t+1 →d M t → 0,
1
An affine module morphism h : M → N is the composition of a linear morphism
h0 : M → N with a translation T t : N → N on the codomain N of h0 .
8.1 Classification of Local Compositions 43

where ∆M is the diagonal submodule and d(x1 , . . . xt+1 ) = (x2 − x1 , . . . xt+1 −

x1 ). The group operation of St+1 on M t+1 leaves ∆M invariant, therefore St+1
∼
operates on M t → M t+1 /∆M . On the other hand, the linear group GL(M )
operates on M t , and these two operations are compatible with each other. This
entails the

Lemma 1 Let M t be the subset of M t such that its elements are (1) differ-
ent from zero, and (2) the sequences (y1 , . . . yt ) generate M . Then the set of
isomorphism classes of t + 1-element generating local compositions in M is in
bijection with the orbit set GL(M )\M t /St+1 .
Proof. If (K, M ) is a t + 1-element local composition, it defines an orbit
by taking its sequence (x2 − x1 , . . . xt+1 − x1 ) in M t and then stepping over
to the orbit by permutation of the sequence and the GL(M )-orbit. Every such
local composition can be found in the orbit space. If two local compositions are
associated with the same orbit, this means that their original element sequences
are transformed by an element of GL(M ) and possibly by a permutation of their
sequences of differences. That we have to take linear maps in GL(M ) is due to
the fact that transpositions cancel out for differences of K elements. So they
are isomorphic, QED.

8.1.3 Transforming Sequences to Linear Maps

This third step deals with the restatement of sequences of module elements
as linear maps between modules. This will be used to introduce subspaces of
modules instead of sequences, and thereby representing isomorphism classes in
geometric spaces (Grassmanians) whose points are such subspaces of modules,
a technique that is standard in modern algebraic geometry.
∼
We use the basic fact that M t → Hom(Rt , M ) sending a t-tuple (mi ) ∈
M to the linear map which evaluates at ei = (0, . . . 0, 1, 0 . . . 0) with the 1 at
t

position i to mi . The above operation of St+1 on M t caries over to an operation

on2 Hom(Rt , M ) and thereby in particular to Rt .
Now consider the sets

Xt (R) = {V ⊂ Rt , V a submodule with ei , ei − ej 6∈ V all i 6= j}

and
∼
XtM (R) = {V ∈ Xt (R), Rt /V → M }.
Evidently, St+1 operates on both, Xt (R) and XtM (R), and we have

Lemma 2 The orbit space XtM (R)/St+1 is in bijection to the space LgM,t+1 of
generating t + 1-element local compositions on a module isomorphic to M .
2
The R-linear maps.
44 8 Local Composition and Gesture Classification

Proof. This bijection is induced by the map f 7→ Ker(f ). Let us verify

that this map generates a bijection. To begin with, by Lemma 1, such a local
composition is identified with an orbit in GL(M )\M t /St+1 , which means that
it is defined by the orbit of a surejctive linear map f : Rt → M . Then Ker(f )
is an element of XtM (R), and the action of the permutation group identifies
the orbits in both spaces. The action of GL(M ) identifies the GL(M ) orbits by
definition of LgM,t+1 . This association injects the space LgM,t+1 via kernels into
XtM (R)/St+1 . But it is also surjective since a subspace V defines a surjection
Rt → M with the defining properties in XtM (R)/St+1 .
This identification means that the set of isomorphism classes of generating
local compositions can be identified with the set
a
Xt (R)/St+1 = XtM (R)/St+1
M ∈M

where M is a set of representatives of module isomorphism classes. This may

be stated as a theorem:

Theorem 3 The orbit space Xt (R)/St+1 is in bijection with the set of isomor-
phism classes of generating t + 1-element local compositions over R.
Since every such local composition is isomorphic to a generating local
composition by simply taking the generated module RK and shifting K into
RK if necessary, this classification also holds without the “generating” attribute.

8.1.4 The Grassmann Scheme for Local Compositions

In this final section, we shall interpret the orbit space Xt (R)/St+1 as a set
of points in an algebraic variety (also called scheme in modern algebraic ge-
ometry), the Grassmannian. This discussion presupposes some knowledge in
algebraic geometry, but it means that isomorphism classes may be viewed as
points in a geometric space.
The scheme Grass(r,t) is the scheme, whose point set Grass(r,t) (R) for the
ring R is defined as the set of direct summands V ⊂ Rt which have a quotient
Rt /V that is of rank r. The latter means that such a quotient is free of rank r
when taking the local modules S ⊗ Rt /V over localizations S of R at its prime
ideals. The operation of St+1 on Grass(r,t) is evidently the restriction of its
operation on Xt (R). This means that the conditions ei − ej 6= 0, ei 6= 0 define
an open subscheme of Grass(r,t) . In other words

Lemma 3 The subfunctor Xtr ⊂ Grass(r,t) is represented by an open, St+1 -

invariant subscheme.
This open subscheme has its orbits which are finite. This implies in view of gen-
eral theorems from algebraic gometry that there is a quotient scheme Xtr /St+1
whose orbits are the isomorphism classes.
8.2 Classification Problems for Local Gestures 45

Theorem 4 The set of isomorphism classes of local, t+1-element compositions

over R is represented as the set of R-valued points of a scheme of finite type
(defined by a finite number of indeterminates over the base field).

8.2 Classification Problems for Local Gestures

In opposition to local compositions, there is not (yet) a classification theo-

→
−
rem for local gestures, i.e., digraph maps g : ∆ → X for a digraph ∆ and a
topological space/category X.
Classifying such objects in their given category turns out to be an open
mathematical problem. Let us focus on a critical example: ∆ = . A local
gesture over this digraph is essentially (equivalent to) a continuous map from
the unit circle in R2 to X. Such a map is called a knot, and the classification
of knots is an open problem in mathematics, but see Chapter 21.1 for more
details. This means that there is no classification theorem for local gestures.
9
Classification of Chords

Summary. This chapter deals with chords, i.e., zero-addressed local composi-
tion denotators K = {c1 , . . . ck } ⊂ Z12 , i.e. K : 0@P itchClass{c1 . . . ck } for the
form P itchClass : Id.Simple(Z12 ). Such denotators can also be interpreted as
a 12-periodic rhythms, we shall occasionally consider this latter point of view.
–Σ–
This classification follows the general concepts, but uses some mathemat-
ical tools which we cannot develop here in our elementary approach. We may
however state that two local compositions K : 0@P itchClass{c1 . . . ck }, L :
0@P itchClass{d1 . . . dl } are isomorphic if and only if there is an affine auto-
morphism f = T t .r, r = 1, 5, 7, 11 such that f (K) = L. The list of all represen-
tatives of isomorphism classes is displayed in Section 23.1. Let us discuss some
immediate consequences of this classification.
It is a well known fact that when we step from C Major to F Major to
B[ Major to E[ Major to A[ Major to D[ Major to G[ Major, the number of
[s increases by one for each step, starting with no [ for C Major and ending
with six [s on G[ Major. Such a regular increase would not be the case if we
had to step from C harmonic minor to F harmonic minor, etc. Why is such and
increase happening?
The reason is that in our list of chord classes we recognize that the C
Major scale is the complement of class 38.1, and the latter is isomorphic to
class 38, i.e., the C Major scale is isomorphic to the complement of class 38.
∼
The isomorphism is given by f : {0, 2, 4, 5, 7, 9, 11} → {11, 0, 1, 2, 3, 4, 5} with
f = f −1 = .7, yielding 7.11 = 5, 7.0 = 0, 7.1 = 7, 7.2 = 2, 7.3 = 9, 7.4 =
4, 7.5 = 11. Now, stepping from C Major to F Major etc., always is per-
formed by a transposition T 5 by a fourth. Then F M ajor = T 5 (CM ajor)
etc. until G[M ajor = T 5 (D[M ajor). Transforming these relationships we get
f (F M ajor) = .7(5 + CM ajor) = 11 + {11, 0, . . . 5} = {10, 11, 0, 1, 2, 3, 4}. This
is just a transposition of class 38 representative by one unit, i.e., the fourth
transposition morphs to a minor second transposition. But hereby, we lose one
point at one end of the chromatic point sequence and add another point at

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 47

the other end. This operation is repeated for every fourth transposition of the
original major scale, we always add a new note by a [ and take away a note
from the previous scale. One could simply state that the increasing sequence
of [s is due to the fact that a major scale is an uninterrupted sequence of fifths
or, in the other direction, of fourths.
A second, even more important, fact from this classification is given by
∼
class 87, which is autocomplementary, there is an isomorphism f : 87 →
ˆ
87,where ˆ denotes the complement of 87. To understand this, observe that
87
class representative 87 = {0, 1, 3, 4, 7, 8} is a transposition of {1, 2, 5, 6, 10, 11}
by 2. But the latter represents the set of dissonances in classical Fuxian coun-
terpoint, and the autocomplementarity function of the dissonances with the
consonances {0, 3, 4, 7, 8, 9} is T 2 .5. It is also the unique function that connects
consonances to dissonances, since the automorphisms of 87 are trivial. This re-
sult is the basis of a mathematical counterpoint model that derives important
rules, e.g., forbidden parallels of fifths, and the dissonant qualification of the
fourth, see [3].
10
Motif Classes

Summary. This chapter sets a perspective onto the classification of motives,

ie., local compositions K ⊂ Z212 . The general case is not made explicit yet, but
for card(K) ≤ 4, we have complete lists and some creative applications.
–Σ–
Motives are the two-dimensional elements of melodic structures. We shall
deal with this gluing process that generates melodies and more general “global
compositions” in Chapter 18. Here we look at the classification of motives with
the aim of constructing global compositions, i.e., set-theoretically speaking, the
union of local motives to build melodies of a certain character with respect to
the classes of local motives.

10.1 Three Element Motives in Analysis

To begin with, the isomorphism classes of three element motives K ⊂ Z212 are
26 in number, their list is shown in Section 23.3. We also show the list in a
geometric, two-dimensional way in Figure 23.1. And we show the Hasse diagram
of these classes in Figure 23.2, which means the following: There is a dominance
arrow from class m to class n iff there is an affine map d : Z212 → Z212 such that
d(m) = n, and d cannot be an isomorphism since the classes m, n are supposed
to be different. For example, class 10 dominates class 18 via a multiplication
by 2 of its coordinates.
The Hasse diagram denotes all shortest dominance arrows, i.e., those
which cannot be decomposed into a sequence of two or more arrows. For ex-
ample, the dominance 10 → 18 decomposes into a dominance 10 → 12 and a
dominance 12 → 18. It results that class 10 is a unique “generic” three-element
motif, all three-element motives are dominated by class 10.
It is remarkable that three-element classes are characterized by two items:
the classes of their two-element submotives and the volume of a three-element
class. The volume is the Z12 element derived from the second exterior power

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 49

Λ2 (Z12 K) of the motif K. Unfortunately, this characterization holds no longer

for four- and higher -element motif classes, see our list in Section 23.4.

X
Xj. b Xk Xj X XjXj
j Xk @ K @ XK
I Mit-ten im Schim-mer der spie-geln-den Wel-len
14 14 19

II glei-tet wie Schwä-ne der wan-ken-de Kahn

14 14 6

III1 Ach auf der Freu-de sanft schim-mern-den Wel-len

23 10 11

III2 ... ... ... ...

11 10 11

IV1 glei-tet die

XJ . b X Xj
See-le da- hin wie der Kahn
3 10 K 11

IV2 ... ... ... ...

3 10 11

V Denn von dem Him-mel her- ab auf die Wel-len

14 10 14

VI1 tan-zet das A-bend-rot rund um den Kahn

10 10 10

VI2 ... ... ... ...

10 14 10

Fig. 10.1: The highly symmetric arrangement of motive classes in the dactylus grid
of the voice score in Schubert’s setting of Stolberg’s poem Lied auf dem Wasser zu
singen, für meine Agnes.

10.1.1 An Example from Classical Music

We consider the analysis for Schubert’s composition Auf dem Wasser zu sin-
gen, für meine Agnes. The composition shows a highly symmetric arrangement
10.2 Three Element Motives and a Jazz Example 51

of three-element melodic motives, see Figure 10.1, the numbers refer to the
numbers in our classification list in Section 23.3.3.

6
& 8 X X X b X b X X X X b X b X b X X X X X X X X X b X X XJ b X b X X X Xj
X
15

19
24
25
14

26
3 16
22
13 6

23 4
11

5
10 8 7
2
21 17 18
1 20

Fig. 10.2: The germinal melody of the Synthesis concert for piano, percussion, and
bass is a patchwork of 26 three-element motives, each representing one motif class in
OnP iM od12,12 . The melody is shown as a score and as local composition (red-colored
points).

10.2 Three Element Motives and an Example of a Melody

from this Repertory in the Jazz Composition Synthesis
For the construction of the composition Synthesis, a germinal melody was con-
structed as a patchwork/union of representatives of all 26 classes, see Figure
10.2. This procedure guarantees the presence of all of 26 classes and thereby
52 10 Motif Classes

generates a melody that is as generic as possible. Here the classification is used

for a melodic construction.

10.3 A General Algorithm for the Construction of

Generic Melodies Composed by Motives of n Elements
Each
This section proposes an algorithm to construct a melody M that is the union of
representatives of all classes of n-element motives. The idea is as follows. Such
an algorithm is mandatory because there are hundreds of classes in general,
for example for n = 4 we already have 215 classes, see the list in Section 23.4.
We suppose that we have constructed a partial melody M ∗ that is the union
of the first k representatives, and we then choose the first n-element motif
that is not contained in M ∗ . We add this new motif to M ∗ , and we continue
until representatives of all classes are added. This method should enable us
to generate melodies that are generic in the sense that they are the union of
representatives of all classes of n-element motives.
Such a method can be realized in many ways, e.g., adding disjoint new
representatives, but we want better solutions, i.e., solutions that add a mini-
mal number of new notes. We first want to check all the classes that are not
contained in M ∗ (modulo Z12 2
). Suppose we found a first such candidate N .
Then, we want to look for the maximal part of N 0 ⊂ N that is contained in
M ∗ . We then add the missing elements of N to M ∗ . We could do so by taking
the first representatives and adding them to M ∗ . We should do so by taking
the first representatives of these elements with the smallest values in Z × Z12 .
The open question so far is to find the first missing motives N . This
can be achieved by taking in the Hasse diagram of n-element motives the first
missing N . What does this mean? If we look at the distance function on this
Hasse diagram, defined by the minimal length sequence of arrows from the
M ∗ n-motives. Within the missing motives, we take one with such a minimal
distance to the M ∗ motives and in this selection taking the first for an a priori
given linear ordering among the n-element motives.
This selection method guarantees that we add a N that is nearest to the
given set of n-motives in M ∗ . This method depends on the a priori given linear
order and on the Hasse diagram, in particular, given the linear ordering, we
may start the entire procedure with the first candidate to get off the ground.
Let us write down this algorithm a bit more formally: Suppose we are
given a linear ordering of the set Class(n) of n-element motive classes in Z12 2
.
1. Start the procedure with a representative M0 of the first class.
2. Suppose we have defined a partial melody M ∗ that is the union of k repre-
sentatives of Class(n). Take the set C(M ∗ ) of classes that are not contained
in M ∗ . Take in C(M ∗ ) the first candidate N with a maximal number of
elements, i.e., a maximal submotif N 0 ⊂ N ∩ M ∗ and with minimal Hasse
distance, and therein the first of the given order.
2
10.4 Calculation of the Classes of n-element Motives in Z12 53

3. Take elements representing missing points in N and add them to M ∗ .

Thereby a representative of the class of N is added to M ∗ .

10.3.1 A RUBETTEr for Generic Melodies

For three-element motives in Z12 2

, we have the complete information: all 26
classes are calculated, the Hasse diagram is known, and for every class, the
classes of its two-element submotives are given. Also, the isomorphism of two-
element subcompositions, together with the motives’ volumes, are sufficient for
the determination of isomorphism classes. This implies that the Java program-
ming of a RUBETTEr is feasible for the task of generating generic melodies
from three-element motives on the RUBATOr platform. The linear ordering of
the 26 classes is the only variable in this project, but 26! is nevertheless a big
number, more than the number of possible 12! dodecaphonic rows.
The musical idea behind the concept of a generic melody is to produce
a melody that subsumes all possible classes of motives of a given size. This
generates musical objects that present a maximal variability of its submotives
of a given size. The construction of “interesting” melodies is a still unsolved
problem: When is a melody interesting, interesting under what perspective? Of
course, historical examples tend to be characterized by a property of becom-
ing “ear worms”. Our approach here will very probably not comply with this
psychological condition, but we believe that our construction yields melodies,
whose anatomy is as rich as possible.

10.4 Calculation of the Classes of n-element Motives

in Z212
Despite the general classification of local comopositions, one might also try to
approach classification from a recursive point of view. Here is the method. Given
a local composition K ⊂ M , one can define its simplicial weight as follows. On
looks at al the proper subcompositions L ⊂ K and then exhibits all their non-
empty intersections, called simplices. So a simplex isTa set σ = {L1 , . . . Lk } of
proper subcompositions with non-empty intersection σ. This configuration is
called the composition’s nerve n(K). For every σ, one exhibits its isomorphism
class cl(σ). The map σ 7→ cl(σ) is called the simplicial weight, it defines the
totality of isomorphism classes of proper subcompositions.
∼
It is an open question, whether two nerves being isomorphic n(K) → n(L)
∼
implies K → L. If this can be proved, one would get a recursive method to deter-
mine isomorphism classes of local compositions. The isomorphism of simplicial
weights is necessary for an isomorphism of two local compositions, but it may
be insufficient. The basic problem is the continuation of isomorphisms: If we
have two isomorphic local compositions K, L with isomorphic subcompositions
∼
K 0 , L0 of K, L, respectively, can we extend a given isomorphism φ0 : K 0 → L0
∼
to an isomorphism φ : K → L?
54 10 Motif Classes

We see that such extensions are not always possible. In particular, this
means that isomorphisms of simplicial weights are necessary but not sufficient
for isomorphisms of the underlying local compositions. An example is shown in
Figure 10.3. Here intervals of length one are isomorphic, but no automorphism
of the motif extends their isomorphisms in general, the only automorphism of
the motif being the diagonal exchange.

Fig. 10.3: A motif which has only the diagonal exchange as an automorphism since
its horizontal and vertical parts are rigid.
11
Third Chain Classes

Summary. Third chains are chords that are built from a concatenation of
minor or major thirds. A major triad 0, 4, 7 is a third chain, concatenating a
major third 0, 4 with a minor third 4, 7. Third chains are important in harmony,
and we shall demonstrate how they are applied to the management of harmonic
contents of general chords. This approach is important for the harmonic analysis
of musical compositions.
–Σ–
In Section 23.2 all third chains are listed. They are denotators of the fol-
lowing type. Let Interval : id.Simple(Z12 ) be the simple form for intervals.
Let Base : id.Simple(Z12 ) be a starting pitch form. Let T hirdChain.n :
id.Limit(Base, Interval, . . . Interval) be the form with n intervals. Then a
third chain with two intervals, such as (3, 4), and based upon starting pitch 0, is
given by the denotator Cm : 0@T hirdChain(B, I1 , I2 ) with B : 0@Base(0), I1 :
0@Interval(3), I2 : 0@Interval(4).
Third chains are a finer construction mode than local compositions. The
construction of pitch classes via T 3 , T 4 and a variable initial pitch class follows
the idea of David Lewin and the author, where sets are replaced by operations
that generate sets, and objects plus such operations may be compared in a finer
approach. For example, the minor triad third chain 2, 5, 9 is isomorphic to the
major triad third chain 11, 3, 6, but as chains they are different.
In the next chapter, we shall discuss a harmony in Hugo Riemann’s sense
that uses third chains to define harmonic functions of local compositions.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 55

Summary. We discuss the application of third chain classification to Rieman-

nian harmony for general chords.
–Σ–
Third chains are important for harmonic theory by the following method.
We suppose that third chain chords are given a determined harmonic function.
Then a general chord Ch may be embedded in third chains, e.g. Ch = {2, 9},
the fifth interval chord over d = 2. We may embed Ch in the third chains
Dm = 2 + (0, 3, 7) or DM = 2 + (0, 4, 7). The harmonic function of Ch may
then be determined by a combination of he harmonic functions of Dm and DM .
Every chord can be embedded in at least one third chain, e.g. the third chain
of the entire pitch class set, namely (0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11). With this in
mind, one may define the third chain spectrum T hirdChainSpectrum(Chord)
as being the set of minimal third chains that contain Chord. For the above Ch,
we have T hirdChainSpectrum(Ch) = {Dm, DM }.
Therefore we should discuss potential harmonic functions of third chains.
If we construct harmonic functions f (T hirdStream) that are numerical, we
may define for a chord Chord its function by
1 X
f (Chord) = f (Ch),
N
Ch∈T hirdChainSpectrum(Chord)

N being the number of third chains in T hirdChainSpectrum(Chord).

The point now is to define f (X) for third chains X. To this end, we first
define the tonality for which this functions should be defined. Let us suppose
that we have two choice: Y major or Y minor. With such a choice, that would
be symbolize by the Y + for major an Y − for minor. Then, we define a harmonic
value of harm(x, Y ±) for pitch class x and tonality Y major (+) or Y minor
(−). With these values, we define
X
harm(X, Y, ±) = harm(x, Y ±).
x∈X

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 57

This defines the value f (Chord) with the values of the third chains in its spec-
trum and the tonality information (Y ±). One might also refine this information
by defining harmonic values

harm(x, Y ±, T onic), harm(x, Y ±, Dominant), harm(x, Y ±, Subdominant),

and then calculating the chord’s value with respect to the tonality parameters
Y, R, R = T onic, Dominant, Subdominant. All in all, this defines a so-called
Riemann matrix Riemann(Chord)(Y, R) with twelve values for Y and six val-
ues for R, three in major and three in minor mode.
The point of this approach is that a chord may have values for all 12 ×
6 = 72 matrix entries, but not in the sense of “yes” or “no”, the values are
a fuzzy logic representation of the harmonic value of a chord. This approach
was implemented in RUBATOr ’s HarmoRubette, a software component for
Riemannian harmonic analysis.
Therefore the classification of third chains offers a tool for Riemannian
harmony with fuzzy numerical values.
13
Counterpoint Worlds

Summary. The classification of strong dichotomies yields a basis for the rules
of counterpoint. We discuss some “exotic” counterpoint worlds (five besides the
Fux world) that are generated by this classification.
–Σ–
This chapter discusses some counterpoint topics from the perspective of
classification. We shall however not penetrate the mathematical model of coun-
terpoint, for which we refer to [27, Part VII] and to [3].
Our approach to counterpoint gets off the ground with a redefinition of
consonant and dissonant intervals. Classically, interval qualities are derived
from the historical approach as introduced by the Pythagorean school under
the form of their tetractys, see Figure 13.1.

Fig. 13.1: The Greek tetractys symbol.

This form is a triangular display of 1,2,3, and 4 points, starting from the
top. It is the expression of a “world formula” in this Greek cosmology. Music
was then understood as an audible expression of this tetractys in the sense that
successive fractions, 1/2, 2/3, 3/4 would represent the intervals of an octave,
fifth, and fourth, where these fractions were meant to refer to relations of the
length of a string of the monochord, i.e., 1/2 represents going to half the full
length of the string, producing an octave higher, 2/3 represents the fraction of

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 59

the full length representing the fifth, and 3/4 represents the fraction of the full
string representing the fourth. Equivalently, the inverse fractions 2/1, 3/2, 4/3
would represent the relative frequencies of the fractions of the monochord’s
string. Later, with Zarlino’s approach in the 16th century, the number 5 was
added, generating the pentactys and enabling the fraction of 4/5 that represents
the major third.
These fractions were understood to define consonant intervals via fre-
quency ratios. This means that an interval would be classified as a consonant
if it represented one of these fractions, which were also the basis of the just
tuning. This classification however had the disadvantage that the fourth 3/4
was no longer understood as being consonant in the 16th century, when coun-
terpoint theory had fully developed since its first appearance around 900 with
the Gregorian chorale. After 600 years of contrapuntal experiments with vari-
able concepts of consonances vs. dissonances, the final shape of counterpoint in
composition and theory had produced an interval concept that was no longer
congruent with the tetractys and pentactys.
The point here is that the contrapuntal concept of a consonance, 0,3,4,7,8,9
semitones between the two notes of the interval, i.e., prime, two thirds, fifth,
two sixths, was no longer a consequence of the Pythagorean approach. In par-
ticular, the fourth had become a dissonant interval. The simple fractions were
no longer a valid criterion!
There have been several attempts, also in the 20th century, to redefine
consonances using fractions, but none could create the contrapuntal conceptu-
alization, which had been set up with Palestrina and later, in 1725, canonized
by Johann Joseph Fux in his famous Gradus ad parnassum. Moreover, the con-
trapuntal rules also forbid the “parallels of fifths”, i.e., the immediate succession
of a fifth after a fifth. The prestigious German music theorist Carl Dahlhaus
argued that this mysterious situation in counterpoint theory was probably due
to an undiscovered argument from musical syntax in composition.
Our approach to counterpoint therefore started with a fundamental cri-
tique of the traditional definition of a consonance. The idea was that in compo-
sition, a consonant interval is not consonant by an attribute of a single interval,
but that such a quality is derived from the fact that in composition, consonances
are defined as collections of intervals that are interesting when creating a mu-
sical work. In other words, we argued that being consonant is a property of
being a member of a distinguished collection of intervals, and not a property
of a single interval without considering other intervals.
The solution of our approach was a criterion relating to the dichotomy
(K = {0, 3, 4, 7, 8, 9}, D = {1, 2, 5, 6, 10, 11}) of consonant v. dissonant inter-
vals. In fact, it turned out that the dichotomy (K, D) is strong, which by
definition means that there is a unique automorphism p of Z12 , which ex-
changes K and D. This automorphism is called polarity, and its formula is
p = T 2 .5. We have p(K) = D, and, by the way, also p(D) = K. For example,
p(3) = 5.3 + 2 = 5.
13 Counterpoint Worlds 61

This polarity is used to derive all contrapuntal rules of Fux, in particular

the forbidden parallels of fifths. See [27, Part VII] and [3]. This theory does not
refer to any psychological argument, it is a purely mathematical model, a fact
that opposes it to the usual psychology of contrapuntal arguments.
The problem here is the question, why the dichotomy (K, D) would be
distinguished among all possible strong dichotomies. This is a classification
problem: Describe all possible strong dichotomies and find arguments for the
choice of (K, D). Such arguments can be determined, when one considers the
∼
geometry of the interval space Z12 → Z3 × Z4 . In the geometry of this space,
the Fux dichotomy has a maximal separation of consonances vs. dissonances.
14
Strong Interval Dichotomies

Summary. Classical Fuxian counterpoint is based upon a dichotomy of the set

of 12 intervals within the set of pitch classes, specifying six consonances versus
six dissonances. We characterize this dichotomy by its unique autocomplemen-
tarity symmetry that exchanges consonances and dissonances, and we calculate
all possible dichotomies with this property.
–Σ–

Definition 4 In the module is Z2n , a dichotomy is an ordered pair (K, D)

where each part K, D has cardinality n and they are disjoint. In other words,
Z2n = K ∪ D, and K ∩ D = ∅, and card(K) = card(D) = n.
The most famous example is the Fux dichotomy

({0, 3, 4, 7, 8, 9}, {1, 2, 5, 6, 10, 11})

from classical counterpoint, where the first half denotes the consonant intervals
of a prime, two thirds, the fifth, and two sixths.
To develop a counterpoint theory, one has to establish rules that define
allowed and forbidden successive intervals. For example, parallels of fifths are
forbidden, i.e., a fifth may not immediately succeed another fifth. To derive the
set of allowed successions k1 → k2 of consonances, one needs to apply another
property of dichotomies:

Definition 5 A dichotomy (K, D) is called strong if its parts are isomorphic

by a unique symmetry, which is called the dichotomy’s polarity.

To exhibit all strong dichotomies, one inspects the list or chord classes 23.1
and searches for those classes C of six-element chords, which are isomorphic to
their complement. Evidently, if a dichotomy (K, D) is strong, all its transfor-
mations (g(K), g(D) by isomorphisms of Z12 are also strong. These classes are
six in number, namely classes C = 64, 71, 75, 78, 82, 87. The Fux dichotomy is

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 63

class 87. It is remarkable that some of these dichotomies arise in natural musical
contexts. For example, class 64 is represented by the dichotomy {2, 4, 5, 7, 9, 11}
of all proper intervals in the major scale when calculated from the major tonic.
Scriabin’s mystical chord {0, 2, 4, 6, 9, 10} represents dichotomy class 78. And
Boulez’s pitch class series in his Structures pour deux pianos has its first six
pitch classes defining a strong dichotomy of class 71.
Conceptually speaking, counterpoint theory only needs the specification
of a strong dichotomy class, all technical methods work without additional
hypotheses. This means that we have a counterpoint theory for each of the
six strong classes. The Fux theory is just the case of class 87. It should be
mentioned that for the Fux theory, the allowed transition k1 → k2 are exactly
those described by Fux in his famous Gradus ad parnassum [9].
There is a mathematically defined geometric reason for selecting class 87.
To understand this fact, one represents the first half of a strong dichotomy as
a subset of the third torus Z3 × Z4 , see Figure 14.1.

Fig. 14.1: The third torus separates consonances (K) from dissonances (D) in an
optimal way.

The Fux dichotomy shows a maximal separation of K and D on the geom-

etry of the torus. To select the Fux dichotomy within the class 87, one observes,
that the Fux selection is defined by a consonance part that is closed under mul-
tiplication, it is a multiplicative monoid. This implies that the Fux dichotomy is
uniquely determined by a set of purely mathematical conditions that are derived
from the classification of chords, no psychology is required.
15
Microtonal Contrapuntal Theories

Summary. We shortly present the extension of counterpoint worlds to mi-

crotonal environments, an extension that would be very problematic, if at all
possible, without the theory exposed above.
–Σ–
After the classical environment of pitch classes in Z12 had been investi-
gated relating to counterpoint, Octavio Alberto Agustín-Aquino, a Mexican
mathematician, started investigating counterpoint theories in finer environ-
ments Z2n for any non-trivial n, especially quarter-tone space Z24 . It turned
out that such an extension is feasible in a remarkably coherent way: Not only
is it possible to perform counterpoint in such refined spaces, but it is also pos-
sible to build “towers” of contrapuntal theories, connecting theories in smaller
spaces to theories in larger spaces. It may be possible with an increasing so-
phistication of music technology to write compositions that live in such towers.
The computer scientist Julien Junod has written software in this direction. All
these efforts were published in a joint book [3].
We should recall that all of this was only possible because of the new defi-
nition of consonances qua sets of intervals with properties that are generic from
the mathematical point of view and relate to classification.
In this chapter, we want to account for this revolutionary development
of a counterpoint theory that transcends all traditional efforts to understand,
what counterpoint is about.

15.1 The Category of Strong Dichotomies

The category SD of strong dichotomies has as objects the strong dichotomies
X = (C, D) on a given Zn , n an even number, together with their polarities
pX , which are affine automorphisms of Zn exchanging C and D.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 65

For two strong dichotomies X1 = (C1 , D1 ) on Z2k1 with polarity p1 , and

X2 = (C2 , D2 ) on Z2k2 with polarity p2 , a morphism f : X1 → X2 is an affine
morphism f : Z2k1 → Z2k2 such that the following diagram commutes:
f
C1 −−−−→ C2
 
p1 y
 p2
y
f
D1 −−−−→ D2
Morphisms of strong dichotomies are connecting counterpoint theories
that live in different domains, and in particular are refinement of each other.

15.2 Towers of Strong Dichotomies

The basic theorem (proved by Agustín-Aquino, see [3]) states that there exist
arbitrary long sequences of such morphisms:

Theorem 5 There is an infinite sequence of the canonical injective morphisms

of strong dichotomies (a “tower”)
X1 X2 . . . X m . . .
with Xm on Z3.2m for all m = 1, 2, 3 . . ..
This means that we start on Z6 , the go to the traditional Z12 , then to
microtones on Z24 , etc. We may in particular generate such a tower with the
Fux dichotomy on Z12 .
16
Dodecaphonic Rows

Summary. This section concludes the discussion of local classification with a

presentation of the variety of dodecaphonic series.
–Σ–
The dodecaphonic composition method was introduced by Arnold Schoön-
berg around 1921. It is a radical deviation from the classical structures of tonal
compositions, such as consonant intervals, standard chord (triads etc.) and
harmonic classification, for example via cadences and modulations.
Schönberg proposes to base a dodecaphonic composition on the selection
of a row, i.e., a determined sequence r = (p1 , p2 , . . . p12 ) of all pitch classes in
Z12 . There are 12! = 479, 001, 600 such rows. Together with this selected row,
−→∗
Schönberg allows the composer to include all rows f (r), where f ∈ GL , the
group of all 48 affine automorphisms f of Z212 which are generated by inversions,
retrogrades, retrograde inversions and transpositions. Such a composition would
−→∗
consist of a patchwork of rows in GL (r), a set of 48 transformed rows starting
from r. In general, due to possible symmetries of r, not all 48 derived rows will
be different. In terms of classification, one considers the set R of all rows and
−→∗
then takes one of the orbits GL (r). Schönberg also allows for successive row
elements to have the same onset.
The problem of this approach is that there is no rule that would define a
global organization of such a dodecaphonic patchwork. When listening to such
a composition, it often turns out to be very difficult—if not impossible—to
recognize the basic row and its transformations. The method has no determined
syntax, which would be required for tonal compositions.
In terms of denotators, a dodecaphonic row is a denotator r : Z11 → Z12
with address Z11 , where the images r(ei ) are given for the affine basis e0 =
(0, 0, . . . 0), e1 = (1, 0, 0, . . . 0), . . . e11 = (0, 0, . . . 0, 1). A sequence of d + 1 rows
would be a denotator in Zd @Z11 @Z12 , a hyperdenotator of address Zd . In the
generalized dodecaphonic composition method, the serial method, one has to
define such a sequence of rows, each row mapping into a set of parameters, such

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 67

pitch, duration, loudness, and attack in the case of a piano composition, with
d = 3. Pierre Boulez has constructed such a composition Structures pour deux
pianos around 1952.
In the tradition of dodecaphonic composition, they have tried to ex-
hibit rows, which have special properties. A first property is what is called
an all interval row. Such a row is characterized by the fact that the inter-
vals of successive row elements pi , pi+1 are all possible non-zero intervals, an
example is r = (1, 4, 11, 5, 10, 0, 9, 8, 6, 2, 3, 7), defining the interval sequence
3, 7, 6, 5, 2, 9, 11, 10, 8, 1, 4. Composer and theorist Herbert Eimert [7] was the
first to present the total collection of all 3, 856 all interval rows. Musically speak-
ing, an all interval row is generic since it generates all intervals. Many rows ex-
hibit a quasi-tonal infrastructure, e.g., r = (0, 3, 6, 9, 1, 4, 7, 10, 2, 5, 8, 11), where
the diminished seventh chord (0, 3, 6, 9) is represented three times in succession.
The idea to consider specific infrastructures in rows is a special case of the
fundamental approach to music objects, which S deals with coverings of musical
objects o by specific subobjects oi , i.e., o = i oi . Boulez in his seminal book
Musidenken heute [4, vol. 1] discusses such global objects with respect to their
musical signification. This perspective opens up a vast theoretical field of so-
called global compositions, which we will discuss in the following chapters, and
in particular addressing the classification topic for global compositions.
Part IV

Global Classification
17
Global Composition and Gesture Classification

Summary. This chapter opens up the concept of a global musical structure

with a number of typical examples from theory and composition. It turns out
that most situations in “musical structuralism” are essentially related to global
phenomena.
–Σ–

17.1 Why Global Compositions and Gestures

17.1.1 Global Compositions

Here we summarize the rationales for introducing global compositions. Intu-

itively speaking, global compositions are patchworks of local compositions, i.e.,
compositions that are not local compositions but compositions that are covered
by local compositions to generate structures that are only “locally” represented
by local compositions.

17.1.1.1 Boulez and Webern

A musical composition is never created as a local composition. In classical

European literature, e.g., for sonatas, string quartets, or symphonies, which
consist of 104 to 105 tone events, it would be impossible to start from such
large local compositions. Rather does the composer start from ‘small’ local
compositions, such as motives, themes, chords, rhythms, and similar elements
as a basis of a ‘creative combinatorics’ and then merges these parts by use of
various transformations to build a compound totality.
An excellent example is Boulez’ example of a dodecaphonic series in [138,
I]: He describes a series together with its internal structure, i.e., its composition
from partial series and their transformations, see Figure 17.1. Boulez’ discussion
refers to the local composition in ambient space of integer onset and common

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 71

pitch classes. We see local subcompositions A;A1;B;G;H of this row and some
transformations, drawn as subsets of points which are surrounded by rectangles,
defining a total of 15 partial series. These subcompositions are not disjoint in
general, they define a sophisticated covering of the given series.

( ).G
1 0
0 2
U.B1
b)
U.A1

A
U.A1 K.G
G
( ).U.A1
1 0
0 7

K.U.B1 B1 B

A1 H
C

( ).U.A1 = B1
1 0
0 8

Fig. 17.1: The covering of a dodecaphonic row by Boulez.

In his discussion of compositional principles of Schönberg, Berg, and We-

bern, the idea of hierarchies of local compositions is made explicit by Boulez
[138, I, p. 86]:
Bei Webern findet sich der Keim einer äusserst fruchtbaren Idee,
die die Reihe als Einigungsfaktor von Untergruppen und Obergruppen
betrachtet. Im Effekt sind alle isomorphen Figuren einer Grundstruk-
tur von der Tatsache abhängig, dass sie sich immer innerhalb der gle-
17.1 Why Global Compositions and Gestures 73

ichen Ordnung abwickeln, entsprechend den gegebenen Transpositionen

und Umkehrungen; sie integrieren sich in die Conditio sine qua non
der Zwölftonreihe: die chromatische Totalität. Diese isomorphen Fig-
uren bilden die Basis von privilegierten Mengen, die ihrerseits auf einer
höheren Ebene wiederum das vorstellen, was die isomorphen Figuren
selbst innerhalb der Reihe bedeuten. Durch Verkettung fügen sich Rei-
henformen, welche festgelegte Privilegien besitzen, zu einer Ganzheit,
einer in gewisser Weise “höheren Reihe” zusammen. Die Grundreihe
kann dann als strukturelle Kraft der Vermittlung zwischen Unter- und
Obergruppen betrachtet werden.
Boulez had learned from the second Viennese school that there is a strong
local-global principle in serial composition.

17.1.1.2 Uhde/Wieland and Marek

Performance is essential in the constitution of a musical work; this was already
pointed out by Adorno [1]:
Musikalische Interpretation ist der Vollzug, der als Synthesis die
Sprachähnlichkeit festhält und zugleich alles einzelne Sprach-ähnliche
tilgt. Darum gehört die Idee der Interpretation zur Musik selber und ist
ihr nicht akzidentiell.
According to Uhde’s and Wieland’s comment in [47], the meaning of “alles
einzelne Sprach-ähnliche” is a semantic moment the meaning of involved signs.
Therefore, music need not be decoded, but requires “imitation” of itself. Semi-
otically speaking, Uhde seems to refer to Jakobson’s poetical function (see [14])
as a projection of the paradigmatic axis to the syntagmatic axis of the sign sys-
tem. A semantic which doesn’t consider the poetical function is excluded. This
view of musical performance as a poetical oriented activity is essentially syn-
tactical articulation and paradigmatic intertwining. It establishes an articulated
whole, built from local and elementary parts of the composition.
In practical performance theory, e.g., in Ceslav Marek’s standard work
“Lehre des Klavierspiels” [24], the basic insights of Adorno and Uhde/Wieland
are realized in the artisanal details.
Here, the metrical grouping of the melody creates successive parts which
induce the shaping in performance by dynamical and phrasing (legato/staccato)
prescriptions. This level of grouping into local units is also essential for
mnemotechnical purposes and in order to give the fingering strategy a sup-
port. In fact, it would be bad piano playing to phrase against the fingering
strategy.

17.1.1.3 Hoffmann and Kaiser

After the foundation of music criticism by Mattheson, Rousseau, or Avison, to
name some of the important contributors, it was the merit of Ernst Theodor
74 17 Global Composition and Gesture Classification

Amadeus Hoffmann—above all in his famous review of Beethoven’s Fifth sym-

phony in the Leipzig-based “Allgemeine Musikalische Zeitung” in 1810—of hav-
ing given music criticism importance as a contribution to the aesthesic identi-
fication of the work. Since this achievement, music criticism was cultivated by
profiled critics, such as Eduard Hanslick or Joachim Kaiser, and has contributed
to the self-estimation of music and to the history of its reception.
It happens only as nonprofessional side-effect that this criticism is eager to
celebrate the “uniquely valid” interpretation and performance, or else to lament
its vanishing. Although Kaiser may be right in his Beethoven book [17] that
whereas it is a rare event in our days when a fundamentally new reading of the
32 Beethoven sonatas is presented, it is a never-ending objective in this field of
classical literature to set forth new approaches to the spiritual torso of these
sonatas, approaches which add to the given ones new perspectives, variations
of explanatory power. In the media and concert business, the music critics
should do exactly the job of commenting on this process and of comparing,
questioning and evaluating the singular approaches in the spirit of a work in
progress towards understanding an infinite evolution.
If the poietic work of the artist is expressed by articulation and correlation
of the composition’s local parts, then this a fortiori is the work of the critic—
only on the level of aesthesis. The overall impression of a performance integrates
knowledge, prejudice, and personal disposition, which may result in tiny local
effects on agogics, dynamics, articulation, and tuning. This is a central effort
to open the access of a broader public to the present work.
With the technology of saving works on LP, CD, MC and other media, the
articulated listening to music which was founded by music criticism has been
enriched by a new aspect: multiply repeated perception. Listening repeatedly
to a work on CD changes one’s articulation and grouping activity. Every new
listening changes or questions the relevant local compositions and their mu-
tual relations, some are eliminated, others are added. Successively, the listener
accumulates a patchwork of elements of comprehension.

17.1.1.4 Graeser, Ruwet, and Nattiez

Opposed to the perspectives of composers are positioned the musicologists
whose efforts for an adequate analysis are characterized by the need to re-
trace the composer’s thoughts. It is not astonishing that in the analysis of
musical works, organically composed hierarchies are common structures. They
start from small local elements, such as chords, degrees, tonalities, tonal func-
tions, voice leading, contrapuntal and harmonic progressions, and end up with
large local compositions, such as exposition, development, recapitulation, and
coda in the sonata form. This principle of hierarchical organisms was explicitly
evidenced in 1924 by Wolfgang Graeser in his analysis of Bach’s Kunst der Fuge
[10]. He describes a contrapuntal form as follows ([10, p.17]):
Bezeichnen wir die Zusammenfassung irgendwelcher Dinge zu ei-
nem Ganzen als eine Menge dieser Dinge und die Dinge selber als
17.1 Why Global Compositions and Gestures 75

Elemente der Menge, so bekommen wir etwa das folgende Bild einer
kontrapunktischen Form: eine kontrapunktische Form ist eine Menge
von Mengen von Mengen.
Das klingt etwas abstrus, wir wollen aber gleich sehen, was wir uns
darunter vorzustellen haben. Bauen wir einmal ein kontrapunktisches
Werk auf. Da haben wir zunächst ein Thema. Dies ist eine Zusammen-
fassung gewisser Töne, also eine Menge, deren Elemente Töne sind.
Aus diesem Thema bilden wir eine Durchführung in irgendeiner Form.
Immer wird dies Durchführung die Zusammenfassung gewisser The-
maeinsätze zu einem Ganzen sein, also eine Menge, deren Elemente
Themen sind. Da die Themen selber Mengen von Tönen sind, so ist
die Durchführung eine Menge von Mengen. Und eine kontrapunktische
Form, ein kontrapunktisches Musikstück ist die Zusammenfassung ge-
wisser Durchführungen zu einem Ganzen, also ein Menge, deren Ele-
mente Mengen von Mengen sind, wir können also sagen: eine Menge
von Mengen von Mengen.
The explicit reference to set theory is historically interesting since set theory
was a new language in mathematics in 1924. The text is somewhat misleading
since it suggests that tones are abstract objects. This is however not the case:
Graeser views tones as points in a geometric space, and he also recognizes the
role of symmetries, such as rotations and reflections, in such a space, transfor-
mations which may be applied to alter sets of tones or to compare different
tone-sets ([10, p.13]):
Gegenstand der Untersuchungen sind aber nicht die Töne selbst,
denn deren spezielle Beschaffenheit spielt gar keine Rolle, sondern die
Verknüpfungen und Verbindungen der Töne untereinander.
Es wird uns interessieren, ob wir gewisse Analogien zwischen den
Gebilden, die man in der Geometrie aus Punkten aufbaut, und unseren
aus Tönen hergestellten erkennen können.
Das wichtigste Grundprinzip der festen Körper, und die Geometrie
ist nichts anderes als das Studium der festen Körper, ist die Eigenschaft
der Symmetrie.
The Paris school of structuralist linguistics has applied results of semiology
after Saussure and Jakobson to musical analysis, above all in the investigations
of Nicolas Ruwet [42] and Jean-Jacques Nattiez [39]. The method of neutral
analysis which was developed by these authors starts from a hierarchical or-
dering into units and subunits which are a function of the given work. These
units are associated with each other by certain equivalence relations [39]:
Les divers unités ont entre elles des rapports d’équivalence de toutes
sortes, rapports qui peuvent unir, par exemple, des segments de longueur
inégale — tel segment apparaîtra comme une expansion, ou comme une
contraction, de tel autre — et aussi des segments empiétinant les uns
sur les autres.
76 17 Global Composition and Gesture Classification

The equivalence relations are not arbitrary but realized by specific transforma-
tions:
Les unités paradigmatiquement associées sont équivalentes d’un point
de vue donné (le thème paradigmatique), rarement identiques, et reliées
entre elles par des transformations qui décrivent les variants par rapport
a des invariants.
This language resembles the one which Graeser seems to aim at. However
the geometric aspect—embedding tones in spaces which admit symmetry
transformations—is more radical with Graeser, though less flexible regard-
ing the transformations which may be applied (Graeser limits his approach
to transformations of the “rigid” geometry, i.e., isometries).

17.1.1.5 Jackendoff and Lerdahl

Explicit grouping concepts are described by Ray Jackendoff and Fred Lerdahl in
[15]. Grouping is described from an aesthesic point of view of music psychology.
It deals with portions of notes which are heard as building an auditory unit. In
this approach, a group is always defined by a time interval, i.e., a group con-
sists of strictly time-adjacent notes and cannot be restricted to proper subsets
within time-slices, and which would be defined by voice splitting or parametric
splitting of pitch or loudness, for example. Also this grouping is a nearly perfect
hierarchy under inclusion: Overlapping neighboring groups may only contain
one common onset and are treated as very special situations in this theory. A
more general “web of motivic associations” would be beyond the theory because
it is not hierarchical [15, p.17].
This approach resembles Graeser’s concept of a contrapuntal structure
which is also a hierarchical grouping of parts, but Graeser’s idea was more gen-
eral insofar as it did not strictly ask for time slices. It is also less paradigmatic
than Graeser’s, Ruwet’s, and Nattiez’ because no significant statements are
made concerning the association of groups under symmetry transformations.
From the remarks in [15, p.286] it follows that these authors have no concept
of the transformation groups which may be adapted to specific contexts, as it
is explicitly proposed by Ruwet and Nattiez under the flag of “paradigmatic
theme” (see [39] for this concept). But it is precisely the psychological claim of
grouping—even in its strictly hierarchical appearance of the Jackendoff-Lerdahl
theory—as a cognitive basic which undermines the fact of global structures in
music.

17.1.1.6 Schaeffer and Cage

In the last years of the 1940s of the 20th century, the new technology of tape
music became a paradigm for local-global constructs. The American music for
tape movement was initiated by Otto Luening and Vladimir Ussachevsky and
17.1 Why Global Compositions and Gestures 77

applied by John Cage in the early fifties1 since his first music for tape Imagi-
nary landscape. At the same time, in Paris, Pierre Schaeffer and Pierre Henry
initiated the “musique concrète” movement, also based on tape as a flexible
medium of syntactical combination and recombination.
The new tape medium became a tool for concretely cutting and merging
time-slices of music. This was a starting point for an entire group of technolog-
ical realizations of local-global patchworks. At present, it is largely extended
and refined in various software for musical composition, notation, and postpro-
duction as compared to hard disk recording. The local parts are termed tracks,
parts, global and local scores, etc.
But the tape music movement had not only consequences for music soft-
ware, it also changed the concept of a score. In particular, Cage realized compo-
sitions, such as his concert for piano and orchestra (1957), where the modular
structure of the composition, showing a number of relatively autonomous local
parts, and an intended ambiguity of identifying such parts, became an explicit
and primordial feature of his poiesis.
In view of these rich traces of a local-global paradigm in music, Chris-
tian von Ehrenfels’ approach to gestalt which stresses super-summativity (in
fact a warmed-up version of the Aristotelian principle that “the whole is more
than the sum of its parts”) does not look very original. But it does provoke
the question how much more the added value exceeds the sum of the parts.
The above examples show that this may be a very complex question, in fact
the only interesting point of super-summativity. How is the whole constructed
from the collection of its parts? When are two wholes different, though having
identical parts? How can we compare wholes if we suppose that their parts are
comparable?

Fact 1 As long as no precise structure theory of wholes qua constructs from

local parts is available, no real understanding of music is possible. By the arsenal
of the preceding examples, this is a problem which touches all levels of the
communicative axis. And it is a problem which involves interpretative activity
and its innate ambiguities.

17.1.1.7 Musical and Mathematical Manifolds

In history of science, the mathematical and musicological concepts of global

gestalts is situated around 1854 when Eduard Hanslick (Figure 17.3) published
his famous treatise “Vom Musikalisch Schönen” [11]. Musical content was rec-
ognized as being “tönend bewegte Formen”. Hanslick added that these forms
are by no means elementary but composed in an artistic way, and building a
unity within the manifold, as restated by music theorist Hugo Riemann.
1
In the Project of Music for Magnetic Tape, together with Morton Feldman, David
Tudor, and Christian Wolff.
78 17 Global Composition and Gesture Classification

Fig. 17.2: Bernhard Riemann in- Fig. 17.3: Eduard Hanslick de-
troduced global structures (mani- scribed music as a compound
folds) as compound mathematical structure, built artistically from
spaces. parts.

It is remarkable that in the same year, the mathematician Bernhard Rie-

mann (Figure 17.2) conceived a far reaching generalization of the mathemati-
cal concept of space to so-called manifolds [40]. These are understood as being
patchworks of locally cartesian charts, similar to geographic atlases.
The shared conceptual basis of these approaches is that locally trivial
structures can add up to aesthetically valid configurations if united in a non-
trivial way. A simple and well-known example of such a global shape is the
Möbius strip. Its fascination stems from gluing together the ends of an ordinary
‘belt’-shaped strip after a rotation of half the full circle of one end. For example,
in musicology, the Möbius strip is realized by the harmonic strip of triadic
degrees within a diatonic scale, we shall discuss this item below.
The basic difference between musical and mathematical manifolds is that
musical manifolds are defined with a fixed atlas, whereas mathematical man-
ifolds are not tied to fixed atlases. Intuitively speaking, the sphere of our
globe, viewed as a mathematical manifold, is the same if we add new small
or large charts as long as they are compatible with the given ones. If I add a
regional map, nobody will complain that the geographic identity of the globe
has changed. Mathematically speaking, we may go to the colimit of all atlases.
But in music, this is completely different: A given covering of a com-
position by a determined set of charts, such as chords, motives, or periods, is
essential for the identification of the composition, two different coverings change
the composition qua global structure or gestalt, to use Christian von Ehrenfels’
concept. Mathematically speaking, the colimit of charts (going to all possi-
17.1 Why Global Compositions and Gestures 79

ble additional local perspectives) is not allowed, the individual interpretational

activity is an integral part of the object’s identity.

17.1.1.8 Global Gestures

In practice, gestures are not compactly defined, but reveal a union of local
gestures, which coincide on determined subgestures. A typical example would
be the dancing gesture, where two persons are represented by their local gesture,
but their holding hands share a common subgestural configuration, see Figure
17.4.

Fig. 17.4: Two dancers holding each other by their hands.

17.1.1.9 The Definition of Global Compositions and Global

Gestures

Definition 6 A (global) objective composition is defined by the following data:

(i) A set G and a finite, non-empty covering I of G by non-emty subsets,
(ii) a family (Kt , Ft )t∈T of local objective compositions,
(iii) a surjection I? : T → I : t 7→ It ,
∼
(iv) a bijection φt : Kt → It for each t ∈ T ,
(v) for each couple s, t ∈ T such that Is ∩ It 6= ∅, the induced bijection

φs,t := φ−1 −1 −1
t ◦ φs : φs (Is ∩ It ) → φt (Is ∩ It )
80 17 Global Composition and Gesture Classification

(restricted to the respective domains and codomains) defines an isomor-

phism
∼ −1
φs,t /1 : (φ−1
s (Is ∩ It ), Fs ) → (φt (Is ∩ It ), Ft )

of local compositions (we only discuss zero-addressed compositions here, but

a general approach is feasible, see [28, vol. 1]).
The data (iii) to (v) are called an A-addressed atlas Φ for the covering I of
G. The bijections φt (or—by abuse of language—the local compositions Kt ) are
called the charts of the atlas Φ.
Two A-addressed atlases Φ, Ψ for the covering I of G are called equivalent
iff their disjoint sum Φ q Ψ is an atlas for the covering I of G. An objective
global composition is a covering I of G (often abbreviated by GI ) together with
an equivalence class of atlases for GI . If no confusion is likely, we may abbre-
viate the entire data by saying that we are given an objective global composition
G, i.e., by just naming the support set G of the objective global composition.
The surjection I? refers to the option to have several charts for the same
local composition part.
Global gestures are defined in the same sprit as follow, similar to the
definition of global compositions:

Definition 7 A global gesture consists of these components:

1. a digraph Γ , the support, and a finite covering I of Γ by non-empty sub-
digraphs,
2. a surjection I? : T → I : t 7→ It ,
−
→
3. a family (kt : It → Kt )T of local gestures,
4. for each couple s, t ∈ I such that Is ∩ It 6= ∅, the identity Ids,t of digraphs
Is ∩ It extends to an isomorphism fs,t of local gestures,
∼
fs,t : ks |Is ∩ It → kt |Is ∩ It .

of local gestures.
The data in 2. and 3. are called an atlas of the global gesture, which is often
denoted by Γ I . The local gestures kt are called the charts of the atlas. Two
atlases are equivalent if their disjoint union is also an atlas for Γ I . The global
gesture is defined by an equivalence class of atlases.

17.1.1.10 Examples of Global Compositions and Gestures

Global compositions and gestures may be derived from corresponding local

→
−
compositions K ⊂ M and gestures ∆ → X by a simple covering of the ambient
spaces, i.e., of K or of ∆. Given such a covering K I or ∆J , for each subset
Ii ∈ I, we get a restricted local composition Ki ⊂ M , and for each subset
→
−
Jj ∈ J, we get a restricted local gesture ∆j → X . The non-empty intersections
17.2 Classification through Modules of Affine Functions 81

in both cases are connected by the identities of the ambient spaces. Such global
structures are called interpretations of the given local structures.
It is remarkable that not every global composition is isomorphic to the
interpretation of a local composition. Here is such a “pathological” example.
It means that this case cannot be played in a musical performance since there
is not ambient module of musical parameters, where the performance could
take place. Nevertheless, non-interpretable global compositions are conceptu-
ally important as “germs” of interpretable global compositions. We shall show
in Section 17.2.3 how this can be achieved.

17.1.2 The Nerve of a Global Composition or Gesture

If GI or Γ I is a global composition or gesture, its nerve n(GI ) or n(Γ I ) is a

subset of the powerset of covering index set I defined by the following proper-
ties:
1. Every element i of I has the singleton {i} in the nerve.
T a finite set of chartes σ = {i1 , i2 , . . . ik } ⊂ I has a non-empty intersection
2. If
σ 6= ∅, then σ is in the nerve.
The elements σ of the nerve with k elements are called k − 1-simplices. So the
0-simplices are the charts. The nerve is also called the simplicial complex of
the global structure. One may represent the nerve by writing the 0-simplices as
points in Rn , the 1-simplices as straight lines connecting their two 0-simplices,
the 2-simplices as triangular surfaces limited by their 1-simplices lines, the 3-
simplices as full tetrahedra limited by their 2-simplices triangles. The construc-
tion goes on for higher dimensions, but these cannot be visualized in 3-space.
See [27, Ch. H.2.1] for a complete mathematical definition.
A classical example is the global composition of a triadic interpretation
X (3) of a major scale X defined by its seven triadic degrees as charts. Its nerve
is a Möbius strip, as shown in Figure 5.1.

17.2 Classification through Modules of Affine Functions

The general classification of global compositions and gestures uses functions on
such global structures to describe them in the context of function spaces and
their transformations by natural group operations on standard structures.

17.2.1 Local Compositions and Functions

We describe the classification method of global compositions using affine func-

tion modules. The idea is a development starting from the tools, which were
used in classification of local compositions. Let us look first at the local situa-
tion. We may first consider the so-called standard local composition of dimen-
sion n, ∆n ⊂ Rn consisting of the n + 1 points e0 = 0, e1 = (1, 0 . . . 0), . . . en =
82 17 Global Composition and Gesture Classification

(0, . . . , 0, 1). For every local n + 1-element composition K ⊂ M , we have a

one-to-one morphism f : ∆n → K sending e0 to k0 , etc., and en to kn , for an
enumeration K = {k0 , . . . kn } of K’s elements. Recall that we used such a func-
tion in local classification, when considering the module morphism Rn → M .
This map f can be used to induce an injective linear map K@R → ∆n @R,
sending an affine function g ∈ K@R to the composed function g ◦ f . Here,
K@R denotes the R-module (addition and scalar multiplication element-wise)
of local composition morphisms from K to the full local composition R ⊂ R.
This identifies the module of affine functions K@R with a submodule of the
module of affine functions ∆n @R.
This affine function methods will be sufficient to create a tool for global
classification. The idea is to classify the affine functions qua functions on the
standard composition and to generate orbits of such function sets under the
action of an adequate permutation group.
If we take a function module K 0 ⊂ K@R ← M @R, we may create a
local composition K 0∗ ⊂ K 0 @R by sending the base elements ei of ∆n to
their image ei∗ in K 0 @R via e∗i (x) = x(ei ) = x(f (ei )), x ∈ K 0 . If for any
two different ei , ej there is a function x that has different images, we call K 0
separating. In this case, the new local composition K 0∗ is the bijective image
of the standard composition ∆n . This new K 0∗ has been constructed from
the module K 0 ⊂ ∆n @R of affine functions, and if we can show that this new
composition is also isomorphic to the initial K, we may think of classification via
modules of affine functions over the standard local composition. Classification
using this method can be achieved if the ambient module M of K ⊂ M is finitely
generated projective, i.e., if we M is a direct summand of a free module Rm .
In this case, the coordinate functions pi : Rm → R yield separating candidates.
Such local compositions are separating, and K 0∗ is in bijection with the original
K. The bijection K → K 0∗ is defined by sending k ∈ K to the evident k ∗ ∈
K 0 @R ⊂ M ∗ @R, M ∗ = Hom(M, R).

17.2.2 Morphisms and Function Spaces

We should add to our previous constructions of local compositions and gestures

from function spaces the behavior of morphisms with respect to this function
space approach.
Let us first consider a morphism f : K → L of local compositions
(K, M ), (L, N ) that is (by definition) induced by an affine module morphism
h : M → N . We have this commutative diagram of local compositions:
17.2 Classification through Modules of Affine Functions 83
∼
→
K - K 0∗

f
∗

-
-
∼
→
L - L0∗

? ∼ ?
→
M - M ∗∗

-
-

? ∼ ?
→
N - N ∗∗
with bijections K → K 0∗ , L → L0∗ , where K 0∗ is defined by the module of
affine functions that are restrictions of linear functions M → R, where L0∗ is
defined by the module of affine functions that are restrictions of linear functions
N → R, and where the codomain local compositions are derived from the above
affine function modules within the standard local compositions and ∆m , ∆n if
m = card(K), n = card(L). In other words, the map K 7→ K 0∗ is a natural
transformation, bijective for separating local compositions.
In the special case, where M is projective and finitely generated, the
bidual map M → M ∗∗ , M ∗∗ the linear bidual of M , is an isomorphism, and
the morphism K → K 0∗ is an isomorphism. In this situation (ambient modules
finitely generated projective) the natural transformation K 7→ K 0∗ can be
used to generate an isomorphism of a global compositions between the given
global composition and a global composition that is derived from affine function
modules.
If in a global composition GI , we have two charts with local compo-
sitions K, L, such that the intersections (subcompositions) are isomorphic,
∼
K|L → L|K, then the bidual of this configuration for finitely generated projec-
tive ambient modules produces intersections (K|L)0∗ ⊂ K 0∗ and (L|K)0∗ ⊂ L0∗
also being isomorphic. In other words, the global composition GI produces an
isomorphic bidual global composition G∗I via the bidual natural transformation.
And the bidual global composition is generated by affine function modules. The
next step deals with the following global affine functions on a global standard
composition.

17.2.3 Global Standard Structures and Resolutions for

Compositions
For a covering GI of a set G by non-empty subsets Gi , i ∈ I, we may cre-
ate a global composition ∆GI together with an obvious bijective morphism
84 17 Global Composition and Gesture Classification

res : ∆GI → GI by the following method. Take the vector space Rn , where G is
supposed to consist of n+1 elements. Take the global composition ∆GI that con-
sist of all n + 1 affine basis vectors e0 = 0, e1 = (1, 0, . . . 0), . . . en = (0, . . . 0, 1),
take any bijection of this basis with G, and define the charts ∆i of ∆GI as being
the inverse image of the charts Gi . This defines an evident bijective morphism
res : ∆GI → GI , and we call this the resolution of GI .
18
The Classification Theorem for Global
Compositions

Summary. This theorem is very involved conceptually and mathematically.

We first develop the conceptual framework, which is based on the construction
of the resolution of a global composition. Every global composition turns out
to be a kind of projection of its resolution. In this context, we discuss sys-
tems of functions on resolutions, and we show, how these function systems are
used to classify global compositions. From this theory we derive the general
classification theorem of global compositions
–Σ–

18.1 Preliminary Remarks

18.1.1 Characterization of Interpretations
To find out whether a global composition is interpretable, i.e., isomorphic to
an interpretation, one has to consider affine functions on global compositions.
An affine function on a global composition GI is a function f : G → R, R being
the underlying ring of its charts. Such a function is, by definition, such that all
of its restrictions to the composition’s charts are affine functions. In general, it
is not true that every affine function f : Gi → R on a chart Gi can be extended
to an affine function on G. If every chart affine function can be extended to all
of G, the global composition is called flasque. It can be proved that

Theorem 6 A global composition G is interpretable, iff all its affine functions

on charts can be extended to functions on all of G, i.e., iff it is flasque.
See [27, Ch. 16] for a proof.

18.1.2 An Example of a Non-interpretable Global Composition

The theory of global structures exhibits global structures that are not inter-
pretable, i.e., they are not isomorphic to interpretations.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 85

An example of a non-interpretable global composition is shown in Figure

18.2.

Example 1 We discuss a non-interpretable global composition GI in the

ˇ “*
zero address, and over the field of real numbers, see Figure 18.2. It has six
points,
G = {x1 , x2 , x3 , x4 , x5 , x6 },
and three charts for its covering, i.e., I = {U1 , U2 , U3 } with

U1 = {x1 , x2 , x3 , x4 },
U2 = {x1 , x2 , x5 , x6 },
U3 = {x3 , x4 , x5 , x6 }.

Fig. 18.1: A non-interpretable composition together with its resolution.

The three covering charts are all in bijection with a ‘square vertex’ com-
position S = {a = (0, 0), b = (1, 0), c = (1, 1), d = (0, 1)} ⊂ 0@R R2 as follows:
18.2 Non-Interpretable Compositions 87
∼
U1 → S : x1 7→ a, x2 7→ d, x3 →7 b, x4 7→ c,
∼
U2 → S : x1 7→ a, x2 7→ d, x5 7→ c, x6 7→ b,
∼
U3 → S : x3 7→ b, x4 7→ c, x5 7→ a, x6 7→ d,

so that the intersections of any two of these charts yield an isomorphism of

a two-point composition in R2 . If we visualize this configuration by means of
a subdivision of each chart as a union of two triangular surfaces, we obtain
a Moebius strip, see Figure 18.2. The resolution of GI in the same surface
representation is also shown. Here, the charts are no longer plane compositions
but tetrahedra in three space.
Let us show that GI is not interpretable. It is evidently sufficient to see
that the global affine functions on this composition do not separate points.
Let f : G → R be an affine function (we suppress the zero address and just
work in the respective ambient spaces), set f (xi ) = fi for i = 1, 2, 3, 4, 5, 6 and
fi,j = fi − fj . Since f is affine on each chart, we have on chart U1 : f1,2 = f4,3 ,
on chart U2 : f1,2 = f5,6 , on chart U3 : f4,3 = f6,5 . Therefore f5,6 = f6,5 , and
f5,6 + f6,5 = 0, by definition, so 0 = f1,2 = f4,3 = f6,5 , and no one of the pairs
(x1 , x2 ), (x3 , x4 ), (x5 , x6 ) can be separated by f .

18.2 The Musical Meaning of Non-Interpretable

Compositions: Varèse’s Program and Yoneda’s Lemma
Summary. A non-interpretable global composition cannot be “played” in a
score since there are not enough affine functions to describe the composition’s
coordinates, a condition that is essential for a composition to be represented
as a set of points with coordinates in a “score space”.
Implicitly, the part of Yoneda’s Lemma dealing with variations of perspec-
tives is of primordial importance in compositional concepts. In Edgar Varèse’s
programmatic writings, a thoroughly geometric approach to the Yoneda phi-
losophy is sketched. It accomplishes the classical variational principle in com-
position; we give an overview to this central connection between modern math-
ematics and music.
–Σ–
Let us first recapitulate the impact of the resolution of a (commutative)
global composition on the esthetics of music. The resolution ∆GI of a global
composition GI generates points which are in general position in every chart
of the atlas. The resolution is interpretable, and GI can be rebuilt from the
retracted function module in the resolution.
Since the resolution’s nerve is isomorphic to the original nerve, and since
the resolution projects bijectively onto the original composition, no note event
and no overlapping relation of charts is destroyed in the resolution. So we virtu-
ally have the same note set, but they are enriched by a collection of parameters
88 18 The Classification Theorem for Global Compositions

Fig. 18.2: Edgar Varèse (1883 – 1965).

which allow us to place these events in optimal relative position. So the resolu-
tion can—in principle—be performed by physical instruments, and it can—in
principle—be played such that the old idea may be heard since old parameters
are preserved.
The freedom of choice for an optimal realization of the resolution (which
after all is only determined up to isomorphism) is also mandatory since a good
physical performance has to meet a number of additional conditions of human
cognition. The auditory system, the instrumental skill of an artist, the mate-
rial possibilities, the time frame at disposition, etc., all these conditions impose
serious value boundaries on the allowed parameter values which are acceptable
in a good realization. Typically, these parameters must reflect the syntactical
structure of the composition, and not only the resolution’s general position con-
text. So time must be given a crucial role in the parametrization of events. And
the distinction of events must also be optimized when the temporal unfolding
of a performance represents a good communication stream. Nonetheless, the
resolution classification technique yields very important necessary conditions
for a comprehensible performative parameter setting.
What could now be the program of classification? Its core objective is that
it deals with understanding musical works. And we should stress that our con-
cept of a musical work is not the narrow one which restricts to those individual
opera which—especially in Europe—started to flourish in the Renaissance. It
includes as well general musical corpora such as scales, systems, everything that
can be represented by means of global compositions, and—in the limit—any
denotator if we admit the most general topos of this theory.
18.2 Non-Interpretable Compositions 89

From the precise parametric description of a work and its ambiguities,

this work appears as a point configuration in a more or less complex space (or
form, if one prefers the denotator terminology). However this configuration is
already a defined perspective which enables many relations among its ingredi-
ents. It is the composer’s perspective (now including an abstract ‘composer’ or
creator of a general musical structure like a scale). For example, the choice of
tonality, instrumentation, tempo, etc., are points of view which may or may
not pertain to the composition, this is a question of the epoch of creation. But
their character can undoubtedly be subject to variation. Among others, here
we do address the question of historical instrumentation for early music.
In order to understand the relations among different parts of a compo-
sition, and even to simply recognize them, a change of the given perspective
is mandatory. If a never seen object must be inspected, what should we do?
You walk around it. This is the most common version of Yoneda’s Lemma.
The analogy to cartography is straightforward: The natural perspective of the
landscape in which we live does not coincide with the perspective which best
meets our need for orientation. To reach this end, we build maps which show
the landscape from an infinitely far vertical viewpoint.
The same is the case in music. You play a piece in slow motion ‘from very
near’, in a zoomed perspective, a complex chord is arpeggiated, i.e., viewed
from a skew angle, and so forth. This idea of variation of the perspective has
also been integrated in the compositional thinking of the 20th century. We want
to illustrate this remarkable fact by a citation from Edgar Varèse’s comments
on his composition “Intégrales” [49, p.67]:
Die Intégrales wurden für eine räumliche Projektion entworfen. (...)
Während wir in unserem musikalischen System Klänge anordnen, deren
Werte festgelegt sind, suchte ich eine Verwirklichung, bei der die Werte
fortwährend im Verhältnis zu einer Konstanten verändert werden. (...)
Um dies besser zu begreifen, übertragen wir, da das Auge viel schnel-
ler und geübter ist als das Ohr, diese Vorstellungen ins Optische und
betrachten die wechselnde Projektion einer geometrischen Figur auf ei-
ne Fläche, wobei Figur und Fläche sich beide im Raum bewegen, aber
jede nach ihren eigenen Geschwindigkeiten, die veränderlich und ver-
schieden sind, die sich verschieben und rotieren. Die augenblickliche
Form der Projektion ist durch die Relation zwischen Figur und Fläche
in diesem Augenblick bestimmt. Aber wenn man erlaubt, daß die Figur
und die Fläche ihre eigenen Bewegungen haben, ist es möglich, mit der
Projektion ein äußerst komplexes und scheinbar unvorhersehbares Bild
zu erhalten. Diese Qualitäten können noch vermehrt werden, wenn man
die Form der geometrischen Figur ebenso wie ihre Geschwindigkeiten
variiert. (...) Ich hoffe, innerhalb kurzer Zeit einen Apparat zur Ver-
fügung zu haben, der es erlauben wird, ein räumliches Relief zu geben.
Nur des Beweises wegen würde ich daran interessiert sein, die Intégra-
90 18 The Classification Theorem for Global Compositions

les eimal so zu realisieren, wie sie ursprünglich konzipiert worden sind.

In 1960 Maurizio Kagel transferred these principles on paper strips and

discs of the score for Transición II for piano, percussion, and two tapes.
Varèse’s idea basically is a remake of the classical variation princi-
ple. Bach’s Goldberg Variations (BWV988), Beethoven’s Diabelli Variations
(op.120), or Webern’s Variationen für Klavier (op.27) are compositions that
follow this spirit. The subject is always an artistically interwoven change of
perspectives of a theme: the variation in the parts of the theme and their rela-
tions.

Principle 1 Variation as a principle of musical shaping is nothing else than

the identification of an idea—such as the theme—as a integral of its perspec-
tives.
For Webern a composition is a cellular organism, a connected manifold of trans-
formations, of ever changing perspectives, of metamorphoses of a single cellular
germ (in the sense of Goethe), which in fact is Schönberg’s dodecaphonic rows.
In front of this historical background the classification problem of global
compositions—together with its central process of resolution—appears as a
canonical program. In particular, the nerve of a resolution, a concept related
to that of a “cell complex” from algebraic topology, reminds us of the cellu-
lar organism alluded to by Webern. The projections described by Varèse in a
visionary fashion show a surprisingly similar geometry to the projections of a
resolution onto the original composition, projections which are distinguished in
that they project a general position onto specializations. Finally, the variation
of these projections corresponds to the variation of the modules of affine func-
tions, i.e., the variation of the compositions which are distinguished from each
other via their retracted function modules on one and the same resolution.
The variational principle is not only a compositional strategy, it even
more dramatically applies to the level of performance. Performance deals with
a transformation from the mental score space to the physical space of an acous-
tic realization. But this transformation locally is a deformation of the “rigid”
parameter values set out on the score. Why should the artist deform a perfect
opus? Wouldn’t this be blasphemy or at least a lack of respect? No, the added
value of such a deformation is not a destruction of given structure, it is a sub-
tle change of parametric perspectives which let the auditory still recognize the
written relations, but on top of that puts configurations into general position
such that their generic, or better: resolved, structure becomes ‘visible’ on the
auditory level—to restate it in the wording of Varèse.

Principle 2 So the structural rationale of performance is a strategy of small

changes of the composer’s perspective to make the resolution of the composition
audible and thereby ease understanding of the underlying composition class (in
the strict sense of classification).
18.3 Classification of Global Compositions 91

18.2.1 Global Dodecaphonic Classes

Dodecaphonic rows are also interpreted by “Binnenstrukturen” (“inland struc-

tures”), especially in Boulez’ theory.

18.3 Classification of Global Compositions

The classification of global compositions is based upon a special type of global
compositions, global standard compositions. These objects are “free” global
compositions, they are always interpretable. And every global composition is
a kind of image of a global standard composition, which is called the global
composition’s resolution.
The information about a global composition may be “retracted” into its
resolution, where it is represented as a module of affine functions. The represen-
tation can be used to calculate all isomorphism classes of global compositions
qua points in adequate algebraic varieties (schemes). Technically speaking, this
classification method is quite demanding, so we shall limit our presentation to
some key points without delving into mathematical subtleties. We refer to [28,
Vol.I, Ch. 15] for a thorough trajectory.

18.3.1 The Resolution of a Global Composition

The resolution of a global composition GI runs as follows. We first introduce

the standard global composition ∆GI . We suppose that G consists of n + 1
points, and that its charts Gi , i ∈ I have ni points. Then we look at the local
composition ∆n ⊂ Rn that consists of the (affine) base points e0 = 0, e1 =
(1, 0, . . . 0), . . . en = (0, . . . 0, 1) ∈ Rn . We now enumerate the points of GI
and associate the chart points of GI with the n + 1 base points of ∆. This
defines a covering of ∆ by corresponding charts ∆i , each being defined by the
corresponding points of the GI charts. This global composition ∆GI evidently
interpretable, and we get a bijective morphism p : ∆GI → GI , this is the
resolution of GI .
The idea now is to represent GI by its affine functions, which are repre-
sented on ∆GI as follows. We first consider the affine functions on GI . This
means that on each chart Gi of GI , we take the R-module of affine functions
a : Gi → R, i.e., the set Γi = Gi @R. We also look at the affine function
modules Γσ = Gσ @R, where Gσ denotes the subcomposition for a non-empty
intersection σ of a number of charts. We have evident restriction module ho-
momorphisms Γσ → Γτ whenever you σ ⊂ τ . De note this system of functions
by N (GI ), and call it a function complex. The bijection (not an isomorphism
in general!) p induces an injection N (GI ) → N (∆GI ). It is this injection of
functions that we will use to reconstruct GI .
The complex N (GI ) is used to generate a new global composition. On the
subompositions σ, we consider the liner dual modules N (GI )(σ)∗ and within
92 18 The Classification Theorem for Global Compositions

these modules, we define points by taking base points ei in σ and then consider
their duals ei∗ ∈ N (GI )(σ)∗ , sending f ∈ N (GI )(σ) to the e∗i (f ) = f (ei ),
a construction, which we discussed in Section 17.2.1. To reconstruct GI , we
must be assured that the ei∗ , e∗j are different if you i 6= j. This is guaranteed
if there are ‘enough’ functions in every N (GI )(σ), i.e., if there is a function
f such that ei (f ) 6= ej (f ). In this case, we call N (GI ) separating. It can be
shown (referring to the discussion in Section 17.2.1) that the complex N (GI ) is
separating whenever the carrier modules of GI are finitely generated projective.
Under this latter hypothesis, we now constructed a global composition built
by the ei∗ , which are defined over function modules N (GI )(σ). This global
composition, which we denote by ∆/GI is isomorphic to the original GI , when
sending its points to the corresponding e∗i under the given bijection p, we omit
the very technical proof of this fact and refer to the technical details in [28,
Vol.I, Ch. 15].
This construction may be embedded in a commutative triangle
∆GI

(18.1)
p
-

∆/GI - GI
∼
→

Moreover, if we are given a morphism F : GI → H J , this one extends to the

above triangle:
∆ GI

pG
-

-
∆F ∆/GI ∼
- GI
→

?
∆H J F

pH
-

? - ?
∆/H J ∼
- HJ
→

This result means that we may reconstruct separating global compositions

from their complexes of affine functions, as symbolized by ∆/GI . Moreover,
18.3 Classification of Global Compositions 93

these complexes are defined on the standard compositions, and one can prove
that isomorphic global compositions are defined by isomorphisms of function
complexes that are derived from permutations of the standard composition’s
points. In other words:
Isomorphism classes of global compositions are defined by the permutation
orbits (on standard global compositions) of separating function complexes.
Similar to the classification theorem of local compositions, see Section
8.1.4, we have this classification theorem:

Theorem 7 The isomorphism classes of global compositions over finitely gen-

erated projective modules can be parametrized as rational points on schemes of
finite type.
The technically precise content of this theorem is displayed in [28, Section
15.3.2, Theorem 18], but we cannot discuss this level because it requires some
higher algebraic geometry. Nevertheless, our presentation captures the main
ideas leading to this result.
19
The Classification Problem of Global Gestures

Summary. According to the description of local gestures using function spaces

as described in Section 19.1, it seems reasonable to go parallel to the classi-
fication of global compositions. We may consider the system (a complex) of
functions Im(|gi |)∗ ⊂ |∆i |∗ and then step over to the orbits of such function
complexes according to the action of the finite permutation group of an under-
lying global digraph. However, as already shown for the local gesture situation,
classification of gestures is not feasible yet. Our discussion will therefore focus
on techniques that could help classify gestures in future developments.
–Σ–

19.1 Local Gesture and Functions

The method of functions for compositions can also be applied to local gestures
→
−
g : ∆ → X . The point here is to find continuous functions f : Im(g) → D,
where Im(g) is the topological subspace of X of the curves of g, with a codomain
topological space D such that it separates points in Im(g). In this case, we may
proceed for local gestures in a similar way as with local compositions: Take
the associated continuous map |g| : |∆| → Im(g) ⊂ X and then the inverse
image of Im(g)∗ = Im(g)@D in |∆|∗ = |∆|@D. We get a canonical continuous
map Im(g) → Im(g)∗∗ and its extension to X → X ∗∗ . We now get a gesture
g ∗∗ : |∆| → Im(g)∗∗ ⊂ X ∗∗ , the values x ∈ |∆| being sent to g ∗∗ (x) : h 7→ h(x).
We get a commutative diagram for gestures as follows. Take a morphism
of gestures:
g →
−
∆ −−−−→ X
 
→
y−f

ty

h →
−
Γ −−−−→ Y
which corresponds to a commutative diagram of topological spaces:

|g|
|∆| −−−−→ X
 
|t|y
 
yf
|h|
|Γ | −−−−→ Y
Taking the “bidual” |∆|∗∗ = (|∆|@T op D)@T op D etc., with a suitable codomain
topological space D, we get a commutative diagram
|g|∗∗
|∆|∗∗ −−−−→ X ∗∗
 
|t|∗∗ y
 f ∗∗
y
|h|∗∗
|Γ |∗∗ −−−−→ Y ∗∗
This induces the following commutative diagram for gestures, analogous to the
previous one for local compositions, and where Imageg , Imageh are the images
in the biduals, respectively:
|g| ⊂
|∆| - Im(|g|) - X
bi

f
j.

-
-

⊂
|t| f Imageg - X ∗∗

? ? ?
|h| ⊂
|Γ | - Im(|h|) - Y f ∗∗
bi
j.

-
-

? ⊂
?
Imageh - Y ∗∗

The critical difference with respect to local compositions is that there is no

analogy to finitely generated projective modules in topology, i.e., there is no
chance to have the biduals being isomorphic to the original objects.

19.1.1 Restriction of Gesture Types

We may try to choose a codomain D that would work for a good selection of
gestures. The topological problem is that there are too many fundamentally
different topologies to enable a generic approach. For example, one may con-
sider topologies deduced from metrics, which are very different from topologies
deduced from algebraic geometry, such as Zariski topologies, where points are
19.1 Local Gesture and Functions 97

not comparable, there are thick and thin points, a situation that would not
occur for metrically defined topologies.
We therefore decide to consider the following framework that is fre-
quently encountered in gestural practice. We first suppose that the topologi-
cal spaces, where gestures are projected, are compact sets within Rn , which
means that they are closed and limited. We then consider the codomain
D = [0, 1] = I, the real unit interval, which is also compact. Under these
conditions, the space X@Top D is compact metric with the distance function
d(ξ, η) = maxx∈X ξ(x) − η(x). The same construction yields a compact metric
space X ∗∗ = (X@Top D)@Top D.
Let us now look at the bidual map q : Im(|g|) → Im(|g|)∗∗ for a gesture
g, and Im(|g|)∗∗ being the image of Im(|g|) in X ∗∗ . To begin with, this map is
continuous. In fact for x, y ∈ Im(|g|), d(x∗∗ , y ∗∗ ) = maxf ∈X ∗ x∗∗ (f ) − y ∗∗ (f ) =
maxf ∈X ∗ f (x) − f (y). If x and y are near to each other in the given metric, the
distance d(x∗∗ , y ∗∗ ) is small since all f are continuous, and by compactness we
may choose a finite number of such f . Conversely, the map is also open, i.e.,
the inverse is continuous. This follows from the observation that if d(x∗∗ , y ∗∗ ) is
small, then all function distances f (x)−f (y) are small, especially the coordinate
functions of x, y must have a small difference. This means that we have a
∼
homeomorphism q : Im(|g|) → Im(q) ⊂ Im(|g|)∗∗ .
The function q can also be obtained from the embedding |g|∗ : Im(|g|)∗ →
|∆| by stepping over to the dual of this map, and then composing it with the
∗

embedding |∆| → |∆|∗∗ according to the following commutative diagram.

|g| surjective
|∆| −−−−−−−−→ Im(|g|)
 
yq
 
y
|g|∗∗
|∆|∗∗ −−−−→ Im(|g|∗∗

19.1.2 Global Standard Structures and Resolutions for Gestures

To construct a global standard gesture ∆I for a global gesture DI , we consider
the standard topological space |∆| deduced from the digraph ∆ of DI , its parts
are the unit interval lines [0, 1] ⊂ R, one for each arrow of ∆, and glued together
(the colimit) on the shared vertices. This topological space is the union of its
subspaces that are defined by the global gesture’s charts and |∆i |, i ∈ I. For
−−→
every chart ∆i , we have the map ∆i → |∆i | defining the ith chart of the global
gesture ∆I . This is the global standard gesture for digraph’s ∆ covering I. For
a global gesture DI on ∆ with a covering I, we obtain the evident morphism
∆I → DI sending the standard charts on ∆i to the charts Di of DI . We call
this morphism to resDI : ∆I → DI the resolution of the global gesture DI .
The continuous functions in Im(g)@D should have a codomain that is
good enough as a separator for all gestures. To this end one could take a set D
with cardinality larger than all relevant domains and take the indiscrete topol-
ogy, further take as Im(g)@D all constant functions, which are automatically
98 19 The Classification Problem of Global Gestures

continuous and thereby separate points of the domain. The problem here is that
the bidual Im(g)∗∗ will not be homeomorphic to Im(g). This is a significant
difference to the compositions’ situation, where the bidual for “good” modules
is isomorphic to the original.

19.2 A Conjectured Classification Theorem

So far the classification runs parallel to the business for global compositions.
But the next step: viewing such function complexes as module complexes, fails.
The function spaces |∆i |∗ are not algebraic structures, they are compact topo-
logical spaces without algebraic specifications. A fortiori, the Grassmann for-
malism does not apply. And we do not recognize an analog mechanism in the
topological realm. This suggests/conjectures

Theorem 8 (Conjecture) The global gestures over real numbers as described

above can be classified by finite group orbits of topological spaces of continuous
functions on standard spaces |∆| together with the orbits of colimit homeomor-
phism actions on the topological carrier spaces.
These orbits are not recognized as being points on manifolds analog to the
Grassmann schemes used in the classification of global compositions.

19.2.1 A Hypergesture for Human Bodies

In this Section, we want to develop the global hypergesture for a human body.
We construct it as a hypergesture of lines of loops. Figures 19.1 and 19.2 show
the idea. Figures 19.3 and 19.4 show the namings of parts and the measured
coordinate values from a superposition of the human body image with our
system of parts. Figure 19.5 shows the line skeleton, and Figure 19.6 shows the
Mathematicar file with these coordinates in form of sequences of line endpoints.
Finally, Figure 19.2 shows the Mathematicar surface generated from the given
data. Each line of loops is first generated as a model at the origin of the 3D
coordinate system and then mapped to its position on the corresponding line
via straightforward linear algebra operations.
For the global gesture example, we suppose that two human bodies qua
hypergestures have been constructed. They can be combined to a global hy-
pergesture by identifying a hand of each body that would hold a hand of the
other body, which would be an isomorphism of hands’ parts of the two (local)
hypergestural bodies, see Figure 19.7
19.2 A Conjectured Classification Theorem 99

Fig. 19.2: The human body’s com-

Fig. 19.1: The figure of a human
putergraphical representation via
body’s hypergesture.
Mathematicar software.
100 19 The Classification Problem of Global Gestures

Fig. 19.3: The symbols of defined Fig. 19.4: The measured coordi-
parts of the hypergestural skele- nates of the parts from Figure
ton. 19.3.
19.2 A Conjectured Classification Theorem 101

Fig. 19.5: The line skeleton.

Fig. 19.6: The Mathematicar coordinates

of the parts from Figure 19.3.
102 19 The Classification Problem of Global Gestures

Fig. 19.7: Two connected human Fig. 19.8: Two connected human
bodies in a computergraphical bodies in a computergraphical
representation via Mathematicar representation via Mathematicar
software; the two connected hands software; the two connected hands
would generate isomorphic parts are identified, when holding each
of the two local hypergestures. other.
20
Singular Homology of Hypergestures

Summary. A preliminary investigation of hypergestural classification is the

introduction of singular homology and its signification for counterpoint.
–Σ–
Homology is an algebraic theory that gives a powerful tool for classification
of topological spaces. It does not provide us with classification, but produces
a number of important “invariants”, i.e., mathematical objects that are only
a function of the isomorphism class of an object in a category. Classification
would mean to have a complete set of invariants. This is a very difficult task
in the category Top.
The idea of homology can be connected to Yoneda’s Lemma. This lemma,
valid for every category C, states (among others) that two objects X, Y of
C are isomorphic iff their presheaves (contravaiant functors) @X, @Y are so.
This means that the Yoneda philosophy holds: To understand an object X, it is
necessary and sufficient to understand its functor @X, i.e., to understand how
X behaves when “viewed” from all objects S of the category via the morphisms
f : S → X, the morphism set S@X is the set of perspectives when ‘looking’ at
X from the ‘point of view’ S.
Singular homology is the special Yoneda point of view in C = Top taken
from the objects I n = [0, 1]n defined by the closed unit interval I ⊂ R of
real numbers, i.e., one considers the set I n @X of so-called singular n-chains
c : I n → X with values in X. Evidently, the choice of I stresses the topol-
ogy of R and neglects topologies, such as the Zariski topology in algebraic
geometry, which are radically different from R. The mathematician Alexan-
der Grothendieck (among others) has extended homological theories to se-
tups which are significant for algebraic geometry (e.g. étale cohomology) and
which requires the introduction of new types of topologies, which are finer
than Zariski’s topology, and eventually led to the solution of the famous Weil
conjectures by Grothendieck’s student Pierre Deligne in 1974.
The strong statement of the Yoneda Lemma is restricted to useful domains
I n in singular homology, and we now want to interpret this homology as a

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 103
L. Zheng, G. Mazzola, Classification of Musical Objects for Analysis and Composition,
Computational Music Science, https://doi.org/10.1007/978-3-031-30183-4_20
104 20 Singular Homology of Hypergestures

special case in hypergesture theory. So it is important to recall that homology

deals with a preliminary approach to classification via Yoneda.

20.1 Homology via Hypergestures

Singular homology can be defined by continuous functions with I n as domain
or with n-dimensional simplices. Both approaches yield the same results, and
we choose the “cubic” approach with I n since it harmonizes better with our
hypergestural formalism.
The singular n-chains are the continuous functions c : I n → X, elements
of the set I n @X. This set is a topological space which is homeomorphic to the
space I@I n−1 @X, and so forth, we have
∼
I n @X → I@I@ . . . I@X

the n-fold iteration X 7→ I@X 7→ I@I@X 7→ . . . I@I . . . I@X up to the result

that is homeomorphic to I n @X. This fact is trivial set-theoretically, and also
easy to prove as a fact about topological spaces of continuous functions. This
latter result was used in the Escher Theorem.
To step over to hypergestures, we observe that I n @X identifies with the
∼
hypergesture space ↑ @ ↑ @ . . . ↑ @X =↑n @X since I@X →↑ @X. To de-
fine homology for a hypergesture space Γn @Γn−1 @ . . . Γ1 @X, we first choose a
→
−
commutative ring R. We define the R-module RΓn @Γn−1 @ . . . Γ1 @ X as being
the free R-module over the basis set Γn @Γn−1 @ . . . Γ1 @X. The elements of this
module are called gestural n-chains.
In homology, one considers boundary homomorphisms ∂n : RI n @X →
RI n−1
@X. For our hypergestural situation, we consider an infinite sequence
of digraphs Γ0 , Γ1 , . . . but we suppose that only a finite number of digraphs
Γ0 , Γ1 , . . . Γd−1 define this sequence. This is of course the case if all digraphs
are ↑, as in the classical case. With this condition, the sequence of digraphs can
be described as a d-adic number X.c1 c2 c3 . . . with 0 ≤ ci < d with X as the
topological space that defines our context. For example, if we set Γ0 = ∅, Γ1 =↑,
then the classical approach would be described by X.111 . . . 1000 . . ..
To define gestural n-chain modules, we take the direct sum Cn of all
→
−
modules RΓi1 @Γi2 @ . . . Γin @ X, where i1 , . . . in is a partial (not necessarily
contiguous) sequence of indices in the set 0.c1 c2 c3 . . .. For the classical case
→
−
X.1111 . . . we would have Cn as the module R ↑ @ ↑ @ . . . ↑ @ X with n copies
of ↑. In case two sequences yield the same sequence of digraphs, we only take
the sequence once. In particular, C0 = RX, and for negative n the n-chain
modules are defined to be the zero module.
The next step consists in defining the boundary maps ∂n : Cn → Cn−1 .
It is evidently sufficient to define ∂g for an n-chain g. to this end, let Γ be a
digraph. Take an arrow a of Γ . Then we define Γ |a− to be the digraph obtained
by removing the tail ta of a and all arrows connected to ta . We define by Γ |a+
20.2 Homology and Counterpoint 105

to be the digraph obtained by removing the head ha and all arrows connected
to ha .
→
−
Then we define g ◦ to be the following: Let g : Γ → X be a gesture. Then
1. if AΓ = ∅, then X
g◦ = g(v).
v∈VΓ

2. The general case is defined recursively on the number of arrows:

X
g◦ = (g|a− )◦ − (g|a+ )◦ .
a∈AΓ

For a general chain, we extend this definition by linearity. Now, if g ∈

Γ0 @Γ1 . . . Γn−1 @X, then call gi the hypergesture obtained by the Escher The-
orem transforming g to the corresponding hypergesture in

Γi @Γ0 . . . Γi−1 @Γi+1 @ . . . Γn−1 @X.

Again, we extend this definition linearly to a linear combination of hyperges-

tures. Finally, we define X
∂n g = (−1)i gi◦ ,
which we extend linearly to any linear combination of hypergestures.
With these definitions, we have

∂n−1 ◦ ∂n = 0.

A proof thereof can be found in [30, Ch. 63.3]. This implies that Im(∂n ) ⊂
Ker(∂n−1 ). This enables us to define the n-th homology group

Hn = Ker(∂n )/Im(∂n+1 ).

This invariant of a hypergesture is an important information, however it is not

yet a complete characterization of the isomorphism class of that hypergesture.

20.2 Homology and Counterpoint

It is remarkable that this hypergestural homology theory applies to our coun-
terpoint theory. Recall that in our theory, we deal with maximizing the inter-
section K[] ∩ g(K[]) of the set of consonant intervals in Z12 [] = Z12 [x]/(x2 )
and its g-transformed version g(Z12 [] = Z12 [x]/(x2 )) for a symmetry g of the
interval ring Z12 []. This ring can be given a topology, and we may consider
the closure K[] ∩ g(K[]). Then we have this theorem [27, Ch. 79.4]:

Theorem 9 with the above notations, we have

dim(H1 (K[] ∩ g(K[])) = card(K[] ∩ g(K[])).

106 20 Singular Homology of Hypergestures

This means that we may view that critical cardinality of the intersection
as being the dimension of a module. The latter is no longer restricted to the
finiteness of the intersection. We may now generalize this counterpoint condi-
tion to cases, where the cardinality is infinite, and only consider the dimension,
which is an essential perspective when focusing on towers of dichotomies as
describe in Section 15.2.
21
Local Gestures, Structures of Knots, and Local
Gestures as Local Compositions

Summary. This section discusses the classification problem of gestures for the
special case of circular gestures, which in mathematics are known as knots.
Knots cannot be classified yet. We compare the category of local compositions
to the category of gestures and discuss essential differences. For the special case
of topological R-vector spaces, we present a method to envisage and classify
gestures and local compositions on a common ground.
–Σ–

21.1 Local Gestures and Knot Theory

→
−
Local gestures are complex. A special example is the type of gestures g : → X
with the loop digraph . A special case is a gesture, whose image Im(g) is a
circular curve without self-intersection, except the first and last values being
identical. It can be shown that the isomorphism classes of such circular gestures
have isomorphic fundamental groups of their complement X − Im(g).
Let us recall the concept of a fundamental group π1,x (X) of a topological
space X at a point x ∈ X. It consists of equivalence classes of closed curves
c : I → X that start and end in x. Two such curves c, d are equivalent if there
is a continuous map h : I 2 → X such that h|I × {0} is c, and h|I × {1} is d
and all values are h|{0} × I and all values are h|{1} × I are constant equal x.
Such a map is called isotopy. This defines an equivalence relation among the
closed curves in x. Two such equivalence classes can be composed with each
other by connecting the end of the first with the start of the second, and the
resulting structure is the group π1,x (X). In particular, the inverse of a curve
is the curve obtained by traversing the values in reversed order. If the space is
pathwise connected, this group is independent (up to isomorphism) of the base
point x, it is denoted by π1 (X) and it is called the fundamental group of X.
The knot theory theorem here states that two isomorphic gestures (with-
out self-intersection in their images) have isomorphic fundamental groups of

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 107
L. Zheng, G. Mazzola, Classification of Musical Objects for Analysis and Composition,
Computational Music Science, https://doi.org/10.1007/978-3-031-30183-4_21
108 21 Local Gestures, Structures of Knots, and Local Gestures as Local Compositions

their images’ complements. This however does not mean that the fundamen-
tal groups determine these knot isomorphism classes, but the groups may be
isomorphic without isomorphic knots.
In knot theory, a complete set of invariants is not known. Knots and their
combinations (so-called links) have not been classified yet, only the first six
billion classes have been described, see [16], for example. In particular, the
conductor’s gestures, which are knots, are not classified yet.
The dimension of the space X is crucial for the knot classification. Usually,
knots are embedded in X = R3 . But if we view them as being embedded in
R4 , the situation changes dramatically. In fact, in four dimensions, any non-
intersecting closed loop of one-dimensional string is equivalent to an unknot,
i.e., the knot without any non-trivial property, the closed circle knot.

21.2 Local Gestures as Local Compositions

The classical type of musical objects, called local compositions, is defined in
its simplest shape as an ordered pair (K, M ), where K ⊂ M is a subset of a
module M over a commutative ring R. Local compositions may be combined as
“charts of an atlas” generating a global composition. See [29, vol. I] and Chapter
8 for the technical details of this conceptualization. Local and global compo-
sitions can be connected by determined ‘functions’, so-called morphisms, and
thereby define categories (Loc) of local and categories (Glob) of global compo-
sitions. For quite general system parameters, the isomorphism classes of local
and global compositions have been determined, a fact that is also know as
“classification” of such structures. The global classification theorem is [29, vol.
I, 15.3.2, Theorem 18], see also Chapter 18. Local and global compositions are
essentially algebraic concepts, the R-modules M that carry local/global compo-
sitions K have no topological specification, and the classification of local/global
compositions refers to thoroughly algebraic affine module morphisms.
The algebraic nature of local and global compositions was recognized
around 2002, ironically the year of the first publication of [29], as a concep-
tual limitation that could not embrace the gestural and embodied ‘dancing’
phenomena of music, in particular of performance by human musicians. This
topological aspect of music was also stressed in free jazz pianist Cecil Taylor’s
statement that “I try to imitate on the piano the leaps in space a dancer makes.”
[8].
It should be observed that classical performance theory, as developed by
Johan Sundberg’s school in Stockholm and Guerino Mazzola’s school in Zürich,
among others, did not focus on gestural dimensions, but only on the perfor-
mance transformation ℘ : SymbolicReality → P hysicalReality of sound ele-
ments that are defined by scores (notes, chords, melodies, etc.). It abstracts
from the concrete traditional genealogy of performed sound, namely the trans-
formation of score symbols into gestures, which in turn would interact with an
instrumental interface to generate the sounding output. Cecil Taylor’s dance
21.2 Local Gestures as Local Compositions 109

approach would even avoid the symbolic start and get off the ground via ges-
tures that do not represent ‘frozen’ score symbols.
For these reasons of conceptual deficiency in music and performance the-
ory, a mathematical theory of musical gestures and hypergestures (gestures of
gestures) was initiated in [33] and elaborated in [30, vol. III], giving the theory
of free jazz creativity [34] a powerful conceptualization.
By an adjunction theorem [33], the set Γ @X of gestures from Γ to X is
in a canonical bijection to the set Top(|Γ |, X) of continuous functions starting
from the topological space |Γ |, which is the gluing (colimit) of copies of I,
one for each arrow of Γ , and singletons {v}, one for each vertex of Γ , with
projections I → {v} to heads and tails of Γ .
Gestures are easily classified in an abstract setup (but not in a ge-
ometric model, see Chapter 19). In fact, there is a left group action of
Aut(X) × Aut(Γ )opp , which acts upon the X values of a gesture and the
digraph according to the definition of gesture morphisms. The isomorphism
classes of gestures in Γ @X are this group action’s orbits, the elements of
Γ @X/Aut(X) × Aut(Γ )opp .
In this chapter, we want to investigate some relationships between the
theories of local compositions and gestures. The main result will be that these
two categories are fundamentally different, there is no reasonable functor that
would relate isomorphism classes of local compositions to isomorphism classes
of gestures. We can nevertheless construct canonical local compositions from
gestures for distinguished body spaces.
The only reasonable common ground of these two categories must be a
type of spaces that are simultaneously modules and topological spaces. We
want to investigate a natural mathematical type of such shared spaces, namely
the category of topological vector spaces over the coefficient ring R of reals.
A topological R-vector space is a R-vector space M whose addition and scalar
multiplication are both continuous for the usual topology on R and a given
topology on M . A typical space of this type is the 4-dimensional space REHLD
parametrizing musical events with the four real coordinates E for onset, H for
pitch, L for loudness, and D for duration.
For a topological vector space M over R, the space Γ @M of gestures
g : Γ → M becomes a topological R-vector space by its description as a limit
of the following diagram of R-vector spaces: For every arrow a of Γ , we take
Ma = Top(I, M ) with its canonical R-vector space structure. For every vertex
v of Γ , we take the space Mv = M . For every head map ha : a 7→ v, we take
the linear map Mha : Ma → Mv sending f ∈ Ma to Mha (f ) = f (1), and for
the tail map ta : a 7→ v, we take Mta (f ) = f (0). The limit of this diagram of
topological vector spaces defines the topological vector space Γ @R M . Thereby,
the set of gestures is given the structure of a topological vector space, which we
call the topological gesture space with skeleton Γ and body M . With this setup,
a local gestural composition over Γ, M is a local composition K ⊂ Γ @R M .
110 21 Local Gestures, Structures of Knots, and Local Gestures as Local Compositions

21.2.1 Characterstic Differences

In this section we shall discuss some important differences between gesture and
local composition theories.
To begin with, suppose that gesture g ∈ Γ @R M is isomorphic to h ∈
∼
Γ @R M by the homeomorphism q : M → M . Then there is a gesture t ∈ Γ @R M
such that h = g + t. This means that isomorphism classes of gestures in Γ @R M
are also translation classes, i.e., the algebraic translation classes are finer than
the topological classes. The critical gesture h is defined as follows. Suppose
∼
that there is a homeomorphism f : M → M such that h = f.g. This means
that for any parameter x ∈ |Γ |, we have h(x) = f (g(x)). Then the gesture
t(x) = h(x) − g(x) answers our question.—However, the reverse direction is
false, i.e., if h = g + t, it is not true in general that h and g are related by a
homeomorphism of M . This is evident for the case where g(x) = g(y) for two
different digraph parameters x, y. A homeomorphism would send this point
to one single point, whereas the translation by t could send the point to two
difference values. Also, if a local gestural composition K with more than one
element is isomorphic by a single homeomorphism (topological isomorphism)
to a local gestural composition L, they will not be translations of each other in
general.
Let us look at a special example for the musical standard space M =
REHLD and Γ = • → • → • → •. Gestures of this data are the typical case as
imagined by David Lewin [19]. Local gestural compositions K ⊂ Γ @R M might
describe sets of gesturally interpreted seventh chords, e.g., in a jazz lead sheet.
Local gestural compositions on REHLD can be transformed in a classical
way: Adding a constant gesture g, i.e., a gesture with a single overall value,
will generate a transformation of K by that single value, e.g., adding a pitch
or onset, for example, to K’s coordinates. Also, retrograde or inversion can be
applied by a scalar multiplication of K by −1 in its relevant coordinates. The
general addition of a gesture may generate ‘deformations’ of K’s gestures, an
operation that is not possible in the traditional setup of such gestures.—The
digraph’s arrows may also be used to represent glissandi or crescendi. More
generally, these gestures can be taken to represent functions of one variable
(the digraph’s arguments), and one may use addition or scalar multiplication
of gestures to represent these operations on such functions, e.g., the linear
combination of sinusoidal functions in Fourier analysis.
In any case,

Theorem 10 For a digraph Γ and a topological R-vector space M , local ges-

tural compositions in Γ @R M can be classified, i.e., their ‘linear’ isomorphism
classes (as distinguished from the topological classes) are completely calculated
following the classification theorem in Section 8.1.4.
Observe that the action of Aut(Γ ) on a local gestural composition is
absorbed by the corresponding affine automorphisms of Γ @R M .
21.2 Local Gestures as Local Compositions 111

21.2.2 Conclusion and Future Topics

The next theoretical step would be to investigate and hopefully classify global
gestural compositions over topological R-vector spaces. Also, more extensive
applications and examples in this context are to be expected, specifically to
understand the structural and dynamic aspect of gesturally driven improvisa-
tion.
Part V

Classification and Creativity

22
Gestural Interpretation of Harmonic Dynamics
in Tonal Modulation and Future Developments

Summary. We give a gestural interpretation of tonal modulations. We discuss

some music software which uses classification results, especially the RUBATO r
components for rhythmical and harmonic analysis. This chapter gives some
perspectives relating to future directions in theory and software, especially
regarding classification in performance.
–Σ–

22.1 Gestural Modulation and Creativity

Fig. 22.1: The force field generated by the Lie bracket of two tonalities.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 115
L. Zheng, G. Mazzola, Classification of Musical Objects for Analysis and Composition,
Computational Music Science, https://doi.org/10.1007/978-3-031-30183-4_22
116 22 Gestural Interpretation and Future Developments

The classical approach to tonal modulation, as described in Section 5.2, in-

volves classification of tonalities via symmetries of their triadic interpretations
of X 3 . These symmetries are however not embodied as movements among mu-
sical objects. The symmetries are only present via modulation quanta, i.e., sets
of notes that are invariant under these symmetries. The symmetries are not
‘visible’, but their movements only appear as abstract relationships connecting
corresponding note objects.
When introducing gestures, one would like to materialize these symme-
tries as gestural movements and thereby map the abstract setup into a gestural
movement. This goal could be achieved by a new gestural model of tonal mod-
ulation, where notes are gesturally moved to describe the modulatory process.
The gestural modulation model gets off the ground by the creation of a
force field by a mathematical tool in differential geometry, called Lie bracket of
vector fields. Figure 22.1 shows the field generated from tonality C major and
tonality F major. The circle shows the pitch classes as immersed in the force
field.
This force field generates force curves that define gestures connecting notes
of triads, as shown in Figure 22.2.

Fig. 22.2: The gestural connection moving a triad into another, thereby defining
modulation.

Without going into details of this gestural modulation model, as devel-

oped in [30, Ch. 80], we can state that it provides us with the same results as
the classical symmetry model. There is however an important advantage of the
gestural approach, namely the fact that the gestural model does not a priori
presuppose that the involved tonalities are in one and the same isomorphism
class. This classical condition disappears in the gestural model. In particular,
22.2 Classification Problems for Performance Gestures 117

this new modulation model could be used to construct a theory, where mod-
ulation moves between tonalities of different classes, e.g., between major and
harmonic minor tonalities, or even between exotic tonalities of general type.
This shift from the category of compositions to the category of gestures
generates an enrichment of theory, which should be applied to musical com-
position. The general principle of composition, namely the projection from a
mathematical category into the realm of musical objects and relations, is now
enriched by the switch from algebraic categories (compositions) to topological
categories (gestures). The intuitive meaning thereof is a transformation from
‘rigid’ abstract concepts to more ‘elastic’ ones of gestural embodiment!

22.2 Classification Problems for Performance Gestures

We discussed the gestural interpretation of performance integral curves of per-
formance transformations ℘ : REHLD → Rehld , see Figure 22.3, in Section
6.2.

∆
X x
Ts

℘
x0
X0

Fig. 22.3: The performance transformation specifies integral curves, which are gestures
in REHLD and their isomorphic images in Rehld .

Classification of such performance curve gestures goes beyond the abstract

classification of these ↑-gestures. We have to recall that they are the result
of a performance vector field qua integral curves thereof, together with the
associated performance transformation ℘.
If we can prove that two such gestures (curves) (ending at the same sym-
bolic point of X in Figure 22.3) are isomorphic, we have to investigate the
associated inverse problem, namely to determine the changes in generating pa-
rameters of the two corresponding performance fields. This is a semantic prob-
lem: How are field parameters related if their integral curves are isomorphic?
We are far from a solution, but we now understand the deep consequences of
118 22 Gestural Interpretation and Future Developments

gestural classification. This topic is evidently very important for understanding

the geneology of performance, to understand, why performances are generated
to produce the audible shape the listener may perceive.

22.3 Mirror Logic

Summary. We present a logical approach to the construction of hypergestures
in musical composition. This logic is built upon a “mirror” construction using
fiber products and operations on the musical time domain. We discuss four
examples: Temporality of harmonic structures, counterpoint, lead sheets, and
dance hypergestures.
–Σ–

22.3.1 Introduction

The mathematical theory of musical gestures, as developed in [29, vol. III],

provides us with an adequate description of the category of gestures and hy-
pergestures, i.e., gestures of gestures. However, the creative process of a musical
composition (including improvisation) has not been dealt with in Mazzola’s ges-
ture theory. The musical logic that drives the construction of hypergestures is
not embodied in the “static” categories of hypergestures. It is evidently impos-
sible to include the limitless variety of creative musical procedures in a single
theory, we understand this infinity from our own practice as a free jazz pianist
(Guerino Mazzola).
Nevertheless, we claim that it is feasible to describe some logical methods
that are omnipresent in the basic construction mechanisms of complex musical
hypergestures. The naive gestural context of a given composition is first of
all a complex space of hypergestures Γ1 @Γ2 @ . . . Γn @X that are built starting
from a basic topological space X (we don’t extend to topological categories
in this chapter). But this hypergestural setup is often too simple. Instead of
descending along a sequence of digraphs, the hypergestures are built from a
ramification of spaces that is defined by cartesian products, e.g., Γ1 @(X1 ×X2 ×
. . . Xm ), where X1 = Γ1,1 @(X1,1 × . . . X1,m1 ) etc. Such cartesian products may
occur, when the gestures involve a number of simultaneous musical parameters,
such as pitch and time, for example. This means that the gestural codomain
involves compound musical spaces. And the cartesian factors of such spaces
are often also hypergestural constructions. In this case, we have the canonical
∼
isomorphism Γ1 @(X1 × X2 × . . . Xm ) → Γ1 @X1 × Γ1 @X2 × . . . Γ1 @Xm , and so
on for its factors if they are cartesian products. Dually, codomains of gestures
may also be coproducts, i.e., disjoint unions of spaces for alternatives of choice,
among orchestral instruments, for example. But we shall not discuss this dual,
but not fundamentally different structural situation here.
Our claim means that in such compound hypergestural architectures, the
factors (or cofactors) are not generated by a single stroke of creativity, but that
22.3 Mirror Logic 119

there often might be a logical order that determines a hierarchy of construction,

i.e., that some factors of hypergestures may be constructed before others are,
and that the subordinate later factors are determined by the earlier and higher
ranked factors in a functional logical dependence.

22.3.2 The Logic of Spatial Mirrors

Our language of gestures has two adjoint setups: either a gesture is a digraph
→
−
morphism g : Γ → X from a digraph Γ (the skeleton) to the body digraph
→
−
X = Top(I, X), I = [0, 1] the real unit interval, or it is by adjunction of
functors a continuous map |g| : |Γ | → X. We shall use the adjoint version here.
The musical time [32] is defined by the topological space |Γ |, and by a limit of
topological digraph spaces for a diagram of hypergestures. Let us focus on the
simpler case |Γ |@X.
A gesture f : |Γ | → X is described by knowing which symbolic time t
goes to which point of X. This association can be deduced from the functor
Y @X = Top(Y, X) → Top(Y ×X |Γ |, |Γ |) defined by the fiber product. In
fact, if Y = {x} is a singleton, i.e., a point of X, then Y ×X |Γ | is the fiber
of f over x, and these fibers identify f . We view this procedure as a kind of
“mirror” at X, yielding the musical times associated with spatial objects over
X.
In the following sections, we describe such a mirroring procedure to un-
derstand a type of logic for the creation of hypergestures. Our examples focus
on four situations: (1) temporal specification of chords, (2) construction of first
species counterpoint, (3) lead sheets in jazz, and (4) dance gestures associated
with musical structures.

22.3.3 Temporal Specification of Chords

This situation starts from a gesture f : |Γ | → X1 × X2 , where X1 is a space

of harmonic structures (chords) and X2 is a space of symbolic onset times,
in quarter-note units, say. The logical challenge here would be to associate
onset times with given harmonic structures, more concretely, to specify onset
times for chords. The logic would work as follows: One specifies a number of
harmonic structures, which are given as a set P = {p1 , p2 , . . . pn } of points in
X1 . We now suppose that the gesture f1 : |Γ | → X1 is given, i.e. the musical
time projecting to the harmonic structure space X1 . Then the fiber over P
would be the set F (P ) = {t1,1 , t1,2 . . . t2,1 , . . . tn,1 , . . .} of pairwise different
musical times, where ti,j is the musical time of index j over structure pi . In
other words, we have a possibly multiple musical time over every structure
pi . On this set, we may now define a set function f2 from |Γ | into the time
space X2 , and, in order to avoid any topological restriction, we may suppose
that X2 is given the indiscrete topology. The two maps f1 , f2 now define the
desired gesture f = (f1 , f2 ) : |Γ | → X1 × X2 . Here, the structural space
120 22 Gestural Interpretation and Future Developments

part with f1 is hierarchically superior to the temporal part, where the onsets
of chords are defined using their musical time for f1 . Of course, the onset
assignment may follow a musical reason (or logic) that the composer should
specify in the concrete situation. For example, a cadence would determine the
onset sequence of its chords of type I, IV, V, I, say. If desired, one could also
specify the harmonic space X1 as being the space of gestures with values in a
pitch space, following David Lewin’s idea of a gestural interpretation of chords
[19].

22.3.4 Construction of First Species Counterpoint

Traditional Fuxian counterpoint [9] in its elementary first species is defined by

a cantus firmus line CF = (c1 , c2 , . . . cm ) of pitch classes with increasing onset
times plus some normalized durations, which we don’t consider here. The points
ci are events with onset and pitch class coordinates, possibly within a given
scale.
This CF line is complemented by a discantus line D = (d1 , d2 , . . . dm ) of
the same onset and pitch class parameter space, where the onsets of a ci and
di coincide, and where the difference of pitch classes of ci and di should be
consonances (pitch class differences 0,3,4,7,8,9), see also [3].
The terminology “punctus contra punctum” does not refer to the D set
as an opposition (“contra”) to CF , it refers to the suspence-packed succession
of intervals Ii defined by ci , di , see [43]. But the CF, D setup defines a simpler
constructive description of first species counterpoint.
This situation can be described by a gesture f : |Γ | → XCF × XD ,
where the space XCF consists of points with onset and pitch class coordinates,
while XD consists of the six consonant pitch class numbers. Both spaces carry
the indiscrete topology. We may suppose that CF is represented by a gesture
f1 : |Γ | → XCF , where the elements of CF are generating the fiber set T in
|Γ |. Observe that a time t projected to a CF point c needn’t be single, f1 may
show a number of times sitting over c. This fiber set T may now be projected
via gesture map f2 : |Γ | → XD to a set of consonances in XD , which means
that for every t ∈ T , we have a discantus dt = f1 (t) + f2 (t). The hierarchy here
is dominated by the cantus firmus line defined by f1 , whereas the consonances
associated with musical times of CF values are chosen as subordinate data.

22.3.4.1 Hypergestures for Counterpoint

Summary. We discuss the application of the Escher Theorem of hypergestures

to counterpoint.
–Σ–
An application of the Escher Theorem deals with first species counter-
point. There are two well-known views of this contrapuntal situation: The first
22.3 Mirror Logic 121

views such a counterpoint as a hypegestural movement from the gesture of can-

tus firmus to the gesture of the discantus. The second views the counterpoint
as a hypergesture of the sequence of interval gestures. The Escher Theorem in-
terprets this duality as the result of exchanging the two digraphs. This means
that both perspectives are equivalent under that digraph permutation. These
perspectives are isomorphic with each other, but one should recall that only
the second, a hypergesture of interval gestures, is historically and theoretically
valid. The “contra” in counterpoint is an opposition of successive interval ges-
tures.

22.3.5 Lead Sheet in Jazz

Lead sheets in jazz have to components: the sequence of chords (defining the
chord changes) and the basic melodic line flying over the chord changes, see also
[37]. This construction may be viewed as a hierarchy, where the chord sequence
is the dominant factor, while the melodic line is derived in accordance with the
harmonic path of chord changes.
We may interpret this data as a gesture f = (f1 , f2 ) : |Γ | → XCh × XM el ,
where XCh is a space of Lewin-type gestures for chords, but enriched by a
symbolic time coordinate, following the ideas of the first example.
The space XM el should consist of melodic gestures with values in pitch-
onset-duration space. The mirror logic would then define a gesture of f2 : |Γ | →
XM el with specific values on the fiber in |Γ | over the chord set defined by f1 .

22.3.6 Dance Gestures Associated with Musical Structures

A dance standard setup has two components: dance gestures and musical struc-
tures that are related to these dance gestures. Accordingly, we have to consider
gestures f = (fdance , fmusic ) : |Γ | → Xdance × Xmusic . The space Xdance is
a space of hypergestures that eventually terminate in a space with spatial 3D
and time coordinates. The space Xmusic should be a hypergesture space whose
elements describe musical structures by hypergestures with values in typical
onset, pitch, loudness, duration coordinates of a common score. The mirror
logic here associates dance gestures with musical times taken from the fiber
of fmusic . The hierarchy is generating dance information from the fibers of
musical gesturality. Of course, as with the above examples, the details of the
compositional logic deduced from the mirror logic are to be elaborated, but
they all must be realized in the mirror logic’s framework. In modern dance, the
logic is also reversed, meaning that the dance gestures are given first, and then
musical structures are deduced from dance information.

22.3.7 Conclusion and Future Topics

We have shown with four typical examples how musical composition in com-
plex hypergestural configurations may be created following a mirror logic, where
122 22 Gestural Interpretation and Future Developments

fiber products are key. Future topics should include an investigation of char-
acteristic differences between composition of scores and improvisation. In the
latter case, psychological logic may play a key role for the selection of hierar-
chically secondary layers of musical creativity.

22.4 Perspectives of Future Theory

The role of classification for musical creativity should be a dominant topic
in future theory and software. From our discussion in this book it appears
that creativity should focus on the transformation of classification, which is a
mathematical theme, into musical structures, compositions and gestures.
The overall image is this: We are given a set of musical structures, i.e.,
compositions or gestures. Their classification generates a corresponding set of
orbits, i.e., classes of equivalent structures. The underlying equivalence rela-
tions are mathematical objects, generated by transformations among members
of such orbits. The musical creativity deals with the translation of these trans-
formations into musical objects.
As we have learned from the examples in this book, from tonal modula-
tion to serial approaches and counterpoint, important compositional techniques
are designed to “musicalize” classifying transformations, i.e., representing them
within a creative compositional strategy.
For tonal modulation, one musicalizes the transformational symmetry be-
tween two tonalities, which means to represent that symmetry in terms of a
modulation quantum, whose triadic degrees materialize the symmetry as a set
of notes. This model also includes a syntax of the modulatory architecture,
which means that not only the materialization is given, but also its temporal
unfolding.
For dodecaphonic and serial composition, one materializes the orbit of
a row by the display of all its 48 equivalent transformed rows. However, this
brute method lacks any syntactical explication.
For counterpoint, the orbit of a strong dichotomy is used to determine
successive intervals, in fact, the deformations of a dichotomy generate successive
interval options.
In this spirit, musical/compositional creativity consists of a musically com-
prehensible projection of abstract logical structures into the material ontology
of sensible (audible, gestural) musical objects.
In the case of improvisation, the dialogue between its performers (includ-
ing the dialogue with oneself) creates new logical expressions, materializes them
and receives one’s collaborator’s reply on the flight. It is a compositional cre-
ation in the moment of it’s making. But it follows the same general methodology
as composition.
In terms of the concept of creativity, what is the “box” that has to be
opened and transgressed in our situation? This box is the logically closed ar-
chitecture of musical constructions in the language of mathematics (or just
22.4 Perspectives of Future Theory 123

the mental formalism of compositional strategies). Often, musical constructors

believe this framework is sufficient for a valid composition. Their composi-
tional work follows this setup without taking care of its transformation into
musical objects. This is an often hidden box, where the difference between
thinking and making/creating music is not recognized. Creativity here stems
from the insight that transforming thoughts into music must build musical ob-
jects, sounds, notes, gestures, etc., that are a sufficiently faithful image of initial
thoughts. This is what “making music” means. Figure 22.4 shows this projec-
tion from mathematical categories to musical realm with its musical parts and
transformations. Classification is involved since the most prominent examples
deal with morphisms that are isomorphisms, i.e., those concepts, which define
classification.

Fig. 22.4: Musical creativity is a projection from mathematical categories (and in par-
ticular isomorphisms) to the realm of music with its musical parts and transformation
qua musical entities.
Part VI

References, Index
23
Classification Lists

Summary. This chapter displays lists of classified objects.

–Σ–

23.1 List of Local Denotators in Z12 , Chord Classes

This section contains the list of all isomorphism classes of zero-addressed chords
in P iM od12 . Here we give a short definition.
• Class Nr. is the number of the isomorphism class, numbers with extension
“.1” indicate the class number for classification under symmetries from Z
(no fifth or fourth transformations). Autocomplementary classes have a star
after the number.
• Representative of Nr. without hat is the number’s representative in full
circles, the one with hat is the complementary chord.
• Group of symmetries is Sym(N r.). To keep notation readable, we use the
notation with linear factor to the left.
• Conj. Class denotes the conjugacy class symbol of Sym(N r.) and refers to
the numbering 1, 2, . . . 19 from [38].
• Card. End. Cl. N r.|N d r. is the pair of numbers of conjugacy classes of en-
domorphisms in Nr. and in its complement N d r., respectively.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 127
L. Zheng, G. Mazzola, Classification of Musical Objects for Analysis and Composition,
Computational Music Science, https://doi.org/10.1007/978-3-031-30183-4_23
128 23 Classification Lists

Chord Classes
Class Representative Group of Conj. ] End.
Nr. N r. = •, N r. = ◦ Symmetries
d Class N r.|N
d r.
−→
1 • • • • • • • • • • •• GL(Z12 ) 19 28|28
One/Eleven Element

2 • ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦◦ Z×
12 8 1|31
Two/Ten Elements

3 • • ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦◦ h−1e−1 i 3 3|23
3.1 • ◦ ◦ ◦ ◦ • ◦ ◦ ◦ ◦ ◦◦
4 • ◦ • ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦◦ {1, 7, −1e−2 , 5e−2 } 8 3|25
5 • ◦ ◦ • ◦ ◦ ◦ ◦ ◦ ◦ ◦◦ {1, 5, 7e −3
, −1e −3
} 8 3|19
6 8
• ◦ ◦ ◦ • ◦ ◦ ◦ ◦ ◦ ◦◦ {1, 7, 5e , −1e } 8
8 3|31
7 • ◦ ◦ ◦ ◦ ◦ • ◦ ◦ ◦ ◦◦ Z×
12 ne 6Z12
13 3|28
Three/Nine Elements

8 • • • ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦◦ h−1e−2 i 2 4|14
8.1 • ◦ • ◦ ◦ ◦ ◦ • ◦ ◦ ◦◦
9 • • ◦ • ◦ ◦ ◦ ◦ ◦ ◦ ◦◦ {1} 1 4|30
9.1 • ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦ ◦◦
10 • • ◦ ◦ • ◦ ◦ ◦ ◦ ◦ ◦◦ {1} 1 8|36
10.1 • ◦ ◦ • ◦ ◦ ◦ • ◦ ◦ ◦◦
11 • • ◦ ◦ ◦ • ◦ ◦ ◦ ◦ ◦◦ h5i 4 4|20
12 • • ◦ ◦ ◦ ◦ • ◦ ◦ ◦ ◦◦ h7e i 6
6 5|29
13 • ◦ • ◦ • ◦ ◦ ◦ ◦ ◦ ◦◦ {1, 7, −1e , 5e }8 8
8 4|18
14 • ◦ • ◦ ◦ ◦ • ◦ ◦ ◦ ◦◦ h7i 6 8|31
15 • ◦ ◦ • ◦ ◦ • ◦ ◦ ◦ ◦◦ {1, 5, −1e , 7e }6 6
8 5|32
16 • ◦ ◦ ◦ • ◦ ◦ ◦ • ◦ ◦◦ Z×
12 ne 4Z12
15 4|20
Four/Eight Elements

17 • • • • ◦ ◦ ◦ ◦ ◦ ◦ ◦◦ h−1e−3 i 3 4|8
17.1 • ◦ • ◦ ◦ • ◦ • ◦ ◦ ◦◦
18 • • • ◦ • ◦ ◦ ◦ ◦ ◦ ◦◦ {1} 1 5|19
18.1 • ◦ • ◦ • ◦ ◦ • ◦ ◦ ◦◦
19 • • • ◦ ◦ • ◦ ◦ ◦ ◦ ◦◦ {1} 1 5|19
19.1 • • ◦ • ◦ ◦ ◦ ◦ • ◦ ◦◦
20 • • • ◦ ◦ ◦ • ◦ ◦ ◦ ◦◦ {1} 1 7|23
23.1 List of Local Denotators in Z12 , Chord Classes 129

Chord Classes—Continued
Class Representative Group of Conj. ] End.
Nr. N r. = •, N r. = ◦ Symmetries
d Class N r.|N
d r.
20.1 • • ◦ ◦ ◦ • ◦ • ◦ ◦ ◦◦
21 • • • ◦ ◦ ◦ ◦ • ◦ ◦ ◦◦ {1, 7, −1e−2 , 5e−2 } 9 7|9
22 • • ◦ • • ◦ ◦ ◦ ◦ ◦ ◦◦ h−1e −4
i 2 6|20
22.1 • ◦ • ◦ ◦ • ◦ ◦ ◦ • ◦◦
23 • • ◦ • ◦ • ◦ ◦ ◦ ◦ ◦◦ h5i 4 5|13
24 • • ◦ • ◦ ◦ • ◦ ◦ ◦ ◦◦ h7e i 6
6 6|17
25 • • ◦ • ◦ ◦ ◦ • ◦ ◦ ◦◦ {1} 1 0|3
25.1 • • ◦ • ◦ • ◦ ◦ ◦ ◦ ◦◦
26 • • ◦ • ◦ ◦ ◦ ◦ ◦ • ◦◦ {1} 1 12|31
26.1 • ◦ • ◦ ◦ • ◦ ◦ • ◦ ◦◦
27 • • ◦ • ◦ ◦ ◦ ◦ ◦ ◦ •◦ {1, 7e−3 , 5e2 , −1e−1 } 11 5|13
28 • • ◦ ◦ • • ◦ ◦ ◦ ◦ ◦◦ h−1e i 7
3 6|14
28.1 • • ◦ ◦ ◦ • ◦ ◦ • ◦ ◦◦
29 • • ◦ ◦ • ◦ ◦ • ◦ ◦ ◦◦ h7i 6 10|23
30 • • ◦ ◦ • ◦ ◦ ◦ • ◦ ◦◦ h5e i 4
4 11|23
31 • • ◦ ◦ • ◦ ◦ ◦ ◦ • ◦◦ {1, −1e −1
, 5e −4
, 7e } 10
3
9|19
32 • • ◦ ◦ ◦ • • ◦ ◦ ◦ ◦◦ {1, 5, −1e6 , 7e6 } 8 7|15
33 • • ◦ ◦ ◦ ◦ • • ◦ ◦ ◦◦ {1, 7, −1e , 5e , −1 −1
14 7|14
e6 , 7e6 , 5e5 , −1e5 }
34 • ◦ • ◦ • ◦ • ◦ ◦ ◦ ◦◦ {1, 7, −1e6 , 5e6 } 9 6|17
35 • ◦ • ◦ • ◦ ◦ ◦ • ◦ ◦◦ {1, 7, −1e , 5e } 4 4
8 11|19
36 • ◦ • ◦ ◦ ◦ • ◦ • ◦ ◦◦ {1, 7, −1e , 5e , −2 −2
13 9|28
e6 , 7e6 , 5e4 , −1e4 }
37 • ◦ ◦ • ◦ ◦ • ◦ ◦ • ◦◦ Z×
12 n e
3Z12
17 7|21
Five/Seven Elements

38 • • • • • ◦ ◦ ◦ ◦ ◦ ◦◦ h−1e−4 i 2 5|7
38.1 • ◦ • ◦ • ◦ ◦ • ◦ • ◦◦
39 • • • • ◦ • ◦ ◦ ◦ ◦ ◦◦ {1} 1 6|10
39.1 • • ◦ • ◦ • ◦ ◦ ◦ ◦ •◦
40 • • • • ◦ ◦ • ◦ ◦ ◦ ◦◦ {1} 1 8|12
40.1 • • ◦ • ◦ ◦ • ◦ • ◦ ◦◦
41 • • • • ◦ ◦ ◦ • ◦ ◦ ◦◦ {1} 1 8|12
130 23 Classification Lists

Chord Classes—Continued
Class Representative Group of Conj. ] End.
Nr. N r. = •, N r. = ◦ Symmetries
d Class N r.|N
d r.
41.1 • • • ◦ ◦ • ◦ • ◦ ◦ ◦◦
42 • • • ◦ • • ◦ ◦ ◦ ◦ ◦◦ {1} 1 6|16
42.1 • • ◦ • ◦ • ◦ ◦ • ◦ ◦◦
43 • • • ◦ • ◦ • ◦ ◦ ◦ ◦◦ {1} 1 8|20
43.1 • • ◦ • ◦ • ◦ • ◦ ◦ ◦◦
44 • • • ◦ • ◦ ◦ • ◦ ◦ ◦◦ h7i 6 7|9
45 • • • ◦ • ◦ ◦ ◦ • ◦ ◦◦ {1} 1 16|22
45.1 • • ◦ ◦ • ◦ • ◦ • ◦ ◦◦
46 • • • ◦ • ◦ ◦ ◦ ◦ • ◦◦ h5e−4 i 4 5|12
47 • • • ◦ • ◦ ◦ ◦ ◦ ◦ •◦ h−1e −2
i 2 8|14
47.1 • ◦ • ◦ • ◦ • ◦ ◦ • ◦◦
48 • • • ◦ ◦ • • ◦ ◦ ◦ ◦◦ {1} 1 8|18
48.1 • • ◦ • ◦ ◦ ◦ • • ◦ ◦◦
49 • • • ◦ ◦ • ◦ ◦ • ◦ ◦◦ {1} 1 10|18
49.1 • • ◦ • ◦ ◦ ◦ ◦ • • ◦◦
50 • • • ◦ ◦ • ◦ ◦ ◦ • ◦◦ h−1e−2 i 2 9|13
50.1 • • ◦ • • ◦ ◦ ◦ • ◦ ◦◦
51 • • • ◦ ◦ ◦ • • ◦ ◦ ◦◦ h7i 6 9|11
52 • • • ◦ ◦ ◦ • ◦ • ◦ ◦◦ {1, −1e −2
, 7e , 5e } 8
6 4
7|17
53 • • ◦ • • ◦ • ◦ ◦ ◦ ◦◦ {1} 1 10|20
53.1 • • ◦ • ◦ ◦ • ◦ ◦ ◦ •◦
54 • • ◦ • • ◦ ◦ • ◦ ◦ ◦◦ {1} 1 14|26
54.1 • • ◦ ◦ • ◦ • ◦ ◦ • ◦◦
55 • • ◦ • ◦ • • ◦ ◦ ◦ ◦◦ {1, 5, −1e6 , 7e6 } 8 8|8
56 • • ◦ • ◦ • ◦ ◦ ◦ • ◦◦ h5i 4 16|16
57 • • ◦ • ◦ ◦ • • ◦ ◦ ◦◦ h7e i 6
6 12|16
58 • • ◦ • ◦ ◦ • ◦ ◦ • ◦◦ h7e i 6
6 18|23
59 • • ◦ • ◦ ◦ ◦ • ◦ • ◦◦ h7i 6 13|29
60 • • ◦ ◦ • • ◦ ◦ • ◦ ◦◦ h5i 4 11|19
61 • • ◦ ◦ • ◦ ◦ • • ◦ ◦◦ {1, 7, −1e4 , 5e4 } 8 14|14
62 • ◦ • ◦ • ◦ • ◦ • ◦ ◦◦ {1, 7, −1e , 7e } 4 4
8 11|19
23.1 List of Local Denotators in Z12 , Chord Classes 131

Chord Classes—Continued
Class Representative Group of Conj. ] End.
Nr. N r. = •, N r. = ◦ Symmetries
d Class N r.|N
d r.
Six/Six Elements

63* • • • • • • ◦ ◦ ◦ ◦ ◦◦ h−1e−5 i 3 5|5

63.1* • • ◦ • ◦ • ◦ ◦ • ◦ •◦
64* • • • • • ◦ • ◦ ◦ ◦ ◦◦ {1} 1 9|9
64.1* • • ◦ • ◦ • ◦ • ◦ ◦ •◦
65 • • • • • ◦ ◦ • ◦ ◦ ◦◦ {1} 1 9|9
65.1 • • • ◦ • ◦ ◦ • ◦ • ◦◦
66 • • • • • ◦ ◦ ◦ • ◦ ◦◦ h−1e−4 i 2 12|6
66.1 • • • ◦ ◦ • ◦ • ◦ • ◦◦
67* • • • • ◦ • ◦ • ◦ ◦ ◦◦ h5e−2 i 5 6|6
68* • • • • ◦ • ◦ ◦ • ◦ ◦◦ {1} 1 9|9
69 • • • • ◦ • ◦ ◦ ◦ • ◦◦ {1} 1 15|11
69.1 • • ◦ • • ◦ • ◦ • ◦ ◦◦
70* • • • • ◦ • ◦ ◦ ◦ ◦ •◦ {1, 5, −1e−3 , 7e−3 } 10 6|6
71* • • • • ◦ ◦ • • ◦ ◦ ◦◦ {1} 1 11|11
71.1* • • • ◦ ◦ • ◦ • • ◦ ◦◦
72 • • • • ◦ ◦ • ◦ • ◦ ◦◦ h7e6 i 6 8|10
73 • • • • ◦ ◦ • ◦ ◦ • ◦◦ h−1e −3
i 3 13|9
73.1 • • ◦ • ◦ ◦ • ◦ • • ◦◦
74* • • • • ◦ ◦ ◦ • • ◦ ◦◦ h−1e−3 i 3 7|7
75* • • • ◦ • • ◦ ◦ • ◦ ◦◦ {1} 1 17|17
75.1* • • ◦ ◦ • • ◦ ◦ • ◦ •◦
76* • • • ◦ • • ◦ ◦ ◦ • ◦◦ h5e−5 i 4 10|10
77* • • • ◦ • ◦ • ◦ • ◦ ◦◦ h5e4 i 4 14|14
78* • • • ◦ • ◦ • ◦ ◦ ◦ •◦ {1} 1 23|23
78.1* • • ◦ • ◦ • ◦ • ◦ • ◦◦
79 • • • ◦ • ◦ • • ◦ ◦ ◦◦ h7i 6 18|10
80 • • • ◦ • ◦ ◦ • ◦ ◦ •◦ {1, 7, 5e −2
, −1e −2
} 9 15|11
81* • • • ◦ • ◦ ◦ ◦ • • ◦◦ h5e −4
i 4 11|11
82* • • • ◦ ◦ • • ◦ ◦ • ◦◦ {1} 1 17|17
82.1* • • ◦ • • ◦ ◦ • • ◦ ◦◦
83* • • • ◦ ◦ ◦ • • • ◦ ◦◦ {1, −1e−2 , 5e−2 , 7} 13 12|12
132 23 Classification Lists

Chord Classes—Continued
Class Representative Group of Conj. ] End.
Nr. N r. = •, N r. = ◦ Symmetries
d Class N r.|N
d r.
ne6Z12
84* • • ◦ • • ◦ • ◦ ◦ • ◦◦ h7e3 i 7 14|14
85 • • ◦ • • ◦ ◦ • ◦ • ◦◦ {1, −1e −4
, 5e −4
, 7} 8 15|23
86* • • ◦ • ◦ ◦ • • ◦ • ◦◦ h7i n e 6Z12
12 20|20
87* • • ◦ • • ◦ ◦ • • ◦ ◦◦ {1} 16 12|12
88* • ◦ • ◦ • ◦ • ◦ • ◦ •◦ Z×
12 ne Z12
18 12|12
23.2 Third Chain Classes 133

23.2 Third Chain Classes

The following list of third chain translation classes shows the class number in
the first column, where equivalence (∼) means that the same pc set is generated.
The second column shows the pitch classes in the order of appearance along the
third chain. The third column shows the third chain, the fourth column shows
the chord class of the pc set, and the fifth column shows lead-sheet symbols.

Third Chains
Chain Nr. Pitch Classes Third Chord Lead-Sheet
∼ equiv. from 0 Chain Class Symbols
Two Pitch Classes
1 0,3 3 5 trd
2 0,4 4 6 T rd
Three Pitch Classes
3 0,3,6 33 15 C0, Cm5-
4 0,3,7 34 10.1 Cm
5 0,4,7 43 10.1 C
6 0,4,8 44 16 C+, C5+
Four Pitch Classes
7 0,3,6,9 333 37 C07-
8 0,3,6,10 334 26.1 C07
9 0,3,7,10 343 22.1 Cm7
10 0,3,7,11 344 30 Cm7+
11 0,4,7,10 433 26.1 C7
12 0,4,7,11 434 28.1 C7+
13 0,4,8,11 443 30 C+7+
Five Pitch Classes
14 0,3,6,9,1 3334 58 C07-/9-
15 0,3,6,10,1 3343 53.1 C09-
16 0,3,6,10,2 3344 56 C09
17 0,3,7,10,1 3433 53.1 Cm9-
18 0,3,7,10,2 3434 42.1 Cm9
19 0,3,7,11,2 3443 59 Cm7+/9, Cmmaj7/9
20 0,4,7,10,1 4333 58 C9-
21 0,4,7,10,2 4334 47.1 C9
22 0,4,7,11,2 4343 42.1 C7+/9, Cmaj7/9
23 0,4,7,11,3 4344 60 C7+/9+
24 0,4,8,11,2 4433 56 C+7+/9
134 23 Classification Lists

Third Chains—Continued
Chain Nr. Pitch Classes Third Chord Lead-Sheet
∼ equiv. from 0 Chain Class Symbols
25 0,4,8,11,3 4434 60 C+7+/9+
Six Pitch Classes
26 0,3,6,9,1,4 33343 84* C07-/9-/11-
27 0,3,6,9,1,5 33344 85b C07-/9-/11
28 0,3,6,10,1,4 33433 79 C09-/11-
29 0,3,6,10,1,5 33434 65.1b C09-/11
30 0,3,6,10,2,5 33443 69.1 C011
31 0,3,7,10,1,4 34333 84* Cm9-/11-
32 0,3,7,10,1,5 34334 64.1* Cm9-/11
33 0,3,7,10,2,5 34343 63.1* Cm11
34 0,3,7,10,2,6 34344 75.1* Cm11+
35 0,3,7,11,2,5 34433 69.1 Cm7+/11
36 0,3,7,11,2,6 34434 82* Cm7+/11+
37 0,4,7,10,1,5 43334 73.1 C9-/11
38 0,4,7,10,2,5 43343 64.1* C11
39 0,4,7,10,2,6 43344 78.1* C11+
40 0,4,7,11,2,5 43433 65.1b C7+/11
41 0,4,7,11,2,6 43434 66.1b C7+/11+
42 0,4,7,11,3,6 43443 82* C7+/9+/11+
43 0,4,8,11,2,5 44333 85b C+7+/11
44 0,4,8,11,2,6 44334 78.1* C+7+/11+
45 0,4,8,11,3,6 44343 75.1* C+7+/9+/11+
46 0,4,8,11,3,7 44344 87* C+7+/9+/(11)/13-
Seven Pitch Classes
47 0,3,6,9,1,4,7 333433 58b C07-/9-/11-/13-
48 0,3,6,9,1,4,8 333434 54.1b C07-/9-/11-/13
49 0,3,6,9,1,5,8 333443 54.1b C07-/9-/13
50 0,3,6,10,1,4,7 334333 58b C09-/11-/13-
51 0,3,6,10,1,4,8 334334 47.1 C09-/11-/13
52 0,3,6,10,1,5,8 334343 38.1 C09-/13
53 0,3,6,10,1,5,9 334344 54.1b C09-/13+
54 0,3,6,10,2,5,8 334433 47.1b C013
55 0,3,6,10,2,5,9 334434 54.1b C013+
56 0,3,7,10,1,4,8 343334 54.1b Cm9-/11-/13
57 0,3,7,10,1,5,8 343343 38.1b Cm9-/13
23.2 Third Chain Classes 135

Third Chains—Continued
Chain Nr. Pitch Classes Third Chord Lead-Sheet
∼ equiv. from 0 Chain Class Symbols
58 0,3,7,10,1,5,9 343344 47.1b Cm9-/13+
59 0,3,7,10,2,5,8 343433 38.1b Cm13
60 0,3,7,10,2,5,9 343434 38.1b Cm13+
61 0,3,7,10,2,6,9 343443 54.1b Cm11+/13+
62 0,3,7,11,2,5,8 344333 54.1b Cm7+/13
63 0,3,7,11,2,5,9 344334 47.1b Cm7+/13+
64 0,3,7,11,2,6,9 344343 54.1b Cm7+/11+/13+
65 0,3,7,11,2,6,10 344344 60b Cm7+/11+/(13)/15-
66 0,4,7,10,1,5,8 433343 54.1b C9-/13
67 0,4,7,10,1,5,9 433344 54.1b C9-/13+
68 0,4,7,10,2,5,8 433433 47.1b C13
69 0,4,7,10,2,5,9 433434 38.1b C13+
70 0,4,7,10,2,6,9 433443 47.1b C11+/13+
71 0,4,7,11,2,5,8 434333 54.1b C7+/13
72 0,4,7,11,2,5,9 434334 38.1b C7+/13+
73 0,4,7,11,2,6,9 434343 38.1b C7+/11+/13+
74 0,4,7,11,2,6,10 434344 45.1b C7+/11+/(13)/15-
75 0,4,7,11,3,6,9 434433 54.1b C7+/9+/11+/13+
76 0,4,7,11,3,6,10 434434 55b C7+/9+/11+/(13)/15-
77 0,4,8,11,2,5,9 443334 54.1b C+7+/13+
78 0,4,8,11,2,6,9 443343 47.1b C+7+/11+/13+
79 0,4,8,11,2,6,10 443344 62b C+7+/11+/(13)/15-
80 0,4,8,11,3,6,9 443433 54.1b C+7+/9+/11+/13+
81 0,4,8,11,3,6,10 443434 45.1b C+7+/9+/11+/(13)/15-
82 0,4,8,11,3,7,10 443443 60b C+7+/9+/(11)/13-/15-
Eight Pitch Classes
83 0,3,6,9,1,4,7,10 3334333 37b C07-/9-/11-/13- . . .
84 0,3,6,9,1,4,7,11 3334334 26.1b C07-/9-/11-/13- . . .
85 0,3,6,9,1,4,8,11 3334343 22.1b C07-/9-/11-/13 . . .
86 0,3,6,9,1,5,8,11 3334433 26.1b C07-/9-/13 . . .
87 0,3,6,10,1,4,7,11 3343334 29b C09-/11-/13- . . .
88 0,3,6,10,1,4,8,11 3343343 18.1b C09-/11-/13 . . .
89 0,3,6,10,1,5,8,11 3343433 17.1b C09-/13 . . .
90 0,3,6,10,2,5,8,11 3344333 26.1b C013 . . .
91 0,3,6,10,2,5,9,1 3344344 31b C013+ . . .
136 23 Classification Lists

Third Chains—Continued
Chain Nr. Pitch Classes Third Chord Lead-Sheet
∼ equiv. from 0 Chain Class Symbols
92 0,3,7,10,1,4,8,11 3433343 31b Cm9-/11-/13 . . .
93 0,3,7,10,1,5,8,11 3433433 18.1b Cm9-/13 . . .
94 0,3,7,10,2,5,8,11 3434333 22.1b Cm13 . . .
95 0,3,7,10,2,5,9,1 3434344 18.1b Cm13+ . . .
96 0,3,7,10,2,6,9,1 3434434 29b Cm11+/13+ . . .
97 0,3,7,11,2,5,9,1 3443344 34b Cm7+/13+ . . .
98 0,3,7,11,2,6,9,1 3443434 29b Cm7+/11+/13+ . . .
99 0,3,7,11,2,6,10,1 3443443 28b Cm7+/11+/(13)/15- . . .
100 0,4,7,10,1,5,8,11 4333433 29b C9-/13 . . .
101 0,4,7,10,2,5,8,11 4334333 18.1b C15
102 0,4,7,10,2,5,9,1 4334344 22.1b C13+ . . .
103 0,4,7,10,2,6,9,1 4334434 26.1b C11+/13+ . . .
104 0,4,7,11,2,5,9,1 4343344 18.1b C7+/13+ . . .
105 0,4,7,11,2,6,9,1 4343434 17.1b C7+/11+/13+ . . .
106 0,4,7,11,2,6,10,1 4343443 25.1b C7+/11+/(13)/15- . . .
107 ∼ 84 0,4,7,11,3,6,9,1 4344334 26.1b C7+/9+/11+/13+ . . .
108 ∼ 87 0,4,7,11,3,6,10,1 4344343 29b C7+/9+/11+/(13)/15- . . .
109 0,4,7,11,3,6,10,2 4344344 30b C7+/9+/11+/(13)/15- . . .
110 0,4,8,11,2,5,9,1 4433344 31b C+7+/13+ . . .
111 0,4,8,11,2,6,9,1 4433434 18.1b C+7+/11+/13+ . . .
112 0,4,8,11,2,6,10,1 4433443 34b C+7+/11+/(13)/15- . . .
113 ∼ 85 0,4,8,11,3,6,9,1 4434334 22.1b C+7+/9+/11+/13+ . . .
114 ∼ 88 0,4,8,11,3,6,10,1 4434343 18.1b C+7+/9+/11+/(13)/15- . . .
115 0,4,8,11,3,6,10,2 4434344 35b C+7+/9+/11+/(13)/15- . . .
116 ∼ 92 0,4,8,11,3,7,10,1 4434433 31b C+7+/9+/(11)/13-/15- . . .
117 0,4,8,11,3,7,10,2 4434434 30b C+7+/9+/(11)/13-/15- . . .
Nine Pitch Classes
118 0,3,6,9,1,4,7,10,2 33343334 15b C07-/9-/11-/13- . . .
119 0,3,6,9,1,4,7,11,2 33343343 9.1b C07-/9-/11-/13- . . .
120 0,3,6,9,1,4,8,11,2 33343433 9.1b C07-/9-/11-/13 . . .
121 0,3,6,9,1,5,8,11,2 33344333 15b C07-/9-/13 . . .
122 0,3,6,10,1,4,7,11,2 33433343 10b C09-/11-/13- . . .
123 0,3,6,10,1,4,8,11,2 33433433 13b C09-/11-/13 . . .
124 0,3,6,10,1,5,8,11,2 33434333 9.1b C09-/13 . . .
125 0,3,6,10,2,5,9,1,4 33443443 10b C013+ . . .
23.2 Third Chain Classes 137

Third Chains—Continued
Chain Nr. Pitch Classes Third Chord Lead-Sheet
∼ equiv. from 0 Chain Class Symbols
126 0,3,7,10,1,4,8,11,2 34333433 10b Cm9-/11-/13 . . .
127 0,3,7,10,1,5,8,11,2 34334333 9.1b Cm9-/13 . . .
128 0,3,7,10,2,5,9,1,4 34343443 9.1b Cm13+ . . .
129 0,3,7,10,2,6,9,1,4 34344343 15b Cm11+/13+ . . .
130 0,3,7,11,2,5,9,1,4 34433443 13b Cm7+/13+ . . .
131 ∼ 119 0,3,7,11,2,6,9,1,4 34434343 9.1b Cm7+/11+/13+ . . .
132 0,3,7,11,2,6,9,1,5 34434344 14b Cm7+/11+/13+ . . .
133 ∼ 122 0,3,7,11,2,6,10,1,4 34434433 10b Cm7+/11+/(13)/15- . . .
134 0,3,7,11,2,6,10,1,5 34434434 11b Cm7+/11+/(13)/15- . . .
135 0,4,7,10,1,5,8,11,2 43334333 15b C9-/13 . . .
136 0,4,7,10,1,5,8,11,3 43334334 10.1b C9-/13 . . .
137 0,4,7,10,2,5,8,11,3 43343334 10.1b C17
138 0,4,7,10,2,6,9,1,5 43344344 10.1b C11+/13+ . . .
139 0,4,7,11,2,6,9,1,5 43434344 8.1b C7+/11+/13+ . . .
140 0,4,7,11,2,6,10,1,5 43434434 12b C7+/11+/(13)/15- . . .
141 0,4,7,11,3,6,9,1,5 43443344 14b C7+/9+/11+/13+ . . .
142 0,4,7,11,3,6,10,1,5 43443434 12b C7+/9+/11+/(13)/15- . . .
143 0,4,7,11,3,6,10,2,5 43443443 11b C7+/9+/11+/(13)/15- . . .
144 0,4,8,11,2,6,9,1,5 44334344 10.1b C+7+/11+/13+ . . .
145 0,4,8,11,2,6,10,1,5 44334434 14b C+7+/11+/(13)/15- . . .
146 0,4,8,11,3,6,9,1,5 44343344 10.1b C+7+/9+/11+/13+ . . .
147 0,4,8,11,3,6,10,1,5 44343434 8.1b C+7+/9+/11+/(13)/15- . . .
148 0,4,8,11,3,6,10,2,5 44343443 14b C+7+/9+/11+/(13)/15- . . .
149 ∼ 136 0,4,8,11,3,7,10,1,5 44344334 10.1b C+7+/9+/(11)/13-/15- . . .
150 ∼ 137 0,4,8,11,3,7,10,2,5 44344343 10.1b C+7+/9+/(11)/13-/15- . . .
151 0,4,8,11,3,7,10,2,6 44344344 16b C+7+/9+/(11)/13-/15- . . .
Ten Pitch Classes
152 0,3,6,9,1,4,7,10,2,5 333433343 5b C07-/9-/11-/13- . . .
153 0,3,6,9,1,4,7,11,2,5 333433433 4b C07-/9-/11-/13- . . .
154 0,3,6,9,1,4,8,11,2,5 333434333 5b C07-/9-/11-/13 . . .
155 0,3,6,10,1,4,7,11,2,5 334333433 3b C09-/11-/13- . . .
156 0,3,6,10,1,4,8,11,2,5 334334333 4b C09-/11-/13 . . .
157 ∼ 152 0,3,6,10,2,5,9,1,4,7 334434433 5b C013+ . . .
158 0,3,6,10,2,5,9,1,4,8 334434434 6b C013+ . . .
159 0,3,7,10,1,4,8,11,2,5 343334333 5b Cm9-/11-/13 . . .
138 23 Classification Lists

Third Chains—Continued
Chain Nr. Pitch Classes Third Chord Lead-Sheet
∼ equiv. from 0 Chain Class Symbols
160 0,3,7,10,1,4,8,11,2,6 343334334 6b Cm9-/11-/13 . . .
161 0,3,7,10,1,5,8,11,2,6 343343334 3.1b Cm9-/13 . . .
162 0,3,7,10,2,5,9,1,4,8 343434434 3.1b Cm13+ . . .
163 0,3,7,10,2,6,9,1,4,8 343443434 7b Cm11+/13+ . . .
164 0,3,7,11,2,5,9,1,4,8 344334434 6b Cm7+/13+ . . .
165 0,3,7,11,2,6,9,1,4,8 344343434 3.1b Cm7+/11+/13+ . . .
166 0,3,7,11,2,6,9,1,5,8 344343443 7b Cm7+/11+/13+ . . .
167 ∼ 160 0,3,7,11,2,6,10,1,4,8 344344334 6b Cm7+/11+/(13)/15- . . .
168 ∼ 161 0,3,7,11,2,6,10,1,5,8 344344343 3.1b Cm7+/11+/(13)/15- . . .
169 0,3,7,11,2,6,10,1,5,9 344344344 6b Cm7+/11+/(13)/15- . . .
170 0,4,7,10,1,5,8,11,2,6 433343334 7b C9-/13 . . .
171 0,4,7,10,1,5,8,11,3,6 433343343 3.1b C9-/13 . . .
172 0,4,7,10,2,5,8,11,3,6 433433343 6b C19
173 0,4,7,10,2,6,9,1,5,8 433443443 6b C11+/13+ . . .
174 0,4,7,11,2,6,9,1,5,8 434343443 3.1b C7+/11+/13+ . . .
175 ∼ 170 0,4,7,11,2,6,10,1,5,8 434344343 7b C7+/11+/(13)/15- . . .
176 0,4,7,11,2,6,10,1,5,9 434344344 3.1b C7+/11+/(13)/15- . . .
177 ∼ 171 0,4,7,11,3,6,9,1,5,8 434433443 3.1b C7+/9+/11+/13+ . . .
178 0,4,7,11,3,6,10,1,5,8 434434343 6b C7+/9+/11+/(13)/15- . . .
179 0,4,7,11,3,6,10,1,5,9 434434344 7b C7+/9+/11+/(13)/15- . . .
180 ∼ 172 0,4,7,11,3,6,10,2,5,8 434434433 6b C7+/9+/11+/(13)/15- . . .
181 0,4,7,11,3,6,10,2,5,9 434434434 3.1b C7+/9+/11+/(13)/15- . . .
182 0,4,8,11,2,6,10,1,5,9 443344344 6b C+7+/11+/(13)/15- . . .
183 0,4,8,11,3,6,10,1,5,9 443434344 3.1b C+7+/9+/11+/(13)/15- . . .
184 0,4,8,11,3,6,10,2,5,9 443434434 7b C+7+/9+/(11)/13-/15- . . .
185 0,4,8,11,3,7,10,1,5,9 443443344 6b C+7+/9+/(11)/13-/15- . . .
186 0,4,8,11,3,7,10,2,5,9 443443434 3.1b C+7+/9+/(11)/13-/15- . . .
187 0,4,8,11,3,7,10,2,6,9 443443443 6b C+7+/9+/(11)/13-/15- . . .
Eleven Pitch Classes
188 0,3,6,9,1,4,7,10,2,5,8 3334333433 2b C07-/9-/11-/13- . . .
189 0,3,6,9,1,4,7,11,2,5,8 3334334333 2b C07-/9-/11-/13- . . .
190 0,3,6,10,1,4,7,11,2,5,8 3343334333 2b C09-/11-/13- . . .
191 0,3,6,10,1,4,7,11,2,5,9 3343334334 2b C09-/11-/13- . . .
192 0,3,6,10,1,4,8,11,2,5,9 3343343334 2b C09-/11-/13 . . .
193 ∼ 191 0,3,6,10,2,5,9,1,4,7,11 3344344334 2b C013+ . . .
2
23.3 List of Local Denotators of Cardinality Two and Three in Z12 139

Third Chains—Continued
Chain Nr. Pitch Classes Third Chord Lead-Sheet
∼ equiv. from 0 Chain Class Symbols
194 ∼ 192 0,3,6,10,2,5,9,1,4,8,11 3344344343 2b C013+ . . .
195 0,3,7,10,1,4,8,11,2,5,9 3433343334 2b Cm9-/11-/13 . . .
196 0,3,7,10,1,4,8,11,2,6,9 3433343343 2b Cm9-/11-/13 . . .
197 0,3,7,10,1,5,8,11,2,6,9 3433433343 2b Cm9-/13 . . .
198 ∼ 195 0,3,7,10,2,5,9,1,4,8,11 3434344343 2b Cm13+ . . .
199 0,4,7,10,1,5,8,11,2,6,9 4333433343 2b C9-/13 . . .
200 0,4,7,10,1,5,8,11,3,6,9 4333433433 2b C9-/13 . . .
201 0,4,7,10,2,5,8,11,3,6,9 4334333433 2b C21
202 ∼ 199 0,4,7,10,2,6,9,1,5,8,11 4334434433 2b C11+/13+ . . .
203 ∼ 191 0,4,7,11,3,6,10,2,5,9,1 4344344344 2b C7+/9+/11+/(13)/15- . . .
204 ∼ 192 0,4,8,11,3,6,10,2,5,9,1 4434344344 2b C+7+/9+/(11)/13-/15- . . .
205 ∼ 195 0,4,8,11,3,7,10,2,5,9,1 4434434344 2b C+7+/9+/(11)/13-/15- . . .
206 ∼ 196 0,4,8,11,3,7,10,2,6,9,1 4434434434 2b C+7+/9+/(11)/13-/15- . . .
Twelve Pitch Classes
207 0,3,6,9,1,4,7,10,2,5,8,11 33343334333 1b C07-/9-/11-/13- . . .
208 ∼ 207 0,4,7,10,2,5,8,11,3,6,9,1 43343334334 1b C23
209 ∼ 207 0,4,7,10,2,6,9,1,5,8,11,3 43344344334 1b C11+/13+ . . .
210 ∼ 207 0,4,8,11,3,7,10,2,6,9,1,5 44344344344 1b C+7+/9+/(11)/13-/15- . . .

23.3 List of Local Denotators of Cardinality Two and

Three in Z212

23.3.1 Two Tone Motifs in Z212

ClassN r. Representative
1 (0, 0), (0, 1)
2 (0, 0), (0, 2)
3 (0, 0), (0, 3)
4 (0, 0), (0, 4)
5 (0, 0), (0, 6)

23.3.2 Two Tone Motifs in Z5 × Z12

140 23 Classification Lists

ClassN r. Representative
1 (0, 0), (0, 1)
2 (0, 0), (0, 2)
3 (0, 0), (0, 3)
4 (0, 0), (0, 4)
5 (0, 0), (0, 6)
6 (0, 0), (1, 0)
7 (0, 0), (1, 1)
8 (0, 0), (1, 2)
9 (0, 0), (1, 2)
10 (0, 0), (1, 4)
11 (0, 0), (1, 6)
2
23.3 List of Local Denotators of Cardinality Two and Three in Z12 141

23.3.3 Three Tone Motifs in Z212

The order of these representatives is a historical one. After this table, the
representatives are also visualized on a 12 × 12 square in list 23.1.

Three-Element Motif Classes in OnP iM od12,12

Class Nr. Representative Kernel Class Weight Volume
1 (0, 0), (1, 0), (2, 0) Z.(1, 2) × Z.(1, 1) (1, 1, 2) 0
2 (0, 0), (1, 0), (3, 0) Z.(1, 2) × Z.(0, 1) (1, 2, 3) 0
3 (0, 0), (1, 0), (4, 0) Z.(1, 0) × Z.(0, 1) (1, 3, 4) 0
4 (0, 0), (1, 0), (5, 0) Z.(1, 2) × Z.(1, 1) (1, 1, 4) 0
5 (0, 0), (1, 0), (6, 0) Z.(1, 2) × Z.(1, 0) (1, 1, 6) 0
6 (0, 0), (2, 0), (4, 0) (Z4 × 2Z4 ) × Z.(1, 1) (2, 2, 4) 0
7 (0, 0), (2, 0), (6, 0) (Z4 × 2Z4 ) × Z.(0, 1) (2, 4, 6) 0
8 (0, 0), (3, 0), (6, 0) Z.(1, 2) × Z23 (3, 3, 6) 0
9 (0, 0), (4, 0), (8, 0) (Z24 ) × Z.(1, 1) (4, 4, 4) 0
10 (0, 0), (1, 0), (0, 1) 0 × 0 (1, 1, 1) 1
11 (0, 0), (2, 0), (0, 1) Z.(2, 0) × 0 (1, 1, 2) 2
12 (0, 0), (3, 0), (0, 1) 0 × Z.(1, 0) (1, 1, 3) 3
13 (0, 1), (0, 2), (3, 0) 0 × Z.(1, 1) (1, 1, 1) 3
14 (0, 0), (0, 1), (4, 0) Z.(1, 0) × 0 (1, 1, 4) 4
15 (0, 0), (1, 2), (2, 0) Z.(1, 2) × 0 (1, 1, 2) 4
16 (0, 0), (2, 0), (0, 2) 2Z24 ×0 (2, 2, 2) 4
17 (0, 0), (6, 0), (0, 1) Z.(2, 0) × Z.(1, 0) (1, 1, 6) 6
18 (0, 0), (3, 0), (0, 2) Z.(2, 0) × Z.(0, 1) (1, 2, 3) 6
19 (0, 0), (0, 2), (3, 1) Z.(2, 0) × Z.(1, 1) (1, 1, 2) 6
20 (0, 0), (4, 0), (0, 2) (Z4 × 2Z4 ) × 0 (2, 2, 4) 4
21 (0, 0), (4, 0), (0, 4) 0 × Z23 (3, 3, 3) 3
22 (0, 0), (6, 0), (0, 2) 2Z24 × Z.(1, 0) (2, 2, 6) 0
23 (0, 2), (0, 4), (6, 0) 2Z24 × Z.(1, 1) (2, 2, 2) 0
24 (0, 0), (4, 0), (0, 4) Z24 × 0 (4, 4, 4) 4
25 (0, 0), (6, 0), (0, 3) Z.(2, 0) × Z23 (3, 3, 6) 6
26 (0, 0), (6, 0), (0, 6) 2Z24 × Z23 (6, 6, 6) 0
142 23 Classification Lists

1 2 3 4

5 6 7 8

9 10 11 12

13 14 15 16

17 18 19 20 21

22 23 24 25 26

Fig. 23.1: Three Tone Motifs in Z212

23.4 List of Local Denotators of Cardinality Four in Z212

This list was calculated by Straub in [45]. The list’s numbering follows Straub’s
algorithm; * denotes classes which are not determined by volume and class
weight.
23.4 List of Local Denotators of Cardinality Four in Z212 143

10
11 12
13
21

19 18 17
14
15
25 3
16
2
20 5

24 23
22
1

6 26 8 7
9

Fig. 23.2: Hasse diagram of dominance/specialization among the 26 isomorphism

classes of motives in Z212 .

Four-Element Motif Classes

Class Nr. Representative Class Weight Volume
0 (0,0),(0,1),(0,2),(0,7) (1,1,5,5) 0
1 (0,0),(0,1),(0,2),(0,3) (1,1,2,2) 0
2 (0,0),(0,1),(0,2),(0,6) (1,4,5,7) 0
3 (0,0),(0,1),(0,2),(0,5) (1,4,2,3) 0
4 (0,0),(0,1),(0,2),(0,4) (1,2,3,6) 0
5 (0,0),(1,0),(0,5),(0,6) (4,4,5,5) 0
6 (0,0),(0,1),(0,4),(0,5) (4,4,3,3) 0
7 (0,0),(0,1),(0,3),(0,5) (4,2,2,6) 0
8 (0,0),(0,1),(0,4),(0,8) (4,3,3,9) 0
9 (0,0),(0,1),(0,6),(0,7) (5,5,5,5) 0
10 (0,0),(0,1),(0,3),(0,6) (5,2,2,8) 0
144 23 Classification Lists

Four-Element Motif Classes—Continued

Class Nr. Representative Class Weight Volume
11 (0,0),(0,1),(0,3),(0,7) (5,2,3,7) 0
12 (0,0),(0,1),(0,4),(0,7) (5,3,3,8) 0
13 (0,0),(0,1),(0,3),(0,10) (2,2,2,2) 0
14 (0,0),(0,1),(0,3),(0,4) (2,2,3,3) 0
15 (0,0),(0,1),(0,3),(0,9) (2,3,7,8) 0
16 (0,0),(0,1),(0,4),(0,9) (3,3,3,3) 0
17 (0,0),(0,2),(6,0),(6,10) (23,23,22,22) 0
18 (0,0),(0,2),(0,4),(6,0) (23,6,22,7) 0
19 (0,0),(0,2),(0,4),(6,2) (6,6,22,22) 0
20 (0,0),(0,2),(0,4),(0,6) (6,6,7,7) 0
21 (0,0),(0,2),0,4),(0,8) (6,7,7,9) 0
22 (0,0),(0,2),(6,0),(6,2) (22,22,22,22) 0
23 (0,0),(0,2),(6,0),(6,6) (22,22,22,26) 0
24 (0,0),(0,2),(0,6),(6,2) (22,22,7,7) 0
25 (0,0),(0,2),(0,6),(6,0) (22,7,7,26) 0
26 (0,0),(0,2),(0,6),(0,8) (7,7,7,7) 0
27 (0,0),(0,3),(0,6),(0,9) (8,8,8,8) 0
28 (0,0),(0,6),(6,0),(6,6) (26,26,26,26) 0
29 (0,0),(0,1),(0,2),(1,0) (1,10,10,11) 1
30 (0,0),(0,1),(0,5),(1,0) (4,10,10,14) 1
31 (0,0),(0,1),(0,6),(1,0) (5,10,10,17) 1
32 (0,0),(0,1),(0,3),(1,0) (2,10,11,12) 1
33 (0,0),(0,1),(0,4),(1,0) (3,10,12,14) 1
34* (0,0),(0,1),(1,0),(1,5) (10,10,10,10) 1
35* (0,0),(0,1),(1,0),(7,7) (10,10,10,10) 1
36* (0,0),(0,1),(1,0),(1,1) (10,10,10,10) 1
37* (0,0),(0,1),(1,0),(3,5) (10,10,10,13) 1
38* (0,0),(0,1),(1,0),(3,11) (10,10,10,13) 1
39* (0,0),(0,1),(1,0),(1,2) (10,10,11,11) 1
40* (0,0),(0,1),(1,0),(5,10) (10,10,11,11) 1
41* (0,0),(0,1),(1,0),(4,10) (10,10,11,15) 1
42* (0,0),(0,1),(1,0),(2,4) (10,10,11,15) 1
43 (0,0),(0,1),(1,0),(2,5) (10,10,11,19) 1
44* (0,0),(0,1),(1,0),(1,3) (10,10,12,12) 1
23.4 List of Local Denotators of Cardinality Four in Z212 145

Four-Element Motif Classes—Continued

Class Nr. Representative Class Weight Volume
45* (0,0),(0,1),(1,0),(7,9) (10,10,12,12) 1
46* (0,0),(0,1),(1,0),(3,3) (10,10,12,12) 1
47 (0,0),(0,1),(1,0),(6,8) (10,10,15,19) 1
48* (0,0),(0,1),(1,0),(1,4) (10,10,14,14) 1
49* (0,0),(0,1),(1,0),(4,4) (10,10,14,14) 1
50 (0,0),(0,1),(1,0),(1,6) (10,10,17,17) 1
51 (0,0),(0,1),(1,0),(2,2) (10,11,11,13) 1
52 (0,0),(0,1),(1,0),(2,3) (10,11,13,15) 1
53 (0,0),(0,1),(1,0),(6,9) (10,11,12,18) 1
54 (0,0),(0,1),(1,0),(8,8) (10,13,14,14) 1
55 (0,0),(0,1),(1,0),(3,4) (10,12,15,18) 1
56 (0,0),(0,1),(2,0),(3,1) (11,11,12,12) 1
57 (0,0),(0,1),(2,0),(3,4) (11,12,12,15) 1
58 (0,0),(0,1),(3,0),(4,1) (12,12,14,14) 1
59 (0,0),(0,1),(0,2),(2,1) (1,11,11,15) 2
60 (0,0),(0,1),(0,2),(2,0) (1,11,11,16) 2
61 (0,0),(0,1),(0,5),(2,0) (4,11,11,14) 2
62 (0,0),(0,1),(0,5),(2,1) (4,11,11,20) 2
63 (0,0),(0,1),(0,6),(2,1) (5,5,11,11) 2
64 (0,0),(0,1),(0,6),(2,0) (5,22,11,11) 2
65 (0,0),(0,1),(0,3),(2,0) (2,11,15,18) 2
66 (0,0),(0,1),(0,3),(2,1) (2,11,16,18) 2
67 (0,0),(0,1),(0,4),(2,1) (3,11,14,18) 2
68 (0,0),(0,1),(0,4),(2,0) (3,11,20,18) 2
69 (0,0),(0,1),(2,0),(4,6) (23,11,11,15) 2
70 (0,0),(0,1),(2,0),(4,0) (6,11,11,14) 2
71 (0,0),(0,1),(2,0),(6,6) (22,11,15,17) 2
72 (0,0),(0,1),(2,0),(6,0) (7,11,14,17) 2
73* (0,0),(0,1),(2,0),(2,1) (11,11,11,11) 2
74* (0,0),(0,1),(2,0),(2,5) (11,11,11,11) 2
75* (0,0),(0,1),(2,0),(2,7) (11,11,11,11) 2
76* (0,0),(0,1),(2,0),(2,11) (11,11,11,11) 2
77* (0,0),(0,1),(2,0),(6,5) (11,11,11,19) 2
78* (0,0),(0,1),(2,0),(6,11) (11,11,11,19) 2
146 23 Classification Lists

Four-Element Motif Classes—Continued

Class Nr. Representative Class Weight Volume
79* (0,0),(0,1),(2,0),(4,7) (11,11,15,15) 2
80* (0,0),(0,1),(2,0),(8,11) (11,11,15,15) 2
81* (0,0),(0,1),(2,0),(2,2) (11,11,15,16) 2
82* (0,0),(0,1),(2,0),(8,10) (11,11,15,16) 2
83* (0,0),(0,1),(2,0),(4,1) (11,11,14,14) 2
84* (0,0),(0,1),(2,0),(8,5) (11,11,14,14) 2
85* (0,0),(0,1),(2,0),(8,4) (11,11,14,20) 2
86* (0,0),(0,1),(2,0),(2,4) (11,11,14,20) 2
87* (0,0),(0,1),(2,0),(6,7) (11,11,17,17) 2
88* (0,0),(0,1),(2,0),(6,1) (11,11,17,17) 2
89* (0,0),(0,1),(2,0),(2,9) (11,11,18,18) 2
90* (0,0),(0,1),(2,0),(2,3) (11,11,18,18) 2
91 (0,0),(0,1),(2,0),(4,3) (11,15,15,19) 2
92 (0,0),(0,1),(2,0),(4,2) (11,15,16,19) 2
93 (0,0),(0,1),(2,0),(4,9) (11,14,14,19) 2
94 (0,0),(0,1),(2,0),(4,8) (11,14,20,19) 2
95 (0,0),(0,1),(4,2),(6,1) (15,15,17,17) 2
96 (0,0),(0,1),(4,2),(6,4) (15,16,18,18) 2
97 (0,0),(0,1),(4,0),(6,1) (14,14,17,17) 2
98 (0,0),(0,1),(4,0),(6,4) (14,20,18,18) 2
99 (0,0),(0,1),(0,2),(3,0) (1,13,12,18) 3
100 (0,0),(0,1),(0,2),(3,1) (1,12,12,19) 3
101 (0,0),(0,1),(0,5),(3,0) (4,4,12,12) 3
102 (0,0),(0,1),(0,4),(3,2) (4,3,13,12) 3
103 (0,0),(0,1),(0,6),(3,2) (5,13,13,17) 3
104 (0,0),(0,1),(0,6),(3,1) (5,12,12,17) 3
105 (0,0),(0,1),(0,6),(3,0) (5,12,12,25) 3
106 (0,0),(0,1),(0,3),(3,2) (2,13,12,19) 3
107 (0,0),(0,1),(0,3),(3,1) (2,12,12,18) 3
108 (0,0),(0,1),(0,3),(3,0) (2,12,21,18) 3
109 (0,0),(0,1),(0,4),(3,0) (3,3,12,12) 3
110 (0,0),(0,1),(0,4),(3,1) (3,3,12,21) 3
111 (0,0),(0,1),(3,0),(6,0) (8,12,12,17) 3
112 (0,0),(0,3),(0,6),(3,0) (8,21,21,25) 3
23.4 List of Local Denotators of Cardinality Four in Z212 147

Four-Element Motif Classes—Continued

Class Nr. Representative Class Weight Volume
113* (0,0),(0,1),(3,0),(3,5) (13,13,12,12) 3
114* (0,0),(0,1),(3,0),(3,11) (13,13,12,12) 3
115* (0,0),(0,1),(3,0),(9,11) (13,13,12,12) 3
116 (0,0),(0,1),(3,2),(3,8) (13,13,17,17) 3
117 (0,0),(0,1),(3,0),(3,2) (13,12,19,18) 3
118* (0,0),(0,1),(3,0),(9,7) (12,12,12,12) 3
119* (0,0),(0,1),(3,0),(3,7) (12,12,12,12) 3
120* (0,0),(0,1),(3,0),(3,1) (12,12,12,12) 3
121* (0,0),(0,1),(3,0),(9,3) (12,12,12,21) 3
122* (0,0),(0,1),(3,0),(3,3) (12,12,12,21) 3
123 (0,0),(0,1),(3,0),(6,5) (12,12,19,19) 3
124 (0,0),(0,1),(3,0),(6,1) (12,12,17,17) 3
125 (0,0),(0,1),(3,0),(3,6) (12,12,17,25) 3
126 (0,0),(0,1),(3,0),(3,10) (12,12,18,18) 3
127 (0,0),(0,1),(3,0),(6,9) (12,21,18,18) 3
128* (0,0),(0,3),(3,0),(3,3) (21,21,21,21) 3
129* (0,0),(0,3),(3,0),(9,9) (21,21,21,21) 3
130 (0,0),(0,3),(3,0),(3,6) (21,21,25,25) 3
131 (0,0),(0,1),(0,2),(4,3) (1,15,15,15) 4
132 (0,0),(0,1),(0,2),(4,1) (1,15,14,14) 4
133 (0,0),(0,1),(0,2),(4,0) (1,15,14,20) 4
134 (0,0),(0,1),(0,5),(4,2) (4,15,15,14) 4
135 (0,0),(0,1),(0,5),(4,3) (4,15,15,20) 4
136 (0,0),(0,1),(0,5),(4,0) (4,14,14,14) 4
137 (0,0),(0,1),(0,5),(4,1) (4,14,14,24) 4
138 (0,0),(0,1),(0,6),(4,3) (5,5,15,15) 4
139 (0,0),(0,1),(0,6),(4,1) (5,5,14,14) 4
140 (0,0),(0,1),(0,6),(4,0) (5,7,15,14) 4
141 (0,0),(0,1),(0,3),(4,2) (2,2,15,15) 4
142 (0,0),(0,1),(0,3),(4,1) (2,2,14,20) 4
143 (0,0),(0,1),(0,3),(4,0) (2,3,15,14) 4
144 (0,0),(0,1),(0,3),(4,3) (2,3,15,20) 4
145 (0,0),(0,1),(0,4),(4,1) (3,3,14,14) 4
146 (0,0),(0,1),(0,4),(4,0) (3,3,14,24) 4
148 23 Classification Lists

Four-Element Motif Classes—Continued

Class Nr. Representative Class Weight Volume
147 (0,0),(0,2),(2,0),(6,10) (23,16,16,16) 4
148 (0,0),(0,2),(2,0),(4,4) (23,16,20,20) 4
149 (0,0),(0,1),(4,0),(8,6) (6,15,15,14) 4
150 (0,0),(0,2),(0,4),(2,0) (6,16,16,20) 4
151 (0,0),(0,2),(0,4),(4,2) (6,20,20,20) 4
152 (0,0),(0,2),(0,4),(4,0) (6,20,20,24) 4
153 (0,0),(0,2),(2,0),(2,6) (22,22,16,16) 4
154 (0,0),(0,2),(4,0),(6,2) (22,22,20,20) 4
155 (0,0),(0,2),(0,6),(2,0) (22,7,16,20) 4
156 (0,0),(0,2),(0,6),(4,0) (7,7,20,20) 4
157 (0,0),(0,2),(0,6),(4,2) (7,7,20,24) 4
158 (0,0),(0,1),(4,0),(8,0) (9,14,14,14) 4
159 (0,0),(0,2),(4,0),(8,0) (9,20,20,20) 4
160 (0,0),(0,4),(0,8),(4,0) (9,24,24,24) 4
161* (0,0),(0,1),(4,2),(4,7) (15,15,15,15) 4
162* (0,0),(0,1),(4,2),(4,3) (15,15,15,15) 4
163* (0,0),(0,1),(4,0),(4,7) (15,15,14,14) 4
164* (0,0),(0,1),(4,0),(4,11) (15,15,14,14) 4
165* (0,0),(0,1),(4,0),(4,2) (15,15,14,20) 4
166* (0,0),(0,1),(4,0),(4,10) (15,15,14,20) 4
167* (0,0),(0,1),(4,0),(4,1) (14,14,14,14) 4
168* (0,0),(0,1),(4,0),(4,5) (14,14,14,14) 4
169 (0,0),(0,1),(4,0),(4,4) (14,14,14,24) 4
170 (0,0),(0,2),(2,0),(2,2) (16,16,16,16) 4
171* (0,0),(0,2),(2,0),(8,8) (16,16,20,20) 4
172* (0,0),(0,2),(2,0),(2,4) (16,16,20,20) 4
173* (0,0),(0,2),(4,0),(4,2) (20,20,20,20) 4
174* (0,0),(0,2),(4,0),(4,10) (20,20,20,20) 4
175 (0,0),(0,2),(4,0),(4,4) (20,20,20,24) 4
176 (0,0),(0,4),(4,0),(4,4) (24,24,24,24) 4
177 (0,0),(0,1),(0,2),(6,1) (1,1,17,17) 6
178 (0,0),(0,1),(0,2),(6,3) (1,2,19,18) 6
179 (0,0),(0,1),(0,2),(6,4) (1,23,18,18) 6
180 (0,0),(0,1),(0,2),(6,0) (1,22,19,17) 6
23.4 List of Local Denotators of Cardinality Four in Z212 149

Four-Element Motif Classes—Continued

Class Nr. Representative Class Weight Volume
181 (0,0),(0,1),(0,5),(6,0) (4,4,17,17) 6
182 (0,0),(0,1),(0,4),(6,5) (4,3,19,18) 6
183 (0,0),(0,1),(0,5),(6,3) (4,6,18,18) 6
184 (0,0),(0,1),(0,5),(6,1) (4,7,19,17) 6
185 (0,0),(0,1),(0,6),(6,5) (5,5,19,19) 6
186 (0,0),(0,1),(0,6),(6,1) (5,5,17,17) 6
187 (0,0),(0,1),(0,6),(6,2) (5,22,19,19) 6
188 (0,0),(0,1),(0,6),(6,4) (5,22,18,18) 6
189 (0,0),(0,1),(0,6),(6,3) (5,8,18,18) 6
190 (0,0),(0,1),(0,6),(6,0) (5,26,17,17) 6
191 (0,0),(0,1),(0,3),(6,0) (2,2,17,25) 6
192 (0,0),(0,1),(0,3),(6,4) (2,2,18,18) 6
193 (0,0),(0,1),(0,3),(6,5) (2,23,19,18) 6
194 (0,0),(0,1),(0,3),(6,1) (2,22,17,18) 6
195 (0,0),(0,1),(0,3),(6,3) (2,22,18,25) 6
196 (0,0),(0,1),(0,4),(6,1) (3,3,17,25) 6
197 (0,0),(0,1),(0,4),(6,3) (3,3,18,18) 6
198 (0,0),(0,1),(0,4),(6,2) (3,6,19,18) 6
199 (0,0),(0,1),(0,4),(6,0) (3,7,17,18) 6
200 (0,0),(0,1),(0,4),(6,4) (3,7,18,25) 6
201 (0,0),(0,1),(6,3),(6,9) (22,8,18,18) 6
202 (0,0),(0,3),(0,6),(6,3) (8,8,25,25) 6
203 (0,0),(0,3),(0,6),(6,0) (8,26,25,25) 6
204* (0,0),(0,1),(6,0),(6,5) (19,19,17,17) 6
205* (0,0),(0,1),(6,0),(6,11) (19,19,17,17) 6
206* (0,0),(0,1),(6,2),(6,3) (19,19,18,18) 6
207* (0,0),(0,1),(6,2),(6,5) (19,19,18,18) 6
208* (0,0),(0,1),(6,0),(6,1) (17,17,17,17) 6
209* (0,0),(0,1),(6,0),(6,7) (17,17,17,17) 6
210* (0,0),(0,1),(6,0),(6,3) (17,18,18,25) 6
211* (0,0),(0,1),(6,0),(6,9) (17,18,18,25) 6
212* (0,0),(0,1),(6,3),(6,4) (18,18,18,18) 6
213* (0,0),(0,1),(6,3),(6,10) (18,18,18,18) 6
214* (0,0),(0,3),(6,0),(6,3) (25,25,25,25) 6
150 23 Classification Lists

Four-Element Motif Classes—Continued

Class Nr. Representative Class Weight Volume
215* (0,0),(0,3),(6,0),(6,9) (25,25,25,25) 6

23.5 List of Modulation Chords (Pivots)

Quanta and Pivots for the Modulations Between Diatonic Major Scales
Transl. p Cadence Quantum Modulator Pivots
1 {II, V } • ◦ • • ◦ • • • • • •• e5 11 {II, III, V, V II}
1 {II, III} • ◦ • • ◦ • • • • • •• e5 11 {II, III, V, V II}
2 {V II} ◦ • • ◦ • • ◦ • ◦ ◦ ◦• e6 11 {II, IV, V II}
2 {II, V } 6
◦ • • ◦ • • ◦ • ◦ • ◦• e 11 {II, IV, V, V II}
2 {IV, V } ◦ • • ◦ • • ◦ • ◦ • ◦• e6 11 {II, IV, V, V II}
3 {II, V } • ◦ • ◦ ◦ • ◦ • • • •• e7 11 {II, III, V, V II}
3 7
{II, III} • ◦ • ◦ ◦ • ◦ • • • •• e 11 {II, III, V, V II}
4 {V II} 8
◦ ◦ • • ◦ • • ◦ ◦ • ◦• e 11 {II, IV, V, V II}
4 {IV, V } ◦ • • • • • • • ◦ • ◦• e 118
{II, III, V, V II}
4 8
{II, III} • • • • ◦ • • • • • ◦• e 11 {V, V II}
5 {V II} 9
◦ ◦ • ◦ • • ◦ • ◦ ◦ •• e 11 {II, IV, V II}
6 {II, III} ◦ • • • • • ◦ • • • •• e6 {II, III, V, V II}
6 {IV, V } ◦ • • • • • • • • • ◦• e 1110
{II, IV, V, V II}
6 {IV, V } • • • • ◦ • • • • • ◦• e6 {II, IV, V, V II}
6 10
{II, III} • • • • ◦ • ◦ • • • •• e 11 {II, III, V, V II}
7 {V II} • ◦ • ◦ ◦ • • ◦ ◦ • ◦• e11 11 {III, V, V II}
8 {V II} 0
◦ • • ◦ ◦ • ◦ • ◦ ◦ •• e 11 {II, V II}
8 {IV, V } ◦ • • • • • ◦ • • • •• e 110
{II, IV, V, V II}
8 0
{II, III} • • • • ◦ • ◦ • ◦ • •• e 11 {II, III, V, V II}
9 {II, V } 1
◦ ◦ • ◦ • • • • • • ◦• e 11 {II, IV, V, V II}
9 {IV, V } ◦ ◦ • ◦ • • • • • • ◦• e 111
{II, IV, V, V II}
10 {V II} • ◦ • • ◦ • ◦ ◦ ◦ • ◦• e2 11 {III, V, V II}
10 {II, V } 2
• ◦ • • ◦ • ◦ • ◦ • ◦• e 11 {II, III, V, V II}
10 2
{II, III} • ◦ • • ◦ • ◦ • ◦ • ◦• e 11 {II, III, V, V II}
23.5 List of Modulation Chords (Pivots) 151

Quanta and Pivots for Diatonic Major Scales—Continued

Class Transl. p Cadence Quantum Modulator Pivots

11 {II, V } ◦ • • ◦ • • • • • • •• e3 11 {II, IV, V, V II}

11 3
{IV, V } ◦ • • ◦ • • • • • • •• e 11 {II, IV, V, V II}
References

1. Adorno Th W: Fragment über Musik und Sprache. Stuttgart, Jahresring 1956

2. Lang S: Algebra. Springer 2002
3. Agustín-Aquino O A, J Junod, G Mazzola: Computational Counterpoint Worlds.
Springer Series Computational Music Science, Heidelberg 2015
4. Boulez P: Musikdenken heute I,II; Darmstädter Beiträge V, VI. Schott, Mainz
1963, 1985
5. Boulez P: structures, premier livre. UE, London 1953
6. Boulez P: structures, deuxiéme livre. UE, London 1967
7. Eimert H: Grundlagen der musikalischen Reihentechnik. Universal Edition,
Wien 1964
8. Funkhouser Ch: being matter ignited...an interview with Cecil Taylor. Hambone
1994
9. Fux J J: Gradus ad Parnassum (1725). Dt. und kommentiert von L. Mitzler,
Leipzig 1742; English edition: The Study of Counterpoint. Translated and edited
by A Mann. Norton & Company, New York, London 1971
https://www.thegreatcoursesdaily.com/kurt-godel-wanted-to-revise-
our-concept-of-time
10. Graeser, W: Bachs “Kunst der Fuge”. In: Bach-Jahrbuch 1924
11. Hanslick E: Vom Musikalisch-Schönen. Breitkopf und Härtel (1854), Wiesbaden
1980
12. Hecquet S and A Prokhoris: Fabriques de la danse. PUF, Paris 2007
13. Hjelmslev L: Prolégomènes á une théorie du langage. Minuit, Paris 1966
14. Jakobson R: Linguistics and Poetics. In: Seboek, TA (ed.): Style in Language.
Wiley, New York 1960
15. Jackendoff R and Lerdahl F: A Generative Theory of Tonal Music. MIT Press,
Cambridge MA, 1983
16. Johnson I and Henrich A K: Knot Theory. Aurora 2017
17. Kaiser J: Beethovens 32 Klaviersonaten und ihre Interpreten. Fischer, Frank-
furt/Main 1979
18. Kant I: Kritik der reinen Vernunft. Meiner, Hamburg 1956
19. Lewin D: Generalized Musical Intervals and Transformations (1987). Cambridge
University Press 1987
20. Ligeti G: Pierre Boulez: Entscheidung und Automatik in der Structure Ia. Die
Reihe IV, UE, Wien 1958

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 153
L. Zheng, G. Mazzola, Classification of Musical Objects for Analysis and Composition,
Computational Music Science, https://doi.org/10.1007/978-3-031-30183-4
154 References

21. Kelley J L: General Topology. Van Nostrand 1955

22. Mac Lane S: Categories for the Working Mathematician. Springer 1998
23. Mac Lane S and Moerdjik I: Sheaves in Geometry and Logic. Springer 1992
24. Marek C: Lehre des Klavierspiels. Atlantis 1977
25. Mazzola G: Gruppen und Kategorien in der Musik. Heldermann, Berlin 1985
26. Mazzola G: www.rubato.org
27. Mazzola G et al.: The Topos of Music—Geometric Logic of Concepts, Theory,
and Performance. Birkhäuser, Basel et al. 2002
28. Mazzola G et al.: The Topos of Music I: Theory. (First volume of second edition
of [27]) Springer, Heidelberg 2018
29. Mazzola G et al.: The Topos of Music II: Performance. (Second volume of second
edition of [27]) Springer, Heidelberg 2018
30. Mazzola G et al.: The Topos of Music III: Gestures. (Third volume of second
edition of [27]) Springer, Heidelberg 2018
31. Mazzola G et al.: The Topos of Music IV: Roots. (Fourth volume of second
edition of [27]) Springer, Heidelberg 2018
32. Mazzola G et. al.: Making Musical Time. Springer 2021
33. Mazzola G and M Andreatta: Diagrams, Gestures, and Formulas in Music. Jour-
nal of Mathematics and Music 2007, Vol. 1, no. 1, 2007
34. Mazzola G and P B Cherlin: Flow, Gesture, and Spaces in Free Jazz—Towards
a Theory of Collaboration. Springer Series Computational Music Science, Hei-
delberg 2009
35. Mazzola G: Categorical Gestures, the Diamond Conjecture, Lewin’s Question,
and the Hammerklavier Sonata. Journal of Mathematics and Music Vol. 3, no.
1, 2009
36. Mazzola G, J Park, F Thalmann: Musical Creativity. Springer, Heidelberg 2011
37. Mazzola G, M Mannone, Y Pang: Cool Math for Hot Music. Springer, Heidelberg
2016
38. Noll Th: http://www.cs.tu-berlin.de/ noll/ChordDictionary.sea.hqx, TU Berlin
1996
social and cultural context (pp. 33–56). Gainesville: University Presses of
Florida, 1981
39. Nattiez J-J: Fondements d’une Sémiologie de la Musique. Edition 10/18 Paris
1975
40. Riemann B (1867): Über die Hypothesen, welche der Geometrie zugrunde liegen
(Habilitationsreferat 1854). Gott. Abh. No.13, 1867
41. Riemann H: System der musikalischen Rhythmik und Metrik. Breitkopf und
Härtel, Leipzig 1903
42. Ruwet N: Langage, Musique, Poesie. Seuil, Paris 1972
43. Sachs K-J: Der Contrapunctus im 14. und 15. Jahrhundert. AMW, Franz
Steiner, Wiesbaden 1974
44. Schönberg A: Harmonielehre (1911). Universal Edition, Wien 1966
45. Straub H: Beiträge zur modultheoretischen Klassifikation musikalischer Motive.
Diplomarbeit ETH-Zürich, Zürich 1989
46. Tymoczko D: A Geometry of Music. Oxford University Press, 2011
47. Uhde J and Wieland R: Denken und Spielen. Bärenreiter, Kassel et al. 1988
48. Uhde J and R Wieland: Forschendes Üben. Bärenreiter 2002
49. Varèse E: Erinnerungen und Gedanken. In: Darmstädter Beiträge III. Schott,
Mainz 1960
References 155

50. Wilson R A: The Finite Simple Groups. Springer 2009

51. Zoon J: Density 21.5. Edgard Varèse: The Complete Works. CD. Decca London
Polygram 289 460 208–2, https://www.youtube.com/watch?v=s6-TBeDLlnA
1998
Index

C(F ), 16 agogics, 74
CT (D), 17 ambiguity, 77
Deno(Cat), 18 analysis, neutral -, 75
F (D), 17 articulated, listening, 74
GI , 80 articulation, 74
I(F ), 16 ASCII, 20
N (D), 17 aspect, psychological -, vii
N (F ), 15 atlas, 78
T (F ), 16 A-addressed -, 80
T ±, 28 Avison, Charles, 73
Cat@ , 15
Mod, 15 B
TopCat, 25 Bach, Johann Sebastian, 74, 90
∇, 25 Beethoven, Ludwig van, 30, 74, 90
Colimit, 18 Boulez, Pierre, 14, 22
Digraph, 15, 23
Limit, 18 C
Power, 18 cadence, 30
Ens, 15 Cage, John, 76
Simple, 17 canonical program, 90
STRG, 19 category
Syn, 17 topological -, 25
RUBATOr software, 34 graph -, 25
Colimit, 16 cell complex, 90
Limit, 16 cellular organism, 90
Power, 16 change of perspective, 89
Simple, 16 chart, 78, 80
Syn, 16 chord, 74
3M, 10 classification
creative value of -, 13
A definition of -, 3
activity, interpretative -, 77, 79 of gestures, 34
address change, 15, 23 scientific value of -, 13
Adorno, Theodor Wiesengrund, 73 semantics of -, 4

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 157
L. Zheng, G. Mazzola, Classification of Musical Objects for Analysis and Composition,
Computational Music Science, https://doi.org/10.1007/978-3-031-30183-4
158 Index

coda, 74 Feldman, Morton, 77

colimit, 78 Fifth symphony, 74
communication, 7 fingering, 73
complex, cell -, 90 form, 3, 15
composition circular -, 21
global -, 28 contrapuntal -, 74
resolution of a -, 87 coordinator, 16
global objective -, 79 functor, 16
local -, 27 identifier, 16
modular -, 77 sonata -, 74
concept space, 16
grouping -, 76 type, 16
score -, 77 coordinator of a -, 15
concert for piano and orchestra, 77 identifier of a -, 15
context, semiotic -, 10 name of a -, 15
contrapuntal form, 74 type of a -, 15
coordinator, form -, 16 Fourier decomposition, 21
covering, 78 frame space, 16
equivalent -, 80 frequency modulation, 22
creativity, 9 Fry, Arthur, 11
function
D poetical -, 73
dactylus grid, 50 tonal -, 74
degree, 29, 74 functor, form -, 16
pivotal -, 30
denotator, 3, 16 G
development, 74 general note, 21
Diabelli Variations, 90 germinal melody, 51
diffeomorphism, 34 gestalt, global -, 77
digraph, 23 gesture, 4, 8, 23, 33
dodecaphonic global -, 28, 79, 80
series, 90
global
row, 38
gestalt, 77
DX7, 22
objective composition, 79
dynamics, 74
score, 77
E Goethe, Johann Wolfgang von, 90
Einstein, Albert, 12 Goldberg Variations, 90
embodiment, 7, 8 Graeser, Wolfgang, 74
equivalence relation, 75 grid, dactylus -, 50
equivalent covering, 80 group, finite simple -, 4
Escher grouping
category, 38 concept, 76
Theorem, 37 metrical -, 73
aesthesic identification, 74
aesthetics of music, 87 H
exposition, 74 Hammerklavier Sonata, 30
Hanslick, Eduard, 74, 77
F hierarchical organism, 74
fact, 8 hierarchy, 76
Index 159

HiHat, 20 mode, 28
historical instrumentation, 89 modular composition, 77
Hofmann, Ernst Theodor Amadeus, 73 modulation
homology theories, 38 model, 29
hyperdenotator, 24, 37 quantized -, 30
hypergesture, 37 quantum, 30
tonal -, 28
I modulator, 30
identification, esthesic -, 74 music
identifier, form -, 16 aesthetics of -, 87
instrumentation, historical -, 89 characterization of -, 9
Intégrales, 89 critic, 74
integral of perspectives, 90 criticism, 74
interpretative activity, 77, 79 ontology, 7
isometry, 76 psychology, 76
isomorphism class, 4 software, 77
tape -, 76
J musique concrète, 77
Jackendoff, Ray, 76
Jakobson, Roman, 75 N
Java, 34 naming policy, 17
jazz improvisation, 38
Nattiez, Jean-Jacques, 74
neutral analysis, 75
K
note, 20
Kagel, Maurizio, 90
Kaiser, Joachim, 73, 74
O
knot, 4, 34
object, 3
Kunst der Fuge, 74
objective global composition, 79
L oniontology, 7, 9
Lanier, Sidney, 12 onset, 20
Lerdahl, Fred, 76 ontology, music -, 7
limits, ontological -, 3 organism
linguistics, structuralist -, 75 cellular, 90
listening, articulated -, 74 hierarchical -, 74
local score, 77
local-global patchwork, 77 P
locally trivial structure, 78 part, 77
loop, 34 patchwork, local-global -, 77
loudness, 19 pause, 20
Luening, Otto, 76 performance
structural
M rationale of -, 90
Möbius strip, 29, 78 performance theory, 34
macroscore, 21 perspective, change of -, 89
manifold, 78 perspectives, integral of -, 90
Marek, Ceslav, 73 pitch, 20
Mattheson, Johann, 73 poetical function, 73
melody, germinal -, 51 Post-It, 10
metrical grouping, 73 presheaf, 34
160 Index

principle, variation -, 90 Hammerklavier -, 30

process, 8 space, form -, 16
program, canonical -, 90 Straub, Hans, 142
progression structural rationale of performance, 90
harmonic -, 74 structuralist linguistics, 75
contrapuntal -, 74 structure, locally trivial -, 78
Project of Music for Magnetic Tape, 77 Structures pour deux pianos, 14, 22, 38
psychology, music -, 76 support set, 80
symmetry transformation, 75, 76
Q
question, open -, 10 T
tönend bewegte Formen, 77
R
tape music, 76
realities, 7
theory, set -, 75
recapitulation, 74
time-slice, 77
relation, equivalence -, 75
tonal function, 74
representation, formal -, 13
tonality, 28, 74
resolution of a global composition, 87
track, 77
Riemann, Bernhard, 78
transformation, symmetry -, 75, 76
Riemann, Hugo, 29, 77
Transición II, 90
Rousseau, Jean-Jacques, 73
Tudor, David, 77
Rubato software, 14
tuning, 74
Ruwet, Nicolas, 74
type, form -, 16
S
Saussure, Ferdinand de, 75 U
scale, 28 understanding, 90
major -, 28 musical works, 88
Schönberg, Arnold, 29, 90 Unicode, 20
Schaeffer, Pierre, 76 Ussachevsky, Vladimir, 76
Schenker analysis, 21
score V
concept, 77 Varèse, Edgar, 87, 89
global -, 77 variation principle, 90
local -, 77 Variationen für Klavier, 90
semiotics, 7 voice leading, 74
serial composition, 22 von Ehrenfels, Christian, 78
series, dodecaphonic -, 90
set W
support -, 80 walls, 10
theory, 75 extended -, 10
sign, critical -, 10 Webern, Anton von, 90
Silver, Spencer, 10 Wolff, Christian, 77
software, 34
music -, 77 Y
sonata Yoneda Lemma, 19, 34, 89
form, 74 Yoneda, Nobuo, 87